UNITED STATES REPORTS VOLUME 491 CASES ADJUDGED IN THE SUPREME COURT AT OCTOBER TERM, 1988 June 15 Through June 22, 1989 FRANK D. WAGNER REPORTER OF DECISIONS WASHINGTON : 1993 Printed on Uncoated Permanent Printing Paper For sale by the U.S. Government Printing Office Superintendent of Documents, Mail Stop: SSOP, Washington, DC 20402-9328 ISBN 0-16-041785-6 Errata 444 U. S. xlvii, line 6 from bottom: “87” should be “871”. 464 U. S. 92, note 4, line 4: “7102” should be “7103”. JUSTICES OF THE SUPREME COURT DURING THE TIME OF THESE REPORTS WILLIAM H. REHNQUIST, Chief Justice. WILLIAM J. BRENNAN, Jr., Associate Justice. BYRON R. WHITE, Associate Justice. THURGOOD MARSHALL, Associate Justice. HARRY A. BLACKMUN, Associate Justice. JOHN PAUL STEVENS, Associate Justice. SANDRA DAY O’CONNOR, Associate Justice. ANTONIN SCALIA, Associate Justice. ANTHONY M. KENNEDY, Associate Justice. RETIRED WARREN E. BURGER, Chief Justice. LEWIS F. POWELL, Jr., Associate Justice. OFFICERS OF THE COURT RICHARD L. THORNBURGH, Attorney General. KENNETH W. STARR, Solicitor General. JOSEPH F. SPANIOL, Jr., Clerk. FRANK D. WAGNER, Reporter of Decisions. ALFRED WONG, Marshal. SHELLEY L. DOWLING, Librarian. in SUPREME COURT OF THE UNITED STATES Allotment of Justices It is ordered that the following allotment be made of the Chief Justice and Associate Justices of this Court among the circuits, pursuant to Title 28, United States Code, Section 42, and that such allotment be entered of record, effective February 18, 1988, viz.: For the District of Columbia Circuit, William H. Rehnquist, Chief Justice. For the First Circuit, William J. Brennan, Jr., Associate Justice. For the Second Circuit, Thurgood Marshall, Associate Justice. For the Third Circuit, William J. Brennan, Jr., Associate Justice. For the Fourth Circuit, William H. Rehnquist, Chief Justice. For the Fifth Circuit, Byron R. White, Associate Justice. For the Sixth Circuit, Antonin Scalia, Associate Justice. For the Seventh Circuit, John Paul Stevens, Associate Justice. For the Eighth Circuit, Harry A. Blackmun, Associate Justice. For the Ninth Circuit, Sandra Day O’Connor, Associate Justice. For the Tenth Circuit, Byron R. White, Associate Justice. For the Eleventh Circuit, Anthony M. Kennedy, Associate Justice. For the Federal Circuit, William H. Rehnquist, Chief Justice. February 18, 1988. (For next previous allotment, and modifications, see 479 U. S., p. v, 483 U. S., pp. v, vi, and 484 U. S., pp. v, vi.) IV TABLE OF CASES REPORTED Note: All undesignated references herein to the United States Code are to the 1982 edition. Cases reported before page 901 are those decided with opinions of the Court or decisions per curiam. Cases reported on page 901 et seq. are those in which orders were entered. Page Adamo v. Hotel Workers........................................ 911 Alabama; Thomas v. ........................................... 910 Alaska v. Kenaitze Indian Tribe............................... 905 Al-Hakim v. Tampa City Council................................ 908 Ameri-Manage, Inc.; Skoblow v. ............................... 904 Anniston; Brooks v............................................ 908 Arizona; Beaty v.............................................. 910 Arkansas v. Gibson............................................ 910 Armontrout; Gilmore v. ....................................... 911 Asberry v. U. S. Postal Service............................... 911 Atlanta; Meade-Murphy v....................................... 907 Attorney General; Michigan Citizens for Independent Press v. .... 903 Attorney General; Randall v. ................................. 904 Attorney Registration and Disciplinary Comm’n of Ill.; Karzov v... 906 Attorneys’ Title Guaranty Fund, Inc. v. Richards.............. 906 Ayala v. United States........................................ 907 Baker; Northeast Ill. Regional Commuter Railroad Corp. v. .... 901 Barbour; Halverson v.......................................... 908 Barron v. Salt Lake City...................................... 911 Barrow v. Wahl................................................ 907 Beaty v. Arizona.............................................. 910 Beer Institute; Healy v. ...... ¿ k. 324 Beer Institute; Wine & Spirits Wholesalers of Conn., Inc. v.. 324 B. F. Goodrich Co.; Worrell v................................ 907 Bishop v. Gwaltney............................................ 904 B. J. F.; Florida Star v..................................... 524 Black; Edwards v.............................................. 913 Board of Ed. of Millville, N. J.; S. T. v..................... 901 Bobrow, Greenapple & Skolnik; Woods v........................ 905 Bonjorno v. Kaiser Aluminum & Chemical Corp................... 903 VI TABLE OF CASES REPORTED Page Bonjorno; Kaiser Aluminum & Chemical Corp. v................... 903 Brathwaite v. United States.................................... 911 Brooks v. Anniston............................................. 908 Bruscino v. Carlson............................................ 907 California; Carella v.......................................... 263 California; Hernandez v........................................ 910 Campos-Fuentes v. United States................................ 909 Caplin & Drysdale, Chartered v. United States.................. 617 Cardenas v. United States...................................... 909 Carella v. California.......................................... 263 Carlson; Bruscino v............................................ 907 Cartwright, In re ............................................. 903 Caruthers v. Tennessee......................................... 910 Catawba Indian Tribe of S. C., Inc.; South Carolina v.......... 906 Central La. Telephone Co. v. McCarroll......................... 905 Chilton Corp. v. Hyde ......................................... 910 City. See name of city. Clemons v. Mississippi......................................... 904 Colonial American Life Ins. Co. v. Commissioner................ 244 Commissioner; Colonial American Life Ins. Co. v................ 244 Commissioner; Dunning v........................................ 911 Commissioner; L & B Corp. v.................................. 905 Commissioner; Sterner v........................................ 907 Commissioner of Internal Revenue. See Commissioner. Commonwealth. See name of Commonwealth. Connaughton; Harte-Hankes Communications, Inc. v............... 657 Connecticut; Velez v........................................... 906 Connecticut; Velilla v......................................... 908 Consolidated Rail Corp. v. Railway Labor Executives’ Assn..... 299 Cooper v. Eisenman ............................................ 907 Corrections Commissioner. See name of commissioner. Council of New Orleans; New Orleans Public Service, Inc. v.... 350 County. See name of county. Court of Appeals of Md.; Martin v.............................. 908 Credit Bureau Services-New Orleans v. Hyde..................... 910 Culberson v. Veterans Administration........................... 911 D.; Michael H. v............................................... 110 Daisley v. General Electric Co................................. 906 Dallas Independent School Dist. v. Jett........................ 701 Dallas Independent School Dist.; Jett v....................... 701 Daniels v. Indiana............................................. 902 Davenport v. United States..................................... 907 Dawson v. Montana.............................................. 910 Dean v. West Virginia ......................................... 908 TABLE OF CASES REPORTED vn Page Dellmuth v. Muth............................................. 223 Delo v. Gilmore.............................................. 913 Delo; Laws v................................................. 913 DeLorenzo v. New Jersey Dept, of Environmental Protection... 905 Dennis, In re ............................................... 902 Department of Justice; Public Citizen v...................... 440 Department of Justice; Washington Legal Foundation v......... 440 Department of Navy; Gonzalez v............................... 911 Director of penal or correctional institution. See name or title of director. District Court of Wyo., First Judicial Dist.; Johnson v...... 908 District of Columbia; District of Columbia Nurses Assn. v. .. 906 District of Columbia Nurses Assn. v. District of Columbia... 906 Doe; Roe v................................................... 904 Dole; Harman v............................................... 911 Dudley v. Missouri:.......................................... 907 Duffel v. United States...................................... 909 Dugger; Williams v........................................... 908 Dunning v. Commissioner ..................................... 911 Eagle-Picher Industries, Inc. v. United States............... 904 Edwards v. Black............................................. 913 Edwards v. Mississippi....................................... 901 Edwards; Schoemehl v......................................... 906 Edwards; S. H. v............................................. 905 Eisenman; Cooper v........................................... 907 Embry v. Kentucky............................................ 907 Enewetak, Rongelap, and Other Marshall Islands v. United States 909 Evatt; Woomer v. ............................................ 901 F.; Florida Star v........................................... 524 Fahy v. United States........................................ 909 Federal Deposit Ins. Corp.; Leach v.......................... 905 F. G. v. Lawton............................................. 909 Flight Attendants v. Zipes................................... 754 Floria; Wemhoff v............................................ 906 Florida v. Wells............................................. 903 Florida Star v. B. J. F...................................... 524 Furnari v. United States..................................... 907 G. v. Lawton................................................ 909 General Electric Co.; Daisley v.............................. 906 Gerald D.; Michael H. v. .................................... 110 Gibson; Arkansas v........................................... 910 Gilmore v. Armontrout........................................ 911 Gilmore; Delo v.............................................. 913 Gladden v. Roach............................................. 907 VIII TABLE OF CASES REPORTED Page Gonzalez v. Department of Navy................................ 911 Gonzalez-Alvarez v. United States............................. 909 Goodrich Co.; Worrell v. ..................................... 907 Grammer v. Patterson Services, Inc............................ 906 Gwaltney; Bishop v............................................ 904 H. v. Edwards................................................ 905 H. v. Gerald D................................................ 110 Halverson v. Barbour.......................................... 908 Hamilton v. Stovall........................................... 905 Hampton v. Michigan........................................... 904 Harman v. Dole................................................ 911 Harper v. Nix................................................. 908 Harte-Hankes Communications, Inc. v. Connaughton.............. 657 Hayes v. Lockhart............................................. 902 Healy v. Beer Institute....................................... 324 Hernandez v. California....................................... 910 Hoffman v. Petsock............................................ 908 Hoke v. Virginia.............................................. 910 Hotel Workers; Adamo v. ...................................... 911 Hughes; Raley v............................................... 910 Humphrey, In re............................................... 903 Hyde; Chilton Corp. v......................................... 910 Hyde; Credit Bureau Services-New Orleans v.................... 910 Idaho; McKinney v............................................. 913 Indiana; Daniels v............................................ 902 In re. See name of party. Jackson v. Maine.............................................. 904 Jenkins; Missouri v........................................... 274 Jett v. Dallas Independent School Dist........................ 701 Jett; Dallas Independent School Dist. v....................... 701 Johnson v. District Court of Wyo., First Judicial Dist........ 908 Johnson; Texas v.............................................. 397 Jones v. Thomas............................................... 376 Judge, Second Administrative Judicial Dist. of Tex.; Hamilton v. .. 905 Kaiser Aluminum & Chemical Corp. v. Bonjorno.................. 903 Kaiser Aluminum & Chemical Corp.; Bonjorno v.................. 903 Karzov v. Attorney Registration and Disciplinary Comm’n of Ill. .. 906 Kenaitze Indian Tribe; Alaska v............................... 905 Kentucky; Embry v............................................ 907 Klein v. United States ....................................... 909 Kosko v. United States........................................ 909 Labor Union. See name of trade. Laws v. Delo.................................................. 913 Lawton; F. G. v. ............................................. 909 TABLE OF CASES REPORTED IX Page L & B Corp. v. Commissioner ................................... 905 Leach v. Federal Deposit Ins. Corp............................. 905 Leech v. United States......................................... 907 Lillback, In re ............................................... 902 Local. For labor union, see name of trade. Lockhart; Hayes v. ............................................ 902 Lombardo v. United States...................................... 905 Macurdy v. Pennsylvania........................................ 906 Maine; Jackson v............................................... 904 Mandel; United States v........................................ 906 Martin v. Court of Appeals of Md............................... 908 Massachusetts v. Oakes......................................... 576 McCarroll; Central La. Telephone Co. v......................... 905 McCormick v. Texas Commerce Bank, N. A......................... 910 McKinney v. Idaho.............................................. 913 McLean Credit Union; Patterson v............................... 164 Meade-Murphy v. Atlanta........................................ 907 Meridian Life Ins. Co.; Steffen v. ............................ 907 Merriman; Walton v............................................. 905 Michael H. v. Gerald D......................................... 110 Michigan; Hampton v............................................ 904 Michigan Citizens for Independent Press v. Thornburgh.......... 903 Michigan Dept, of State Police; Will v.......................... 58 Millsap; Quinn v. .............................................. 95 Mississippi; Clemons v......................................... 904 Mississippi; Edwards v. ....................................... 901 Missouri; Dudley v............................................. 907 Missouri v. Jenkins............................................ 274 Monsanto; United States v...................................... 600 Montana; Dawson v. ............................................ 910 Montemayor, In re.............................................. 903 Moser v. United States......................................... 909 Muth; Dellmuth v............................................... 223 Naiyavaj v. United States...................................... 909 New Jersey Dept, of Environmental Protection; DeLorenzo v. .... 905 New Orleans Public Service, Inc. v. Council of New Orleans... 350 Nittayanupap v. United States.................................. 902 Nix; Harper v.................................................. 908 Norfin; Ricoh Co. v............................................ 910 Northeast Ill. Regional Commuter Railroad Corp. v. Baker..... 901 Norton v. United States........................................ 907 Oakes; Massachusetts v......................................... 576 Ohio; Walker v................................................. 908 Parish. See name of parish. X TABLE OF CASES REPORTED Page Patterson v. McLean Credit Union................................... 164 Patterson v. Transworld Drilling Co................................ 908 Patterson Services, Inc.; Grammer v................................ 906 Pennsylvania; Macurdy v. .......................................... 906 Pennsylvania; Travaglia v.......................................... 910 Pennsylvania v. Union Gas Co......................................... 1 Pennsylvania; Yarris v. ........................................... 910 Petsock; Hoffman v. ............................................... 908 Pittsburgh & Lake Erie R. Co. v. Railway Labor Executives’ Assn. 490 Powell, In re...................................................... 907 Prevenslik v. Rossko............................................... 902 Public Citizen v. Department of Justice............................ 440 Quinn v. Millsap.................................................... 95 Railway Labor Executives’ Assn.; Consolidated Rail Corp. v.. 299 Railway Labor Executives’ Assn.; Pittsburgh & Lake Erie R. Co. v. 490 Raley v. Hughes.................................................... 910 Ramirez v. United States........................................... 909 Randall v. Thornburgh.............................................. 904 Raymark Industries, Inc. v. United States.......................... 904 Reidt v. United States............................................. 911 Rice v. Washington................................................. 910 Richards; Attorneys’ Title Guaranty Fund, Inc. v. ................. 906 Ricoh Co. v. Norfin................................................ 910 Ricoh Co. v. Snellman.............................................. 910 Rivera v. United States............................................ 909 Roach; Gladden v. ................................................. 907 Rock Against Racism; Ward v........................................ 781 Roe v. Doe......................................................... 904 Roe v. Superior Court of Cal., Marin County........................ 904 Ross v. Zimmerman.................................................. 908 Rossko; Prevenslik v............................................... 902 Rumney v. United States............................................ 908 St. Tammany Parish Hospital; Tasher v.............................. 906 Salazar v. United States........................................... 902 Sales Tax Collector, St. Charles v. Westside Sand Co............... 905 Salt Lake City; Barron v........................................... 911 Sams v. United States ............................................. 905 Schaefer, In re.................................................... 902 Schafer; Vest v. .................................................. 904 Schoemehl v. Edwards............................................... 906 Schoo v. U. S. Postal Service...................................... 904 Secretary of Labor; Harman v....................................... 911 S. H. v. Edwards.................................................. 905 Sierra Club; Simkins Industries, Inc. v............................ 904 TABLE OF CASES REPORTED XI Page Simkins Industries, Inc. v. Sierra Club......................... 904 Skoblow v. Ameri-Manage, Inc.................................... 904 Snellman; Ricoh Co. v........................................... 910 Solon v. United States.......................................... 911 South Carolina v. Catawba Indian Tribe of S. C., Inc............ 906 S. T. v. Board of Ed. of Millville, N. J........................ 901 State. See name of State. Steffen v. Meridian Life Ins. Co................................ 907 Stengel v. United States ....................................... 909 Sterner v. Commissioner......................................... 907 Stovall; Hamilton v............................................. 905 Superintendent of penal or correctional institution. See name or title of superintendent. Superior Court of Cal., Marin County; Roe v..................... 904 Superior Court of Sacramento County; Walker v................... 905 T. v. Board of Ed. of Millville, N. J.......................... 901 Tampa City Council; Al-Hakim v. ................................ 908 Tasher v. St. Tammany Parish Hospital........................... 906 Teamsters v. Terry.............................................. 903 Tennessee; Caruthers v.......................................... 910 Terry; Teamsters v.............................................. 903 Texas v. Johnson................................................ 397 Texas; Webb v................................................... 910 Texas Commerce Bank, N. A.; McCormick v......................... 910 Thomas v. Alabama............................................... 910 Thomas; Jones v................................................. 376 Thornburgh; Michigan Citizens for Independent Press v........... 903 Thornburgh; Randall v........................................... 904 Thorpe v. United States......................................... 909 Townley; Waye v. ............................................... 910 Transworld Drilling Co.; Patterson v............................ 908 Travaglia v. Pennsylvania....................................... 910 Trice; United States v.......................................... 914 Trousdale v. United States...................................... 909 Union. For labor union, see name of trade. Union Gas Co.; Pennsylvania v..................................... 1 United States. See name of other party. U. S. Postal Service; Asberry v. .............................. 911 U. S. Postal Service; Schoo v................................... 904 Upjohn Co.; Wilsmann v.......................................... 906 Velez v. Connecticut............................................ 906 Velilla v. Connecticut.......................................... 908 Verdugo-Urquidez; United States v............................... 903 Vest v. Schafer................................................. 904 XII TABLE OF CASES REPORTED Page Veterans Administration; Culberson v.......................... 911 Video Int’l Productions, Inc. v. Warner-Amex Cable Com., Inc. ... 906 Virginia; Hoke v.............................................. 910 Wagstaff v. United States..................................... 907 Wahl; Barrow v................................................ 907 Walker v. Ohio................................................ 908 Walker v. Superior Court of Sacramento County................. 905 Walton v. Merriman............................................ 905 Ward v. Rock Against Racism................................... 781 Warden. See name of warden. Warner-Amex Cable Com., Inc.; Video Int’l Productions, Inc. v. .. 906 Washington; Rice v............................................ 910 Washington Legal Foundation v. Department of Justice.......... 440 Waye v. Townley................................................ 910 Webb v. Texas.................................................. 910 Wells; Florida v.............................................. 903 Wemhoff v. Floria.............................................. 906 Westside Sand Co.; Sales Tax Collector, St. Charles v......... 905 West Virginia; Dean v.......................................... 908 Will v. Michigan Dept, of State Police ........................ 58 Williams v. Dugger............................................. 908 Wilsmann v. Upjohn Co.......................................... 906 Wine & Spirits Wholesalers of Conn., Inc. v. Beer Institute .. 324 Woods v. Bobrow, Greenapple & Skolnik.......................... 905 Woomer v. Evatt................................................ 901 Worrell v. B. F. Goodrich Co................................... 907 Wright v. United States........................................ 909 Yarris v. Pennsylvania......................................... 910 Zimmerman; Ross v.............................................. 908 Zipes; Flight Attendants v..................................... 754 Zolin; United States v......................................... 554 TABLE OF CASES CITED Page Abrams v. Baylor College of Medicine, 805 F. 2d 528 285 Abrams v. United States, 250 U.S. 616 419 Afro-American Publishing Co. v. Jaffe, 125 U.S. App. D. C. 70 691 Aikens v. California, 406 U.S. 813 585 Air Line Pilots Assn., Int’l v. Eastern Air Lines, Inc., 276 U.S. App. D. C. 199 304 Akron Center for Reproductive Health v. Akron, 604 F. Supp. 1268 774, 778 Alabama v. Pugh, 438 U.S. 781 86, 92 Alabama Pub. Serv. Comm’n v. Southern R. Co., 341 U.S. 341 360-362 Albemarle Paper Co. v. Moody, 422 U.S. 405 759, 772 Allen v. McCurry, 449 U.S. 90 721 Aluminum Co. of America v. Central Lincoln Peoples’ Utility Dist., 467 U.S. 380 464 Alyeska Pipeline Service Co. v. Wilderness Society, 421 U.S. 240 200, 204, 758, 773 Amerada Hess Corp. v. Director, N. J. Div. of Taxation, 490 U.S. 66 348 American Fire & Casualty Co. v. Finn, 341 U.S. 6 26 American Foreign Service Assn. v. Garfinkel, 490 U.S. 153 466 Anderson v. Brand, 303 U.S. 95 102 Anderson v. Celebrezze, 460 U.S. 780 799 Anderson v. Liberty Lobby, Inc., 477 U.S. 242 530 Page Andrews v. Louisville & Nash- ville R. Co., 406 U.S. 320 173 Antitrust Grand Jury, In re, 805 F. 2d 155 569 Apex Hosiery Co. v. Leader, 310 U.S. 469 205 Argersinger v. Hamlin, 407 U.S. 25 646 Arizona v. Rumsey, 467 U.S. 203 172 Arkansas Electric Cooperative Corp. v. Arkansas Pub. Serv. Comm’n, 461 U.S. 375 365 Arkansas Writers’ Project, Inc. v. Ragland, 481 U.S. 221 541 Armstrong v. Manzo, 380 U.S. 545 163 AS ARCO Inc. v. Kadish, 490 U.S. 605 103 Ashe v. Swenson, 397 U.S. * 436 388 Associated Enterprises, Inc. v. Toltec Watershed Improvement Dist., 410 U.S. 743 104, 105 Atascadero State Hospital v. Scanlon, 473 U.S. 234 7, 16,23-25,28, 36,39,44, 46,47,49,56,65, 75, 76, 85, 227, 228, 230-233, 238-243, 279, 294, 295 Attorney General of N. J. v. First Family Mortgage Corp, of Fla., 487 U.S. 1213 585 Aubuchon v. Missouri, 631F. 2d 581 62 Autry v. Estelle, 464 U.S. 1 912, 913 Avco Delta Corporation Canada Ltd. v. United States, 484 F. 2d 692 631 Ayers, In re, 123 U.S. 443 71 Bacchus Imports, Ltd. v. Dias, 468 U.S. 263 344 XIII XIV TABLE OF CASES CITED Page Bachellar v. Maryland, 397 U.S. 564 414 Bacon v. Rutland R. Co., 232 U.S. 134 372 Badaracco v. Commissioner, 464 U.S. 386 297 Baker v. Carr, 369 U.S. 186 102 Baldwin v. G. A. F. Seelig, Inc., 294 U.S. 511 332- 334, 336, 345, 346 Ball v. James, 451 U.S. 355 104, 105, 109 Bankers Life & Casualty Co. v. Crenshaw, 486 U.S. 71 117 Bantam Books, Inc. v. Sullivan, 372 U.S. 58 809 Bates v. State Bar of Ariz., 433 U.S. 350 531 Beckley Newspapers Corp. v. Hanks, 389 U.S. 81 667, 685 Beers v. Arkansas, 20 How. 527 67 Bell v. Burson, 402 U.S. 535 152 Bell v. Hood, 327 U.S. 678 742 Bell v. Milwaukee, 746 F. 2d 1205 745 Berkley & Co., In re, 629 F. 2d 548 556 Bigelow v. Virginia, 421 U.S. 809 582, 583, 587 Bird v. State Dept, of Public Safety, 375 N. W. 2d 36 62 Bivens v. Six Unknown Fed. Narcotics Agents, 403 U.S. 388 735 Blanchard v. Bergeron, 489 U.S. 87 283 Blockburger v. United States, 284 U.S. 299 385 Blum v. Bacon, 457 U.S. 132 480 Blum v. Stenson, 465 U.S. 886 283, 285, 286 Board of Airport Comm’rs of Los Angeles v. Jews for Jesus, Inc., 482 U.S. 569 581, 595 Board of Ed. of Hendrick Hudson Central School Dist. v. Rowley, 458 U.S. 176 234 Board of Liquidation v. McComb, 92 U.S. 531 71 Page Board of Regents of State Colleges v. Roth, 408 U.S. 564 138 Bob Jones Univ. v. United States, 461 U.S. 574 174, 191, 205 Boettger v. Loverro, 555 A. 2d 1234 551 Boldt v. State, 101 Wis. 2d 566 62 Bolger v. Youngs Drug Products Corp., 463 U.S. 60 414 Bollenbach v. United States, 326 U.S. 607 269 Boos v. Barry, 485 U.S. 312 401, 407, 408, 410-412, 792, 798, 804 Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485 659, 685, 686, 689, 694, 697 Boston Sand & Gravel Co. v. United States, 278 U.S. 41 455 Boudreaux v. Baton Rouge Marine Contracting Co., 437 F. 2d 1011 202 Bowers v. Hardwick, 478 U.S. 186 127 Boys Markets, Inc. v. Retail Clerks, 398 U.S. 235 172, 173, 513 Bozza v. United States, 330 U.S. 160 387, 394, 395 Braden v. 30th Judicial Circuit Ct. of Ky., 410 U.S. 484 173 Bradley, In re, 318 U.S. 50 379, 380, 382- 384, 386-391, 395, 396 Brady v. United States, 397 U.S. 742 634 Brandenburg v. Ohio, 395 U.S. 444 409, 410, 418 Brandon v. Holt, 469 U.S. 464 71, 91 Broadrick v. Oklahoma, 413 U.S. 601 581, 583, 588-590, 595 Brockett v. Spokane Arcades, Inc., 472 U.S. 491 581, 586 Brotherhood of Locomotive Engineers v. Burlington Northern R. Co., 838 F. 2d 1087 301, 306 Brotherhood of Maintenance of Way Employees, Lodge 16 v. Burlington Northern R. Co., 802 F. 2d 1016 301, 307 TABLE OF CASES CITED xv Page Brown v. Board of Ed., 347 U.S. 483 174, 191 Brown v. GSA, 425 U.S. 820 733, 734, 748 Brown v. Louisiana, 383 U.S. 131 404, 414 Brown-Forman Distillers Corp, v. New York State Liquor Authority, 476 U.S. 573 330- 339, 342, 343 Buckley v. Valeo, 424 U.S. 1 414, 487, 488 Bullock v. Carter, 405 U.S. 134 107 Burford v. Sun Oil Co., 319 U.S. 315 355, 358-362, 364, 374 Burkey v. Southern Ohio Correctional Facility, 38 Ohio App. 3d 170 62 Burlington Northern R. Co. v. Maintenance of Way Employes, 481 U.S. 429 311, 518 Caban v. Mohammed, 441 U.S. 380 123, 129, 133, 142-144, 153, 158, 160 Cabana v. Bullock, 474 U.S. 376 269 Cabell v. Markham, 148 F. 2d 737 455 Caldwell v. District Court, 644 P. 2d 26 572 Caldwell v. National Brewing Co., 443 F. 2d 1044 202 Cale v. Covington, 586 F. 2d 311 735 Calero-Toledo v. Pearson Yacht Leasing Co., 416 U.S. 663 615, 653 Califano v. Sanders, 430 U.S. 99 201 California v. LaRue, 409 U.S. 109 26 California Retail Liquor Dealers Assn. v. Midcal Aluminum, Inc., 445 U.S. 97 349 Cameo Convalescent Center, Inc. v. Senn, 738 F. 2d 836 288 Cannon v. University of Chicago, 441 U.S. 677 721, 732, 743, 746 Canton v. Harris, 489 U.S. 378 711 Page Cantwell v. Connecticut, 310 U.S. 296 409 Cape Publications, Inc. v. Hitchner, 514 So. 2d 1136 550 Caplin & Drysdale, Chartered v. United States, 491 U.S. 617 604, 614-616 Carey v. Brown, 447 U.S. 455 414, 417, 796 Carpenters v. United States, 330 U.S. 395 268, 270 Carrington v. Rash, 380 U.S. 89 121 Carter v. Jury Comm’n of Greene County, 396 U.S. 320 103 Case Co. v. Borak, 377 U.S. 426 742 Central Hudson Gas & Electric Corp. v. Public Serv. Comm’n of N. Y., 447 U.S. 557 366 Chandler v. Fretag, 348 U.S. 3 644 Chaplinsky v. New Hampshire, 315 U.S. 568 409, 430 Chapman v. California, 386 U.S. 18 271 Chapman v. Houston Welfare Rights Organization, 441 U.S. 600 724, 730, 731 Chappelle v. Greater Baton Rouge Airport Dist., 431 U.S. 159 98- 100, 104, 106, 108, 109 Charles v. Daley, 846 F. 2d 1057 771, 778 Chicot County v. Sherwood, 148 U.S. 529 358 Chisholm v. Georgia, 2 Dall. 419 32, 36, 37, 78 Christiansburg Garment Co. v. EEOC, 434 U.S. 412 307, 759-764, 766-768, 771, 772, 776, 779 Church of Holy Trinity v. United States, 143 U.S. 457 454, 470-474 City. See name of city. City Council of Los Angeles v. Taxpayers for Vincent, 466 U.S. 789 414, 432, 595, 796, 802 XVI TABLE OF CASES CITED Page Civil Rights Cases, 109 U.S. 3 82, 722 CSC v. Letter Carriers, 413 U.S. 548 436 Clark v. Barnard, 108 U.S. 436 26, 293 Clark v. Community for Creative Non-Violence, 468 U.S. 288 406, 407, 791, 796-802, 805 Clark v. United States, 289 U.S. 1 563, 564, 567 Cleveland Bd. of Ed. v. LaFleur, 414 U.S. 632 120, 121, 151 Coates v. Cincinnati, 402 U.S. 611 409 Cody v. Union Electric, 518 F. 2d 978 210 Cohens v. Virginia, 6 Wheat. 264 28, 358 Coker v. Georgia, 433 U.S. 584 542 Collector v. Day, 11 Wall. 113 728, 729 Colorado Interstate Gas Co. v. FPC, 324 U.S. 581 371 Colorado River Water Conservation Dist. v. United States, 424 U.S. 800 359, 361, 362, 368 Commissioner v. Idaho Power Co., 418 U.S. 1 253 Commissioner v. Standard Life & Accident Ins. Co., 433 U.S. 148 254 Commissioner of Internal Revenue. See Commissioner. Commonwealth. See name of Commonwealth. Community for Creative Non- Violence v. Watt, 227 U.S. App. D. C. 19 406 Cone Memorial Hospital v. Mercury Construction Corp., 460 U.S. 1 368 Connecticut v. Johnson, 460 U.S. 73 266, 269, 270 Consolidated Rail Corporation v. Railway Labor Executives’ Assn., 491 U.S. 299 516 Page Construction Laborers v. Curry, 371 U.S. 542 173 Consumer Product Safety Comm’n v. GTE Sylvania, Inc., 447 U.S. 102 751 Continental T. V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 173 Cort v. Ash, 422 U.S. 66 743 Cotton v. United States, 11 How. 229 78 County. See name of county. Cox v. Louisiana, 379 U.S. 536 409 Cox Broadcasting Corp. v. Cohn, 420 U.S. 469 529-535, 537, 538, 543-547, 553 Craig v. Boren, 429 U.S. 190 121, 349, 624 Crawford Fitting Co. v. J. T. Gibbons, Inc., 482 U.S. 437 297 Crowell v. Benson, 285 U.S. 22 466, 481, 643 CTS Corp. v. Dynamics Corp, of America, 481 U.S. 69 337 Curtis Publishing Co. v. Butts, 388 U.S. 130 659, 664, 665, 688, 692, 693, 695, 696 Dallas v. Stanglin, 490 U.S. 19 406 Dandridge v. Williams, 397 U.S. 471 740 David D. v. Dartmouth School Committee, 775 F. 2d 411 227, 235 Davis v. Gray, 16 Wall. 203 88 Davis v. Michigan Dept, of Treasury, 489 U.S. 803 27 Deakins v. Monaghan, 484 U.S. 193 359 DeBartolo Corp. v. Florida Gulf Coast Building & Construction Trades Council, 485 U.S. 568 466, 611, 636, 643 DeFunis v. Odegaard, 416 U.S. 312 585 Delaware State College v. Ricks, 449 U.S. 250 190, 741 De Lima v. Bidwell, 182 U.S. 1 742 Della Grotta v. Rhode Island, 781 F. 2d 343 61, 86 TABLE OF CASES CITED XVII Page Dellmuth v. Muth, 491 U.S. 223 16, 50, 75, 279 Denton Construction Co. v. Missouri State Highway Comm’n, 454 S. W. 2d 44 293 Department of Air Force v. Rose, 425 U.S. 352 449 Department of Justice v. Julian, 486 U.S. 1 449 Department of Justice v. Reporters Comm, for Freedom of Press, 489 U.S. 749 449, 552 Detroit & T. S. L. R. Co. v. Transportation Union, 396 U.S. 142 303, 305, 503- 508, 517-520, 522, 523 Diamond v. Charles, 476 U.S. 54 778 Dister v. The Continental Group, Inc., 859 F. 2d 1108 216 District of Columbia v. Carter, 409 U.S. 418 197, 713 District of Columbia Court of Appeals v. Feldman, 460 U.S. 462 371 Ditson v. California, 372 U.S. 933 585 Division No. 1, Detroit, Brotherhood of Locomotive Eng’rs v. Consolidated Rail Corp., 844 F. 2d 1218 303, 304 Doe v. Maher, 792 F. 2d 1470 227 Doe v. Sarasota-Bradenton Florida Television Co., 436 So. 2d 328 529 Doe*Corp., In re, 675 F. 2d 482 571 Dollar Savings Bank v. United States, 19 Wall. 227 73 Dombrowski v. Pfister, 380 U.S. 479 584 Downum v. United States, 372 U.S. 734 390 Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749 530 Duncan v. Louisiana, 391 U.S. 145 268 Edelman v. Jordan, 415 U.S. 651 26, 27, 55, 63, 89, 93, 238, 239, 241, 278, 279 Page Edgar v. MITE Corp., 457U.S. 624 333, 336 Edgar v. State, 92 Wash. 2d 217 62, 63 Edward J. DeBartolo Corp. v. Florida Gulf Coast Building & Construction Trades Council, 485 U.S. 568 466, 611, 636, 643 Edwards v. National Audubon Society, Inc., 556 F. 2d 113 660, 661, 692, 694 Edwards v. South Carolina, 372 U.S. 229 686 Eisenstadt v. Baird, 405 U.S. 438 128, 132, 139, 624 Elgin, J. & E. R. Co. v. Burley, 325 U.S. 711 302, 303, 305 Ellis v. Blum, 643 F. 2d 68 735 Employees v. Missouri Dept, of Public Health and Welfare, 411 U.S. 279 9, 11, 14, 15,17,25,29,41,48,49, 52, 54, 76, 239, 241, 279 Erznoznik v. City of Jacksonville, 422 U.S. 205 591, 594, 598, 599 Estate. See name of estate. Eu v. San Francisco County Democratic Central Comm., 489 U.S. 214 453 Evans v. Jeff D., 475 U.S. 717 633 Ex parte. See name of party. Fairmont Creamery Co. v. Minnesota, 275 U.S. 70 279 Faretta v. California, 422 U.S. 806 628, 645 FBI v. Abramson, 456 U.S. 615 449, 470 FCC v. Pacifica Foundation, 438 U.S. 726 409, 414, 438 FDIC v. Philadelphia Gear Corp., 476 U.S. 426 454 FEC v. National Conservative Political Action Committee, 470 U.S. 480 809 Felder v. Casey, 487 U.S. 131 66 Firefighters Institute for Racial Equality v. St. Louis, 549 F. 2d 506 180 XVIII TABLE OF CASES CITED Page Firestone v. Time, Inc., 460 F. 2d 712 661 First National Maintenance Corp. v. NLRB, 452 U.S. 666 508, 522, 523 Fisher v. United States, 425 U.S. 391 562 Fitzpatrick v. Bitzer, 427 U.S. 445 14-18, 22, 23, 26, 41, 62, 70, 227, 279, 294 Fleischmann Corp. v. Maier Brewing Co., 386 U.S. 714 732 Fletcher v. Peck, 6 Cranch 87 435 Flood v. Kuhn, 407 U.S. 258 201 Food Employees v. Logan Val- ley Plaza, Inc., 391 U.S. 308 404 Ford Motor Co. v. EEOC, 458 U.S. 219 763, 764 Forfeiture Hearing as to Caplin & Drysdale, Chartered, In re, 837 F. 2d 637 606, 607, 652 Forrester v. White, 484 U.S. 219 86 Fox Film Corp. v. Muller, 296 U.S. 207 101 Francis v. Franklin, 471 U.S. 307 265 Franks v. Bowman Transporta- tion Co., 424 U.S. 747 756, 779 Freedman v. Maryland, 380 U.S. 51 793, 809 Frisby v. Schultz, 487 U.S. 474 411, 791, 796, 799, 801, 802, 804 Fuchilla v. Layman, 109 N. J. 319 62 Furnco Construction Corp. v. Waters, 438 U.S. 567 186, 187, 216, 217 Garcia v. San Antonio Metropol- itan Transit Authority, 469 U.S. 528 28 Garland, Ex parte, 4 Wall. 333 485 Garrison v. Louisiana, 379 U.S. 64 530- 532, 667, 681, 687, 688 Gary A. v. New Trier High School Dist. No. 203, 796 F. 2d 940 227 Page Gay v. State, 730 S. W. 2d 154 62 Gay Student Services v. Texas A&M Univ., 612 F. 2d 160 61 Geier v. Richardson, 871 F. 2d 1310 774 General Building Contractors Assn., Inc. v. Pennsylvania, 458 U.S. 375 186, 190, 198, 215, 721, 722, 741, 763 General Electric Co. v. Gilbert, 429 U.S. 125 201, 464, 465, 479, 480 Georgia v. Rachel, 384 U.S. 780 176, 721 Gertz v. Robert Welch, Inc., 418 U.S. 323 530, 539, 548, 666, 686 Gibbons v. Ogden, 9 Wheat. 1 20 Gideon v. Wainwright, 372 U.S. 335 644, 646 Ginsberg v. New York, 390 U.S. 629 592 Girouard v. United States, 328 U.S. 61 200 Globe Newspaper Co. v. Superior Court of Norfolk County, 457 U.S. 596 539 Glona v. American Guarantee & Liability Ins. Co., 391 U.S. 73 145 Goldberg v. Kelly, 397 U.S. 254 138 Goldwater v. Ginzburg, 414 F. 2d 324 661 Gomez v. Perez, 409 U.S. 535 131, 145 Goodman v. Lukens Steel Co., 482 U.S. 656 177, 183, 190, 212, 221, 222, 741 Governor of Ga. v. Madrazo, 1 Pet. 110 93 Grand Jury Subpoena Duces Tecum Dated September 15, 1983, In re, 731 F. 2d 1032 564 Grand Jury Witness, In re, 695 F. 2d 359 569 Grano v. Barry, 251 U.S. App. D. C. 289 765 Grayned v. Rockford, 408 U.S. 104 414, 794, 796, 804 TABLE OF CASES CITED XIX Page Great Northern Ins. Co. v. Read, 322 U.S. 47 38 Green v. Bock Laundry Ma- chine Co., 490 U.S. 504 454 Green v. Mansour, 474 U.S. 64 15, 27, 41, 89 Green v. United States, 9 Wall. 655 74 Greenbelt Cooperative Publishing Assn., Inc. v. Bresler, 398 U.S. 6 530, 685, 695, 696 Greenwood v. Ross, 778 F. 2d 448 745 Gregg v. Georgia, 428 U.S. 153 901, 911, 913, 914 Grier v. Specialized Skills, Inc., 326 F. Supp. 856 211 Griswold v. Connecticut, 381 U.S. 479 123, 128, 132, 139 Grove City College v. Bell, 465 U.S. 555 201 Guarantee Title & Trust Co. v. Title Guaranty & Surety Co., 224 U.S. 152 74 Gumbhir v. Kansas State Bd. of Pharmacy, 231 Kan. 507 61 Guste v. Council of New Orleans, 309 So. 2d 290 372 Hagans v. Lavine, 415 U.S. 528 63 Hall v. Bio-Medical Application, Inc., 671 F. 2d 300 210 Hall v. Pennsylvania State Police, 570 F. 2d 86 210 Halliburton Oil Well Cementing Co. v. Reily, 373 U.S. 64 27 Halter v. Nebraska, 205 U.S. 34 415, 429 Hankerson v. North Carolina, 432 U.S. 233 740 Hanrahan v. Hampton, 446 U.S. 754 759 Hans v. Louisiana, 134 U.S. 1 7, 18, 19, 23, 25,27,28, 30-33, 35-37, 39, 44, 45, 57, 229, 233 Hardy v. United States, 292 F. 2d 192 386 Haroco, Inc. v. American Nat. Bank & Trust Co. of Chicago, 747 F. 2d 384 609 Page Harrington v. California, 395 U.S. 250 267 Harris v. Missouri Court of Appeals, 787 F. 2d 427 62 Harris v. Oklahoma, 433 U.S. 682 384 Harris v. Reed, 489 U.S. 255 28 Hawaii Housing Authority v. Midkiff, 467 U.S. 229 359 Hayes v. Michigan Central R. Co., Ill U.S. 228 742 Heckler v. Campbell, 461 U.S. 458 740 Heffron v. International Society for Krishna Consciousness, Inc., 452 U.S. 640 791, 793 Heim v. McCall, 239 U.S. 175 78 Henderson’s Tobacco, 11 Wall. 652 746 Henry v. Collins, 380 U.S. 356 530, 667 Hensley v. Eckerhart, 461 U.S. 424 286, 758, 759, 766 Herbert v. Lando, 441 U.S. 153 530, 668 Hess v. Indiana, 414 U.S. 105 685 Heublien, Inc. v. South Carolina Tax Comm’n, 409 U.S. 275 27 Hill v. Florida Dept, of Corrections, 513 So. 2d 129 62 Hishon v. King & Spaulding, 467 U.S. 69 185 Hoffman Estates v. The Flipside, Hoffman Estates, Inc., 455 U.S. 489 796 Honig v. Doe, 484 U.S. 305 235 Hood & Sons, Inc. v. Du Mond, 336 U.S. 525 326 Houston v. Hill, 482 U.S. 451 595-598 Howard Security Services, Inc. v. Johns Hopkins Hospital, 516 F. Supp. 508 210 H. P. Hood & Sons, Inc. v. Du Mond, 336 U.S. 525 326 Huffman v. Pursue, Ltd., 420 U.S. 592 27, 368, 369, 372, 373 Hughes v. Oklahoma, 441 U.S. 322 326 XX TABLE OF CASES CITED Page Humphrey’s Executor v. United States, 295 U.S. 602 487 Hunt v. Blackburn, 128 U.S. 464 562 Hunt v. Liberty Lobby, 720 F. 2d 631 688 Hurd v. Hodge, 334 U.S. 24 713, 722, 745, 747, 750 Hustler Magazine, Inc. v. Falwell, 485 U.S. 46 409, 414, 667 Hutchinson v. Proxmire, 443 U.S. Ill 530 Hutto v. Finney, 437 U.S. 678 62, 63, 90, 91, 227, 241, 278-282, 284, 292, 294 Illinois v. Milwaukee, 406 U.S. 91 20 Illinois Brick Co. v. Illinois, 431 U.S. 720 173 INS v. Chadha, 462 U.S. 919 486 Indiana ex rel. Anderson v. Brand, 303 U.S. 95 102 Ingraham v. Wright, 430 U.S. 651 139 In re. See name of party or proceeding. International Assn, of Machinists and Aerospace Workers v. Northeast Airlines, Inc., 473 F. 2d 549 304, 309 International Brotherhood of Electrical Workers v. Washington Terminal Co., 154 U. S. App. D. C. 119 307 International News Service v. Associated Press, 248 U.S. 215 429 ICC v. Railway Labor Executives’ Assn., 315 U.S. 373 517 Jackson v. Ashton, 8 Pet. 148 26 Jackson v. Jackson, 67 Cal. 2d 245 162 Jacobs v. Mancuso, 825 F. 2d 559 284 Jefferson County Pharmaceutical Assn. v. Abbott Laboratories, 460 U.S. 150 64, 65, 73 Jenkins v. Georgia, 418 U.S. 153 685 Page Jenkins v. Missouri, 807 F. 2d 657, 855 F. 2d 1295 276 Jett v. Dallas Independent School Dist., 798 F. 2d 748 745 J. I. Case Co. v. Borak, 377 U.S. 426 742 John Doe Corp., In re, 675 F. 2d 482 571 Johnson v. Georgia Highway Express, Inc., 488 F. 2d 714 288 Johnson v. Railway Express Agency, Inc., 421 U.S. 454 171, 181, 190, 196, 197, 204, 211, 212, 733, 741 Johnson v. Robison, 415 U.S. 361 231 Johnson v. Transportation Agency, Santa Clara County, 480 U.S. 616 175, 200 Joint Anti-Fascist Refugee Committee v. McGrath, 341 U.S. 123 645 Jones v. Alfred H. Mayer Co., 392 U.S. 409 176, 190, 192-199, 201, 202, 720, 741, 744, 745, 747 Joseph E. Seagram & Sons, Inc. v. Hostetter, 384 U.S. 35 328, 330, 331, 334, 335, 341-344, 348 Juidice v. Vail, 430 U.S. 327 27, 368, 373 Katz v. United States, 389 U.S. 347 551 Kazenbach v. McClung, 379 U.S. 294 20 Kendall v. United States, 12 Pet. 524 742 Kenosha v. Bruno, 412 U.S. 507 71, 93 Kentucky v. Dennison, 24 How. 66 728, 729 Kentucky v. Graham, 473 U.S. 159 63, 71, 89, 91, 762, 763, 777 Kerr v. United States District Court for Northern Dist. of Cal., 426 U.S. 394 569 Kline v. Burke Construction Co., 260 U.S. 226 359 TABLE OF CASES CITED XXI Page Kovacs v. Cooper, 336 U.S. 77 794, 796, 803 Lakewood v. Plain Dealer Publishing Co., 486 U.S. 750 793, 794, 796, 809, 811 Landmark Communications, Inc. v. Virginia, 435 U.S. 829 531, 533, 535, 537, 538 Lane v. Wilson, 307 U.S. 268 372 Lange, Ex parte, 18 Wall. 163 379, 380, 382-385, 388-391 Laska v. United States, 82 F. 2d 672 626 Laurens Federal Savings & Loan Assn. v. South Carolina Tax Comm’n, 365 U.S. 517 28 Lehr v. Robertson, 463 U.S. 248 123, 128, 129, 133, 142-144, 158-161, 163 Leonard v. Frankfort Electric and Water Plant Bd., 752 F. 2d 189 710, 745 Levy v. Louisiana, 391 U.S. 68 145 Lewis v. BT Investment Man- agers, Inc., 447 U.S. 27 340 Lewis v. United States, 92 U.S. 618 74 Library of Congress v. Shaw, 478 U.S. 310 278, 280-282, 290-292, 295 Lisa R., In re, 13 Cal. 3d 636 135, 149 Little v. Streater, 452 U.S. 1 161 Local 1477 United Transportation Union v. Baker, 482 F. 2d 228 306 Locomotive Engineers v. Mis-souri-K.-T. R. Co., 363 U.S. 528 304 Loeb v. Textron, Inc., 600 F. 2d 1003 216 Loft v. Fuller, 408 So. 2d 619 550 Logan v. Zimmerman Brush Co., 455 U.S. 422 145 Los Angeles v. Lyons, 461 U.S. 95 27 Loving v. Virginia, 388 U.S. 1 132, 145 Page Lubin v. Panish, 415 U.S. 709 107 Machinists v. Central Airlines, Inc., 372 U.S. 682 304 Machinists v. Street, 367 U.S. 740 466 Maher v. Gagne, 448 U.S. 122 776 Mahone v. Waddle, 564 F. 2d 1018 745 Maine v. Taylor, 477 U.S. 131 341 Maine v. Thiboutot, 448 U.S. 1 63, 64, 92, 280 Maine Central R. Co. v. United Transportation Union, 787 F. 2d 780 307, 731, 747 Marbury v. Madison, 1 Cranch 137 742 Mardan Corp. v. C. G. C. Music, Ltd., 804 F. 2d 1454 22 Marriage of Sharyne and Stephen B., In re, 124 Cal. App. 3d 524 140 Martin v. Hunter’s Lessee, 1 Wheat. 304 28, 104 Martin v. Struthers, 319 U.S. 141 799, 805 Martin v. Wilks, 490 U.S. 755 762, 764, 768, 777 Martinez v. California, 444 U.S. 277 63, 87, 92 Massachusetts Mut. Life Ins. Co. v. Ludwig, 426 U.S. 479 740 Massey Motors, Inc. v. United States, 364 U.S. 92 253 Mathews v. Eldridge, 424 U.S. 319 154 Mayor of Philadelphia v. Educational Equality League, 415 U.S. 605 103 McDonald v. Santa Fe Trail Transp. Co., 427 U.S. 273 190, 741 McDonnell Douglas Corp. v. Green, 411 U.S. 792 186, 187, 215-217 McNeese v. Board of Ed. for Community Unit School Dist. 187, Cahokia, 373 U.S. 668 361 McPherson v. Blacker, 146 U.S. 1 78 XXII TABLE OF CASES CITED Page McVey Trucking, Inc., In re, 812 F. 2d 311 15, 17 Memphis v. Greene, 451 U.S. 100 207 Merit Life Ins. Co. v. Commissioner, 853 F. 2d 1435 249, 262 Meritor Savings Bank v. Vinson, 477 U.S. 57 180, 209 Merrill Lynch, Pierce, Fenner & Smith, Inc. v. Curran, 456 U.S. 353 743 Merritt v. State, 108 Idaho 20 62 Merritt ex rel. Merritt v. State, 108 Idaho 20 62 Metrocare v. Washington Metropolitan Area Transit Auth., 220 U.S. App. D. C. 104 745 Meyer v. Nebraska, 262 U.S. 390 121, 142 Mezey v. State, 161 Cal. App. 3d 1060 62 Michelle W. v. Ronald W., 39 Cal. 3d 354 135, 146, 149 Michigan v. Long, 463 U.S. 1032 28, 102 Michigan v. Shabaz, 478 U.S. 1017 585 Middlesex County Ethics Comm. v. Garden State Bar Assn., 457 U.S. 423 364, 369, 370, 374 Middlesex County Sewerage Authority v. National Sea Clammers Assn., 453 U.S. 1 748, 749, 753 Miener v. Missouri, 673 F. 2d 969 227 Milk Control Bd. v. Eisenberg Farm Products, 306 U.S. 346 346 Miller v. California, 413 U.S. 15 591, 685 Milliken v. Bradley, 433 U.S. 267 26, 90, 290 Mills Music, Inc. v. Arizona, 591 F. 2d 1278 15 Milton v. Wainwright, 407 U.S. 371 267 Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U.S. 575 541, 549 Page Minnesota v. Clover Leaf Creamery Co., 449 U.S. 456 106 Mississippi Power & Light Co. v. Mississippi ex rel. Moore, 487 U.S. 354 352, 353, 355 Missouri v. Hunter, 459 U.S. 359 381, 382, 386, 392 Missouri Pacific R. Co. v. Stroud, 267 U.S. 404 20 Mistretta v. United States, 488 U.S. 361 487 Mitchell v. Maurer, 293 U.S. 237 26 Mitchum v. Foster, 407 U.S. 225 724 Mobile v. Bolden, 446 U.S. 55 201 Modem American Life Ins. Co. v. Commissioner, 830 F. 2d 110 259 Molina v. Richardson, 578 F. 2d 846 735 Monaco v. Mississippi, 292 U.S. 313 19, 33, 40 Monell v. New York City Dept. of Social Services, 436 U.S. 658 12, 34, 62, 64, 67- 71, 77, 78,82,83,89,92, 200, 201, 710, 712, 726-729, 735-737, 748-750 Monitor Patriot Co. v. Roy, 401 U.S. 265 530, 687 Monroe v. Pape, 365 U.S. 167 62, 70, 76, 82, 724, 726, 749 Monroe v. State Court of Fulton County, 739 F. 2d 568 408 Monroe County v. Florida, 678 F. 2d 1124 15 Moor v. County of Alameda, 411 U.S. 693 733, 744 Moore v. East Cleveland, 431 U.S. 494 121-123, 137, 145, 163 Moore v. Sims, 442 U.S. 415 368 Moore Ice Cream Co. v. Rose, 289 U.S. 373 467 Morissette v. United States, 342 U.S. 246 268 Morris v. Mathews, 475 U.S. 237 380, 384, 385, 387, 388, 396 Morris v. Slappy, 461 U.S. 1 630, 651 TABLE OF CASES CITED XXIII Page Morrison v. Olson, 487 U.S. 654 483, 484 Morton v. Mancari, 417 U.S. 535 510, 746 Moses H. Cone Memorial Hospital v. Mercury Construction Corp., 460 U.S. 1 368 Mt. Healthy Bd. of Ed. v. Doyle, 429 U.S. 274 70 Mullane v. Central Hanover Bank & Trust Co., 339 U.S. 306 154 Murray v. Wilson Distilling Co., 213 U.S. 151 238 Myers v. United States, 272 U.S. 52 484 Nantahala Power & Light Co. v. Thornburg, 476 U.S. 953 352, 353, 357 Nashville, C. & St. L. R. Co. v. Wallace, 288 U.S. 249 103 NAACP v. Button, 371 U.S. 415 686 National Ins. Co. v. Tidewater Co., 337 U.S. 582 138 NLRB v. American National Ins. Co., 343 U.S. 395 308, 309 NLRB v. Catholic Bishop of Chicago, 440 U.S. 490 466, 611, 643 NLRB v. Jones & Laughlin Steel Corp., 301 U.S. 1 20 National Mut. Ins. Co. v. Tidewater Transfer Co., 337 U.S. 582 27 National Railroad Passenger Corp. v. National Assn, of Railroad Passengers, 414 U.S. 453 732, 748 National Railway Labor Conf. v. Int’l Assn, of Machinists and Aerospace Workers, 830 F. 2d 741 305, 306 Nearv. Minnesota ex rel. Olson, 283 U.S. 697 532, 808 Nebraska Press Assn. v. Stuart, 427 U.S. 539 531 Nevada v. Hall, 440 U.S. 410 28, 92 New Energy Co. of Ind. v. Limbach, 486 U.S. 269 340, 344 Page Newman v. Piggie Park Enterprises, Inc., 390 U.S. 400 758, 759, 761, 767, 772, 773 New Orleans Pub. Serv., Inc. v. New Orleans, 782 F. 2d 1236 356, 358, 361 Newport v. Fact Concerts, Inc., 453 U.S. 247 67, 723 Newport News Shipbuilding & Dry Dock Co. v. EEOC, 462 U.S. 669 201 New York, Ex parte, 256 U.S. 490 22, 38 New York v. Ferber, 458 U.S. 747 588, 589, 591, 592, 594, 595, 598 New York v. Shore Realty Co., 759 F. 2d 1032 53 New York State Club Assn. v. City of New York, 487 U.S. 1 435, 590 New York Times Co. v. Sullivan, 376 U.S. 254 530, 659, 663, 665-667, 685, 686, 688, 689, 693, 694 New York Times Co. v. United States, 403 U.S. 713 531, 535 New York Trust Co. v. Eisner, 256 U.S. 345 421 Niemotko v. Maryland, 340 U.S. 268 686, 809 Nixon v. Administrator of General Services, 433 U.S. 425 484 Norfolk & Portsmouth Belt Line R. Co. v. Brotherhood of Trainmen, Lodge No. 514, 248 F. 2d 34 306 North Carolina v. Pearce, 395 U.S. 711 381, 382, 384 Northcross v. Board of Ed. of Memphis City Schools, 611 F. 2d 624 285 Northcross v. Memphis Bd. of Ed., 412 U.S. 427 758, 759 Northwest Airlines, Inc. v. Air Line Pilots Assn., Int’l, 442 F. 2d 251 311 Northwest Central Pipeline Corp. v. State Corp. Comm’n of Kan., 489 U.S. 493 20 XXIV TABLE OF CASES CITED Page Norwegian Nitrogen Products Co. v. United States, 288 U.S. 294 465 Ocala Star-Banner Co. v. Damron, 401 U.S. 295 686 Ohio v. Helvering, 292 U. S. 360 82 Ohio v. Johnson, 467 U.S. 493 381, 392 Ohio Civil Rights Comm’n v. Dayton Christian Schools, Inc., 477 U.S. 619 365, 369, 370, 374 Oklahoma City v. Tuttle, 471 U.S. 808 753 Oklahoma Publishing Co. v. Oklahoma County District Court, 43 U.S. 308 530, 534, 535, 537, 538, 543, 545-547 Oilman v. Evans, 242 U.S. App. D. C. 301 687 Oneida County v. Oneida Indian Nation of N. Y., 470 U.S. 226 15, 41 One 1958 Plymouth Sedan v. Pennsylvania, 380 U.S. 693 653 Order of Conductors v. Pitney, 326 U.S. 561 310 O’Rourke v. Darbishire, [1920] A. C. 581 563 Osborn v. Bank of United States, 9 Wheat. 738 71, 92 Osborn v. Ozlin, 310 U.S. 53 348 Osterneck v. Ernst & Whinney, 489 U.S. 169 294 Owen v. City of Independence, 445 U.S. 622 12, 67, 87 Owen v. Independence, 589 F. 2d 335 735 Owen Equipment & Erection Co. v. Kroger, 437 U.S. 365 26 Pacific Gas & Elec. Co. v. Public Utilities Comm’n of Cal., 475 U.S. 1 804 Pacific Gas & Elec. Co. v. State Energy Resources Conservation and Dev. Comm’n, 461 U.S. 190 365 Palko v. Connecticut, 302 U.S. 319 128 Page Papasan v. Allain, 478 U.S. 265 89, 90, 290 Parden v. Terminal Railway of Ala. Docks Dept., 377 U.S. 184 14, 23, 25, 41-43 Patsy v. Board of Regents of Fla., 457 U.S. 496 66, 205 Patterson v. McLean Credit Union, 485 U.S. 617 171, 172 Patterson v. McLean Credit Union, 491 U.S. 164 711, 739, 741 Peel v. Florida Dept, of Transportation, 600 F. 2d 1070 15 Pembaur v. Cincinnati, 475 U.S. 469 711, 737 Pennekamp v. Florida, 328 U.S. 331 686, 688 Pennhurst State School and Hospital v. Halderman, 465 U.S. 89 16, 26, 27, 65, 74, 89, 227, 241, 294 Pennsylvania v. Delaware Valley Citizens’ Council, 483 U.S. 711 282, 283, 286 Pennsylvania v. Goldhammer, 474 U.S. 28 394 Pennsylvania v. Union Gas Co., 491 U.S. 1 76, 243, 279 Pennzoil Co. v. Texaco Inc., 481 U.S. 1 359, 368, 373 Perry Ed. Assn. v. Perry Local Educator’s Assn., 460 U.S. 37 791, 804 Petitioner F. v. Respondent R., 430 A. 2d 1075 150 Petty v. Tennessee-Missouri Bridge Comm’n, 359 U.S. 275 43 Pfizer Inc. v. India, 434 U.S. 308 74, 81 Philadelphia v. New Jersey, 437 U.S. 617 20, 21 Philadelphia Newspapers, Inc. v. Hepps, 475 U.S. 767 530, 694 Pierce v. Society of Sisters, 268 U.S. 510 121 Pierson v. Ray, 386 U.S. 547 67 Plessy v. Ferguson, 163 U.S. 537 174, 191 Poe v. Ullman, 367 U.S. 497 132 TABLE OF CASES CITED xxv Page Poindexter v. Greenhow, 114 U.S. 270 78, 79, 88, 93 Police Dept, of Chicago v. Mosley, 408 U.S. 92 414 Pollard v. Bailey, 20 Wall. 520 742 Pope v. Illinois, 481 U.S. 497 270-272, 584 Posadas v. National City Bank, 296 U.S. 497 746 Poulos v. New Hampshire, 345 U.S. 395 796 Powell v. Alabama, 287 U.S. 45 616, 644 Power Reactor Co. v. Electricians, 367 U.S. 396 464 Prairie States Life Ins. Co. v. United States, 828 F. 2d 1222 249 Prate v. Freedman, 583 F. 2d 42 765 Preiser v. Rodriguez, 411 U.S. 475 748 Prentis v. Atlantic Coast Line Co., 211 U.S. 210 370-373 Prigg v. Pennsylvania, 16 Pet. 539 728, 729, 752 Prince v. Massachusetts, 321 U.S. 158 142 Public Util. Comm’n of Ohio v. United Fuel Gas Co., 317 U.S. 456 362, 366 Quern v. Jordan, 440 U.S. 332 15-17,41,45,55,63,64,66-69,76,77,84,86,90,241 Quilloin v. Walcott, 434 U.S. 246 123, 142-144, 147, 158 Rahmah Navajo School Bd., Inc. v. Bureau of Revenue, 104 N. M. 302 61 Railroad Co. v. Tennessee, 101 U.S. 337 67 Railroad Telegraphers v. Chicago & N. W. R. Co., 362 U.S. 330 508, 517-523 Railway Clerks v. Florida East Coast R. Co., 384 U.S. 238 504, 519 Railway Labor Executives Assn. v. Norfolk & Western R. Co., 833 F. 2d 700 301 Page Ramos v. Lamm, 713 F. 2d 546 285, 286 Reeves v. Harrell, 791 F. 2d 1481 765 Regan v. Time, Inc., 468 U.S. 641 797, 805 Renton v. Playtime Theatres, Inc., 475 U.S. 41 791, 802, 804 Revere v. Massachusetts Gen- eral Hospital, 463 U.S. 239 106 Rice v. Santa Fe Elevator Corp., 331 U.S. 218 65, 74, 238, 239 Richardson v. Alaska Airlines, Inc., 750 F. 2d 763 765 Richardson v. Byrd, 709 F. 2d 1016 285 Richmond v. J. A. Croson Co., 488 U.S. 469 799, 806 Riverside v. Rivera, 477 U.S. 561 283, 285, 286 Rizzo v. Goode, 423 U.S. 362 27 Rodriguez de Quijas v. Shear-son/American Express, Inc., 490 U.S. 477 173 Roe v. Wade, 410 U.S. 113 127 Rogers v. EEOC, 454 F. 2d 234 180 Rogers v. Okin, 821 F. 2d 22 278, 295 Rome v. United States, 446 U.S. 156 294 Rose v. Clark, 478 U.S. 570 266, 267, 269, 271-273 Rosenblatt v. Baer, 383 U.S. 75 530 Rosenbloom v. Metromedia, Inc., 403 U.S. 29 530, 666 Ruby v. Taca International Air- lines, S. A., 439 F. 2d 1359 307 Ruiz v. Estelle, 679 F. 2d 1115 62 Runyon v. McCrary, 427 U.S. 160 171-175, 177, 178, 181, 182, 185, 190, 191, 196-201, 203-206, 210, 219-222, 722, 739, 741 Russello v. United States, 464 U.S. 16 630 Rutland Railway Corp. v. Brotherhood of Locomotive Engineers, 307 F. 2d 21 308, 309 XXVI TABLE OF CASES CITED Page Sacher v. United States, 343 U.S. 1 647 St. Amant v. Thompson, 390 U.S. 727 667, 686, 688 Saint Francis College v. Al- Khazraji, 481 U.S. 604 190, 741 St. Louis v. Praprotnik, 485 U.S. 112 710, 737, 738 St. Martin Evangelical Lutheran Church v. South Dakota, 451 U.S. 772 466 St. Mary’s Hospital and Health Center v. State, 150 Ariz. 8 62 Salyer Land Co. v. Tulare Lake Basin Water Storage Dist., 410 U.S. 719 104, 105 Sanders v. Dobbs Houses, Inc., 431 F. 2d 1097 202 Sandstrom v. Montana, 442 U.S. 510 265, 266, 268, 271 San Francisco Arts & Athletics, Inc. v. United States Olympic Committee, 483 U.S. 522 415, 429 Santa Fe Industries, Inc. v. Green, 430 U.S. 462 183 Save Our Cumberland Mountains, Inc. v. Hodel, 263 U.S. App. D. C. 409 284 Schacht v. United States, 398 U.S. 58 404, 417 Schad v. Mount Ephraim, 452 U.S. 61 591 Schaumburg v. Citizens for a Better Environment, 444 U.S. 620 581 Schenck v. United States, 249 U.S. 47 430 Scheuer v. Rhodes, 416 U.S. 232 67 Schick v. Reed, 419 U.S. 256 485 Schneider v. State, 308 U.S. 147 805, 808 School Committee of Burlington v. Department of Ed. of Mass., 471 U.S. 359 225, 235, 243 Schultz v. Newsweek, Inc., 668 F. 2d 911 688 Schwegmann Brothers Giant Super Markets v. Louisiana Milk Comm’n, 365 F. Supp. 1144 333 Page Scott v. Young, 307 F. Supp. 1005 211 Seagram & Sons, Inc. v. Hostetter, 384 U.S. 35 328, 330, 331, 334, 335, 341-344, 348 Sealed Case, In re, 219 U.S. App. D. C. 195 571 Sedima, S. P. R. L. v. Imrex Co., 473 U.S. 479 183, 609, 610 Sethy v. Alameda County Water Dist., 545 F. 2d 1157 745 Shaare Tefila Congregation v. Cobb, 481 U.S. 615 190, 207, 741 Shaffer v. Heitner, 433 U.S. 186 334, 336 Shaw v. St. Louis, 664 S. W. 2d 572 62 Shearson/American Express Inc. v. McMahon, 482 U.S. 220 201 Shively v. Lankford, 174 Mo. 535 106 Shuttlesworth v. Birmingham, 394 U.S. 147 809 Sims v. Amos, 340 F. Supp. 691 773 Singleton v. Wulff, 428 U.S. 106 623, 624 Skinner v. Oklahoma, 316 U.S. 535 142 Skinner v. Railway Labor Exec- utives’ Assn., 489 U.S. 602 314, 315, 319 Smith v. Daily Mail Publishing Co., 443 U.S. 97 531, 533-538, 540-546, 549 Smith v. Goguen, 415 U.S. 566 403, 405, 411, 413, 415, 433, 434 Smith v. Reeves, 178 U.S. 436 28 Smith v. Robinson, 468 U.S. 992 201, 234, 235, 237 Snyder v. Massachusetts, 291 U.S. 97 122, 128 Sosna v. Iowa, 419 U.S. 393 26 South Dakota v. Dole, 483 U.S. 203 65, 74 South Dakota v. North Carolina, 192 U.S. 286 33 TABLE OF CASES CITED xxvn Page Southeastern Promotions, Ltd. v. Conrad, 420 U.S. 546 790, 795, 808, 809 Spanish Action Committee of Chicago v. Chicago, 811 F. 2d 1129 824 Special September 1978 Grand Jury, In re, 640 F. 2d 49 571 Spence v. Washington, 418 U.S. 405 403^07, 410, 415, 416, 418, 433 Sporhase v. Nebraska ex rel. Douglas, 458 U.S. 941 340 Springer v. Seamen, 821 F. 2d 871 710, 745 Square D Co. v. Niagara Frontier Tariff Bureau, Inc., 476 U.S. 409 173 Stanley v. Illinois, 405 U.S. 645 120, 123, 133, 139, 142-144, 153, 158, 160 Stanley v. Schwalby, 147 U.S. 508 82 Stanton v. Godfrey, 415 N. E. 2d 103 61 State. See also name of State. State v. Cerny, 248 S. W. 2d 844 379 State v. Green, 633 P. 2d 1381 62, 63 State v. Mitchell, 32 Ohio App. 2d 16 431 State v. Morgan, 612 S. W. 2d 1 379, 381, 386, 391 State v. Olds, 603 S. W. 2d 501 379, 381, 386, 391 State v. Royal, 113 N. H. 224 431 State v. Waterman, 190 N. W. 2d 809 431 State ex rel. Guste v. Council of New Orleans, 309 So. 2d 290 372 Steelworkers v. Warrior & Gulf Navigation Co., 363 U.S. 574 308, 311, 312 Steppelman v. State Highway Comm’n of Mo., 650 S. W. 2d 343 293 Strauder v. West Virginia, 100 U.S. 303 721 Page Street v. New York, 394 U.S. 576 402, 414, 432, 433, 437, 685 Strickland v. Washington, 466 U.S. 668 631-633, 645 Stromberg v. California, 283 U.S. 359 404, 414 Stump v. Sparkman, 435 U.S. 349 67 Sullivan v. Hudson, 490 U.S. 877 774 Sullivan v. Little Hunting Park, Inc., 396 U.S. 229 190, 721, 741, 747 Summers v. United States, 250 F. 2d 132 631 Supreme Court of Va. v. Consumers Union of U. S., Inc., 446 U.S. 719 91, 763, 777, 778 Sutherland v. DeWulf, 323 F. Supp. 740 419 Swift & Co. v. Wickman, 382 U.S. Ill 173 Takahashi v. Fish and Game Comm’n, 334 U.S. 410 713, 745 Tarpley v. Greene, 221 U.S. App. D. C. 227 735 Tavoulareas v. Piro, 260 U.S. App. D. C. 39 661, 666, 668, 691, 697 Taylor v. Jones, 653 F. 2d 1193 745 Teamsters v. United States, 431 U.S. 324 764, 779 TVA v. Hill, 437 U.S. 153 522 Tenney v. Brandhove, 341 U.S. 367 67, 729 Terminiello v. Chicago, 337 U.S. 1 409 Testa v. Katt, 330 U.S. 386 87 Texas v. White, 7 Wall. 700 424 Texas Dept, of Community Affairs v. Burdine, 450 U.S. 248 186, 187, 215, 216, 218 Texas & N. O. R. Co. v. Railway Clerks, 281 U.S. 548 742 Texas & Pacific R. Co. v. Rigsby, 241 U.S. 33 742, 743 Texas State Teachers Assn. v. Garland Independent School Dist., 489 U.S. 782 283, 771 XXVIII TABLE OF CASES CITED Page Textile Workers v. Darlington Mfg. Co., 380 U.S. 263 507- 509, 522 Thomas v. Shipka, 818 F. 2d 496 735 Thornberry v. Delta Air Lines, Inc., 676 F. 2d 1240 285 Thornhill v. Alabama, 310 U.S. 88 793 Tillman v. Wheaton-Haven Recreation Assn., Inc., 410 U.S. 431 171, 190, 196, 197, 204, 210, 741 Time, Inc. v. Hill, 385 U.S. 374 530, 533, 553 Time, Inc. v. Pape, 401 U.S. 279 530, 685, 695, 696 Tinker v. Des Moines Independent Community School Dist., 393 U.S. 503 404, 409 Tiverton Bd. of License Comm’rs v. Pastore, 469 U.S. 238 585 Toledo, P. & W. R. Co. v. Illinois, 744 F. 2d 1296 62 Trainmen v. Chicago R. & I. R. Co., 353 U.S. 30 304, 305, 513 Trainmen v. Howard, 343 U.S. 768 513 Trainmen v. Jacksonville Terminal Co., 394 U.S. 369 496 Trainor v. Hernandez, 431 U.S. 434 27, 368, 373 Trammel v. United States, 445 U.S. 40 562 Transportation Union v. Union Pacific R. Co., 385 U.S. 157 311 Trans World Airlines, Inc. v. Flight Attendants, 489 U.S. 426 311, 765 Traynor v. Turnage, 485 U.S. 535 317 Turner v. Fouche, 396 U.S. 346 98-100, 103-109 Turner v. Safley, 482 U.S. 78 132 Uberoi v. University of Colo., 713 P. 2d 894 61 Udall v. Tailman, 380 U.S. 1 464 Ungar v. Sarafite, 376 U.S. 575 651 Page Union Pacific R. Co. v. Sheehan, 439 U.S. 89 304 United Indus. Workers, Seafarers Union v. Bd. of Trustees, Galveston Wharves, 351F. 2d 183 306 United States v. Albertini, 472 U.S. 675 611, 797, 799-801, 805 United States v. American Trucking Assns., Inc., 310 U.S. 534 455 United States v. Arrellano- Rios, 799 F. 2d 520 395 United States v. Author Serv- ices, Inc., 804 F. 2d 1520 556 United States v. Badalamenti, 614 F. Supp. 194 635, 636, 648-650 United States v. Barrett, 837 F. 2d 1341 556 United States v. Bass, 404 U.S. 336 65, 74, 238 United States v. Bassett, 632 F. Supp. 1308 635, 641, 646, 650 United States v. Bello, 470 F. Supp. 723 609 United States v. Bissell, 866 F. 2d 1343 607 United States v. Bliss, 667 F. Supp. 1298 53 United States v. Brodson, 241 F. 2d 107 631 United States v. Brown, 333 U.S. 18 470 United States v. California, 297 U.S. 175 73, 74 United States v. California, 332 U.S. 19 11, 54 United States v. Classic, 313 U.S. 299 744 United States v. Cooper Corp., 312 U.S. 600 64, 74 United States v. DiFrancesco, 449 U.S. 117 383, 394, 395 United States v. Edmondson, 792 F. 2d 1492 395 United States v. $8,850, 461 U.S. 555 615 United States v. Estevez, 645 F. Supp. 869 635 TABLE OF CASES CITED XXIX Page United States v. Fausto, 484 U.S. 439 30, 181, 510 United States v. Fogel, 264 U.S. App. D. C. 292 394 United States v. Fox, 94 U.S. 315 69, 80 United States v. Friedman, 445 F. 2d 1076 561 United States v. Grace, 461 U.S. 171 404, 791, 796 United States v. Guthrie, 17 How. 284 427 United States v. Halper, 490 U.S. 435 381, 393 United States v. Harvey, 814 F. 2d 905 622, 645, 647, 653 United States v. Hasting, 461 U.S. 499 267 United States v. Herron, 20 Wall. 251 73 United States v. lanniello, 644 F. Supp. 452 635, 648 United States v. Jones, 837 F. 2d 1332 606, 607 United States v. King, 395 U.S. 1 10 United States v. Klein, 13 Wall. 128 485, 488, 489 United States v. Laura, 607 F. 2d 52 648 United States v. Lawless, 709 F. 2d 485 569 United States v. Leather, 271 F. 2d 80 387 United States v. Lee, 106 U.S. 196 71, 93 United States v. Leon, 468 U.S. 897 711 United States v. Locke, 471 U.S. 84 467 United States v. Lowden, 308 U.S. 225 516 United States v. Martin Linen Supply Co., 430 U.S. 564 268 United States v. Maurice, 2 Brock. 96 78 United States v. Mine Workers, 330 U.S. 258 64 United States v. Mississippi, 380 U.S. 128 11, 33, 34 Page United States v. Mitchell, 445 U.S. 535 36 United States v. Monsanto, 491 U.S. 600 623, 624, 626, 632, 636, 637, 639, 642, 652, 654, 761, 772 United States v. Monsanto, 852 F. 2d 1400 623, 637, 640, 642, 643, 650, 652 United States v. Moya-Gomez, 860 F. 2d 706 606, 607 United States v. Naftalin, 441 U.S. 768 611 United States v. Nichols, 841 F. 2d 1485 606, 607 United States v. Nixon, 418 U.S. 683 485 United States v. North Carolina, 136 U.S. 211 293 United States v. O’Brien, 391 U.S. 367 403, 404, 406, 407, 410, 414, 789, 797, 798 United States v. Price, 361 U.S. 304 610, 752 United States v. Price, 383 U.S. 787 84 United States v. Reckmeyer, 631 F. Supp. 1191 635 United States v. Reynolds, 345 U.S. 1 570, 571 United States v. Rogers, 602 F. Supp. 1332 630, 635, 646, 650 United States v. Salerno, 481 U.S. 739 616, 634, 654 United States v. Shewfelt, 455 F. 2d 836 556, 560, 561 United States v. Shirey, 359 U.S. 255 82 United States v. Stanley, 483 U.S. 669 132 United States v. Stowell, 133 U.S. 1 627, 638 United States v. Testan, 424 U.S. 392 10, 33, 36, 48 United States v. Texas, 143 U.S. 621 33 United States v. Thier, 801 F. 2d 1463 647 United States v. Turkette, 452 U.S. 576 606 XXX TABLE OF CASES CITED Page United States v. Union Gas Co., 792 F. 2d 372 6, 46, 56 United States v. Union Gas Co., 575 F. Supp. 949 56 United States v. United States Gypsum Co., 438 U.S. 422 268 United States v. Weber Aircraft Corp., 465 U.S. 792 449 United States v. Weisman, 111 F. 2d 260 571 United States Brewers Assn., Inc. v. Healy, 692 F. 2d 275 328-331, 333, 347 United States Brewers Assn., Inc. v. Healy, 532 F. Supp. 1312 326-328 U. S. Postal Service Bd. of Governors v. Aikens, 460 U.S. 711 187 Upjohn Co. v. United States, 449 U.S. 383 562 Upper Midwest Booksellers Assn. v. Minneapolis, 602 F. Supp. 1361 584 Van Brocklin v. Tennessee, 117 U.S. 151 78 Vargas, In re, 723 F. 2d 1461 569 Vasquez v. Hillery, 474 U.S. 254 172, 191 Verlinden B. V. v. Central Bank of Nigeria, 461 U.S. 480 19 Vincent B. v. Joan R., 126 Cal. App. 3d 619 116, 119, 120, 134, 135, 146, 149, 150, 152, 155 Virgil v. Time, Inc., 527 F. 2d 1122 552 Virginia, Ex parte, 100 U.S. 339 16, 42 Virginia v. American Booksellers Assn., Inc., 484 U.S. 383 594 Virginia v. West Virginia, 238 U.S. 202 293 Virginia Pharmacy Bd. v. Virginia Citizens Consumer Council, Inc., 425 U.S. 748 531, 791 Virginia R. Co. v. Railway Employees, 300 U.S. 515 311 Vitek v. Jones, 445 U.S. 480 139 Page Vlandis v. Kline, 412 U.S. 441 120, 121, 151 Vulcan Society of Westchester Co., Inc. v. Fire Dept, of White Plains, 533 F. Supp. 1054 774 Wabash, S. L. & P. R. Co. v. Illinois, 118 U.S. 557 516 Walls v. Waste Resource Corp., 761 F. 2d 311 22 Walters v. National Assn, of Radiation Survivors, 473 U.S. 305 626, 644 Wardius v. Oregon, 412 U.S. 470 633 Washington v. Yakima Indian Nation, 439 U.S. 463 740 Waters v. Wisconsin Steel Works, 427 F. 2d 476 202 Watseka v. Illinois Public Action Council, 796 F. 2d 1547 799 Watson v. Buck, 313 U.S. 387 366 Watt v. Alaska, 451 U.S. 259 454, 510 Weinberger v. Salfi, 422 U.S. 749 121, 151, 153 Welch v. Texas Dept, of Highways and Public Transportation, 483 U.S. 468 15, 30, 33, 34,39,41,43,45,47,49, 56, 57, 66,172,233,240, 279, 743 Westmoreland v. CBS Inc., 596 F. Supp. 1170 666 West Virginia v. United States, 479 U.S. 305 293 West Virginia Bd. of Ed. v. Barnette, 319 U.S. 624 401,404, 405, 414-416, 418, 437 Wheat v. United States, 486 U.S. 153 624, 625, 634, 644, 651 Whitney v. California, 274 U.S. 357 419 Wickard v. Filburn, 317 U.S. Ill 20 Will v. Michigan Dept, of State Police, 491 U.S. 58 236, 724 Willcox v. Consolidated Gas Co., 212 U.S. 19 359 TABLE OF CASES CITED XXXI Page Williams v. Vermont, 472 U.S. 14 27 Wilson v. Omaha Indian Tribe, 442 U.S. 653 64, 72, 74, 83 Winship, In re, 397 U.S. 358 265, 266 Woodbridge v. Worcester State Hospital, 384 Mass. 38 62, 63 Woodward v. Commissioner, 397 U.S. 572 253 Wooley v. Maynard, 430 U.S. 705 372 Wyatt v. Security Inn Food & Beverage, Inc., 819 F. 2d 69 210 Wygant v. Jackson Bd. of Ed., 476 U.S. 267 798 Page Young, Ex parte, 209 U.S. 123 26, 34, 71, 278 Young v. American Mini Theatres, Inc., 427 U.S. 50 414,438 Young v. International Tel. & Tel. Co., 438 F. 2d 757 202 Younger v. Harris, 401 U.S. 37 27, 356, 358- 360, 364-370, 372-375 Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579 468 Zablocki v. Redhail, 434 U.S. 374 363 Zacchini v. Scripps-Howard Broadcasting Co., 433 U.S. 552 102 Zenith Radio Corp. v. United States, 437 U.S. 443 464 CASES ADJUDGED IN THE SUPREME COURT OF THE UNITED STATES AT OCTOBER TERM, 1988 PENNSYLVANIA v. UNION GAS CO. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT No. 87-1241. Argued October 31, 1988—Decided June 15, 1989 Respondent’s predecessors operated a coal gasification plant, which produced coal tar as a by-product, along a creek in Pennsylvania. Shortly after acquiring easements in the property along the creek, and while excavating to control flooding, the State struck a large deposit of coal tar which began to seep into the creek. Finding the tar to be a hazardous substance, the Environmental Protection Agency declared the site the Nation’s first Superfund site, and the State and the Federal Government together cleaned up the area. The Government reimbursed the State for cleanup costs and sued respondent to recoup those costs under §§ 104 and 106 of the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U. S. C. §§ 9604 and 9606, claiming that respondent was liable because it and its predecessors had deposited the tar in the ground. Respondent filed a third-party complaint against the State, asserting, inter alia, that it was liable as an “owner and operator” of the site under § 107(a) of CERCLA. The District Court dismissed this complaint on the ground that the State’s Eleventh Amendment immunity barred the suit. The Court of Appeals affirmed, finding no clear expression of intent to hold States liable in monetary damages under CERCLA. However, after this Court vacated that decision and remanded for reconsideration in light of subsequent amendments to CERCLA made by the Superfund Amendments and Reauthorization Act of 1986 (SARA), the Court of Appeals held that the statute’s amended language clearly rendered States liable for 1 2 OCTOBER TERM, 1988 Syllabus 491 U. S. monetary damages and that Congress had the power to do so under the Commerce Clause. Held: The judgment is affirmed, and the case is remanded. 832 F. 2d 1343, affirmed and remanded. Justice Brennan delivered the opinion of the Court with respect to Parts I and II, concluding that CERCLA, as amended by SARA, clearly expresses an intent to hold States liable in damages in federal court. Pp. 7-13. (a) The statute’s plain language authorizes such suits. Section 101(21)’s express inclusion of States within its definition of “persons,” and § 101(20)(D)’s plain statement that state and local governments are to be considered “owners or operators” in all but very narrow circumstances, together establish that Congress intended that States be liable for cleanup costs under § 107 along with everyone else responsible for creating hazardous waste sites. The fact that § 101(20)(D) uses language virtually identical to § 120(a)(l)’s waiver of the Federal Government’s sovereign immunity is highly significant, demonstrating that Congress must have intended to override the States’ immunity from suit. This conclusion is not contradicted by § 101(20)(D)’s exclusion of States from the category of “owners and operators” when they acquire ownership or control of a site involuntarily by virtue of their function as sovereign, by § 107(d)(2)’s general exemption of States from liability for actions taken during cleanup of contamination generated by other persons’ facilities, or by 42 U. S. C. § 9659(a)(l)’s express reservation of States’ Eleventh Amendment rights in citizen suits, since those provisions would be unnecessary unless suits against States were otherwise permitted by the statute. Pp. 7-10. (b) Pennsylvania’s arguments to the contrary are not persuasive. If accepted, the contention that CERCLA creates state liability only to the Federal Government would render meaningless the § 101(20)(D) language making States liable “to the same extent ... as any nongovernmental entity, including liability for [damages],” since no explicit authorization is necessary before the Federal Government may sue a State for damages. Moreover, § 101(20)(D) obviously explains and qualifies the entire definition of “owner or operator,” and does not, as Pennsylvania suggests, render States liable only if they acquire property involuntarily and then contribute to contamination there. Nor can it be decisive that § 101(20)(D) mentions local governments, which do not enjoy immunity, in the same breath as States, since it was natural for Congress to discuss governmental entities together. Pp. 11-13. Justice Brennan, joined by Justice Marshall, Justice Black-mun, and Justice Stevens, concluded in Part III that Congress has PENNSYLVANIA v. UNION GAS CO. 3 1 Syllabus the authority to render States liable for money damages in federal court when legislating pursuant to the Commerce Clause. Pp. 13-23. (a) This Court’s decisions indicate that Congress has the authority to override States’ immunity when legislating pursuant to the Commerce Clause. See, e. g., Parden v. Terminal Railway of Alabama Docks Dept., 377 U. S. 184; Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279. This conclusion is confirmed by a consideration of the special nature of the plenary power conferred by the Clause, which expands federal power by taking power away from the States. Cf. Fitzpatrick n. Bitzer, 427 U. S. 445, 454-456; Ex parte Virginia, 100 U. S. 339, 346. Pp. 14-19. (b) By giving Congress plenary authority to regulate commerce, the States relinquished their immunity where Congress finds it necessary, in exercising this authority, to render them liable. Since the commerce power can displace State regulation, a conclusion that Congress may not create a damages remedy against the States would sometimes mean that no one could do so. Indeed, this Court has recognized that the general problem of environmental harm is often not susceptible to a local solution. See Illinois v. Milwaukee, 406 U. S. 91; Philadelphia v. New Jersey, 437 U. S. 617. Moreover, in many situations, it is only money damages that will effectuate Congress’ legitimate Commerce Clause objectives. Here, for example, after failing to solve the hazardous-substances problem through preventive measures, Congress chose to extend liability to everyone potentially responsible for contamination, and, because of the enormous costs of cleanups and the finite nature of Government resources, sought to encourage private parties to help out by allowing them to recover for their own cleanup efforts. There is no merit to Pennsylvania’s contention that the allowance of damages suits by private citizens against unconsenting States impermissibly expands the jurisdiction of federal courts beyond the bounds of Article III, since, by ratifying the Constitution containing the Commerce Clause, the States consented to suits against them based on congressionally created causes of action. Cf. Fitzpatrick v. Bitzer, supra. Pp. 19-23. Justice White agreed with the plurality’s conclusion that Congress has the authority under Article I to abrogate the States’ Eleventh Amendment immunity, but disagreed with the reasoning supporting that conclusion. P. 57. Brennan, J., announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I and II, in which Marshall, Blackmun, Stevens, and Scalia, JJ., joined, and an opinion with respect to Part III, in which Marshall, Blackmun, and Stevens, JJ., joined. Stevens, J., filed a concurring opinion, post, p. 23. Scalia, J., 4 OCTOBER TERM, 1988 Syllabus 491 U. S. filed an opinion concurring in part and dissenting in part, in Parts II, III, and IV of which Rehnquist, C. J., and O’Connor and Kennedy, JJ., joined, post, p. 29. White, J., filed an opinion concurring in the judgment in part and dissenting in part, in Part I of which Rehnquist, C. J., and O’Connor and Kennedy, JJ., joined, post, p. 45. O’Connor, J., filed a dissenting opinion, post, p. 57. John G. Knorr III, Chief Deputy Attorney General of Pennsylvania, argued the cause for petitioner. With him on the briefs were LeRoy S. Zimmerman, Attorney General, and Gregory R. Neuhauser, Senior Deputy Attorney General. Robert A. Swift argued the cause for respondent. With him on the brief were Marguerite R. Goodman and Lawrence Demase.* *A brief of amici curiae urging reversal was filed for the State of New York et al. by Robert Abrams, Attorney General of New York, 0. Peter Sherwood, Solicitor General, Elaine Gail Suchman, Assistant Attorney General, John K. Van de Kamp, Attorney General of California, Clifford L. Rechtschaffen and J. Matthew Rodriquez, Deputy Attorneys General, Joseph I. Lieberman, Attorney General of Connecticut, Kenneth N. Tedford, Assistant Attorney General, Michael Bowers, Attorney General of Georgia, Barbara H. Gallo, Assistant Attorney General, Neil F. Hartigan, Attorney General of Illinois, Rosalyn Kaplan, Linley E. Pearson, Attorney General of Indiana, Harry John Watson III, Thomas J. Miller, Attorney General of Iowa, John P. Sarcone, Assistant Attorney General, Arthur L. Williams, J. Joseph Curran, Jr., Attorney General of Maryland, Andrew H. Baida, Richard M. Hall, and Michael C. Powell, Assistant Attorneys General, Hubert H. Humphrey III, Attorney General of Minnesota, John R. Tunheim, Chief Deputy Attorney General, William L. Webster, Attorney General of Missouri, Shelley A. Woods, Assistant Attorney General, W. Cary Edwards, Attorney General of New Jersey, John J. Maiorana, Deputy Attorney General, Hal Stratton, Attorney General of New Mexico, Alicia Mason, Assistant Attorney General, Lacy H. Thornburg, Attorney General of North Carolina, Gay IM. Manthei, Assistant Attorney General, Robert H. Henry, Attorney General of Oklahoma, Sara J. Drake, Assistant Attorney General, T. Travis Medlock, Attorney General of South Carolina, Walton J. McLeod III, Jacquelyn S. Dickman, W. J. Michael Cody, Attorney General of Tennessee, Michael W. Catalano, Deputy Attorney General, David L. Wilkinson, Attorney General of Utah, Fred G. Nelson, Assistant Attorney General, Jeffrey L. Amestoy, Attorney General of Vermont, Conrad W. Smith, Assistant PENNSYLVANIA v. UNION GAS CO. 5 1 Opinion of the Court Justice Brennan announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I and II, and an opinion with respect to Part III, in which Justice Marshall, Justice Blackmun, and Justice Stevens join. This case presents the questions whether the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U. S. C. §9601 et seq., as amended by the Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. 99-499, 100 Stat. 1613, permits a suit for monetary damages against a State in federal court and, if so, whether Congress has the authority to create such a cause of action when legislating pursuant to the Commerce Clause. The answer to both questions is “yes.” I For about 50 years, the predecessors of respondent Union Gas Co. operated a coal gasification plant near Brodhead Creek in Stroudsburg, Pennsylvania, which produced coal tar as a by-product. The plant was dismantled around 1950. A few years later, Pennsylvania took part in major flood-control efforts along the creek. In 1980, shortly after acquiring easements to the property along the creek, the Commonwealth struck a large deposit of coal tar while excavating the creek. The coal tar began to seep into the creek, and the Attorney General, Charles G. Brown, Attorney General of West Virginia, C. William Ullrich, First Deputy Attorney General, Donald J. Hanaway, Attorney General of Wisconsin, and Charles D. Hoomstra, Assistant Attorney General. Briefs of amici curiae urging affirmance were filed for the American Federation of Labor and Congress of Industrial Organizations by Robert M. Weinberg and Laurence Gold; for the Association of American Publishers, Inc., et al. by Jon Baumgarten, Christopher A. Meyer, and Charles S. Sims; for the Chemical Manufacturers Association by Neil J. King and Carol F. Lee; and for the National Music Publishers’ Association, Inc., et al. by Steven B. Rosenfeld. Ronald A. Zumbrun and Robin L. Rivett filed a brief for the Pacific Legal Foundation as amicus curiae. 6 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Environmental Protection Agency determined that the tar was a hazardous substance and declared the site the Nation’s first emergency Superfund site. Working together, Pennsylvania and the Federal Government cleaned up the area, and the Federal Government reimbursed the State for cleanup costs of $720,000. To recoup these costs, the United States sued Union Gas under §§104 and 106 of CERCLA, 42 U. S. C. §§9604 and 9606, claiming that Union Gas was liable for such costs because the company and its predecessors had deposited coal tar into the ground near Brodhead Creek. Union Gas filed a third-party complaint against Pennsylvania, asserting that the Commonwealth was responsible for at least a portion of the costs because it was an “owner or operator” of the hazardous-waste site, 42 U. S. C. § 9607(a), and because its flood-control efforts had negligently caused or contributed to the release of the coal tar into the creek. The District Court dismissed the complaint, accepting Pennsylvania’s claim that its Eleventh Amendment immunity barred the suit. A divided panel of the Court of Appeals for the Third Circuit affirmed, finding no clear expression of congressional intent to hold States liable in monetary damages under CERCLA. United States v. Union Gas Co., 792 F. 2d 372 (1986). While Union Gas’ petition for certiorari was pending, Congress amended CERCLA by passing SARA. We granted certiorari, vacated the Court of Appeals’ opinion, and remanded for reconsideration in light of these amendments. 479 U. S. 1025 (1987). On remand, the Court of Appeals held that the language of CERCLA, as amended, clearly rendered States liable for monetary damages and that Congress had the power to do so when legislating pursuant to the Commerce Clause. United States v. Union Gas Co., 832 F. 2d 1343 (1986). We granted certiorari, 485 U. S. 958 (1988), and now affirm. PENNSYLVANIA v. UNION GAS CO. 7 1 Opinion of the Court II In Hans n. Louisiana, 134 U. S. 1 (1890), this Court held that the principle of sovereign immunity reflected in the Eleventh Amendment rendered the States immune from suits for monetary damages in federal court even where jurisdiction was premised on the presence of a federal question. Congress may override this immunity when it acts pursuant to the power granted it under § 5 of the Fourteenth Amendment, but it must make its intent to do so “unmistakably clear.” See Atascadero State Hospital v. Scanlon, 473 U. S. 234, 242 (1985). Before turning to the question whether Congress possesses the same power of abrogation under the Commerce Clause, we must first decide whether CERCLA, as amended by SARA, clearly expresses an intent to hold States liable in damages for conduct described in the statute. If we decide that it does not, then we need not consider the constitutional question. CERCLA both provides a mechanism for cleaning up hazardous-waste sites, 42 U. S. C. §§9604, 9606 (1982 ed. and Supp. IV), and imposes the costs of the cleanup on those responsible for the contamination, §9607. Two general terms, among others, describe those who may be liable under CERCLA for the costs of remedial action: “persons” and “owners or operators.” §9607(a). “States” are explicitly included within the statute’s definition of “persons.” § 9601(21). The term “owner or operator” is defined by reference to certain activities that a “person” may undertake. § 9601(20)(A). Section 101(20)(D) of SARA excludes from the category of “owners or operators” States that “acquired ownership or control involuntarily through bankruptcy, tax delinquency, abandonment, or other circumstances in which the government involuntarily acquires title by virtue of its function as 8 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. sovereign.” §9601(20)(D).1 However, §101(20)(D) continues, “[t]he exclusion provided under this paragraph shall not apply to any State or local government which has caused or contributed to the release or threatened release of a hazardous substance from the facility, and such a State or local government shall be subject to the provisions of this chapter in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under section 9607 of this title.” Ibid. The express inclusion of States within the statute’s definition of “persons,” and the plain statement that States are to be considered “owners or operators” in all but very narrow circumstances, together convey a message of unmistakable clarity: Congress intended that States be liable along with everyone else for cleanup costs recoverable under CERCLA. Section 101(20)(D) is an express acknowledgment of Congress’ background understanding—evidenced first in its inclusion of States as “persons”—that States would be liable in any circumstance described in § 107(a) from which they were not expressly excluded. The “exclusion” furnished to the States in § 101(20)(D) would be unnecessary unless such a background understanding were at work.2 1 Section 101(20)(D), as set forth in 42 U. S. C. 9601(20)(D), provides in full: “(D) The term ‘owner or operator’ does not include a unit of State or local government which acquired ownership or control involuntarily through bankruptcy, tax delinquency, abandonment, or other circumstances in which the government involuntarily acquires title by virtue of its function as sovereign. The exclusion provided under this paragraph shall not apply to any State or local government which has caused or contributed to the release or threatened release of a hazardous substance from the facility, and such a State or local government shall be subject to the provisions of this chapter in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under section 9607 of this title.” 2Justice White’s attack on the notion that the definition of the word “persons,” standing alone, abrogates the States’ immunity from suit, see post, at 46-50, is directed at an argument that we do not make. We do not say that CERCLA’s definition of “persons” alone overrides the States’ im- PENNSYLVANIA v. UNION GAS CO. 9 1 Opinion of the Court The plain language of another section of the statute reinforces this conclusion. Section 107(d)(2) of CERCLA, as set forth in 42 U. S. C. § 9607(d)(2) (1982 ed., Supp. IV), headed “State and local governments,” provides: “No State or local government shall be liable under this subchapter for costs or damages as a result of actions taken in response to an emergency created by the release or threatened release of a hazardous substance generated by or from a facility owned by another person. This paragraph shall not preclude liability for costs or damages as a result of gross negligence or inten- munity, but instead read CERCLA and SARA together, and argue that SARA’s wording must inform our understanding of the other definitional sections of the statute. The failure to appreciate this point leads to four mistakes. First, in his “judicial headcount,” post, at 46-47, Justice White counts the votes as to the wrong statute. The judges who ruled that CERCLA did not render States liable did so when they considered the unamended version of CERCLA; as to CERCLA as amended by SARA, the three-judge panel unanimously agreed that it clearly abrogated the States’ immunity. (This headcounting approach is flawed for another, more fundamental reason: surely judges can disagree about the content and rigor of the standard of “unmistakable clarity,” and if they do, they are likely to reach different results on States’ amenability to suit for reasons having nothing to do with the statutory language itself.) Second, Justice White asserts that our reading of CERCLA is inconsistent with the Court’s conclusion in Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279 (1973), that a statute literally including the States as “persons” subject to the statute was not clear enough to abrogate the States’ immunity. Post, at 48-49. This claim ignores SARA’s more specific language. Third, Justice White claims that our reading of CERCLA renders § 107(g)—which overrides the United States’ sovereign immunity from suit—redundant. Post, at 47. However, since we do not argue here that the inclusion of the States and the Federal Government in § 101(21)’s definition of “persons,” standing alone, overrides these entities’ immunity, our position does not make § 107(g) superfluous. Finally, only a failure to recognize that we rely on § 101(21) and § 101(20) (D) in combination could lead to the suggestion that States would enjoy § 101(20)(D)’s more favorable standard of liability even if they voluntarily acquired a site. Post, at 53, n. 5. 10 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. tional misconduct by the State or local government.” This section is, needless to say, an explicit recognition of the potential liability of States under this statute; Congress need not exempt States from liability unless they would otherwise be liable. Similarly, unless suits against the States were elsewhere permitted, Congress would have had no reason to specify that citizen suits—as opposed to the kind of lawsuit involved here—could be brought “against any person (including the United States and any other governmental instrumentality or agency, to the extent permitted by the eleventh amendment to the Constitution).” 42 U. S. C. § 9659(a)(1). The reservation of States’ rights under the Eleventh Amendment would be unnecessary if Congress had not elsewhere in the statute overridden the States’ immunity from suit. It is also highly significant that, in § 101(20)(D), Congress used language virtually identical to that it chose in waiving the Federal Government’s immunity from suits for damages under CERCLA. Section 120(a)(1) of CERCLA, as set forth in 42 U. S. C. § 9620(a)(1), provides: “Each department, agency, and instrumentality of the United States (including the executive, legislative, and judicial branches of government) shall be subject to, and comply with, this chapter in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under section 9607 of this title.” This is doubtless an “ ‘unequivocal] express[ion]’” of the Federal Government’s waiver of its own sovereign immunity, United States v. Testan, 424 U. S. 392, 399 (1976), quoting United States v. King, 395 U. S. 1,4 (1969), since we cannot imagine any other plausible explanation for this unqualified language. It can be no coincidence that in describing the potential liability of the States in § 101(20)(D), Congress chose language mirroring that of § 120(a)(1). In choosing this mirroring language in § 101(20)(D), therefore, Congress must have intended to override the States’ immunity from suit, just as it waived the Federal Government’s immunity in § 120(a)(1). PENNSYLVANIA v. UNION GAS CO. 11 1 Opinion of the Court This cascade of plain language does not, however, impress Pennsylvania. In the face of such clarity, the Commonwealth bravely insists that CERCLA merely makes clear that States may be liable to the United States, not that they may be liable to private entities such as Union Gas. The Commonwealth relies principally on this Court’s decision in Employees n. Missouri Dept, of Public Health and Welfare, 411 U. S. 279 (1973). We held there that Congress had not abrogated the States’ immunity from suit in the Fair Labor Standards Act. Nevertheless, we found, the statute’s explicit inclusion of state-run hospitals among those to whom the law would apply was not meaningless: since the statute allowed the United States to sue, the inclusion of States within the entities covered by the statute served to permit suits by the United States against the States. Id., at 285-286. Although it is true that the inclusion of States within CERCLA’s definition of “persons” would not be rendered meaningless if we held that CERCLA did not subject the States to suits brought by private citizens, it is equally certain that such a holding would deprive the last portion of § 101(20)(D) of all meaning. Congress would have had no cause to stress that States would be liable “to the same extent ... as any nongovernmental entity,” § 101(20)(D), if it had meant only that they could be liable to the United States. In United States v. Mississippi, 380 U. S. 128, 140-141 (1965), we recognized that the Constitution presents no barrier to lawsuits brought by the United States against a State. For purposes of such lawsuits, States are naturally just like “any nongovernmental entity”; there are no special rules dictating when they may be sued by the Federal Government, nor is there a stringent interpretive principle guiding construction of statutes that appear to authorize such suits. Indeed, this Court has gone so far as to hold that no explicit statutory authorization is necessary before the Federal Government may sue a State. See United States v. California, 12 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. 332 U. S. 19, 26-28 (1947). Unless Congress intended to permit suits brought by private citizens against the States, therefore, the highly specific language of §101(20)(D) was unnecessary. The same can be said about the clause of § 101(20)(D) specifying that States would be subject to CERCLA’s provisions, “including liability under section 9607 of this title.” Section 9607 provides for liability in damages, and liability in damages is considered a special remedy, requiring special statutory language, only where the States’ immunity from suits by private citizens is involved. In light of § 101(20)(D)’s very precise language, it would be exceedingly odd to interpret this provision as merely a signal that the United States — rather than private citizens—could sue the States for damages under CERCLA.3 Moreover, § 101(20)(D) does not, as Pennsylvania suggests, render States liable only if they acquire property involuntarily and then contribute to a release of harmful substances at that property. Section 101(20)(D) obviously explains and qualifies the entire definition of “owner or operator”—not 3 Justice White’s response to this point is unconvincing. After claiming that our reading renders a part of the statute redundant—an accusation without merit, see n. 2, supra—Justice White resorts to a reading of § 101(20)(D) that, he admits, renders the phrase “as any nongovernmental entity” superfluous. Post, at 55, n. 6. To say that this phrase can be explained as a “statutory ‘exclamation point,’ ” post, at 54-55, n. 6, is just another way of describing redundancy. Nor is it possible to explain this passage as an effort to pre-empt state-law immunity for local governments. See post, at 55, n. 6. Given our recognition that “there is no tradition of immunity for municipal corporations,” Owen v. City of Independence, 445 U. S. 622, 638 (1980), and our refusal in the past to allow state-law immunities to define the scope of federal statutes, see, e. g., Monell v. New York City Dept, of Social Services, 436 U. S. 658, 695, n. 59 (1978), Congress would see no need to use emphatic language to override this kind of immunity. Unless we conclude, therefore, that the phrase “as any nongovernmental entity” is superfluous, this clause demonstrates that § 101(20)(D) was designed to do more than render the States liable in damages to the Federal Government. PENNSYLVANIA v. UNION GAS CO. 13 1 Opinion of Brennan, J. just that part of the definition applicable to involuntary owners. Nor can it be decisive that § 101(20)(D) mentions local governments as well as States. The Commonwealth argues that, because local governments do not enjoy immunity from suit, § 101(20)(D)’s reference to local governments means that the section shows no intent to abrogate States’ immunity. It was natural, however, for Congress to describe the potential liability of States and local governments in the same breath, since both are governmental entities and both enjoy special exemptions from liability under CERCLA. See §§ 101(20)(D), 107(d)(2). Pennsylvania also argues that § 101(20)(D) demonstrates no intent to hold the States liable because this provision limits the States’ liability. It is true that this section rescues the States from liability where they obtained ownership of cleanup sites involuntarily. The Commonwealth fails to grasp, however, that a limitation of liability is nonsensical unless liability existed in the first place. We thus hold that the language of CERCLA as amended by SARA clearly evinces an intent to hold States liable in damages in federal court.4 Ill Our conclusion that CERCLA clearly permits suits for money damages against States in federal court requires us to decide whether the Commerce Clause grants Congress the power to enact such a statute. Pennsylvania argues that the principle of sovereign immunity found in the Eleventh 4 The language of the Rehabilitation Act Amendments of 1986, Pub. L. 99-506, 100 Stat. 1807, is indeed more pointed on the subject of abrogation than is CERCLA, since it mentions the Eleventh Amendment by name. See post, at 56, n. 7. It is surprising that Justice White’s opinion lays so much stress on this difference in wording, however, because it expressly disclaims any intent to require that the words “Eleventh Amendment” appear in a statute in order to find abrogation. Ibid. If no magic words are required for abrogation, then each statute must be evaluated on its own terms, not defeated by reference to another statute that uses more specific language. 14 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. Amendment precludes such congressional authority. We do not agree. A Though we have never squarely resolved this issue of congressional power, our decisions mark a trail unmistakably leading to the conclusion that Congress may permit suits against the States for money damages. The trail begins with Parden v. Terminal Railway of Alabama Docks Dept., 377 U. S. 184 (1964). There, in responding to a state-owned railway’s argument that Congress had no authority to subject the railway to suit, we concluded that “the States surrendered a portion of their sovereignty when they granted Congress the power to regulate commerce,” id., at 191, and that “[b]y empowering Congress to regulate commerce, . . . the States necessarily surrendered any portion of their sovereignty that would stand in the way of such regulation,” id., at 192. Although it is true that we have referred to Parden as a case involving a waiver of immunity, Fitzpatrick n. Bitzer, 427 U. S. 445, 451 (1976), the statements quoted above lay a firm foundation for the argument that Congress’ authority to regulate commerce includes the authority directly to abrogate States’ immunity from suit. The path continues in Employees n. Missouri Dept, of Public Health and Welfare, 411 U. S., at 286, in which we again acknowledged, quoting Parden, that “‘the States surrendered a portion of their sovereignty when they granted Congress the power to regulate commerce.’” Although we declined “to extend Parden to cover every exercise by Congress of its commerce power,” we did so in Employees itself only because “the purpose of Congress to give force to the Supremacy Clause by lifting the sovereignty of the States and putting the States on the same footing as other employers [was] not clear.” 411 U. S., at 286-287. Employees1 message is plain: the power to regulate commerce includes the power to override States’ immunity from suit, but we will PENNSYLVANIA v. UNION GAS CO. 15 1 Opinion of Brennan, J. not conclude that Congress has overridden this immunity unless it does so clearly. Since Employees, we have twice assumed that Congress has the authority to abrogate States’ immunity when acting pursuant to the Commerce Clause. See Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468, 475-476, and n. 5 (1987); County of Oneida n. Oneida Indian Nation of New York, 470 U. S. 226, 252 (1985). See also Green v. Mansour, 474 U. S. 64, 68 (1985) (“States may not be sued in federal court... unless Congress, pursuant to a valid exercise of power, unequivocally expresses its intent to abrogate the immunity”); Quern v. Jordan, 440 U. S. 332, 343 (1979) (referring to congressional power recognized in Employees as power “to abrogate Eleventh Amendment immunity”). It is no accident, therefore, that every Court of Appeals to have reached this issue has concluded that Congress has the authority to abrogate States’ immunity from suit when legislating pursuant to the plenary powers granted it by the Constitution. See, e. g., United States v. Union Gas Co., 832 F. 2d 1343 (CA3 1987) (case below); In re McVey Trucking, Inc., 812 F. 2d 311 (CA7), cert, denied, 484 U. S. 895 (1987); County of Monroe v. Florida, 678 F. 2d 1124 (CA2 1982), cert, denied, 459 U. S. 1104 (1983); Peel v. Florida Dept, of Transportation, 600 F. 2d 1070 (CA5 1979); Mills Music, Inc. v. Arizona, 591 F. 2d 1278 (CA9 1979). Even if we never before had discussed the specific connection between Congress’ authority under the Commerce Clause and States’ immunity from suit, careful regard for precedent still would mandate the conclusion that Congress has the power to abrogate immunity when exercising its plenary authority to regulate interstate commerce. In Fitzpatrick v. Bitzer, supra, we held that Congress may subject States to suits for money damages in federal court when legislating under § 5 of the Fourteenth Amendment, and further held that Congress had done so in the 1972 Amendments to Title VII of the Civil Rights Act of 1964. Subsequent cases 16 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. hold firmly to the principle that Congress can override States’ immunity under §5. See, e. g., Dellmuth v. Muth, post, p. 223; Atascadero State Hospital v. Scanlon, 473 U. S., at 238; Pennhurst State School and Hospital v. Halderman, 465 U. S. 89, 99 (1984); Quern v. Jordan, supra. Fitzpatrick’s rationale is straightforward: “When Congress acts pursuant to § 5, not only is it exercising legislative authority that is plenary within the terms of the constitutional grant, it is exercising that authority under one section of a constitutional Amendment whose other sections by their own terms embody limitations on state authority.” 427 U. S., at 456. In so reasoning, we emphasized the “shift in the federal-state balance” occasioned by the Civil War Amendments, id., at 455, and in particular quoted extensively from Ex parte Virginia, 100 U. S. 339 (1880). The following passage from Ex parte Virginia is worth quoting here as well: “Such enforcement [of the prohibitions of the Fourteenth Amendment] is no invasion of State sovereignty. No law can be, which the people of the States have, by the Constitution of the United States, empowered Congress to enact. . . . [I]n exercising her rights, a State cannot disregard the limitations which the Federal Constitution has applied to her power. Her rights do not reach to that extent. Nor can she deny to the general government the right to exercise all its granted powers, though they may interfere with the full enjoyment of rights she would have if those powers had not been thus granted. Indeed, every addition of power to the general government involves a corresponding diminution of the governmental powers of the States. It is carved out of them.” Id., at 346, quoted in Fitzpatrick, supra, at 454-455. Each of these points is as applicable to the Commerce Clause as it is to the Fourteenth Amendment. Like the Fourteenth Amendment, the Commerce Clause with one hand gives power to Congress while, with the other, it takes power away from the States. It cannot be relevant that the Four- PENNSYLVANIA v. UNION GAS CO. 17 1 Opinion of Brennan, J. teenth Amendment accomplishes this exchange in two steps (§§ 1-4, plus §5), while the Commerce Clause does it in one. The important point, rather, is that the provision both expands federal power and contracts state power; that is the meaning, in fact, of a “plenary” grant of authority, and the lower courts have rightly concluded that it makes no sense to conceive of § 5 as somehow being an “ultraplenary” grant of authority. See, e. g., In re McVey Trucking, supra, at 316. See also Quern, supra, at 343 (distinguishing Employees (Commerce Clause) from Fitzpatrick (§5) only by reference to the clarity of the congressional intent expressed in the relevant statutes). Pennsylvania attempts to bring this case outside Fitzpatrick by asserting that “[t]he Fourteenth Amendment . . . alters what would otherwise be the proper constitutional balance between federal and state governments.” Brief for Petitioner 39. The Commonwealth believes, apparently, that the “constitutional balance” existing prior to the Fourteenth Amendment did not permit Congress to override the States’ immunity from suit. This claim, of course, begs the very question we face. For its part, Justice Scalia’s opinion casually announces: “Nothing in [Fitzpatrick’s] reasoning justifies limitation of the principle embodied in the Eleventh Amendment through appeal to antecedent provisions of the Constitution.” Post, at 42. The operative word here is, it would appear, “antecedent”; and it is important to emphasize that, according to Justice Scalia, the Commerce Clause is antecedent, not to the Eleventh Amendment, but to “the principle embodied in the Eleventh Amendment.” But, according to Part II of Justice Scalia’s opinion, this “principle” has been with us since the days before the Constitution was ratified—since the days, in other words, before the Commerce Clause. In describing the “consensus that the doctrine of sovereign immunity . . . was part of the understood background against which the Constitution was adopted, and which its jurisdic 18 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. tional provisions did not mean to sweep away,” post, at 31-32, Justice Scalia clearly refers to a state of affairs that existed well before the States ratified the Constitution. Justice Scalia, therefore, has things backwards: it is not the Commerce Clause that came first, but “the principle embodied in the Eleventh Amendment” that did so. Antecedence takes this case closer to, not further from, Fitzpatrick. Even if “the principle embodied in the Eleventh Amendment” made its first appearance at the same moment as the Commerce Clause, and not before, Justice Scalia could no longer rely on chronology in distinguishing Fitzpatrick. Only if it were the Eleventh Amendment itself that introduced the principle of sovereign immunity into the Constitution would the Commerce Clause have preceded this principle. Even then, the order of events would matter only if the Amendment changed things; that is, it would matter only if, before the Eleventh Amendment, the Commerce Clause did authorize Congress to abrogate sovereign immunity. But if Congress enjoyed such power prior to the enactment of this Amendment, we would require a showing far more powerful than Justice Scalia can muster that the Amendment was intended to obliterate that authority. The language of the Eleventh Amendment gives us no hint that it limits congressional authority; it refers only to “the judicial power” and forbids “constru[ing]” that power to extend to the enumerated suits —language plainly intended to rein in the Judiciary, not Congress. It would be a fragile Constitution indeed if subsequent amendments could, without express reference, be interpreted to wipe out the original understanding of congressional power. Justice Scalia attempts to avoid the pull of our prior decisions by claiming that Hans answered this constitutional question over 100 years ago. Because Hans was brought into federal court via the Judiciary Act of 1875 and because the Court there held that the suit was barred by the Eleventh Amendment, Justice Scalia argues, that case disposed PENNSYLVANIA v. UNION GAS CO. 19 1 Opinion of Brennan, J. of the question whether Congress has the authority to abrogate States’ immunity when legislating pursuant to the powers granted it by the Constitution. See post, at 36-37. This argument depends on the notion that, in passing the Judiciary Act, “Congress . . . sought to eliminate [the] state sovereign immunity” that Article III had not eliminated. Post, at 36 (emphasis in original). As Justice Scalia is well aware, however, the Judiciary Act merely gave effect to the grant of federal-question jurisdiction under Article III, which was not self-executing. Thus, if Article III did not “automatically eliminate” sovereign immunity, see post, at 33, then neither did the Judiciary Act of 1875. That unsurprising conclusion does not begin to address the question whether other congressional enactments, not designed simply to implement Article Ill’s grants of jurisdiction, may override States’ immunity. When one recalls, in addition, our conclusion that “Art[icle] III ‘arising under’ jurisdiction is broader than federal-question jurisdiction under § 1331,” Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480, 495 (1983), Justice Scalia’s conception of Hans’ holding looks particularly exaggerated. Our prior cases thus indicate that Congress has the authority to override States’ immunity when legislating pursuant to the Commerce Clause. This conclusion is confirmed by a consideration of the special nature of the power conferred by that Clause. B We have recognized that the States enjoy no immunity where there has been “‘a surrender of this immunity in the plan of the convention.’” Monaco v. Mississippi, 292 U. S. 313, 322-323 (1934), quoting The Federalist No. 81, p. 657 (H. Dawson ed. 1876) (A. Hamilton). Because the Commerce Clause withholds power from the States at the same time as it confers it on Congress, and because the congressional power thus conferred would be incomplete without the authority to render States liable in damages, it must be that, to the ex 20 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. tent that the States gave Congress the authority to regulate commerce, they also relinquished their immunity where Congress found it necessary, in exercising this authority, to render them liable. The States held liable under such a congressional enactment are thus not “unconsenting”; they gave their consent all at once, in ratifying the Constitution containing the Commerce Clause, rather than on a case-by-case basis. It would be difficult to overstate the breadth and depth of the commerce power. See, e. g., NLRB v. Jones & Laughlin Steel Corp., 301 U. S. 1 (1937); Wickard v. Filbum, 317 U. S. Ill, 127-128 (1942); Katzenbach v. McClung, 379 U. S. 294 (1964). It is not the vastness of this power, however, that is so important here: it is its effect on the power of the States. The Commerce Clause, we long have held, displaces state authority even where Congress has chosen not to act, see Gibbons v. Ogden, 9 Wheat. 1 (1824); Missouri Pacific R. Co. v. Stroud, 267 U. S. 404, 408 (1925); Northwest Central Pipeline Corp. v. State Corp. Comm’n of Kansas, 489 U. S. 493 (1989), and it sometimes precludes state regulation even though existing federal law does not pre-empt it, see Philadelphia n. New Jersey, 437 U. S. 617, 621, n. 4, 628-629 (1978); Northwest Central Pipeline Corp., supra. Since the States may not legislate at all in these last two situations, a conclusion that Congress may not create a cause of action for money damages against the States would mean that no one could do so. And in many situations, it is only money damages that will carry out Congress’ legitimate objectives under the Commerce Clause. The case before us brilliantly illuminates these points. The general problem of environmental harm is often not susceptible of a local solution. See Illinois v. Milwaukee, 406 U. S. 91 (1972) (recognizing authority of federal courts to create federal “common law” of nuisance to apply to interstate water pollution, displacing state nuisance laws). We have, in fact, invalidated one State’s effort to deal with the problem PENNSYLVANIA v. UNION GAS CO. 21 1 Opinion of Brennan, J. of waste disposal on a local level. See Philadelphia v. New Jersey, supra. N New Jersey statute prohibited the treatment and disposal, within the State, of any solid or liquid wastes generated outside the State. Indicating that a law applicable to all wastes would have survived under the Commerce Clause, id., at 626, we held that the exemption of locally produced wastes doomed the statute, id., at 626-629. As a practical matter, however, it is difficult to imagine that a State could forbid the disposal of all wastes. Hence, the Commerce Clause as interpreted in Philadelphia v. New Jersey ensures that we often must look to the Federal Government for environmental solutions. And often those solutions, to be satisfactory, must include a cause of action for money damages. The cause of action under consideration, for example, came about only after Congress had tried to solve the problem posed by hazardous substances through other means. Prior statutes such as the Resource Conservation and Recovery Act of 1976, 90 Stat. 2796, as amended, 42 U. S. C. §6901 et seq., had failed in large part because they focused on preventive measures to the exclusion of remedial ones. See Note, Superfund and California’s Implementation: Potential Conflict, 19 C. W. L. R. 373, 376, n. 23 (1983). The remedy that Congress felt it needed in CERCLA is sweeping: everyone who is potentially responsible for hazardous-waste contamination may be forced to contribute to the costs of cleanup. See, e. g., 42 U. S. C. § 9613(f)(1) (1986 ed., Supp. IV). Congress did not think it enough, moreover, to permit only the Federal Government to recoup the costs of its own cleanups of hazardous-waste sites; the Government’s resources being finite, it could neither pay up front for all necessary cleanups nor undertake many different projects at the same time. Some help was needed, and Congress sought to encourage that help by allowing private parties who voluntarily cleaned up hazardous-waste sites to recover a proportionate amount of the costs of cleanup from the other 22 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. potentially responsible parties. See ibid.; Mardan Corp. n. C. G. C. Music, Ltd., 804 F. 2d 1454, 1457, n. 3 (CA9 1986); Walls v. Waste Resource Corp., 761 F. 2d 311, 318 (CA6 1985). If States, which comprise a significant class of owners and operators of hazardous-waste sites, see Brief for Respondent 8, need not pay for the costs of cleanup, the overall effect on voluntary cleanups will be substantial. This case thus shows why the space carved out for federal legislation under the commerce power must include the power to hold States financially accountable not only to the Federal Government, but to private citizens as well. It does not follow that Congress, pursuant to its authority under the Commerce Clause, could authorize suits in federal court that the bare terms of Article III would not permit. No one suggests that if the Commerce Clause confers on Congress the power of abrogation, it must also confer the power to direct that certain state-law suits (not falling under the diversity jurisdiction) be brought in federal court. According to Pennsylvania, however, to decide that Congress may permit suits against States for money damages in federal court is equivalent to holding that Congress may expand the jurisdiction of the federal courts beyond the bounds of Article III. Pennsylvania argues that the federal judicial power as set forth in Article III does not extend to any suits for damages brought by private citizens against unconsenting States. See Brief for Petitioner 35-36, quoting Ex parte New York, 256 U. S. 490, 497 (1921) C“[T]he entire judicial power granted by the Constitution does not embrace authority to entertain a suit brought by private parties against a State without consent given’”). We never have held, however, that Article III does not permit such suits where the States have consented to them. Pennsylvania’s argument thus is answered by our conclusion that, in approving the commerce power, the States consented to suits against them based on congressionally created causes of action. Its claim also is answered by Fitzpatrick n. Bitzer, 427 U. S. 445 PENNSYLVANIA v. UNION GAS CO. 23 1 Stevens, J., concurring (1976) . The Fourteenth Amendment does not purport to expand or even change the scope of Article III. If Pennsylvania were right about the limitations on Article III, then our holding in Fitzpatrick would mean that the Fourteenth Amendment, though silent on the subject, expanded the judicial power as originally conceived. We do not share that view of Fitzpatrick.5 IV We hold that CERCLA renders States liable in money damages in federal court, and that Congress has the authority to render them so liable when legislating pursuant to the Commerce Clause. Given our ruling in favor of Union Gas, we need not reach its argument that Hans v. Louisiana, 134 U. S. 1 (1890), should be overruled. We affirm the judgment of the Court of Appeals for the Third Circuit and remand the case for further proceedings consistent with this opinion. It is so ordered. Justice Stevens, concurring. It is important to emphasize the distinction between our two Eleventh Amendments. There is first the correct and literal interpretation of the plain language of the Eleventh Amendment that is fully explained in Justice Brennan’s dissenting opinion in Atascadero State Hospital n. Scanlon, 473 U. S. 234, 247 (1985). In addition, there is the defense of sovereign immunity that the Court has added to the text of the Amendment in cases like Hans v. Louisiana, 134 U. S. 1 (1890). With respect to the former—the legitimate scope of the Eleventh Amendment limitation on federal judicial power—I do not believe Congress has the power under the 5 Since Union Gas itself eschews reliance on the theory of waiver we announced in Parden v. Terminal Railway of Alabama Docks Dept., 377 U. S. 184 (1964), see Brief for Respondent 31, we neither discuss this theory here nor understand why Justice Scalia feels the need to do so. See post, at 42-44. 24 OCTOBER TERM, 1988 Stevens, J., concurring 491 U. S. Commerce Clause, or under any other provision of the Constitution, to abrogate the States’ immunity. A statute cannot amend the Constitution. With respect to the latter—the judicially created doctrine of state immunity even from suits alleging violation of federally protected rights—I agree that Congress has plenary power to subject the States to suit in federal court. Because Justice Brennan’s opinion in Atascadero and the works of numerous scholars1 have exhaustively and conclusively refuted the contention that the Eleventh Amendment embodies a general grant of sovereign immunity to the States, further explication on this point is unnecessary. Suffice it to say that the Eleventh Amendment carefully mirrors the language of the citizen-state and alien-state diversity clauses of Article III and only provides that “[t]he Judicial power of the United States shall not be construed to extend” to these cases. There is absolutely nothing in the text of the Amendment that in any way affects the other grants of “judicial Power” contained in Article HI.2 Plainer language is seldom, if ever, found in constitutional law. ^ee, e. g., Marshall, Fighting the Words of the Eleventh Amendment, 102 Harv. L. Rev. 1342 (1989); Jackson, The Supreme Court, the Eleventh Amendment, and State Sovereign Immunity, 98 Yale L. J. 1 (1988); Amar, Of Sovereignty and Federalism, 96 Yale L. J. 1425 (1987); Lee, Sovereign Immunity and the Eleventh Amendment: The Uses of History, 18 Urb. Law. 519 (1986); Shapiro, Wrong Turns: The Eleventh Amendment and the Pennhurst Case, 98 Harv. L. Rev. 61 (1984); Gibbons, The Eleventh Amendment and State Sovereign Immunity: A Reinterpretation, 83 Colum. L. Rev. 1889 (1983); Fletcher, A Historical Interpretation of the Eleventh Amendment: A Narrow Construction of an Affirmative Grant of Jurisdiction Rather than a Prohibition Against Jurisdiction, 35 Stan. L. Rev. 1033 (1983); Tribe, Intergovernmental Immunities in Litigation, Taxation, and Regulation: Separation of Powers Issues in Controversies About Federalism, 89 Harv. L. Rev. 682 (1976). 2 The Eleventh Amendment asserts: “The Judicial power of the United States shall not be construed to extend to any suit in law or equity, commenced or prosecuted against one of the PENNSYLVANIA v. UNION GAS CO. 25 1 Stevens, J., concurring In Hans n. Louisiana, supra, however, the Court departed from the plain language, purpose, and history of the Eleventh Amendment, extending to the States immunity from suits premised on the “arising under” jurisdictional grant of Article III. Later adjustments to this rule, as well as the Court’s inability to develop a coherent doctrine of Eleventh Amendment immunity, make clear that this expansion of state immunity is not a matter of Eleventh Amendment law at all, but rather is based on a prudential interest in federal-state comity and a concern for “Our Federalism.” The Eleventh Amendment, as does Article III, speaks in terms of “judicial power.” The question that must therefore animate the inquiry in any actual Eleventh Amendment case is whether the federal court has power to entertain the suit. In cases in which there is no such power, Congress cannot provide it—even through a “clear statement.” Many of this Court’s decisions, however, purporting to apply the Eleventh Amendment, do not deal with judicial power at all. Instead, the issue of immunity is treated as a question of the proper role of the federal courts in the amalgam of federal-state relations. It is in these cases that congressional abrogation is appropriate. Several of this Court’s decisions make clear that much of our state immunity doctrine has absolutely nothing to do with the limit on judicial power contained in the Eleventh Amendment. For example, it is well established that a State may waive its immunity, subjecting itself to possible suit in federal court. See Atascadero, 473 U. S., at 238; Parden v. Terminal Railway of Alabama Docks Dept., 377 U. S. 184, 186 (1964); Employees v. Missouri Dept, of Public Health United States by Citizens of another State, or by Citizens or Subjects of any Foreign State.” This language parallels Article III, which provides in pertinent part: “The judicial Power shall extend ... to Controversies . . . between a State and Citizens of another State . . . and between a State . . . and foreign . . . Citizens or Subjects.” 26 OCTOBER TERM, 1988 Stevens, J., concurring 491 U. S. and Welfare, 411 U. S. 279, 284 (1973); Clark v. Barnard, 108 U. S. 436, 447-448 (1883). Yet, the cases are legion holding that a party may not waive a defect in subject-matter jurisdiction or invoke federal jurisdiction simply by consent. See, e. g., Owen Equipment & Erection Co. v. Kroger, 437 U. S. 365, 377, n. 21 (1978); Sosna n. Iowa, 419 U. S. 393, 398 (1975); California v. LaRue, 409 U. S. 109, 112, n. 3 (1972); American Fire & Casualty Co. n. Finn, 341 U. S. 6, 17-18, and n. 17 (1951); Mitchell v. Maurer, 293 U. S. 237, 244 (1934); Jackson n. Ashton, 8 Pet. 148, 149 (1834). This must be particularly so in cases in which the federal courts are entirely without Article III power to entertain the suit. Our willingness to allow States to waive their immunity thus demonstrates that this immunity is not a product of the limitation of judicial power contained in the Eleventh Amendment. Another striking example of the application of prudential— rather than true jurisdictional—concerns is found in our decision in Edelman v. Jordan, 415 U. S. 651 (1974). There, the Court inexplicably limited the fiction established in Ex parte Young, 209 U. S. 123 (1908), which permits suits against state officials in their official capacities for ultra vires acts, and concluded that the Young fiction only applies to prospective grants of relief. If Edelman simply involved an application of the limitation on judicial power contained in the Eleventh Amendment, once judicial power was found to exist to award prospective relief (even at some monetary cost to the State, see, e. g., Milliken v. Bradley, 433 U. S. 267 (1977)), it is difficult to understand why that same judicial power would not extend to award other forms of relief. See Fitzpatrick n. Bitzer, 427 U. S. 445, 459 (1976) (Stevens, J., concurring in judgment). In Pennhurst State School and Hospital n. Halderman, 465 U. S. 89, 104-106 (1984), the Court made explicit what was implicit in Edelman: the Young fiction “rests on the need to promote the vindication of federal rights,” while Edelman represents an attempt to “accommodate” this protection to the “competing inter- PENNSYLVANIA v. UNION GAS CO. 27 1 Stevens, J., concurring est” in “the constitutional immunity of the States.” Similarly, in Green v. Mansour, 474 U. S. 64, 68 (1985), the Court explained: “Both prospective and retrospective relief implicate Eleventh Amendment concerns, but the availability of prospective relief of the sort awarded in Ex parte Young gives life to the Supremacy Clause. Remedies designed to end a continuing violation of federal law are necessary to vindicate the federal interest in assuring the supremacy of that law. But compensatory or deterrence interests are insufficient to overcome the dictates of the Eleventh Amendment.” (Citations omitted.) The theme that thus emerges from cases such as Edelman, Pennhurst, and Green is one of balancing of state and federal interests. This sort of balancing, however, like waiver, is antithetical to traditional understandings of Article III subject-matter jurisdiction—either the judicial power extends to a suit brought against a State or it does not. See National Mutual Ins. Co. v. Tidewater Transfer Co., 337 U. S. 582, 646-655 (1949) (Frankfurter, J., dissenting). As a result, these cases are better understood as simply invoking the comity and federalism concerns discussed in our abstention cases, see, e. g., Los Angeles v. Lyons, 461 U. S. 95 (1983); Trainor v. Hernandez, 431 U. S. 434 (1977); Juidice v. Vail, 430 U. S. 327 (1977); Rizzo v. Goode, 423 U. S. 362 (1976); Huffman v. Pursue, Ltd., 420 U. S. 592 (1975); Younger v. Harris, 401 U. S. 37 (1971), although admittedly in a slightly different voice.3 In my view, federal courts 3 This understanding of our state immunity cases explains an additional anomaly. Over the years, this Court has repeatedly exercised Article III power to review state-court judgments in cases involving claims that, under our post-Hans decisions, could not have been brought in federal district court. See, e. g., Davis v. Michigan Dept, of Treasury, 489 U. S. 803 (1989); Williams v. Vermont, 472 U. S. 14 (1985); Bacchus Imports, Ltd. v. Dias, 468 U. S. 263 (1984); Heublein, Inc. v. South Carolina Tax Comm’n, 409 U. S. 275 (1972); Halliburton Oil Well Cementing Co. v. 28 OCTOBER TERM, 1988 Stevens, J., concurring 491 U. S. “have a primary obligation to protect the rights of the individual that are embodied in the Federal Constitution” and laws, Harris v. Reed, 489 U. S. 255, 267 (1989) (Stevens, J., concurring), and generally should not eschew this responsibility based on some diffuse, instrumental concern for state autonomy. Yet, even if I were convinced otherwise, I would think it readily apparent that congressional abrogation is entirely appropriate.4 Cf. Garcia v. San Antonio Metropolitan Transit Authority, 469 U. S. 528 (1985). Congress is not superseding a constitutional provision in these cases, but rather is setting aside the Court’s assessment of the extent to which the use of constitutionally prescribed federal authority is prudent. Because Congress has decided that the federal interest in protecting the environment outweighs any countervailing interest in not subjecting States to the possible award of monetary damages in a federal court, and because the “judicial power” of the United States plainly extends to such suits, I join Justice Brennan’s opinion. Even if a majority of this Court might have reached a different assessment of the Reily, 373 U. S. 64 (1963); Laurens Federal Savings & Loan Assn. v. South Carolina Tax Comm’n, 365 U. S. 517 (1961). See also Smith v. Reeves, 178 U. S. 436 (1900); Cohens v. Virginia, 6 Wheat. 264 (1821). To the extent the Eleventh Amendment is broadly construed to have removed all federal power to adjudicate claims against the States regardless of whether or not the claim is one arising under federal law, it is difficult to justify our exercise of power in these cases. See Atascadero State Hospital n. Scanlon, 473 U. S. 234, 256, n. 8 (1985) (Brennan, J., dissenting). See also Jackson, 98 Yale L. J., at 13-39. However, if our post-Hans state immunity cases are instead understood as premised on a prudential balancing of state and federal interests, these cases are easily explained: a state-court decision defining federal law tips the balance in favor of federal review. Cf. Michigan v. Long, 463 U. S. 1032, 1040 (1983); Martin n. Hunter’s Lessee, 1 Wheat. 304, 347-348 (1816). 4 To the extent state immunity from suit in federal court is based on a concern for comity, and not on a limitation on Article III power, Congress is just as free to “declare its will” that this presumption come to an end as are States to decide not to accord one another immunity from suit in state court. See Nevada v. Hall, 440 U. S. 410, 425-426 (1979). PENNSYLVANIA v. UNION GAS CO. 29 1 Opinion of Scalia, J. proper balance of state and federal interests as an original matter, once Congress has spoken, we may not disregard its express decision to subject the States to liability under federal law. Justice Scalia, with whom The Chief Justice, Justice O’Connor, and Justice Kennedy join as to Parts II, III, and IV, concurring in part and dissenting in part. I I join Part II of Justice Brennan’s opinion holding that the text of the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U. S. C. § 9601 et seq., as amended by the Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. 99-499, 100 Stat. 1613, clearly renders States liable for money damages in private suits. Justice White’s contention that there is no clear statement is given plausibility only by his methodology of considering CERCLA and SARA separately, finding that first the one and then the other does not necessarily import monetary liability to private individuals — CERCLA because, as we held in Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279 (1973), the inclusion of States within defined terms is not alone enough to evince clear intent to abrogate Eleventh Amendment immunity, post, at 48-49 (opinion concurring in judgment in part and dissenting in part); and SARA because there the unquestionable reference to liability coextensive with the liability of private persons was set forth in a section dealing with limitation of liability, thus not assuring the intent of the Congress which enacted that provision to extend liability to the States, post, at 51-52. That methodology is appropriate, and Justice White’s conclusion is perhaps correct, if one assumes that the task of a court of law is to plumb the intent of the particular Congress that enacted a particular provision. That methodology is not mine nor, I think, the one that courts have traditionally 30 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. followed. It is our task, as I see it, not to enter the minds of the Members of Congress—who need have nothing in mind in order for their votes to be both lawful and effective—but rather to give fair and reasonable meaning to the text of the United States Code, adopted by various Congresses at various times. See United States v. Fausto, 484 U. S. 439, 454-455 (1988). CERCLA, as amended by SARA, clearly holds the States liable for damages in private suits. The inclusion of States, apparently for all purposes, within the definition of “person,” reinforced by the language of the limitation that assumes state liability equivalent to the liability of private individuals, leaves no fair doubt that States are liable to private persons for money damages. Whether it was the CERCLA Congress that envisioned this, or the SARA Congress, is to me irrelevant. The law does. Finding that the statute renders the States liable in private suits for money damages, I must consider the continuing validity of Hans n. Louisiana, 134 U. S. 1 (1890), which held that the Eleventh Amendment precludes individuals from bringing damages suits against States in federal court even where the asserted basis of jurisdiction is not diversity of citizenship but the existence of a federal question. II Eight Members of the Court addressed the question whether to overrule Hans only two Terms ago—but inconclusively, since they were evenly divided. See Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468 (1987). Since the substantive issue was addressed so extensively by the plurality opinion announcing the judgment of the Court in that case (which I will refer to as the “plurality opinion”), and by the dissent, I will only sketch its outlines here. The Eleventh Amendment states: “The Judicial power of the United States shall not be construed to extend to any suit in law or equity, com- PENNSYLVANIA v. UNION GAS CO. 31 1 Opinion of SCALIA, J. menced or prosecuted against one of the United States by Citizens of another State, or by Citizens or Subjects of any Foreign State.” If this text were intended as a comprehensive description of state sovereign immunity in federal courts—that is, if there were no state sovereign immunity beyond its precise terms — then it would unquestionably be most reasonable to interpret it as providing immunity only when the sole basis of federal jurisdiction is the diversity of citizenship that it describes (which of course tracks some of the diversity jurisdictional grants in U. S. Const., Art. Ill, § 2). For there is no plausible reason why one would wish to protect a State from being sued in federal court for violation of federal law (a suit falling within the jurisdictional grant over cases “arising under . . . the Laws of the United States”) when the plaintiff is a citizen of another State or country, but to permit a State to be sued there when the plaintiff is citizen of the State itself. Thus, unless some other constitutional principle beyond the immediate text of the Eleventh Amendment confers immunity in the latter situation—that is to say, unless the text of the Eleventh Amendment is not comprehensive—even if the parties to a suit fell within its precise terms (for example, a State and the citizen of another State) sovereign immunity would not exist so long as one of the other, nondiversity grounds of jurisdiction existed. About a century ago, in the landmark case of Hans v. Louisiana, the Court unanimously rejected this “comprehensive” approach to the Amendment, finding sovereign immunity where not only a nondiversity basis of jurisdiction was present, but even where the parties did not fit the description of the Eleventh Amendment, the plaintiff being a citizen not of another State or country, but of Louisiana itself. What we said in Hans was, essentially, that the Eleventh Amendment was important not merely for what it said but for what it reflected: a consensus that the doctrine of sovereign immunity, for States as well as for the Federal Govern 32 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. ment, was part of the understood background against which the Constitution was adopted, and which its jurisdictional provisions did not mean to sweep away. “[T]he cognizance of suits and actions [against unconsenting States] was not contemplated by the Constitution when establishing the judicial power of the United States.” 134 U. S., at 15. We noted that the decision of this Court that prompted the Eleventh Amendment, Chisholm v. Georgia, 2 Dall. 419 (1793), permitting a South Carolina citizen to bring an assumpsit action for damages against the State of Georgia in federal court, had “created ... a shock of surprise throughout the country,” 134 U. S., at 11; and we concluded that the Amendment which by its precise terms repudiated that decision reflected as well a repudiation of the premise upon which that decision was based, namely, that Article Ill’s jurisdictional grants over the States are unlimited by the doctrine of sovereign immunity. “The letter [of Article III and the Eleventh Amendment] is appealed to now,” we said, “as [the letter of Article III] was then, as a ground for sustaining a suit brought by an individual against a State.” Id., at 15. We rejected that appeal. The rationale of Hans and of the many cases that have followed it was concisely expressed, again for a unanimous Court, by Chief Justice Hughes in a case which held that, despite Article Ill’s express grant of jurisdiction over suits “between a State . . . and foreign States,” and despite the absence of express grant of sovereign immunity in the Eleventh Amendment, a State could not be sued by a foreign State in federal court: “Manifestly, we cannot rest with a mere literal application of the words of §2 of Article III, or assume that the letter of the Eleventh Amendment exhausts the restrictions upon suits against non-consenting States. Behind the words of the constitutional provisions are postulates which limit and control. There is the essential postulate that the controversies, as contemplated, shall be found to be of a justiciable character. There is also PENNSYLVANIA v. UNION GAS CO. 33 1 Opinion of Scalia, J. the postulate that States of the Union, still possessing attributes of sovereignty, shall be immune from suits, without their consent, save where there has been ‘a surrender of this immunity in the plan of the convention.’ The Federalist, No. 81.” Monaco v. Mississippi, 292 U. S. 313, 322-323 (1934) (footnote omitted). The evidence is strong that the jurisdictional grants in Article III of the Constitution did not automatically eliminate underlying state sovereign immunity, and even stronger that that assumption was implicit in the Eleventh Amendment. What is subject to greater dispute, however, is how much sovereign immunity was implicitly eliminated by what Hamilton called the “plan of the convention.” We have already held that “inherent in the constitutional plan,” Monaco v. Mississippi, supra, at 329, are a waiver of immunity against suits by the United States itself, see United States v. Mississippi, 380 U. S. 128, 140-141 (1965); United States v. Texas, 143 U. S. 621, 641-646 (1892), and a waiver of immunity against suits by other States, see South Dakota v. North Carolina, 192 U. S. 286 (1904). The foremost argument urged in favor of overruling Hans is that a waiver of immunity against suits presenting federal questions is also implicit in the constitutional scheme. On this single point I add a few words to what was so recently said in Welch. The inherent necessity of a tribunal for peaceful resolution of disputes between the Union and the individual States, and between the individual States themselves, is incomparably greater, in my view, than the need for a tribunal to resolve disputes on federal questions between individuals and the States. Undoubtedly the Constitution envisions the necessary judicial means to assure compliance with the Constitution and laws. But since the Constitution does not deem this to require that private individuals be able to bring claims against the Federal Government for violation of the Constitution or laws, see United States v. Testan, 424 U. S. 392, 399-402 (1976); U. S. Const., Art. I, §9, cl. 7 (“No Money 34 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. shall be drawn from the Treasury, but in Consequence of Appropriations made by Law”), it is difficult to see why it must be interpreted to require that private individuals be able to bring such claims against the States. If private initiation of suit against the offending sovereign as such is essential to preservation of the structure, it is difficult to see why it would not be essential at both levels. Indeed if anything it would seem more important at the federal level, since suits against the States for violation of the Constitution or laws can at least be brought by the Federal Government itself, see United States v. Mississippi, supra, at 140-141. In providing federal immunity from private suit, therefore, the Constitution strongly suggests that state immunity exists as well. Of course federal law can give, and has given, the private suitor many means short of actions against the State to assure compliance with federal law. He may obtain a federal injunction against the state officer, which will effectively stop the unlawful action, see Ex parte Young, 209 U. S. 123, 160 (1908), and may obtain money damages against state officers, and even local governments, under 42 U. S. C. § 1983; see Monell n. New York City Dept, of Social Services, 436 U. S. 658 (1978). I think it impossible to find in the scheme of the Constitution a necessity that private remedies be expanded beyond this, to include a remedy not available, for a similar infraction, against the United States itself. Even if I were wrong, however, about the original meaning of the Constitution, or the assumption adopted by the Eleventh Amendment, or the structural necessity for federal-question suits against the States, it cannot possibly be denied that the question is at least close. In that situation, the mere venerability of an answer consistently adhered to for almost a century, and the difficulty of changing, or even clearly identifying, the intervening law that has been based on that answer, strongly argue against a change. As noted by the Welch plurality, “Hans has been reaffirmed in case after case, often unanimously and by exceptionally PENNSYLVANIA v. UNION GAS CO. 35 1 Opinion of Scalia, J. strong Courts”; its reversal “would overrule at least 17 cases, in addition to Hans itself” and cast doubt on “a variety of other cases that were concerned with this Court’s traditional treatment of sovereign immunity.” 483 U. S., at 494, n. 27. Moreover, unlike the vast majority of judicial decisions, Hans has had a pervasive effect upon statutory law, automatically assuring that private damages actions created by federal law do not extend against the States. Forty-nine Congresses since Hans have legislated under that assurance. It is impossible to say how many extant statutes would have included an explicit preclusion of suits against States if it had not been thought that such suits were automatically barred. Indeed, it is not even possible to say that, without Hans, all constitutional amendments would have taken the form they did. The Seventeenth Amendment, eliminating the election of Senators by state legislatures, was ratified in 1913, 23 years after Hans. If it had been known at that time that the Federal Government could confer upon private individuals federal causes of action reaching state treasuries; and if the state legislatures had had the experience of urging the Senators they chose to protect them against the proposed creation of such liability; it is not inconceivable, especially at a time when voluntary state waiver of sovereign immunity was rare, that the Amendment (which had to be ratified by three-quarters of the same state legislatures) would have contained a proviso protecting against such incursions upon state sovereignty. I would therefore decline respondent’s invitation to overrule Hans v. Louisiana. Ill Justice Brennan’s plurality opinion purports to assume the validity of Hans, and yet reaches the result that CERCLA’s imposition of monetary liability is constitutional because Congress has the power to abrogate state sovereign immunity in the exercise of its Commerce Clause power. Justice White, who not merely assumes the validity of 36 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. Hans but actually believes in it, agrees with that disposition. Better to overrule Hans, I should think, than to perpetuate the complexities that it creates, see Atascadero State Hospital v. Scanlon, 473 U. S. 234, 252-258 (1985) (Brennan, J., dissenting), but eliminate all its benefits to the federal system. If Hans means only that federal-question suits for money damages against the States cannot be brought in federal court unless Congress clearly says so, it means nothing at all. We do not need Hans for the “clear statement” rule— just as we do not need to rely on any constitutional prohibition of suits against the Federal Government to require a similar rule for elimination of the sovereign immunity of the United States. See United States v. Mitchell, 445 U. S. 535, 538 (1980); United States v. Testan, 424 U. S., at 399. As far as I can discern, the course the Court today pursues — preserving Hans but permitting Congress to overrule it — achieves the worst of both worlds. And it is a course no more justified by text than by consequences. To begin with, Hans did not merely hold that Article III failed to eliminate state sovereign immunity of its own force, without any congressional action to that end. In Hans, as here, there was a congressional statute that could be pointed to as eliminating state sovereign immunity—namely, the Judiciary Act of 1875, ch. 137, § 1, 18 Stat. 470, which gave United States courts jurisdiction over cases involving federal questions. (The Hans Court was unquestionably aware of that refinement, because it was the statutory ground of interpretation of the Judiciary Act of 1789, ch. 20, § 13, 1 Stat. 80, rather than the constitutional ground, that Justice Iredell had relied upon in his dissent in Chisholm, which the Hans Court discussed at some length.) Thus, the distinction that the Court must rely upon is not one between cases in which Congress has assertedly sought to eliminate state sovereign immunity and cases in which in no such assertion is available; but rather the much more gossamer distinction between cases in which Congress has assertedly sought to eliminate PENNSYLVANIA v. UNION GAS CO. 37 1 Opinion of Scalia, J. state sovereign immunity pursuant to its powers to create and organize courts, and cases in which it has assertedly sought to do so pursuant to some of its other powers. I think it plain that the position adopted by the Court contradicts the rationale of Hans, if not its narrow holding. Hans was not expressing some narrow objection to the particular federal power by which Louisiana had been haled into court, but was rather enunciating a fundamental principle of federalism, evidenced by the Eleventh Amendment, that the States retained their sovereign prerogative of immunity. That is clear throughout the opinion, but particularly in the following passage: “Suppose that Congress, when proposing the Eleventh Amendment, had appended to it a proviso that nothing therein contained should prevent a State from being sued by its own citizens in cases arising under the Constitution or laws of the United States: can we imagine that it would have been adopted by the States? The supposition that it would is almost an absurdity on its face. “The truth is, that the cognizance of suits and actions unknown to the law, and forbidden by the law, was not contemplated by the Constitution when establishing the judicial power of the United States.” 134 U. S., at 15. This rationale is also evident from Hans’ reliance upon the dissenting opinion of Justice Iredell in Chisholm—whose views, the Court said, “were clearly right,—as the people of the United States in their sovereign capacity [by ratifying the Eleventh Amendment] subsequently decided.” 134 U. S., at 14. Iredell’s only words addressed precisely to the constitutional issue were as follows: “So much, however, has been said on the Constitution, that it may not be improper to intimate that my present opinion is strongly against any construction of it, which 38 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. will admit, under any circumstances, a compulsive suit against a State for the recovery of money. I think every word in the Constitution may have its full effect without involving this consequence, and that nothing but express words, or an insurmountable implication (neither of which I consider, can be found in this case) would authorise the deduction of so high a power.” 2 Dall., at 449-450. Our later cases are similarly clear that state immunity from suit in federal courts is a structural component of federalism, and not merely a default disposition that can be altered by action of Congress pursuant to its Article I powers. As we unanimously explained in Ex parte New York, 256 U. S. 490, 497 (1921): “That a State may not be sued without its consent is a fundamental rule of jurisprudence having so important a bearing upon the construction of the Constitution of the United States that it has become established by repeated decisions of this court that the entire judicial power granted by the Constitution does not embrace authority to entertain a suit brought by private parties against a State without consent given: not one brought by citizens of another State, or by citizens or subjects of a foreign State, because of the Eleventh Amendment; and not even one brought by its own citizens, because of the fundamental rule of which the Amendment is but an exemplification. ” In Great Northern Ins. Co. n. Read, 322 U. S. 47, 51 (1944), we said: “A state’s freedom from litigation was established as a constitutional right through the Eleventh Amendment. The inherent nature of sovereignty prevents actions against a state by its own citizens without its consent.” PENNSYLVANIA v. UNION GAS CO. 39 1 Opinion of Scalia, J. In Atascadero, 473 U. S., at 242, we identified this principle as an essential element of the constitutional checks and balances: “The ‘constitutionally mandated balance of power’ between the States and the Federal Government was adopted by the Framers to ensure the protection of ‘our fundamental liberties.’ [Garcia v. San Antonio Metropolitan Transit Authority, 469 U. S. 528, 572 (Powell, J., dissenting)]. By guaranteeing the sovereign immunity of the States against suit in federal court, the Eleventh Amendment serves to maintain this balance.” And in recently refusing to overrule Hans in Welch—-an opinion joined by Justice White—the plurality opinion observed that Hans “established that the Eleventh Amendment embodies a broad constitutional principle of sovereign immunity”; that “ ‘a suit directly against a State by one of its own citizens is not one to which the judicial power of the United States extends, unless the State itself consents to be sued.’” 483 U. S., at 486, quoting Hans, 134 U. S., at 21 (Harlan, J. concurring). The only attempt by either the plurality or Justice White to reconcile today’s holding with the “broad constitutional principle of sovereign immunity” established by these precedents is the plurality’s facile assertion that “in approving the commerce power, the States consented to suits against them based on congressionally created causes of action,” ante, at 22. The suggestion that this is the kind of consent our cases had in mind when reciting the familiar phrase, “the States may not be sued without their consent,” does not warrant response. The Court’s conclusion is not only contrary to the clear understanding of a century of cases regarding the Eleventh Amendment, but it contradicts our unvarying approach to Article III as setting forth the exclusive catalog of permissible federal-court jurisdiction. When we have turned to consider whether “a surrender of [state] immunity [is inherent] in the plan of the convention,” we have discussed that issue 40 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. under the rubric of the various grants of jurisdiction in Article III, seeking to determine which of those grants must reasonably be thought to include suits against the States. See, e. g., Monaco, 292 U. S., at 328-330. We have never gone thumbing through the Constitution, to see what other original grants of authority—as opposed to Amendments adopted after the Eleventh Amendment—might justify elimination of state sovereign immunity. If private suits against States, though not permitted under Article III (by virtue of the understanding represented by the Eleventh Amendment), are nonetheless permitted under the Commerce Clause, or under some other Article I grant of federal power, then there is no reason why the other limitations of Article III cannot be similarly exceeded. That Article would be transformed from a comprehensive description of the permissible scope of federal judicial authority to a mere default disposition, applicable unless and until Congress prescribes more expansive authority in the exercise of one of its Article I powers. That is not the regime the Constitution establishes. The Court’s error is clear enough from the embarrassing frailty of the case support to which the plurality opinion appeals. Justice Brennan refers to “statements . . . [that] lay a firm foundation,” ante, at 14, a “path [that] continues,” ibid., and a “message [that] is plain,” ibid. What he notably does not cite is a single Supreme Court case, over the past 200 years upholding (in absence of a waiver) the congressional exercise of the asserted power—or even a single Supreme Court case finding that such an exercise has occurred. How strange that such a useful power—one that the plurality finds essential to the achievement of congressional objectives, ante, at 20-22—should never have been approved and rarely (if ever) have been asserted. Even the “messagesending” dicta that the plurality describes cannot be taken at face value. When the plurality states, for example, that “we have twice assumed that Congress has the authority to abrogate States’ immunity when acting pursuant to the Com- PENNSYLVANIA v. UNION GAS CO. 41 1 Opinion of Scalia, J. merce Clause,” ante, at 15, it means not that we have assumed it to be true, but that we have assumed it for the sake of argument. See Welch, 483 U. S. at 475 (specifically refraining from even “intimating a view of the question”); County of Oneida v. Oneida Indian Nation 470 U. S. 226, 252 (1985). And of the two cases cited as referring to existence of a congressional power “to abrogate . . . immunity,” ante, at 15, one is plainly discussing abrogation not pursuant to Article I but pursuant to the Fourteenth Amendment, see Quern v. Jordan, 440 U. S. 332, 343 (1979), and the other is ambiguous but surely susceptible of that interpretation, see Green v. Mansour, 474 U. S. 64, 68 (1985). In fact the only dicta even suggesting the position the Court today adopts were contained in Parden v. Terminal Railway of Alabama Docks Dept., 377 U. S. 184, 191-192 (1964), and (because it quoted Parden) in Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S., at 286. As our later cases have made plain, see Fitzpatrick v. Bitzer, 427 U. S. 445, 451 (1976), Par deri s holding was based upon the State’s waiver of its sovereign immunity. One aspect of the case has already been overruled, and another cast in doubt, see infra, at 43; its dicta, and the dicta of a later case quoting its dicta, are hardly substantial support for the new constitutional principle the Court adopts. Finally, the plurality opinion errs in relying on Fitzpatrick v. Bitzer, supra, which upheld a money award against a State under Title VII of the Civil Rights Act of 1964, 78 Stat. 253, as amended, 42 U. S. C. §2000e et seq. The distinction, as we carefully explained in that opinion, is that the Civil Rights Act was enacted pursuant to § 5 of the Fourteenth Amendment. We held that “the Eleventh Amendment, and the principle of state sovereignty which it embodies, see Hans v. Louisiana, . . . are necessarily limited” by the later Amendment, 427 U. S., at 456, whose substantive provisions were “by express terms directed at the States,” id., at 453, and “ ‘were intended to be, what they really are, limitations of the 42 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. power of the States and enlargements of the power of Congress,”’ id., at 454, quoting Ex parte Virginia, 100 U. S. 339, 345 (1880). Nothing in this reasoning justifies limitation of the principle embodied in the Eleventh Amendment through appeal to antecedent provisions of the Constitution. The plurality asserts that it is no more impossible for provisions of the Constitution adopted concurrently with Article III to permit abrogation of state sovereign immunity than it is for provisions adopted subsequently. We do not dispute that that is possible, but only that it happened. As suggested above, if the Article I commerce power enables abrogation of state sovereign immunity, so do all the other Article I powers. An interpretation of the original Constitution which permits Congress to eliminate sovereign immunity only if it wants to renders the doctrine a practical nullity and is therefore unreasonable. The Fourteenth Amendment, on the other hand, was avowedly directed against the power of the States, and permits abrogation of their sovereign immunit1 only for a limited purpose. IV It remains for me to consider whether the doctrine of waiver applies here. The basis for application of a waiver theory would be that, subsequent to enactment of CERCLA, Pennsylvania acted as the “owner and operator of. . .a facility,” 42 U. S. C. § 9607(a)(1), which latter term includes a “site or area where a hazardous substance has been deposited, stored, disposed of, or placed, or otherwise come to be located,” §9601(9)(B); and that, by so acting, Pennsylvania voluntarily assumed the state liability for private suit that the legislation (assertedly) contains. Parden is the only case in which we have held that the Federal Government can demand, as a condition to its permission of state action regulable under the Commerce PENNSYLVANIA v. UNION GAS CO. 43 1 Opinion of Scalia, J. Clause, the waiver of state sovereign immunity.1 Two Terms ago, in Welch, we overruled Parden insofar as that case spoke to the clarity of language necessary to constitute such a demand. See 483 U. S., at 478 (plurality opinion); id., at 496 (Scalia, J., concurring in part and concurring in judgment). We explicitly declined to address, however, the continuing validity of Parden^ holding that the Commerce Clause provided the constitutional power to make such a demand, 483 U. S., at 478, n. 8. I would drop the other shoe. There are obvious and fatal difficulties in acknowledging such a power if no Commerce Clause power to abrogate state sovereign immunity exists. All congressional creations of private rights of action attach recovery to the defendant’s commission of some act, or possession of some status, in a field where Congress has authority to regulate conduct. Thus, all federal prescriptions are, insofar as their prospective application is concerned, in a sense conditional, and—to the extent that the objects of the prescriptions consciously engage in the activity or hold the status that produces liability-can be redescribed as invitations to “waiver.” For example, one is not liable for damages to private parties under the federal securities laws, see the Securities Exchange Act of 1934, § 10(b), 48 Stat. 891, 15 U. S. C. § 78j(b), unless one participates in the activity of purchasing or selling securities affecting interstate commerce; and it is possible to describe that liability as not having been categorically imposed, but rather as being the result of a “waiver” of one’s immunity, in 1 In Petty v. Tennessee-Missouri Bridge Comm’n, 359 U. S. 275 (1959), we said that a condition of suability of the Bridge Commission, which we interpreted Congress to have attached to its approval of the interstate compact creating the Commission, was accepted by the States when they implemented the compact. That was an alternative holding, since we also found that the terms of the compact itself made the Commission suable. Obviously, moreover, what Congress may exact with respect to new entities created by compacts that the States have no constitutional power to make without its explicit consent, see U. S. Const., Art. I, § 10, cl. 3, may be much greater than what it may exact in other contexts. 44 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. exchange for federal permission to engage in that activity. At bottom, then, to acknowledge that the Federal Government can make the waiver of state sovereign immunity a condition to the State’s action in a field that Congress has authority to regulate is substantially the same as acknowledging that the Federal Government can eliminate state sovereign immunity in the exercise of its Article I powers2—that is, to adopt the very principle I have just rejected. There is little more than a verbal distinction between saying that Congress can make the Commonwealth of Pennsylvania liable to private parties for hazardous-waste cleanup costs on sites that the Commonwealth owns and operates, and saying the same thing but adding at the end “if the Commonwealth chooses to own and operate them.” If state sovereign immunity has any reality, it must mean more than this. * * * The Court’s holding today can be applauded only by those who think state sovereign immunity so constitutionally insignificant that Hans itself might as well be abandoned. It is only the Court’s steadfast refusal to accept the fundamental structural importance of that doctrine, reflected in Hans and the other cases discussed above, that permits it to regard abrogation through Article I as an open question, and enables the plurality to fight the Hans-Atascadero battle all over again—but this time to win it—on the field of the Commerce Clause. It is a particularly unhappy victory, since instead of cleaning up the allegedly muddled Eleventh Amendment jurisprudence produced by Hans, the Court leaves that in 2 A “waiver” theory would not support retroactive imposition of liability—but that is rare in any event. Moreover, it could be held that waiver cannot occur when the State is unaware of the facts that trigger its liability, or of the law that imposes it. It is difficult to imagine how ignorance of the facts could ever be found, unless (as is most unlikely) we should decline to attribute the knowledge of the State’s agents to the State itself. Our cases discussing waiver have displayed no interest in “actual” state knowledge of either facts or law. PENNSYLVANIA v. UNION GAS CO. 45 1 Opinion of White, J. place, and adds to the clutter the astounding principle that Article III limitations can be overcome by simply exercising Article I powers. It is an unstable victory as well, since that principle is too much at war with itself to endure. We shall either overrule Hans in form as well as in fact, or return to its genuine meaning. I would reverse the judgment of the Court of Appeals on the ground that federal courts have no power to entertain the present suit against the Commonwealth of Pennsylvania. Justice White, with whom The Chief Justice, Justice O’Connor, and Justice Kennedy join as to Part I, concurring in the judgment in part and dissenting in part. I find no “unmistakably clear language,” Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468, 478 (1987), in either CERCLA or SARA that expresses Congress’ intent to abrogate the States’ Eleventh Amendment immunity. However, a majority of the Court concludes otherwise, and therefore I reach the constitutional issue presented here. On that question, I concur in Justice Brennan’s conclusion, but not his reasoning. I Our cases make it plain that only the most direct expression of Congress’ intent to make the States subject to suit will suffice to abrogate their sovereign immunity as recognized in the Eleventh Amendment. Thus, we have said that Congress must “explicitly and by clear language indicate on [the] face [of an enactment] an intent to sweep away the immunity of the States”; and that any such law must “have a history which focuses directly on the question of state liability and which shows that Congress considered and firmly decided to abrogate the Eleventh Amendment immunity of the States.” Quern v. Jordan, 440 U. S. 332, 345 (1979). As we put it more recently: “Congress must express its intention to abrogate the Eleventh Amendment in unmistakable language in 46 OCTOBER TERM, 1988 Opinion of White, J. 491 U. S. the statute itself.” Atascadero State Hospital v. Scanlon, 473 U. S. 234, 243 (1985). Two statutes are offered by the Court as providing the “unmistakable language” required by our cases to abrogate the States’ Eleventh Amendment immunity: the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U. S. C. §9601 et seq. (1982 ed. and Supp. IV), and the 1986 Amendments to CERCLA, found in the Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. 99-499, 100 Stat. 1613. I consider both of these statutes in turn. A I begin by examining CERCLA, in the form in which Congress originally adopted it in 1980. In its initial consideration of this case—under CERCLA before the SARA amendments were added in 1986—the Third Circuit concluded that the statute did not contain an “unmistakable” abrogation of the Eleventh Amendment. United States v. Union Gas Co., 792 F. 2d 372, 378-382 (1986). The Court disagrees, however, suggesting that because CERCLA includes “States” within its definition of “persons,” 42 U. S. C. §9601(21), and because the statute makes “persons” who are “owners or operators,” 42 U. S. C. §9601(20) (1982 ed., Supp. IV), liable under § 9607, Congress expressed in CERCLA an “unmistakably” clear intent to make the States liable to suit by private parties in federal court. Ante, at 7-8. I reject this conclusion for several reasons. First, I note that of the four federal judges who examined this question under CERCLA, only one—Judge Higginbotham in dissent in the Third Circuit’s initial consideration of this case, 792 F. 2d, at 383-386—found in this statutory scheme the requisite clear statement of Congress’ intent to abrogate the States’ immunity. See n. 7, infra. While such a “judicial headcount” is, of course, not dispositive, it does suggest that whatever one can say about CERCLA, it did PENNSYLVANIA v. UNION GAS CO. 47 1 Opinion of White, J. not include an “unmistakable” declaration of abrogation of state immunity. If we are going to be faithful to Atascadero and Welch as providing our standard for this sort of case, then the fact that experienced jurists could disagree about Congress’ intent under CERCLA is relevant, because the disagreement suggests that the statute’s provisions about state liability were certainly not “unmistakably clear.” Second, the significance that the Court draws from CERCLA’s inclusion of States within its definition of persons is suspect for its impact on other portions of the statute. The definitional section the Court relies on also includes the “United States Government” within the term “person.” 42 U. S. C. §9601(21). Yet Congress also adopted, in CERCLA, an entirely separate statutory provision rendering the Federal Government suable under the statute’s liability provision, see § 9607(g). If the Court’s views about the significance of including States within the definition of persons is correct, then § 9607(g) was wholly redundant, because—by including the United States Government within the definition of persons—Congress had already stripped the Federal Government of its sovereign immunity.1 1 In an effort to avoid the force of this observation, the Court unleashes its oft-repeated statement that it relies on a “combination” of CERCLA and SARA to reach its conclusion. Ante, at 9, n. 2. The Court says that it is my “failure to recognize” this quality in its analysis that leads to my “confusion” about this case. Ibid. I do not “fail to recognize” the Court’s approach—I reject it outright. The search for an “unmistakable statement” of abrogation is the search for unmistakable proof that Congress purposefully intended to set aside the States’ immunity. It is, therefore, the search for a historical fact that either was or was not true at the time Congress legislated. The Court’s “combination” analysis loses sight of this underlying theory behind our cases and, unfortunately, substantially undermines our precedents. As I see it, the analysis must be this: either Congress abrogated the Eleventh Amendment when it enacted CERCLA—in which case, § 9607(g) was superfluous when adopted—or Congress did not do so until it adopted SARA—which is a peculiar view, for reasons I explain in Part I-B below— or Congress did not have an intent to abrogate in either instance. Blur- 48 OCTOBER TERM, 1988 Opinion of White, J. 491 U. S. Rather than assuming that Congress wrote a wholly redundant subsection of § 9607, however, it seems more likely to conclude that Congress did not think that including the United States Government or the States within § 9601(21)’s general definition of “persons” subject to CERCLA’s regime was enough to abrogate the sovereign immunity of either for damages awards.2 Cf. United States v. Testan, 424 U. S. 392, 399 (1976). With respect to the Federal Government, Congress went on to enact a separate provision executing the requisite waiver of immunity, § 9607(g). However, with respect to the States, Congress made no such additional provision: the conclusion to be drawn is obvious. Finally, and most importantly, the Court’s reading of CERCLA employs the precise analytical approach we rejected in Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279 (1973). There, as is true here, the relevant statutory term that described who was covered by the Act (in Employees, it was the term “employers” in the Fair Labor Standards Act (FLSA)), expressly included the state defendant (in Employees, it was the State as an employer of “employees of a State . . . hospital”); invoking these provisions, a private litigant sought to hold the State liable under the statute’s damages remedy. Id., at 282-283. Nonetheless, in Employees, we held that Congress had not thereby abrogated the States’ Eleventh Amendment immu- ring the choice among these possible historical facts by resting on a “combination” analysis is only an effort to make this difficult case artificially easier. 2 This conclusion is also supported by the fact that in two other places in § 9607, where Congress wished a particular provision to apply to private persons and the United States and the States, it used the phrases “[n]o person (including the United States or any State) ...” and “any person (including the United States or any State).” See §§9607(i), (j). If Congress believed (as the Court contends that it did) that its inclusion of States within CERCLA’s definition of “person” was adequate to bring the States fully within the operation of § 9607, then the parenthetical phrases I quote here would have been wholly redundant. PENNSYLVANIA v. UNION GAS CO. 49 1 Opinion of White, J. nity; instead, we concluded, Congress had meant only to make the States subject to enforcement actions brought by the Federal Government. Id., at 285-286. In all relevant respects, the portion of CERCLA on which the Court relies and the portion of the FLSA that was before us in Employees are indistinguishable, as are the arguments made for considering the statutes to have abrogated the States’ immunity. In Employees, we rejected these arguments; the same result should attach here. Instead, we should conclude, as we did in Employees, that Congress’ intent could have been to let the Act’s policies be achieved through enforcement actions taken by the Federal Government against the States. As we observed in Employees, supra, at 286: “The policy of the Act so far as the States are concerned is wholly served by allowing the delicate federal-state relationship to be managed through” enforcement actions directed by the Federal Executive Branch—and not through litigation by private parties against the States. Nor is the Court’s result supported by reference to the purposes of CERCLA. Respondent finds much significance in the fact that this statute was designed to be “comprehensive” in nature. 792 F. 2d, at 381 (summarizing respondent’s contention below). But surely the Federal Employers’ Liability Act (Welch), the Rehabilitation Act (Atascadero), and the FLSA (Employees) were all “comprehensive” statutes in their respective fields, and yet this was not enough to deem the Eleventh Amendment abrogated in those cases. Nor is it true that CERCLA’s “comprehensiveness” will be substantially lessened by deeming the States’ immunity to have survived intact. The States remain subject to liability at the hands of the Federal Government; this provides a viable means of achieving CERCLA’s ends. See Reply Brief for Petitioner 10.3 3 Respondent approaches the policy question with the view that limitless state liability under CERCLA is the best means to achieve the statute’s ends. However, Congress clearly did not think so: it limited state 50 OCTOBER TERM, 1988 Opinion of White, J. 491 U. S. Above all, the entire purpose of our “clear statement” rule would be obliterated if this Court were to imply Eleventh Amendment abrogation from our sense of what would best serve the general policy ends Congress was trying to achieve in a statute. Such arguments based on the statute’s general goals, whatever weight they might have under a normal exercise in statutory construction, have no bearing on our analysis of congressional abrogation. Cf. Dellmuth v. Muth, post, at 230-231. If Congress believes that making the States liable to private parties is critical to the scheme it has created in CERCLA, it is up to Congress to say so in unmistakable language. Since it has not, I believe that our “clear statement” precedents bar us from implying such a policy choice—even if it is “latent” in the statutory scheme, or an advisable means of achieving the statute’s ends. B The question then becomes whether, as the Court of Appeals found, United States v. Union Gas Co., 832 F. 2d 1343 (1987), the 1986 amendments to CERCLA (known as SARA) added such an “unmistakable” statement of abrogation to the statute. and local governmental liability under § 9607 in several respects. First, there is the involuntary-ownership exclusion of § 9601(20)(D), adopted in the 1986 SARA amendments, that is discussed in detail in Part I-B, infra. In addition, Congress also adopted in SARA a limitation on state and local government liability (to the Federal Government) for actions taken at toxic waste sites in response to emergencies. Pub. L. 99-499, § 107(d)(2), 100 Stat. 1629; 42 U. S. C. § 9607(d)(2) (1982 ed., Supp. IV). As the House Commerce Committee observed, this legislative exemption was designed to “remov[e] a disincentive for governments to respond to emergencies covered by CERCLA.” H. R. Rep. No. 99-253, pt. 1, p. 73 (1985). Thus, Congress did not view ever expanding governmental liability as the only way to achieve CERCLA’s ends. Of course, even if policy reasons did counsel expansive state liability under CERCLA, our “clear statement” rule mandates that the choice is to be left to Congress—to resolve with an explicit declaration of its decision— and not to be implied by this Court. PENNSYLVANIA v. UNION GAS CO. 51 1 Opinion of White, J. The text of the relevant portion of SARA (now codified at 42 U. S. C. § 9601(20) (D) (1982 ed., Supp. IV)) states, in full: “State or Local Government Limitation—Paragraph (20) of [42 U. S. C. § 9601] (defining ‘owner or operator’) is amended as follows: “(1) Add the following new subparagraph at the end thereof: “‘(D) The term “owner or operator” does not include a unit of State or local government which acquired ownership or control involuntarily through bankruptcy, tax delinquency, abandonment, or other circumstances in which the government involuntarily acquires title by virtue of its function as sovereign. The exclusion provided under this paragraph shall not apply to any State or local government which has caused or contributed to the release ... of a hazardous substance from the facility, and such a State or local government shall be subject to the provisions of this Act in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under [42 U. S. C. §9607].’” Pub. L. 99-499, 100 Stat. 1615. Although Congress entitled the amendment “State or Local Government Limitation,” the Court disparages the idea that §9601(20)(D) was enacted solely as a limitation on governmental unit liability. The Court asserts that such a view ignores that § 101(20)(D) “would be unnecessary unless” the States could be liable under §9607. Ante, at 8. But everyone agrees that States may be liable under §9607: the liability of the Commonwealth of Pennsylvania to the United States. Section 9601(20)(D) provides a significant reduction of that potential liability, as it limits the circumstances under which state and local governments will be forced to pay the United States Government for cleanups at involuntarily acquired sites. Given this fact, §9601(20)(D) makes 52 OCTOBER TERM, 1988 Opinion of White, J. 491 U. S. perfectly good sense without any contortion of it to imply an intent of Congress to abrogate the Eleventh Amendment.4 There is a second fact about the relevant part of SARA that makes it an odd candidate for an Eleventh Amendment abrogation provision: it only applies to facilities acquired by state and local governments “involuntarily ... by virtue of [their] function[s] as sovereign.” See § 9601(20)(D). If this amendment is the means by which Congress intended to make the States liable to suit, it did so only with respect to those properties which a State acquired involuntarily; States would remain immune for sites which they owned and operated by choice. A State would be immune from private suit under § 9607 for costs associated with the cleanup of a state-created, owned, and operated hazardous-waste dump, but it would be liable for discharges at sites it acquired when an owner abandoned his property. Surely if the two cases are to be distinguished, the logical distinction would be exactly the opposite one. Recognizing that Congress could not have intended such a result, the Court avoids this conclusion by saying that this part of SARA “explains and qualifies the entire definition of ‘owner or operator’—not just that part of the definition applicable to involuntary owners.” Ante, at 12-13. But this is plainly wrong: the portion of the sentence which the 4 A similar observation explains another section of SARA which the Court, ante, at 9-10, attempts to use as support for its reading of § 9601 (20)(D): § 9607(d)(2), which was enacted by Congress to encourage state and local governments to conduct emergency cleanups of waste sites by exempting them from potential liability for those cleanup activities. See 42 U. S. C. § 9607(d)(2) (1982 ed., Supp. IV); H. R. Conf. Rep. No. 99-962, pp. 203-204 (1986). About this amendment, the Court again suggests that “Congress need not exempt States from liability unless they would otherwise be liable.” Ante, at 10. As with § 9601(20)(D), however, this limitation is best understood as a limit on state liability to the United States; it need not be read as an implicit statement that elsewhere the Eleventh Amendment has been waived for private lawsuits, in order to make it a vital part of the statute. Cf. Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279, 285-287 (1973). PENNSYLVANIA v. UNION GAS CO. 53 1 Opinion of White, J. Court says renders the States liable (“a State or local government shall be subject . . . ”) is introduced by the words, “[t]he exclusion provided under this paragraph shall not apply. . . .” § 9601(20)(D). Thus, the liability-creating portion of § 9601(20)(D) exists only as a “limit” on the liabilitylimiting portion of §9601(20)(D).5 Under the Court’s reading of the statute, we are left with the paradox of Congress being tougher on States that find themselves involuntary operators of waste sites, than it was on those that had owned and operated such facilities on their own accord. The Court argues that the last clause of the last sentence of §9601(20)(D)—making involuntary-owner state and local governments that cause the release of toxic chemicals “subject to the provisions of [CERCLA] in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity”—provides the clear statement of abrogation required by our cases. But like the Court’s reliance on the inclusion of States within CERCLA’s definition of “persons” subject to the Act (which I discussed above), 5 The Court also rejects this conclusion by saying that the inclusion of the liability-creating exception to the liability-limiting exception of §9601(20)(D) serves to enlighten us as to Congress’ “background understanding” of the effect of CERCLA in the first place: that States would be liable under § 9607. In this instance, and throughout, see n. 1, supra, the Court does not make it clear whether it is the SARA amendments of 1986, or CERCLA itself, that renders the States liable to suit under § 9607. Yet the difference may be a significant one. Section 9607 is a strict-liability provision. See, e. g., New York v. Shore Realty Co., 759 F. 2d 1032, 1042 (CA2 1985); United States v. Bliss, 667 F. Supp. 1298, 1304 (ED Mo. 1987). If CERCLA as originally enacted—without any help from SARA—rendered States liable to private suits under §9607, then they must be subject to that section’s strict-liability rule as well. But under § 9601(20)(D), state and local governments are liable only if they have “caused or contributed” to a release of toxic materials. If § 9601(20)(D) is the source of the Eleventh Amendment waiver, and if, as the Court contends, its provisions are meant to address all state and local governments that own or operate toxic sites, then perhaps Congress abrogated the Eleventh Amendment only far enough to make States liable under this less stringent rule—whether they are voluntary or involuntary owners of a site. 54 OCTOBER TERM, 1988 Opinion of White, J. 491 U. S. this method of analysis is directly contrary to the approach we took in Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279 (1973). The Court insists that its reliance on this part of SARA is correct because, if the statute is interpreted to mean something other than abrogating state immunity, the provision is rendered redundant and meaningless. Ante, at 11-12. The provision, however, has meaning as something less than an abrogation provision because, like the statute in question in Employees, it exists to make the States liable to the Federal Government. While the Court is surely correct when it observes that, under United States v. California, 332 U. S. 19, 26-27 (1947), no statutory provision is required as a general matter to permit the United States to sue a State, here, the Congress forbade such actions in the first part of §9601(20)(D) with respect to some States (i e., involuntary owners of waste sites). Thus, the portion of §9601(20)(D) on which the Court rests its case is precisely like the 1966 amendment to § 3(d) of the FLSA that was before us in Employees: it operates to put some States back into the class of entities that may be liable to the United States, after Congress had previously exempted them from such actions. See Employees, supra, at 282-283. As in Employees, the statute should be read as only authorizing suits by the United States against the States, absent a more clear statement of an authorization of private actions.6 6 The Court goes on to observe, however, that even if this interpretation is accepted as explaining almost all of the last sentence of § 9601(20)(D), it still does not account for Congress “stress[ing] that States would be liable ‘to the same extent ... as any nongovernmental entity,”’ ante, at 11. The Court contends that the first part of the last sentence of § 9601(20)(D) (i. e., “such a State . . . shall be subject”) would have been enough to accomplish the end of merely making involuntary-owner States liable to actions by the United States; the addition of the phrase “as any nongovernmental entity” means that Congress must have intended something more. To this I have three responses. First, Congress may have added the phrase in which the Court puts so much stock (“as any nongovernmental entity”) as a statutory “exclamation PENNSYLVANIA v. UNION GAS CO. 55 1 Opinion of White, J. In Edelman v. Jordan, 415 U. S. 651, 673 (1974), we said of the related question of interpreting a state statute to find a waiver of Eleventh Amendment immunity, that such a waiver would only be found “where stated ‘by the most express language or by such overwhelming implications from the text as [will] leave no room for any other reasonable construction’ ” of the statute in question. Here, there is room for a “reasonable construction” of SARA that does not entail an Eleventh Amendment abrogation; i. e., that Congress intended it as a modification of the liability of the States to the Federal Government. Even if the Court’s interpretation of § 9601(20)(D) were itself “reasonable,” the existence of an alternative, nonabrogating “reasonable” interpretation of the section dictates rejection of its view. Consequently, I do not think that SARA’s liability-limiting amendment to CERCLA contains an “unmistakably clear” statement by Congress that it wanted to abrogate the point”: Congress may have reasoned that while state and local governments that are involuntary owners should be exempted from liability under CERCLA, those that actually cause subsequent discharges should be liable under the statute, with their involuntary ownership no defense or excuse whatsoever when the United States seeks recovery. In this view, Congress simply added the relevant phrase to strongly emphasize that involuntary ownership is no defense if a state or local government causes a discharge. Put another way, it is incongruous to attribute such sweeping significance—an Eleventh Amendment abrogation, something we have found present in only the most extraordinary circumstances—to this one phrase in the definitional portion of SARA/CERCLA. Second, Congress could have used the phrase “as any nongovernmental entity” to insure that local governments that cause discharges at involuntarily acquired sites would be liable under §9607. Congress may have merely wanted to be forceful in using its pre-emptive power to set aside any state-law immunity doctrines for such local government entities, without necessarily going so far as to execute an “unmistakably clear” abrogation of state government immunity. Cf. Quern v. Jordan, 440 U. S. 332, 338-341 (1979). Finally, even if my reading of this phrase makes it somewhat superfluous to the statute, the redundancy created by my interpretation of this one clause is not nearly as severe as the redundancy created by the Court’s reading of the statute, and discussed in the text, supra, at 47. 56 OCTOBER TERM, 1988 Opinion of White, J. 491 U. S. States’ solemn immunity to private suit under the Eleventh Amendment.7 II My view on the statutory issue has not prevailed, however; a majority of the Court has ruled that the statute, as amended, plainly intended to abrogate the immunity of the 7 One additional observation concerning SARA may be made. At the time SARA was enacted, one Court of Appeals—the Third Circuit, in its initial decision in this case, United States v. Union Gas Co., 792 F. 2d 372 (1986)—and one District Court—also as part of this litigation, United States v. Union Gas Co., 575 F. Supp. 949 (ED Pa. 1983)—had ruled on the question whether CERCLA as it was then written abridged States’ Eleventh Amendment immunity. Both of these courts held that it did not; no federal court had ruled to the contrary. The Court’s view of SARA is that, in enacting § 9601(20)(D), Congress had an “unmistakably clear” intent to amend CERCLA so as to reverse the force of these holdings finding a lack of abrogation in CERCLA’s original text. Yet just eight days after it adopted SARA, Congress enacted the Rehabilitation Act Amendments of 1986, Pub. L. 99-506, 100 Stat. 1807, which included a provision setting aside the force of our holding in Atascadero State Hospital v. Scanlon, 473 U. S. 234 (1985), that Congress had failed to provide a clear statement of abrogation of the Eleventh Amendment. The words Congress chose in that Act are instructive: “A State shall not be immune under the Eleventh Amendment. . . from suit in Federal court for a violation of [portions of the Act].” 100 Stat. 1845. While I would not go so far as to hold that Congress must use these precise words (i. e., make reference to the Eleventh Amendment) before it will be deemed to have abrogated States’ immunity, the words used by Congress to set aside Atascadero are legions more “unmistakably clear” than the tangled mess in § 9601(20)(D), which the Court concludes set aside the then-existing case law with respect to CERCLA. Of course, I do not believe that only the “magic words” found in the Rehabilitation Act amendment will suffice to achieve abrogation. Cf. ante, at 13, n. 4. Instead, my view (based on our prior decisions in Atascadero and Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468 (1987)) is that Congress’ intent to abrogate must be expressed clearly, in a plain statement in the text of the enactment—and is not to be derived by parsing together various fragments scattered about a statute, as if it were a legislative quote acrostic. See also n. 1, supra. PENNSYLVANIA v. UNION GAS CO. 57 1 O’Connor, J., dissenting States from suit in the federal courts. I accept that judgment. This brings me to the question whether Congress has the constitutional power to abrogate the States’ immunity.8 In that respect, I agree with the conclusion reached by Justice Brennan in Part III of his opinion, that Congress has the authority under Article I to abrogate the Eleventh Amendment immunity of the States, although I do not agree with much of his reasoning. Accordingly, I would affirm the judgment of the Court of Appeals. Justice O’Connor, dissenting. I agree with Justice Scalia that a faithful interpretation of the Eleventh Amendment embodies a concept of state sovereignty which limits the power of Congress to abrogate States’ immunity when acting pursuant to the Commerce Clause. But that view does not command a majority of the Court, thus necessitating an inquiry whether Congress intended in CERCLA, 42 U. S. C. §9601 et seq., and SARA, Pub. L. 99-499, 100 Stat. 1613, to abrogate the States’ Eleventh Amendment immunity. On that question, I join Part I of Justice White’s opinion. I also join Parts II, III, and IV of Justice Scalia’s opinion concurring in part and dissenting in part. 8 As a preliminary matter, I reiterate my view that, for the reasons stated by the plurality in Welch v. Texas Dept, of Highways, supra, at 478-488, Hans v. Louisiana, 134 U. S. 1 (1890), should not be overruled. 58 OCTOBER TERM, 1988 Syllabus 491 U. S. WILL v. MICHIGAN DEPARTMENT OF STATE POLICE et al. CERTIORARI TO THE SUPREME COURT OF MICHIGAN No. 87-1207. Argued December 5, 1988—Decided June 15, 1989 Petitioner filed Michigan state-court suits under 42 U. S. C. § 1983 alleging that respondents, the Department of State Police and the Director of State Police in his official capacity, had denied him a promotion for an improper reason. The state-court judge ruled for petitioner, finding that both respondents were “persons” under § 1983, which provides that any person who deprives an individual of his or her constitutional rights under color of state law shall be liable to that individual. However, the State Court of Appeals vacated the judgment against the Department, holding that a State is not a person under § 1983, and remanded the case for a determination of the Director’s possible immunity. The State Supreme Court affirmed in part and reversed in part, agreeing that the State is not a person under § 1983, but holding that a state official acting in his or her official capacity also is not such a person. Held: Neither States nor state officials acting in their official capacities are “persons” within meaning of § 1983. Pp. 62-71. (a) That a State is not a person under § 1983 is supported by the statute’s language, congressional purpose, and legislative history. In common usage, the term “person” does not include a State. This usage is particularly applicable where it is claimed that Congress has subjected the States to liability to which they had not been subject before. Reading § 1983 to include States would be a decidedly awkward way of expressing such a congressional intent. The statute’s language also falls short of satisfying the ordinary rule of statutory construction that Congress must make its intention to alter the constitutional balance between the States and the Federal Government unmistakably clear in a statute’s language. Moreover, the doctrine of sovereign immunity is one of the well-established common-law immunities and defenses that Congress did not intend to override in enacting § 1983. Cf. Newport v. Fact Concerts, Inc., 453 U. S. 247; Railroad Co. v. Tennessee, 101 U. S. 337. The “Dictionary Act” provision that a “person” includes “bodies politic and corporate” fails to evidence such an intent. This Court’s ruling in Monell n. New York City Dept, of Social Services, 436 U. S. 658—which held that a municipality is a person under § 1983—is not to the contrary, since States are protected by the Eleventh Amendment while municipalities are not. Pp. 63-70. WILL v. MICHIGAN DEPT. OF STATE POLICE 59 58 Syllabus (b) A suit against state officials in their official capacities is not a suit against the officials but rather is a suit against the officials’ offices and, thus, is no different from a suit against the State itself. Pp. 70-71. 428 Mich. 540, 410 N. W. 2d 749, affirmed. White, J., delivered the opinion of the Court, in which Rehnquist, C. J., and O’Connor, Scalia, and Kennedy, JJ., joined. Brennan, J., filed a dissenting opinion, in which Marshall, Blackmun, and Stevens, JJ., joined, post, p. 71. Stevens, J., filed a dissenting opinion, post, p. 87. William Burnham argued the cause for petitioner. With him on the briefs were Clark Cunningham, Paul D. Rein-gold, John A. Powell, Helen Hershkoff, and Steven R. Shapiro. George H. Weller, Assistant Attorney General of Michigan, argued the cause for respondents. With him on the brief were Frank J. Kelley, Attorney General, Louis J. Caruso, Solicitor General, and Thomas L. Casey, Assistant Solicitor General. * ^William A. Bradford, Jr., Conrad K. Harper, Stuart J. Land, Norman Redlich, William L. Robinson, and Antonia Hernandez filed a brief for the Lawyers’ Committee for Civil Rights Under Law et al. as amici curiae urging reversal. Briefs of amici curiae urging affirmance were filed for the State of Tennessee et al. by W. J. Michael Cody, Attorney General of Tennessee, and Michael W. Catalano, Deputy Attorney General, and by the Attorneys General for their respective jurisdictions as follows: Don Siegelman of Alabama, Robert K. Corbin of Arizona, John Steven Clark of Arkansas, John Van de Kamp of California, Duane Woodard of Colorado, Joseph Lieberman of Connecticut, Charles M. Oberly of Delaware, Robert Butterworth of Florida, Warren Pries III of Hawaii, Neil F. Hartigan of Illinois, Linley E. Pearson of Indiana, Thomas J. Miller of Iowa, Robert T. Stephan of Kansas, Frederic J. Cowan of Kentucky, William J. Guste, Jr., of Louisiana, J. Joseph Curran, Jr., of Maryland, Hubert H. Humphrey III of Minnesota, Michael C. Moore of Mississippi, William L. Webster of Missouri, Mike Greely of Montana, Robert M. Spire of Nebraska, Stephen E. Merrill of New Hampshire, Hal Stratton of New Mexico, Lacy H. Thornburg of North Carolina, Nicholas Spaeth of North Dakota, Anthony J. Celebrezze, Jr., of Ohio, Robert Henry of Oklahoma, LeRoy S. Zimmerman of Pennsylvania, Hector Rivera-Cruz of Puerto Rico, Travis Medlock of South 60 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Justice White delivered the opinion of the Court. This case presents the question whether a State, or an official of the State while acting in his or her official capacity, is a “person” within the meaning of Rev. Stat. § 1979, 42 U. S. C. § 1983. Petitioner Ray Will filed suit in Michigan Circuit Court alleging various violations of the United States and Michigan Constitutions as grounds for a claim under §1983? He alleged that he had been denied a promotion to a data systems analyst position with the Department of State Police for an improper reason, that is, because his brother had been a student activist and the subject of a “red squad” file maintained by respondent. Named as defendants were the Department of State Police and the Director of State Police in his official capacity, also a respondent here.2 The Circuit Court remanded the case to the Michigan Civil Service Commission for a grievance hearing. While the grievance was pending, petitioner filed suit in the Michigan Carolina, Roger A. Tellinghuisen of South Dakota, David L. Wilkinson of Utah, Jeffrey Amestoy of Vermont, Mary Sue Terry of Virginia, Kenneth 0. Eikenberry of Washington, Charlie Brown of West Virginia, Don J. Hanaway of Wisconsin, and Joseph B. Meyer of Wyoming; and for the National Governors’ Association et al. by Benna Ruth Solomon, Kenneth S. Geller, and Andrew J. Pincus. 1 Section 1983 provides as follows: “Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory or the District of Columbia, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in an action at law, suit in equity, or other proper proceeding for redress. For the purposes of this section, any Act of Congress applicable exclusively to the District of Columbia shall be considered to be a statute of the District of Columbia.” 42 U. S. C. § 1983. 2 Also named as defendants were the Michigan Department of Civil Service and the State Personnel Director, but those parties were subsequently dismissed by the state courts. WILL v. MICHIGAN DEPT. OF STATE POLICE 61 58 Opinion of the Court Court of Claims raising an essentially identical § 1983 claim. The Civil Service Commission ultimately found in petitioner’s favor, ruling that respondents had refused to promote petitioner because of “partisan considerations.” App. 46. On the basis of that finding, the state-court judge, acting in both the Circuit Court and the Court of Claims cases, concluded that petitioner had established a violation of the United States Constitution. The judge held that the Circuit Court action was barred under state law but that the Claims Court action could go forward. The judge also ruled that respondents were persons for purposes of § 1983. The Michigan Court of Appeals vacated the judgment against the Department of State Police, holding that a State is not a person under § 1983, but remanded the case for determination of the possible immunity of the Director of State Police from liability for damages. The Michigan Supreme Court granted discretionary review and affirmed the Court of Appeals in part and reversed in part. Smith v. Department of Pub. Health, 428 Mich. 540, 410 N. W. 2d 749 (1987). The Supreme Court agreed that the State itself is not a person under § 1983, but held that a state official acting in his or her official capacity also is not such a person. The Michigan Supreme Court’s holding that a State is not a person under § 1983 conflicts with a number of state- and federal-court decisions to the contrary.3 We granted certiorari to resolve the conflict. 485 U. S. 1005 (1988). 8 The courts in the following cases have taken the position that a State is a person under § 1983. See Della Grotta v. Rhode Island, 781 F. 2d 343, 349 (CAI 1986); Gay Student Services v. Texas A&M University, 612 F. 2d 160, 163-164 (CA5), cert, denied, 449 U. S. 1034 (1980); Uberoi v. University of Colorado, 713 P. 2d 894, 900-901 (Colo. 1986); Stanton v. Godfrey, 415 N. E. 2d 103, 107 (Ind. App. 1981); Gumbhir v. Kansas State Bd. of Pharmacy, 231 Kan. 507, 512-513, 646 P. 2d 1078, 1084 (1982), cert, denied, 459 U. S. 1103 (1983); Rahmah Navajo School Bd., Inc. n. Bureau of Revenue, 104 N. M. 302, 310, 720 P. 2d 1243, 1251 (App.), cert, denied, 479 U. S. 940 (1986). A larger number of courts have agreed with the Michigan Supreme Court that a State is not a person under § 1983. See Ruiz v. Estelle, 679 62 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Prior to Monell v. New York City Dept, of Social Services, 436 U. S. 658 (1978), the question whether a State is a person within the meaning of § 1983 had been answered by this Court in the negative. In Monroe v. Pape, 365 U. S. 167, 187-191 (1961), the Court had held that a municipality was not a person under § 1983. “[T]hat being the case,” we reasoned, § 1983 “could not have been intended to include States as parties defendant.” Fitzpatrick v. Bitzer, 427 U. S. 445, 452 (1976). But in Monell, the Court overruled Monroe, holding that a municipality was a person under § 1983. 436 U. S., at 690. Since then, various members of the Court have debated whether a State is a person within the meaning of § 1983, see Hutto v. Finney, 437 U. S. 678, 700-704 (1978) (Brennan, J., concurring); id., at 708, n. 6 (Powell, J., concurring in F. 2d 1115, 1137 (CA5), modified on other grounds, 688 F. 2d 266 (1982), cert, denied, 460 U. S. 1042 (1983); Toledo, P. & W. R. Co. v. Illinois, 744 F. 2d 1296, 1298-1299, and n. 1 (CA7 1984), cert, denied, 470 U. S. 1051 (1985); Harris v. Missouri Court of Appeals, 787 F. 2d 427, 429 (CA8), cert, denied, 479 U. S. 851 (1986); Aubuchon v. Missouri, 631 F. 2d 581, 582 (CA8 1980) (per curiam), cert, denied, 450 U. S. 915 (1981); State v. Green, 633 P. 2d 1381, 1382 (Alaska 1981); St. Mary’s Hospital and Health Center v. State, 150 Ariz. 8, 11, 721 P. 2d 666, 669 (App. 1986); Mezey n. State, 161 Cal. App. 3d 1060, 1065, 208 Cal. Rptr. 40, 43 (1984); Hill v. Florida Dept, of Corrections, 513 So. 2d 129, 132 (Fla. 1987), cert, denied, 484 U. S. 1064 (1988); Merritt ex rel. Merritt v. State, 108 Idaho 20, 26, 696 P. 2d 871, 877 (1985); Woodbridge v. Worcester State Hospital, 384 Mass. 38, 44-45, n. 7, 423 N. E. 2d 782, 786, n. 7 (1981); Bird v. State Dept, of Public Safety, 375 N. W. 2d 36, 43 (Minn. App. 1985); Shaw v. St. Louis, 664 S. W. 2d 572, 576 (Mo. App. 1983), cert, denied, 469 U. S. 849 (1984); Fuchilla v. Layman, 109 N. J. 319, 323-324, 537 A. 2d 652, 654, cert, denied, 488 U. S. 826 (1988); Burkey v. Southern Ohio Correctional Facility, 38 Ohio App. 3d 170, 170-171, 528 N. E. 2d 607, 608 (1988); Gay v. State, 730 S. W. 2d 154, 157-158 (Tex. App. 1987); Edgar v. State, 92 Wash. 2d 217, 221, 595 P. 2d 534, 537 (1979), cert, denied, 444 U. S. 1077 (1980); Boldt v. State, 101 Wis. 2d 566, 584, 305 N. W. 2d 133, 143-144, cert, denied, 454 U. S. 973 (1981). WILL v. MICHIGAN DEPT. OF STATE POLICE 63 58 Opinion of the Court part and dissenting in part), but this Court has never expressly dealt with that issue.4 Some courts, including the Michigan Supreme Court here, have construed our decision in Quern v. Jordan, 440 U. S. 332 (1979), as holding by implication that a State is not a person under § 1983. See Smith v. Department of Pub. Health, supra, at 581, 410 N. W. 2d, at 767. See also, e. g., State v. Green, 633 P. 2d 1381, 1382 (Alaska 1981); Woodbridge v. Worcester State Hospital, 384 Mass. 38, 44-45, n. 7, 423 N. E. 2d 782, 786, n. 7 (1981); Edgar v. State, 92 Wash. 2d 217, 221, 595 P. 2d 534, 537 (1979), cert, denied, 444 U. S. 1077 (1980). Quern held that § 1983 does not override a State’s Eleventh Amendment immunity, a holding that the concurrence suggested was “patently dicta” to the effect that a State is not a person, 440 U. S., at 350 (Brennan, J., concurring in judgment). Petitioner filed the present § 1983 actions in Michigan state court, which places the question whether a State is a person under § 1983 squarely before us since the Eleventh Amend 4 Petitioner cites a number of cases from this Court that he asserts have “assumed” that a State is a person. Those cases include ones in which a State has been sued by name under § 1983, see, e. g., Maine v. Thiboutot, 448 U. S. 1 (1980); Martinez v. California, 444 U. S. 277 (1980), various cases awarding attorney’s fees against a State or a state agency, Maine v. Thiboutot, supra; Hutto v. Finney, 437 U. S. 678 (1978), and various cases discussing the waiver of Eleventh Amendment immunity by States, see, e. g., Kentucky v. Graham, 473 U. S. 159, 167, n. 14 (1985); Edelman v. Jordan, 415 U. S. 651 (1974). But the Court did not address the meaning of person in any of those cases, and in none of the cases was resolution of that issue necessary to the decision. Petitioner’s argument evidently rests on the proposition that whether a State is a person under § 1983 is “jurisdictional” and “thus could have been raised by the Court on its own motion” in those cases. Brief for Petitioner 25, n. 15. Even assuming that petitioner’s premise and characterization of the cases is correct, “this Court has never considered itself bound [by prior sub silentio holdings] when a subsequent case finally brings the jurisdictional issue before us.” Hagans v. Lavine, 415 U. S. 528, 535, n. 5 (1974). 64 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ment does not apply in state courts. Maine n. Thiboutot, 448 U. S. 1, 9, n. 7 (1980). For the reasons that follow, we reaffirm today what we had concluded prior to Monell and what some have considered implicit in Quern: that a State is not a person within the meaning of § 1983. We observe initially that if a State is a “person” within the meaning of § 1983, the section is to be read as saying that “every person, including a State, who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory or the District of Columbia, subjects . . . .” That would be a decidedly awkward way of expressing an intent to subject the States to liability. At the very least, reading the statute in this way is not so clearly indicated that it provides reason to depart from the often-expressed understanding that “‘in common usage, the term ‘person’ does not include the sovereign, [and] statutes employing the [word] are ordinarily construed to exclude it.’” Wilson v. Omaha Tribe, 442 U. S. 653, 667 (1979) (quoting United States v. Cooper Corp., 312 U. S. 600, 604 (1941)). See also United States v. Mine Workers, 330 U. S. 258, 275 (1947). This approach is particularly applicable where it is claimed that Congress has subjected the States to liability to which they had not been subject before. In Wilson v. Omaha Tribe, supra, we followed this rule in construing the phrase “white person” contained in 25 U. S. C. § 194, enacted as Act of June 30, 1834, 4 Stat. 729, as not including the “sovereign States of the Union.” 442 U. S., at 667. This common usage of the term “person” provides a strong indication that “person” as used in § 1983 likewise does not include a State.5 3 Jefferson County Pharmaceutical Assn. v. Abbott Laboratories, 460 U. S. 150 (1983), on which petitioner relies, is fully reconcilable with our holding in the present case. In Jefferson County, the Court held that States were persons that could be sued under the Robinson-Patman Act, 15 U. S. C. §§ 13(a) and 13(f). 460 U. S., at 155-157. But the plaintiff there was seeking only injunctive relief and not damages against the State WILL v. MICHIGAN DEPT. OF STATE POLICE 65 58 Opinion of the Court The language of § 1983 also falls far short of satisfying the ordinary rule of statutory construction that if Congress intends to alter the “usual constitutional balance between the States and the Federal Government,” it must make its intention to do so “unmistakably clear in the language of the statute.” Atascadero State Hospital n. Scanlon, 473 U. S. 234, 242 (1985); see also Pennhurst State School and Hospital v. Halderman, 465 U. S. 89, 99 (1984). Atascadero was an Eleventh Amendment case, but a similar approach is applied in other contexts. Congress should make its intention “clear and manifest” if it intends to pre-empt the historic powers of the States, Rice v. Santa Fe Elevator Corp., 331 U. S. 218, 230 (1947), or if it intends to impose a condition on the grant of federal moneys, Pennhurst State School and Hospital v. Halderman, 451 U. S. 1, 16 (1981); South Dakota v. Dole, 483 U. S. 203, 207 (1987). “In traditionally sensitive areas, such as legislation affecting the federal balance, the requirement of clear statement assures that the legislature has in fact faced, and intended to bring into issue, the critical matters involved in the judicial decision.” United States v. Bass, 404 U. S. 336, 349 (1971). Our conclusion that a State is not a “person” within the meaning of § 1983 is reinforced by Congress’ purpose in en- defendant, the Board of Trustees of the University of Alabama; the District Court had dismissed the plaintiff’s damages claim as barred by the Eleventh Amendment. Id., at 153, n. 5. Had the present § 1983 action been brought in federal court, a similar disposition would have resulted. Of course, the Court would never be faced with a case such as Jefferson County that had been brought in a state court because the federal courts have exclusive jurisdiction over claims under the federal antitrust laws. 15 U. S. C. §§ 15 and 26. Moreover, the Court in Jefferson County was careful to limit its holding to “state purchases for the purpose of competing against private enterprise ... in the retail market.” 460 U. S., at 154. It assumed without deciding “that Congress did not intend the Act to apply to state purchases for consumption in traditional governmental functions,” ibid., which presents a more difficult question because it may well “affec[t] the federal balance.” See United States v. Bass, 404 U. S. 336, 349 (1971). 66 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. acting the statute. Congress enacted § 1 of the Civil Rights Act of 1871, 17 Stat. 13, the precursor to § 1983, shortly after the end of the Civil War “in response to the widespread deprivations of civil rights in the Southern States and the inability or unwillingness of authorities in those States to protect those rights or punish wrongdoers.” Felder n. Casey, 487 U. S. 131, 147 (1988). Although Congress did not establish federal courts as the exclusive forum to remedy these deprivations, ibid., it is plain that “Congress assigned to the federal courts a paramount role” in this endeavor, Patsy n. Board of Regents of Florida, 457 U. S. 496, 503 (1982). Section 1983 provides a federal forum to remedy many deprivations of civil liberties, but it does not provide a federal forum for litigants who seek a remedy against a State for alleged deprivations of civil liberties. The Eleventh Amendment bars such suits unless the State has waived its immunity, Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468, 472-473 (1987) (plurality opinion), or unless Congress has exercised its undoubted power under § 5 of the Fourteenth Amendment to override that immunity. That Congress, in passing § 1983, had no intention to disturb the States’ Eleventh Amendment immunity and so to alter the federal-state balance in that respect was made clear in our decision in Quern. Given that a principal purpose behind the enactment of § 1983 was to provide a federal forum for civil rights claims, and that Congress did not provide such a federal forum for civil rights claims against States, we cannot accept petitioner’s argument that Congress intended nevertheless to create a cause of action against States to be brought in state courts, which are precisely the courts Congress sought to allow civil rights claimants to avoid through § 1983. This does not mean, as petitioner suggests, that we think that the scope of the Eleventh Amendment and the scope of § 1983 are not separate issues. Certainly they are. But in deciphering congressional intent as to the scope of § 1983, the WILL v. MICHIGAN DEPT. OF STATE POLICE 67 58 Opinion of the Court scope of the Eleventh Amendment is a consideration, and we decline to adopt a reading of § 1983 that disregards it.6 Our conclusion is further supported by our holdings that in enacting §1983, Congress did not intend to override well-established immunities or defenses under the common law. “One important assumption underlying the Court’s decisions in this area is that members of the 42d Congress were familiar with common-law principles, including defenses previously recognized in ordinary tort litigation, and that they likely intended these common-law principles to obtain, absent specific provisions to the contrary.” Newport v. Fact Concerts, Inc., 453 U. S. 247, 258 (1981). Stump n. Sparkman, 435 U. S. 349, 356 (1978); Scheuer v. Rhodes, 416 U. S. 232, 247 (1974); Pierson n. Ray, 386 U. S. 547, 554 (1967); and Tenney v. Brandhove, 341 U. S. 367, 376 (1951), are also to this effect. The doctrine of sovereign immunity was a familiar doctrine at common law. “The principle is elementary that a State cannot be sued in its own courts without its consent.” Railroad Co. v. Tennessee, 101 U. S. 337, 339 (1880). It is an “established principle of jurisprudence” that the sovereign cannot be sued in its own courts without its consent. Beers v. Arkansas, 20 How. 527, 529 (1858). We cannot conclude that § 1983 was intended to disregard the well-established immunity of a State from being sued without its consent.7 fi Petitioner argues that Congress would not have considered the Eleventh Amendment in enacting § 1983 because in 1871 this Court had not yet held that the Eleventh Amendment barred federal-question cases against States in federal court. This argument is no more than an attempt to have this Court reconsider Quern v. Jordan, 440 U. S. 332 (1979), which we decline to do. ' Our recognition in Monell v. New York City Dept, of Social Services, 436 U. S. 658 (1978), that a municipality is a person under § 1983, is fully consistent with this reasoning. In Owen v. City of Independence, 445 U. S. 622 (1980), we noted that by the time of the enactment of § 1983, municipalities no longer retained the sovereign immunity they had previously shared with the States. “[B]y the end of the 19th century, courts 68 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. The legislative history of § 1983 does not suggest a different conclusion. Petitioner contends that the congressional debates on § 1 of the 1871 Act indicate that § 1983 was intended to extend to the full reach of the Fourteenth Amendment and thereby to provide a remedy “ ‘against all forms of official violation of federally protected rights.’” Brief for Petitioner 16 (quoting Monell, 436 U. S., at 700-701). He refers us to various parts of the vigorous debates accompanying the passage of § 1983 and revealing that it was the failure of the States to take appropriate action that was undoubtedly the motivating force behind § 1983. The inference must be drawn, it is urged, that Congress must have intended to subject the States themselves to liability. But the intent of Congress to provide a remedy for unconstitutional state action does not without more include the sovereign States among those persons against whom § 1983 actions would lie. Construing § 1983 as a remedy for “official violation of federally protected rights” does no more than confirm that the section is directed against state action—action “under color of” state law. It does not suggest that the State itself was a person that Congress intended to be subject to liability. Although there were sharp and heated debates, the discussion of § 1 of the bill, which contained the present § 1983, was not extended. And although in other respects the impact on state sovereignty was much talked about, no one suggested that § 1 would subject the States themselves to a damages suit under federal law. Quern, 440 U. S., at 343. There was complaint that § 1 would subject state officers to damages liability, but no suggestion that it would also expose the States themselves. Cong. Globe, 42d Cong., 1st Sess., regularly held that in imposing a specific duty on the municipality either in its charter or by statute, the State had impliedly withdrawn the city’s immunity from liability for the nonperformance or misperformance of its obligation,” id., at 646, and, as a result, municipalities had been held liable for damages “in a multitude of cases” involving previously immune activities, id., at 646-647. WILL v. MICHIGAN DEPT. OF STATE POLICE 69 58 Opinion of the Court 366, 385 (1871). We find nothing substantial in the legislative history that leads us to believe that Congress intended that the word “person” in § 1983 included the States of the Union. And surely nothing in the debates rises to the clearly expressed legislative intent necessary to permit that construction. Likewise, the Act of Feb. 25, 1871, §2, 16 Stat. 431 (the “Dictionary Act”),8 on which we relied in Moneti, supra, at 688-689, does not counsel a contrary conclusion here. As we noted in Quern, that Act, while adopted prior to § 1 of the Civil Rights Act of 1871, was adopted after §2 of the Civil Rights Act of 1866, from which § 1 of the 1871 Act was derived. 440 U. S., at 341, n. 11. Moreover, we disagree with Justice Brennan that at the time the Dictionary Act was passed “the phrase ‘bodies politic and corporate’ was understood to include the States.” Post, at 78. Rather, an examination of authorities of the era suggests that the phrase was used to mean corporations, both private and public (municipal), and not to include the States.9 In our view, the 8 The Dictionary Act provided that “in all acts hereafter passed . . . the word ‘person’ may extend and be applied to bodies politic and corporate . . . unless the context shows that such words were intended to be used in a more limited sense.” Act of Feb. 25, 1871, §2, 16 Stat. 431. 9 See United States v. Fox, 94 U. S. 315, 321 (1877); 1 B. Abbott, Dictionary of Terms and Phrases Used in American or English Jurisprudence 155 (1879) (“most exact expression” for “public corporation”); W. Anderson, A Dictionary of Law 127 (1893) (“most exact expression for a public corporation or corporation having powers of government”); Black’s Law Dictionary 143 (1891) (“body politic” is “term applied to a corporation, which is usually designated as a ‘body corporate and politic’ ” and “is particularly appropriate to a public corporation invested with powers and duties of government”); 1 A. Burrill, A Law Dictionary and Glossary 212 (2d ed. 1871) (“body politic” is “term applied to a corporation, which is usually designated as a body corporate and politic”). A public corporation, in ordinary usage, was another term for a municipal corporation, and included towns, cities, and counties, but not States. See 2 Abbott, supra, 70 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Dictionary Act, like § 1983 itself and its legislative history, fails to evidence a clear congressional intent that States be held liable. Finally, Monell itself is not to the contrary. True, prior to Monell the Court had reasoned that if municipalities were not persons then surely States also were not. Fitzpatrick n. Bitzer, 427 U. S., at 452. And Monell overruled Monroe, undercutting that logic. But it does not follow that if municipalities are persons then so are States. States are protected by the Eleventh Amendment while municipalities are not, Monell, 436 U. S., at 690, n. 54, and we consequently limited our holding in Monell “to local government units which are not considered part of the State for Eleventh Amendment purposes,” ibid. Conversely, our holding here does not cast any doubt on Monell, and applies only to States or governmental entities that are considered “arms of the State” for Eleventh Amendment purposes. See, e. g., Mt. Healthy Bd. of Ed. v. Doyle, 429 U. S. 274, 280 (1977). Petitioner asserts, alternatively, that state officials should be considered “persons” under § 1983 even though acting in their official capacities. In this case, petitioner named as defendant not only the Michigan Department of State Police but also the Director of State Police in his official capacity. at 347; Anderson, supra, at 264-265; Black, supra, at 278; 2 Burrill, supra, at 352. Justice Brennan appears to confuse this precise definition of the phrase with its use “in a rather loose way,” see Black, supra, at 143, to refer to the state (as opposed to a State). This confusion is revealed most clearly in Justice Brennan’s reliance on the 1979 edition of Black’s Law Dictionary, which defines “body politic or corporate” as “[a] social compact by which the whole people covenants with each citizen, and each citizen with the whole people, that all shall be governed by certain laws for the common good.” Post, at 79. To the extent Justice Brennan’s citation of other authorities does not suffer from the same confusion, those authorities at best suggest that the phrase is ambiguous, which still renders the Dictionary Act incapable of supplying the necessary clear intent. WILL v. MICHIGAN DEPT. OF STATE POLICE 71 58 Brennan, J., dissenting Obviously, state officials literally are persons. But a suit against a state official in his or her official capacity is not a suit against the official but rather is a suit against the official’s office. Brandon v. Holt, 469 U. S. 464, 471 (1985). As such, it is no different from a suit against the State itself. See, e. g., Kentucky n. Graham, 473 U. S. 159, 165-166 (1985); Monell, supra, at 690, n. 55. We see no reason to adopt a different rule in the present context, particularly when such a rule would allow petitioner to circumvent congressional intent by a mere pleading device.10 We hold that neither a State nor its officials acting in their official capacities are “persons” under § 1983. The judgment of the Michigan Supreme Court is affirmed. It is so ordered. Justice Brennan, with whom Justice Marshall, Justice Blackmun, and Justice Stevens join, dissenting. Because this case was brought in state court, the Court concedes, the Eleventh Amendment is inapplicable here. See ante, at 63-64. Like the guest who would not leave, 10 Of course a state official in his or her official capacity, when sued for injunctive relief, would be a person under § 1983 because “official-capacity actions for prospective relief are not treated as actions against the State.” Kentucky v. Graham, 473 U. S., at 167, n. 14; Ex parte Young, 209 U. S. 123, 159-160 (1908). This distinction is “commonplace in sovereign immunity doctrine,” L. Tribe, American Constitutional Law § 3-27, p. 190, n. 3 (2d ed. 1988), and would not have been foreign to the 19th-century Congress that enacted § 1983, see, e. g., In re Ayers, 123 U. S. 443, 506-507 (1887); United States v. Lee, 106 U. S. 196, 219-222 (1882); Board of Liquidation v. McComb, 92 U. S. 531, 541 (1876); Osborn v. Bank of United States, 9 Wheat. 738 (1824). City of Kenosha v. Bruno, 412 U. S. 507, 513 (1973), on which Justice Stevens relies, see post, at 93, n. 8, is not to the contrary. That case involved municipal liability under § 1983, and the fact that nothing in § 1983 suggests its “bifurcated application to municipal corporations depending on the nature of the relief sought against them,” 412 U. S., at 513, is not surprising, since by the time of the enactment of § 1983 municipalities were no longer protected by sovereign immunity. Supra, at 67-68, n. 7. 72 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. however, the Eleventh Amendment lurks everywhere in today’s decision and, in truth, determines its outcome. I Section 1 of the Civil Rights Act of 1871, 42 U. S. C. §1983, renders certain “persons” liable for deprivations of constitutional rights. The question presented is whether the word “person” in this statute includes the States and state officials acting in their official capacities. One might expect that this statutory question would generate a careful and thorough analysis of the language, legislative history, and general background of § 1983. If this is what one expects, however, one will be disappointed by today’s decision. For this case is not decided on the basis of our ordinary method of statutory construction; instead, the Court disposes of it by means of various rules of statutory interpretation that it summons to its aid each time the question looks close. Specifically, the Court invokes the following interpretative principles: the word “persons” is ordinarily construed to exclude the sovereign; congressional intent to affect the federal-state balance must be “clear and manifest”; and intent to abrogate States’ Eleventh Amendment immunity must appear in the language of the statute itself. The Court apparently believes that each of these rules obviates the need for close analysis of a statute’s language and history. Properly applied, however, only the last of these interpretative principles has this effect, and that principle is not pertinent to the case before us. The Court invokes, first, the “often-expressed understanding” that “‘in common usage, the term “person” does not include the sovereign, [and] statutes employing the [word] are ordinarily construed to exclude it.’” Ante, at 64, quoting Wilson v. Omaha Tribe, 442 U. S. 653, 667 (1979). This rule is used both to refute the argument that the language of § 1983 demonstrates an intent that States be included as defendants, ante, at 64, and to overcome the argu- WILL v. MICHIGAN DEPT. OF STATE POLICE 73 58 Brennan, J., dissenting ment based on the Dictionary Act’s definition of “person” to include bodies politic and corporate, ante, at 69-70. It is ironic, to say the least, that the Court chooses this interpretive rule in explaining why the Dictionary Act is not decisive, since the rule is relevant only when the word “persons” has no statutory definition. When one considers the origins and content of this interpretive guideline, moreover, one realizes that it is inapplicable here and, even if applied, would defeat rather than support the Court’s approach and result. The idea that the word “persons” ordinarily excludes the sovereign can be traced to the “familiar principle that the King is not bound by any act of Parliament unless he be named therein by special and particular words.” Dollar Savings Bank v. United States, 19 Wall. 227, 239 (1874). As this passage suggests, however, this interpretive principle applies only to “the enacting sovereign.” United States v. California, 297 U. S. 175, 186 (1936). See also Jefferson County Pharmaceutical Assn., Inc. v. Abbott Laboratories, 460 U. S. 150, 161, n. 21 (1983). Furthermore, as explained in United States v. Herron, 20 Wall. 251, 255 (1874), even the principle as applied to the enacting sovereign is not without limitations: “Where an act of Parliament is made for the public good, as for the advancement of religion and justice or to prevent injury and wrong, the king is bound by such act, though not particularly named therein; but where a statute is general, and thereby any prerogative, right, title, or interest is divested or taken from the king, in such case the king is not bound, unless the statute is made to extend to him by express words.” It would be difficult to imagine a statute more clearly designed “for the public good,” and “to prevent injury and wrong,” than § 1983. Even if this interpretive principle were relevant to this case, the Court’s invocation of it to the exclusion of careful statutory analysis is in error. As we have made clear, this principle is merely “an aid to consistent construction of statutes of the enacting sovereign when their purpose is in 74 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. doubt, but it does not require that the aim of a statute fairly to be inferred be disregarded because not explicitly stated.” United States v. California, supra, at 186. Indeed, immediately following the passage quoted by the Court today, ante, at 64, to the effect that statutes using the word “person” are “ordinarily construed to exclude” the sovereign, we stated: “But there is no hard and fast rule of exclusion. The purpose, the subject matter, the context, the legislative history, and the executive interpretation of the statute are aids to construction which may indicate an intent, by the use of the term, to bring state or nation within the scope of the law. “Decision is not to be reached by a strict construction of the words of the Act, nor by the application of artificial canons of construction. On the contrary, we are to read the statutory language in its ordinary and natural sense, and if doubts remain, resolve them in the light, not only of the policy intended to be served by the enactment, but, as well, by all other available aids to construction.” United States v. Cooper Corp., 312 U. S. 600, 604-605 (1941). See also Wilson v. Omaha Indian Tribe, supra, at 667 (“There is . . . ‘no hard and fast rule of exclusion,’ United States v. Cooper Corp., [312 U. S. 600,] 604-605 [(1941)]; and much depends on the context, the subject matter, legislative history, and executive interpretation”); Pfizer Inc. v. India, 434 U. S. 308, 315-318 (1978); Guarantee Title & Trust Co. n. Title Guaranty & Surety Co., 224 U. S. 152, 155 (1912); Lewis v. United States, 92 U. S. 618, 622 (1875); Green v. United States, 9 Wall. 655, 658 (1870). The second interpretive principle that the Court invokes comes from cases such as Rice v. Santa Fe Elevator Corp., 331 U. S. 218, 230 (1947); Pennhurst State School and Hospital n. Halderman, 451 U. S. 1, 16 (1981); South Dakota n. Dole, 483 U. S. 203, 207-208 (1987); and United States v. WILL v. MICHIGAN DEPT. OF STATE POLICE 75 58 Brennan, J., dissenting Bass, 404 U. S. 336, 349 (1971), which require a “clear and manifest” expression of congressional intent to change some aspect of federal-state relations. Ante, at 65. These cases do not, however, permit substitution of an absolutist rule of statutory construction for thorough statutory analysis. Indeed, in each of these decisions the Court undertook a careful and detailed analysis of the statutory language and history under consideration. Rice is a particularly inapposite source for the interpretive method that the Court today employs, since it observes that, according to conventional pre-emption analysis, a “clear and manifest” intent to pre-empt state legislation may appear in the “scheme” or “purpose” of the federal statute. See 331 U. S., at 230. The only principle of statutory construction employed by the Court that would justify a perfunctory and inconclusive analysis of a statute’s language and history is one that is irrelevant to this case. This is the notion “that if Congress intends to alter the ‘usual constitutional balance between the States and the Federal Government,’ it must make its intention to do so ‘unmistakably clear in the language of the statute.’” Ante, at 65, quoting Atascadero State Hospital n. Scanlon, 473 U. S. 234, 242 (1985). As the Court notes, Atascadero was an Eleventh Amendment case; the “constitutional balance” to which Atascadero refers is that struck by the Eleventh Amendment as this Court has come to interpret it. Although the Court apparently wishes it were otherwise, the principle of interpretation that Atascadero announced is unique to cases involving the Eleventh Amendment. Where the Eleventh Amendment applies, the Court has devised a clear-statement principle more robust than its requirement of clarity in any other situation. Indeed, just today, the Court has intimated that this clear-statement principle is not simply a means of discerning congressional intent. See Dellmuth v. Muth, post, at 232 (concluding that one may not rely on a “permissible inference” from a statute’s language and structure in finding abrogation of immunity); post, 76 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. at 238-239 (Brennan, J., dissenting); but see Pennsylvania v. Union Gas Co., ante, p. 1. Since this case was brought in state court, however, this strict drafting requirement has no application here. The Eleventh Amendment can hardly be “a consideration,” ante, at 67, in a suit to which it does not apply. That this Court has generated a uniquely daunting requirement of clarity in Eleventh Amendment cases explains why Quern v. Jordan, 440 U. S. 332 (1979), did not decide the question before us today. Because only the Eleventh Amendment permits use of this clear-statement principle, the holding of Quern v. Jordan that § 1983 does not abrogate States’ Eleventh Amendment immunity tells us nothing about the meaning of the term “person” in § 1983 as a matter of ordinary statutory construction. Quern’s conclusion thus does not compel, or even suggest, a particular result today. The singularity of this Court’s approach to statutory interpretation in Eleventh Amendment cases also refutes the Court’s argument that, given Quern’s holding, it would make no sense to construe §1983 to include States as “persons.” See ante, at 66. This is so, the Court suggests, because such a construction would permit suits against States in state but not federal court, even though a major purpose of Congress in enacting § 1983 was to provide a federal forum for litigants who had been deprived of their constitutional rights. See, e. g., Monroe v. Pape, 365 U. S. 167 (1961). In answering the question whether § 1983 provides a federal forum for suits against the States themselves, however, one must apply the clear-statement principle reserved for Eleventh Amendment cases. Since this principle is inapplicable to suits brought in state court, and inapplicable to the question whether States are among those subject to a statute, see Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279, 287 (1973); Atascadero, supra, at 240, n. 2, the answer to the question whether § 1983 provides a federal forum for suits against the States may be, and most often will WILL v. MICHIGAN DEPT. OF STATE POLICE 77 58 Brennan, J., dissenting be, different from the answer to the kind of question before us today. Since the question whether Congress has provided a federal forum for damages suits against the States is answered by applying a uniquely strict interpretive principle, see supra, at 75, the Court should not pretend that we have, in Quern, answered the question whether Congress intended to provide a federal forum for such suits, and then reason backwards from that “intent” to the conclusion that Congress must not have intended to allow such suits to proceed in state court. In short, the only principle of statutory interpretation that permits the Court to avoid a careful and thorough analysis of § 1983’s language and history is the clear-statement principle that this Court has come to apply in Eleventh Amendment cases—a principle that is irrelevant to this state-court action. In my view, a careful and detailed analysis of § 1983 leads to the conclusion that States are “persons” within the meaning of that statute. II Section 1983 provides: “Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory or the District of Columbia, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in an action at law, suit in equity, or other proper proceeding for redress.” Although § 1983 itself does not define the term “person,” we are not without a statutory definition of this word. “Any analysis of the meaning of the word ‘person’ in § 1983 . . . must begin . . . with the Dictionary Act.” Monell v. New York City Dept, of Social Services, 436 U. S. 658, 719 (1978) (Rehnquist, J., dissenting). Passed just two months be 78 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. fore § 1983, and designed to “suppl[y] rules of construction for all legislation,” ibid., the Dictionary Act provided: “That in all acts hereafter passed . . . the word ‘person’ may extend and be applied to bodies politic and corporate . . . unless the context shows that such words were intended to be used in a more limited sense . . . .” Act of Feb. 25, 1871, §2, 16 Stat. 431. In Monell, we held this definition to be not merely allowable but mandatory, requiring that the word “person” be construed to include “bodies politic and corporate” unless the statute under consideration “by its terms called for a deviation from this practice.” 436 U. S., at 689-690, n. 53. Thus, we concluded, where nothing in the “context” of a particular statute “call[s] for a restricted interpretation of the word ‘person,’ the language of that [statute] should prima facie be construed to include ‘bodies politic’ among the entities that could be sued.” Ibid. Both before and after the time when the Dictionary Act and § 1983 were passed, the phrase “bodies politic and corporate” was understood to include the States. See, e. g., J. Bouvier, 1 A Law Dictionary Adapted to the Constitution and Laws of the United States of America 185 (11th ed. 1866); W. Shumaker & G. Longsdorf, Cyclopedic Dictionary of Law 104 (1901); Chisholm v. Georgia, 2 Dall. 419, 447 (1793) (Iredell, J.); id., at 468 (Cushing, J.); Cotton v. United States, 11 How. 229, 231 (1851) (“Every sovereign State is of necessity a body politic, or artificial person”); Poindexter v. Greenhow, 114 U. S. 270, 288 (1885); McPherson n. Blacker, 146 U. S. 1, 24 (1892); Heim n. McCall, 239 U. S. 175, 188 (1915). See also United States v. Maurice, 2 Brock. 96, 109 (CC Va. 1823) (Marshall, C. J.) (“The United States is a government, and, consequently, a body politic and corporate”); Van Brocklin v. Tennessee, 117 U. S. 151, 154 (1886) (same). Indeed, the very legislators who passed § 1 referred to States in these terms. See, e. g., Cong. Globe, 42d Cong., 1st Sess., 661-662 (1871) (Sen. Vickers) (“What is a State? Is WILL v. MICHIGAN DEPT. OF STATE POLICE 79 58 Brennan, J., dissenting it not a body politic and corporate?”); id., at 696 (Sen. Edmunds) (“A State is a corporation”). The reason why States are “bodies politic and corporate” is simple: just as a corporation is an entity that can act only through its agents, “[t]he State is a political corporate body, can act only through agents, and can command only by laws.” Poindexter v. Greenhow, supra, at 288. See also Black’s Law Dictionary 159 (5th ed. 1979) (“[B]ody politic or corporate”: “A social compact by which the whole people covenants with each citizen, and each citizen with the whole people, that all shall be governed by certain laws for the common good”). As a “body politic and corporate,” a State falls squarely within the Dictionary Act’s definition of a “person.” While it is certainly true that the phrase “bodies politic and corporate” referred to private and public corporations, see ante, at 69, and n. 9, this fact does not draw into question the conclusion that this phrase also applied to the States. Phrases may, of course, have multiple referents. Indeed, each and every dictionary cited by the Court accords a broader realm—one that comfortably, and in most cases explicitly, includes the sovereign—to this phrase than the Court gives it today. See 1 B. Abbott, Dictionary of Terms and Phrases Used in American or English Jurisprudence 155 (1879) (“[T]he term body politic is often used in a general way, as meaning the state or the sovereign power, or the city government, without implying any distinct express incorporation”); W. Anderson, A Dictionary of Law 127 (1893) (“[B]ody politic”: “The governmental, sovereign power: a city or a State”); Black’s Law Dictionary 143 (1891) (“[B]ody politic”: “It is often used, in a rather loose way, to designate the state or nation or sovereign power, or the government of a county or municipality, without distinctly connoting any express and individual corporate charter”); 1 A. Burrill, A Law Dictionary and Glossary 212 (2d ed. 1871) (“[B]ody politic”: “A body to take in succession, framed by policy”', “[p]articu- 80 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. larly applied, in the old books, to a corporation sole”); id., at 383 (“[C]orporation sole” includes the sovereign in England). Because I recognize that both uses of this phrase were deemed valid when § 1983 and the Dictionary Act were passed, the Court accuses me of “confus[ing] [the] precise definition of [this] phrase with its use ‘in a rather loose way,’ ” “to refer to the state (as opposed to a State).” Ante, at 70, n. 9, quoting Black, supra, at 143. It had never occurred to me, however, that only “precise” definitions counted as valid ones. Where the question we face is what meaning Congress attached to a particular word or phrase, we usually—and properly—are loath to conclude that Congress meant to use the word or phrase in a hypertechnical sense unless it said so. Nor does the Court’s distinction between “the state” and “a State” have any force. The suggestion, I take it, is that the phrase “bodies politic and corporate” refers only to nations rather than to the states within a nation; but then the Court must explain why so many of the sources I have quoted refer to states in addition to nations. In an opinion so utterly devoted to the rights of the States as sovereigns, moreover, it is surprising indeed to find the Court distinguishing between our sovereign States and our sovereign Nation. In deciding what the phrase “bodies politic and corporate” means, furthermore, I do not see the relevance of the meaning of the term “public corporation.” See ante, at 69-70, n. 9. That is not the phrase chosen by Congress in the Dictionary Act, and the Court’s suggestion that this phrase is coterminous with the phrase “bodies politic and corporate” begs the question whether the latter one includes the States. Nor do I grasp the significance of this Court’s decision in United States v. Fox, 94 U. S. 315 (1877), in which the question was whether the State of New York, by including “persons” and “corporations” within the class of those to whom land could be devised, had intended to authorize devises to the United States. Ante, at 69-70, n. 9. Noting that “[t]he question is to be determined by the laws of [New York],” the WILL v. MICHIGAN DEPT. OF STATE POLICE 81 58 Brennan, J., dissenting Court held that it would require “an express definition” to hold that the word “persons” included the Federal Government, and that under state law the term “corporations” applied only to corporations created under the laws of New York. 94 U. S., at 320-321. The pertinence of these statelaw questions to the issue before us today escapes me. Not only do we confront an entirely different, federal statute, but we also have an express statement, in the Dictionary Act, that the word “person” in § 1 includes “bodies politic and corporate.” See also Pfizer Inc. v. India, 434 U. S., at 315, n. 15. The relevance of the fact that §2 of the Civil Rights Act of 1866, 14 Stat. 27,—the model for § 1 of the 1871 Act—was passed before the Dictionary Act, see ante, at 69, similarly eludes me. Congress chose to use the word “person” in the 1871 Act even after it had passed the Dictionary Act, presumptively including “bodies politic and corporate” within the category of “persons.” Its decision to do so—and its failure to indicate in the 1871 Act that the Dictionary Act’s presumption was not to apply—demonstrate that Congress did indeed intend “persons” to include bodies politic and corporate. In addition, the Dictionary Act’s definition of “person” by no means dropped from the sky. Many of the authorities cited above predate both the Dictionary Act and the 1866 Act, indicating that the word “persons” in 1866 ordinarily would have been thought to include “bodies politic and corporate,” with or without the Dictionary Act. This last point helps to explain why it is a matter of small importance that the Dictionary Act’s definition of “person” as including bodies politic and corporate was retroactively withdrawn when the federal statutes were revised in 1874. See T. Durant, Report to Joint Committee on Revision of Laws 2 (1873). Only two months after presumptively designating bodies politic and corporate as “persons,” Congress chose the word “person” for § 1 of the Civil Rights Act. For the purpose of determining Congress’ intent in using this 82 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. term, it cannot be decisive that, three years later, it withdrew this presumption. In fact, both the majority and dissent in Monel I emphasized the 1871 version of the Dictionary Act, but neither saw fit even to mention the 1874 revision of this statute. 436 U. S., at 688-689, and nn. 51, 53 (opinion for the Court); id., at 719 (Rehnquist, J., dissenting). Even in cases, moreover, where no statutory definition of the word “persons” is available, we have not hesitated to include bodies politic and corporate within that category. See Stanley v. Schwalby, 147 U. S. 508, 517 (1893) (“[T]he word ‘person’ in the statute would include [the States] as a body politic and corporate”); Ohio v. Helvering, 292 U. S. 360, 370 (1934); United States v. Shirey, 359 U. S. 255, 257, n. 2 (1959). Thus, the question before us is whether the presumption that the word “person” in § 1 of the Civil Rights Act of 1871 included bodies politic and corporate—and hence the States—is overcome by anything in the statute’s language and history. Certainly nothing in the statutory language overrides this presumption. The statute is explicitly directed at action taken “under color of” state law, and thus supports rather than refutes the idea that the “persons” mentioned in the statute include the States. Indeed, for almost a century—until Monroe v. Pape, 365 U. S. 167 (1961)—it was unclear whether the statute applied at all to action not authorized by the State, and the enduring significance of the first cases construing the Fourteenth Amendment, pursuant to which § 1 was passed, lies in their conclusion that the prohibitions of this Amendment do not reach private action. See Civil Rights Cases, 109 U. S. 3 (1883). In such a setting, one cannot reasonably deny the significance of § 1983’s explicit focus on state action. Unimpressed by such arguments, the Court simply asserts that reading “States” where the statute mentions “person” would be “decidedly awkward.” Ante, at 64. The Court does not describe the awkwardness that it perceives, but I take it that its objection is that the under-color-of-law WILL v. MICHIGAN DEPT. OF STATE POLICE 83 58 Brennan, J., dissenting requirement would be redundant if States were included in the statute because States necessarily act under color of state law. But § 1983 extends as well to natural persons, who do not necessarily so act; in order to ensure that they would be liable only when they did so, the statute needed the under-color-of-law requirement. The only way to remove the redundancy that the Court sees would have been to eliminate the catchall phrase “person” altogether, and separately describe each category of possible defendants and the circumstances under which they might be liable. I cannot think of a situation not involving the Eleventh Amendment, however, in which we have imposed such an unforgiving drafting requirement on Congress. Taking the example closest to this case, we might have observed in Monell that § 1983 was clumsily written if it included municipalities, since these, too, may act only under color of state authority. Nevertheless, we held there that the statute does apply to municipalities. 436 U. S., at 690. Similarly, we have construed the statutory term “white persons” to include “‘corporations, companies, associations, firms, partnerships, societies, and joint stock companies, as well as individuals,’” see Wilson v. Omaha Tribe, 442 U. S., at 666, quoting 1 U. S. C. § 1, despite the evident awkwardness in doing so. Indeed, virtually every time we construe the word “person” to include corporate or other artificial entities that are not individual, flesh-and-blood persons, some awkwardness results. But given cases like Monell and Wilson, it is difficult to understand why mere linguistic awkwardness should control where there is good reason to accept the “awkward” reading of a statute. The legislative history and background of the statute confirm that the presumption created by the Dictionary Act was not overridden in § 1 of the 1871 Act, and that, even without such a presumption, it is plain that “person” in the 1871 Act must include the States. I discussed in detail the legislative history of this statute in my opinion concurring in the judg 84 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. ment in Quern v. Jordan, 440 U. S., at 357-365, and I shall not cover that ground again here. Suffice it to say that, in my view, the legislative history of this provision, though spare, demonstrates that Congress recognized and accepted the fact that the statute was directed at the States themselves. One need not believe that the statute satisfies this Court’s heightened clear-statement principle, reserved for Eleventh Amendment cases, in order to conclude that the language and legislative history of § 1983 show that the word “person” must include the States. As to the more general historical background of § 1, we too easily forget, I think, the circumstances existing in this country when the early civil rights statutes were passed. “[V]iewed against the events and passions of the time,” United States v. Price, 383 U. S. 787, 803 (1966), I have little doubt that § 1 of the Civil Rights Act of 1871 included States as “persons.” The following brief description of the Reconstruction period is illuminating: “The Civil War had ended in April 1865. Relations between Negroes and whites were increasingly turbulent. Congress had taken control of the entire governmental process in former Confederate States. It had declared the governments in 10 ‘unreconstructed’ States to be illegal and had set up federal military administrations in their place. Congress refused to seat representatives from these States until they had adopted constitutions guaranteeing Negro suffrage, and had ratified the Fourteenth Amendment. Constitutional conventions were called in 1868. Six of the 10 States fulfilled Congress’ requirements in 1868, the other four by 1870. “For a few years ‘radical’ Republicans dominated the governments of the Southern States and Negroes played a substantial political role. But countermeasures were swift and violent. The Ku Klux Klan was organized by southern whites in 1866 and a similar organization appeared with the romantic title of the Knights of the WILL v. MICHIGAN DEPT. OF STATE POLICE 85 58 Brennan, J., dissenting White Camellia. In 1868 a wave of murders and assaults was launched including assassinations designed to keep Negroes from the polls. The States themselves were helpless, despite the resort by some of them to extreme measures such as making it legal to hunt down and shoot any disguised man. “Within the Congress pressures mounted in the period between the end of the war and 1870 for drastic measures. A few months after the ratification of the Thirteenth Amendment on December 6, 1865, Congress, on April 9, 1866, enacted the Civil Rights Act of 1866 .... On June 13, 1866, the Fourteenth Amendment was proposed, and it was ratified in July 1868. In February 1869 the Fifteenth Amendment was proposed, and it was ratified in February 1870. On May 31, 1870, the Enforcement Act of 1870 was enacted.” Id., at 803-805 (footnotes omitted). This was a Congress in the midst of altering the “ ‘balance between the States and the Federal Government.”’ Ante, at 65, quoting Atascadero State Hospital v. Scanlon, 473 U. S., at 242. It was fighting to save the Union, and in doing so, it transformed our federal system. It is difficult, therefore, to believe that this same Congress did not intend to include States among those who might be liable under § 1983 for the very deprivations that were threatening this Nation at that time. Ill To describe the breadth of the Court’s holding is to demonstrate its unwisdom. If States are not “persons” within the meaning of § 1983, then they may not be sued under that statute regardless of whether they have consented to suit. Even if, in other words, a State formally and explicitly consented to suits against it in federal or state court, no § 1983 plaintiff could proceed against it because States are not within the statute’s category of possible defendants. 86 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. This is indeed an exceptional holding. Not only does it depart from our suggestion in Alabama v. Pugh, 438 U. S. 781, 782 (1978), that a State could be a defendant under § 1983 if it consented to suit, see also Quern v. Jordan, supra, at 340, but it also renders ineffective the choices some States have made to permit such suits against them. See, e. g., Della Grotta v. Rhode Island, 781 F. 2d 343 (CAI 1986). I do not understand what purpose is served, what principle of federalism or comity is promoted, by refusing to give force to a State’s explicit consent to suit. The Court appears to be driven to this peculiar result in part by its view that “in enacting § 1983, Congress did not intend to override well-established immunities or defenses under the common law.” Ante, at 67. But the question whether States are “persons” under § 1983 is separate and distinct from the question whether they may assert a defense of common-law sovereign immunity. In our prior decisions involving common-law immunities, we have not held that the existence of an immunity defense excluded the relevant state actor from the category of “persons” liable under § 1983, see, e. g., Forrester v. White, 484 U. S. 219 (1988), and it is a mistake to do so today. Such an approach entrenches the effect of common-law immunity even where the immunity itself has been waived. For my part, I would reverse the judgment below and remand for resolution of the question whether Michigan would assert common-law sovereign immunity in defense to this suit and, if so, whether that assertion of immunity would preclude the suit. Given the suggestion in the court below that Michigan enjoys no common-law immunity for violations of its own Constitution, Smith v. Department of Public Health, 428 Mich. 540, 641-642, 410 N. W. 2d 749, 793-794 (1987) (Boyle, J., concurring), there is certainly a possibility that that court would hold that the State also lacks immunity against § 1983 suits for violations of the Federal Constitution. WILL v. MICHIGAN DEPT. OF STATE POLICE 87 58 Stevens, J., dissenting Moreover, even if that court decided that the State’s waiver of immunity did not apply to § 1983 suits, there is a substantial question whether Michigan could so discriminate between virtually identical causes of action only on the ground that one was a state suit and the other a federal one. Cf. Testa v. Katt, 330 U. S. 386 (1947); Martinez v. California, 444 U. S. 277, 283, n. 7 (1980). Finally, even if both of these questions were resolved in favor of an immunity defense, there would remain the question whether it would be reasonable to attribute to Congress an intent to allow States to decide for themselves whether to take cognizance of § 1983 suits brought against them. Cf. Martinez, supra, at 284, and n. 8; Owen v. City of Independence, 445 U. S. 622, 647-648 (1980). Because the court below disposed of the case on the ground that States were not “persons” within the meaning of § 1983, it did not pass upon these difficult and important questions. I therefore would remand this case to the state court to resolve these questions in the first instance. Justice Stevens, dissenting. Legal doctrines often flourish long after their raison d’etre has perished.1 The doctrine of sovereign immunity rests on the fictional premise that the “King can do no wrong.”2 Even though the plot to assassinate James I in 1605, the exe 1 “A very common phenomenon, and one very familiar to the student of history, is this. The customs, beliefs, or needs of a primitive time establish a rule or a formula. In the course of centuries the custom, belief, or necessity disappears, but the rule remains. The reason which gave rise to the rule has been forgotten and ingenious minds set themselves to inquire how it is to be accounted for. Some ground of policy is thought of, which seems to explain it and to reconcile it with the present state of things; and then the rule adapts itself to the new reasons which have been found for it, and enters on a new career. The old form receives a new content, and in time even the form modifies itself to fit the meaning which it has received.” 0. Holmes, The Common Law 8 (M. Howe ed. 1963). 2 See 1 W. Blackstone, Commentaries *246 (“The king, moreover, is not only incapable of doing wrong, but even of thinking wrong; he can never mean to do an improper thing”). 88 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. cution of Charles I in 1649, and the Colonists’ reaction to George Ill’s stamp tax made rather clear the fictional character of the doctrine’s underpinnings, British subjects found a gracious means of compelling the King to obey the law rather than simply repudiating the doctrine itself. They held his advisers and his agents responsible.3 In our administration of § 1983, we have also relied on fictions to protect the illusion that a sovereign State, absent consent, may not be held accountable for its delicts in federal court. Under a settled course of decision, in contexts ranging from school desegregation to the provision of public 3 In the first chapter of his classic History of England, published in 1849, Thomas Macaulay wrote: “Of these kindred constitutions the English was, from an early period, justly reputed the best. The prerogatives of the sovereign were undoubtedly extensive. “But his power, though ample, was limited by three great constitutional principles, so ancient that none can say when they began to exist, so potent that their natural development, continued through many generations, has produced the order of things under which we now live. “First, the King could not legislate without the consent of his Parliament. Secondly, he could impose no tax without the consent of his Parliament. Thirdly, he was bound to conduct the executive administration according to the laws of the land, and, if he broke those laws, his advisers and his agents were responsible.” 1 T. Macaulay, History of England 28-29. In the United States as well, at the time of the passage of the Civil Rights Act of 1871, actions against agents of the sovereign were the means by which the State, despite its own immunity, was required to obey the law. See, e. g., Poindexter v. Greenhow, 114 U. S. 270, 297 (1885) (“The fancied inconvenience of an interference with the collection of its taxes by the govenment of Virginia, by suits against its tax collectors, vanishes at once upon the suggestion that such interference is not possible, except when that government seeks to enforce the collection of its taxes contrary to the law and contract of the State, and in violation of the Constitution of the United States”); Davis v. Gray, 16 Wall. 203, 220 (1873) (“Where the State is concerned, the State should be made a party, if it could be done. That it cannot be done is a sufficient reason for the omission to do it, and the court may proceed to decree against the officers of the State in all respects as if the State were a party to the record”). WILL v. MICHIGAN DEPT. OF STATE POLICE 89 58 Stevens, J., dissenting assistance benefits to the administration of prison systems and other state facilities, we have held the States liable under § 1983 for their constitutional violations through the artifice of naming a public officer as a nominal party. Once one strips away the Eleventh Amendment overlay applied to actions in federal court, it is apparent that the Court in these cases has treated the State as the real party in interest both for the purposes of granting prospective and ancillary relief and of denying retroactive relief. When suit is brought in state court, where the Eleventh Amendment is inapplicable, it follows that the State can be named directly as a party under § 1983. An official-capacity suit is the typical way in which we have held States responsible for their duties under federal law. Such a suit, we have explained, “ ‘generally represent[s] only another way of pleading an action against an entity of which an officer is an agent.’” Kentucky n. Graham, 473 U. S. 159, 165 (1985) (quoting Monell v. New York City Dept, of Social Services, 436 U. S. 658, 690, n. 55 (1978)); see also Pennhurst State School and Hospital v. Halderman, 465 U. S. 89, 101 (1984). In the peculiar Eleventh Amendment analysis we have applied to such cases, we have recognized that an official-capacity action is in reality always against the State and balanced interests to determine whether a particular type of relief is available. The Court has held that when a suit seeks equitable relief or money damages from a state officer for injuries suffered in the past, the interests in compensation and deterrence are insufficiently weighty to override the State’s sovereign immunity. See Papasan n. Allain, 478 U. S. 265, 278 (1986); Green v. Mansour, 474 U. S. 64, 68 (1985); Edelman v. Jordan, 415 U. S. 651, 668 (1974). On the other hand, although prospective relief awarded against a state officer also “implicate[s] Eleventh Amendment concerns,” Mansour, 474 U. S., at 68, the interests in “end[ing] a continuing violation of federal law,” ibid., outweigh the interests in state sovereignty and justify 90 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. an award under § 1983 of an injunction that operates against the State’s officers or even directly against the State itself. See, e. g., Papas an, supra, at 282; Quern v. Jordan, 440 U. S. 332, 337 (1979); Milliken v. Bradley, 433 U. S. 267, 289 (1977). In Milliken n. Bradley, supra, for example, a unanimous Court upheld a federal-court order requiring the State of Michigan to pay $5,800,000 to fund educational components in a desegregation decree “notwithstanding [its] direct and substantial impact on the state treasury.” Id., at 289 (emphasis added).4 As Justice Powell stated in his opinion concurring in the judgment, “the State [had] been adjudged a participant in the constitutional violations, and the State therefore may be ordered to participate prospectively in a remedy otherwise appropriate.” Zd.,at295. Subsequent decisions have adhered to the position that equitable relief—even “a remedy that might require the expenditure of state funds,” Papasan, supra, at 282—may be awarded to ensure future compliance by a State with a substantive federal question determination. See also Quern v. Jordan, 440 U. S., at 337. Our treatment of States as “persons” under § 1983 is also exemplified by our decisions holding that ancillary relief, such as attorney’s fees, may be awarded directly against the State. We have explained that “liability on the merits and responsibility for fees go hand in hand; where a defendant has not been prevailed against, either because of legal immunity 4 We noted in Hutto v. Finney, 437 U. S. 678, 692, n. 20 (1978): “In Milliken n. Bradley, [433 U. S. 267 (1977)], we affirmed an order requiring a state treasurer to pay a substantial sum to another litigant, even though the District Court’s opinion explicitly recognized that ‘this remedial decree will be paid for by the taxpayers of the City of Detroit and the State of Michigan,’ App. to Pet. for Cert, in Milliken v. Bradley, O. T. 1976, No. 76-447, pp. 116a-117a, and even though the Court of Appeals, in affirming, stated that ‘the District Court ordered that the State and Detroit Board each pay one-half the costs’ of relief. Bradley v. Milliken, 540 F. 2d 229, 245 (CA6 1976).” WILL v. MICHIGAN DEPT. OF STATE POLICE 91 58 Stevens, J., dissenting or on the merits, § 1988 does not authorize a fee award against that defendant.” Kentucky v. Graham, supra, at 165. Nonetheless, we held in Hutto v. Finney, 437 U. S. 678 (1978), a case challenging the administration of the Arkansas prison system, that a Federal District Court could award attorney’s fees directly against the State under §1988,5 id., at 700; see Brandon v. Holt, 469 U. S. 464, 472 (1985), and could assess attorney’s fees for bad-faith litigation under § 1983 “ ‘to be paid out of Department of Corrections funds.’” 437 U. S., at 692. In Supreme Court of Virginia v. Consumers Union of United States, Inc., 446 U. S. 719, 739 (1980), Justice White reaffirmed for a unanimous Court that an award of fees could be entered against a State or state agency, in that case a State Supreme Court, in an injunctive action under § 1983.6 In suits commenced in state court, in which there is no independent reason to require parties to sue nominally a state officer, we have held that attor 5 We explained that the legislative history evinced Congress’ intent that attorney’s fees be assessed against the State: “The legislative history is equally plain: ‘[I]t is intended that the attorneys’ fees, like other items of costs, will be collected either directly from the official, in his official capacity, from funds of his agency or under his control, or from the State or local government (whether or not the agency or government is a named party).’ S. Rep. No. 94-1011, p. 5 (1976) (footnote omitted). The House Report is in accord: ‘The greater resources available to governments provide an ample base from which fees can be awarded to the prevailing plaintiff in suits against governmental officials or entities.’ H. R. Rep. No. 94-1558, p. 7 (1976). The Report added in a footnote that: ‘Of course, the 11th Amendment is not a bar to the awarding of counsel fees against state governments. Fitzpartick v. Bitzer.’ Id., at 7, n. 14. Congress’ intent was expressed in deeds as well as words. It rejected at least two attempts to amend the Act and immunize state and local governments from awards.” Hutto, supra, at 694. “The Court is surely incorrect to assert that a determination that a State is a person under § 1983 was unnecessary to our decisions awarding attorney’s fees against a State or state agency. Ante, at 63, n. 4. If there was no basis for liability because the State or state agency was not a party under § 1983, it is difficult to see how there was a basis for imposition of fees. 92 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. ney’s fees can be awarded against the State in its own name. See Maine v. Thiboutot, 448 U. S. 1, 10-11 (1980).7 The Civil Rights Act of 1871 was “intended to provide a remedy, to be broadly construed, against all forms of official violation of federally protected rights.” Monell v. New York City Dept, of Social Services, 436 U. S., at 700-701. Our holdings that a § 1983 action can be brought against state officials in their official capacity for constitutional violations properly recognize and are faithful to that profound mandate. If prospective relief can be awarded against state officials under § 1983 and the State is the real party in interest in such suits, the State must be a “person” which can be held liable under § 1983. No other conclusion is available. Eleventh Amendment principles may limit the State’s capacity to be sued as such in federal court. See Alabama v. Pugh, 438 U. S. 781 (1978). But since those principles are not applicable to suits in state court, see Thiboutot, supra, at 9, n. 7; Nevada v. Hall, 440 U. S. 410 (1979), there is no need to resort to the fiction of an official-capacity suit and the State may and should be named directly as a defendant in a § 1983 action. The Court concludes, however, that “a state official in his or her official capacity, when sued for injunctive relief, would be a person under § 1983,” ante, at 71, n. 10, while that same party sued in the same official capacity is not a person when the plaintiff seeks monetary relief. It cites in support of this proposition cases such as Osborn v. Bank of United States, 9 Wheat. 738 (1824), in which the Court through Chief Justice Marshall held that an action against a state auditor to recover taxes illegally collected did not constitute an action against the State. This line of authority, the Court states, “would 7 Indeed, we have never questioned that a State is a proper defendant in a § 1983 action when the State has consented to being joined in its own name in a suit in federal court, see Alabama v. Pugh, 438 U. S. 781 (1978), or has been named as a defendant in an action in state court, see Maine v. Thiboutot, 448 U. S. 1 (1980); Martinez v. California, 444 U. S. 277 (1980). WILL v. MICHIGAN DEPT. OF STATE POLICE 93 58 Stevens, J., dissenting not have been foreign to the 19th-century Congress that enacted §1983.” Ante, at 71, n. 10. On the Court’s supposition, the question would be whether the complaint against a state official states a claim for the type of relief sought, not whether it will have an impact on the state treasury. See, e. g., Governor of Georgia n. Madrazo, 1 Pet. 110, 124 (1828). At least for actions in state court, as to which there could be no constitutional reason to look to the effect on the State, see Edelman v. Jordan, 415 U. S. 651 (1974), the Court’s analysis would support actions for the recovery of chattel and real property against state officials both of which were well known in the 19th century. See Poindexter v. Greenhow, 114 U. S. 270 (1884); United States v. Lee, 106 U. S. 196 (1882). Although the conclusion that a state officer sued for damages in his or her official capacity is not a “person” under § 1983 would not quite follow,8 it might nonetheless be permissible to assume that the 1871 Congress did not contemplate an action for damages payable not by the officer personally but by the State. The Court having constructed an edifice for the purposes of the Eleventh Amendment on the theory that the State is always the real party in interest in a § 1983 official-capacity action against a state officer, I would think the majority would be impelled to conclude that the State is a “person” under § 1983. As Justice Brennan has demonstrated, there is also a compelling textual argument that States are persons under § 1983. In addition, the Court’s construction draws an illogical distinction between wrongs committed by county or municipal officials on the one hand, and those committed by state officials on the other. Finally, there is no necessity to 8Cf. City of Kenosha v. Bruno, 412 U. S. 507, 513 (1973) (“We find nothing in the legislative history discussed in Monroe [v. Pape, 365 U. S. 167 (1961)], or in the language actually used by Congress, to suggest that the generic word ‘person’ in § 1983 was intended to have a bifurcated application to municipal corporations depending on the nature of the relief sought against them”). 94 OCTOBER TERM, 1988 491 U. S. Stevens, J., dissenting import into this question of statutory construction doctrine created to protect the fiction that one sovereign cannot be sued in the courts of another sovereign. Aside from all of these reasons, the Court’s holding that a State is not a person under § 1983 departs from a long line of judicial authority based on exactly that premise. I respectfully dissent. QUINN v. MILLSAP 95 Syllabus QUINN et al. v. MILLSAP et al. APPEAL FROM THE SUPREME COURT OF MISSOURI No. 88-1048. Argued April 25, 1989—Decided June 15, 1989 Article VI, § 30, of the Missouri Constitution (hereafter § 30) provides that the governments of the city of St. Louis and St. Louis County may be reorganized by a vote of the electorate upon a plan of reorganization drafted by a “board of freeholders.” The State Circuit Court interpreted “freeholder” as not entailing a condition of property ownership and, with only a tentative discussion of the Equal Protection Clause, entered a declaratory judgment that § 30 is valid both on its face and as applied to the present board of freeholders. The Missouri Supreme Court affirmed, but relied exclusively on its interpretation of the Equal Protection Clause and held that that Clause had no relevancy because the board does not exercise general governmental powers. Held: 1. This Court has jurisdiction over the appeal. Pp. 101-104. 2. The Missouri Supreme Court’s ruling that the Equal Protection Clause had no relevancy to the case because the board of freeholders exercises no general governmental power reflects a significant misreading of this Court’s precedents. The fact that the board serves only to recommend a plan of reorganization to the voters and does not enact any laws of its own cannot immunize it from equal protection scrutiny. Pp. 104-106. 3. A land-ownership requirement for appointment to the board of freeholders violates the Equal Protection Clause, Turner v. Fouche, 396 U. S. 346; Chappelle v. Greater Baton Rouge Airport District, 431 U. S. 159; it is a form of invidious discrimination to require land ownership of all appointees to a body authorized to propose reorganization of local government. Pp. 106-109. 757 S. W. 2d 591, reversed. Blackmun, J., delivered the opinion for a unanimous Court. Kevin M. O’Keefe argued the cause for appellants. With him on the briefs were Charles W. Bobinette, Jess W. Ullom, and Mark D. Mittleman. Simon B. Buckner, Assistant Attorney General of Missouri, argued the cause for appellees. With him on the brief 96 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. were William L. Webster, Attorney General, Thomas W. Wehrle, Andrew J. Minardi, and Eugene P. Freeman. * Justice Blackmun delivered the opinion of the Court. The Constitution of the State of Missouri provides that the governments of the city of St. Louis and St. Louis County may be reorganized by a vote of the electorate of the city and county upon a plan of reorganization drafted by a “board of freeholders.” Appellants contend that this provision violates the Equal Protection Clause of the Fourteenth Amendment to the United States Constitution because it requires that every member of this official board own real property. The Supreme Court of Missouri, without disputing appellants’ premise that ownership of real property is a prerequisite for appointment to the board of freeholders, ruled that “the Equal Protection Clause has no relevancy here” because the board “exercises no general governmental powers.” 757 S. W. 2d 591, 595 (1988). This ruling reflects a significant misreading of our precedents, and, accordingly, we reverse. I In 1987, pursuant to Art. VI, §30, of the Missouri Constitution,1 a sufficient number of voters signed petitions “to * Stanley E. Goldstein, Kathleen L. Wilde, Laughlin McDonald, and Neil Bradley filed a brief for the American Civil Liberties Union of Eastern Missouri et al. as amici curiae urging reversal. ‘Art. VI, §30(a) provides: “The people of the city of St. Louis and the people of the county of St. Louis shall have power (1) to consolidate the territories and governments of the city and county into one political subdivision under the municipal government of the city of St. Louis; or, (2) to extend the territorial boundaries of the county so as to embrace the territory within the city and to reorganize and consolidate the county governments of the city and the county, and adjust their relations as thus united, and thereafter the city may extend its limits in the manner provided by law for other cities; or, (3) to enlarge the present or future limits of the city by annexing thereto part of the territory of the county, and to confer upon the city exclusive jurisdiction of the territory so annexed to the city; or, (4) to establish a metropolitan district or districts for the functional administration of QUINN v. MILLSAP 97 95 Opinion of the Court establish a board of St. Louis area property owners (freeholders)” to consider the reorganization of “governmental structures and responsibilities” for the city and county. App. 20, 30. As a result, under § 30, the city’s mayor and the county executive were required each to appoint nine members to this board, and the Governor was required to appoint one.2 After the mayor had chosen nine individuals based on several criteria, including a history of community service and demonstrated leadership ability, he was informed by the city’s counsel that ownership of real property was a prerequisite for board membership. One of the persons selected by the mayor, the Reverend Paul C. Reinert,3 did not own real property. He was removed from the mayor’s list and replaced with an appointee who satisfied the real-property requirement. The county executive similarly was told by the county’s counsel that real property ownership was a necessary condition for board membership. The Governor also considered services common to the area included therein; or, (5) to formulate and adopt any other plan for the partial or complete government of all or any part of the city and the county. The power so given shall be exercised by the vote of the people of the city and county upon a plan prepared by a board of freeholders consisting of nineteen members, nine of whom shall be electors of the city and nine electors of the county and one an elector of some other county.” 2 Section 30(a) further provides: “Upon the filing with the officials in general charge of elections in the city of a petition proposing the exercise of the powers hereby granted, . . . the mayor shall, with the approval of a majority of the board of aidermen, appoint the city’s nine members of the board, not more than five of whom shall be members of or affiliated with the same political party.” The section contains a similar provision regarding the appointment of the county’s nine members. Section 30(b) provides that “the governor shall appoint one member of the board who shall be a resident of the state, but shall not reside in either the city or the county.” 8 Father Reinert, a Jesuit priest, has been affiliated with St. Louis University since at least 1948. He has served there as professor, dean, president, and university chancellor. See Who’s Who in America 2567 (45th ed. 1988). 98 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. real property ownership as a necessary qualification. Thus, all 19 members appointed to the board of freeholders in 1987 owned real property, as was inevitable given the prevailing belief that § 30 required this result. In November 1987, appellants Robert J. Quinn, Jr., and Patricia J. Kampsen filed in the United States District Court for the Western District of Missouri a class-action complaint on behalf of all Missouri voters who did not own real property. Appellants claimed that §30 violated the Equal Protection Clause of the Fourteenth Amendment on its face, insofar as it required ownership of real property in order to serve on the board that was to consider proposals for reorganizing the St. Louis city and county governments. Quinn v. Missouri, 681 F. Supp. 1422, 1433 (1988). Appellants also claimed that § 30 violated the Equal Protection Clause as applied, because in this instance “appointment to the board [of freeholders] was actually limited to those who were ascertained to be owners of real property.” Ibid. Relying on this Court’s decisions in Turner n. Fouche, 396 U. S. 346 (1970), and Chappelle v. Greater Baton Rouge Airport Dist., 431 U. S. 159 (1977), appellants asserted that the requirement that members of the board own real property—whether contained within § 30 itself or resulting from a misinterpretation of that provision—is not rationally related to any legitimate state purpose. Appellants’ federal-court complaint, as amended, named as defendants the mayor, the county executive, the Governor, and the members of the board of freeholders, as well as the State of Missouri itself. These defendants, all appellees here, in turn sued appellants in a Missouri Circuit Court for a declaratory judgment that § 30 does not violate the Federal Constitution. Appellants counterclaimed in the state court, raising the same claims they presented in their federal-court complaint. QUINN v. MILLSAP 99 95 Opinion of the Court Once the property qualification issue became embroiled in litigation, the official view of §30 changed. Whereas the mayor, the county executive, and the Governor all had assumed during the appointment process that ownership of real property was a prerequisite for board membership, they (together with the other appellees) have argued in court that the use of the term “freeholder” in §30—contrary to its generally accepted meaning—does not entail a condition of property ownership. Because § 30(a) states that “a board of freeholders” shall consist of “nine . . . electors of the city and nine electors of the county and one . . . elector of some other county,” appellees contend that the only qualification necessary for appointment to a board of freeholders is that one be an “elector” of a relevant jurisdiction. Based on their contention that the meaning of “freeholder” in § 30 is an unsettled question of state law, appellees urged the Federal District Court to abstain from adjudicating the merits of appellants’ complaint while the state-court proceeding was pending. The District Court refused to abstain, 681 F. Supp., at 1427-1432, finding appellees’ interpretation of the term “freeholder” to be “strained at best,” id., at 1430, and contrary both to the generally recognized meaning of the term and to its use in Missouri decisional law. Reaching the merits of appellants’ constitutional claim, the court agreed with appellants that Turner and Chappelle required the conclusion that §30 (construed to contain a property requirement) violates the Equal Protection Clause. 681 F. Supp., at 1433-1436. The Federal Court of Appeals, after a preliminary order, see 839 F. 2d 425 (CA8 1988), reversed, holding that the District Court should have abstained. App. to Juris. Statement 61; 855 F. 2d 856 (CA8 1988). Thereafter, in an unpublished memorandum, the State Circuit Court adopted appellees’ interpretation of §30. Although in property law the term “freeholder” means some 100 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. one with a fee or similar estate in land, the court reasoned that in “public law” the phrase “board of freeholders” was equivalent to “board of commissioners.” App. to Juris. Statement 17-18. Additionally, the court suggested that, notwithstanding Turner and Chappelle, §30 might not violate the Equal Protection Clause even if it imposes a real-property-ownership requirement. Speculating about a possible rational basis for this, the court suggested that land ownership might enhance the work of the board because one of the issues it faces is whether to change the boundaries between the city and the county. App. to Juris. Statement 19. The court’s discussion of the Equal Protection Clause remained tentative, however, and the court did not specifically explain the constitutionality of § 30 as applied to the present board of freeholders. Nonetheless, in an order accompanying its memorandum, the state court entered a declaratory judgment that §30 is valid both on its face and as applied to the present board. Id., at 20-21.4 The Missouri Supreme Court affirmed this judgment, but relied exclusively on its interpretation of the Equal Protection Clause. The court did not address the argument that § 30 does not impose a property-ownership requirement, except to say: “We recognize membership on the Board of Freeholders was restricted to owners of real property.” 757 S. W. 2d, at 595. The court continued: “However, we hold that the composition of the Board of Freeholders does not violate the Equal Protection Clause because the Board of Freeholders does not exercise general governmental powers.” Ibid. Thus, the Missouri Supreme Court rejected both the facial and as-applied challenges to § 30 based on its belief that the Equal Protection Clause was inapplicable to the board of freeholders. 4 In its order, the state court also certified as defendants the class of all Missouri voters who do not own real property. App. to Juris. Statement 20. Appellants Quinn and Kampsen have appealed, as class representatives, the declaratory judgment against the class. QUINN v. MILLSAP 101 95 Opinion of the Court Contesting the Missouri Supreme Court’s interpretation of the Equal Protection Clause, appellants filed the appeal now before us, and we noted probable jurisdiction. 489 U. S. 1009 (1989).5 II Appellees dispute this Court’s power to hear the appeal, offering four separate arguments in an attempt to avoid a decision on the merits. First, in an effort to rely on the adequate and independent state ground doctrine, see Fox Film Corp. v. Muller, 296 U. S. 207, 210 (1935), appellees would persuade us that the Missouri Supreme Court actually accepted their interpretation of § 30. They point to the following passage from that court’s opinion: “Following certification of the petitions, section 30 required both the mayor of St. Louis and the county supervisor of St. Louis County to appoint nine ‘electors’ to the Board. In addition the Governor of Missouri was required to appoint one elector to the Board.” 757 S. W. 2d, at 592 (footnote omitted). This passage, in the introductory section of the opinion, simply repeats the language of §30 itself. See n. 1, supra. It cannot reasonably be considered as a holding that “freeholder” means no more than “elector” and that ownership of real property is not a prerequisite for sitting on the board of freeholders. We are not convinced that the Missouri Supreme Court interpreted § 30 as urged by appellees. Rather, as explained in Part I, supra, the judgment of the Missouri Supreme Court rests solely on its belief that “the Equal Protection Clause has no relevancy” to this case. 757 S. W. 2d, at 595. In these circumstances, there can be no dispute about our power to consider the federal issue decided by the state court: “Where the state court does not decide “Since then, the State Circuit Court has stayed a vote, scheduled for June 20, 1989, on a plan proposed by the board of freeholders. Tr. of Oral Arg. 17, 46; Brief for Appellants 11; Brief for Appellees 5. 102 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. against a petitioner or appellant upon an independent state ground, but deeming the federal question to be before it, actually entertains and decides that question adversely to the federal right asserted, this Court has jurisdiction to review the judgment if, as here, it is a final judgment.” Indiana ex rel. Anderson v. Brand, 303 U. S. 95, 98 (1938). “That the [state] court might have, but did not, invoke state law does not foreclose jurisdiction here.” Zacchini n. Scripps-Howard Broadcasting Co., 433 U. S. 562, 568 (1977).6 Appellees’ remaining three jurisdictional arguments are rather surprising given the fact that it was they who brought this declaratory judgment action against appellants. Appellees argue that the validity of § 30 under the Equal Protection Clause is a nonjusticiable political question, although they filed this lawsuit seeking a judicial determination of § 30’s validity under the Federal Constitution. See App. 6. In any event, their political question argument—that the Guarantee Clause7 precludes review of the equal protection issue— was expressly rejected in Baker n. Carr, 369 U. S. 186, 228 (1962). Next, appellees argue that appellants lack Article III standing to bring this appeal, although appellees stated in their petition for a declaratory judgment that a “controversy” exists between “adverse” parties involving “legally protectable interests.” App. 5. While appellees now might wish to repudiate this view, we have no doubt that the appeal “re 6 Moreover, the passage cited by appellees certainly does not qualify as a “plain statement” of the court’s reliance on an alternative state-law holding. See Michigan n. Long, 463 U. S. 1032, 1041 (1983). In the absence of such a “plain statement,” we have jurisdiction to review the federal ground on which the Missouri Supreme Court’s judgment rests. Id., at 1042. 7 Art. IV, §4, of the Federal Constitution provides: “The United States shall guarantee to every State in this Union a Republican Form of Government, and shall protect each of them against Invasion; and on Application of the Legislature, or of the Executive (when the Legislature cannot be convened) against domestic Violence.” QUINN v. MILLSAP 103 95 Opinion of the Court tains the essentials of an adversary proceeding, involving a real, not a hypothetical, controversy,” Nashville, C. & St. L. R. Co. v. Wallace, 288 U. S. 249, 264 (1933), and therefore qualifies as a “Cas[e]” for the purposes of Article III, §2. See also ASARCO Inc. v. Kadish, 490 U. S. 605 (1989). Indeed, in Turner v. Fouche, we specifically held that a person who does not own real property has Article III standing to challenge under the Equal Protection Clause a state-law requirement that one own real property in order to serve on a particular government board. 396 U. S., at 361-362, n. 23. Given Turner, appellants necessarily have standing to appeal the Missouri Supreme Court’s determination that, even if Missouri law requires that members of the board of freeholders own real property, the Equal Protection Clause is inapplicable.8 Finally, appellees contend that an adjudication of appellants’ appeal would interfere with the power of executive officials to make discretionary appointments, although, again, they filed this state-court action seeking a declaration of the legal validity of §30 and the present board of freeholders. In any event, the argument is frivolous. Appellees rely on dicta in two cases, in which this Court suggested that federal district courts might lack the authority to order executive officials to make discretionary appointments in a particular way. See Mayor of Philadelphia v. Educational Equality League, 415 U. S. 605, 615 (1974); Carter v. Jury Comm’n of 8 Appellees concede that under Turner appellants have standing to appeal insofar as they challenge the facial validity of § 30. Appellees contend, however, that appellants lack Article III standing insofar as they challenge § 30 as applied. Brief for Appellees 27. This contention is beside the point, however, since the federal question decided by the Missouri Supreme Court—whether the board of freeholders is exempt from equal protection scrutiny—concerns the validity of § 30 on its face, in addition to its validity as applied. Thus, as long as appellants have Article III standing to challenge the facial validity of § 30 (as they undoubtedly do under Turner), they have sufficient standing to appeal the judgment of the Missouri Supreme Court in this case. 104 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Greene County, 396 U. S. 320, 338 (1970). Whatever the limits of a federal court’s power to remedy violations of the Equal Protection Clause, however, those limits are plainly irrelevant when this Court is asked to review a state-court judgment that no violation of the Equal Protection Clause has occurred or, as here, that the Equal Protection Clause is inapplicable to the state action in question. When a state supreme court denies the existence of a federal right and rests its decision on that basis, this Court unquestionably has jurisdiction to review the federal issue decided by the state court. To suggest otherwise would contradict principles laid down in the Judiciary Act of 1789, 1 Stat. 73, 85, and settled since Martin n. Hunter’s Lessee, 1 Wheat. 304 (1816). Satisfied of our jurisdiction over this appeal, we turn to the merits. Ill A In Turner v. Fouche, supra, the Court applied the Equal Protection Clause to a requirement that members of a local school board own real property and held the requirement unconstitutional because it was not rationally related to any legitimate state interest. 396 U. S., at 362-364. Subsequently, we applied the holding in Turner to strike down a requirement of local-property ownership for membership on a local airport commission. Chappelle v. Greater Baton Rouge Airport Dist., 431 U. S. 159 (1977), summarily rev’g 329 So. 2d 810 (La. App. 1976). Here, the Missouri Supreme Court held that “Turner does not control. . . because Turner dealt with a unit of local government which had general governmental powers.” 757 S. W. 2d., at 594. The Missouri Supreme Court, instead, turned to our decisions in Ball v. James, 451 U. S. 355 (1981), Salyer Land Co. n. Tulare Lake Basin Water Storage Dist., 410 U. S. 719 (1973), and Associated Enterprises, Inc. v. Toltec Watershed Improvement Dist., 410 U. S. 743 (1973), believing those decisions to support its conclusion that “the Equal Protection Clause has no QUINN v. MILLSAP 105 95 Opinion of the Court relevancy here.” 757 S. W. 2d, at 595. They do not support that conclusion. In each of these cases, the Court sustained the constitutionality of a water-district voting scheme based on land ownership. But the Court did not reach that result by ruling, as the Missouri Supreme Court held here, that the Equal Protection Clause was irrelevant because of the kind of functions performed by the water-district officials. On the contrary, the Court expressly applied equal protection analysis and concluded that the voting qualifications at issue passed constitutional scrutiny. Ball, 451 U. S., at 371; Salyer, 410 U. S., at 730-731; Toltec, 410 U. S., at 744. Precisely because the water-district cases applied equal protection analysis, they cannot stand for the proposition that the Equal Protection Clause is inapplicable “when the local unit of government in question [has no] general governmental powers.” 757 S. W. 2d, at 595. Thus, the Missouri Supreme Court erred in thinking that the three water-district cases allowed it to avoid an application of the Equal Protection Clause. In holding the board of freeholders exempt from the constraints of the Equal Protection Clause, the Missouri Supreme Court also relied on the fact that the “Board of Freeholders serves only to recommend a plan of reorganization to the voters of St. Louis City and St. Louis County” and does not enact any laws of its own. Ibid. But this fact cannot immunize the board of freeholders from equal protection scrutiny. As this Court in Turner explained, the Equal Protection Clause protects the “right to be considered for public service without the burden of invidiously discriminatory disqualifications.” 396 U. S., at 362. Membership on the board of freeholders is a form of public service, even if the board only recommends a proposal to the electorate and does not enact laws directly. Thus, the Equal Protection Clause protects appellants’ right to be considered for appointment to the board without the burden of “invidiously discriminatory disqualifications.” 106 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. The rationale of the Missouri Supreme Court’s contrary decision would render the Equal Protection Clause inapplicable even to a requirement that all members of the board be white males. This result, and the reasoning that leads to it, are obviously untenable. Thus, we conclude that it is incorrect to say, as that court did, that the Equal Protection Clause does not apply to the board of freeholders because the electorate votes on its proposals and it “does not exercise general governmental powers.” 757 S. W. 2d, at 595. The board in this case—like the school board in Turner and the airport commission in Chappelle—is subject to the constraints of the Equal Protection Clause. B The question, of course, remains whether the land-ownership requirement in this particular case passes or fails equal protection scrutiny. We could remand this question to the Missouri Supreme Court, but there is no good reason to delay the resolution of this issue any further. The parties have briefed and argued the issue throughout this litigation, first in federal court, then in state court, and now in this Court. Cf. Revere v. Massachusetts General Hospital, 463 U. S. 239, 244, n. 6 (1983); Minnesota v. Clover Leaf Creamery Co., 449 U. S. 456, 470-471, n. 14 (1981). Indeed, there already has been an adjudication of the merits of this issue by the United States District Court. Quinn v. Missouri, 681 F. Supp., at 1433-1436.9 Moreover, the resolution of this issue 9 Nor must we remand this issue just because the Missouri Supreme Court failed to settle the parties’ dispute over the meaning of § 30. The court assumed the existence of a land-ownership requirement, as shall we. Our assumption is especially reasonable in the peculiar circumstances of this case. First, the term “freeholder,” when used elsewhere in the Missouri Constitution, carries its usual meaning of land ownership. See, e. g., Shively v. Lankford, 174 Mo. 535, 548, 74 S. W. 835, 838 (1903) (defining “freeholder” to mean “one who owns ‘a freehold estate, that is, an estate in lands, tenements, or hereditaments of an indeterminate duration, other than an estate at will or by sufferance’ ”); see also Tr. of Oral Arg. 48 (conceding that “freeholder” means “owner of real property” for purposes of QUINN v. MILLSAP 107 95 Opinion of the Court is straightforward: it is a form of invidious discrimination to require land ownership of all appointees to a body authorized to propose reorganization of local government. We need apply no more than the rationality review articulated in Turner to reach this conclusion.10 In their brief, appellees offer two justifications for a real-property requirement in this case. First, they contend that owners of real estate have a “first-hand knowledge of the value of good schools, sewer systems and the other problems and amenities of urban life.” Brief for Appellees 41 (footnote omitted). Second, they assert that a real-property owner “has a tangible stake in the long term future of his area.” Ibid. These two arguments, however, were precisely the ones that this Court rejected in Turner itself. other provisions of the Missouri Constitution); see generally Quinn v. Missouri, 681 F. Supp., at 1430-1431 (reviewing Missouri authorities). Second, there is no indication that anyone in Missouri (at least prior to this litigation) understood the term “freeholder” in § 30 to mean something other than its ordinary usage. See Tr. of Oral Arg. 50-51. On the contrary, the mayor, the county executive, and the Governor all made their appointments to the present board of freeholders with a belief that real-property ownership was a necessary qualification for membership on the board, and the petitions to establish the present board of freeholders expressly referred to “a board of St. Louis area property owners (freeholders).” App. 30 (emphasis added). While the Missouri Supreme Court retains the final authority to interpret § 30, we have no substantial reason to believe that appellees’ interpretation might be accepted. Third, even if the appointing officials misinterpreted § 30, the very fact that they did so means, in effect, that all members of the board were required to own real property. Father Reinert, who is a member of the class represented by appellants, was removed from the mayor’s list just because he did not own real property. Accordingly, in the posture that this case comes before this Court, it is appropriate for us to assume that land ownership was a prerequisite for all positions on the board. 10 Because we conclude that a land-ownership requirement for all members of the board of freeholders cannot survive Turner’s rationality review, we need not consider appellants’ argument that a strict standard of review applies by virtue of such cases as Bullock v. Carter, 405 U. S. 134 (1972), and Lubin v. Panish, 415 U. S. 709 (1974). See also Turner, 396 U. S., at 362 (declining to consider whether a higher level of scrutiny applies). 108 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. As to the first, the Court explained that an ability to understand the issues concerning one’s community does not depend on ownership of real property. "It cannot be seriously urged that a citizen in all other respects qualified to sit on a school board must also own real property if he is to participate responsibly in educational decisions.” 396 U. S., at 363-364. Similarly indefensible is the proposition that someone otherwise qualified to sit on the board that proposes a reorganization of St. Louis government must be removed from consideration just because he does not own real property. The Court in Turner also squarely rejected appellees’ second argument by recognizing that persons can be attached to their community without owning real property. “However reasonable the assumption that those who own realty do possess such an attachment, [the State] may not rationally presume that that quality is necessarily wanting in all citizens of the county whose estates are less than freehold.” Id., at 364. Thus, Turner plainly forecloses Missouri’s reliance on this justification for a land-ownership requirement.11 At oral argument, counsel for appellees adopted the suggestion of the State Circuit Court that a land-ownership requirement might be justifiable in this case because the board of freeholders considers issues that may relate to land. Tr. of Oral Arg. 39.12 Of course, the airport commission in Chappelle may have made decisions affecting real estate in its vicinity. Nonetheless, we held in Chappelle that excluding from service on the airport commission anyone who did not own local property was unconstitutional under Turner. Thus, the mere fact that the board of freeholders considers “The absurdity of appellees’ position is vividly demonstrated in this case by the property-based exclusion of Father Reinert, whose long experience as a professor and officer of a local university gave him a sufficient stake in the community and knowledge of local conditions to make him an appropriate choice for appointment to the board. See, n. 3, supra. 12 The State Circuit Court referred specifically to a possible change of boundaries between the city and county. App. to Juris. Statement 19. QUINN v. MILLSAP 109 95 Opinion of the Court land-use issues cannot suffice to sustain a land-ownership requirement in this case. Moreover, the board of freeholders here is unlike any of the governmental bodies at issue in the three water-district cases. Whereas it was rational for the States in those cases to limit voting rights to landowners, Ball, 451 U. S., at 371, the “constitutionally relevant fact” there was “that all water delivered by [those districts was] distributed according to land ownership,” id., at 367. The purpose of the board of freeholders, however, is not so directly linked with land ownership. Cf. id., at 357 (emphasizing “the peculiarly narrow function of [the] local government body” in Ball and its “special relationship” to the class of landowners). Even if the board of freeholders considers land-use issues, the scope of its mandate is far more encompassing: it has the power to draft and submit a plan to reorganize the entire governmental structure of St. Louis city and county. The work of the board of freeholders thus affects all citizens of the city and county, regardless of land ownership. Consequently, Missouri cannot entirely exclude from eligibility for appointment to this board all persons who do not own real property, regardless of their other qualifications and their demonstrated commitment to their community. In sum, we cannot agree with appellees that under the Equal Protection Clause, as previously construed by this Court, landowners alone may be eligible for appointment to a body empowered to propose a wholesale revision of local government. “Whatever objectives” Missouri may wish “to obtain by [a] ‘freeholder’ requirement must be secured, in this instance at least, by means more finely tailored to achieve the desired goal.” Turner, 396 U. S., at 364. Accordingly, a land-ownership requirement is unconstitutional here, just as it was in Turner and in Chappelle. The judgment of the Missouri Supreme Court is reversed. It is so ordered. 110 OCTOBER TERM, 1988 Syllabus 491 U. S. MICHAEL H. ET AL. v. GERALD D. APPEAL FROM THE COURT OF APPEAL OF CALIFORNIA, SECOND APPELLATE DISTRICT No. 87-746. Argued October 11, 1988—Decided June 15, 1989 In May 1981, appellant Victoria D. was bom to Carole D., who was married to, and resided with, appellee Gerald D. in California. Although Gerald was listed as father on the birth certificate and has always claimed Victoria as his daughter, blood tests showed a 98.07% probability that appellant Michael H., with whom Carole had had an adulterous affair, was Victoria’s father. During Victoria’s first three years, she and her mother resided at times with Michael, who held her out as his own, at times with another man, and at times with Gerald, with whom they have lived since June 1984. In November 1982, Michael filed a filiation action in California Superior Court to establish his paternity and right to visitation. Victoria, through her court-appointed guardian ad litem, filed a cross-complaint asserting that she was entitled to maintain filial relationships with both Michael and Gerald. The court ultimately granted Gerald summary judgment on the ground that there were no triable issues of fact as to paternity under Cal. E vid. Code § 621, which provides that a child born to a married woman living with her husband, who is neither impotent nor sterile, is presumed to be a child of the marriage, and that this presumption may be rebutted only by the husband or wife, and then only in limited circumstances. Moreover, the court denied Michael’s and Victoria’s motions for visitation pending appeal under Cal. Civ. Code §4601, which provides that a court may, in its discretion, grant “reasonable visitation rights ... to any . . . person having an interest in the [child’s] welfare.” The California Court of Appeal affirmed, rejecting Michael’s procedural and substantive due process challenges to §621 as well as Victoria’s due process and equal protection claims. The court also rejected Victoria’s assertion of a right to continued visitation with Michael under §4601, on the ground that California law denies visitation against the wishes of the mother to a putative father who has been prevented by § 621 from establishing his paternity. Held: The judgment is affirmed. 191 Cal. App. 3d 995, 236 Cal. Rptr. 810, affirmed. Justice Scalia, joined by The Chief Justice, and in part by Justice O’Connor and Justice Kennedy, concluded that: MICHAEL H. v. GERALD D. Ill 110 Syllabus 1. The §621 presumption does not infringe upon the due process rights of a man wishing to establish his paternity of a child born to the wife of another man. Pp. 118-130. (a) Michael’s contention that procedural due process requires that he be afforded an opportunity to demonstrate his paternity in an evidentiary hearing fundamentally misconceives the nature of § 621. Although phrased in terms of a presumption, §621 expresses and implements a substantive rule of law declaring it to be generally irrelevant for paternity purposes whether a child conceived during, and born into, an existing marriage was begotten by someone other than the husband and had a prior relationship with him, based on the state legislature’s determination as a matter of overriding social policy that the husband should be held responsible for the child and that the integrity and privacy of the family unit should not be impugned. Because Michael’s complaint is that the statute categorically denies all men in his circumstances an opportunity to establish their paternity, his challenge is not accurately viewed as procedural. Pp. 119-121. (b) There is no merit to Michael’s substantive due process claim that he has a constitutionally protected “liberty” interest in the parental relationship he has established with Victoria, and that protection of Gerald’s and Carole’s marital union is an insufficient state interest to support termination of that relationship. Michael has failed to meet his burden of proving that his claimed “liberty” interest is one so deeply imbedded within society’s traditions as to be a fundamental right. Not only has he failed to demonstrate that the interest he seeks to vindicate has traditionally been accorded protection by society, but the common-law presumption of legitimacy, and even modern statutory and decisional law, demonstrate that society has historically protected, and continues to protect, the marital family against the sort of claim Michael asserts. Pp. 121-130. 2. The §621 presumption does not infringe upon any constitutional right of a child to maintain a relationship with her natural father. Victoria’s assertion that she has a due process right to maintain filial relationships with both Michael and Gerald is, at best, the obverse of Michael’s claim and fails for the same reasons. Nor is there any merit to her claim that her equal protection rights have been violated because, unlike her mother and presumed father, she had no opportunity to rebut the presumption of her legitimacy, since the State’s decision to treat her differently from her parents pursues the legitimate end of preventing the disruption of an otherwise peaceful union by the rational means of 112 OCTOBER TERM, 1988 Syllabus 491 U. S. not allowing anyone but the husband or wife to contest legitimacy. Pp. 130-132. Justice Stevens, although concluding that a natural father might have a constitutionally protected interest in his relationship with a child whose mother was married to, and cohabiting with, another man at the time of the child’s conception and birth, also concluded that the California statutory scheme, as applied in this case, is consistent with the Due Process Clause, since it did not deprive Michael of a fair opportunity to prove that he is an “other person having an interest in the welfare of the child” to whom “reasonable visitation rights” may be awarded in the trial judge’s discretion under § 4601. The plurality’s interpretation of § 621 as creating an absolute bar to such a determination is not only an unnatural reading of the statute’s plain language but is also not consistent with the reading given by the courts below and California courts in other cases, all of which, after deciding that the §621 presumption barred a natural father from proving paternity, have nevertheless gone on to consider the separate question whether it would be proper to allow the natural father visitation as an “other person” based on the best interests of the child in the circumstances of the particular case. Here, where the record shows that, after its shaky start, the marriage between Carole and Gerald developed a stability that now provides Victoria with a loving and harmonious family home, there was nothing fundamentally unfair in the trial judge’s exercise of his discretion to allow the mother to decide whether the child’s best interests would be served by allowing the natural father visitation privileges. Pp. 132-136. Scalia, J., announced the judgment of the Court and delivered an opinion, in which Rehnquist, C. J., joined, and in all but n. 6 of which O’Connor and Kennedy, JJ., joined. O’Connor, J., filed an opinion concurring in part, in which Kennedy, J., joined, post, p. 132. Stevens, J., filed an opinion concurring in the judgment, post, p. 132. Brennan, J., filed a dissenting opinion, in which Marshall and Blackmun, JJ., joined, post, p. 136. White, J., filed a dissenting opinion, in which Brennan, J., joined, post, p. 157. Robert A. W. Boraks argued the cause for appellants. With him on the briefs for appellant Michael H. were George Kaufmann, Ronald K. Henry, Paul R. Taskier, and Joel S. Aaronson. Leslie Ellen Shear filed briefs for appellant Victoria D. MICHAEL H. v. GERALD D. 113 110 Opinion of Scalia, J. Larry M. Hoffman argued the cause for appellee. With him on the brief was Glen H. Schwartz. * Justice Scalia announced the judgment of the Court and delivered an opinion, in which The Chief Justice joins, and in all but footnote 6 of which Justice O’Connor and Justice Kennedy join. Under California law, a child born to a married woman living with her husband is presumed to be a child of the marriage. Cal. Evid. Code Ann. § 621 (West Supp. 1989). The presumption of legitimacy may be rebutted only by the husband or wife, and then only in limited circumstances. Ibid. The instant appeal presents the claim that this presumption infringes upon the due process rights of a man who wishes to establish his paternity of a child born to the wife of another man, and the claim that it infringes upon the constitutional right of the child to maintain a relationship with her natural father. I The facts of this case are, we must hope, extraordinary. On May 9, 1976, in Las Vegas, Nevada, Carole D., an international model, and Gerald D., a top executive in a French oil company, were married. The couple established a home in Playa del Rey, California, in which they resided as husband and wife when one or the other was not out of the country on business. In the summer of 1978, Carole became involved in an adulterous affair with a neighbor, Michael H. In September 1980, she conceived a child, Victoria D., who was born on May 11, 1981. Gerald was listed as father on the birth certificate and has always held Victoria out to the world as his * Michael L. Oddenino filed a brief for the National Council for Children’s Rights as amicus curiae urging reversal. Paul Hoffman, Joan Howarth, John A. Powell, Helen Hershkoff, Steven R. Shapiro, and Isabelle Katz Pinzler filed a brief for the American Civil Liberties Union Foundation et al. as amici curiae. 114 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. daughter. Soon after delivery of the child, however, Carole informed Michael that she believed he might be the father. In the first three years of her life, Victoria remained always with Carole, but found herself within a variety of quasifamily units. In October 1981, Gerald moved to New York City to pursue his business interests, but Carole chose to remain in California. At the end of that month, Carole and Michael had blood tests of themselves and Victoria, which showed a 98.07% probability that Michael was Victoria’s father. In January 1982, Carole visited Michael in St. Thomas, where his primary business interests were based. There Michael held Victoria out as his child. In March, however, Carole left Michael and returned to California, where she took up residence with yet another man, Scott K. Later that spring, and again in the summer, Carole and Victoria spent time with Gerald in New York City, as well as on vacation in Europe. In the fall, they returned to Scott in California. In November 1982, rebuffed in his attempts to visit Victoria, Michael filed a filiation action in California Superior Court to establish his paternity and right to visitation. In March 1983, the court appointed an attorney and guardian ad litem to represent Victoria’s interests. Victoria then filed a cross-complaint asserting that if she had more than one psychological or de facto father, she was entitled to maintain her filial relationship, with all of the attendant rights, duties, and obligations, with both. In May 1983, Carole filed a motion for summary judgment. During this period, from March through July 1983, Carole was again living with Gerald in New York. In August, however, she returned to California, became involved once again with Michael, and instructed her attorneys to remove the summary judgment motion from the calendar. For the ensuing eight months, when Michael was not in St. Thomas he lived with Carole and Victoria in Carole’s apartment in Los Angeles and held Victoria out as his daughter. In April 1984, Carole and Michael signed a stipulation that MICHAEL H. v. GERALD D. 115 110 Opinion of Scalia, J. Michael was Victoria’s natural father. Carole left Michael the next month, however, and instructed her attorneys not to file the stipulation. In June 1984, Carole reconciled with Gerald and joined him in New York, where they now live with Victoria and two other children since bom into the marriage. In May 1984, Michael and Victoria, through her guardian ad litem, sought visitation rights for Michael pendente lite. To assist in determining whether visitation would be in Victoria’s best interests, the Superior Court appointed a psychologist to evaluate Victoria, Gerald, Michael, and Carole. The psychologist recommended that Carole retain sole custody, but that Michael be allowed continued contact with Victoria pursuant to a restricted visitation schedule. The court concurred and ordered that Michael be provided with limited visitation privileges pendente lite. On October 19, 1984, Gerald, who had intervened in the action, moved for summary judgment on the ground that under Cal. E vid. Code § 621 there were no triable issues of fact as to Victoria’s paternity. This law provides that “the issue of a wife cohabiting with her husband, who is not impotent or sterile, is conclusively presumed to be a child of the marriage.” Cal. Evid. Code Ann. § 621(a) (West Supp. 1989). The presumption may be rebutted by blood tests, but only if a motion for such tests is made, within two years from the date of the child’s birth, either by the husband or, if the natural father has filed an affidavit acknowledging paternity, by the wife. §§ 621(c) and (d). On January 28, 1985, having found that affidavits submitted by Carole and Gerald sufficed to demonstrate that the two were cohabiting at conception and birth and that Gerald was neither sterile nor impotent, the Superior Court granted Gerald’s motion for summary judgment, rejecting Michael’s and Victoria’s challenges to the constitutionality of §621. The court also denied their motions for continued visitation pending the appeal under Cal. Civ. Code §4601, which provides that a court may, in its discretion, grant “reasonable 116 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. visitation rights ... to any . . . person having an interest in the welfare of the child.” Cal. Civ. Code Ann. §4601 (West Supp. 1989). It found that allowing such visitation would “violat[e] the intention of the Legislature by impugning the integrity of the family unit.” Supp. App. to Juris. Statement A-91. On appeal, Michael asserted, inter alia, that the Superior Court’s application of §621 had violated his procedural and substantive due process rights. Victoria also raised a due process challenge to the statute, seeking to preserve her de facto relationship with Michael as well as with Gerald. She contended, in addition, that as § 621 allows the husband and, at least to a limited extent, the mother, but not the child, to rebut the presumption of legitimacy, it violates the child’s right to equal protection. Finally, she asserted a right to continued visitation with Michael under §4601. After submission of briefs and a hearing, the California Court of Appeal affirmed the judgment of the Superior Court and upheld the constitutionality of the statute. 191 Cal. App. 3d 995, 236 Cal. Rptr. 810 (1987). It interpreted that judgment, moreover, as having denied permanent visitation rights under §4601, regarding that as the implication of the Superior Court’s reliance upon § 621 and upon an earlier California case, Vincent B. v. Joan R., 126 Cal. App. 3d 619, 179 Cal. Rptr. 9 (1981), appeal dism’d, 459 U. S. 807 (1982), which had held that once an assertion of biological paternity is “determined to be legally impossible” under § 621, visitation against the wishes of the mother should be denied under § 4601. 126 Cal. App. 3d, at 627-628, 179 Cal. Rptr., at 13. The Court of Appeal denied Michael’s and Victoria’s petitions for rehearing, and, on July 30, 1987, the California Supreme Court denied discretionary review. On February 29, 1988, we noted probable jurisdiction of the present appeal. 485 U. S. 903. Before us, Michael and Victoria both raise equal protection and due process challenges. We do not reach Michael’s equal protection claim, however, as it MICHAEL H. v. GERALD D. 117 110 Opinion of Scalia, J. was neither raised nor passed upon below. See Bankers Life & Casualty Co. v. Crenshaw, 486 U. S. 71 (1988). II The California statute that is the subject of this litigation is, in substance, more than a century old. California Code of Civ. Proc. § 1962(5), enacted in 1872, provided that “[t]he issue of a wife cohabiting with her husband, who is not impotent, is indisputably presumed to be legitimate.” In 1955, the legislature amended the statute by adding the preface: “Notwithstanding any other provision of law.” 1955 Cal. Stats., ch. 948, p. 1835, §3. In 1965, when California’s Evidence Code was adopted, the statute was codified as §621, with no substantive change except replacement of the word “indisputably” with “conclusively,” 1965 Cal. Stats., ch. 299, §2, pp. 1297, 1308. When California adopted the Uniform Parentage Act, 1975 Cal. Stats., ch. 1244, §11, pp. 3196-3201, codified at Cal. Civ. Code Ann. §7000 et seq. (West 1983), it amended §621 by replacing the word “legitimate” with the phrase “a child of the marriage” and by adding nonsterility to nonimpotence and cohabitation as a predicate for the presumption. 1975 Cal. Stats., ch. 1244, §13, p. 3202. In 1980, the legislature again amended the statute to provide the husband an opportunity to introduce blood-test evidence in rebuttal of the presumption, 1980 Cal. Stats., ch. 1310, p. 4433; and in 1981 amended it to provide the mother such an opportunity, 1981 Cal. Stats., ch. 1180, p. 4761. In their present form, the substantive provisions of the statute are as follows: “§ 621. Child of the marriage; notice of motion for blood tests “(a) Except as provided in subdivision (b), the issue of a wife cohabiting with her husband, who is not impotent or sterile, is conclusively presumed to be a child of the marriage. 118 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. “(b) Notwithstanding the provisions of subdivision (a), if the court finds that the conclusions of all the experts, as disclosed by the evidence based upon blood tests performed pursuant to Chapter 2 (commencing with Section 890) of Division 7 are that the husband is not the father of the child, the question of paternity of the husband shall be resolved accordingly. “(c) The notice of motion for blood tests under subdivision (b) may be raised by the husband not later than two years from the child’s date of birth. “(d) The notice of motion for blood tests under subdivision (b) may be raised by the mother of the child not later than two years from the child’s date of birth if the child’s biological father has filed an affidavit with the court acknowledging paternity of the child. “(e) The provisions of subdivision (b) shall not apply to any case coming within the provisions of Section 7005 of the Civil Code [dealing with artificial insemination] or to any case in which the wife, with the consent of the husband, conceived by means of a surgical procedure.” Ill We address first the claims of Michael. At the outset, it is necessary to clarify what he sought and what he was denied. California law, like nature itself, makes no provision for dual fatherhood. Michael was seeking to be declared the father of Victoria. The immediate benefit he evidently sought to obtain from that status was visitation rights. See Cal. Civ. Code Ann. § 4601 (West 1983) (parent has statutory right to visitation “unless it is shown that such visitation would be detrimental to the best interests of the child”). But if Michael were successful in being declared the father, other rights would follow—most importantly, the right to be considered as the parent who should have custody, Cal. Civ. Code Ann. § 4600 (West 1983), a status which “embrace[s] the sum of parental rights with respect to the rearing of a child, including the child’s care; the right to the child’s services and MICHAEL H. v. GERALD D. 119 110 Opinion of Scalia, J. earnings; the right to direct the child’s activities; the right to make decisions regarding the control, education, and health of the child; and the right, as well as the duty, to prepare the child for additional obligations, which includes the teaching of moral standards, religious beliefs, and elements of good citizenship.” 4 California Family Law § 60.02[l][b] (C. Markey ed. 1987) (footnotes omitted). All parental rights, including visitation, were automatically denied by denying Michael status as the father. While Cal. Civ. Code Ann. § 4601 places it within the discretionary power of a court to award visitation rights to a nonparent, the Superior Court here, affirmed by the Court of Appeal, held that California law denies visitation, against the wishes of the mother, to a putative father who has been prevented by § 621 from establishing his paternity. See 191 Cal. App. 3d, at 1013, 236 Cal. Rptr., at 821, citing Vincent B. v. Joan R., 126 Cal. App. 3d, at 627-628 179 Cal. Rptr., at 13. Michael raises two related challenges to the constitutionality of §621. First, he asserts that requirements of procedural due process prevent the State from terminating his liberty interest in his relationship with his child without affording him an opportunity to demonstrate his paternity in an evidentiary hearing. We believe this claim derives from a fundamental misconception of the nature of the California statute. While § 621 is phrased in terms of a presumption, that rule of evidence is the implementation of a substantive rule of law. California declares it to be, except in limited circumstances, irrelevant for paternity purposes whether a child conceived during, and born into, an existing marriage was begotten by someone other than the husband and had a prior relationship with him. As the Court of Appeal phrased it: “‘The conclusive presumption is actually a substantive rule of law based upon a determination by the Legislature as a matter of overriding social policy, that given a certain relationship between the husband and wife, the husband is to be held responsible for the child, and that 120 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. the integrity of the family unit should not be impugned.’” 191 Cal. App. 3d, at 1005, 236 Cal. Rptr., at 816, quoting Vincent B. n. Joan R., supra, at 623, 179 Cal. Rptr., at 10. Of course the conclusive presumption not only expresses the State’s substantive policy but also furthers it, excluding inquiries into the child’s paternity that would be destructive of family integrity and privacy.1 This Court has struck down as illegitimate certain “irrebuttable presumptions.” See, e. g., Stanley v. Illinois, 405 U. S. 645 (1972); Vlandis v. Kline, 412 U. S. 441 (1973); Cleveland Board of Education n. LaFleur, 414 U. S. 632 (1974). Those holdings did not, however, rest upon procedural due process. A conclusive presumption does, of course, foreclose the person against whom it is invoked from demonstrating, in a particularized proceeding, that applying the presumption to him will in fact not further the lawful governmental policy the presumption is designed to effectuate. But the same can be said of any legal rule that establishes general classifications, whether framed in terms of a presumption or not. In this respect there is no difference between a rule which says that the marital husband shall be irrebuttably presumed to be the father, and a rule which says that the adulterous natural father shall not be recognized as the legal father. Both rules deny someone in Michael’s situation a hearing on whether, in the particular circumstances of his case, California’s policies would best be served by giving him parental rights. Thus, as many commentators have observed, see, e. g., Bezanson, Some Thoughts on the Emerging Irrebuttable Presumption Doctrine, 7 Ind. L. Rev. 644 (1974); Nowak, Realigning 1 In those circumstances in which California allows a natural father to rebut the presumption of legitimacy of a child born to a married woman, e. g., where the husband is impotent or sterile, or where the husband and wife have not been cohabiting, it is more likely that the husband already knows the child is not his, and thus less likely that the paternity hearing will disrupt an otherwise harmonious and apparently exclusive marital relationship. MICHAEL H. v. GERALD D. 121 110 Opinion of Scalia, J. the Standards of Review Under the Equal Protection Guarantee-Prohibited, Neutral, and Permissive Classifications, 62 Geo. L. J. 1071, 1102-1106 (1974); Note, Irrebuttable Presumptions: An Illusory Analysis, 27 Stan. L. Rev. 449 (1975); Note, The Irrebuttable Presumption Doctrine in the Supreme Court, 87 Harv. L. Rev. 1534 (1974), our “irrebuttable presumption” cases must ultimately be analyzed as calling into question not the adequacy of procedures but — like our cases involving classifications framed in other terms, see, e. g., Craig v. Boren, 429 U. S. 190 (1976); Carrington v. Rash, 380 U. S. 89 (1965)—the adequacy of the “fit” between the classification and the policy that the classification serves. See LaFleur, supra, at 652 (Powell, J., concurring in result); Vlandis, supra, at 456-459 (White, J., concurring), 466-469 (Rehnquist, J., dissenting); Weinberger v. Salfi, 422 U. S. 749 (1975). We therefore reject Michael’s procedural due process challenge and proceed to his substantive claim. Michael contends as a matter of substantive due process that, because he has established a parental relationship with Victoria, protection of Gerald’s and Carole’s marital union is an insufficient state interest to support termination of that relationship. This argument is, of course, predicated on the assertion that Michael has a constitutionally protected liberty interest in his relationship with Victoria. It is an established part of our constitutional jurisprudence that the term “liberty” in the Due Process Clause extends beyond freedom from physical restraint. See, e. g., Pierce v. Society of Sisters, 268 U. S. 510 (1925); Meyer v. Nebraska, 262 U. S. 390 (1923). Without that core textual meaning as a limitation, defining the scope of the Due Process Clause “has at times been a treacherous field for this Court,” giving “reason for concern lest the only limits to . . . judicial intervention become the predilections of those who happen at the time to be Members of this Court.” Moore v. East Cleveland, 431 U. S. 494, 502 (1977). The need for restraint has been cogently expressed by Justice White: 122 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. “That the Court has ample precedent for the creation of new constitutional rights should not lead it to repeat the process at will. The Judiciary, including this Court, is the most vulnerable and comes nearest to illegitimacy when it deals with judge-made constitutional law having little or no cognizable roots in the language or even the design of the Constitution. Realizing that the present construction of the Due Process Clause represents a major judicial gloss on its terms, as well as on the anticipation of the Framers . . . , the Court should be extremely reluctant to breathe still further substantive content into the Due Process Clause so as to strike down legislation adopted by a State or city to promote its welfare. Whenever the Judiciary does so, it unavoidably pre-empts for itself another part of the governance of the country without express constitutional authority.” Moore, supra, at 544 (dissenting opinion). In an attempt to limit and guide interpretation of the Clause, we have insisted not merely that the interest denominated as a “liberty” be “fundamental” (a concept that, in isolation, is hard to objectify), but also that it be an interest traditionally protected by our society.2 As we have put it, the Due Process Clause affords only those protections “so rooted in the traditions and conscience of our people as to be ranked as fundamental.” Snyder v. Massachusetts, 291 U. S. 97, 105 (1934) (Cardozo, J.). Our cases reflect “continual insistence upon respect for the teachings of history [and] solid recogni 2 We do not understand what Justice Brennan has in mind by an interest “that society traditionally has thought important . . . without protecting it.” Post, at 140. The protection need not take the form of an explicit constitutional provision or statutory guarantee, but it must at least exclude (all that is necessary to decide the present case) a societal tradition of enacting laws denying the interest. Nor do we understand why our practice of limiting the Due Process Clause to traditionally protected interests turns the Clause “into a redundancy,” post, at 141. Its purpose is to prevent future generations from lightly casting aside important traditional values—not to enable this Court to invent new ones. MICHAEL H. v. GERALD D. 123 110 Opinion of Scalia, J. tion of the basic values that underlie our society. ...” Griswold v. Connecticut, 381 U. S. 479, 501 (1965) (Harlan, J., concurring in judgment). This insistence that the asserted liberty interest be rooted in history and tradition is evident, as elsewhere, in our cases according constitutional protection to certain parental rights. Michael reads the landmark case of Stanley v. Illinois, 405 U. S. 645 (1972), and the subsequent cases of Quilloin v. Walcott, 434 U. S. 246 (1978), Caban v. Mohammed, 441 U. S. 380 (1979), and Lehr n. Robertson, 463 U. S. 248 (1983), as establishing that a liberty interest is created by biological fatherhood plus an established parental relationship-factors that exist in the present case as well. We think that distorts the rationale of those cases. As we view them, they rest not upon such isolated factors but upon the historic respect—indeed, sanctity would not be too strong a term—traditionally accorded to the relationships that develop within the unitary family.3 See Stanley, supra, at 651; Quilloin, supra, at 254-255; Caban, supra, at 389; Lehr, supra, at 261. In Stanley, for example, we forbade the destruction of such a family when, upon the death of the mother, the State had sought to remove children from the custody of a father who had lived with and supported them and their mother for 18 years. As Justice Powell stated for the plurality in Moore v. East Cleveland, supra, at 503: “Our 3 Justice Brennan asserts that only a “pinched conception of‘the family’ ” would exclude Michael, Carole, and Victoria from protection. Post, at 145. We disagree. The family unit accorded traditional respect in our society, which we have referred to as the “unitary family,” is typified, of course, by the marital family, but also includes the household of unmarried parents and their children. Perhaps the concept can be expanded even beyond this, but it will bear no resemblance to traditionally respected relationships—and will thus cease to have any constitutional significance—if it is stretched so far as to include the relationship established between a married woman, her lover, and their child, during a 3-month sojourn in St. Thomas, or during a subsequent 8-month period when, if he happened to be in Los Angeles, he stayed with her and the child. 124 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. decisions establish that the Constitution protects the sanctity of the family precisely because the institution of the family is deeply rooted in this Nation’s history and tradition.” Thus, the legal issue in the present case reduces to whether the relationship between persons in the situation of Michael and Victoria has been treated as a protected family unit under the historic practices of our society, or whether on any other basis it has been accorded special protection. We think it impossible to find that it has. In fact, quite to the contrary, our traditions have protected the marital family (Gerald, Carole, and the child they acknowledge to be theirs) against the sort of claim Michael asserts.4 The presumption of legitimacy was a fundamental principle of the common law. H. Nicholas, Adulturine Bastardy 1 (1836). Traditionally, that presumption could be rebutted only by proof that a husband was incapable of procreation or had had no access to his wife during the relevant period. Id., at 9-10 (citing Bracton, De Legibus et Consuetudinibus Angliae, bk. i, ch. 9, p. 6; bk. ii, ch. 29, p. 63, ch. 32, p. 70 (1569)). As explained by Blackstone, nonaccess could only be proved “if the husband be out of the kingdom of England (or, as the law somewhat loosely phrases it, extra quatuor maria [beyond the four seas]) for above nine months. . . .” 1 Blackstone’s Commentaries 456 (J. Chitty ed. 1826). And, under the common law both in England and here, “neither 4 Justice Brennan insists that in determining whether a liberty interest exists we must look at Michael’s relationship with Victoria in isolation, without reference to the circumstance that Victoria’s mother was married to someone else when the child was conceived, and that that woman and her husband wish to raise the child as their own. See post, at 145-146. We cannot imagine what compels this strange procedure of looking at the act which is assertedly the subject of a liberty interest in isolation from its effect upon other people—rather like inquiring whether there is a liberty interest in firing a gun where the case at hand happens to involve its discharge into another person’s body. The logic of Justice Brennan’s position leads to the conclusion that if Michael had begotten Victoria by rape, that fact would in no way affect his possession of a liberty interest in his relationship with her. MICHAEL H. v. GERALD D. 125 110 Opinion of Scalia, J. husband nor wife [could] be a witness to prove access or nonaccess.” J. Schouler, Law of the Domestic Relations §225, p. 306 (3d ed. 1882); R. Graveson & F. Crane, A Century of Family Law: 1857-1957, p. 158 (1957). The primary policy rationale underlying the common law’s severe restrictions on rebuttal of the presumption appears to have been an aversion to declaring children illegitimate, see Schouler, supra, §225, at 306-307; M. Grossberg, Governing the Hearth 201 (1985), thereby depriving them of rights of inheritance and succession, 2 J. Kent, Commentaries on American Law *175, and likely making them wards of the state. A secondary policy concern was the interest in promoting the “peace and tranquillity of States and families,” Schouler, supra, §225, at 304, quoting Boullenois, Traité des Status, bk. 1, p. 62, a goal that is obviously impaired by facilitating suits against husband and wife asserting that their children are illegitimate. Even though, as bastardy laws became less harsh, “[j]udges in both [England and the United States] gradually widened the acceptable range of evidence that could be offered by spouses, and placed restraints on the ‘four seas rule’ ...[,] the law retained a strong bias against ruling the children of married women illegitimate.” Grossberg, supra, at 202. We have found nothing in the older sources, nor in the older cases, addressing specifically the power of the natural father to assert parental rights over a child born into a woman’s existing marriage with another man. Since it is Michael’s burden to establish that such a power (at least where the natural father has established a relationship with the child) is so deeply embedded within our traditions as to be a fundamental right, the lack of evidence alone might defeat his case. But the evidence shows that even in modern times — when, as we have noted, the rigid protection of the marital family has in other respects been relaxed—the ability of a person in Michael’s position to claim paternity has not been generally acknowledged. For example, a 1957 annotation on the subject: “Who may dispute presumption of legitimacy of 126 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. child conceived or born during wedlock,” 53 A. L. R. 2d 572, shows three States (including California) with statutes limiting standing to the husband or wife and their descendants, one State (Louisiana) with a statute limiting it to the husband, two States (Florida and Texas) with judicial decisions limiting standing to the husband, and two States (Illinois and New York) with judicial decisions denying standing even to the mother. Not a single decision is set forth specifically according standing to the natural father, and “express indications of the nonexistence of any . . . limitation” upon standing were found only “in a few jurisdictions.” Id., at 579. Moreover, even if it were clear that one in Michael’s position generally possesses, and has generally always possessed, standing to challenge the marital child’s legitimacy, that would still not establish Michael’s case. As noted earlier, what is at issue here is not entitlement to a state pronouncement that Victoria was begotten by Michael. It is no conceivable denial of constitutional right for a State to decline to declare facts unless some legal consequence hinges upon the requested declaration. What Michael asserts here is a right to have himself declared the natural father and thereby to obtain parental prerogatives.5 What he must establish, therefore, is not that our society has traditionally allowed a natural father in his circumstances to establish paternity, but that it has traditionally accorded such a father parental rights, or at least has not traditionally denied them. Even if the law in all States had always been that the entire world could chal 5 According to Justice Brennan, Michael does not claim—and in order to prevail here need not claim—a substantive right to maintain a parental relationship with Victoria, but merely the right to “a hearing on the issue” of his paternity. Post, at 156, n. 12. “Michael’s challenge . . . does not depend,” we are told, “on his ability ultimately to obtain visitation rights.” Post, at 147. To be sure it does not depend upon his ability ultimately to obtain those rights, but it surely depends upon his asserting a claim to those rights, which is precisely what Justice Brennan denies. We cannot grasp the concept of a “right to a hearing” on the part of a person who claims no substantive entitlement that the hearing will assertedly vindicate. MICHAEL H. v. GERALD D. 127 110 Opinion of Scalia, J. lenge the marital presumption and obtain a declaration as to who was the natural father, that would not advance Michael’s claim. Thus, it is ultimately irrelevant, even for purposes of determining current social attitudes towards the alleged substantive right Michael asserts, that the present law in a number of States appears to allow the natural father—including the natural father who has not established a relationship with the child—the theoretical power to rebut the marital presumption, see Note, Rebutting the Marital Presumption: A Developed Relationship Test, 88 Colum. L. Rev. 369, 373 (1988). What counts is whether the States in fact award substantive parental rights to the natural father of a child conceived within, and born into, an extant marital union that wishes to embrace the child. We are not aware of a single case, old or new, that has done so. This is not the stuff of which fundamental rights qualifying as liberty interests are made.6 6 Justice Brennan criticizes our methodology in using historical traditions specifically relating to the rights of an adulterous natural father, rather than inquiring more generally “whether parenthood is an interest that historically has received our attention and protection.” Post, at 139. There seems to us no basis for the contention that this methodology is “nove[l],” post, at 140. For example, in Bowers v. Hardwick, 478 U. S. 186 (1986), we noted that at the time the Fourteenth Amendment was ratified all but 5 of the 37 States had criminal sodomy laws, that all 50 of the States had such laws prior to 1961, and that 24 States and the District of Columbia continued to have them; and we concluded from that record, regarding that very specific aspect of sexual conduct, that “to claim that a right to engage in such conduct is ‘deeply rooted in this Nation’s history and tradition’ or ‘implicit in the concept of ordered liberty’ is, at best, facetious.” Id., at 194. In Roe v. Wade, 410 U. S. 113 (1973), we spent about a fifth of our opinion negating the proposition that there was a longstanding tradition of laws proscribing abortion. Id., at 129-141. We do not understand why, having rejected our focus upon the societal tradition regarding the natural father’s rights vis-à-vis a child whose mother is married to another man, Justice Brennan would choose to focus instead upon “parenthood.” Why should the relevant category not be even more general—perhaps “family relationships”; or “personal relationships”; or even “emotional attachments in general”? Though the dissent has no basis for the level of generality it would select, we do: We refer 128 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. In Lehr v. Robertson, a case involving a natural father’s attempt to block his child’s adoption by the unwed mother’s new husband, we observed that “[t]he significance of the biological connection is that it offers the natural father an opportunity that no other male possesses to develop a relationship to the most specific level at which a relevant tradition protecting, or denying protection to, the asserted right can be identified. If, for example, there were no societal tradition, either way, regarding the rights of the natural father of a child adulterously conceived, we would have to consult, and (if possible) reason from, the traditions regarding natural fathers in general. But there is such a more specific tradition, and it unqualifiedly denies protection to such a parent. One would think that Justice Brennan would appreciate the value of consulting the most specific tradition available, since he acknowledges that “[e]ven if we can agree . . . that ‘family’ and ‘parenthood’ are part of the good life, it is absurd to assume that we can agree on the content of those terms and destructive to pretend that we do.” Post, at 141. Because such general traditions provide such imprecise guidance, they permit judges to dictate rather than discern the society’s views. The need, if arbitrary decisionmaking is to be avoided, to adopt the most specific tradition as the point of reference—or at least to announce, as Justice Brennan declines to do, some other criterion for selecting among the innumerable relevant traditions that could be consulted—is well enough exemplified by the fact that in the present case Justice Brennan’s opinion and Justice O’Connor’s opinion, post, p. 132, which disapproves this footnote, both appeal to tradition, but on the basis of the tradition they select reach opposite results. Although assuredly having the virtue (if it be that) of leaving judges free to decide as they think best when the unanticipated occurs, a rule of law that binds neither by text nor by any particular, identifiable tradition is no rule of law at all. Finally, we may note that this analysis is not inconsistent with the result in cases such as Griswold v. Connecticut, 381 U. S. 479 (1965), or Eisenstadt v. Baird, 405 U. S. 438 (1972). None of those cases acknowledged a longstanding and still extant societal tradition withholding the very right pronounced to be the subject of a liberty interest and then rejected it. Justice Brennan must do so here. In this case, the existence of such a tradition, continuing to the present day, refutes any possible contention that the alleged right is “so rooted in the traditions and conscience of our people as to be ranked as fundamental,” Snyder v. Massachusetts, 291 U. S. 97, 105 (1934), or “implicit in the concept of ordered liberty,” Palko v. Connecticut, 302 U. S. 319, 325 (1937). MICHAEL H. v. GERALD D. 129 110 Opinion of Scalia, J. with his offspring,” 463 U. S., at 262, and we assumed that the Constitution might require some protection of that opportunity, id., at 262-265. Where, however, the child is born into an extant marital family, the natural father’s unique opportunity conflicts with the similarly unique opportunity of the husband of the marriage; and it is not unconstitutional for the State to give categorical preference to the latter. In Lehr we quoted approvingly from Justice Stewart’s dissent in Caban v. Mohammed, 441 U. S., at 397, to the effect that although “‘[i]n some circumstances the actual relationship between father and child may suffice to create in the unwed father parental interests comparable to those of the married father,’” “‘the absence of a legal tie with the mother may in such circumstances appropriately place a limit on whatever substantive constitutional claims might otherwise exist.’” 463 U. S., at 260, n. 16. In accord with our traditions, a limit is also imposed by the circumstance that the mother is, at the time of the child’s conception and birth, married to, and cohabitating with, another man, both of whom wish to raise the child as the offspring of their union.7 It is a question of legislative policy and not constitutional law whether 'Justice Brennan chides us for thus limiting our holding to situations in which, as here, the husband and wife wish to raise her child jointly. The dissent believes that without this limitation we would be unable to “rely on the State’s asserted interest in protecting the ‘unitary family’ in denying that Michael and Victoria have been deprived of liberty.” Post, at 147. As we have sought to make clear, however, and as the dissent elsewhere seems to understand, see post, at 139, 140-141,145, 147, we rest our decision not upon our independent “balancing” of such interests, but upon the absence of any constitutionally protected right to legal parentage on the part of an adulterous natural father in Michael’s situation, as evidenced by long tradition. That tradition reflects a “balancing” that has already been made by society itself. We limit our pronouncement to the relevant facts of this case because it is at least possible that our traditions lead to a different conclusion with regard to adulterous fathering of a child whom the marital parents do not wish to raise as their own. It seems unfair for those who disagree with our holding to include among their criticisms that we have not extended the holding more broadly. 130 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. California will allow the presumed parenthood of a couple desiring to retain a child conceived within and born into their marriage to be rebutted. We do not accept Justice Brennan’s criticism that this result “squashes” the liberty that consists of “the freedom not to conform.” Post, at 141. It seems to us that reflects the erroneous view that there is only one side to this controversy—that one disposition can expand a “liberty” of sorts without contracting an equivalent “liberty” on the other side. Such a happy choice is rarely available. Here, to provide protection to an adulterous natural father is to deny protection to a marital father, and vice versa. If Michael has a “freedom not to conform” (whatever that means), Gerald must equivalently have a “freedom to conform.” One of them will pay a price for asserting that “freedom”—Michael by being unable to act as father of the child he has adulter-ously begotten, or Gerald by being unable to preserve the integrity of the traditional family unit he and Victoria have established. Our disposition does not choose between these two “freedoms,” but leaves that to the people of California. Justice Brennan’s approach chooses one of them as the constitutional imperative, on no apparent basis except that the unconventional is to be preferred. IV We have never had occasion to decide whether a child has a liberty interest, symmetrical with that of her parent, in maintaining her filial relationship. We need not do so here because, even assuming that such a right exists, Victoria’s claim must fail. Victoria’s due process challenge is, if anything, weaker than Michael’s. Her basic claim is not that California has erred in preventing her from establishing that Michael, not Gerald, should stand as her legal father. Rather, she claims a due process right to maintain filial relationships with both Michael and Gerald. This assertion merits little discussion, for, whatever the merits of the guardian MICHAEL H. v. GERALD D. 131 110 Opinion of Scalia, J. ad litem’s belief that such an arrangement can be of great psychological benefit to a child, the claim that a State must recognize multiple fatherhood has no support in the history or traditions of this country. Moreover, even if we were to construe Victoria’s argument as forwarding the lesser proposition that, whatever her status vis-à-vis Gerald, she has a liberty interest in maintaining a filial relationship with her natural father, Michael, we find that, at best, her claim is the obverse of Michael’s and fails for the same reasons. Victoria claims in addition that her equal protection rights have been violated because, unlike her mother and presumed father, she had no opportunity to rebut the presumption of her legitimacy. We find this argument wholly without merit. We reject, at the outset, Victoria’s suggestion that her equal protection challenge must be assessed under a standard of strict scrutiny because, in denying her the right to maintain a filial relationship with Michael, the State is discriminating against her on the basis of her illegitimacy. See Gomez v. Perez, 409 U. S. 535, 538 (1973). Illegitimacy is a legal construct, not a natural trait. Under California law, Victoria is not illegitimate, and she is treated in the same manner as all other legitimate children: she is entitled to maintain a filial relationship with her legal parents. We apply, therefore, the ordinary “rational relationship” test to Victoria’s equal protection challenge. The primary rationale underlying §621’s limitation on those who may rebut the presumption of legitimacy is a concern that allowing persons other than the husband or wife to do so may undermine the integrity of the marital union. When the husband or wife contests the legitimacy of their child, the stability of the marriage has already been shaken. In contrast, allowing a claim of illegitimacy to be pressed by the child—or, more accurately, by a court-appointed guardian ad litem—may well disrupt an otherwise peaceful union. Since it pursues a legitimate end by rational means, California’s de 132 OCTOBER TERM, 1988 Stevens, J., concurring in judgment 491 U. S. cision to treat Victoria differently from her parents is not a denial of equal protection. The judgment of the California Court of Appeal is Affirmed. Justice O’Connor, with whom Justice Kennedy joins, concurring in part. I concur in all but footnote 6 of Justice Scalia’s opinion. This footnote sketches a mode of historical analysis to be used when identifying liberty interests protected by the Due Process Clause of the Fourteenth Amendment that may be somewhat inconsistent with our past decisions in this area. See Griswold v. Connecticut, 381 U. S. 479 (1965); Eisenstadt v. Baird, 405 U. S. 438 (1972). On occasion the Court has characterized relevant traditions protecting asserted rights at levels of generality that might not be “the most specific level” available. Ante, at 127-128, n. 6. See Loving n. Virginia, 388 U. S. 1, 12 (1967); Turner v. Safley, 482 U. S. 78, 94 (1987); cf. United States v. Stanley, 483 U. S. 669, 709 (1987) (O’Connor, J., concurring in part and dissenting in part). I would not foreclose the unanticipated by the prior imposition of a single mode of historical analysis. Poe v. Ullman, 367 U. S. 497, 542, 544 (1961) (Harlan, J., dissenting). Justice Stevens, concurring in the judgment. As I understand this case, it raises two different questions about the validity of California’s statutory scheme. First, is Cal. Evid. Code Ann. §621 (West Supp. 1989) unconstitutional because it prevents Michael and Victoria from obtaining a judicial determination that he is her biological father— even if no legal rights would be affected by that determination? Second, does the California statute deny appellants a fair opportunity to prove that Victoria’s best interests would be served by granting Michael visitation rights? On the first issue I agree with Justice Scalia that the Federal Constitution imposes no obligation upon a State to MICHAEL H. v. GERALD D. 133 110 Stevens, J., concurring in judgment “declare facts unless some legal consequence hinges upon the requested declaration.” Ante, at 126. “The actions of judges neither create nor sever genetic bonds.” Lehr v. Robertson, 463 U. S. 248, 261 (1983). On the second issue I do not agree with Justice Scalia’s analysis. He seems to reject the possibility that a natural father might ever have a constitutionally protected interest in his relationship with a child whose mother was married to, and cohabiting with, another man at the time of the child’s conception and birth. I think cases like Stanley v. Illinois, 405 U. S. 645 (1972), and Caban v. Mohammed, 441 U. S. 380 (1979), demonstrate that enduring “family” relationships may develop in unconventional settings. I therefore would not foreclose the possibility that a constitutionally protected relationship between a natural father and his child might exist in a case like this. Indeed, I am willing to assume for the purpose of deciding this case that Michael’s relationship with Victoria is strong enough to give him a constitutional right to try to convince a trial judge that Victoria’s best interest would be served by granting him visitation rights. I am satisfied, however, that the California statute, as applied in this case, gave him that opportunity. Section 4601 of the California Civil Code Annotated (West Supp. 1989) provides: “[R]easonable visitation rights [shall be awarded] to a parent unless it is shown that the visitation would be detrimental to the best interests of the child. In the discretion of the court, reasonable visitation rights may be granted to any other person having an interest in the welfare of the child.” (Emphasis added.) The presumption established by §621 denied Michael the benefit of the first sentence of § 4601 because, as a matter of law, he is not a “parent.” It does not, however, prevent him from proving that he is an “other person having an interest in the welfare of the child.” On its face, therefore, the statute 134 OCTOBER TERM, 1988 Stevens, J., concurring in judgment 491 U. S. plainly gave the trial judge the authority to grant Michael “reasonable visitation rights.” I recognize that my colleagues have interpreted §621 as creating an absolute bar that would prevent a California trial judge from regarding the natural father as either a “parent” within the meaning of the first sentence of § 4601 or as “any other person” within the meaning of the second sentence. See ante, at 116, 119; post, at 148-151 (Brennan, J., dissenting). That is not only an unnatural reading of the statute’s plain language, but it is also not consistent with the California courts’ reading of the statute. Thus, in Vincent B. v. JoanR., 126 Cal. App. 3d 619,179 Cal. Rptr. 9 (1981), appeal dism’d, 459 U. S. 807 (1982), the California Court of Appeal, after deciding that the §621 presumption barred a natural father from proving paternity, went on to consider the separate question whether it would be proper to allow visitation pursuant to the second sentence of §4601: “Finally, appellant contends that even if Frank is conclusively presumed to be Z.’s father, appellant should be allowed visitation rights, since Civil Code section 4601 gives discretion to grant visitation rights to ‘any other person having an interest in the welfare of the child.’ We think it obvious that in the circumstances of this case such court-ordered visitation would be detrimental to the best interests of the child. Appellant’s interest in visiting the child is based on his claim that appellant is Z.’s father. Such claim is now determined to be legally impossible. The mother does not wish the child to be visited by appellant. Confusion, uncertainty, and embarrassment to the child would likely result from a court order that appellant, who claims to be Z.’s biological father, is entitled to visitation against the wishes of the mother. (Petitioner F. v. Respondent R., supra, 430 A. 2d 1075, 1080.)” 126 Cal. App. 3d, at 627-628, 179 Cal. Rptr., at 13 (emphasis added). MICHAEL H. v. GERALD D. 135 110 Stevens, J., concurring in judgment Supporting the court’s decision that granting visitation rights to Vincent would be contrary to the child’s best interests was the fact that “unlike the putative fathers in Stanley [v. Illinois, 405 U. S. 645 (1972),] and [In re] Lisa R. [, 13 Cal. 3d 636, 532 P. 2d 123 (1975)], appellant has never lived with the mother and child, nor has he ever supported the child.” 126 Cal. App. 3d, at 626, 179 Cal. Rptr., at 12. Similarly, in this case, the trial judge not only found the conclusive presumption applicable, but also separately considered the effect of § 4601 and expressly found “that, at the present time, it is not in the best interests of the child that the Plaintiff have visitation. The Court believes that the existence of two (2) ‘fathers’ as male authority figures will confuse the child and be counter-productive to her best interests.” Supp. App. to Juris. Statement A-90—A-91. In its opinion, the Court of Appeal also concluded that Michael “is not entitled to rights of visitation under section 4601,” see 191 Cal. App. 3d 995, 1013, 236 Cal. Rptr. 810, 821 (1987), and then quoted the above excerpt from the opinion in Vincent B. v. Joan R. As I read that opinion, it does not support the view that a natural father cannot be an “other person” within the meaning of §4601; rather, it indicates that the outcome depends largely on “the circumstances of th[e] case.”* Under the circumstances of the case before us, Michael was given a fair opportunity to show that he is Victoria’s natural father, that he had developed a relationship with her, and that her interests would be served by granting him visitation rights. On the other hand, the record also shows that after its rather shaky start, the marriage between Carole and Gerald developed a stability that now provides Victoria with *For cases showing the California courts’ willingness to decide §621 cases on a case-by-case basis, see, e. g., Michelle W. v. Ronald W., 39 Cal. 3d 354, 703 P. 2d 88 (1985), app. dism’d, 474 U. S. 1043 (1986); In re Lisa R; 13 Cal. 3d 636, 532 P. 2d 123, cert, denied sub nom. Porzuczek v. Towner, 421 U. S. 1014 (1975). 136 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. a loving and harmonious family home. In the circumstances of this case, I find nothing fundamentally unfair about the exercise of a judge’s discretion that, in the end, allows the mother to decide whether her child’s best interests would be served by allowing the natural father visitation privileges. Because I am convinced that the trial judge had the authority under state law both to hear Michael’s plea for visitation rights and to grant him such rights if Victoria’s best interests so warranted, I am satisfied that the California statutory scheme is consistent with the Due Process Clause of the Fourteenth Amendment. I therefore concur in the Court’s judgment of affirmance. Justice Brennan, with whom Justice Marshall and Justice Blackmun join, dissenting. In a case that has yielded so many opinions as has this one, it is fruitful to begin by emphasizing the common ground shared by a majority of this Court. Five Members of the Court refuse to foreclose “the possibility that a natural father might ever have a constitutionally protected interest in his relationship with a child whose mother was married to, and cohabiting with, another man at the time of the child’s conception and birth.” Ante, at 133 (Stevens, J., concurring in judgment); see infra, at 141-147; post, at 157 (White, J., dissenting). Five Justices agree that the flaw inhering in a conclusive presumption that terminates a constitutionally protected interest without any hearing whatsoever is a procedural one. See infra, at 153; post, at 163 (White, J., dissenting); ante, at 132 (Stevens, J., concurring in judgment). Four Members of the Court agree that Michael H. has a liberty interest in his relationship with Victoria, see infra, at 143; post, at 157 (White, J., dissenting), and one assumes for purposes of this case that he does, see ante, at 133 (Stevens, J., concurring in judgment). In contrast, only one other Member of the Court fully endorses Justice Scalia’s view of the proper method of analyzing questions arising under the Due Process Clause. MICHAEL H. v. GERALD D. 137 110 Brennan, J., dissenting See ante, at 113; ante, at 132 (O’Connor, J., concurring in part). Nevertheless, because the plurality opinion’s exclusively historical analysis portends a significant and unfortunate departure from our prior cases and from sound constitutional decisionmaking, I devote a substantial portion of my discussion to it. I Once we recognized that the “liberty” protected by the Due Process Clause of the Fourteenth Amendment encompasses more than freedom from bodily restraint, today’s plurality opinion emphasizes, the concept was cut loose from one natural limitation on its meaning. This innovation paved the way, so the plurality hints, for judges to substitute their own preferences for those of elected officials. Dissatisfied with this supposedly unbridled and uncertain state of affairs, the plurality casts about for another limitation on the concept of liberty. It finds this limitation in “tradition.” Apparently oblivious to the fact that this concept can be as malleable and as elusive as “liberty” itself, the plurality pretends that tradition places a discernible border around the Constitution. The pretense is seductive; it would be comforting to believe that a search for “tradition” involves nothing more idiosyncratic or complicated than poring through dusty volumes on American history. Yet, as Justice White observed in his dissent in Moore v. East Cleveland, 431 U. S. 494, 549 (1977): “What the deeply rooted traditions of the country are is arguable.” Indeed, wherever I would begin to look for an interest “deeply rooted in the country’s traditions,” one thing is certain: I would not stop (as does the plurality) at Bracton, or Blackstone, or Kent, or even the American Law Reports in conducting my search. Because reasonable people can disagree about the content of particular traditions, and because they can disagree even about which traditions are relevant to the definition of “liberty,” the plurality has not found the objective boundary that it seeks. 138 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. Even if we could agree, moreover, on the content and significance of particular traditions, we still would be forced to identify the point at which a tradition becomes firm enough to be relevant to our definition of liberty and the moment at which it becomes too obsolete to be relevant any longer. The plurality supplies no objective means by which we might make these determinations. Indeed, as soon as the plurality sees signs that the tradition upon which it bases its decision (the laws denying putative fathers like Michael standing to assert paternity) is crumbling, it shifts ground and says that the case has nothing to do with that tradition, after all. “[W]hat is at issue here,” the plurality asserts after canvassing the law on paternity suits, “is not entitlement to a state pronouncement that Victoria was begotten by Michael.” Ante, at 126. But that is precisely what is at issue here, and the plurality’s last-minute denial of this fact dramatically illustrates the subjectivity of its own analysis. It is ironic that an approach so utterly dependent on tradition is so indifferent to our precedents. Citing barely a handful of this Court’s numerous decisions defining the scope of the liberty protected by the Due Process Clause to support its reliance on tradition, the plurality acts as though English legal treatises and the American Law Reports always have provided the sole source for our constitutional principles. They have not. Just as common-law notions no longer define the “property” that the Constitution protects, see Goldberg v. Kelly, 397 U. S. 254 (1970), neither do they circumscribe the “liberty” that it guarantees. On the contrary, “‘[l]ib-erty’ and ‘property’ are broad and majestic terms. They are among the ‘[g]reat [constitutional] concepts . . . purposely left to gather meaning from experience. . . . [T]hey relate to the whole domain of social and economic fact, and the statesmen who founded this Nation knew too well that only a stagnant society remains unchanged.’” Board of Regents of State Colleges n. Roth, 408 U. S. 564, 571 (1972), quoting Na MICHAEL H. v. GERALD D. 139 110 Brennan, J., dissenting tional Ins. Co. v. Tidewater Co., 337 U. S. 582, 646 (1949) (Frankfurter, J., dissenting). It is not that tradition has been irrelevant to our prior decisions. Throughout our decisionmaking in this important area runs the theme that certain interests and practices — freedom from physical restraint, marriage, childbearing, childrearing, and others—form the core of our definition of “liberty.” Our solicitude for these interests is partly the result of the fact that the Due Process Clause would seem an empty promise if it did not protect them, and partly the result of the historical and traditional importance of these interests in our society. In deciding cases arising under the Due Process Clause, therefore, we have considered whether the concrete limitation under consideration impermissibly impinges upon one of these more generalized interests. Today’s plurality, however, does not ask whether parenthood is an interest that historically has received our attention and protection; the answer to that question is too clear for dispute. Instead, the plurality asks whether the specific variety of parenthood under consideration—a natural father’s relationship with a child whose mother is married to another man—has enjoyed such protection. If we had looked to tradition with such specificity in past cases, many a decision would have reached a different result. Surely the use of contraceptives by unmarried couples, Eisenstadt v. Baird, 405 U. S. 438 (1972), or even by married couples, Griswold v. Connecticut, 381 U. S. 479 (1965); the freedom from corporal punishment in schools, Ingraham v. Wright, 430 U. S. 651 (1977); the freedom from an arbitrary transfer from a prison to a psychiatric institution, Vitek v. Jones, 445 U. S. 480 (1980); and even the right to raise one’s natural but illegitimate children, Stanley v. Illinois, 405 U. S. 645 (1972), were not “interest[s] traditionally protected by our society,” ante, at 122, at the time of their consideration by this Court. If we had asked, therefore, in Eisenstadt, Griswold, Ingraham, Vitek, or Stanley itself whether 140 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. the specific interest under consideration had been traditionally protected, the answer would have been a resounding “no.” That we did not ask this question in those cases highlights the novelty of the interpretive method that the plurality opinion employs today. The plurality’s interpretive method is more than novel; it is misguided. It ignores the good reasons for limiting the role of “tradition” in interpreting the Constitution’s deliberately capacious language. In the plurality’s constitutional universe, we may not take notice of the fact that the original reasons for the conclusive presumption of paternity are out of place in a world in which blood tests can prove virtually beyond a shadow of a doubt who sired a particular child and in which the fact of illegitimacy no longer plays the burdensome and stigmatizing role it once did. Nor, in the plurality’s world, may we deny “tradition” its full scope by pointing out that the rationale for the conventional rule has changed over the years, as has the rationale for Cal. Evid. Code Ann. §621 (West Supp. 1989);1 instead, our task is simply to identify a rule denying the asserted interest and not to ask whether the basis for that rule—which is the true reflection of the values undergirding it—has changed too often or too recently to call the rule embodying that rationale a “tradition.” Moreover, by describing the decisive question as whether Michael’s and Victoria’s interest is one that has been “traditionally protected by our society,” ante, at 122 (emphasis added), rather than one that society traditionally has thought important (with or without protecting it), and by suggesting that our sole function is to “discern the society’s views,” ante, at 128, n. 6 (emphasis added), the plurality acts as if the only pur 1 See In re Marriage of Sharyne and Stephen B., 124 Cal. App. 3d 524, 528-531, 177 Cal. Rptr. 429, 431-433 (1981) (noting that California courts initially justified conclusive presumption of paternity on the ground that biological paternity was impossible to prove, but that the preservation of family integrity became the rule’s paramount justification when paternity tests became reliable). MICHAEL H. v. GERALD D. 141 110 Brennan, J., dissenting pose of the Due Process Clause is to confirm the importance of interests already protected by a majority of the States. Transforming the protection afforded by the Due Process Clause into a redundancy mocks those who, with care and purpose, wrote the Fourteenth Amendment. In construing the Fourteenth Amendment to offer shelter only to those interests specifically protected by historical practice, moreover, the plurality ignores the kind of society in which our Constitution exists. We are not an assimilative, homogeneous society, but a facilitative, pluralistic one, in which we must be willing to abide someone else’s unfamiliar or even repellent practice because the same tolerant impulse protects our own idiosyncracies. Even if we can agree, therefore, that “family” and “parenthood” are part of the good life, it is absurd to assume that we can agree on the content of those terms and destructive to pretend that we do. In a community such as ours, “liberty” must include the freedom not to conform. The plurality today squashes this freedom by requiring specific approval from history before protecting anything in the name of liberty. The document that the plurality construes today is unfamiliar to me. It is not the living charter that I have taken to be our Constitution; it is instead a stagnant, archaic, hidebound document steeped in the prejudices and superstitions of a time long past. This Constitution does not recognize that times change, does not see that sometimes a practice or rule outlives its foundations. I cannot accept an interpretive method that does such violence to the charter that I am bound by oath to uphold. II The plurality’s reworking of our interpretive approach is all the more troubling because it is unnecessary. This is not a case in which we face a “new” kind of interest, one that requires us to consider for the first time whether the Constitution protects it. On the contrary, we confront an interest — that of a parent and child in their relationship with each 142 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. other—that was among the first that this Court acknowledged in its cases defining the “liberty” protected by the Constitution, see, e. g., Meyer v. Nebraska, 262 U. S. 390, 399 (1923); Skinner n. Oklahoma, 316 U. S. 535, 541 (1942); Prince n. Massachusetts, 321 U. S. 158, 166 (1944), and I think I am safe in saying that no one doubts the wisdom or validity of those decisions. Where the interest under consideration is a parent-child relationship, we need not ask, over and over again, whether that interest is one that society traditionally protects. Thus, to describe the issue in this case as whether the relationship existing between Michael and Victoria “has been treated as a protected family unit under the historic practices of our society, or whether on any other basis it has been accorded special protection,” ante, at 124, is to reinvent the wheel. The better approach—indeed, the one commanded by our prior cases and by common sense—is to ask whether the specific parent-child relationship under consideration is close enough to the interests that we already have protected to be deemed an aspect of “liberty” as well. On the facts before us, therefore, the question is not what “level of generality” should be used to describe the relationship between Michael and Victoria, see ante, at 127, n. 6, but whether the relationship under consideration is sufficiently substantial to qualify as a liberty interest under our prior cases. On four prior occasions, we have considered whether unwed fathers have a constitutionally protected interest in their relationships with their children. See Stanley v. Illinois, 405 U. S. 645 (1972); Quilloin n. Walcott, 434 U. S. 246 (1978); Caban n. Mohammed, 441 U. S. 380 (1979); and Lehr n. Robertson, 463 U. S. 248 (1983). Though different in factual and legal circumstances, these cases have produced a unifying theme: although an unwed father’s biological link to his child does not, in and of itself, guarantee him a constitutional stake in his relationship with that child, such a link combined with a substantial parent-child relationship will do MICHAEL H. v. GERALD D. 143 110 Brennan, J., dissenting so.2 “When an unwed father demonstrates a full commitment to the responsibilities of parenthood by ‘com[ing] forward to participate in the rearing of his child,’ ... his interest in personal contact with his child acquires substantial protection under the Due Process Clause. At that point it may be said that he ‘act[s] as a father toward his children.’” Lehr v. Robertson, supra, at 261, quoting Caban v. Mohammed, supra, at 392, 389, n. 7. This commitment is why Mr. Stanley and Mr. Caban won; why Mr. Quilloin and Mr. Lehr lost; and why Michael H. should prevail today. Michael H. is almost certainly Victoria D.’s natural father, has lived with her as her father, has contributed to her support, and has from the beginning sought to strengthen and maintain his relationship with her. Claiming that the intent of these cases was to protect the “unitary family,” ante, at 123, the plurality waves Stanley, Quilloin, Caban, and Lehr aside. In evaluating the plurality’s dismissal of these precedents, it is essential to identify its conception of the “unitary family.” If, by acknowledging that Stanley et al. sought to protect “the relationships that develop within the unitary family,” ibid., the plurality meant only to describe the kinds of relationships that develop when parents and children live together (formally or informally) as a family, then the plurality’s vision of these cases would be correct. But that is not the plurality’s message. Though it pays lipservice to the idea that marriage is not the crucial fact in denying constitutional protection to the relationship between Michael and Victoria, ante, at 123, n. 3, the plurality cannot mean what it says. The evidence is undisputed that Michael, Victoria, and Carole did live together as a family; that is, they shared the 2 The plurality’s claim that “[t]he logic of [my] position leads to the conclusion that if Michael had begotten Victoria by rape, that fact would in no way affect his possession of a liberty interest in his relationship with her,” ante, at 124, n. 4, ignores my observation that a mere biological connection is insufficient to establish a liberty interest on the part of an unwed father. 144 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. same household, Victoria called Michael “Daddy,” Michael contributed to Victoria’s support, and he is eager to continue his relationship with her. Yet they are not, in the plurality’s view, a “unitary family,” whereas Gerald, Carole, and Victoria do compose such a family. The only difference between these two sets of relationships, however, is the fact of marriage. The plurality, indeed, expressly recognizes that marriage is the critical fact in denying Michael a constitutionally protected stake in his relationship with Victoria: no fewer than six times, the plurality refers to Michael as the “adulterous natural father” (emphasis added) or the like. Ante, at 120; 127, n. 6; 128, n. 6; 129, n. 7; 130. See also ante, at 124 (referring to the “marital family” of Gerald, Carole, and Victoria) (emphasis added); ante, at 129 (plurality’s holding limited to those situations in which there is “an extant marital family”).3 However, the very premise of Stanley and the cases following it is that marriage is not decisive in answering the question whether the Constitution protects the parental relationship under consideration. These cases are, after all, important precisely because they involve the rights of unwed fathers. It is important to remember, moreover, that in Quilloin, Caban, and Lehr, the putative father’s demands would have disrupted a “unitary family” as the plurality defines it; in each case, the husband of the child’s mother sought to adopt the child over the objections of the natural father. Significantly, our decisions in those cases in no way relied on the need to protect the marital family. Hence the plurality’s claim that Stanley, Quilloin, Caban, and Lehr 3 In one place, the plurality opinion appears to suggest that the length of time that Michael and Victoria lived together is relevant to the question whether they have a liberty interest in their relationship with each other. See ante, at 123, n. 3. The point is not pursued, however, and in any event I am unable to find in the traditions on which the plurality otherwise exclusively relies any emphasis on the duration of the relationship between the putative father and child. MICHAEL H. v. GERALD D. 145 110 Brennan, J., dissenting were about the “unitary family,” as that family is defined by today’s plurality, is surprising indeed. The plurality’s exclusive rather than inclusive definition of the “unitary family” is out of step with other decisions as well. This pinched conception of “the family,” crucial as it is in rejecting Michael’s and Victoria’s claims of a liberty interest, is jarring in light of our many cases preventing the States from denying important interests or statuses to those whose situations do not fit the government’s narrow view of the family. From Loving v. Virginia, 388 U. S. 1 (1967), to Levy n. Louisiana, 391 U. S. 68 (1968), and Glona v. American Guarantee & Liability Ins. Co., 391 U. S. 73 (1968), and from Gomez v. Perez, 409 U. S. 535 (1973), to Moore v. East Cleveland, 431 U. S. 494 (1977), we have declined to respect a State’s notion, as manifested in its allocation of privileges and burdens, of what the family should be. Today’s rhapsody on the “unitary family” is out of tune with such decisions. The plurality’s focus on the “unitary family” is misdirected for another reason. It conflates the question whether a liberty interest exists with the question what procedures may be used to terminate or curtail it. It is no coincidence that we never before have looked at the relationship that the unwed father seeks to disrupt, rather than the one he seeks to preserve, in determining whether he has a liberty interest in his relationship with his child. To do otherwise is to allow the State’s interest in terminating the relationship to play a role in defining the “liberty” that is protected by the Constitution. According to our established framework under the Due Process Clause, however, we first ask whether the person claiming constitutional protection has an interest that the Constitution recognizes; if we find that he or she does, we next consider the State’s interest in limiting the extent of the procedures that will attend the deprivation of that interest. See, e. g., Logan v. Zimmerman Brush Co., 455 U. S. 422, 428 (1982). By stressing the need to preserve the “unitary 146 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. family” and by focusing not just on the relationship between Michael and Victoria but on their “situation” as well, ante, at 124, today’s plurality opinion takes both of these steps at once. The plurality’s premature consideration of California’s interests is evident from its careful limitation of its holding to those cases in which “the mother is, at the time of the child’s conception and birth, married to, and cohabitating with, another man, both of whom wish to raise the child as the offspring of their union.” Ante, at 129 (emphasis added). See also ante, at 127 (describing Michael’s liberty interest as the “substantive parental rights [of] the natural father of a child conceived within, and born into, an extant marital union that wishes to embrace the child”). The highlighted language suggests that if Carole or Gerald alone wished to raise Victoria, or if both were dead and the State wished to raise her, Michael and Victoria might be found to have a liberty interest in their relationship with each other.4 But that would be to say that whether Michael and Victoria have a liberty interest varies with the State’s interest in recognizing that interest, for it is the State’s interest in protecting the marital family— and not Michael and Victoria’s interest in their relationship with each other—that varies with the status of Carole and Gerald’s relationship. It is a bad day for due process when 4 Note that the plurality presumably would disapprove the California courts’ holdings in Vincent B. v. Joan R., 126 Cal. App. 3d 619, 179 Cal. Rptr. 9 (1981) (§ 621 defeated putative father’s interest even where husband and wife divorced at the time of the paternity action), and Michelle W. n. Ronald W., 39 Cal. 3d 354, 703 P. 2d 88 (1985) (§621 defeated putative father’s interest even where mother had married putative father and divorced man to whom she had been married at time of conception and birth). To suggest, moreover, that “it is at least possible that our traditions lead to a different conclusion” in cases such as Vincent B. and Michelle W., ante, at 129, n. 7, is to express an optimism about our ability to identify “traditions” with microscopic precision that I do not share, and a willingness to slice society up into minuscule pieces, based only on tradition, that I cannot endorse. MICHAEL H. v. GERALD D. 147 110 Brennan, J., dissenting the State’s interest in terminating a parent-child relationship is reason to conclude that that relationship is not part of the “liberty” protected by the Fourteenth Amendment. The plurality has wedged itself between a rock and a hard place. If it limits its holding to those situations in which a wife and husband wish to raise the child together, then it necessarily takes the State’s interest into account in defining “liberty”; yet if it extends that approach to circumstances in which the marital union already has been dissolved, then it may no longer rely on the State’s asserted interest in protecting the “unitary family” in denying that Michael and Victoria have been deprived of liberty. The plurality’s confusion about the proper analysis of claims involving procedural due process also becomes obvious when one examines the plurality’s shift in emphasis from the putative father’s standing to his ability to obtain parental prerogatives. See ante, at 126. In announcing that what matters is not the father’s ability to claim paternity, but his ability to obtain “substantive parental rights,” ante, at 127, the plurality turns procedural due process upside down. Michael’s challenge in this Court does not depend on his ability ultimately to obtain visitation rights; it would be strange indeed if, before one could be granted a hearing, one were required to prove that one would prevail on the merits. The point of procedural due process is to give the litigant a fair chance at prevailing, not to ensure a particular substantive outcome. Nor does Michael’s challenge depend on the success of fathers like him in obtaining parental rights in past cases; procedural due process is, by and large, an individual guarantee, not one that should depend on the success or failure of prior cases having little or nothing to do with the claimant’s own suit.5 5 One need only look as far as Quilloin v. Walcott, 434 U. S. 246, 255 (1978), to understand why an unwed father might lose for reasons having nothing to do with his own relationship with the child: there, we approved the use of a “best interest” standard, rather than an “unfitness” standard, 148 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. Ill Because the plurality decides that Michael and Victoria have no liberty interest in their relationship with each other, it need consider neither the effect of § 621 on their relationship nor the State’s interest in bringing about that effect. It is obvious, however, that the effect of §621 is to terminate the relationship between Michael and Victoria before affording any hearing whatsoever on the issue whether Michael is Victoria’s father. This refusal to hold a hearing is properly analyzed under our procedural due process cases, which instruct us to consider the State’s interest in curtailing the procedures accompanying the termination of a constitutionally protected interest. California’s interest, minute in comparison with a father’s interest in his relationship with his child, cannot justify its refusal to hear Michael out on his claim that he is Victoria’s father. A We must first understand the nature of the challenged statute: it is a law that stubbornly insists that Gerald is Victoria’s father, in the face of evidence showing a 98 percent probability that her father is Michael.6 What Michael wants is a chance to show that he is Victoria’s father. By depriving him of this opportunity, California prevents Michael from taking advantage of the best-interest standard embodied in §4601 of California’s Civil Code, which directs that parents be given visitation rights unless “the visitation would be detrimental to the best interests of the child.” Cal. Civ. Code Ann. §4601 (West Supp. 1989).7 for an unwed father who objected to the adoption of his child by another man. 6 Justice Stevens’ claim that “Michael was given a fair opportunity to show that he is Victoria’s natural father,” ante, at 135, ignores the fact that this case is before us precisely because California law refuses to allow men like Michael such an opportunity. 7 Showing a startling misunderstanding of the stakes in this case, the plurality characterizes the issue at the hearing that Michael seeks as MICHAEL H. v. GERALD D. 149 110 Brennan, J., dissenting As interpreted by the California courts, however, §621 not only deprives Michael of the benefits of the best-interest standard; it also deprives him of any chance of maintaining his relationship with the child he claims to be his own. When, as a result of § 621, a putative father may not establish his paternity, neither may he obtain discretionary visitation rights as a “nonparent” under § 4601. See Vincent B. v. Joan R., 126 Cal. App. 3d 619, 627-628, 179 Cal. Rptr. 9, 13 (1981), appeal dism’d, 459 U. S. 807 (1982); see also ante, at 116. Justice Stevens’ assertion to the contrary, ante, at 134-135, is mere wishful thinking. In concluding that the California courts afford putative fathers like Michael a meaningful opportunity to show that visitation rights would be in the best interests of their children, he fastens upon the words “in the circumstances of this case” in Vincent B. n. Joan R., supra, at 627, 179 Cal. Rptr., at 13. Ante, at 134-135. His suggestion is that the court in that case conducted an individualized assessment of the effect on the child of granting visitation rights to Vincent B. “whether, in the particular circumstances of his case, California’s policies would best be served by giving him parental rights.” Ante, at 120. The hearing that the plurality describes is merely one that the California courts hold in response to constitutional challenges such as those lodged here, see, e. g., Michelle W. v. Ronald W., 39 Cal. 3d, at 363, 703 P. 2d, at 93; it is not the hearing that Michael seeks as the end result of this lawsuit. The plurality’s confusion is further evident in its announcement that “what is at issue here is not entitlement to a state pronouncement that Victoria was begotten by Michael.” Ante, at 126 (emphasis added). That is precisely what is at issue in the hearing that Michael seeks. Justice Stevens exhibits the same misunderstanding in pointing to Michelle W. and In re Lisa R., 13 Cal. 3d 636, 532 P. 2d 123 (1975), as evidence of “the California courts’ willingness to decide § 621 cases on a case-by-case basis.” Ante, at 135, n. This “case-by-case” analysis is not the result of a flexible interpretation of § 621, but is the courts’ response to the many constitutional challenges brought against § 621. Similarly, Michael was given an opportunity to show that “he had developed a relationship with [Victoria],” ante, at 135, only because he launched this constitutional attack on § 621. 150 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. The California appellate court’s decision will not support Justice Stevens’ reading, as the court’s reasoning applies to all putative fathers whom § 621 has denied the opportunity to show paternity. The court in Vincent B. began by stressing the fact that the child’s mother objected to visits from Vincent. This circumstance is present in every single case falling under the conclusive presumption of § 621. Granting visitation rights to a person who claimed to be the child’s father, the court went on, also would cause “confusion, uncertainty, and embarrassment.” 126 Cal. App. 3d, at 628, 179 Cal. Rptr., at 13. Again, the notion that unacceptable confusion would result from awarding visitation to a person who claims to be the child’s father is equally applicable to any case in which the “nonparent” under §4601 has lost under §621. Finally, the court in Vincent B. approvingly cited Petitioner F. v. Respondent R., 430 A. 2d 1075, 1080 (1981), in which the Supreme Court of Delaware rejected a putative father’s argument that Delaware’s conclusive presumption of paternity violated the Equal Protection Clause of the Federal Constitution. 126 Cal. App. 3d, at 627, 179 Cal. Rptr., at 13. Emphasizing the “permanent stigma and distress” that would result from granting parental rights to a putative father whose child was born to the wife of another man, the Delaware court decided that, given the State’s interest in “guard-ting] against assaults upon the family unit[,] . . . [t]he application of the presumption of legitimacy of a child born to a married woman would be in the child’s interest in practically all cases.” 430 A. 2d, at 1080 (emphasis added). Vincent B.’s reliance on Petitioner F. sends a clear signal that the California court was issuing a ruling applicable to any case that fit into §621’s conclusive presumption, and that the “rough justice” that prevailed under § 621 also would suffice under §4601. This kind of determination is a far cry from the individualized assessment that Justice Stevens would seem to demand. Ante, at 135.8 8 Justice Stevens incorrectly suggests that the court in Vincent B. based its denial of visitation rights under § 4601 partly on the lack of an MICHAEL H. v. GERALD D. 151 110 Brennan, J., dissenting Likewise, in the case before us, the court’s finding that “the existence of two (2) ‘fathers’ as male authority figures will confuse the child and be counter-productive to her best interests,” Supp. App. to Juris. Statement A-90—A-91, is not an evaluation of the relationship between Michael and Victoria, but a restatement of the policies underlying § 621 itself. It may well be that the California courts’ interpretation of § 4601 as precluding visitation rights for a putative father is “an unnatural reading” of that provision, ante, at 134, but it is not for us to decide what California’s statute means. Section 621 as construed by the California courts thus cuts off the relationship between Michael and Victoria—a liberty interest protected by the Due Process Clause—without affording the least bit of process. This case, in other words, involves a conclusive presumption that is used to terminate a constitutionally protected interest—the kind of rule that our preoccupation with procedural fairness has caused us to condemn. See, e. g., Vlandis v. Kline, 412 U. S. 441 (1973); Cleveland Board of Education v. LaFleur, 414 U. S. 632 (1974); Weinberger v. Salfi, 422 U. S. 749, 770-772 (1975). Gerald D. and the plurality turn a blind eye to the true nature of § 621 by protesting that, instead of being a conclusive presumption, it is a “substantive rule of law.” Ante, at 119. This facile observation cannot save § 621. It may be that all conclusive presumptions are, in a sense, substantive rules of law; but §621 then belongs in that special category of substantive rules that presumes a fact relevant to a certain class of litigation, and it is that feature that renders § 621 suspect under our prior cases. To put the point differently, a conclusive presumption takes the form of “no X’s are Y’s,” and is typically accompanied by a rule such as, “. . . and only Y’s may obtain a driver’s license.” (There would be no need for the presumption unless something hinged on the fact pre- established relationship between Vincent B. and the child. Ante, at 135. In fact, the court did not even mention the specific relationship between these two people in coming to its decision under § 4601. See 126 Cal. App. 3d, at 628, 179 Cal. Rptr., at 13. 152 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. sumed.) Ignoring the fact that §621 takes the form of “no X’s are Y’s,” Gerald D. and the plurality fix upon the rule following §621—only Y’s may assert parental rights—and call § 621 a substantive rule of law. This strategy ignores both the form and the effect of §621. In a further effort to show that §621 is not a conclusive presumption, Gerald D. claims—and the plurality agrees, see ante, at 119—that whether a man is the biological father of a child whose family situation places the putative father within § 621 is simply irrelevant to the State. Brief for Appellee 14. This is, I surmise, an attempt to avoid the implications of our cases condemning the presumption of a fact that a State has made relevant or decisive to a particular decision. See, e. g., Bell v. Burson, 402 U. S. 535 (1971). Yet the claim that California does not care about factual paternity is patently false. California cares very much about factual paternity when the husband is impotent or sterile, see Cal. Evid. Code Ann. § 621(a) (West Supp. 1989); it cares very much about it when the wife and husband do not share the same home, see Vincent B. v. Joan R., 126 Cal. App. 3d, at 623-624, 179 Cal. Rptr., at 11; and it cares very much about it when the husband himself declares that he is not the father, see Cal. Evid. Code Ann. § 621(c) (West Supp. 1989). Indeed, under California law as currently structured, paternity is decisive in choosing the standard that will be used in granting or denying custody or visitation. The State, though selective in its concern for factual paternity, certainly is not indifferent to it.9 More fundamentally, California’s purported indifference to factual paternity does not show that § 621 is not a conclu- 9 In this respect, the plurality is mistaken in suggesting that “there is no difference between a rule which says that the marital husband shall be irrebuttably presumed to be the father, and a rule which says that the adulterous natural father shall not be recognized as the legal father.” Ante, at 120. In the latter case, the State has not made paternity the predominant concern in child-custody disputes and then told some putative fathers that they may not prove their paternity. MICHAEL H. v. GERALD D. 153 110 Brennan, J., dissenting sive presumption. To say that California does not care about factual paternity in the limited circumstances of this case—where the husband is neither impotent nor sterile nor living apart from his wife—is simply another way of describing its conclusive presumption. Not content to rest on its assertion that § 621 does not, in fact, establish a conclusive presumption, the plurality goes on to argue that a challenge to a conclusive presumption must rest on substantive rather than procedural due process. See ante, at 120-121. This is simply not so. In Weinberger v. Salfi, supra, the Court identified two lines of cases involving challenges to social-welfare legislation: those in which a legislative classification was challenged as arbitrary and those in which a conclusive presumption was attacked. The Court fit the complaint in Salfi into the former category on the ground that the challenged law did not deprive anyone of a constitutionally protected interest. 422 U. S., at 772. Today’s plurality, in contrast, classifies this case as one invoking substantive due process before it considers the nature of the interest at stake. Its support for this innovation includes several law-review commentaries, two concurrences in the judgment, a dissent, and Salfi itself. Ante, at 120-121. Even more disturbing than the plurality’s reliance on these infirm foundations is its failure to recognize that the defect from which conclusive presumptions suffer is a procedural one: the State has declared a certain fact relevant, indeed controlling, yet has denied a particular class of litigants a hearing to establish that fact. This is precisely the kind of flaw that procedural due process is designed to correct.10 10 We recognized as much in Caban v. Mohammed, 441 U. S. 380, 385, n. 3 (1979), in which we explicitly described Stanley v. Illinois, 405 U. S. 645 (1972), as a case involving procedural due process. The plurality’s bald statement that the holding in Stanley did not rely on procedural due process is therefore incorrect. See ante, at 120. 154 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. B The question before us, therefore, is whether California has an interest so powerful that it justifies granting Michael no hearing before terminating his parental rights. “Many controversies have raged about the cryptic and abstract words of the Due Process Clause but there can be no doubt that at a minimum they require that deprivation of life, liberty or property by adjudication be preceded by notice and opportunity for hearing appropriate to the nature of the case.” Mullane v. Central Hanover Bank & Trust Co., 339 U. S. 306, 313 (1950). When a State seeks to limit the procedures that will attend the deprivation of a constitutionally protected interest, it is only the State’s interest in streamlining procedures that is relevant. See, e. g., Mathews v. Eldridge, 424 U. S. 319, 335 (1976). A State may not, in other words, justify abbreviated procedures on the ground that it wishes to pay welfare benefits to fewer people or wants to reduce the number of tenured professors on its payroll. It would be strange indeed if a State could curtail procedures with the explanation that it was hostile to the underlying, constitutionally protected interest. The purported state interests here, however, stem primarily from the State’s antagonism to Michael’s and Victoria’s constitutionally protected interest in their relationship with each other and not from any desire to streamline procedures. Gerald D. explains that §621 promotes marriage, maintains the relationship between the child and presumed father, and protects the integrity and privacy of the matrimonial family. Brief for Appellee 24. It is not, however, §621, but the best-interest principle, that protects a stable marital relationship and maintains the relationship between the child and presumed father. These interests are implicated by the determination of who gets parental rights, not by the determination of who is the father; in the hearing that Michael seeks, parental rights are not the issue. Of the objectives that Gerald stresses, therefore, only the preservation of fam MICHAEL H. v. GERALD D. 155 110 Brennan, J., dissenting ily privacy is promoted by the refusal to hold a hearing itself. Yet § 621 furthers even this objective only partially. Gerald D. gives generous proportions to the privacy protected by § 621, asserting that this provision protects a couple like Gerald and Carole from answering questions on such matters as “their sexual habits and practices with each other and outside their marriage, their finances, and their thoughts, beliefs, and opinions concerning their relationship with each other and with Victoria.” Id., at 25. Yet invalidation of § 621 would not, as Gerald suggests, subject Gerald and Carole to public scrutiny of all of these private matters. Family finances and family dynamics are relevant, not to paternity, but to the best interests of the child—and the child’s best interests are not, as I have stressed, in issue at the hearing that Michael seeks. The only private matter touching on the paternity presumed by § 621 is the married couple’s sex life. Even there, § 621 as interpreted by California’s intermediate appellate courts pre-empts inquiry into a couple’s sexual relations, since “cohabitation” consists simply of living under the same roof together; the wife and husband need not even share the same bed. See, e. g., Vincent B. v. Joan R., 126 Cal. App. 3d 619, 179 Cal. Rptr. 9 (1981). Admittedly, §621 does not foreclose inquiry into the husband’s fertility or virility—matters that are ordinarily thought of as the couple’s private business. In this day and age, however, proving paternity by asking intimate and detailed questions about a couple’s relationship would be decidedly anachronistic. Who on earth would choose this method of establishing fatherhood when blood tests prove it with far more certainty and far less fuss? The State’s purported interest in protecting matrimonial privacy thus does not measure up to Michael’s and Victoria’s interest in maintaining their relationship with each other.11 11 Thus, in concluding that § 621 “exclud[es] inquiries into the child’s paternity that would be destructive of family integrity and privacy,” ante, at 120, the plurality exaggerates the extent to which these interests would be threatened by the elimination of § 621’s presumption. On the other hand, 156 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. Make no mistake: to say that the State must provide Michael with a hearing to prove his paternity is not to express any opinion of the ultimate state of affairs between Michael and Victoria and Carole and Gerald. In order to change the current situation among these people, Michael first must convince a court that he is Victoria’s father, and even if he is able to do this, he will be denied visitation rights if that would be in Victoria’s best interests. See Cal. Civ. Code Ann. §4601 (West Supp. 1989). It is elementary that a determination that a State must afford procedures before it terminates a given right is not a prediction about the end result of those procedures.12 IV The atmosphere surrounding today’s decision is one of make-believe. Beginning with the suggestion that the situa if the State’s foremost interest is in protecting the husband from discovering that he may not be the father of his wife’s children, as the plurality suggests, see ante, at 120, n. 1, then § 621 is unhelpful indeed. Since “cohabitation” under California law includes sharing the same roof but not the same bed and since a person need only make a phone call in order to unsettle a husband’s certainty in the paternity of his wife’s children, § 621 will do little to prevent such discoveries. See also post, at 162 (White, J., dissenting). 12 The plurality’s failure to see this point causes it to misstate Michael’s claim in the following way: “Michael contends as a matter of substantive due process that, because he has established a parental relationship with Victoria, protection of Gerald’s and Carole’s marital union is an insufficient state interest to support termination of that relationship.” Ante, at 121. Michael does not claim that the State may not, under any circumstance, terminate his relationship with Victoria; instead, he simply claims that the State may not do so without affording him a hearing on the issue—paternity—that it deems vital to the question whether their relationship may be discontinued. The plurality makes Michael’s claim easier to knock down by turning it into such a big target. The plurality’s misunderstanding of Michael’s claim also leads to its assertion that “to provide protection to an adulterous natural father is to deny protection to a marital father.” Ante, at 130. To allow Michael a chance to prove his paternity, however, in no way guarantees that Gerald’s relationship with Victoria will be changed. MICHAEL H. v. GERALD D. 157 110 White, J., dissenting tion confronting us here does not repeat itself every day in every corner of the country, ante, at 113, moving on to the claim that it is tradition alone that supplies the details of the liberty that the Constitution protects, and passing finally to the notion that the Court always has recognized a cramped vision of “the family,” today’s decision lets stand California’s pronouncement that Michael—whom blood tests show to a 98 percent probability to be Victoria’s father—is not Victoria’s father. When and if the Court awakes to reality, it will find a world very different from the one it expects. Justice White, with whom Justice Brennan joins, dissenting. California law, as the plurality describes it, ante, at 119, tells us that, except in limited circumstances, California declares it to be “irrelevant for paternity purposes whether a child conceived during, and born into, an existing marriage was begotten by someone other than the husband” (emphasis in original). This I do not accept, for the fact that Michael H. is the biological father of Victoria is to me highly relevant to whether he has rights, as a father or otherwise, with respect to the child. Because I believe that Michael H. has a liberty interest that cannot be denied without due process of the law, I must dissent. I Like Justices Brennan, Marshall, Blackmun, and Stevens, I do not agree with the plurality opinion’s conclusion that a natural father can never “have a constitutionally protected interest in his relationship with a child whose mother was married to, and cohabiting with, another man at the time of the child’s conception and birth.” Ante, at 133 (Stevens, J., concurring in judgment). Prior cases here have recognized the liberty interest of a father in his relationship with his child. In none of these cases did we indicate that the father’s rights were dependent on the marital status of the mother or biological father. The basic principle enun 158 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. ciated in the Court’s unwed father cases is that an unwed father who has demonstrated a sufficient commitment to his paternity by way of personal, financial, or custodial responsibilities has a protected liberty interest in a relationship with his child.1 We have not before faced the question of a biological father’s relationship with his child when the child was born while the mother was married to another man. On several occasions however, we have considered whether a biological father has a constitutionally cognizable interest in an opportunity to establish paternity. Stanley n. Illinois, 405 U. S. 645 (1972), recognized the biological father’s right to a legal relationship with his illegitimate child, holding that the Due Process Clause of the Fourteenth Amendment entitled the biological father to a hearing on his fitness before his illegitimate children could be removed from his custody. We rejected the State’s treatment of Stanley “not as a parent but as a stranger to his children.” Id., at 648. Quilloin n. Walcott, 434 U. S. 246, 255 (1978), also expressly recognized due process rights in the biological father, even while holding that those rights were not impermissibly burdened by the State’s application of a “best interests of the child” standard. Caban v. Mohammed, 441 U. S. 380 'Lehr v. Robertson, 463 U. S. 248, 259-260 (1983), emphasized the distinction between “a mere biological relationship and an actual relationship of parental responsibility.” In the dissent to Lehr, I said: “As Jessica’s biological father, Lehr either had an interest protected by the Constitution or he did not. If the entry of the adoption order in this case deprived Lehr of a constitutionally protected interest, he is entitled to notice and an opportunity to be heard before the order can be accorded finality.” Id., at 268 (footnote omitted). I rejected the majority’s approach which purported to analyze the particular facts of the case in order to determine whether Mr. Lehr had a constitutionally protected liberty interest. I stressed the interest that a natural parent has in his child, “one that has long been recognized and accorded constitutional protection.” Id., at 270. Whether or not the majority in Lehr was in error, on the facts of the instant case, even LeWs more demanding standard is clearly satisfied. MICHAEL H. v. GERALD D. 159 110 White, J., dissenting (1979), invalidated on equal protection grounds a statute under which a man’s children could be adopted by their natural mother and her husband without the natural father’s consent. In Lehr v. Robertson, 463 U. S. 248, 261-262 (1983), though holding against the father in that case, the Court said clearly that fathers who have participated in raising their illegitimate children and have developed a relationship with them have constitutionally protected parental rights. Indeed, the Court in Lehr suggested that States must provide a biological father of an illegitimate child the means by which he may establish his paternity so that he may have the opportunity to develop a relationship with his child. The Court upheld a stepparent adoption over the natural father’s objections, but acknowledged that “the existence or nonexistence of a substantial relationship between parent and child is a relevant criterion in evaluating both the rights of the parent and the best interests of the child.” Id., at 266-267. There, however, the father had never established a custodial, personal, or financial relationship with his child. Lehr had never lived with the child or the child’s mother after the birth of the child and had never provided any financial support. In the case now before us, Michael H. is not a father unwilling to assume his responsibilities as a parent. To the contrary, he is a father who has asserted his interests in raising and providing for his child since the very time of the child’s birth. In contrast to the father in Lehr, Michael had begun to develop a relationship with his daughter. There is no dispute on this point. Michael contributed to the child’s support. Michael and Victoria lived together (albeit intermittently, given Carole’s itinerant lifestyle). There is a personal and emotional relationship between Michael and Victoria, who grew up calling him “Daddy.” Michael held Victoria out as his daughter and contributed to the child’s financial support. (Even appellee concedes that Michael has “made greater efforts and had more success in establishing a 160 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. father-child relationship” than did Mr. Lehr. Brief for Appellee 13, n. 6.) The mother has never denied, and indeed has admitted, that Michael is Victoria’s father.2 Lehr was predicated on the absence of a substantial relationship between the man and the child and emphasized the “difference between the developed parent-child relationship that was implicated in Stanley and Caban, and the potential relationship involved in Quilloin and [Lehr]” Lehr, supra, at 261. “When an unwed father demonstrates a full commitment to the responsibilities of parenthood by ‘com[ing] forward to participate in the rearing of his child,’ Caban, supra, at 392, his interest in personal contact with his child acquires substantial protection under the Due Process Clause.” Lehr, supra, at 261. The facts in this case satisfy the Lehr criteria, which focused on the relationship between father and child, not on the relationship between father and mother. Under Lehr a “mere biological relationship” is not enough, but in light of Carole’s vicissitudes, what more could Michael have done? It is clear enough that Michael more than meets the mark in establishing the constitutionally protected liberty interest discussed in Lehr and recognized in Stanley v. Illinois, supra, and Caban v. Mohammed, supra. He therefore has a liberty interest entitled to protection under the Due Process Clause of the Fourteenth Amendment. II California plainly denies Michael this protection, by refusing him the opportunity to rebut the State’s presumption that the mother’s husband is the father of the child. California law not only deprives Michael of a legal parent-child relationship with his daughter Victoria but even denies him the opportunity to introduce blood-test evidence to rebut the de 2 As the plurality concedes, Carole signed a stipulation in April 1984 acknowledging that Michael was Victoria’s father. Ante, at 114-115. MICHAEL H. v. GERALD D. 161 110 White, J., dissenting monstrable fiction that Gerald is Victoria’s father.3 Unlike Lehr, Michael has not been denied notice. He has, most definitely, however, been denied any real opportunity to be heard. The grant of summary judgment against Michael was based on the conclusive presumption of Cal. Evid. Code Ann. § 621 (West Supp. 1989), which denied him the opportunity to prove that he is Victoria’s biological father. The Court gives its blessing to §621 by relying on the State’s asserted interests in the integrity of the family (defined as Carole and Gerald) and in protecting Victoria from the stigma of illegitimacy and by balancing away Michael’s interest in establishing that he is the father of the child. The interest in protecting a child from the social stigma of illegitimacy lacks any real connection to the facts of a case where a father is seeking to establish, rather than repudiate, paternity. The “stigma of illegitimacy” argument harks back to ancient common law when there were no blood tests to ascertain that the husband could not “by the laws of nature” be the child’s father. Judicial process refused to declare that a child born in wedlock was illegitimate unless the proof was positive. The only such proof was physical absence or impotency. But we have now clearly recognized the use of blood tests as an authoritative means of evaluating allegations of paternity. See, e. g., Little v. Streater, 452 U. S. 1, 6-7 (1981). I see no reason to debate the plurality’s multilingual explorations into “spousal nonaccess” and ancient policy concerns behind bastardy laws. It may be true that a child conceived in an extramarital relationship would 3 While the ultimate resolution of Michael’s case, were he permitted to introduce such evidence, might well be visitation rights or even custody of the child, it is important to keep in mind that the question at issue here is not whether he should be granted visitation or custody but simply whether he can take the first step in any such proceeding. Whatever the end result, Michael is simply asking that he be permitted to offer proof that he is Victoria’s father. In the instant case, that is likely to mean that he would introduce the blood tests that he and Carole took and which show that Michael is Victoria’s father. 162 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. be considered a “bastard” in the literal sense of the word, but whatever stigma remains in today’s society is far less compelling in the context of a child of a married mother, especially when there is a father asserting paternity and seeking a relationship with his child. It is hardly rare in this world of divorce and remarriage for a child to live with the “father” to whom her mother is married, and still have a relationship with her biological father. The State’s professed interest in the preservation of the existing marital unit is a more significant concern. To be sure, the intrusion of an outsider asserting that he is the father of a child whom the husband believes to be his own would be disruptive to say the least. On the facts of this case, however, Gerald was well aware of the liaison between Carole and Michael. The conclusive presumption of evidentiary rule §621 virtually eliminates the putative father’s chances of succeeding in his effort to establish paternity, but it by no means prevents him from asserting the claim. It may serve as a deterrent to such claims but does not eliminate the threat. Further, the argument that the conclusive presumption preserved the sanctity of the marital unit had more sway in a time when the husband was similarly prevented from challenging paternity.4 4 Even in the last quarter century, under California law, a husband whose blood test definitively showed he could not be the father of the child born to his wife was nonetheless not permitted to present this evidence to a court in order to refute the conclusive presumption of paternity. In 1967, however, the California courts began to erode the presumption as it applied to the husband, providing the husband with at least some opportunity to demonstrate that he was not the child’s father. Jackson v. Jackson, 67 Cal. 2d 245, 430 P. 2d 289 (1967). In 1980, the California Legislature amended §621 of its Evidence Code in order to permit the husband an opportunity to overcome the presumption that he is the father of his wife’s child if he raises the notice of motion for blood tests not later than two years from the birth of the child. (So much for the State’s interest in protecting the child from the stigma of illegitimacy!) MICHAEL H. v. GERALD D. 163 110 White, J., dissenting “The emphasis of the Due Process Clause is on ‘process.’” Moore v. East Cleveland, 431 U. S. 494, 542 (1977) (White, J., dissenting). I fail to see the fairness in the process established by the State of California and endorsed by the Court today. Michael has evidence which demonstrates that he is the father of young Victoria. Yet he is blocked by the State from presenting that evidence to a court. As a result, he is foreclosed from establishing his paternity and is ultimately precluded, by the State, from developing a relationship with his child. “A fundamental requirement of due process is ‘the opportunity to be heard.’ Grannis v. Or dean, 234 U. S. 385, 394. It is an opportunity which must be granted at a meaningful time and in a meaningful manner.” Armstrong v. Manzo, 380 U. S. 545, 552 (1965). I fail to see how Michael was granted any meaningful opportunity to be heard when he was precluded at the very outset from introducing evidence which would support his assertion of paternity. Michael has never been afforded an opportunity to present his case in any meaningful manner. As the Court has said: “The significance of the biological connection is that it offers the natural father an opportunity that no other male possesses to develop a relationship with his offspring. If he grasps that opportunity and accepts some measure of responsibility for the child’s future, he may enjoy the blessings of the parent-child relationship and make uniquely valuable contributions to the child’s development.” Lehr, 463 U. S., at 262. It is as if this passage was addressed to Michael. Yet the plurality today recants. Michael eagerly grasped the opportunity to have a relationship with his daughter (he lived with her; he declared her to be his child; he provided financial support for her) and still, with today’s opinion, his opportunity has vanished. He has been rendered a stranger to his child. Because Cal. Evid. Code Ann. §621, as applied, should be held unconstitutional under the Due Process Clause of the Fourteenth Amendment, I respectfully dissent. 164 OCTOBER TERM, 1988 Syllabus 491 U. S. PATTERSON v. McLEAN CREDIT UNION CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT No. 87-107. Argued February 29, 1988—Reargued October 12, 1988— Decided June 15, 1989 Petitioner, a black woman, was employed by respondent credit union as a teller and file coordinator for 10 years until she was laid off. Thereafter, she brought this action in District Court under 42 U. S. C. § 1981, alleging that respondent had harassed her, failed to promote her to accounting clerk, and then discharged her, all because of her race. The District Court determined that a claim for racial harassment is not actionable under § 1981 and declined to submit that part of the case to the jury. The court instructed the jury, inter alia, that in order to prevail on her promotion-discrimination claim, petitioner had to prove that she was better qualified than the white employee who allegedly had received the promotion. The jury found for respondent on this claim, as well as on petitioner’s discriminatory-discharge claim. The Court of Appeals affirmed the judgment in favor of respondent. Held: 1. This Court will not overrule its decision in Runyon v. McCrary, 427 U. S. 160, that § 1981 prohibits racial discrimination in the making and enforcement of private contracts. Stare decisis compels the Court to adhere to that interpretation, absent some “special justification” not to do so. The burden borne by a party advocating the abandonment of an established precedent is greater where the Court is asked to overrule a point of statutory construction, which, unlike constitutional interpretation, may be altered by Congress. Here, no special justification has been shown for overruling Runyon, which has not been undermined by subsequent changes or development in the law, has not proved to be unworkable, and does not pose an obstacle to the realization of objectives embodied in other statutes, particularly Title VII of the Civil Rights Act of 1964. Furthermore, Runyon is entirely consistent with society’s deep commitment to the eradication of race-based discrimination. Pp. 171-175. 2. Racial harassment relating to the conditions of employment is not actionable under § 1981, which provides that “[a]ll persons . . . shall have the same right ... to make and enforce contracts ... as is enjoyed by white citizens,” because that provision does not apply to conduct which PATTERSON v. McLEAN CREDIT UNION 165 164 Syllabus occurs after the formation of a contract and which does not interfere with the right to enforce established contract obligations. Pp. 175-185. (a) Since § 1981 is restricted in its scope to forbidding racial discrimination in the “mak[ing] and enforce[ment]” of contracts, it cannot be construed as a general proscription of discrimination in all aspects of contract relations. It provides no relief where an alleged discriminatory act does not involve the impairment of one of the specified rights. The “right ... to make . . . contracts” extends only to the formation of a contract, such that § 1981’s prohibition encompasses the discriminatory refusal to enter into a contract with someone, as well as the offer to make a contract only on discriminatory terms. That right does not extend to conduct by the employer after the contract relation has been established, including breach of the contract’s terms or the imposition of discriminatory working conditions. The “right . . . to . . . enforce contracts” embraces only protection of a judicial or nonjudicial legal process, and of a right of access to that process, that will address and resolve contractlaw claims without regard to race. It does not extend beyond conduct by an employer which impairs an employee’s ability to enforce through legal process his or her established contract rights. Pp. 176-178. (b) Thus, petitioner’s racial harassment claim is not actionable under § 1981. With the possible exception of her claim that respondent’s refusal to promote her was discriminatory, none of the conduct which she alleges—that her supervisor periodically stared at her for minutes at a time, gave her more work than white employees, assigned her to demeaning tasks not given to white employees, subjected her to a racial slur, and singled her out for criticism, and that she was not afforded training for higher level jobs and was denied wage increases—involves either a refusal to make a contract with her or her ability to enforce her established contract rights. Rather, the conduct alleged is postformation conduct by the employer relating to the terms and conditions of continuing employment, which is actionable only under the more expansive reach of Title VII. Interpreting § 1981 to cover postformation conduct unrelated to an employee’s right to enforce her contract is not only inconsistent with the statute’s limitations, but also would undermine Title Vil’s detailed procedures for the administrative conciliation and resolution of claims, since § 1981 requires no administrative review or opportunity for conciliation. Pp. 178-182. (c) There is no merit to the contention that § 1981’s “same right” phrase must be interpreted to incorporate state contract law, such that racial harassment in the conditions of employment is actionable when, and only when, it amounts to a breach of contract under state law. That theory contradicts Runyon by assuming that § 1981’s prohibitions are limited to state-law protections. Moreover, racial harassment amount 166 OCTOBER TERM, 1988 Syllabus 491 U. S. ing to breach of contract, like racial harassment alone, impairs neither the right to make nor the right to enforce a contract. In addition, the theory would unjustifiably federalize all state-law breach of contract claims where racial animus is alleged, since § 1981 covers all types of contracts. Also without merit is the argument that § 1981 should be interpreted to reach racial harassment that is sufficiently “severe or pervasive” as effectively to belie any claim that the contract was entered into in a racially neutral manner. Although racial harassment may be used as evidence that a divergence in the explicit terms of particular contracts is explained by racial animus, the amorphous and manipulable “severe or pervasive” standard cannot be used to transform a nonactionable challenge to employment conditions into a viable challenge to the employer’s refusal to contract. Pp. 182-185. 3. The District Court erred when it instructed the jury that petitioner had to prove that she was better qualified than the white employee who allegedly received the accounting clerk promotion. Pp. 185-188. (a) Discriminatory promotion claims are actionable under § 1981 only where the promotion rises to the level of an opportunity for a new and distinct relation between the employer and the employee. Here, respondent has never argued that petitioner’s promotion claim is not cognizable under § 1981. Pp. 185-186. (b) The Title VII disparate-treatment framework of proof applies to claims of racial discrimination under § 1981. Thus, to make out a prima facie case, petitioner need only prove by a preponderance of the evidence that she applied for and was qualified for an available position, that she was rejected, and that the employer then either continued to seek applicants for the position, or, as is alleged here, filled the position with a white employee. The establishment of a prima facie case creates an inference of discrimination, which the employer may rebut by articulating a legitimate, nondiscriminatory reason for its action. Here, respondent did so by presenting evidence that it promoted the white applicant because she was better qualified for the job. Thereafter, however, petitioner should have had the opportunity to demonstrate that respondent’s proffered reasons for its decision were not its true reasons. There are a variety of types of evidence that an employee can introduce to show that an employer’s stated reasons are pretextual, and the plaintiff may not be limited to presenting evidence of a certain type. Thus, the District Court erred in instructing the jury that petitioner could carry her burden of persuasion only by showing that she was in fact better qualified than the person who got the job. Pp. 186-188. 805 F. 2d 1143, affirmed in part, vacated in part, and remanded. PATTERSON v. McLEAN CREDIT UNION 167 164 Syllabus Kennedy, J., delivered the opinion of the Court, in which Rehnquist, C. J., and White, O’Connor, and Scalia, JJ., joined. Brennan, J., filed an opinion concurring in the judgment in part and dissenting in part, in which Marshall and Blackmun, JJ., joined, and in Parts II-B, II-C, and III of which Stevens, J., joined, post, p. 189. Stevens, J., filed an opinion concurring in the judgment in part and dissenting in part, post, p. 219. Julius LeVonne Chambers reargued the cause for petitioner. Penda D. Hair argued the cause for petitioner on the original argument. With them on the briefs were Charles Stephen Ralston, Gail J. Wright, Eric Schnapper, Ronald L. Ellis, Harold L. Kennedy III, and Harvey L. Kennedy. Roger S. Kaplan reargued the cause for respondent. H. Lee Davis, Jr., argued the cause for respondent on the original argument. With them on the briefs were George E. Doughton, Jr., Anthony H. Atlas, Gary R. Kessler, and Earl M. Maltz* *Briefs of amici curiae urging reversal were filed for the United States by Solicitor General Fried, Assistant Attorney General Reynolds, Deputy Solicitor General Ayer, Deputy Assistant Attorney General Clegg, Glen D. Nager, and Jessica Dunsay Silver; and for the American Civil Liberties Union Foundation et al. by Steven R. Shapiro, John A. Powell, Helen Hershkoff, and Adam Stein. Robert E. Williams, Douglas S. McDowell, and Lorence L. Kessler filed a brief for the Equal Employment Advisory Council as amicus curiae urging affirmance. Briefs of amici curiae were filed for 66 Members of the United States Senate et al. by John H. Pickering, Timothy B. Dyk, James E. Coleman, Jr., John Payton, Kerry W. Kircher, Edward H. Levi, Laurence H. Tribe, and William L. Taylor; for the State of New York et al. by Robert Abrams, Attorney General of New York, 0. Peter Sherwood, Solicitor General, Suzanne M. Lynn and Sanford M. Cohen, Assistant Attorneys General, James M. Shannon, Attorney General of Massachusetts, Barbara B. Dickey and Douglas T. Shwarz, Assistant Attorneys General, Hubert H. Humphrey III, Attorney General of Minnesota, Robert M. Spire, Attorney General of Nebraska, Dave Frohnmayer, Attorney General of Oregon, T. Travis Medlock, Attorney General of South Carolina, W. J. Michael Cody, Attorney General of Tennessee, Don Siegelman, Attorney General of Alabama, Grace Berg Schaible, Attorney General of Alaska, John Steven Clark, Attorney General of Arkansas, John K. Van 168 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Justice Kennedy delivered the opinion of the Court. In this case, we consider important issues respecting the meaning and coverage of one of our oldest civil rights statutes, 42 U. S. C. § 1981. de Kamp, Attorney General of California, Duane Woodard, Attorney General of Colorado, Joseph Lieberman, Attorney General of Connecticut, Charles M. Oberly, Attorney General of Delaware, Robert Butterworth, Attorney General of Florida, Michael J. Bowers, Attorney General of Georgia, Warren Price III, Attorney General of Hawaii, Jim Jones, Attorney General of Idaho, Neil F. Hartigan, Attorney General of Illinois, Linley E. Pearson, Attorney General of Indiana, Thomas J. Miller, Attorney General of Iowa, Robert T. Stephan, Attorney General of Kansas, Frederick J. Cowan, Attorney General of Kentucky, William J. Guste, Jr., Attorney General of Louisiana, James E. Tierney, Attorney General of Maine, J. Joseph Curran, Jr., Attorney General of Maryland, Frank J. Kelley, Attorney General of Michigan, Michael C. Moore, Attorney General of Mississippi, William L. Webster, Attorney General of Missouri, Mike Greely, Attorney General of Montana, Brian McKay, Attorney General of Nevada, Stephen E. Merrill, Attorney General of New Hampshire, Cary Edwards, Attorney General of New Jersey, Lacy H. Thornburg, Attorney General of North Carolina, Nicholas Spaeth, Attorney General of North Dakota, Anthony J. Celebrezze, Jr., Attorney General of Ohio, Robert Henry, Attorney General of Oklahoma, LeRoy S. Zimmerman, Attorney General of Pennsylvania, James E. O'Neil, Attorney General of Rhode Island, Roger A. Tellinghuisen, Attorney General of South Dakota, Jim Mattox, Attorney General of Texas, Jeffrey Amestoy, Attorney General of Vermont, Mary Sue Terry, Attorney General of Virginia, Kenneth 0. Eikenberry, Attorney General of Washington, Charles G. Brown, Attorney General of West Virginia, Don Hanaway, Attorney General of Wisconsin, Joseph B. Meyer, Attorney General of Wyoming, Godfrey R. deCastro, Acting Attorney General of the Virgin Islands, Frederick D. Cooke, Corporation Counsel of the District of Columbia, Hector Rivera-Cruz, Attorney General of Puerto Rico, and Elizabeth Barrett-Anderson, Attorney General of Guam; for the American Bar Association by Robert MacCrate, William H. Allen, and Mitchell F. Dolin; for the American Jewish Congress et al. by Marvin E. Frankel and Marc D. Stem; for the Association of the Bar of the City of New York et al. by Jonathan Lang, Howard J. Aibel, and Charles S. Sims; for the Center for Civil Rights by Clint Bolick, Jerald L. Hill, and Mark J. Bredemeier; for the Center for Constitutional Rights et al. by Esmeralda Simmons, Arthur Kinoy, Frank E. Deale, and Wilhelm Joseph; for the Lawyers’ PATTERSON v. McLEAN CREDIT UNION 169 164 Opinion of the Court I Petitioner Brenda Patterson, a black woman, was employed by respondent McLean Credit Union as a teller and a file coordinator, commencing in May 1972. In July 1982, she was laid off. After the termination, petitioner commenced this action in the United States District Court for the Middle District of North Carolina. She alleged that respondent, in violation of 14 Stat. 27, 42 U. S. C. § 1981, had harassed her, failed to promote her to an intermediate accounting clerk position, and then discharged her, all because of her race. Petitioner also claimed this conduct amounted to an intentional infliction of emotional distress, actionable under North Carolina tort law. The District Court determined that a claim for racial harassment is not actionable under § 1981 and declined to sub- committee for Civil Rights Under Law by Thomas D. Barr, Robert F. Mullen, Conrad K. Harper, Stuart J. Land, Norman Redlich, William L. Robinson, Judith A. Winston, Richard T. Seymour, Stephen L Spitz, Albert E. Arent, Thomas I. Atkins, St. John Barrett, Wiley A. Branton, Sr., Paul A. Brest, David R. Brink, William H. Brown III, Ramsey Clark, Jerome A. Cooper, Michael A. Cooper, Lloyd N. Cutler, James T. Danaher, Drew S. Days III, Armand Derfner, Paul R. Dimond, John W. Douglas, Victor M. Earle III, Robert Ehrenbard, Fred N. Fishman, MacDonald Flinn, Laurence S. Fordham, Eleanor M. Fox, John D. French, Lloyd K. Garrison, A. Spencer Gilbert III, Joan Hall, Herbert J. Hansell, JohnB. Jones, Stuart L. Kadison, Robert H. Kapp, Nicholas deB. Katzenbach, Robert M. Landis, Jerome B. Libin, John V. Lindsay, Hans F. Loeser, Henry L. Marsh III, Robert W. Meserve, Robert B. McKay, Peter P. Mullen, Robert A. Murphy, John E. Nolan, Jr., Kenneth Penegar, Charles S. Rhyne, Elliot L. Richardson, James Robertson, Mitchell Rogovin, Edwin A. Rothschild, Stephen H. Sachs, Bernard G. Segal, Jerome G. Shapiro, Jerome J. Shestack, Asa D. Sokolow, Chesterfield Smith, David S. Tatel, Randolph W. Thrower, John E. Tobin, Michael Traynor, Mama S. Tucker, Harold R. Tyler, Jr., Herbert M. Wachtell, and Howard P. Willens; for the Washington Legal Foundation et al. by Daniel J. Popeo and Paul D. Kamenar; for J. Philip Anderegg, pro se; for Carol L. Bisharat et al. by Eva Jefferson Paterson, Nathaniel Colley, William C. McNeill III, and Robert L. Harris; for Curtis McCrary et al. by Gary T. Brown; and for Eric Foner et al. by Richard D. Parsons. 170 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. mit that part of the case to the jury. The jury did receive and deliberate upon petitioner’s § 1981 claims based on alleged discrimination in her discharge and the failure to promote her, and it found for respondent on both claims. As for petitioner’s state-law claim, the District Court directed a verdict for respondent on the ground that the employer’s conduct did not rise to the level of outrageousness required to state a claim for intentional infliction of emotional distress under applicable standards of North Carolina law. In the Court of Appeals, petitioner raised two matters which are relevant here. First, she challenged the District Court’s refusal to submit to the jury her § 1981 claim based on racial harassment. Second, she argued that the District Court erred in instructing the jury that in order to prevail on her § 1981 claim of discriminatory failure to promote, she must show that she was better qualified than the white employee who she alleges was promoted in her stead. The Court of Appeals affirmed. 805 F. 2d 1143 (1986). On th racial harassment issue, the court held that, while instances of racial harassment “may implicate the terms and conditions of employment under Title VII [of the Civil Rights Act of 1964, 78 Stat. 253, 42 U. S. C. §2000e et seq.} and of course may be probative of the discriminatory intent required to be shown in a § 1981 action,” id., at 1145 (citation omitted), racial harassment itself is not cognizable under § 1981 because “racial harassment does not abridge the right to ‘make’ and ‘enforce’ contracts,” id., at 1146. On the jury instruction issue, the court held that once respondent had advanced superior qualification as a legitimate nondiscriminatory reason for its promotion decision, petitioner had the burden of persuasion to show that respondent’s justification was a pretext and that she was better qualified than the employee who was chosen for the job. Id., at 1147. We granted certiorari to decide whether petitioner’s claim of racial harassment in her employment is actionable under § 1981, and whether the jury instruction given by the Dis- PATTERSON v. McLEAN CREDIT UNION 171 164 Opinion of the Court trict Court on petitioner’s § 1981 promotion claim was error. 484 U. S. 814 (1987). After oral argument on these issues, we requested the parties to brief and argue an additional question: “Whether or not the interpretation of 42 U. S. C. § 1981 adopted by this Court in Runyon v. McCrary, 427 U. S. 160 (1976), should be reconsidered.” Patterson v. McLean Credit Union, 485 U. S. 617 (1988). We now decline to overrule our decision in Runyon v. McCrary, 427 U. S. 160 (1976). We hold further that racial harassment relating to the conditions of employment is not actionable under § 1981 because that provision does not apply to conduct which occurs after the formation of a contract and which does not interfere with the right to enforce established contract obligations. Finally, we hold that the District Court erred in instructing the jury regarding petitioner’s burden in proving her discriminatory promotion claim. II In Runyon, the Court considered whether § 1981 prohibits private schools from excluding children who are qualified for admission, solely on the basis of race. We held that § 1981 did prohibit such conduct, noting that it was already well established in prior decisions that § 1981 “prohibits racial discrimination in the making and enforcement of private contracts.” Id., at 168, citing Johnson v. Railway Express Agency, Inc., 421 U. S. 454, 459-460 (1975); Tillman v. Wheaton-Haven Recreation Assn., Inc., 410 U. S. 431, 439-440 (1973). The arguments about whether Runyon was decided correctly in light of the language and history of the statute were examined and discussed with great care in our decision. It was recognized at the time that a strong case could be made for the view that the statute does not reach private conduct, see 427 U. S., at 186 (Powell, J., concurring); id., at 189 (Stevens, J., concurring); id., at 192 (White, J., dissenting), but that view did not prevail. Some Members of 172 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. this Court believe that Runyon was decided incorrectly, and others consider it correct on its own footing, but the question before us is whether it ought now to be overturned. We conclude after reargument that Runyon should not be overruled, and we now reaffirm that § 1981 prohibits racial discrimination in the making and enforcement of private contracts. The Court has said often and with great emphasis that “the doctrine of stare decisis is of fundamental importance to the rule of law.” Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468, 494 (1987). Although we have cautioned that “stare decisis is a principle of policy and not a mechanical formula of adherence to the latest decision,” Boys Markets, Inc. v. Retail Clerks, 398 U. S. 235, 241 (1970), it is indisputable that stare decisis is a basic self-governing principle within the Judicial Branch, which is entrusted with the sensitive and difficult task of fashioning and preserving a jurisprudential system that is not based upon “an arbitrary discretion.” The Federalist, No. 78, p. 490 (H. Lodge ed. 1888) (A. Hamilton). See also Vasquez v. Hillery, 474 U. S. 254, 265 (1986) (stare decisis ensures that “the law will not merely change erratically” and “permits society to presume that bedrock principles are founded in the law rather than in the proclivities of individuals”). Our precedents are not sacrosanct, for we have overruled prior decisions where the necessity and propriety of doing so has been established. See Patterson v. McLean Credit Union, supra, at 617-618 (citing cases). Nonetheless, we have held that “any departure from the doctrine of stare decisis demands special justification.” Arizona v. Rumsey, 467 U. S. 203, 212 (1984). We have said also that the burden borne by the party advocating the abandonment of an established precedent is greater where the Court is asked to overrule a point of statutory construction. Considerations of stare decisis have special force in the area of statutory interpretation, for here, unlike in the context of constitutional interpretation, the legislative power is implicated, PATTERSON v. McLEAN CREDIT UNION 173 164 Opinion of the Court and Congress remains free to alter what we have done. See, e. g., Square D Co. v. Niagara Frontier Tariff Bureau, Inc., 476 U. S. 409, 424 (1986); Illinois Brick Co. n. Illinois, 431 U. S. 720, 736 (1977). We conclude, upon direct consideration of the issue, that no special justification has been shown for overruling Runyon. In cases where statutory precedents have been overruled, the primary reason for the Court’s shift in position has been the intervening development of the law, through either the growth of judicial doctrine or further action taken by Congress. Where such changes have removed or weakened the conceptual underpinnings from the prior decision, see, e. g., Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U. S. 477, 480-481 (1989); Andrews v. Louisville & Nashville R. Co., 406 U. S. 320, 322-323 (1972), or where the later law has rendered the decision irreconcilable with competing legal doctrines or policies, see, e. g., Braden v. 30th Judicial Circuit Ct. of Ky., 410 U. S. 484, 497-499 (1973); Construction Laborers v. Curry, 371 U. S. 542, 552 (1963), the Court has not hesitated to overrule an earlier decision. Our decision in Runyon has not been undermined by subsequent changes or development in the law. Another traditional justification for overruling a prior case is that a precedent may be a positive detriment to coherence and consistency in the law, either because of inherent confusion created by an unworkable decision, see, e. g., Continental T. V., Inc. v. GTE Sylvania, Inc., 433 U. S. 36, 47-48 (1977); Swift & Co. v. Wickham, 382 U. S. Ill, 124-125 (1965), or because the decision poses a direct obstacle to the realization of important objectives embodied in other laws, see, e. g., Rodriguez de Quijas, supra, at 484; Boys Markets, Inc. v. Retail Clerks, supra, at 240-241. In this regard, we do not find Runyon to be unworkable or confusing. Respondent and various amici have urged that Runyon’s interpretation of § 1981, as applied to contracts of employment, frustrates the objectives of Title VII. The argument is that 174 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. a substantial overlap in coverage between the two statutes, given the considerable differences in their remedial schemes, undermines Congress’ detailed efforts in Title VII to resolve disputes about racial discrimination in private employment through conciliation rather than litigation as an initial matter. After examining the point with care, however, we believe that a sound construction of the language of § 1981 yields an interpretation which does not frustrate the congressional objectives in Title VII to any significant degree. See Part III, infra. Finally, it has sometimes been said that a precedent becomes more vulnerable as it becomes outdated and after being “‘tested by experience, has been found to be inconsistent with the sense of justice or with the social welfare.’” Runyon, 427 U. S., at 191 (Stevens, J., concurring), quoting B. Cardozo, The Nature of the Judicial Process 149 (1921). Whatever the effect of this consideration may be in statutory cases, it offers no support for overruling Runyon. In recent decades, state and federal legislation has been enacted to prohibit private racial discrimination in many aspects of our society. Whether Runyon’s interpretation of § 1981 as prohibiting racial discrimination in the making and enforcement of private contracts is right or wrong as an original matter, it is certain that it is not inconsistent with the prevailing sense of justice in this country. To the contrary, Runyon is entirely consistent with our society’s deep commitment to the eradication of discrimination based on a person’s race or the color of his or her skin. See Bob Jones University n. United States, 461 U. S. 574, 593 (1983) (“[E]very pronouncement of this Court and myriad Acts of Congress and Executive Orders attest a firm national policy to prohibit racial segregation and discrimination”); see also Brown v. Board of Education, 347 U. S. 483 (1954); Plessy n. Ferguson, 163 U. S. 537, 559 (1896) (Harlan, J., dissenting) (“The law regards man as man, and takes no account of his . . . PATTERSON v. McLEAN CREDIT UNION 175 164 Opinion of the Court color when his civil rights as guaranteed by the supreme law of the land are involved”).1 We decline to overrule Runyon and acknowledge that its holding remains the governing law in this area. Ill Our conclusion that we should adhere to our decision in Runyon that § 1981 applies to private conduct is not enough to decide this case. We must decide also whether the con 1 Justice Brennan chides us for ignoring what he considers “two very obvious reasons” for adhering to Runyon. Post, at 191. First, he argues at length that Runyon was correct as an initial matter. See post, at 191-199. As we have said, however, see supra, at 171-172, it is unnecessary for us to address this issue because we agree that, whether or not Runyon was correct as an initial matter, there is no special justification for departing here from the rule of stare decisis. Justice Brennan objects also to the fact that our stare decisis analysis places no reliance on the fact that Congress itself has not overturned the interpretation of § 1981 contained in Runyon, and in effect has ratified our decision in that case. See post, at 200-205. This is no oversight on our part. As we reaffirm today, considerations of stare decisis have added force in statutory cases because Congress may alter what we have done by amending the statute. In constitutional cases, by contrast, Congress lacks this option, and an incorrect or outdated precedent may be overturned only by our own reconsideration or by constitutional amendment. See supra, at 172-173. It does not follow, however, that Congress’ failure to overturn a statutory precedent is reason for this Court to adhere to it. It is “impossible to assert with any degree of assurance that congressional failure to act represents” affirmative congressional approval of the Court’s statutory interpretation. Johnson v. Transportation Agency, Santa Clara County, 480 U. S. 616, 671-672 (1987) (Scalia, J., dissenting). Congress may legislate, moreover, only through the passage of a bill which is approved by both Houses and signed by the President. See U. S. Const., Art. I, §7, cl. 2. Congressional inaction cannot amend a duly enacted statute. We think also that the materials relied upon by Justice Brennan as “more positive signs of Congress’ views,” which are the failure of an amendment to a different statute offered before our decision in Runyon, see post, at 201-204, and the passage of an attorney’s fee statute having nothing to do with our holding in Runyon, see post, at 204-205, demonstrate well the danger of placing undue reliance on the concept of congressional “ratification.” 176 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. duct of which petitioner complains falls within one of the enumerated rights protected by § 1981. A Section 1981 reads as follows: “All persons within the jurisdiction of the United States shall have the same right in every State and Territory to make and enforce contracts, to sue, be parties, give evidence, and to the full and equal benefit of all laws and proceedings for the security of persons and property as is enjoyed by white citizens, and shall be subject to like punishment, pains, penalties, taxes, licenses, and exactions of every kind, and to no other.” Rev. Stat. § 1977. The most obvious feature of the provision is the restriction of its scope to forbidding discrimination in the “mak[ing] and enforce[ment]” of contracts alone. Where an alleged act of discrimination does not involve the impairment of one of these specific rights, § 1981 provides no relief. Section 1981 cannot be construed as a general proscription of racial discrimination in all aspects of contract relations, for it expressly prohibits discrimination only in the making and enforcement of contracts. See also Jones n. Alfred H. Mayer Co., 392 U. S. 409, 436 (1968) (§ 1982, the companion statute to § 1981, was designed “to prohibit all racial discrimination, whether or not under color of law, with respect to the rights enumerated therein”) (emphasis added); Georgia v. Rachel, 384 U. S. 780, 791 (1966) (“The legislative history of the 1866 Act clearly indicates that Congress intended to protect a limited category of rights”). By its plain terms, the relevant provision in § 1981 protects two rights: “the same right ... to make . . . contracts” and “the same right . . . to . . . enforce contracts.” The first of these protections extends only to the formation of a contract, but not to problems that may arise later from the conditions of continuing employment. The statute prohibits, PATTERSON v. McLEAN CREDIT UNION 177 164 Opinion of the Court when based on race, the refusal to enter into a contract with someone, as well as the offer to make a contract only on discriminatory terms. But the right to make contracts does not extend, as a matter of either logic or semantics, to conduct by the employer after the contract relation has been established, including breach of the terms of the contract or imposition of discriminatory working conditions. Such postformation conduct does not involve the right to make a contract, but rather implicates the performance of established contract obligations and the conditions of continuing employment, matters more naturally governed by state contract law and Title VII. See infra, at 179-180. The second of these guarantees, “the same right. . . to . . . enforce contracts ... as is enjoyed by white citizens,” embraces protection of a legal process, and of a right of access to legal process, that will address and resolve contract-law claims without regard to race. In this respect, it prohibits discrimination that infects the legal process in ways that prevent one from enforcing contract rights, by reason of his or her race, and this is so whether this discrimination is attributed to a statute or simply to existing practices. It also covers wholly private efforts to impede access to the courts or obstruct nonjudicial methods of adjudicating disputes about the force of binding obligations, as well as discrimination by private entities, such as labor unions, in enforcing the terms of a contract. Following this principle and consistent with our holding in Runyon that § 1981 applies to private conduct, we have held that certain private entities such as labor unions, which bear explicit responsibilities to process grievances, press claims, and represent member in disputes over the terms of binding obligations that run from the employer to the employee, are subject to liability under § 1981 for racial discrimination in the enforcement of labor contracts. See Goodman v. Lukens Steel Co., 482 U. S. 656 (1987). The right to enforce contracts does not, however, extend beyond conduct by an employer which impairs an employee’s ability 178 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. to enforce through legal process his or her established contract rights. As Justice White put it with much force in Runyon, one cannot seriously “contend that the grant of the other rights enumerated in § 1981, [that is, other than the right to “make” contracts,] i. e., the rights ‘to sue, be parties, give evidence,’ and ‘enforce contracts’ accomplishes anything other than the removal of legal disabilities to sue, be a party, testify or enforce a contract. Indeed, it is impossible to give such language any other meaning.” 427 U. S., at 195, n. 5 (dissenting opinion) (emphasis in original). B Applying these principles to the case before us, we agree with the Court of Appeals that petitioner’s racial harassment claim is not actionable under § 1981. Petitioner has alleged that during her employment with respondent, she was subjected to various forms of racial harassment from her supervisor. As summarized by the Court of Appeals, petitioner testified that “[her supervisor] periodically stared at her for several minutes at a time; that he gave her too many tasks, causing her to complain that she was under too much pressure; that among the tasks given her were sweeping and dusting, jobs not given to white employees. On one occasion, she testified, [her supervisor] told [her] that blacks are known to work slower than whites. According to [petitioner, her supervisor] also criticized her in staff meetings while not similarly criticizing white employees.” 805 F. 2d, at 1145. Petitioner also alleges that she was passed over for promotion, not offered training for higher level jobs, and denied wage increases, all because of her race.2 2 In addition, another of respondent’s managers testified that when he recommended a different black person for a position as a data processor, petitioner’s supervisor stated that he did not “need any more problems PATTERSON v. McLEAN CREDIT UNION 179 164 Opinion of the Court With the exception perhaps of her claim that respondent refused to promote her to a position as an accountant, see Part IV, infra, none of the conduct which petitioner alleges as part of the racial harassment against her involves either a refusal to make a contract with her or the impairment of her ability to enforce her established contract rights. Rather, the conduct which petitioner labels as actionable racial harassment is postformation conduct by the employer relating to the terms and conditions of continuing employment. This is apparent from petitioner’s own proposed jury instruction on her § 1981 racial harassment claim: . The plaintiff has also brought an action for harassment in employment against the defendant, under the same statute, 42 USC § 1981. An employer is guilty of racial discrimination in employment where it has either created or condoned a substantially discriminatory work environment. An employee has a right to work in an environment free from racial prejudice. If the plaintiff has proven by a preponderance of the evidence that she was subjected to racial harassment by her manager while employed at the defendant, or that she was subjected to a work environment not free from racial prejudice which was either created or condoned by the defendant, then it would be your duty to find for plaintiff on this issue.” 1 Record, Doc. No. 18, p. 4 (emphasis added). Without passing on the contents of this instruction, it is plain to us that what petitioner is attacking are the conditions of her employment. This type of conduct, reprehensible though it be if true, is not actionable under § 1981, which covers only conduct at the initial formation of the contract and conduct which impairs the right to enforce contract obligations through legal proc around here,” and that he would “search for additional people who are not black.” Tr. 2-160 to 2-161. 180 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ess. Rather, such conduct is actionable under the more expansive reach of Title VII of the Civil Rights Act of 1964. The latter statute makes it unlawful for an employer to “discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment.” 42 U. S. C. §2000e-2(a)(l). Racial harassment in the course of employment is actionable under Title Vil’s prohibition against discrimination in the “terms, conditions, or privileges of employment.” “[T]he [Equal Employment Opportunity Commission (EEOC)] has long recognized that harassment on the basis of race ... is an unlawful employment practice in violation of § 703 of Title VII of the Civil Rights Act.” See 2 EEOC Compliance Manual §615.7 (1982). While this Court has not yet had the opportunity to pass directly upon this interpretation of Title VII, the lower federal courts have uniformly upheld this view,3 and we implicitly have approved it in a recent decision concerning sexual harassment, Meritor Savings Bank v. Vinson, 477 U. S. 57, 65-66 (1986). As we said in that case, “harassment [which is] sufficiently severe or pervasive ‘to alter the conditions of [the victim’s] employment and create an abusive working environment,’” id., at 67 (citation omitted), is actionable under Title VII because it “affects a ‘term, condition, or privilege’ of employment,” ibid. Interpreting § 1981 to cover postformation conduct unrelated to an employee’s right to enforce his or her contract, such as incidents relating to the conditions of employment, is not only inconsistent with that statute’s limitation to the making and enforcement of contracts, but would also undermine the detailed and well-crafted procedures for conciliation and resolution of Title VII claims. In Title VII, Congress set up an elaborate administrative procedure, implemented through the EEOC, that is designed to assist in the investi- 3 See, e. g., Firefighters Institute for Racial Equality v. St. Louis, 549 F. 2d 506, 514-515 (CA8), cert, denied sub nom. Banta v. United States, 434 U. S. 819 (1977); Rogers v. EEOC, 454 F. 2d 234 (CA5 1971), cert, denied, 406 U. S. 957 (1972). PATTERSON v. McLEAN CREDIT UNION 181 164 Opinion of the Court gation of claims of racial discrimination in the workplace and to work towards the resolution of these claims through conciliation rather than litigation. See 42 U. S. C. § 2000e-5(b). Only after these procedures have been exhausted, and the plaintiff has obtained a “right to sue” letter from the EEOC, may he or she bring a Title VII action in court. See 42 U. S. C. §2000e-5(f)(l). Section 1981, by contrast, provides no administrative review or opportunity for conciliation. Where conduct is covered by both § 1981 and Title VII, the detailed procedures of Title VII are rendered a dead letter, as the plaintiff is free to pursue a claim by bringing suit under § 1981 without resort to those statutory prerequisites. We agree that, after Runyon, there is some necessary overlap between Title VII and § 1981, and that where the statutes do in fact overlap we are not at liberty “to infer any positive preference for one over the other.” Johnson v. Railway Express Agency, Inc., 421 U. S., at 461. We should be reluctant, however, to read an earlier statute broadly where the result is to circumvent the detailed remedial scheme constructed in a later statute. See United States v. Fausto, 484 U. S. 439 (1988). That egregious racial harassment of employees is forbidden by a clearly applicable law (Title VII), moreover, should lessen the temptation for this Court to twist the interpretation of another statute (§ 1981) to cover the same conduct. In the particular case before us, we do not know for certain why petitioner chose to pursue only remedies under § 1981, and not under Title VII. See 805 F. 2d, at 1144, n.; Tr. of Oral Arg. 15-16, 23 (Feb. 29, 1988). But in any event, the availability of the latter statute should deter us from a tortuous construction of the former statute to cover this type of claim. By reading § 1981 not as a general proscription of racial discrimination in all aspects of contract relations, but as limited to the enumerated rights within its express protection, specifically the right to make and enforce contracts, we may preserve the integrity of Title Vil’s procedures without sacrific 182 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ing any significant coverage of the civil rights laws.4 Of course, some overlap will remain between the two statutes: specifically, a refusal to enter into an employment contract on the basis of race. Such a claim would be actionable under Title VII as a “refus[al] to hire” based on race, 42 U. S. C. §2000e-2(a), and under § 1981 as an impairment of “the same right... to make . . . contracts . . . as . . . white citizens,” 42 U. S. C. §1981. But this is precisely where it would make sense for Congress to provide for the overlap. At this stage of the employee-employer relation Title Vil’s mediation and conciliation procedures would be of minimal effect, for there is not yet a relation to salvage. C The Solicitor General and Justice Brennan offer two alternative interpretations of § 1981. The Solicitor General argues that the language of § 1981, especially the words “the same right,” requires us to look outside § 1981 to the terms of particular contracts and to state law for the obligations and covenants to be protected by the federal statute. Under this view, § 1981 has no actual substantive content, but instead mirrors only the specific protections that are afforded under the law of contracts of each State. Under this view, racial harassment in the conditions of employment is actionable when, and only when, it amounts to a breach of contract under state law. We disagree. For one thing, to the extent that it assumes that prohibitions contained in § 1981 incorporate only those protections afforded by the States, this theory is directly inconsistent with Runyon, which we today 4 Unnecessary overlap between Title VII and § 1981 would also serve to upset the delicate balance between employee and employer rights struck by Title VII in other respects. For instance, a plaintiff in a Title VII action is limited to a recovery of backpay, whereas under § 1981 a plaintiff may be entitled to plenary compensatory damages, as well as punitive damages in an appropriate case. Both the employee and employer will be unlikely to agree to a conciliatory resolution of the dispute under Title VII if the employer can be found liable for much greater amounts under § 1981. PATTERSON v. McLEAN CREDIT UNION 183 164 Opinion of the Court decline to overrule. A more fundamental failing in the Solicitor’s argument is that racial harassment amounting to breach of contract, like racial harassment alone, impairs neither the right to make nor the right to enforce a contract. It is plain that the former right is not implicated directly by an employer’s breach in the performance of obligations under a contract already formed. Nor is it correct to say that racial harassment amounting to a breach of contract impairs an employee’s right to enforce his contract. To the contrary, conduct amounting to a breach of contract under state law is precisely what the language of § 1981 does not cover. That is because, in such a case, provided that plaintiff’s access to state court or any other dispute resolution process has not been impaired by either the State or a private actor, see Goodman v. Lukens Steel Co., 482 U. S. 656 (1987), the plaintiff is free to enforce the terms of the contract in state court, and cannot possibly assert, by reason of the breach alone, that he has been deprived of the same right to enforce contracts as is enjoyed by white citizens. In addition, interpreting § 1981 to cover racial harassment amounting to a breach of contract would federalize all state-law claims for breach of contract where racial animus is alleged, since § 1981 covers all types of contracts, not just employment contracts. Although we must do so when Congress plainly directs, as a rule we should be and are “reluctant to federalize” matters traditionally covered by state common law. Santa Fe Industries, Inc. v. Green, 430 U. S. 462, 479 (1977); see also Sedima, S. P. R. L. v. Imrex Co., 473 U. S. 479, 507 (1985) (Marshall, J., dissenting). By confining § 1981 to the impairment of the specific rights to make and enforce contracts, Congress cannot be said to have intended such a result with respect to breach of contract claims. It would be no small paradox, moreover, that under the interpretation of § 1981 offered by the Solicitor General, the more a State extends its own contract law to protect employees in general and minorities in particular, the greater 184 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. would be the potential displacement of state law by § 1981. We do not think § 1981 need be read to produce such a peculiar result. Justice Brennan, for his part, would hold that racial harassment is actionable under § 1981 when “the acts constituting harassment [are] sufficiently severe or pervasive as effectively to belie any claim that the contract was entered into in a racially neutral manner.” See post, at 208. We do not find this standard an accurate or useful articulation of which contract claims are actionable under § 1981 and which are not. The fact that racial harassment is “severe or pervasive” does not by magic transform a challenge to the conditions of employment, not actionable under § 1981, into a viable challenge to the employer’s refusal to make a contract. We agree that racial harassment may be used as evidence that a divergence in the explicit terms of particular contracts is explained by racial animus.5 Thus, for example, if a potential employee is offered (and accepts) a contract to do a job for less money than others doing like work, evidence of racial harassment in the workplace may show that the employer, at the time of formation, was unwilling to enter into a nondiscriminatory contract. However, and this is the critical point, the question under § 1981 remains whether the employer, at the time of the formation of the contract, in fact intentionally refused to enter into a contract with the employee on racially neutral terms. The plaintiff’s ability to plead that the racial harassment is “severe or pervasive” should not allow him to bootstrap a challenge to the conditions of employment (actionable, if at all, under Title VII) into a claim under § 1981 that the employer refused to offer petitioner the “same right ... to make” a contract. We think it clear that the conduct challenged by petitioner relates not to her employer’s refusal to 5 This was the permissible use of evidence of racial harassment that the Fourth Circuit, in its decision below, envisioned for § 1981 cases. See 805 F. 2d 1143, 1145 (1986). PATTERSON v. McLEAN CREDIT UNION 185 164 Opinion of the Court enter into a contract with her, but rather to the conditions of her employment.6 IV Petitioner’s claim that respondent violated § 1981 by failing to promote her, because of race, to a position as an intermediate accounting clerk is a different matter. As a preliminary point, we note that the Court of Appeals distinguished between petitioner’s claims of racial harassment and discriminatory promotion, stating that although the former did not give rise to a discrete § 1981 claim, “[c]laims of racially discriminatory . . . promotion go to the very existence and nature of the employment contract and thus fall easily within § 1981 ’s protection.” 805 F. 2d, at 1145. We think that somewhat overstates the case. Consistent with what we have said in Part III, supra, the question whether a promotion claim is actionable under § 1981 depends upon whether the nature of the change in position was such that it involved the opportunity to enter into a new contract with the employer. If so, then the employer’s refusal to enter the new contract is actionable under §1981. In making this determination, a lower court should give a fair and natural reading to the statutory phrase “the same right ... to make . . . contracts,” and should not strain in an undue manner the language of § 1981. Only where the promotion rises to the level of an opportunity for a new and distinct relation between the employee and the employer is such a claim actionable under § 1981. Cf. Hishon v. King & Spaulding, 467 U. S. 69 (1984) 6 In his separate opinion, Justice Stevens construes the phrase “the same right ... to make . . . contracts” with ingenuity to cover various postformation conduct by the employer. But our task here is not to construe § 1981 to punish all acts of discrimination in contracting in a like fashion, but rather merely to give a fair reading to scope of the statutory terms used by Congress. We adhere today to our decision in Runyon that § 1981 reaches private conduct, but do not believe that holding compels us to read the statutory terms “make” and “enforce” beyond their plain and commonsense meaning. We believe that the lower courts will have little difficulty applying the straightforward principles that we announce today. 186 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. (refusal of law firm to accept associate into partnership) (Title VII). Because respondent has not argued at any stage that petitioner’s promotion claim is not cognizable under § 1981, we need not address the issue further here. This brings us to the question of the District Court’s jury instructions on petitioner’s promotion claim. We think the District Court erred when it instructed the jury that petitioner had to prove that she was better qualified than the white employee who allegedly received the promotion. In order to prevail under § 1981, a plaintiff must prove purposeful discrimination. General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S. 375, 391 (1982). We have developed, in analogous areas of civil rights law, a carefully designed framework of proof to determine, in the context of disparate treatment, the ultimate issue whether the defendant intentionally discriminated against the plaintiff. See Texas Dept, of Community Affairs v. Burdine, 450 U. S. 248 (1981); McDonnell Douglas Corp. v. Green, 411 U. S. 792 (1973). We agree with the Court of Appeals that this scheme of proof, structured as a “sensible, orderly way to evaluate the evidence in light of common experience as it bears on the critical question of discrimination,” Fumco Construction Corp. v. Waters, 438 U. S. 567, 577 (1978), should apply to claims of racial discrimination under § 1981. Although the Court of Appeals recognized that the McDonnell Douglas!Burdine scheme of proof should apply in § 1981 cases such as this one, it erred in describing petitioner’s burden. Under our well-established framework, the plaintiff has the initial burden of proving, by a preponderance of the evidence, a prima facie case of discrimination. Burdine, 450 U. S., at 252-253. The burden is not onerous. Id., at 253. Here, petitioner need only prove by a preponderance of the evidence that she applied for and was qualified for an available position, that she was rejected, and that after she was rejected respondent either continued to seek applicants for the position, or, as is alleged here, filled the position with a PATTERSON v. McLEAN CREDIT UNION 187 164 Opinion of the Court white employee. See id., at 253, and n. 6; McDonnell Douglas, supra, at 802.7 Once the plaintiff establishes a prima facie case, an inference of discrimination arises. See Burdine, 450 U. S., at 254. In order to rebut this inference, the employer must present evidence that the plaintiff was rejected, or the other applicant was chosen, for a legitimate nondiscrimina-tory reason. See ibid. Here, respondent presented evidence that it gave the job to the white applicant because she was better qualified for the position, and therefore rebutted any presumption of discrimination that petitioner may have established. At this point, as our prior cases make clear, petitioner retains the final burden of persuading the jury of intentional discrimination. See id., at 256. Although petitioner retains the ultimate burden of persuasion, our cases make clear that she must also have the opportunity to demonstrate that respondent’s proffered reasons for its decision were not its true reasons. Ibid. In doing so, petitioner is not limited to presenting evidence of a certain type. This is where the District Court erred. The evidence which petitioner can present in an attempt to establish that respondent’s stated reasons are pretextual may take a variety of forms. See McDonnell Douglas, supra, at 804-805; Furneo Construction Corp., supra, at 578; cf. United States Postal Service Bd. of Governors v. Aikens, 460 U. S. 711, 714, n. 3 (1983). Indeed, she might seek to demonstrate that respondent’s claim to have promoted a better qualified applicant was pretextual by showing that she was in fact 7 Here, respondent argues that petitioner cannot make out a prima facie case on her promotion claim because she did not prove either that respondent was seeking applicants for the intermediate accounting clerk position or that the white employee named to fill that position in fact received a “promotion” from her prior job. Although we express no opinion on the merits of these claims, we do emphasize that in order to prove that she was denied the same right to make and enforce contracts as white citizens, petitioner must show, inter alia, that she was in fact denied an available position. 188 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. better qualified than the person chosen for the position. The District Court erred, however, in instructing the jury that in order to succeed petitioner was required to make such a showing. There are certainly other ways in which petitioner could seek to prove that respondent’s reasons were pretextual. Thus, for example, petitioner could seek to persuade the jury that respondent had not offered the true reason for its promotion decision by presenting evidence of respondent’s past treatment of petitioner, including the instances of the racial harassment which she alleges and respondent’s failure to train her for an accounting position. See supra, at 178. While we do not intend to say this evidence necessarily would be sufficient to carry the day, it cannot be denied that it is one of the various ways in which petitioner might seek to prove intentional discrimination on the part of respondent. She may not be forced to pursue any particular means of demonstrating that respondent’s stated reasons are pretextual. It was, therefore, error for the District Court to instruct the jury that petitioner could carry her burden of persuasion only by showing that she was in fact better qualified than the white applicant who got the job. V The law now reflects society’s consensus that discrimination based on the color of one’s skin is a profound wrong of tragic dimension. Neither our words nor our decisions should be interpreted as signaling one inch of retreat from Congress’ policy to forbid discrimination in the private, as well as the public, sphere. Nevertheless, in the area of private discrimination, to which the ordinance of the Constitution does not directly extend, our role is limited to interpreting what Congress may do and has done. The statute before us, which is only one part of Congress’ extensive civil rights legislation, does not cover the acts of harassment alleged here. PATTERSON v. McLEAN CREDIT UNION 189 164 Opinion of Brennan, J. In sum, we affirm the Court of Appeals’ dismissal of petitioner’s racial harassment claim as not actionable under §1981. The Court of Appeals erred, however, in holding that petitioner could succeed in her discriminatory promotion claim under § 1981 only by proving that she was better qualified for the position of intermediate accounting clerk than the white employee who in fact was promoted. The judgment of the Court of Appeals is therefore vacated insofar as it relates to petitioner’s discriminatory promotion claim, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Justice Brennan, with whom Justice Marshall and Justice Blackmun join, and with whom Justice Stevens joins as to Parts II-B, II-C, and III, concurring in the judgment in part and dissenting in part. What the Court declines to snatch away with one hand, it takes with the other. Though the Court today reaffirms §1981’s applicability to private conduct, it simultaneously gives this landmark civil rights statute a needlessly cramped interpretation. The Court has to strain hard to justify this choice to confine § 1981 within the narrowest possible scope, selecting the most pinched reading of the phrase “same right to make a contract,” ignoring powerful historical evidence about the Reconstruction Congress’ concerns, and bolstering its parsimonious rendering by reference to a statute enacted nearly a century after § 1981, and plainly not intended to affect its reach. When it comes to deciding whether a civil rights statute should be construed to further our Nation’s commitment to the eradication of racial discrimination, the Court adopts a formalistic method of interpretation antithetical to Congress’ vision of a society in which contractual opportunities are equal. I dissent from the Court’s holding that § 1981 does not encompass Patterson’s racial harassment claim. 190 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. I Thirteen years ago, in deciding Runyon v. McCrary, 427 U. S. 160 (1976), this Court treated as already “well established” the proposition that “§ 1 of the Civil Rights Act of 1866, 14 Stat. 27, 42 U. S. C. § 1981, prohibits racial discrimination in the making and enforcement of private contracts,” as well as state-mandated inequalities, drawn along racial lines, in individuals’ ability to make and enforce contracts. Id., at 168, citing Johnson n. Railway Express Agency, Inc., 421 U. S. 454 (1975); Tillman v. Wheaton-Haven Recreation Assn., Inc., 410 U. S. 431 (1973); and Jones n. Alfred H. Mayer Co., 392 U. S. 409 (1968). Since deciding Runyon, we have upon a number of occasions treated as settled law its interpretation of § 1981 as extending to private discrimination. Goodman v. Lukens Steel Co., 482 U. S. 656 (1987); Saint Francis College n. Al-Khazraji, 481 U. S. 604 (1987); General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S. 375 (1982); Delaware State College n. Ricks, 449 U. S. 250 (1980); McDonald v. Santa Fe Trail Transp. Co., 427 U. S. 273 (1976). We have also reiterated our holding in Jones that 42 U. S. C. § 1982 similarly applies to private discrimination in the sale or rental of real or personal property—a holding arrived at through an analysis of legislative history common to both § 1981 and § 1982. Shaare Tefila Congregation v. Cobb, 481 U. S. 615 (1987); Sullivan v. Little Hunting Park, Inc., 396 U. S. 229 (1969). The Court’s reaffirmation of this long and consistent line of precedents establishing that § 1981 encompasses private discrimination is based upon its belated decision to adhere to the principle of stare decisis—a decision that could readily, and would better, have been made before the Court decided to put Runyon and its progeny into question by ordering reargument in this case. While there is an exception to stare decisis for precedents that have proved “outdated, . . . unworkable, or otherwise legitimately vulnerable to serious PATTERSON v. McLEAN CREDIT UNION 191 164 Opinion of Brennan, J. reconsideration,” Vasquez v. Hillery, 474 U. S. 254, 266 (1986), it has never been arguable that Runyon falls within it. Rather, Runyon is entirely consonant with our society’s deep commitment to the eradication of discrimination based on a person’s race or the color of her skin. See Bob Jones University v. United States, 461 U. S. 574, 593 (1983) (“[E]very pronouncement of this Court and myriad Acts of Congress and Executive Orders attest a firm national policy to prohibit racial segregation and discrimination”). That commitment is not bounded by legal concepts such as “state action,” but is the product of a national consensus that racial discrimination is incompatible with our best conception of our communal life, and with each individual’s rightful expectation that her full participation in the community will not be contingent upon her race. In the past, this Court has overruled decisions antagonistic to our Nation’s commitment to the ideal of a society in which a person’s opportunities do not depend on her race, e. g., Brown n. Board of Education, 347 U. S. 483 (1954) (overruling Plessy v. Ferguson, 163 U. S. 537 (1896)), and I find it disturbing that the Court has in this case chosen to reconsider, without any request from the parties, a statutory construction so in harmony with that ideal. Having decided, however, to reconsider Runyon, and now to reaffirm it by appeal to stare decisis, the Court glosses over what are in my view two very obvious reasons for refusing to overrule this interpretation of § 1981: that Runyon was correctly decided, and that in any event Congress has ratified our construction of the statute. A A survey of our cases demonstrates that the Court’s interpretation of § 1981 has been based upon a full and considered review of the statute’s language and legislative history, assisted by careful briefing, upon which no doubt has been cast by any new information or arguments advanced in the briefs filed in this case. 192 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. In Jones v. Alfred H. Mayer Co., supra, this Court considered whether § 1982, which provides that “[a]ll citizens of the United States shall have the same right, in every State and Territory, as is enjoyed by white citizens thereof to inherit, purchase, lease, sell, hold, and convey real and personal property,” prohibits private discrimination on the basis of race, and if so, whether the statute is constitutional. The Court held, over two dissenting votes, that § 1982 bars private, as well as public, racial discrimination, and that the statute was a valid exercise of Congress’ power under § 2 of the Thirteenth Amendment to identify the badges and incidents of slavery and to legislate to end them. The Court began its careful analysis in Jones by noting the expansive language of § 1982, and observing that a black citizen denied the opportunity to purchase property as a result of discrimination by a private seller cannot be said to have the “same right” to purchase property as a white citizen. 392 U. S., at 420-421. The Court also noted that, in its original form, § 1982 had been part of § 1 of the Civil Rights Act of 1866,1 and that §2 of the 1866 Act provided for criminal penalties against any person who violated rights secured or 1 Act of Apr. 9, 1866, ch. 31, § 1, 14 Stat. 27. Section 1 provided: “[C]itizens, of every race and color, without regard to any previous condition of slavery or involuntary servitude, . . . shall have the same right, in every State and Territory in the United States, to make and enforce contracts, to sue, be parties, and give evidence, to inherit, purchase, lease, sell, hold, and convey real and personal property, and to full and equal benefit of all laws and proceedings for the security of person and property, as is enjoyed by white citizens, and shall be subject to like punishment, pains, and penalties, and to none other, any law, statute, ordinance, regulation, or custom, to the contrary notwithstanding.” All members of the Court agreed in Jones v. Alfred H. Mayer Co., 392 U. S. 409 (1968), that intervening revisions in the property clause of § 1— the reenactment of the 1866 Act in § 18 of the Voting Rights Act of 1870, ch. 114, § 18, 16 Stat. 144, the codification of the property clause in § 1978 of the Revised Statutes of 1874, and its recodification as 42 U. S. C. § 1982—had not altered its substance. Jones, 392 U. S., at 436-437 (opinion of the Court); id., at 453 (dissenting opinion). PATTERSON v. McLEAN CREDIT UNION 193 164 Opinion of Brennan, J. protected by the Act “under color of any law, statute, ordinance, regulation, or custom.” 392 U. S., at 424-426. This explicit limitation upon the scope of § 2, to exclude criminal liability for private violations of § 1, strongly suggested that § 1 itself prohibited private discrimination, for otherwise the limiting language of § 2 would have been redundant. Ibid. Although Justice Harlan, in dissent, thought a better explanation of the language of §2 was that it “was carefully drafted to enforce all of the rights secured by § 1,” id., at 454, it is by no means obvious why the dissent’s view should be regarded as the more accurate interpretation of the structure of the 1866 Act.2 The Court then engaged in a particularly thorough analysis of the legislative history of §1 of the 1866 Act, id., at 422-437, which had been discussed at length in the briefs of both parties and their amici.3 While never doubting that the prime targets of the 1866 Act were the Black Codes, in which the Confederate States imposed severe disabilities on the freedmen in an effort to replicate the effects of slavery, see, e. g., 1 C. Fairman, Reconstruction and Reunion 1864-1888, pp. 110-117 (1971) (discussing Mississippi’s Black Codes), the Court concluded that Congress also had intended § 1 to reach private discriminatory conduct. The Court cited 2 In support of its view, the Court in Jones quoted from an exchange during the House debate on the civil rights bill. When Congressman Loan of Missouri asked the Chairman of the House Judiciary Committee why § 2 had been limited to those who acted under color of law, he was told, not that the statute had no application at all to those who had not acted under color of law, but that the limitation had been imposed because it was not desired to make ‘“a general criminal code for the States.’” Id., at 425, n. 33, quoting Cong. Globe, 39th Cong., 1st Sess., 1120 (1866). Justice Harlan in dissent conceded that the Court’s interpretation of this exchange as supporting a broader reading of § 1 was “a conceivable one.” 392 U. S., at 470. 3 See, e. g., Brief for Petitioners 12-16, Brief for Respondents 7-24, Brief for United States as Amicus Curiae 28-35, 38-51, and Brief for National Committee Against Discrimination in Housing et al. as Amici Curiae 9—39, in Jones v. Alfred H. Mayer Co., 0. T. 1967, No. 45. 194 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. a bill (S. 60) to amend the Freedmen’s Bureau Act, introduced prior to the civil rights bill, and passed by both Houses during the 39th Congress (though it was eventually vetoed by President Johnson), as persuasive evidence that Congress was fully aware that any newly recognized rights of blacks would be as vulnerable to private as to state infringement. 392 U. S., at 423, and n. 30. The amendment would have extended the jurisdiction of the Freedmen’s Bureau over all cases in the former Confederate States involving the denial on account of race of rights to make and enforce contracts or to purchase or lease property, “in consequence of any State or local law, ordinance, police, or other regulation, custom, or prejudice.” Cong. Globe, 39th Cong., 1st Sess., 209 (1866) (emphasis added). When the civil rights bill was subsequently introduced, Representative Bingham specifically linked it in scope to S. 60. Id., at 1292. See Jones, 392 U. S., at 424, n. 31. The Court further noted that there had been “an imposing body of evidence [before Congress] pointing to the mistreatment of Negroes by private individuals and unofficial groups, mistreatment unrelated to any hostile state legislation.” Id., at 427. This evidence included the comprehensive report of Major General Carl Schurz on conditions in the Confederate States. This report stressed that laws were only part of the problem facing the freedmen, who also encountered private discrimination and often brutality.4 The con- 4Report of C. Schurz, S. Exec. Doc. No. 2, 39th Cong., 1st Sess. (1865). The Schurz report is replete with descriptions of private discrimination, relating both to the freedmen’s ability to enter into contracts and to their treatment once under contract. It notes, for example, that some planters had initially endeavored to maintain “the relation of master and slave, partly by concealing from [their slaves] the great changes that had taken place, and partly by terrorizing them into submission to their behests.” Id., at 15. It portrays as commonplace the use of “force and intimidation” to keep former slaves on the plantations: “In many instances negroes who walked away from the plantations, or were found upon the roads, were shot or otherwise severely punished, PATTERSON v. McLEAN CREDIT UNION 195 164 Opinion of Brennan, J. gressional debates on the Freedmen’s Bureau and civil rights bills show that legislators were well aware that the rights of former slaves were as much endangered by private action as by legislation. See id., at 427-428, and nn. 37-40. To be sure, there is much emphasis in the debates on the evils of the Black Codes. But there are also passages that indicate that Congress intended to reach private discrimination that posed an equal threat to the rights of the freedmen. See id., at 429-437. Senator Trumbull, for example, promised to introduce a bill aimed not only at “local legislation,” but also at any “prevailing public sentiment” that blacks in the South “should continue to be oppressed and in fact deprived of their free- which was calculated to produce the impression among those remaining with their masters that an attempt to escape from slavery would result in certain destruction.” Id., at 17. In Georgia, Schurz reported, “the reckless and restless characters of that region had combined to keep the negroes where they belonged,” shooting those caught trying to escape. Id., at 18. The effect of this private violence against those who tried to leave their former masters was that “large numbers [of freedmen], terrified by what they saw and heard, quietly remained under the restraint imposed upon them.” Ibid. See Jones, 392 U. S., at 428-429. It must therefore have been evident to members of the 39th Congress that, quite apart from the Black Codes, the freedmen would not enjoy the same right as whites to contract or to own or lease property so long as private discrimination remained rampant. This broad view of the obstacles to the freedmen’s enjoyment of contract and property rights was similarly expressed in the Howard Report on the operation of the Freedmen’s Bureau, H. R. Exec. Doc. No. 11, 39th Cong., 1st Sess. (1865). It likewise appears in the hearings conducted by the Joint Committee on Reconstruction contemporaneously with Congress’ consideration of the civil rights bill. See Report of the Joint Committee on Reconstruction, 39th Cong., 1st Sess., pts. I-IV (1866). These investigations uncovered numerous incidents of violence aimed at restraining southern blacks’ efforts to exercise their new-won freedom, e. g., id., pt. Ill, p. 143, and whippings aimed simply at making them work harder, or handed out as punishment for a laborer’s transgressions, e. g., id., pt. IV, p. 83, as well, for example, as refusals to pay freedmen more than a fraction of white laborers’ wages, 9-, id., pt. II, pp. 12-13, 54-55, 234. 196 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. dom.” Cong. Globe, 39th Cong., 1st Sess., 77 (1866), quoted in Jones, supra, at 431.5 In the Jones Court’s view, which I share, Congress said enough about the injustice of private discrimination, and the need to end it, to show that it did indeed intend the Civil Rights Act to sweep that far. Because the language of both § 1981 and § 1982 appeared traceable to § 1 of the Civil Rights Act of 1866, the decision in Jones was naturally taken to indicate that § 1981 also prohibited private racial discrimination in the making and enforcement of contracts. Thus, in Tillman v. Wheaton-Haven Recreation Assn., Inc., 410 U. S., at 440, the Court held that “[i]n light of the historical interrelationship between § 1981 and § 1982,” there was no reason to construe those sections differently as they related to a claim that a community swimming club denied property-linked membership preferences to blacks; and in Johnson v. Railway Express Agency, Inc., 421 U. S., at 459-460, the Court stated that “§ 1981 affords a federal remedy against discrimination in private employment on the basis of race.” The Court only addressed the scope of § 1981 in any depth, however, in Runyon v. McCrary, 427 U. S. 160 (1976), where we held that § 1981 prohibited racial discrimination in the admissions policy of a private school. That issue was directly presented and fully briefed in Runyon.* 3 Senator Trumbull was speaking here of his Freedmen’s Bureau bill, which was regarded as having the same scope as his later civil rights bill. See supra, at 193-194. For other statements indicating that § 1 reached private conduct, see Cong. Globe, 39th Cong., 1st Sess., 1118 (1866) (“Laws barbaric and treatment inhuman are the rewards meted out by our white enemies to our colored friends. We should put a stop to this at once and forever”) (Rep. Wilson); id., at 1152 (bill aimed at “the tyrannical acts, the tyrannical restrictions, and the tyrannical laws which belong to the condition of slavery”) (emphasis added) (Rep. Thayer). 6 See, e. g., Brief for Petitioners 2, 6-11, Brief for Respondents 13-22, and Brief for United States as Amicus Curiae 13-18, in Runyon v. Me- PATTERSON v. McLEAN CREDIT UNION 197 164 Opinion of Brennan, J. Although the Court in Runyon treated it as settled by Jones, Tillman, and Johnson that § 1981 prohibited private racial discrimination in contracting, it nevertheless discussed in detail the claim that §1981 is narrower in scope than § 1982. The primary focus of disagreement between the majority in Runyon and Justice White’s dissent, a debate renewed by the parties here on reargument, concerns the origins of § 1981. Section 1 of the 1866 Act was expressly reenacted by § 18 of the Voting Rights Act of 1870. Act of May 31, 1870, ch. 114, §18, 16 Stat. 144. Section 16 of the 1870 Act nevertheless also provided that “all persons within the jurisdiction of the United States shall have the same right in every State and Territory in the United States to make and enforce contracts . . . .” Ibid. Section 1 of the 1866 Act, as reenacted by § 18 of the 1870 Act, was passed under Congress’ Thirteenth Amendment power to identify and legislate against the badges and incidents of slavery, and, we held in Jones, applied to private acts of discrimination. The dissent in Runyon, however, argued that § 16 of the 1870 Act was enacted solely under Congress’ Fourteenth Amendment power to prohibit States from denying any person the equal protection of the laws, and could have had no application to purely private discrimination. See Runyon, supra, at 195-201 (White, J., dissenting). But see District of Columbia v. Carter, 409 U. S. 418, 424, n. 8 (1973) (suggesting Congress has the power to proscribe purely private conduct under § 5 of the Fourteenth Amendment). When all existing federal statutes were codified in the Revised Statutes of 1874, the Statutes included but a single provision prohibiting racial discrimination in the making and enforcement of contracts—§ 1977, which was identical to the current §1981. The Runyon dissenters believed that this provision derived solely from § 16 of the 1870 Act, that the analysis of § 1 in Crary, O. T. 1975, No. 75-62; Brief for Petitioner 17-59, in Fairfax-Brewster School, Inc. v. Gonzales, O. T. 1975, No. 75-66. 198 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. Jones was of no application to § 1981, and that § 1981 hence could not be interpreted to prohibit private discrimination. The Court concluded in Runyon, however—correctly, I believe—that §1977 derived both from §1 of the 1866 Act (as reenacted) and from § 16 of the 1870 Act, and thus was to be interpreted, in light of the decision in Jones, as applying to private conduct. See also General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S., at 390, n. 17 (“[Section] 1981, because it is derived in part from the 1866 Act, has roots in the Thirteenth as well as the Fourteenth Amendment”). This result followed, the Court held, from the terms of the 1874 revision of the statutes. The revisers who prepared the codification had authority only to “revise, simplify, arrange, and consolidate” existing laws, to omit “redundant or obsolete” provisions, and to make suggestions for repeal. Act of June 27, 1866, 14 Stat. 74-75. See Runyon, 427 U. S., at 168, n. 8. The revisers made no recommendation that § 1 of the 1866 Act, as reenacted, be repealed, and obviously the broad 1866 provision, applying to private actors, was not made redundant or obsolete by § 16 of the 1870 Act, with its potentially narrower scope. Hence it is most plausible to think that § 1977 was a consolidation of § 1 and of § 16. Id., at 169, n. 8. The Runyon Court explained that a revisers’ note printed alongside § 1977, indicating that it was derived from § 16, but not mentioning § 1 or its reenactment, had to be viewed in light of the terms of the codification as either inadvertent or an error, and declined “to attribute to Congress an intent to repeal a major piece of Reconstruction legislation on the basis of an unexplained omission from the revisers’ marginal notes.” Ibid.1 Respondent has supplied 7 Congress originally entrusted the revision of the laws to three Commissioners appointed under the Act of June 27, 1866, 14 Stat. 74-75. These Commissioners were instructed to draft sidenotes indicating the source of each section of their revision, § 2, id., at 75, and they wrote the marginal note to what became § 1977 of the Revised Statutes, which referred as a source only to § 16 of the 1870 Act. See 1 Revision of the PATTERSON v. McLEAN CREDIT UNION 199 164 Opinion of Brennan, J. no new information suggesting that the Court’s conclusion as to the dual origins of § 1981 was mistaken.8 In sum, I find the careful analysis in both Jones and Runyon persuasive. United States Statutes as Drafted by the Commissioners Appointed for that Purpose 947 (1872). Congress rejected the work of the Commissioners, however, precisely because Members believed it to contain new legislation. See 2 Cong. Rec. 646 (1874). Congress then appointed Thomas Durant to review the Commissioners’ work. See Act of Mar. 3, 1873, § 3, 17 Stat. 580. “[W]herever the meaning of the law had been changed,” Durant was “to strike out such changes.” 2 Cong. Rec. 646 (1874). Durant reported that he had compared the Commissioners’ revision with preexisting statutes, and that “wherever it has been found that a section contained any departure from the meaning of Congress as expressed in the Statutes at Large, such change has been made as was necessary to restore the original signification.” Report to the Joint Committee on the Revision of the Laws 1 (1873). Durant’s revision, H. R. 1215, 43d Cong., 1st Sess. (1874), which was put before Congress in the form of a bill, see 2 Cong. Rec. 819 (1874), contained no marginal notations. See id., at 826-827, 1210. The Commissioners’ reference to § 16 reappeared only after Congress authorized the Secretary of State to publish the Revised Statutes with marginal notations. See Act of June 20, 1874, ch. 333, § 2, 18 Stat, (part 3) 113. Apparently, the Secretary simply lifted notations from the Commissioners’ draft revision. Hence, insofar as Durant might have thought that the Commissioners had changed the law by referring only to § 16 as their source, and that this problem had been cured merely by the omission of the marginal note from his own draft, it seems strained to rely upon the later decision to restore the Commissioners’ marginal notes as evidence that § 1977 derives solely from § 16. This is particularly so in light of criticism directed in Congress to the accuracy of some of the Commissioners’ side-notes. See 2 Cong. Rec. 828 (1874) (citing as an error a marginal note that was “not sufficently comprehensive” to reflect the provision’s source) (Rep. Lawrence). 81 find strong support for our prior holding that § 1981 is derived in part from the 1866 Act in the legislative history of the 1874 codification. Representative Lawrence, a member of the Joint Committee on the Revision of the Laws, specifically commented in the House upon the proposed revision of the 1866 and 1870 Acts. Id., at 827-828. He noted that the plan of revision was “to collate in one title of ‘civil rights’ the statutes which declare them.” Id., at 827. After setting out § 1 and §2 of the 1866 Act, and then § 16 and § 17 of the 1870 Act, Representative Lawrence stated that the revisers had “very properly not treated [the 1870 Act] as super- 200 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. B Even were there doubts as to the correctness of Runyon, Congress has in effect ratified our interpretation of § 1981, a fact to which the Court pays no attention. We have justified our practice of according special weight to statutory precedents, see ante, at 172-173, by reference to Congress’ ability to correct our interpretations when we have erred. To be sure, the absence of legislative correction is by no means in all cases determinative, for where our prior interpretation of a statute was plainly a mistake, we are reluctant to “‘place on the shoulders of Congress the burden of the Court’s own error.’” Monell v. New York City Dept, of Social Services, 436 U. S. 658, 695 (1978), quoting Girouard v. United States, 328 U. S. 61, 70 (1946). Where our prior interpretation of congressional intent was plausible, however—which is the very least that can be said for our construction of § 1981 in Runyon—we have often taken Congress’ subsequent inaction as probative to varying degrees, depending upon the circumstances, of its acquiescence. See Johnson v. Transportation Agency, Santa Clara County, 480 U. S. 616, 629-630, n. 7 (1987). Given the frequency with which Congress has in recent years acted to overturn this Court’s mistaken interpretations of civil rights statutes,9 its failure to enact legislation seding the entire original act.” Id., at 828. Rather, they had “trans-lat[ed] the sections I have cited from the acts of 1866 and 1870, so far as they relate to a declaration of existing rights,” in the provisions that have now become § 1981 and § 1982. Ibid. There is no hint in this passage that any part of the 1866 Act would be lost in the revision, and indeed in other parts of his statement Representative Lawrence makes it plain that he understood the revisers’ task to be that of presenting “the actual state of the law.” Id., at 826. See also id., at 647-649 (general discussion on the aim of the revision to codify existing law without modification), and id., at 1210 (“[W]e do not purpose to alter the law one jot or tittle”) (Rep. Poland). 9 See, e. g., Civil Rights Attorney’s Fees Awards Act of 1976, Pub. L. 94-559, 90 Stat. 2641, 42 U. S. C. § 1988 (overturning Alyeska Pipeline Service Co. v. Wilderness Society, 421 U. S. 240 (1975)); Pregnancy Discrimination Act, Pub. L. 95-555, 92 Stat. 2076, 42 U. S. C. §2000e(k) PATTERSON v. McLEAN CREDIT UNION 201 164 Opinion of Brennan, J. to overturn Runyon appears at least to some extent indicative of a congressional belief that Runyon was correctly decided. It might likewise be considered significant that no other legislative developments have occurred that cast doubt on our interpretation of § 1981. Cf., e. g., Shearson/American Express Inc. n. McMahon, 482 U. S. 220, 233-234 (1987) (regulatory developments); Moneti, supra, at 697-699; Califano v. Sanders, 430 U. S. 99, 105-107 (1977). There is no cause, though, to consider the precise weight to attach to the fact that Congress has not overturned or otherwise undermined Runyon. For in this case we have more positive signs of Congress’ views. Congress has considered and rejected an amendment that would have rendered § 1981 unavailable in most cases as a remedy for private employment discrimination, which is evidence of congressional acquiescence that is “something other than mere congressional silence and passivity.” Flood v. Kuhn, 407 U. S. 258, 283 (1972). In addition, Congress has built upon our interpretation of § 1981 in enacting a statute that provides for the recovery of attorney’s fees in § 1981 actions. After the Court’s decision in Jones v. Alfred H. Mayer Co., Congress enacted the Equal Employment Opportunity Act of 1972, Pub. L. 92-261, 86 Stat. 103, amending Title VII of the Civil Rights Act of 1964, 42 U. S. C. § 2000e et seq. During Congress’ consideration of this legislation—by which time there had been ample indication that § 1981 was being (overturning General Electric Co. v. Gilbert, 429 U. S. 125 (1976); see Newport News Shipbuilding & Dry Dock Co. v. EEOC, 462 U. S. 669 (1983)); Voting Rights Act Amendments of 1982, Pub. L. 97-205, 96 Stat. 131, 42 U. S. C. § 1973 (overturning Mobile v. Bolden, 446 U. S. 55 (1980); see, e. g., H. R. Rep. No. 97-227, pp. 28-30 (1981)); Handicapped Children’s Protection Act of 1986, Pub. L. 99-372, 100 Stat. 796, 20 U. S. C. §§ 1415(e)(4)(B)-(G) (1982 ed., Supp. V) (overturning Smith v. Robinson, 468 U. S. 992 (1984); see e. g., H. R. Rep. No. 99-296, p. 4 (1985)); Civil Rights Restoration Act of 1987, Pub. L. 100-259, 102 Stat. 28, note following 20 U. S. C. § 1687 (overturning Grove City College v. Bell, 465 U. S. 555 (1984); see, e. g., S. Rep. No. 100-64, p. 2 (1987)). 202 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. interpreted to apply to private acts of employment discrimination10—it was suggested that Title VII rendered redundant the availability of a remedy for employment discrimination under provisions derived from the Civil Rights Act of 1866. Some concluded that Title VII should be made, with limited exceptions, the exclusive remedy for such discrimination. See H. R. Rep. No. 92-238, pp. 66-67 (1971) (minority views). Senator Hruska proposed an amendment to that effect. 118 Cong. Rec. 3172 (1972). Speaking for his amendment, Senator Hruska stated his belief that under existing law private employment discrimination would give rise to a § 1981 claim. He complained specifically that without a provision making Title VII an exclusive remedy, “a black female employee [alleging] a denial of either a promotion or pay raise . . . because of her color,” might “completely bypass” Title VII by filing “a complaint in Federal court under the provisions of the Civil Rights Act of 1866 against . . . the employer.” Id., at 3368, 3369. In speaking against the Hruska amendment, Senator Williams, floor manager of the bill, stated that it was not the purpose of the bill “to repeal existing civil rights laws,” and that to do so “would severely weaken our overall effort to combat the presence of employment discrimination.” Id., at 3371. He referred to §1981 as an existing protection that should not be limited by the amendments to Title VII: “The right of individuals to bring suits in Federal courts to redress individual acts of discrimination, including 10 The Court had remarked in Jones upon the close parallel between § 1981 and § 1982. 392 U. S., at 441, n. 78. Moreover, the lower federal courts already had begun to interpret § 1981 to reach private employment discrimination. See, e. g., Waters v. Wisconsin Steel Works, 427 F. 2d 476 (CA7), cert, denied, 400 U. S. 911 (1970); Sanders v. Dobbs Houses, Inc., 431 F. 2d 1097 (CA5 1970), cert, denied, 401 U. S. 948 (1971); FoW v. International Tel. & Tel. Co., 438 F. 2d 757 (CA3 1971); Caldwell v. National Brewing Co., 443 F. 2d 1044 (CA5 1971), cert, denied, 405 U. S. 916 (1972); Boudreaux v. Baton Rouge Marine Contracting Co., 437 F. 2d 1011 (CA5 1971). PATTERSON v. McLEAN CREDIT UNION 203 164 Opinion of Brennan, J. employment discrimination^] was first provided by the Civil Rights Acts of 1866 and 1871, 42 U. S. C. sections 1981, 1983. It was recently stated by the Supreme Court in the case of Jones v. Mayer, that these acts provide fundamental constitutional guarantees. In any case, the courts have specifically held that title VII and the Civil Rights Acts of 1866 and 1871 are not mutually exclusive, and must be read together to provide alternative means to redress individual grievances. “Mr. President, the amendment of [Senator Hruska] will repeal the first major piece of civil rights legislation in this Nation’s history. We cannot do that. “The peculiarly damaging nature of employment discrimination is such that the individual, who is frequently forced to face a large and powerful employer, should be accorded every protection that the law has in its purview, and that the person should not be forced to seek his remedy in only one place.” Id., at 3371-3372.11 The Hruska amendment failed to win passage on a tied vote, id., at 3373, and the Senate later defeated a motion to reconsider the amendment by a vote of 50 to 37, id., at 3964-3965. Though the House initially adopted a similar amendment, 117 Cong. Rec. 31973, 32111 (1971), it eventually agreed with the Senate that Title VII should not preclude other remedies for employment discrimination, see H. R. Conf. Rep. No. 92-899 (1972). Thus, Congress in 1972 assumed that § 1981 reached private discrimination, and declined to alter its availability as an alternative to those remedies provided by Title VII. The Court in Runyon properly relied upon Congress’ refusal to adopt an amendment that 11 See also 118 Cong. Rec. 3370 (1972) (Sen. Javits) (opposing the Hruska amendment because it would “cut off. . . the possibility of using civil rights acts long antedating the Civil Rights Act of 1964 in a given situation which might fall, because of the statute of limitations or other provisions, in the interstices of the Civil Rights Act of 1964”). 204 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. would have made § 1981 inapplicable to racially discriminatory actions by private employers, and concluded, as I do, that “[t]here could hardly be a clearer indication of congressional agreement with the view that § 1981 does reach private acts of racial discrimination.” 427 U. S., at 174-175 (emphasis in original). Events since our decision in Runyon confirm Congress’ approval of our interpretation of § 1981. In 1976—shortly after the decision in Runyon, and well after the Court had indicated in Tillman and Johnson that § 1981 prohibits private discrimination—Congress reacted to the ruling in Alyeska Pipeline Service Co. v. Wilderness Society, 421 U. S. 240 (1975), that attorney’s fees are not ordinarily recoverable absent statutory authorization, by enacting the Civil Rights Attorney’s Fees Awards Act of 1976, Pub. L. 94-559, 90 Stat. 2641, 42 U. S. C. § 1988. A number of civil rights statutes, like §1981, did not provide for the recovery of attorney’s fees, and Congress heard testimony that the decision in Alyeska Pipeline might have a “devastating impact” on litigation under the civil rights laws. H. R. Rep. No. 94-1558, p. 3 (1976). Congress responded by passing an Act to permit the recovery of attorney’s fees in civil rights cases, including those brought under § 1981. Congress was well aware when it passed the 1976 Act that this Court had interpreted § 1981 to apply to private discrimination. The House Judiciary Committee Report had expressly stated: “Section 1981 is frequently used to challenge employment discrimination based on race or color. Johnson v. Railway Express Agency, Inc., 421 U. S. 454 (1975). Under that section the Supreme Court recently held that whites as well as blacks could bring suit alleging racially discriminatory employment practices. McDonald v. Santa Fe Trail Transportation Co. [, 427 U. S. 273 (1976)]. Section 1981 has also been cited to attack exclusionary admissions policies at recreational facilities. PATTERSON v. McLEAN CREDIT UNION 205 164 Opinion of Brennan, J. Tillman v. Wheaton-Haven Recreation Assn., Inc., 410 U. S. 431 (1973).” Id., at 4 (footnote omitted). The House recognized that § 1981, thus interpreted, overlaps significantly with Title VII, and expressed dissatisfaction that attorney’s fees should be available under the latter, but not the former, statute. See also S. Rep. No. 94-1011, p. 4 (1976) (“[F]ees are now authorized in an employment discrimination suit brought under Title VII of the 1964 Civil Rights Act, but not in the same suit brought under 42 U. S. C. § 1981, which protects similar rights but involves fewer technical prerequisites to the filing of an action”). Congress’ action in providing for attorney’s fees in § 1981 actions, intending that successful § 1981 plaintiffs who could have brought their action under Title VII not be deprived of fees, and knowing that this Court had interpreted § 1981 to apply to private discrimination, goes beyond mere acquiescence in our interpretation of § 1981. Congress approved and even built upon our interpretation. Overruling Runyon would be flatly inconsistent with this expression of congressional intent. See Bob Jones University v. United States, 461 U. S., at 601-602; Patsy v. Board of Regents of Florida, 457 U. S. 496, 501 (1982); Apex Hosiery Co. v. Leader, 310 U. S. 469, 488 (1940). II I turn now to the two issues on which certiorari was originally requested and granted in this case. The first of these is whether a plaintiff may state a cause of action under § 1981 based upon allegations that her employer harassed her because of her race. In my view, she may. The Court reaches a contrary conclusion by conducting an ahistorical analysis that ignores the circumstances and legislative history of § 1981. The Court reasons that Title VII or modern state contract law “more naturally govern[s]” harassment actions, ante, at 177—nowhere acknowledging the anachronism attendant upon the implication that the Reconstruction Congress would have viewed state law, or a federal civil rights 206 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. statute passed nearly a century later, as the primary basis for challenging private discrimination. A The legislative history of § 1981—to which the Court does not advert—makes clear that we must not take an overly narrow view of what it means to have the “same right ... to make and enforce contracts” as white citizens. The very same legislative history that supports our interpretation of § 1981 in Runyon also demonstrates that the 39th Congress intended, in the employment context, to go beyond protecting the freedmen from refusals to contract for their labor and from discriminatory decisions to discharge them. Section 1 of the Civil Rights Act was also designed to protect the freedmen from the imposition of working conditions that evidence an intent on the part of the employer not to contract on non-discriminatory terms. See supra, at 194, and n. 4. Congress realized that, in the former Confederate States, employers were attempting to “adher[e], as to the treatment of the laborers, as much as possible to the traditions of the old system, even where the relations between employers and laborers had been fixed by contract.” Report of C. Schurz, S. Exec. Doc. No. 2, 39th Cong., 1st Sess., p. 19 (1865) (emphasis added). These working conditions included the use of the whip as an incentive to work harder—the commonplace result of an entrenched attitude that “[y]ou cannot make the negro work without physical compulsion,” id., at 16—and the practice of handing out severe and unequal punishment for perceived transgressions. See id., at 20 (“The habit [of corporal punishment] is so inveterate with a great many persons as to render, on the least provocation, the impulse to whip a negro almost irresistible”). Since such “acts of persecution” against employed freedmen, ibid., were one of the 39th Congress’ concerns in enacting the Civil Rights Act, it is clear that in granting the freedmen the “same right ... to make PATTERSON v. McLEAN CREDIT UNION 207 164 Opinion of Brennan, J. and enforce contracts” as white citizens, Congress meant to encompass postcontractual conduct. B The Court holds that § 1981, insofar as it gives an equal right to make a contract, “covers only conduct at the initial formation of the contract.” Ante, at 179; see also ante, at 183. This narrow interpretation is not, as the Court would have us believe, ante, at 176-177, the inevitable result of the statutory grant of an equal right “to make contracts.” On the contrary, the language of § 1981 is quite naturally read as extending to cover postformation conduct that demonstrates that the contract was not really made on equal terms at all. It is indeed clear that the statutory language of § 1981 imposes some limit upon the type of harassment claims that are cognizable under § 1981, for the statute’s prohibition is against discrimination in the making and enforcement of contracts; but the Court mistakes the nature of that limit.12 In my view, harassment is properly actionable under the language of § 1981 mandating that all persons “shall have the same right ... to make . . . contracts ... as is enjoyed by white citizens” if it demonstrates that the employer has in 12 The Court’s overly narrow reading of the language of § 1981 is difficult to square with our interpretation of the equal right protected by § 1982 “to inherit, purchase, lease, sell, hold, and convey real and personal property” not just as covering the rights to acquire and dispose of property, but also the “right... to use property on an equal basis with white citizens,” Memphis v. Greene, 451 U. S. 100, 120 (1981) (emphasis added), and “not to have property interests impaired because of. . . race,” id., at 122 (emphasis added). In Shaare Tefila Congregation v. Cobb, 481 U. S. 615 (1987), we reversed the dismissal of a claim by a Jewish congregation alleging that individuals were liable under § 1982 for spraying racist graffiti on the walls of the congregation’s synagogue. Though our holding in that case was limited to deciding that Jews are a group protected by § 1982, our opinion nowhere hints that the congregation’s vandalism claim might not be cognizable under the statute because it implicated the use of property, and not its acquisition or disposal. 208 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. fact imposed discriminatory terms and hence has not allowed blacks to make a contract on an equal basis. The question in a case in which an employee makes a § 1981 claim alleging racial harassment should be whether the acts constituting harassment were sufficently severe or pervasive as effectively to belie any claim that the contract was entered into in a racially neutral manner. Where a black employee demonstrates that she has worked in conditions substantially different from those enjoyed by similarly situated white employees, and can show the necessary racial animus, a jury may infer that the black employee has not been afforded the same right to make an employment contract as white employees. Obviously, as respondent conceded at oral argument, Tr. of Oral Arg. 30 (Feb. 29, 1987), if an employer offers a black and a white applicant for employment the same written contract, but then tells the black employee that her working conditions will be much worse than those of the white hired for the same job because “there’s a lot of harassment going on in this workplace and you have to agree to that,” it would have to be concluded that the white and black had not enjoyed an equal right to make a contract. I see no relevant distinction between that case and one in which the employer’s different contractual expectations are unspoken, but become clear during the course of employment as the black employee is subjected to substantially harsher conditions than her white co-workers. In neither case can it be said that whites and blacks have had the same right to make an employment contract.13 The Court’s failure to consider such examples, and to explain the abundance of legislative history that con- 131 observe too that a company’s imposition of discriminatory working conditions on black employees will tend to deter other black persons from seeking employment. “[W]hen a person is deterred, because of his race, from even entering negotiations, his equal opportunity to contract is denied as effectively as if he were discouraged by an offer of less favorable terms.” Comment, Developments in the Law—Section 1981, 15 Harv. Civ. Rights-Civ. Lib. L. Rev. 29, 101 (1980). PATTERSON v. McLEAN CREDIT UNION 209 164 Opinion of Brennan, J. founds its claim that § 1981 unambiguously decrees the result it favors, underscore just how untenable is the Court’s position.14 Having reached its decision based upon a supposedly literal reading of § 1981, the Court goes on to suggest that its grudging interpretation of this civil rights statute has the benefit of not undermining Title VII. Ante, at 180-182. It is unclear how the interpretation of § 1981 to reach pervasive post-contractual harassment could be thought in any way to undermine Congress’ intentions as regards Title VII. Congress has rejected an amendment to Title VII that would have rendered § 1981 unavailable as a remedy for employment discrimination, and has explicitly stated that §1981 “protects similar rights [to Title VII] but involves fewer technical prerequisites to the filing of an action,” see supra, at 205; that the Acts “provide alternative means to redress individual grievances,” see supra, at 203; and that an employee who is discriminated against “should be accorded 14 In Meritor Savings Bank v. Vinson, 477 U. S. 57 (1986), we addressed the question whether allegations of discriminatory workplace harassment state a claim under §703 of Title VII, 42 U. S. C. § 2000e-2(a)(l), which prohibits discrimination “with respect to [an employee’s] compensation, terms, conditions, or privileges of employment.” We held that sexual harassment creating a hostile workplace environment may ground an action under Title VII. “[N]ot all workplace conduct that may be described as ‘harassment’ affects a ‘term, condition, or privilege’ of employment within the meaning of Title VII,” however. 477 U. S., at 67. “For sexual harassment to be actionable it must be sufficiently severe or pervasive ‘to alter the conditions of [the victim’s] employment and create an abusive working environment.’” Ibid. Similarly, not all workplace conduct that may be described as racial harassment affects an employee’s right to make contracts free of discrimination. But racial harassment of sufficent severity may impinge upon that right, as explained in the text, and should be actionable under § 1981. Petitioner has never argued that the harassment she allegedly suffered amounted to a breach of an express or implied contract under state law, so this case presents no occasion to consider the United States’ view that such a breach is actionable under § 1981 because it deprives a black employee of the same right to make contracts as a white person. 210 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. every protection that the law has in its purview, and . . . the person should not be forced to seek his remedy in only one place,” ibid. Evidently, Title VII and § 1981 provide independent remedies, and neither statute has a preferred status that is to guide interpretation of the other. The Court, indeed, is forced to concede this fact, admitting that where the statutes overlap “we are not at liberty ‘to infer any positive preference for one over the other.’” Ante, at 181. But the Court then goes on to say that the existence of Title VII “should lessen the temptation for this Court to twist the interpretation of [§1981] to cover the same conduct.” Ibid. This, of course, brings us back to the question of what § 1981, properly interpreted, means. The Court’s lengthy discussion of Title VII adds nothing to an understanding of that issue. The Court’s use of Title VII is not only question begging; it is also misleading. Section 1981 is a statute of general application, extending not just to employment contracts, but to all contracts. Thus we have held that it prohibits a private school from applying a racially discriminatory admissions policy, Runyon, and a community recreational facility from denying membership based on race, Tillman. The lower federal courts have found a broad variety of claims of contractual discrimination cognizable under §1981. E.g., Wyatt v. Security Inn Food & Beverage, Inc., 819 F. 2d 69 (CA4 1987) (discriminatory application of hotel bar’s policy of ejecting persons who do not order drinks); Hall n. Bio-Medical Application, Inc., 671 F. 2d 300 (CAS 1982) (medical facility’s refusal to treat black person potentially cognizable under § 1981); Hall v. Pennsylvania State Police, 570 F. 2d 86 (CA3 1978) (bank policy to offer its services on different terms dependent upon race); Cody v. Union Electric, 518 F. 2d 978 (CA8 1975) (discrimination with regard to the amount of security deposit required to obtain service); Howard Security Services, Inc. v. Johns Hopkins Hospital, 516 F. Supp. 508 (Md. 1981) (racially discriminatory award of contract to PATTERSON v. McLEAN CREDIT UNION 211 164 Opinion of Brennan, J. supply services); Grier v. Specialized Skills, Inc., 326 F. Supp. 856 (WDNC 1971) (discrimination in admissions to barber school); Scott v. Young, 307 F. Supp. 1005 (ED Va. 1969) (discrimination in amusement park admissions policy), aff’d, 421 F. 2d 143 (CA4), cert, denied, 398 U. S. 929 (1970). The Court, however, demonstrates no awareness at all that § 1981 is so much broader in scope than Title VII, instead focusing exclusively upon the claim that its cramped construction of § 1981 “preserve[s] the integrity of Title Vil’s procedures,” ante, at 181, and avoids “[u]nnecessary overlap” that would “upset the delicate balance between employee and employer rights struck by Title VII,” ante, at 182, n. 4. Rights as between an employer and employee simply are not involved in many § 1981 cases, and the Court’s restrictive interpretation of §1981, minimizing the overlap with Title VII, may also have the effect of restricting the availability of § 1981 as a remedy for discrimination in a host of contractual situations to which Title VII does not extend. Even as regards their coverage of employment discrimination, § 1981 and Title VII are quite different. As we have previously noted, “the remedies available under Title VII and under § 1981, although related, and although directed to most of the same ends, are separate, distinct, and independent.” Johnson, 421 U. S., at 461. Perhaps most important, §1981 is not limited in scope to employment discrimination by businesses with 15 or more employees, cf. 42 U. S. C. §2000e(b), and hence may reach the nearly 15% of the workforce not covered by Title VII. See Eisenberg & Schwab, The Importance of Section 1981, 73 Cornell L. Rev. 596, 602 (1988). A § 1981 backpay award may also extend beyond the 2-year limit of Title VII. Johnson, 421 U. S., at 460. Moreover, a § 1981 plaintiff is not limited to recovering backpay: she may also obtain damages, including punitive damages in an appropriate case. Ibid. Other differences between the two statutes include the right to a jury trial under § 1981, but not Title VII; a different statute of limitations in 212 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. § 1981 cases, see Goodman v. Lukens Steel Co., 482 U. S. 656 (1987); and the availability under Title VII, but not § 1981, of administrative machinery designed to provide assistance in investigation and conciliation, see Johnson, supra, at 460.15 The fact that § 1981 provides a remedy for a type of racism that remains a serious social ill broader than that available under Title VII hardly provides a good reason to see it, as the Court seems to, as a disruptive blot on the legal landscape, a provision to be construed as narrowly as possible. C Applying the standards set forth above, I believe the evidence in this case brings petitioner’s harassment claim firmly within the scope of § 1981. Petitioner testified at trial that during her 10 years at McLean she was subjected to racial slurs; given more work than white employees and assigned the most demeaning tasks; passed over for promotion, not informed of promotion opportunities, and not offered training 15 The Court suggests that overlap between § 1981 and Title VII interferes with Title Vil’s mediation and conciliation procedures. Ante, at 180-182, and n. 4. In Johnson v. Railway Express Agency, Inc., 421 U. S., at 461, however, we rejected a suggestion that the need for Title VII procedures to continue unimpeded by collateral litigation required that the timely filing of a discrimination charge with the EEOC toll the limitation period for § 1981: “Conciliation and persuasion through the administrative process . . . often constitute a desirable approach to settlement of disputes based on sensitive and emotional charges of invidious employment discrimination. We recognize, too, that the filing of a lawsuit might tend to deter efforts at conciliation, that a lack of success in the legal action could weaken the [EEOC’s] efforts to induce voluntary compliance, and that suit is privately oriented and narrow, rather than broad, in application, as successful conciliation tends to be. But these are the natural effects of the choice Congress has made available to the claimant by its conferring upon him independent administrative and judicial remedies. The choice is a valuable one. Under some circumstances, the administrative route may be highly preferred over the litigatory; under others, the reverse may be true.” (Emphasis added.) PATTERSON v. McLEAN CREDIT UNION 213 164 Opinion of Brennan, J. for higher level jobs; denied wage increases routinely given other employees; and singled out for scrutiny and criticism. Robert Stevenson, the general manager and later president of McLean, interviewed petitioner for a file clerk position in 1972. At that time he warned her that all those with whom she would be working were white women, and that they probably would not like working with a black. Tr. 1-19. In fact, however, petitioner testified that it was Stevenson and her supervisors who subjected her to racial harassment, rather than her co-workers. For example, petitioner testified that Stevenson told her on a number of occasions that “blacks are known to work slower than whites by nature,” id., at 1-87 to 1-88, 2-80 to 2-81, or, as he put it in one instance, that “some animals [are] faster than other animals.” Id., at 2-83. Stevenson also repeatedly suggested that a white would be able to do petitioner’s job better than she could. Id., at 1-83.16 Despite petitioner’s stated desire to “move up and advance” at McLean to an accounting or secretarial position, id., at 1-22, she testified that she was offered no training for a higher level job during her entire tenure at the credit union. Id., at 1-25. White employees were offered training, id., at 1-93, including a white employee at the same level as petitioner but with less seniority. That less senior white employee was eventually promoted to an intermediate accounting clerk position. Id., at 1-48 to 1-49, 2-114 to 2-115. As with every other promotion opportunity that occurred, petitioner was never informed of the opening. Id., at 1-46, 1-91 to 1-92. During the 10 years petitioner worked for McLean, white persons were repeatedly hired for more se 16 A former manager of data processing for McLean testified that when he recommended a black person for a position as a data processor, Stevenson criticized him, saying that he did not “need any more problems around here,” that he would interview the person, but not hire him, and that he would then “search for additional people who are not black.” Tr. 2-160 to 2-161. 214 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. nior positions, without any notice of these job openings being posted, and without petitioner ever being informed of, let alone interviewed for, any of these opportunities. Id., at 1-93 to 1-97. Petitioner claimed to have received different treatment as to wage increases as well as promotion opportunities. Thus she testified that she had been denied a promised pay raise after her first six months at McLean, though white employees automatically received pay raises after six months. Id., at 1-84 to 1-85. See also id., at 1-60 to 1-65 (denial of merit increase). Petitioner testified at length about allegedly unequal work assignments given by Stevenson and her other supervisors, id., at 1-27 to 1-28,1-30, and detailed the extent of her work assignments. Id., at 1-31, 1-101 to 1-120, 2-18, 2-119 to 2-121. When petitioner complained about her workload, she was given no help with it. Id., at 1-82 to 1-83. In fact, she was given more work and was told she always had the option of quitting. Id., at 1-29. Petitioner claimed that she was also given more demeaning tasks than white employees and was the only clerical worker who was required to dust and to sweep. Id., at 1-31. She was also the only clerical worker whose tasks were not reassigned during a vacation. Whenever white employees went on vacation, their work was reassigned; but petitioner’s work was allowed to accumulate for her return. Id., at 1-37, 1-87. Petitioner further claimed that Stevenson scrutinized her more closely and criticized her more severely than white employees. Stevenson, she testified, would repeatedly stare at her while she was working, although he would not do this to white employees. Id., at 1-38 to 1-39, 1-90 to 1-91. Stevenson also made a point of criticizing the work of white employees in private, or discussing their mistakes at staff meetings without attributing the error to a particular individual. But he would chastise petitioner and the only other black employee publicly at staff meetings. Id., at 1-40, 1-89 to 1-90, 2-72 to 2-73. PATTERSON v. McLEAN CREDIT UNION 215 164 Opinion of Brennan, J. The defense introduced evidence at trial contesting each of these assertions by petitioner. But given the extent and nature of the evidence produced by Patterson, and the importance of credibility determinations in assigning weight to that evidence, the jury may well have concluded that petitioner was subjected to such serious and extensive racial harassment as to have been denied the right to make an employment contract on the same basis as white employees of the credit union.17 Ill I agree that the District Court erred when it instructed the jury as to petitioner’s burden in proving her claim that McLean violated § 1981 by failing to promote her, because she is black, to an intermediate accounting clerk position. The District Court instructed the jury that Patterson had to prove not only that she was denied a promotion because of her race, but also that she was better qualified than the white employee who had allegedly received the promotion. That instruction is inconsistent with the scheme of proof we have carefully designed, in analogous cases, “to bring the litigants and the court expeditiously and fairly to [the] ultimate question” whether the defendant intentionally discriminated against the plaintiff. Texas Dept, of Community Affairs v. Burdine, 450 U. S. 248, 253 (1981). A § 1981 plaintiff must prove purposeful discrimination. General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S., at 391. Where the ultimate issue in a disparatetreatment action is whether the defendant intentionally discriminated against the plaintiff, a well-established framework of proof applies if the plaintiff offers only indirect evidence of discriminatory motive. See McDonnell Douglas 11 The proposed jury instruction quoted by the Court, ante, at 179, is scarcely conclusive as to the nature of Patterson’s harassment claim. Indeed, it is precisely harassment so pervasive as to create a discriminatory work environment that will demonstrate that a black plaintiff has been denied an opportunity to contract on equal terms with white employees. 216 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. Corp. v. Green, 411 U. S. 792 (1973) (Title VII); Dister v. The Continental Group, Inc., 859 F. 2d 1108 (CA2 1988) (discriminatory interference with right to benefits, in violation of §510 of the Employee Retirement Income Security Act of 1974, 29 U. S. C. §1140); Loeb v. Textron, Inc., 600 F. 2d 1003 (CAI 1979) (violation of the Age Discrimination in Employment Act, 29 U. S. C. §621 et seq.). There is no reason why this scheme of proof, carefully structured as a “sensible, orderly way to evaluate the evidence in light of common experience as it bears on the critical question of discrimination,” Furnco Construction Corp. v. Waters, 438 U. S. 567, 577 (1978), should not apply to claims of racial discrimination under § 1981. Indeed, the Court of Appeals held below that “[t]he disparate treatment proof scheme developed for Title VII actions in McDonnell Douglas Corp. v. Green, [supra,] and its progeny, may properly be transposed, as here, to the jury trial of a §1981 claim.” 805 F. 2d 1143, 1147 (CA4 1986). The courts below erred, however, in identifying a § 1981 plaintiff’s burden under that framework. A black plaintiff claiming that an employment decision infringed her § 1981 right to make and enforce contracts on the same terms as white persons has the initial burden of establishing a prima facie case. This burden is not an onerous one. Burdine, supra, at 253. The plaintiff need only prove by a preponderence of the evidence that she applied for an available position for which she was qualified, see supra, at 213-214, that she was rejected, and that the employer either continued to seek applicants for the position, or, as allegedly occurred in this case, filled the position with a white employee, see McDonnell Douglas, supra, at 802; Burdine, supra, at 253. We have required at this stage proof only that a plaintiff was qualified for the position she sought, not proof that she was better qualified than other applicants. See McDonnell Douglas, supra, at 802; Burdine, supra, at 253, n. 6. Proof sufficient to make out a prima facie case raises a presumption that the employer acted for impermissi- PATTERSON v. McLEAN CREDIT UNION 217 164 Opinion of Brennan, J. ble reasons, see Fumco Construction Corp., supra, at 577, which the employer may then rebut by articulating “some legitimate, nondiscriminatory reason for the employee’s rejection,” McDonnell Douglas, supra, at 802. In this case, in addition to attacking petitioner’s claim to have made out a prima facie case, respondent introduced evidence tending to show that if it promoted a white employee over petitioner, it did so because the white employee was better qualified for the job. This evidence rebutted any presumption of discrimination raised by petitioner’s prima facie case. Our cases make it clear, however, that a plaintiff must have the opportunity to introduce evidence to show that the employer’s proffered reasons for its decision were not its true reasons. It is equally well established that this evidence may take a variety of forms. McDonnell Douglas, supra, at 804-805; Fumco Construction Corp., supra, at 578. Though petitioner might have sought to prove that McLean’s claim to have promoted a better qualified applicant was not its true reason by showing she was in fact better qualified than the person promoted, the District Court erred in instructing the jury that to succeed petitioner was required to make that showing. Such an instruction is much too restrictive, cutting off other methods of proving pretext plainly recognized in our cases. We suggested in McDonnell Douglas, for example, that a black plaintiff might be able to prove pretext by showing that the employer has promoted white employees who lack the qualifications the employer relies upon, or by proving the employer’s “general policy and practice with respect to minority employment.” 411 U. S., at 804-805. And, of particular relevance given petitioner’s evidence of racial harassment and her allegation that respondent failed to train her for an accounting position because of her race, we suggested that evidence of the employer’s past treatment of the plaintiff would be relevant to a showing that the employer’s proffered legitimate reason was not its true reason. Id., at 804. There are innumerable dif 218 OCTOBER TERM, 1988 Opinion of Brennan, J. 491 U. S. ferent ways in which a plaintiff seeking to prove intentional discrimination by means of indirect evidence may show that an employer’s stated reason is pretextual and not its real reason. The plaintiff may not be forced to pursue any one of these in particular.18 I therefore agree that petitioner’s promotion discrimination claim must be remanded because of the District Court’s erroneous instruction as to petitioner’s burden. It seems to me, however, that the Court of Appeals was correct when it said that promotion-discrimination claims are cognizable under § 1981 because they “go to the very existence and nature of the employment contract.” 805 F. 2d, at 1145. The Court’s disagreement with this commonsense view, and its statement that “the question whether a promotion claim is actionable under § 1981 depends upon whether the nature of the change in position was such that it involved the opportunity to enter into a new contract with the employer,” ante, at 185, display nicely how it seeks to eliminate with technicalities the protection § 1981 was intended to afford—to limit protection to the form of the contract entered into, and not to extend it, as Congress intended, to the substance of the contract as it is worked out in practice. Under the Court’s view, the employer may deny any number of promotions solely on the basis of race, safe from a § 1981 suit, provided it is careful that promotions do not involve new contracts. 18 The Court of Appeals mistakenly held that the instruction requiring petitioner to prove her superior qualifications was necessary in order to protect the employer’s right to choose among equally well-qualified applicants. As we stated in Texas Dept, of Community Affairs n. Burdine, 450 U. S. 248, 259 (1981): “[T]he employer has discretion to choose among equally qualified candidates, provided the decision is not based upon unlawful criteria.” (Emphasis added.) Where a plaintiff proves that an employer’s purported reasons for a promotion decision were all pretextual, the factfinder may infer that the employer’s decision was not based upon lawful criteria; and, as I point out in the text, there are many ways in which a plaintiff can prove pretext other than by proving her superior qualifications. PATTERSON v. McLEAN CREDIT UNION 219 164 Opinion of Stevens, J. It is admittedly difficult to see how a “promotion”—which would seem to imply different duties and employment terms — could be achieved without a new contract, and it may well be as a result that promotion claims will always be cognizable under § 1981. Nevertheless, the same criticisms I have made of the Court’s decision regarding harassment claims apply here: proof that an employee was not promoted because she is black—while all around white peers are advanced— shows that the black employee has in substance been denied the opportunity to contract on the equal terms that § 1981 guarantees. IV In summary, I would hold that the Court of Appeals erred in deciding that petitioner’s racial harassment claim is not cognizable under § 1981. It likewise erred in holding that petitioner could succeed in her promotion-discrimination claim only by proving that she was better qualified for the position of intermediate accounting clerk than the white employee who was in fact promoted. Justice Stevens, concurring in the judgment in part and dissenting in part. When I first confronted the task of interpreting § 1981, I was persuaded by Justice Cardozo’s admonition that it is wise for the judge to “ day one’s own course of bricks on the secure foundation of the courses laid by others who had gone before him.’” Runyon v. McCrary, 427 U. S. 160, 191 (1976) (concurring opinion) (quoting B. Cardozo, The Nature of the Judicial Process 149 (1921)). The Court had already construed the statutory reference to the right “to make and enforce contracts” as a guarantee of equal opportunity, and not merely a guarantee of equal rights. Today the Court declines its own invitation to tear down that foundation and begin to build a different legal structure on its original text. I agree, of course, that Runyon should not be overruled. I am also persuaded, however, that the meaning that had already been 220 OCTOBER TERM, 1988 Opinion of Stevens, J. 491 U. S. given to “the same right... to make and enforce contracts” that “is enjoyed by white citizens”—the statutory foundation that was preserved in Runyon—encompasses an employee’s right to protection from racial harassment by her employer. In Runyon we held that § 1981 prohibits a private school from excluding qualified children because they are not white citizens. Just as a qualified nonwhite child has a statutory right to equal access to a private school, so does a nonwhite applicant for employment have a statutory right to enter into a personal service contract with a private employer on the same terms as a white citizen. If an employer should place special obstacles in the path of a black job applicant—perhaps by requiring her to confront an openly biased and hostile interviewer—the interference with the statutory right to make contracts to the same extent “as is enjoyed by white citizens” would be plain. Similarly, if the white and the black applicants are offered the same terms of employment with just one exception—that the black employee would be required to work in dark, uncomfortable surroundings, whereas the white employee would be given a well-furnished, two-window office—the discrimination would be covered by the statute. In such a case, the Court would find discrimination in the making of the contract because the disparity surfaced before the contract was made. See ante, at 176-177, 179, 180, 184. Under the Court’s understanding of the statute, the black applicant might recover on one of two theories: She might demonstrate that the employer intended to discourage her from taking the job—which is the equivalent of a “refusal to enter into a contract”—or she might show that the employer actually intended to enter a contract, but “only on discriminatory terms.” Ante, at 177. Under the second of these theories of recovery, however, it is difficult to discern why an employer who makes his intentions known has discriminated in the “making” of a contract, while the employer who conceals his discriminatory intent until after the applicant has ac- PATTERSON v. McLEAN CREDIT UNION 221 164 Opinion of Stevens, J. cepted the job, only later to reveal that black employees are intentionally harassed and insulted, has not. It is also difficult to discern why an employer who does not decide to treat black employees less favorably than white employees until after the contract of employment is first conceived is any less guilty of discriminating in the “making” of a contract. A contract is not just a piece of paper. Just as a single word is the skin of a living thought, so is a contract evidence of a vital, ongoing relationship between human beings. An at-will employee, such as petitioner, is not merely performing an existing contract; she is constantly remaking that contract. Whenever significant new duties are assigned to the employee—whether they better or worsen the relationship—the contract is amended and a new contract is made. Thus, if after the employment relationship is formed, the employer deliberately implements a policy of harassment of black employees, it has imposed a contractual term on them that is not the “same” as the contractual provisions that are “enjoyed by white citizens.” Moreover, whether employed at will or for a fixed term, employees typically strive to achieve a more rewarding relationship with their employers. By requiring black employees to work in a hostile environment, the employer has denied them the same opportunity for advancement that is available to white citizens. A deliberate policy of harassment of black employees who are competing with white citizens is, I submit, manifest discrimination in the making of contracts in the sense in which that concept was interpreted in Runyon v. McCrary, supra. I cannot believe that the decision in that case would have been different if the school had agreed to allow the black students to attend, but subjected them to segregated classes and other racial abuse. Indeed, in Goodman v. Lukens Steel Co., 482 U. S. 656 (1987), we built further on the foundation laid in Runyon. We decided that a union’s “toleration and tacit encouragement of racial harassment” violates §1981. 482 U. S., at 222 OCTOBER TERM, 1988 Opinion of Stevens, J. 491 U. S. 665. Although the Court now explains that the Lukens decision rested on the union’s interference with its members’ right to enforce their collective-bargaining agreement, see ante, at 177-178, 183, when I joined that opinion I thought — and I still think—that the holding rested comfortably on the foundation identified in Runyon. In fact, in the section of the Lukens opinion discussing the substantive claim, the Court did not once use the term “enforce” or otherwise refer to that particular language in the statute. 482 U. S., at 664-669. The Court’s repeated emphasis on the literal language of § 1981 might be appropriate if it were building a new foundation, but it is not a satisfactory method of adding to the existing structure. In the name of logic and coherence, the Court today adds a course of bricks dramatically askew from “the secure foundation of the courses laid by others,” replacing a sense of rational direction and purpose in the law with an aimless confinement to a narrow construction of what it means to “make” a contract. For the foregoing reasons, and for those stated in Parts II-B and II-C of Justice Brennan’s opinion, I respectfully dissent from the conclusion reached in Part III of the Court’s opinion. I also agree with Justice Brennan’s discussion of the promotion claim. DELLMUTH v. MUTH 223 Syllabus DELLMUTH, ACTING SECRETARY OF EDUCATION OF PENNSYLVANIA v. MUTH et al. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT No. 87-1855. Argued February 28, 1989—Decided June 15, 1989 The Education of the Handicapped Act (EHA)—which enacts a comprehensive scheme to assure that handicapped children may achieve a free public education appropriate for their needs—provides, inter alia, that parents may challenge the appropriateness of their child’s “individualized education program” (IEP) in an administrative hearing with subsequent judicial review. Respondent Muth (hereinafter respondent) requested a hearing to contest the local school district’s IEP for his son Alex, who is handicapped within the meaning of the EHA. Before the hearing was convened, respondent enrolled Alex in a private school. Alex’s IEP then was revised and declared appropriate by the hearing examiner, and that decision was affirmed by Pennsylvania’s secretary of education more than one year after the original hearing. While the administrative proceedings were under way, respondent brought suit in the Federal District Court against the school district and the secretary challenging the appropriateness of the IEP and the validity of the administrative proceedings and seeking, among other things, reimbursement for Alex’s private-school tuition and attorney’s fees. The court found that, while the revised IEP was appropriate, procedural flaws had delayed the administrative process and that, since the EHA had abrogated the Commonwealth’s Eleventh Amendment immunity from suit, the school district and the Commonwealth were jointly and severally liable for reimbursement of Alex’s tuition and attorney’s fees. The Court of Appeals affirmed. Held: The EHA does not abrogate the States’ Eleventh Amendment immunity from suit, and, thus, the Amendment bars respondent’s attempt to collect tuition reimbursement from Pennsylvania. Pp. 227-232. (a) Congress may abrogate the States’ immunity only by making its intention “unmistakably clear in the language of the statute.” Atascadero State Hospital v. Scanlon, 473 U. S. 234, 242. Pp. 227-228. (b) Respondent’s nontextual arguments—that abrogation is necessary to meet the EHA’s goals and that amendments to the Rehabilitation Act, though not retroactively applicable to respondent’s suit, evince a previous intention to abrogate immunity from EHA suits — have no bearing on the abrogation analysis since congressional intent must be unmistakably 224 OCTOBER TERM, 1988 Syllabus 491 U. S. clear in the statute’s language. Although nontextual evidence might have some weight under a normal exercise in statutory construction, it is generally irrelevant to a judicial inquiry into whether Congress intended to abrogate the Eleventh Amendment. The argument that application of the Atascadero standard is unfair in this case because Congress could not have foreseen that application is premised on an unrealistic view of the legislative process. It is unlikely that the 94th Congress, taking careful stock of the state of Eleventh Amendment law, would drop coy hints but stop short of making its intention manifest. Pp. 228-230. (c) The EH A provisions relied on by the Court of Appeals—the preamble’s statement of purpose, the 1986 amendments dealing with attorney’s fees, and the authorization for judicial review—do not address abrogation even in oblique terms. The statutory structure—which, unlike the Atascadero statute, makes frequent references to States—lends force only to a permissible inference that States are logical defendants and is not an unequivocal declaration of congressional intent to abrogate. Pp. 231-232. 839 F. 2d 113, reversed and remanded. Kennedy, J., delivered the opinion of the Court, in which Rehnquist, C. J., and White, O’Connor, and Scalia, JJ., joined. Scalia, J., filed a concurring opinion, post, p. 233. Brennan, J., filed a dissenting opinion, in which Marshall, Blackmun, and Stevens, JJ., joined, post, p. 233. Blackmun, J., post, p. 243, and Stevens, J., post, p. 243, filed dissenting opinions. Maria Parisi Vickers, Chief Deputy Attorney General of Pennsylvania, argued the cause for petitioner. With her on the briefs were LeRoy S. Zimmerman, former Attorney General, Ernest D. Preate, Jr., Attorney General, John G. Knorr III, Chief Deputy Attorney General, and Gregory R. Neuhauser, Senior Deputy Attorney General. Martha A. Field argued the cause for respondents and filed a brief for respondent Muth. Joanne D. Sommer filed a brief for respondent Central Bucks School District. * *Lacy H. Thornburg, Attorney General, and Edwin M. Speas, Jr., Special Deputy Attorney General, filed a brief for the State of North Carolina as amicus curiae urging reversal. Minna J. Kotkin, Kathleen A. Sullivan, Herbert Semmel, and Elizabeth Lottman Schneider filed a brief for the American Civil Liberties Union et al. as amici curiae urging affirmance. DELLMUTH v. MUTH 225 223 Opinion of the Court Justice Kennedy delivered the opinion of the Court. The question before us is whether the Education of the Handicapped Act abrogates the States’ Eleventh Amendment immunity from suit in the federal courts. I The Education of the Handicapped Act (EHA), 84 Stat. 175, as amended, 20 U. S. C. § 1400 et seq. (1982 ed. and Supp. V.), enacts a comprehensive scheme to assure that handicapped children may receive a free public education appropriate to their needs. To achieve these ends, the Act mandates certain procedural requirements for participating state and local educational agencies. In particular, the Act guarantees to parents the right to participate in the development of an “individualized education program” (IEP) for their handicapped child, and to challenge the appropriateness of their child’s IEP in an administrative hearing with subsequent judicial review. See 20 U. S. C. § 1415 (1982 ed. and Supp. V); School Committee of Burlington v. Department of Education of Massachusetts, 471 U. S. 359, 361 (1985). Alex Muth, the son of respondent Russell Muth (hereinafter respondent), is a bright child, but one handicapped within the meaning of the EHA by a language learning disability and associated emotional problems. Alex was enrolled in public school in the Central Bucks School District of Pennsylvania from 1980 to 1983. In the summer of 1983, respondent requested a statutory administrative hearing to challenge the district’s IEP for Alex. In September, shortly before the hearing convened, respondent enrolled Alex in a private school for learning disabled children for the coming school year. The hearing examiner found that Alex’s original IEP was inappropriate and made a number of recommendations. Both respondent and the school district then appealed .the decision to the secretary of education, as provided under Pennsylvania law, see 22 Pa. Code § 13.32(24) (1988). The secretary remanded the case to the hearing examiner with in 226 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. structions to the school district to revise Alex’s IEP (1988). After the district did so, the hearing examiner issued a decision declaring the revised IEP appropriate, and the secretary affirmed that decision on October 24, 1984, more than a year after the original due process hearing. While the administrative proceedings were underway, respondent brought this suit in the Eastern District of Pennsylvania against the school district and the state secretary of education, whose successor is petitioner here. As amended, respondent’s complaint alleged that the district’s IEP for Alex was inappropriate and that the Commonwealth’s administrative proceedings had violated the procedural requirements of the EHA in two respects: the assignment of review to the secretary, an allegedly partial officer; and the delays occasioned by the secretary’s remand to the hearing examiner. Respondent requested declaratory and injunctive relief, reimbursement for Alex’s private-school tuition in 1983-1984, and attorney’s fees. The District Court found various procedural infirmities in Pennsylvania’s administrative scheme and entered summary judgment on respondent’s procedural claims. The court held a hearing to resolve the remaining issues in the case and to determine the proper remedy for the procedural violations. The court concluded that, while the district’s proposed IEP for 1983-1984 had been appropriate within the meaning of the EHA, respondent was entitled to reimbursement for Alex’s tuition that year because the procedural flaws had delayed the administrative process. The District Court further determined that the school district and the Commonwealth of Pennsylvania were jointly and severally liable, agreeing with respondent that the EHA abrogated Pennsylvania’s Eleventh Amendment immunity from suit. The court also awarded attorney’s fees, assessed jointly and severally against the school district and the Commonwealth. The United States Court of Appeals for the Third Circuit affirmed. Muth n. Central Bucks School Dist., 839 F. 2d DELLMUTH u MUTH 227 223 Opinion of the Court 113 (1988). Most pertinent for this case, the Court of Appeals agreed with the District Court that the Eleventh Amendment did not bar the reimbursement award against the Commonwealth. The court concluded that “the text of EH A and its legislative history leave no doubt that Congress intended to abrogate the 11th amendment immunity of the states.” Id., at 128. To resolve a conflict among the Circuits, we granted certiorari sub nom. Gilhool v. Muth, 488 U. S. 815 (1988), on the question whether the EHA abrogates the States’ sovereign immunity under the Eleventh Amendment. Compare David D. v. Dartmouth School Committee, 775 F. 2d 411 (CAI 1985) (finding abrogation), with Gary A. v. New Trier High School Dist. No. 203, 796 F. 2d 940 (CA7 1986); Doe v. Maher, 793 F. 2d 1470 (CA9 1986); and Miener v. Missouri, 673 F. 2d 969 (CA8 1982) (finding no abrogation). We now reverse. II We have recognized that Congress, acting in the exercise of its enforcement authority under §5 of the Fourteenth Amendment,1 may abrogate the States’ Eleventh Amendment immunity. Fitzpatrick v. Bitzer, 427 U. S. 445, 456 (1976). We have stressed, however, that abrogation of sovereign immunity upsets “the fundamental constitutional balance between the Federal Government and the States,” Atascadero State Hospital v. Scanlon, 473 U. S. 234, 238 (1985), placing a considerable strain on “‘[t]he principles of federalism that inform Eleventh Amendment doctrine,’” Pennhurst State School and Hospital v. Halderman, 465 U. S. 89, 100 (1984), quoting Hutto v. Finney, 437 U. S. 678, 691 (1978). To temper Congress’ acknowledged powers of 1 Petitioner concedes that the EHA was enacted pursuant to Congress’ authority under § 5 of the Fourteenth Amendment, and that Congress has the power to abrogate the Eleventh Amendment with respect to the Act. Tr. of Oral Arg. 14-15; see Atascadero State Hospital v. Scanlon, 473 U. S. 234, 244-245, n. 4 (1985). We decide the case on these assumptions. 228 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. abrogation with due concern for the Eleventh Amendment’s role as an essential component of our constitutional structure, we have applied a simple but stringent test: “Congress may abrogate the States’ constitutionally secured immunity from suit in federal court only by making its intention unmistakably clear in the language of the statute.” Atascadero, supra, at 242. In concluding that the EHA contains the requisite clear statement of congressional intent, the Court of Appeals rested principally on three textual provisions. The court first cited the Act’s preamble, which states Congress’ finding that “it is in the national interest that the Federal government assist State and local efforts to provide programs to meet the education needs of handicapped children in order to assure equal protection of the law.” 20 U. S. C. § 1400(b)(9). Second, and most important for the Court of Appeals, was the Act’s judicial review provision, which permits parties aggrieved by the administrative process to “bring a civil action . . . in any State court of competent jurisdiction or in a district court of the United States without regard to the amount in controversy.” 20 U. S. C. § 1415(e)(2). Finally, the Court of Appeals pointed to a 1986 amendment to the EHA, which states that the Act’s provision for a reduction of attorney’s fees shall not apply “if the court finds that the State or local educational agency unreasonably protracted the final resolution of the action or proceeding or there was a violation of this section.” 20 U. S. C. § 1415(e)(4)(G) (1982 ed., Supp. V). In the view of the Court of Appeals, this amendment represented an express statement of Congress’ understanding that States can be parties in civil actions brought under the EHA. Respondent supplements these points with some nontextual arguments. Most notably, respondent argues that abrogation is “necessary ... to achieve the EHA’s goals,” Brief for Respondent Muth 37; and that the 1986 amendments to another statute, the Rehabilitation Act, 100 Stat. DELLMUTH v. MUTH 229 223 Opinion of the Court 1845, 42 U. S. C. §2000d-7 (1982 ed., Supp. IV), expressly abrogate state immunity from suits brought under the EHA, Brief for Respondent Muth 30. In connection with the latter argument, respondent recognizes that the Rehabilitation Act amendments expressly apply only to “violations that occur in whole or in part after October 21, 1986.” 42 U. S. C. §2000d-7(b) (1982 ed., Supp. IV). Respondent contends, however, that “[a]lthough the amendment became effective after Muth initially filed suit, . . . the overwhelming support for the amendment shows that it reflects Congress’ intent in originally enacting the EHA [in 1975].” Brief for Respondent Muth 32, n. 48.2 We turn first to respondent’s nontextual arguments, because they are the easier to dismiss. It is far from certain that the EHA cannot function if the States retain immunity, or that the 1986 amendments to the Rehabilitation Act are a useful guide to congressional intent in 1975. Indeed, the language of the 1986 amendments to the Rehabilitation Act appears to cut against respondent. Without intending in any way to prejudge the Rehabilitation Act amendments, we note that a comparison of the language in the amendments with the language of the EHA serves only to underscore the difference in the two statutes, and the absence of any clear statement of abrogation in the EHA. The amendments to the Rehabilitation Act read in pertinent part: “A State shall not be immune under the Eleventh Amendment of the Constitution of the United States from suit in Federal court for a violation of [several enumerated provisions] or the provisions of any other Federal statute prohibiting discrimination by recipients of Respondent also offers us another avenue to affirm the result below, which is to overrule the longstanding holding of Hans v. Louisiana, 134 U. S. 1 (1890), that an unconsenting State is immune from liability for damages in a suit brought in federal court by one of its own citizens. We decline this most recent invitation to overrule our opinion in Hans. 230 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Federal financial assistance.” 42 U. S. C. §2000d-7(a) (1) (1982 ed., Supp. IV). When measured against such explicit consideration of abrogation of the Eleventh Amendment, the EHA’s treatment of the question appears ambiguous at best. More importantly, however, respondent’s contentions are beside the point. Our opinion in Atascadero should have left no doubt that we will conclude Congress intended to abrogate sovereign immunity only if its intention is “unmistakably clear in the language of the statute.” Atascadero, 473 U. S., at 242. Lest Atascadero be thought to contain any ambiguity, we reaffirm today that in this area of the law, evidence of congressional intent must be both unequivocal and textual. Respondent’s evidence is neither. In particular, we reject the approach of the Court of Appeals, according to which, “[w]hile the text of the federal legislation must bear evidence of such an intention, the legislative history may still be used as a resource in determining whether Congress’ intention to lift the bar has been made sufficiently manifest.” 839 F. 2d, at 128. Legislative history generally will be irrelevant to a judicial inquiry into whether Congress intended to abrogate the Eleventh Amendment. If Congress’ intention is “unmistakably clear in the language of the statute,” recourse to legislative history will be unnecessary; if Congress’ intention is not unmistakably clear, recourse to legislative history will be futile, because by definition the rule of Atascadero will not be met. The gist of Justice Brennan’s dissent’s argument appears to be that application of the governing law in Atascadero is unfair in this case. The dissent complains that we “resor[t] to an interpretative standard that Congress could have anticipated only with the aid of a particularly effective crystal ball.” Post, at 241. This complaint appears to be premised on an unrealistic and cynical view of the legislative process. We find it difficult to believe that the 94th Congress, taking careful stock of the state of Eleventh Amendment law, decided it DELLMUTH v. MUTH 231 223 Opinion of the Court would drop coy hints but stop short of making its intention manifest. Rather, the salient point in our view is that it cannot be said with perfect confidence that Congress in fact intended in 1975 to abrogate sovereign immunity, and imperfect confidence will not suffice given the special constitutional concerns in this area. Cf. Johnson v. Robison, 415 U. S. 361, 373-374 (1974) (federal statute will not be construed to preclude judicial review of constitutional challenges absent clear and convincing evidence of congressional intent). We now turn our attention to the proper focus of an inquiry into congressional abrogation of sovereign immunity, the language of the statute. We cannot agree that the textual provisions on which the Court of Appeals relied, or any other provisions of the EHA, demonstrate with unmistakable clarity that Congress intended to abrogate the States’ immunity from suit. The EHA makes no reference whatsoever to either the Eleventh Amendment or the States’ sovereign immunity. Cf. supra, at 228. Nor does any provision cited by the Court of Appeals address abrogation in even oblique terms, much less with the clarity Atascadero requires. The general statement of legislative purpose in the Act’s preamble simply has nothing to do with the States’ sovereign immunity. The 1986 amendment to the EHA deals only with attorney’s fees, and does not alter or speak to what parties are subject to suit. Respondent conceded as much at oral argument, acknowledging that “the 1986 EHA Amendments . . . are not directly relevant [here] because they concerned only attorney’s fees.” Tr. of Oral Arg. 28. Finally, 20 U. S. C. § 1415(e)(2), the centerpiece of the Court of Appeals’ textual analysis, provides judicial review for aggrieved parties, but in no way intimates that the States’ sovereign immunity is abrogated. As we made plain in Atascadero: “A general authorization for suit in federal court is not the kind of unequivocal statutory language sufficient to abrogate the Eleventh Amendment.” 473 U. S., at 246. 232 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. At its core, respondent’s attempt to distinguish this case from Atascadero appears to reduce to the proposition that the EH A “is replete with references to the states,” whereas in “Atascadero . . . the statutory language at issue did not include mention of states.” Brief for Respondent Muth 32-33. We recognize that the EHA’s frequent reference to the States, and its delineation of the States’ important role in securing an appropriate education for handicapped children, make the States, along with local agencies, logical defendants in suits alleging violations of the EHA. This statutory structure lends force to the inference that the States were intended to be subject to damages actions for violations of the EHA. But such a permissible inference, whatever its logical force, would remain just that: a permissible inference. It would not be the unequivocal declaration which, we reaffirm today, is necessary before we will determine that Congress intended to exercise its powers of abrogation. Ill We hold that the statutory language of the EHA does not evince an unmistakably clear intention to abrogate the States’ constitutionally secured immunity from suit.3 The Eleventh Amendment bars respondent’s attempt to collect tuition reimbursement from the Commonwealth of Pennsylvania. The judgment of the Court of Appeals is reversed, and the case is remanded for proceedings consistent with this opinion. It is so ordered. 3 Our grant of certiorari also embraced the question whether the EHA precluded petitioner from hearing administrative appeals. Since we conclude that the Commonwealth is not subject to suit under the EHA, and since the school district did not petition for review of the Court of Appeals decision, we have no occasion to reach this question. After oral argument, respondent filed a motion to remand this suit to the District Court for consolidation with another related action. In light of our disposition today, respondent’s motion is denied. DELLMUTH v. MUTH 233 223 Brennan, J., dissenting Justice Scalia, concurring. I join the opinion of the Court, with the understanding that its reasoning does not preclude congressional elimination of sovereign immunity in statutory text that clearly subjects States to suit for monetary damages, though without explicit reference to state sovereign immunity or the Eleventh Amendment. Justice Brennan, with whom Justice Marshall, Justice Blackmun, and Justice Stevens join, dissenting. I respectfully dissent from the Court’s holding that the Commonwealth of Pennsylvania is immune from suit in the federal courts for violations of the Education of the Handicapped Act (EHA), 20 U. S. C. § 1400 et seq. (1982 ed. and Supp. V). For reasons I have set out elsewhere, see Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468, 509-511 (1987) (Brennan, J., dissenting); Atascadero State Hospital v. Scanlon, 473 U. S. 234, 258-302 (1985) (Brennan, J., dissenting), I would accept respondent Muth’s invitation to overrule Hans v. Louisiana, 134 U. S. 1 (1890), as that case has been interpreted in this Court’s recent decisions. Even if I did not hold that view, I would nevertheless affirm the decision of the Court of Appeals on the ground that Congress in the EHA abrogated state immunity. I Applying the standard method for ascertaining congressional intent, I conclude, with the Court of Appeals, that “[t]he text of EHA and its legislative history leave no doubt that Congress intended to abrogate the 11th amendment immunity of the states.” Muth v. Central Bucks School Dist., 839 F. 2d 113, 128 (CA3 1988). The EHA imposes substantial obligations on the States, as well as on local education authorities, as might be expected in an Act authorizing federal financial aid “to assist States and localities to provide for the education of all handicapped chil 234 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. dren.” 20 U. S. C. § 1400(c). To be eligible for federal assistance, a State must develop a plan for the education of all handicapped children and establish the procedural safeguards mandated in § 1415. §§ 1412(2), (5). It is the state educational agency that is “responsible for assuring that the requirements of [EHA Subchapter II, dealing with federal assistance for education of handicapped children] are carried out and that all educational programs for handicapped children within the State, including all such programs administered by any other State or local agency, [are] under the general supervision of the persons responsible for educational programs for handicapped children in the State educational agency.” § 1412(6). See Smith v. Robinson, 468 U. S. 992, 1010 (1984) (“The responsibility for providing the required education remains on the States”); Board of Education of Hendrick Hudson Central School Dist. n. Rowley, 458 U. S. 176, 182-183 (1982). In accord with this overarching responsibility placed upon the States, the EHA contemplates that in a number of situations where a local education authority cannot or will not provide appropriate educational services to the handicapped, the State will do so directly. See 20 U. S. C. § 1411(c)(4)(A)(ii) (State to assure provision of services where local authority barred from receiving federal funds because it has failed to submit a proper application); § 1414(d) (State “to provide special education and related services directly to handicapped children residing in the area served by [a] local educational agency” that is unable or unwilling to establish or maintain programs, or to be merged with other local agencies to enable it to do so, or that has “handicapped children who can best be served by a regional or State center”). And in any event, where a local education authority would be entitled to less than $7,500 in EHA funding for a fiscal year, the State may not distribute the funds, but must use the funds itself to ensure provision of appropriate services. §§ 1411(c)(4)(A)(i), DELLMUTH v. MUTH 235 223 Brennan, J., dissenting (c)(4)(B). Moreover, a State may choose to administer up to 25 percent of its federal funding itself, rather than distributing these funds to local education authorities, and use such funds to provide direct services to the handicapped. §§ 1411(c)(1), (c)(2). “[T]he EH A confers upon disabled students an enforceable substantive right to public education in participating States and conditions federal financial assistance upon a State’s compliance with the substantive and procedural goals of the Act.” Honig v. Doe, 484 U. S. 305, 310 (1988) (emphasis added; citation omitted). See also Smith v. Robinson, supra, at 1010. Thus, § 1415(e)(2) provides that “any party aggrieved by the findings and decision [made in an administrative process] shall have the right to bring a civil action ... in any State court of competent jurisdiction or in a district court of the United States without regard to the amount in controversy.” This provision makes no distinction between civil actions based upon the type of relief sought and hence plainly contemplates tuition-reimbursement actions. See School Committee of Burlington v. Department of Education of Massachusetts, 471 U. S. 359 (1985). In light of the States’ pervasive role under the EHA, and the clarity with which the statute imposes both procedural and substantive obligations on the States, I have no trouble in inferring from the text of the EHA that “Congress intended that the state should be named as an opposing party, if not the sole party, to [a] proceeding” brought under § 1415(e)(2), whatever remedy is sought, and that Congress thereby abrogated Eleventh Amendment immunity from suit in federal court. David D. v. Dartmouth School Comm., 775 F. 2d 411, 422 (CAI 1985), cert, denied, 475 U. S. 1140 (1986). Indeed, in those situations where a State has elected to provide educational services to the handicapped directly, or where under the EHA it is required to provide direct services, the State would appear 236 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. to be the only proper defendant in a federal action to enforce EH A rights.1 This solely textually based interpretation of the EHA is supported by the statute’s legislative history. Senator Williams, a primary author of the EHA, explained to Congress that, under the Act, “it should be clear that a parent or guardian may present a complaint alleging that a State or local educational agency has refused to provide services to which a child may be entitled or alleging that a State or local educational agency has erroneously classified a child as a handicapped child.” 121 Cong. Rec. 37415 (1975) (emphasis added). In addition, he emphasized that “any party aggrieved by the findings and decision rendered in the due process hearing o[r] the State educational agency review of such hearing shall have the right to bring a civil action with respect to the original complaint,” id., at 37416 (emphasis added), that is, with respect to the administrative complaint, which of course may allege EHA violations by the State.2 The text and legisla 1 Moreover, it is not even clear that in those situations where the State is the only proper defendant, an action could always be brought against the State even in state court; for in Will v. Michigan Dept, of State Police, ante, at 66, the Court seems to suggest that the very same rule of interpretation it applies here to decide whether Eleventh Amendment immunity is abrogated is also to be used to determine whether a federal statute requires a State to allow itself to be sued in state court. See ante, at 76 (Brennan, J., dissenting). If the EHA does not guarantee that the State can be sued somewhere, then our previous statements that the statute provides enforceable rights are a mockery. 2 The view that Congress believed it had abrogated state immunity in the EHA is confirmed by the legislative history of the Handicapped Children’s Protection Act of 1985. Congress complained that “[c]ongressional intent was ignored by the U. S. Supreme Court when ... it handed down its decision in Smith v. Robinson, 468 U. S. 992 (1984),” where the Court held that “ ‘the EHA repealed the availability of sections 504 [of the Rehabilitation Act of 1973] and 1983 [of Title 42] to individuals seeking a free appropriate public education,’ ” so that such litigants could no longer obtain DELLMUTH v. MUTH 237 223 Brennan, J., dissenting tive history of the EH A thus make it unmistakably clear that Congress there intended to abrogate state immunity from suit. II The Court does not seem to disagree with this analysis of actual congressional intent. Even without benefit of reference to the legislative history that confirms the obvious interpretation of the text and makes Congress’ purpose undeniably clear—history spurned by the Court because it has devised in this case a novel rule that “[legislative history generally will be irrelevant to a judicial inquiry into whether Congress intended to abrogate the Eleventh Amendment,” ante, at 230—the Court is able to “recognize that the EHA’s frequent reference to the States, and its delineation of the States’ important role in securing an appropriate education for handicapped children, make the States, along with local agencies, logical defendants in suits alleging violations of the EHA. This statutory structure lends force to the inference that the States were intended to be subject to damages actions for violations of the EHA.” Ante, at 232. Nevertheless, although Congress did intend to abrogate the States’ immunity from suit, the Court refuses to give effect to this intention because it was not, in the Court’s view, “unequivocal and textual.” Ante, at 230. attorney’s fees. H. R. Rep. No. 99-296, p. 4 (1985) (quoting Smith, supra, at 1030 (Brennan, J., dissenting). To correct this error, Congress enacted an amendment, codified at 20 U. S. C. § 1415(e)(4) (1982 ed., Supp. V), providing for the award of attorney’s fees under the EHA. The statement in the House Report on this amendment that “[i]n some actions or proceedings in which the parents or guardian prevail, more than one local or State agency may be named as a respondent,” and that in such cases, “it is expected that the court will apportion the award of attorneys’ fees and other expense based on the relative culpability of the agencies,” H. R. Rep. No. 99-296, at 6, clearly demonstrates a belief that Congress had abrogated Eleventh Amendment immunity in the EHA. 238 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. I dispute the Court’s conclusion that the text of the EHA is equivocal. To my mind, immunity is “unequivocally” textually abrogated when state amenability to suit is the logical inference from the language and structure of the text. Cf. Edelman v. Jordan, 415 U. S. 651, 673 (1974) (a clear declaration of a State’s consent to suit in federal court does not require “‘express language,’” but may be found in “‘overwhelming implications from the text [that] leave no room for any other reasonable construction,’” quoting Murray n. Wilson Distilling Co., 213 U. S. 151, 171 (1909)). The Court reaches the conclusion it does only because it requires more than an unequivocal text. In doing so, the Court is far removed from any real concern to discern a “clear and manifest” statement of congressional intent, Rice v. Santa Fe Elevator Corp., 331 U. S. 218, 230 (1947), which is all that the Court has otherwise looked for when inquiring into the meaning of congressional action, even “[i]n traditionally sensitive areas, such as legislation affecting the federal balance,” United States v. Bass, 404 U. S. 336, 349 (1971). Were the Court in fact concerned with Congress’ intent it could not have adopted the strict drafting regulations it devises today, ruling out resort to legislative history and apparently also barring inferential reasoning from text and structure. The Court’s justification for such a rule is that abrogation of immunity “upsets ‘the fundamental constitutional balance between the Federal Government and the States,’. . . placing considerable strain on ‘[t]he principles of federalism that inform Eleventh Amendment doctrine,’” and that a “stringent test” is necessary “[t]o temper Congress’ acknowledged powers of abrogation with due concern for the Eleventh Amendment’s role as an essential component of our constitutional structure.” Ante, at 227-228. I maintain that the Court makes one very basic error here, for “[t]here simply is no constitutional principle of state sovereign immunity.” Atascadero, 473 U. S., at 259 (Brennan, J., dissenting). But quite apart from that, the Court has never explained DELLMUTH v. MUTH 239 223 Brennan, J., dissenting why it is that the constitutional principle it has created should require a novel approach to ascertaining congressional intent. As I said in Atascadero, “special rules of statutory drafting are not justified (nor are they justifiable) as efforts to determine the genuine intent of Congress; no reason has been advanced why ordinary canons of statutory construction would be inadequate to ascertain the intent of Congress.” Id., at 254. I entirely fail to see, for example, why the “clear and manifest purpose of Congress” to pre-empt under Article VI “the historic police powers of the States,” Rice, supra, at 230, may be found in so many and various ways, while the Court in the Eleventh Amendment context insists on setting up ever-tighter drafting regulations that Congress must have followed (though Congress could not have been aware of such requirements when it acted) in order to abrogate immunity. A genuine concern to identify Congress’ purpose would lead the Court to consider both the logical inferences to be drawn from the text and structure of the EHA, cf. Edelman v. Jordan, supra, at 673, and the statute’s legislative history, see Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279, 283-285 (1973) (examining legislative history in order to determine whether Congress abrogated Eleventh Amendment immunity), in deciding whether Congress intended to subject States to suit in federal court. Though the special and strict drafting regulations the Court has now foisted on Congress are unjustifiable, still worse is the Court’s retroactive application of these new rules. It would be one thing to tell Congress how in the future the Court will measure Congress’ intent. That at least would ensure that Congress and this Court were operating under the same rules at the same time. But it makes no sense whatsoever to test congressional intent using a set of interpretative rules that Congress could not conceivably have foreseen at the time it acted—rules altogether different from, and much more stringent than, those with which Congress, reasonably relying upon this Court’s opinions, believed itself 240 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. to be working. See Atascadero, supra, at 255, n. 7 (Brennan, J., dissenting). The effect of retroactively applying the Court’s peculiar rule will be to override congressional intent to abrogate immunity, though such intent was absolutely clear under principles of statutory interpretation established at the time of enactment. Retroactive application of new drafting regulations in such circumstances is simply unprincipled. Cf. Welch, 483 U. S., at 496 (Scalia, J., concurring in part and concurring in judgment) (where Congress has enacted statutes based on an assumption reasonably derived from our cases, “[e]ven if we were now to find that assumption to have been wrong, we could not, in reason, interpret the statutes as though the assumption never existed”). Congress has already had cause to complain of the Court’s changing its interpretative rules in midcourse. After the Court held in Atascadero that § 504 of the Rehabilitation Act of 1973, 29 U. S. C. §794, contained no “unmistakable language” abrogating Eleventh Amendment immunity, 473 U. S., at 243, Congress in 1986 enacted an amendment to the Act providing: “A State shall not be immune under the Eleventh Amendment of the Constitution of the United States from suit in Federal court for a violation of [enumerated provisions of the Rehabilitation Act] or the provisions of any other Federal statute prohibiting discrimination by recipients of Federal financial assistance.” 42 U. S. C. § 2000d-7(a)(l) (1982 ed., Supp. IV). Congress enacted this provision, the Senate Conference Report tells us, because “[t]he Supreme Court’s decision [in Atascadero] misinterpreted congressional intent. Such a gap in Section 504 coverage was never intended. It would be inequitable for Section 504 to mandate state compliance with its provisions and yet deny litigants the right to enforce their rights in Federal courts when State or State agency actions are in issue.” S. Conf. Rep. No. 99-388, pp. 27-28 (1986). See also 132 Cong. Rec. 28623 (1986) (amendment DELLMUTH v. MUTH 241 223 Brennan, J., dissenting “eliminate[s] the court-made barrier to effectuating congressional intent that the holding in the Atascadero case so unwisely has raised”) (Sen. Cranston, a principal author of § 504 of the 1973 Act). Had the Court followed the usual rules for determining legislative intent, as Congress in 1973 had every reason to expect it would, the Court could not have fallen into this error. See Atascadero, 473 U. S., at 248-252 (Brennan, J., dissenting) (examining the text, structure, and legislative history of §504 to conclude that Congress intended that the States be amenable to suit in federal court). It is perfectly clear that again today the Court ignores Congress’ actual intent to abrogate state immunity—an intent that is even plainer here than in the case of § 504, which lacked the EHA’s frequent reference to the obligations of States—instead resorting to an interpretative standard that Congress could have anticipated only with the aid of a particularly effective crystal ball. When § 1415 was enacted in its present form in 1975, Edelman v. Jordan, 415 U. S. 651 (1974), and Employees v. Missouri Dept, of Public Health and Welfare, supra, established that this Court would consider legislative history and make inferences from text and structure in determining whether Congress intended to abrogate Eleventh Amendment immunity. Indeed, in Quern v. Jordan, 440 U. S. 332, 342-345 (1979), the Court evidently remained of the view that legislative history might be taken into account. Cf. Hutto n. Finney, 437 U. S. 678, 693-694 (1978). And later still, in Pennhurst State School and Hospital v. Halderman, 465 U. S. 89, 99 (1984), the Court still was requiring only “an unequivocal expression of congressional intent,” and citing cases in support—Edelman and Quern—that discuss legislative history in assessing whether Congress intended to abrogate immunity. Obviously, there was no rule in 1975 of the sort the Court has devised in this case, and I fail to understand what theory it is that justifies 242 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. the Court now gauging the 94th Congress’ intent by using such a rule.3 Ill Though I would hold that Pennsylvania is not immune from suit in federal court for breaches of its obligations under the EHA, I find it unnecessary to go on to consider the second question upon which certiorari was granted: whether the Court of Appeals erred in ruling that Pennsylvania’s secretary of education is precluded from deciding special education administrative appeals under § 1415(c) because he is an employee of the Commonwealth. There was an alternative ground for the Court of Appeals’ judgment against Pennsylvania—that because of the secretary’s remand to a hearing officer following respondent’s administrative appeal, respondent was deprived of the timely “final” judgment to which he was entitled under 20 U. S. C. § 1415(e) and 34 CFR § 300.512 (1988). 839 F. 2d, at 124-125. Petitioner’s predecessor did not seek review of the Court of Appeals’ decision on this alternative ground, which appears adequate to support the judgment below, and no purpose would be served by our considering whether the secretary’s participation in the appeal was a violation of the EHA’s procedural requirements. I would thus affirm the judgment below. 31 can only express amazement at the Court’s statement that “a comparison of the language in the [Rehabilitation Act] amendments with the language of the EHA serves only to underscore the difference in the two statutes,” ante, at 229, as if the omission from the EHA of the Rehabilitation Act amendments’ provision that “[a] State shall not be immune under the Eleventh Amendment” actually tells us something about Congress’ intent when it enacted the EHA. The 1986 amendment was a response to Atascadero, tailored to overrule a decision that had misinterpreted Congress’ intent in the Rehabilitation Act of 1973 to abrogate state immunity. If Congress’ reaction to Atascadero tells us anything, it is that Congress prior to that decision believed it could effectively express its intent to abrogate immunity without resorting to the degree of textual clarity the Court demands in this case. DELLMUTH v. MUTH 243 223 Stevens, J., dissenting Justice Blackmun, dissenting. I join Justice Brennan’s opinion because he correctly ascertains the unmistakable intent of Congress to subject state agencies to liability for tuition-reimbursement awards under the Education of the Handicapped Act, 20 U. S. C. § 1415(e)(2). See also School Committee of Burlington v. Department of Education of Massachusetts, 471 U. S. 359 (1985). Indeed, as Justice Brennan convincingly demonstrates, this statute passes even the stringent test set forth in Atascadero State Hospital v. Scanlon, 473 U. S. 234 (1985). It is only by resorting to a stricter standard yet that the Court is able to reach the result that it does here. Because the Court never should have started down this road, it certainly should not take today’s additional step. Justice Stevens, dissenting. While I join Justice Brennan’s dissent, I adhere to my view that a “statute cannot amend the Constitution.” Pennsylvania v. Union Gas Co., ante, at 24 (concurring opinion). Because this case deals with the judicially created doctrine of sovereign immunity rather than the real Eleventh Amendment’s limitation on federal judicial power, the congressional decision to confer jurisdiction on the federal courts must prevail. 244 OCTOBER TERM, 1988 Syllabus 491 U. S. COLONIAL AMERICAN LIFE INSURANCE CO. v. COMMISSIONER OF INTERNAL REVENUE CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT No. 88-396. Argued April 18, 1989—Decided June 15, 1989 Insurance companies commonly enter into reinsurance agreements, whereby the reinsurer pays the primary insurer, or “ceding company,” an up-front fee—a “ceding commission”—and agrees to assume the ceding company’s liabilities on the reinsured policies in return for the future income generated from the policies and their associated reserve accounts. Under an “assumption” reinsurance agreement, the reinsurer steps into the ceding company’s shoes, becoming directly liable to the policyholders and receiving all premiums directly. In contrast, under an indemnity reinsurance agreement, the reinsurer assumes no direct liability, instead reimbursing the ceding company for a specified percentage of the claims and expenses attributable to the risks that have been insured and receiving a like percentage of the premiums generated by the insurance of those risks. In 1975 and 1976, petitioner entered into four indemnity reinsurance agreements on life insurance policies written by Transport Life Insurance Company, the ceding company, agreeing to pay Transport ceding commissions. Sections 801-820 of the Internal Revenue Code—which relate to life insurance companies—do not specifically address the tax treatment of indemnity reinsurance ceding commissions. On its income tax returns for the years in question, petitioner claimed deductions for the full amount of the commissions. Respondent Commissioner of Internal Revenue disallowed the deductions on the ground that the commissions had to be capitalized and amortized over the useful life of the reinsurance agreements, a 7-year period, but the Tax Court reversed. The Court of Appeals, in turn, reversed, holding that ceding commissions are not currently deductible. It reasoned that since they represent payments to acquire an asset within an income producing life that extends substantially beyond one year, the payments must be amortized over the estimated life of the asset. Held: Ceding commissions paid under an indemnity reinsurance agreement must be amortized over the anticipated life of the agreement. Pp. 249-260. (a) Such commissions represent an investment in the future income stream from the reinsured policies and, as such, should be treated in the same manner as commissions involved in assumption reinsurance, which COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 245 244 Syllabus must be capitalized and amortized. See 26 CFR § 1.817-4(d). This analogy is appropriate since none of the differences between indemnity and assumption reinsurance goes to the function and purpose of ceding commissions. Whether the reinsurer assumes direct liability to the policy-holder in no way alters the commissions’ economic role. Less compelling is petitioner’s analogy to agents’ commissions incurred by a life insurance company in issuing directly written insurance, which are currently deductible as ordinary and necessary business expenses under § 809(d)(2). The agent’s commission in a direct insurance setting is an administrative expense—akin to a salary and other sales expenses of writing new policies—to remunerate a third party who helps facilitate the sale, whereas the payment in the reinsurance setting is for the asset sold by the ceding company rather than for services. Even accepting that petitioner’s analogy were true, this would not undermine the basic character of ceding commissions as capital expenditures, but would, at most, prove that Congress decided to carve out an exception for agents’ commissions, notwithstanding their arguable character as capital expenses. This Court will not extend such an exception to other capital expenditures where Congress has not so provided. Pp. 249-253. (b) No Code provision requires that the commissions in question be currently deductible. They are not ordinary and necessary business expenses under § 809(d)(12). Nor does § 818(a)—which requires life insurance companies to compute their taxes in accordance with the accounting procedures of the National Association of Insurance Commissioners (NAIC) for preparing an annual statement, except when such procedures would be inconsistent with accrual accounting rules—authorize current deduction even though NAIC prescribes such treatment of ceding commissions. NAIC’s practice is inconsistent with accrual accounting rules, which require that capital expenditures be amortized. Moreover, petitioner’s reading of § 818(a) is unduly expansive, since it is inconceivable that Congress intended to delegate to the insurance industry the core policy determination whether an expense is a capital outlay or a business expense. Ceding commissions also are not “return premiums, and premiums and other consideration arising out of reinsurance ceded,” which § 809(c)(1) permits a company to exclude from the gross premiums included in its tax base, thereby reducing its taxable income. Such commissions—which are up-front, one-time payments to secure a share in a future income stream and bear no resemblance to premiums — fall well outside § 809(c)(l)’s intended purpose, which is to except from the general definition of premium income a small, residual category of payments that resemble premiums but, because they in fact never really accrued to the company that nominally receives them, do not fairly represent income to the recipient. Petitioner’s reading of § 809(c)(1) is 246 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. highly implausible since it is unlikely that Congress would have subsumed a major deduction within the fine details of its definition of premium income rather than including it with the other deductions discussed in § 809(d). Pp. 253-260. 843 F. 2d 201, affirmed. Kennedy, J., delivered the opinion of the Court, in which Rehnquist, C. J., and Brennan, White, Marshall, and Scalia, JJ., joined. Stevens, J., filed a dissenting opinion, in which Blackmun and O’Connor, JJ., joined, post, p. 260. Carolyn P. Chiechi argued the cause for petitioner. With her on the briefs were Margaret Milner Richardson, Francis M. Gregory, Jr., James V. Heffernan, Gordon 0. Pehrson, Jr., and George R. Abramowitz. Michael R. Dreeben argued the cause for respondent. With him on the brief were Acting Solicitor General Bryson, Acting Assistant Attorney General Knapp, Deputy Solicitor General Wallace, Alan I. Horowitz, Gary R. Allen, David English Carmack, and Nancy G. Morgan.* Justice Kennedy delivered the opinion of the Court. The arcane but financially important question before us is whether ceding commissions paid by a reinsurance company to a direct insurer under a contract for indemnity reinsurance are fully deductible in the year tendered or instead must be amortized over the anticipated life of the reinsurance agreements. I This case involves the workings of the reinsurance industry. In order to spread the risks on policies they have written or to reduce required reserves, insurance companies commonly enter into reinsurance agreements. Under these agreements, the reinsurer pays the primary insurer, or “ceding company,” a negotiated amount and agrees to assume the *William B. Harman, Jr., Jack H. Blaine, John W. Holt, and John T. Adney filed a brief for the American Council of Life Insurance et al. as amici curiae urging reversal. COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 247 244 Opinion of the Court ceding company’s liabilities on the reinsured policies. In return, the reinsurer receives the future income generated from the policies and their associated reserve accounts. Reinsurance comes in two basic types, assumption reinsurance and indemnity reinsurance. In the case of assumption reinsurance, the reinsurer steps into the shoes of the ceding company with respect to the reinsured policy, assuming all its liabilities and its responsibility to maintain required reserves against potential claims. The assumption reinsurer thereafter receives all premiums directly and becomes directly liable to the holders of the policies it has reinsured. In indemnity reinsurance, which is at issue in this case, it is the ceding company that remains directly liable to its policy-holders, and that continues to pay claims and collect premiums. The indemnity reinsurer assumes no direct liability to the policyholders. Instead, it agrees to indemnify, or reimburse, the ceding company for a specified percentage of the claims and expenses attributable to the risks that have been reinsured, and the ceding company turns over to it a like percentage of the premiums generated by the insurance of those risks. Both the assumption and the indemnity reinsurer ordinarily pay an up-front fee, known as a “ceding commission,” to the ceding company.1 The issue in this case is whether ceding commissions for indemnity reinsurance may be deducted by the reinsurer in the year in which they are paid, or whether they must be capitalized over the estimated life of the underlying policies. Petitioner writes and reinsures life, accident, and health insurance. In 1975 and 1976, petitioner entered into four indemnity reinsurance agreements to rein 1 There is a form of indemnity reinsurance known as risk-premium, or yearly-renewable-term, reinsurance that does not involve ceding commissions. Under risk-premium reinsurance, much like a normal insurance policy, the ceding company typically pays an annual premium to the reinsurer in return for which the reinsurer promises to reimburse the ceding company should identified losses arise. 248 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. sure blocks of life insurance policies written by Transport Life Insurance Company, the ceding company. The agreements required petitioner to indemnify Transport for 76.6% of Transport’s liabilities under the block of reinsured policies.2 Petitioner also contracted to pay ceding commissions of $680,000 for the 1975 pair of agreements and $852,000 for the 1976 pair of agreements. In addition, petitioner paid Transport a “finder’s fee” of $13,600 in 1975, which the parties agree is subject to the same tax treatment as the ceding commissions. On its federal income tax returns for 1975 and 1976, petitioner claimed deductions for the full amount of the ceding 2 The parties structured the agreements so as to require the actual transfer of only a small amount of cash. To understand this arrangement, it is necessary to touch on the differences between the two types of coinsurance, which is the most common form of indemnity reinsurance. These two types are conventional coinsurance and modified coinsurance. The two differ in their effect on the reserves that insurance companies are required to maintain against potential liabilities, and which represent essentially an estimate of the present value of future benefits less future premiums. In a conventional coinsurance agreement, the ceding company pays a “reinsurance commission” to the reinsurer in an amount equal to the reserves that the reinsurer must establish to support the liabilities assumed; in a modified coinsurance agreement, the ceding company continues to maintain the reserves and transfers to the reinsurer only the investment income that the reserves generate. Insurance companies frequently pair conventional and modified coinsurance agreements in such a proportion that the ceding commission is roughly equal to the reinsurance commission, with the net effect being that very little money changes hands. So it was in this case: petitioner entered into two modified coinsurance agreements covering 70% of a block of policies, and two conventional coinsurance agreements covering 6.6% of the same block of policies; the total ceding commissions were designed to be roughly equal to the reserves petitioner was required to establish under the conventional agreements, with the result being that petitioner actually paid Transport a total of less than $5,000. The parties elected to treat the modified coinsurance agreements for tax purposes as if they were conventional coinsurance agreements, which the Code then permitted. 26 U. S. C. §820 (1976 ed.). Thus, the difference between modified and conventional coinsurance agreements is, mercifully, of no legal significance in this case. COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 249 244 Opinion of the Court commissions and the finder’s fee. The Commissioner disallowed the deductions, concluding that the ceding commissions and finder’s fee had to be capitalized and amortized over the useful life of the reinsurance agreements, a period later stipulated to be seven years. Petitioner then filed for review in the Tax Court, which agreed with petitioner that the ceding commissions could be deducted in full in the year of payment. The Court of Appeals for the Fifth Circuit reversed, holding that ceding commissions are not currently deductible. 843 F. 2d 201 (1988). The Court of Appeals reasoned that ceding commissions represent payments to acquire an asset with an income producing life that extends substantially beyond one year, and that under fundamental principles of taxation law, such payments must be amortized over the estimated life of the asset. To resolve a conflict in the Courts of Appeals,3 we granted certiorari. 488 U. S. 980 (1988). II This case is initially a battle of analogies. The tax treatment of life insurance companies is prescribed in Part I of Subchapter L of the Internal Revenue Code of 1954, 26 U. S. C. §§801-820 (1970 ed. and Supp. V). Given that these provisions do not specify in explicit terms whether ceding commissions for indemnity reinsurance may be taken as current deductions, the parties each argue that the tax treatment of allegedly analogous payments should be controlling. Petitioner analogizes to the tax treatment of “agents’ commissions and other expenses incurred by a life insurance company in issuing directly-written insurance.” Brief for Petitioner 21. Such expenses of primary insurers are currently deductible under § 809(d)(12) of the Code, which incorporates 3 Compare Prairie States Life Ins. Co. v. United States, 828 F. 2d 1222 (CA8 1987) (requiring capitalization), with Merit Life Ins. Co. v. Commissioner, 853 F. 2d 1435 (CA7 1988) (permitting current deduction). 250 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the allowance for “ordinary and necessary” business expenses under § 162(a).4 Petitioner argues that indemnity reinsurance is in effect a direct insurance agreement between the reinsurer and the ceding company. Parties to an indemnity reinsurance agreement, petitioner points out, stand in the same relation to one another as do the parties to a conventional insurance policy: in return for a premium, the reinsurer agrees to reimburse the ceding company in the event the company becomes liable for certain designated risks. Petitioner reasons that just as a direct insurer may currently deduct the commissions it pays to acquire policies, so should an indemnity reinsurer be able to deduct currently the ceding commissions it expends to acquire business. Respondent counters with an analogy to assumption reinsurance, the ceding commissions for which, it is well established, must be capitalized and amortized. See 26 CFR § 1.817-4(d) (1988). “[T]here is essentially no economic difference,” respondent argues, “between a ceding commission paid in an assumption reinsurance transaction and one paid in an indemnity reinsurance transaction.” Brief for Respondent 19-20. In both cases, according to respondent’s analysis, the ceding commission represents payment for the right to share in the future income stream from the reinsured policies. Id., at 18-19. As the parties’ dispute makes clear, indemnity reinsurance bears some formal and functional similarities to both direct insurance and assumption reinsurance. But the salient comparison is between ceding commissions in indemnity reinsurance and their asserted analogues in the other two forms of insurance. At this level of inquiry, we agree with respondent that the analogy to ceding commissions in assumption re- 4 Although the Code does not explicitly permit primary insurers to deduct agent’s commissions in the year in which they are paid, such deductions have been permitted historically, and Congress has recognized and approved of the historic practice. See S. Rep. No. 291, 86th Cong., 1st Sess., 7, 9 (1959); H. R. Rep. No. 98-432, p. 1428 (1984). COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 251 244 Opinion of the Court insurance is the more compelling one. Although indemnity reinsurance is different from assumption reinsurance in some important ways, none of them go to the function and purpose of the ceding commissions. Whether the reinsurer assumes direct liability to the policyholder in no way alters the economic role that the ceding commissions play in both kinds of transactions. The only rational business explanation for the more than $1,500,000 that petitioner paid in ceding commissions to Transport is that petitioner was investing in the future earnings on the reinsured policies. The ceding commissions thus are not administrative expenses on the order of agents’ commissions in direct insurance; rather, they represent part of the purchase price to acquire the right to a share of future profits. The parallels between ceding commissions in indemnity insurance and agents’ commissions in direct insurance, on the other hand, are chiefly nominal. The commission paid to the insurance agent in a direct insurance setting is an administrative expense to remunerate a third party who helps to facilitate the sale; the agent’s commission is akin to a salary, and to other sales expenses of writing new policies, such as administrative overhead. In the reinsurance setting, by contrast, the ceding company owns the asset it is selling, and the reinsurer pays a substantial “commission” as part of the purchase price to induce the ceding company to part with the asset it has created; the payment, in other words, is for the asset itself rather than for services.5 This point is illus- 6 Petitioner suggests that the ceding commission is designed in part to reimburse the ceding company for the deductible, administrative costs it originally incurred in issuing the policies. Even assuming this is so, however, petitioner’s argument confuses the character of the payment to the taxpayer with its function to the seller. Whether the payment represents a partial reimbursement of deductible expenses to the seller is not pivotal, for as respondent points out, that is often the case with capital assets. See Brief for Respondent 19, n. 11. The important point is not how the purchase price breaks down for the seller but whether the taxpayer is investing in an asset or economic interest with an income-producing life that ex 252 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. trated by a comparison with risk-premium insurance, which is in effect like a direct insurance contract between the reinsurer and the ceding company. In risk-premium reinsurance, the reinsurer does not acquire a future stream of income extending beyond the 1-year term of insurance; rather, in exchange for a premium, it agrees to indemnify the ceding company against liability to its policyholders. Not coincidentally, risk-premium reinsurance agreements typically do not involve the payment of ceding commissions. See n. 1, supra. Finally, even if we were to accept petitioner’s arguments about the resemblances between direct insurance and indemnity reinsurance, it would not undermine the basic character of the ceding commissions at issue here as capital expenditures. Petitioner’s argument at most proves only that Congress decided to carve out an exception for agents’ commissions, notwithstanding their arguable character as capital expenditures. We would not take it upon ourselves to extend that exception to other capital expenditures, notwithstanding firmly established tax principles requiring capitalization, where Congress has not provided for the extension.6 We therefore agree with respondent that the ceding commissions paid in respect of indemnity reinsurance, like those involved in assumption reinsurance, represent an investment in a future income stream. The general tax treatment of this sort of expense is well established. Both the Code and our cases long have recognized that amounts expended to acquire an asset with an income-producing life extending substantially beyond the taxable year of acquisition must be capital- tends substantially beyond the taxable year. For this reason, contrary to Justice Stevens’ suggestion, see post, at 261, n., whether the receipt of the ceding commission creates a capital gain for the ceding company is of no relevance in this case. 6 Likewise, we do not mean to imply that other expenses that do bear a greater resemblance to agents’ commissions would be currently deductible, notwithstanding the strictures of the Code. We confront today only the specific tax treatment of ceding commissions for indemnity reinsurance. COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 253 244 Opinion of the Court ized and amortized over the useful life of the asset. See 26 U. S. C. §263 (1970 ed. and Supp. V.); Commissioner v. Idaho Power Co., 418 U. S. 1, 12 (1974); Woodward v. Commissioner, 397 U. S. 572, 575 (1970); see also Massey Motors, Inc. n. United States, 364 U. S. 92, 104 (1960) (the basic purpose of capitalization rules is to “mak[e] a meaningful allocation of the cost entailed in the use ... of the asset to the periods to which it contributes [income] ”). Our agreement with respondent as to the character of ceding commissions therefore resolves this case, absent some specific statutory provision indicating that ceding commissions for indemnity insurance are an exception to the general rule for which Congress has authorized current deduction. Petitioner offers three possible sources in Subchapter L of such a specific authorization. We consider first petitioner’s contention that the commissions are currently deductible under §809(d)(12) of the Code. That provision authorizes deductions for “ordinary and necessary” business expenses as described in § 162(a). It is § 809(d)(12) upon which direct insurers rely in deducting the commissions paid to their agents. Petitioner argues that there is no distinction in Subchapter L between direct insurance and indemnity reinsurance, and therefore that the allowance for direct insurers applies in the latter context as well. This argument, in other words, is the statutory hook upon which petitioner hangs its general submission that its ceding commissions should receive the same tax treatment as the agents’ commissions paid by direct insurers. Were we to agree with petitioner’s general premise, §809(d)(12) would be a logical source of authority to deduct ceding commissions as ordinary and necessary business expenses. But since we have rejected petitioner’s efforts to analogize ceding commissions to agents’ commissions paid in a direct insurance setting, we necessarily reject its argument that § 809(d)(12) authorizes the deduction petitioner claimed. That section does permit petitioner to deduct ordinary and 254 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. necessary business expenses such as salaries and certain administrative costs, but the ceding commissions at issue in this case do not fall in that category. Petitioner also relies on § 818(a) of the Code, which requires a life insurance company to compute its taxes in a manner consistent with the accounting procedures established by the National Association of Insurance Commissioners (NAIC) for purposes of preparing an annual statement, except when such procedures would be inconsistent with accrual accounting rules. See Commissioner v. Standard Life & Accident Ins. Co., 433 U. S. 148, 158-159 (1977). Petitioner points out that NAIC practices prescribe the current deduction of ceding commissions, and argues that the Code, through § 818(a), incorporates the same prescription. In the first place, we think petitioner’s argument begs the question. Treasury Regulations require accrual taxpayers to amortize the expenses of procuring intangible assets that produce economic benefits extending over more than one year. Thus, § 1.461-l(a)(2) of the Treasury Regulations, 26 CFR § 1.461-1(a)(2) (1988), entitled “Taxpayer using an accrual method,” provides that “any expenditure which results in the creation of an asset having a useful life which extends substantially beyond the close of the taxable year may not be deductible, or may be deductible only in part, for the taxable year in which incurred.”7 Since NAIC practices do not apply where their application would be inconsistent with accrual accounting rules, they are inapposite if a ceding commission is prop- 7 Petitioner suggests that § 1.461-l(a)(2) is not an accrual accounting rule because its prescription applies equally to cash-basis taxpayers. Under petitioner’s argument, NAIC accounting principles would dictate all questions of accounting save in those rare instances where Congress or the Commissioner had promulgated a special rule applicable only to accrualbasis taxpayers. We decline to interpret a statutory provision requiring life insurance companies to compute their taxable income “under an accrual method of accounting,” § 818(a)(1), to prescribe application of the rules of accrual accounting only to the extent that they are inconsistent with the rules of cash-basis accounting. COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 255 244 Opinion of the Court erly characterized as an “expenditure which results in the creation of an asset having a useful life which extends substantially beyond the close of the taxable year.” Petitioner’s contention that § 818(a) justifies the deduction therefore loops back into its general contention that ceding commissions are up-front expenses rather than capital expenditures, a contention which we have rejected. More important, petitioner’s argument rests on an unduly expansive reading of the reference to the NAIC in § 818(a), one that would trump many of the precise and careful substantive sections of the Code. Under petitioner’s interpretation, the fundamental question whether an expense is properly characterized as a capital outlay which has to be amortized or instead as an ordinary business expense subject to immediate deduction would be answered by simple reference to accounting procedures in the industry. It is inconceivable that Congress intended to delegate such a core policy determination to the NAIC. Indeed, under petitioner’s argument, it appears that ceding commissions for assumption reinsurance, no less than those for indemnity reinsurance, should be immediately deductible because NAIC accounting principles appear not to distinguish between the two kinds of ceding commissions. See Patterson, Underwriting Income, in Reinsurance 539 (R. Strain ed. 1980). Yet it is common ground among the parties that ceding commissions for assumption reinsurance must be amortized, regardless of the treatment they are accorded under NAIC accounting. As this point suffices to illustrate, petitioner’s interpretation of § 818(a) proves too much. Petitioner’s remaining statutory argument, based on §809 (c)(1) of the Code, is more difficult to dismiss. As part of their computation of gains from operations, life insurance companies must calculate the gains from several designated categories, including “Premiums.” Section 809(c)(1) provides the somewhat complicated formula governing gains 256 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. from premiums. The provision instructs the company to take into account “[t]he gross amount of premiums and other consideration (including advance premiums, deposits, fees, assessments, and consideration in respect of assuming liabilities under contracts not issued by the taxpayer) on insurance and annuity contracts (including contracts supplementary thereto).” From this amount the taxpayer is then to subtract “return premiums, and premiums and other consideration arising out of reinsurance ceded. Except in the case of amounts of premiums or other consideration returned to another life insurance company in respect of reinsurance ceded, amounts returned where the amount is not fixed in the contract but depends on the experience of the company or the discretion of the management shall not be included in return premiums.” 26 U. S. C. § 809(c)(1) (1970 ed.). The sum of the amounts identified in the first clause of the provision minus the amounts excluded in the second part of the provision represents the gross amount of premium income earned by a life insurance company. This figure is then added to the other sources of income identified in §§ 809(b) and (c), and from that total the life insurance company subtracts any allowable deductions identified in § 809(d). The result represents the company’s net gain or loss from operations, which is the basis of its tax bill. In this way, the items identified in the latter part of § 809(c)(1) which are subtracted from premium income contribute eventually to a reduction in the insurance company’s taxable income. Petitioner contends that § 809(c)(1) allows it to subtract the ceding commissions it pays for indemnity reinsurance from its premium income in either of two ways. The commissions, petitioner argues, qualify under the latter part of § 809(c)(1) both as “return premiums” and as “premiums and COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 257 244 Opinion of the Court other consideration arising out of reinsurance ceded.” In construing the statutory phrase “return premiums,” petitioner relies on the definition of that phrase in the Treasury Regulations. Title 26 CFR § 1.809-4(a)(l)(ii) (1988) provides: “The term ‘return premiums’ means amounts returned or credited which are fixed by contract and do not depend on the experience of the company or the discretion of the management. Thus, such term includes amounts refunded due to policy cancellations or erroneously computed premiums. Furthermore, amounts of premiums or other consideration returned to another life insurance company in respect of reinsurance ceded shall be included in return premiums.” Thus, to compress petitioner’s labyrinthine statutory argument, petitioner should prevail in this case if ceding commissions for indemnity reinsurance are fairly encompassed in either the statutory term “premiums and other consideration arising out of reinsurance ceded” or the regulatory definition “consideration returned to another life insurance company in respect of reinsurance ceded.”8 It cannot be denied that the language on which petitioner relies, taken in isolation, could be read to authorize the tax treatment it seeks. Ceding commissions for indemnity reinsurance might loosely be described as consideration “arising out of” or “in respect of reinsurance ceded.” But when the statutory and regulatory language is parsed more carefully, petitioner’s position becomes dubious, and when the language is read against the background of the statutory structure, it becomes untenable. The difficulty with including ceding commissions within the regulatory definition of “return premiums” is that ceding 8 As petitioner points out, 26 CFR § 1.809-4(a)(l)(iii) (1988) specifies that the term “reinsurance ceded” in § 809(c)(1) includes indemnity reinsurance but not assumption reinsurance. 258 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. commissions are not “returned to” the ceding company at all. The commissions never belong to the ceding company until they are paid over in exchange for the right to share in the future income from the reinsured policies. The term “return premiums” more naturally refers to premiums that the insuring or reinsuring company has been paid and then must remit to the individual policyholder or ceding company, as, for example, pursuant to an experience-rated refund clause, which readjusts the amounts of policy premiums paid over to the ceding company to reflect unanticipated savings. As for the statutory language “premiums and other consideration arising out of reinsurance ceded,” ceding commissions do not find a snug fit within this phrase either. Unlike individual policyholders and, in the case of risk-premium reinsurance, ceding companies, reinsurers do not pay premiums. Therefore, a plausible reading of this language is that it refers only to payments from the ceding company to the reinsurer, as, for example, when the ceding company is simply passing on premiums it has received from a policyholder but is obligated to deliver to a reinsurer under an indemnityreinsurance agreement. The “other consideration” phrase, while admittedly open ended, can be read in quite a sensible way as tagalong language that refers to analogous expenditures of this kind, rather than as a broad catchall provision that encompasses payments of any kind from any party. This reading is supported by a comparison with the identical language in the first portion of § 809(c)(1), which also furnishes possible content to “other consideration.” That phrase would appear to refer only to incidental items such as “advance premiums, deposits, [and] fees” paid, like premiums, to the reinsurer from the ceding company. What this closer reading augurs, a broader examination of the statutory structure confirms: ceding commissions are not at all the kind of payments that Congress sought to permit the taxpayer to exclude from gross premiums in § 809(c)(1). In fact, deduction of ceding commissions has nothing to do COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 259 244 Opinion of the Court with the calculations prescribed by that provision. The purpose of § 809(c)(1) is to ensure that “premium income” is included in a company’s tax base and to specify exactly what is and is not encompassed by that term. The provision begins with a general definition of premium income, which it then fine-tunes in the latter part of the section by excluding certain items that might otherwise be considered to come within the general definition. As the Court of Appeals for the Eighth Circuit has written, the latter part of § 809(c)(1) “serves simply to eliminate from the ‘gross amount of premiums and other consideration’ those portions of premiums received which do not, in the end, ‘belong’ to the company in question, but which must either be returned to the policy-holder or turned over to or shared with another company under an indemnity reinsurance agreement. ” Modem American Life Ins. Co. v. Commissioner, 830 F. 2d 110, 113-114 (1987) (footnote omitted). Thus, we read the latter part of § 809(c)(1) as a fine-tuning mechanism that permits the exclusion from premium income of phantom premiums that might be encompassed within a strict definition of premiums but that in fact never really accrued to the company that nominally receives them. This category might include, for example, experience-rated refunds; or premium payments that have been refunded because of an overcharge or the cancellation of a policy; or premiums that the ceding company has received from policy-holders and must pass on to an indemnity reinsurer. See S. Rep. No. 291, 86th Cong., 1st Sess., 39, 54 (1959). But the ceding commissions that are at issue in this case fall well outside what we take to be the intended purpose of the provision, which is to except from the general provision a small, residual category of payments that resemble premiums but do not fairly represent income to the recipient. There is no need for careful delineation of ceding commissions as apart from the general statutory category of premium income, because ceding commissions never would be thought to come 260 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. within that category in the first place. Unlike the above examples, ceding commissions bear no resemblance to premiums; rather, they are an up-front, one-time payment to secure a share in a future income stream. Finally, we note that petitioner’s reading of § 809(c)(1) is highly implausible in light of the intricate attention to detail displayed throughout Subchapter L. To accept petitioner’s submission, we would have to conclude that Congress subsumed a major deduction within the fine details of its definition of premium income. This would be especially surprising given that § 809(c) in its entirety concerns gross income; deductions are treated in a separate subsection, § 809(d). We find it incredible that Congress, with but a whisper, would have tucked away in the fine points of its definition of premium income a deduction of this magnitude. Ill We have concluded that ceding commissions are costs incurred to acquire an asset with an income-producing life that may extend substantially beyond one year. General tax principles provide that such costs must be amortized and capitalized over the useful life of the asset, and no specific provision in the Code dictates a contrary result. The judgment of the Court of Appeals therefore is Affirmed. Justice Stevens, with whom Justice Blackmun and Justice O’Connor join, dissenting. Charting one’s course through the intricacies of the Internal Revenue Code on the basis of first principles rather than statutory text can be hazardous. Intuitively, the Court concludes that the ceding commission a reinsurer pays to indemnify a direct insurer on its policy risks constitutes the purchase price for a capital asset because it produces a stream of future income. The same intuition should lead to the conclusion that the commission a direct insurer pays to acquire policies that will bring future profits constitutes a COLONIAL AM. LIFE INS. CO. v. COMMISSIONER 261 244 Stevens, J., dissenting capital expenditure. Yet everyone agrees that the latter payment is currently deductible. See ante, at 250.* If the Court had begun its analysis with the text of 26 U. S. C. §809 (1970 ed.) and the regulations promulgated thereunder—instead of waiting until after it had decided the case on its view of first principles to respond to the statutory provision—it might well have recognized the merit in the taxpayer’s position. Section 809(c)(1) distinguishes between assumption and indemnity reinsurance, providing that return premiums “arising out of” an indemnity reinsurance transaction are deductible from gross premiums received. See ante, at 256. The Treasury Regulations thus confirm that while payments made by an assumption reinsurer for purchases of policies must be amortized, Treas. Reg. § 1.817-4(d)(2)(ii)(B), 26 CFR § 1.817-4(d)(2)(ii)(B) (1988), “consideration returned to another life insurance company [by an indemnity reinsurer] in respect of reinsurance ceded” is immediately deductible from the reinsurer’s gross premiums. Treas. Reg. § 1.809-4(a)(l)(ii), 26 CFR § 1.809-4(a)(l)(ii) (1988). There is no warrant in the text for the Court’s rulings that assumption reinsurance and indemnity reinsurance should be treated alike for tax purposes, ante, at 251, and that experience refunds constitute return premiums while ceding commissions do not. See ante, at 257-258. In the context of this transaction, in which the ceding commission was netted against the initial reinsurance premium, the commission quite literally is a sum that the “reinsuring company has been paid and then must remit to the . . . ceding company.” Ante, at 258. In all events, for the reasons stated in full in Judge Will’s opinion for the Court of Appeals for the *Similarly, if a ceding commission constituted a capital asset for the purchaser in an indemnity reinsurance transaction, the receipt of a commission should, at least in some circumstances, create a capital gain for the seller. Yet, as the Commissioner conceded at oral argument, the receipt of the ceding commission is taxable as ordinary income. See Tr. of Oral Arg. 20-21. 262 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. Seventh Circuit in Merit Life Ins. Co. n. Commissioner, 853 F. 2d 1435 (1988), cert, pending, No. 88-955,1 would reverse the judgment of the Court of Appeals. CARELLA v. CALIFORNIA 263 Per Curiam CARELLA v. CALIFORNIA APPEAL FROM THE APPELLATE DEPARTMENT, SUPERIOR COURT OF CALIFORNIA, LOS ANGELES COUNTY No. 87-6997. Argued April 26, 1989—Decided June 15, 1989 Appellant Carella was convicted by a California jury of grand theft for failure to return a rented car. At his trial, the judge instructed the jury that a person “shall be presumed to have embezzled” a vehicle if it is not returned within 5 days of the rental agreement’s expiration date and that “intent to commit theft by fraud is presumed” from failure to return the property within 20 days of demand. The Appellate Department held these presumptions valid, even though the prosecution acknowledged that the instructions imposed conclusive presumptions as to core elements of the crime in violation of the Due Process Clause. Held: The jury instructions violated the Due Process Clause of the Fourteenth Amendment. The instructions were mandatory, since they could have been understood by reasonable jurors to require them to find the presumed facts if the State proved certain predicate facts. Thus, they directly foreclosed independent jury consideration of whether the facts proved established certain elements of the offenses with which Carella was charged and relieved the State of its burden, under In re Winship, 397 U. S. 358, of proving by evidence every essential element of Ca-rella’s crime beyond a reasonable doubt. The determination whether the error was harmless is for the lower court to make in the first instance. Reversed and remanded. Christopher D. Cerf, by appointment of the Court, 488 U. S. 992, argued the cause pro hac vice and filed briefs for appellant. Arnold T. Guminski argued the cause for appellee. With him on the brief were Ira Reiner and Harry B. Sondheim. * Per Curiam. On March 24, 1986, after a jury trial in the Municipal Court of Beverly Hills Judicial District, California, appellant Eu * William T. Stephens filed a brief for the American Rental Association as amicus curiae urging affirmance. 264 OCTOBER TERM, 1988 Per Curiam 491 U. S. gene Carella was convicted of grand theft for failure to return a rented car.1 At his trial, the court adopted the prosecution’s requested instructions applying the statutory presumptions in Cal. Veh. Code Ann. § 10855 (West 1987)2 and Cal. Penal Code Ann. § 484(b) (West 1988).3 Specifically, over Carella’s objection, the court charged the jury as follows: (1) “Presumption Respecting Theft by Fraud: “Intent to commit theft by fraud is presumed if one who has leased or rented the personal property of another pursuant to a written contract fails to return the personal property to its owner within 20 days after the owner has made written demand by certified or registered mail following the expiration of the lease or rental agreement for return of the property so leased or rented.” (2) “Presumption Respecting Embezzlement of a Leased or Rented Vehicle: “Whenever any person who has leased or rented a vehicle wilfully and intentionally fails to return the vehicle to its owner within five days after the lease or rental agreement has expired, that person shall be presumed to have embezzled the vehicle.” App. 15. Karelia was acquitted of the charged violation of Cal. Veh. Code Ann. § 10851(a) (West 1987), which provides that the nonconsensual taking or driving of a vehicle is a “public offense” if accomplished with the specific “intent either to permanently or temporarily” deprive the owner of title or possession. 2 California Veh. Code Ann. § 10855 reads: “Whenever any person who has leased or rented a vehicle wilfully and intentionally fails to return the vehicle to its owner within five days after the lease or rental agreement has expired, that person shall be presumed to have embezzled the vehicle.” 3 California Penal Code Ann. § 484(b) reads: “Except as provided in Section 10855 of the Vehicle Code, intent to commit theft by fraud is presumed if one who has leased or rented the personal property of another pursuant to a written contract fails to return the personal property to its owner within 20 days after the owner has made written demand by certified or registered mail following the expiration of the lease or rental agreement for return of the property so leased or rented.” CABELLA v. CALIFORNIA 265 263 Per Curiam On appeal to the Appellate Department of the Superior Court, the prosecution confessed error, acknowledging that these two instructions unconstitutionally imposed conclusive presumptions as to core elements of Carella’s crime. The Appellate Department disagreed, however, and validated the presumptions on the ground that Carella “never offered testimony concerning the nonexistence of the presumed facts....” Id., at 61. This disposition was so plainly at odds with prior decisions of this Court that we noted probable jurisdiction, 488 U. S. 964 (1988), and now reverse. The Due Process Clause of the Fourteenth Amendment denies States the power to deprive the accused of liberty unless the prosecution proves beyond a reasonable doubt every element of the charged offense. In re Winship, 397 U. S. 358, 364 (1970). Jury instructions relieving States of this burden violate a defendant’s due process rights. See Francis v. Franklin, 471 U. S. 307 (1985); Sandstrom v. Montana, 442 U. S. 510 (1979). Such directions subvert the presumption of innocence accorded to accused persons and also invade the truth-finding task assigned solely to juries in criminal cases. We explained in Francis and Sandstrom that courts should ask whether the presumption in question is mandatory, that is, whether the specific instruction, both alone and in the context of the overall charge, could have been understood by reasonable jurors to require them to find the presumed fact if the State proves certain predicate facts. See Sandstrom, supra, at 514. The prosecution understandably does not now dispute that the instructions in this case were phrased as commands, for those instructions were explicit and unqualified to that effect and were not explained elsewhere in the jury charge to be merely permissive. Carella’s jury was told first that a person “shall be presumed to have embezzled” a vehicle if it is not returned within 5 days of the expiration of the rental agreement; and second, that “intent to commit theft by fraud is presumed” from failure to return rented property within 20 days of demand. 266 OCTOBER TERM, 1988 Per Curiam 491 U. S. These mandatory directions directly foreclosed independent jury consideration of whether the facts proved established certain elements of the offenses with which Carella was charged. The instructions also relieved the State of its burden of proof articulated in Winship, namely, proving by evidence every essential element of Carella’s crime beyond a reasonable doubt. The two instructions violated the Fourteenth Amendment. The State insists that the error was in any event harmless. As we have in similar cases, we do not decide that issue here. In Sandstrom v. Montana, supra, at 515, the jury in a murder case was instructed that the “law presumes that a person intends the ordinary consequences of his voluntary acts.” We held that, because the jury might have understood the presumption to be conclusive or as shifting the burden of persuasion, the instruction was constitutional error. There was a claim of harmless error, however, and even though the jury might have considered the presumption to be conclusive, we remanded for the state court to consider the issue if it so chose. In Rose n. Clark, 478 U. S. 570 (1986), we again said that a Sandstrom error is subject to the harmless-error rule. “Nor is Sandstrom error equivalent to a directed verdict for the State. When a jury is instructed to presume malice from predicate facts, it still must find the existence of those facts beyond a reasonable doubt. Connecticut v. Johnson, 460 U. S. 73, 96-97 (1983) (Powell, J., dissenting). In many cases, the predicate facts conclusively establish intent, so that no rational jury could find that the defendant committed the relevant criminal act but did not intend to cause injury. ... In that event the erroneous instruction is simply superfluous: the jury has found, in Winship’s words, ‘every fact necessary’ to establish every element of the offense beyond a reasonable doubt.” Rose, supra, at 580-581 (footnote and citations omitted). We also observed that although we have the authority to make the harmless-error determination CABELLA v. CALIFORNIA 267 263 Scalia, J., concurring in judgment ourselves, we do not ordinarily do so. Hence, we remanded the case for the lower court to make that determination in the first instance. We follow the same course here and reverse the judgment of the California court without deciding here whether no rational jury could find the predicate acts but fail to find the fact presumed. 478 U. S., at 580-581. Accordingly, the judgment of the Appellate Department is reversed, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered. Justice Scalia, with whom Justice Brennan, Justice Marshall, and Justice Blackmun join, concurring in the judgment. I agree with the Court that the decision below must be reversed, and that it is sensible to permit the state court to conduct harmless-error analysis in the first instance. I write separately, however, because the Court has only implicitly acknowledged (by quoting the passage that it does from Rose v. Clark, 478 U. S. 570, 580-581 (1986), see ante, at 266) what should be made explicit—that the harmless-error analysis applicable in assessing a mandatory conclusive presumption is wholly unlike the typical form of such analysis. In the usual case the harmlessness determination requires consideration of “the trial record as a whole,” United States v. Hasting, 461 U. S. 499, 509 (1983), in order to decide whether the fact supported by improperly admitted evidence was in any event overwhelmingly established by other evidence, see, e. g., Milton v. Wainwright, 407 U. S. 371, 372-373 (1972); Harrington n. California, 395 U. S. 250, 254 (1969). Such an expansive inquiry would be error here, and I think it important both to explain why and to describe the mode of analysis that is appropriate. The Court’s mere citation of Rose is inadequate to those ends, since, for reasons I shall describe, infra, at 271-272, that case itself is ambiguous. 268 OCTOBER TERM, 1988 Scalia, J., concurring in judgment 491 U. S. The Court has disapproved the use of mandatory conclusive presumptions not merely because it “ ‘conflicts] with the overriding presumption of innocence with which the law endows the accused,’” Sandstrom v. Montana, 442 U. S. 510, 523 (1979) (quoting Morissette v. United States, 342 U. S. 246, 275 (1952)), but also because it “‘invade[s] [the] fact-finding function’ which in a criminal case the law assigns solely to the jury,” 442 U. S., at 523 (quoting United States v. United States Gypsum Co., 438 U. S. 422, 446 (1978)). The constitutional right to a jury trial embodies “a profound judgment about the way in which law should be enforced and justice administered.” Duncan v. Louisiana, 391 U. S. 145, 155 (1968). It is a structural guarantee that “reflect[s] a fundamental decision about the exercise of official power—a reluctance to entrust plenary powers over the life and liberty of the citizen to one judge or to a group of judges.” Id., at 156. A defendant may assuredly insist upon observance of this guarantee even when the evidence against him is so overwhelming as to establish guilt beyond a reasonable doubt. That is why the Court has found it constitutionally impermissible for a judge to direct a verdict for the State. See United States v. Martin Linen Supply Co., 430 U. S. 564, 572-573 (1977). That is also why in Carpenters v. United States, 330 U. S. 395 (1947), the Court did not treat as harmless a jury instruction that mistakenly did not require express authorization or ratification to hold a union criminally liable for its officers’ participation in an antitrust conspiracy—regardless of how overwhelming the evidence that authorization or ratification in fact existed. We said: “No matter how strong the evidence may be of an association’s or organization’s participation through its agents in the conspiracy, there must be a charge to the jury setting out correctly the limited liability under [the Norris-LaGuardia Act, 47 Stat. 70,] of such association or organization for acts of its agents. For a judge may not direct a verdict of guilty no matter how conclusive the CARELLA v. CALIFORNIA 269 263 Scalia, J., concurring in judgment evidence. There is no way of knowing here whether the jury’s verdict was based on facts within the condemned instructions ... or on actual authorization or ratification of such acts . . . .” Id., at 408-409 (footnotes omitted). In other words, “the question is not whether guilt may be spelt out of a record, but whether guilt has been found by a jury according to the procedure and standards appropriate for criminal trials.” Bollenbach v. United States, 326 U. S. 607, 614 (1946). “Findings made by a judge cannot cure deficiencies in the jury’s findings as to the guilt or innocence of a defendant resulting from the court’s failure to instruct it to find an element of the crime.” Cabana v. Bullock, 474 U. S. 376, 384-385 (1986). These principles necessarily circumscribe the availability of harmless-error analysis when a jury has been instructed to apply a conclusive presumption. If the judge in the present case had instructed the jury, “You are to apply a conclusive presumption that Carella embezzled the rental car if you find that he has blue eyes and lives in the United States,” it would not matter, for purposes of assuring Carella his jury-trial right, whether the record contained overwhelming evidence that he in fact embezzled the car. For nothing in the instruction would have directed the jury, or even permitted it, to consider and apply that evidence in reaching its verdict. And the problem would not be cured by an appellate court’s determination that the record evidence unmistakably established guilt, for that would represent a finding of fact by judges, not by a jury. As with a directed verdict, “the error in such a case is that the wrong entity judged the defendant guilty.” Rose v. Clark, supra, at 578. Four Members of the Court concluded as much in Connecticut v. Johnson, 460 U. S. 73 (1983) (plurality opinion), which considered whether it could be harmless error to instruct a jury that “every person is conclusively presumed to intend the natural and necessary consequences of his act.” Id., at 78. Justice Blackmun wrote for the plurality: 270 OCTOBER TERM, 1988 Scalia, J., concurring in judgment 491 U. S. “An erroneous presumption on a disputed element of a crime renders irrelevant the evidence on the issue because the jury may have relied upon the presumption rather than upon that evidence. If the jury may have failed to consider evidence of intent, a reviewing court cannot hold that the error did not contribute to the verdict. The fact that the reviewing court may view the evidence of intent as overwhelming is then simply irrelevant. To allow a reviewing court to perform the jury’s function of evaluating the evidence of intent, when the jury never may have performed that function, would give too much weight to society’s interest in punishing the guilty and too little weight to the method by which decisions of guilt are to be made.” Id., at 85-86 (footnotes omitted). The plurality therefore determined—I think correctly—that the use of conclusive presumptions could be harmless error only in those “rare situations” when “the reviewing court can be confident that [such an] error did not play any role in the jury’s verdict.” Id., at 87. The opinion mentioned as among those “rare situations” an instruction establishing a conclusive presumption on a charge of which the defendant was acquitted (and not affecting other charges), and an instruction establishing a conclusive presumption with regard to an element of the crime that the defendant in any case admitted. Ibid. Another basis for finding a conclusive-presumption instruction harmless explains our holding two Terms ago in Pope n. Illinois, 481 U. S. 497 (1987). Although the error in instruction held to be harmless there was not a conclusive presumption but rather misdescription of an element of the offense, the latter like the former deprives the jury of its factfinding role, and must be analyzed similarly. (Thus, as noted earlier, misdescription of an element of the offense has similarly been held not curable by overwhelming record evidence of guilt. See Carpenters n. United States, supra, at 408-409.) CARELLA % CALIFORNIA 271 263 Scalia, J., concurring in judgment In both convictions at issue in Pope the juries had been instructed to apply a “community standard]” in deciding whether allegedly obscene magazines, “‘taken as a whole, lac[k] serious literary, artistic, political, or scientific value.’” 481 U. S., at 498-499 (citation omitted). The Court concluded, however, that the First Amendment required a different finding: “whether a reasonable person would find such value in the material, taken as a whole.” Id., at 501. Even though the juries were not instructed to make the precise finding necessary to convict the defendants, the Court held that the error was harmless. I joined that opinion only because I believed that no rational juror could plausibly have found the magazines utterly lacking in value under a community standard and come to a different conclusion under a reasonable person standard. See id., at 504 (Scalia, J., concurring). In an appropriate case, a similar analysis could lead to the conclusion of harmless error for a conclusive presumption: When the predicate facts relied upon in the instruction, or other facts necessarily found by the jury, are so closely related to the ultimate fact to be presumed that no rational jury could find those facts without also finding that ultimate fact, making those findings is functionally equivalent to finding the element required to be presumed. The error is harmless because it is “beyond a reasonable doubt,” Chapman v. California, 386 U. S. 18, 24 (1967), that the jury found the facts necessary to support the conviction. The Court’s opinion does not discuss any of this precedent, but relies exclusively upon citation of, and quotation from, Rose v. Clark.* See ante, at 266-267. In that case we ac * Sandstrom v. Montana, 442 U. S. 510 (1979), is also cited, see ante, at 266, but only (or only properly) for the proposition that we need not conduct harmless-error analysis ourselves, not for the proposition that harmless-error analysis is applicable. In Sandstrom we “decline[d] to reach” not only the State’s claim that the flawed instruction “constituted harmless error,” but also the defendant’s claim that “in any event an unconstitutional jury instruction on an element of the crime can never constitute harmless error.” 442 U. S., at 526-527. 272 OCTOBER TERM, 1988 Scalia, J., concurring in judgment 491 U. S. knowledged the possibility of harmless error (and remanded for determination of that issue) with respect to an instruction that said: “[I]f the State has proven beyond a reasonable . . . doubt that a killing has occurred, then it is presumed that the killing was done maliciously. But this presumption may be rebutted . . . .” 478 U. S., at 574. In explaining why the use of an impermissible presumption, unlike the granting of a directed verdict for the State, can in some circumstances be deemed harmless error, we observed: “When a jury is instructed to presume malice from predicate facts, it still must find the existence of those facts beyond a reasonable doubt. ... In many cases, the predicate facts conclusively establish intent so that no rational jury could find that the defendant committed the relevant criminal act but did not intend to cause injury. . . . In that event . . . the jury has found . . . ‘every fact necessary’ to establish every element of the offense beyond a reasonable doubt.” Id., at 580-581 (emphasis in original). That passage suggests the mode of analysis just discussed in connection with Pope. Were that all which Rose contained on the subject, or were the Court willing to make explicit that the more usual harmless-error analysis does not apply, today’s opinion could be regarded as terse but not misleading. Elsewhere, however, Rose says that usual harmless-error analysis is applicable: “Where a reviewing court can find that the record developed at trial establishes guilt beyond a reasonable doubt, the interest in fairness has been satisfied and the judgment should be affirmed,” 478 U. S., at 579; see id., at 583. I therefore think it at best misleading to suggest without qualification that Rose governs here. Even if Rose’s more expansive description of the sort of harmless-error analysis available is accepted with regard to the type of presumption at issue in that case—a rebuttable presumption—it need not (and for the reasons discussed above cannot) be accepted for conclusive presumptions such CABELLA v. CALIFORNIA 273 263 Scalia, J., concurring in judgment as that in the present case. The Rose jury, instructed regarding a rebuttable presumption of malice, could—indeed, was compelled to—weigh the relevant evidence and decide whether the presumption had been overcome. It had made a finding regarding the elemental fact, and the only difficulty was that the burden of proof had been placed upon the defendant rather than the State. It is one thing to say that the effect of this erroneous burden shifting will be disregarded if “the record developed at trial establishes guilt beyond a reasonable doubt”; it is quite another to say that the jury’s failure to make any factual determination of the elemental fact—because of a conclusive presumption resting upon findings that do not establish beyond a reasonable doubt the elemental fact—will be similarly disregarded. For these reasons, I concur only in the judgment of the Court. 274 OCTOBER TERM, 1988 Syllabus 491 U. S. MISSOURI et al. v. JENKINS, by her friend, AGYEI, ET AL. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE EIGHTH CIRCUIT No. 88-64. Argued February 21, 1989—Decided June 19, 1989 In this major school desegregation litigation in Kansas City, Missouri, in which various desegregation remedies were granted against the State of Missouri and other defendants, the plaintiff class was represented by a Kansas City lawyer (Benson) and by the NAACP Legal Defense and Educational Fund, Inc. (LDF). Benson and the LDF requested attorney’s fees under the Civil Rights Attorney’s Fees Awards Act of 1976 (42 U. S. C. § 1988), which provides with respect to such litigation that the court, in its discretion, may allow the prevailing party, other than the United States, “a reasonable attorney’s fee as part of the costs.” In calculating the hourly rates for Benson’s, his associates’, and the LDF attorneys’ fees, the District Court took account of delay in payment by using current market rates rather than those applicable at the time the services were rendered. Both Benson and the LDF employed numerous paralegals, law clerks, and recent law graduates, and the court awarded fees for their work based on market rates, again using current rather than historic rates in order to compensate for the delay in payment. The Court of Appeals affirmed. Held: 1. The Eleventh Amendment does not prohibit enhancement of a fee award under § 1988 against a State to compensate for delay in payment. That Amendment has no application to an award of attorney’s fees, ancillary to a grant of prospective relief, against a State, Hutto v. Finney, 437 U. S. 678, and it follows that the same is true for the calculation of the amount of the fee. An adjustment for delay in payment is an appropriate factor in determining what constitutes a reasonable attorney’s fee under § 1988. Pp. 278-284. 2. The District Court correctly compensated the work of paralegals, law clerks, and recent law graduates at the market rates for their services, rather than at their cost to the attorneys. Clearly, “a reasonable attorney’s fee” as used in § 1988 cannot have been meant to compensate only work performed personally by members of the bar. Rather, that term must refer to a reasonable fee for an attorney’s work product, and thus must take into account the work not only of attorneys, but also the work of paralegals and the like. A reasonable attorney’s fee under MISSOURI v. JENKINS 275 274 Opinion of the Court § 1988 is one calculated on the basis of rates and practices prevailing in the relevant market and one that grants the successful civil rights plaintiff a “fully compensatory fee,” comparable to what “is traditional with attorneys compensated by a fee-paying client.” In this case, where the practice in the relevant market is to bill the work of paralegals separately, the District Court’s decision to award separate compensation for paralegals, law clerks, and recent law graduates at prevailing market rates was fully in accord with § 1988. Pp. 284-289. 838 F. 2d 260, affirmed. Brennan, J., delivered the opinion of the Court, in which White, Blackmun, Stevens, and Kennedy, JJ., joined, and in Parts I and III of which O’Connor and Scalia, JJ., joined. O’Connor, J., filed an opinion concurring in part and dissenting in part, in which Scalia, J., joined, and in which Rehnquist, C. J., joined in part, post, p. 289. Rehnquist, C. J., filed a dissenting opinion, post, p. 295. Marshall, J., took no part in the consideration or decision of the case. Bruce Farmer, Assistant Attorney General of Missouri, argued the cause for petitioners. With him on the brief were William L. Webster, Attorney General, Terry Allen, Deputy Attorney General, and Michael L. Boicourt and Bart A. Ma-tanic, Assistant Attorneys General. Jay Topkis argued the cause for respondents. With him on the brief were Julius LeVonne Chambers, Charles Stephen Ralston, Arthur A. Benson II, Russell E. Lovell II, and Theodore M. Shaw* Justice Brennan delivered the opinion of the Court. This is the attorney’s fee aftermath of major school desegregation litigation in Kansas City, Missouri. We granted certiorari, 488 U. S. 888 (1988), to resolve two questions relating to fees litigation under 90 Stat. 2641, as amended, 42 U. S. C. § 1988. First, does the Eleventh Amendment prohibit enhancement of a fee award against a State to compensate for delay in payment? Second, should the fee award compensate the work of paralegals and law clerks by applying the market rate for their work? *John A. DeVault III filed a brief for the National Association of Legal Assistants, Inc., as amicus curiae urging affirmance. 276 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. I This litigation began in 1977 as a suit by the Kansas City Missouri School District (KCMSD), the school board, and the children of two school board members, against the State of Missouri and other defendants. The plaintiffs alleged that the State, surrounding school districts, and various federal agencies had caused and perpetuated a system of racial segregation in the schools of the Kansas City metropolitan area. They sought various desegregation remedies. KCMSD was subsequently realigned as a nominal defendant, and a class of present and future KCMSD students was certified as plaintiffs. After lengthy proceedings, including a trial that lasted 7% months during 1983 and 1984, the District Court found the State of Missouri and KCMSD liable, while dismissing the suburban school districts and the federal defendants. It ordered various intradistrict remedies, to be paid for by the State and KCMSD, including $260 million in capital improvements and a magnet-school plan costing over $200 million. See Jenkins n. Missouri, 807 F. 2d 657 (CA8 1986) (en banc), cert, denied, 484 U. S. 816 (1987); Jenkins v. Missouri, 855 F. 2d 1295 (CA8 1988), cert, granted, 490 U. S. 1034 (1989). The plaintiff class has been represented, since 1979, by Kansas City lawyer Arthur Benson and, since 1982, by the NAACP Legal Defense and Educational Fund, Inc. (LDF). Benson and the LDF requested attorney’s fees under the Civil Rights Attorney’s Fees Awards Act of 1976, 42 U. S. C. § 1988? Benson and his associates had devoted 10,875 attorney hours to the litigation, as well as 8,108 hours of paralegal and law clerk time. For the LDF the corresponding 1 Section 1988 provides in relevant part: “In any action or proceeding to enforce a provision of sections 1981, 1982, 1983, 1985, and 1986 of this title, title IX of Public Law 92-318 [20 U. S. C. § 1681 et se^.], or title VI of the Civil Rights Act of 1964 [42 U. S. C. §2000d et seq.}, the court, in its discretion, may allow the prevailing party, other than the United States, a reasonable attorney’s fee as part of the costs.” MISSOURI v. JENKINS 277 274 Opinion of the Court figures were 10,854 hours for attorneys and 15,517 hours for paralegals and law clerks. Their fee applications deleted from these totals 3,628 attorney hours and 7,046 paralegal hours allocable to unsuccessful claims against the suburban school districts. With additions for postjudgment monitoring and for preparation of the fee application, the District Court awarded Benson a total of approximately $1.7 million and the LDF $2.3 million. App. to Pet. for Cert. A22-A43. In calculating the hourly rate for Benson’s fees the court noted that the market rate in Kansas City for attorneys of Benson’s qualifications was in the range of $125 to $175 per hour, and found that “Mr. Benson’s rate would fall at the higher end of this range based upon his expertise in the area of civil rights.” Id., at A26. It calculated his fees on the basis of an even higher hourly rate of $200, however, because of three additional factors: the preclusion of other employment, the undesirability of the case, and the delay in payment for Benson’s services. Id., at A26-A27. The court also took account of the delay in payment in setting the rates for several of Benson’s associates by using current market rates rather than those applicable at the time the services were rendered. Id., at A28-A30. For the same reason, it calculated the fees for the LDF attorneys at current market rates. Id., at A33. Both Benson and the LDF employed numerous paralegals, law clerks (generally law students working part time), and recent law graduates in this litigation. The court awarded fees for their work based on Kansas City market rates for those categories. As in the case of the attorneys, it used current rather than historic market rates in order to compensate for the delay in payment. It therefore awarded fees based on hourly rates of $35 for law clerks, $40 for paralegals, and $50 for recent law graduates. Id., at A29-A31, A34. The Court of Appeals affirmed in all respects. 838 F. 2d 260 (CA8 1988). 278 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. II Our grant of certiorari extends to two issues raised by the State of Missouri. Missouri first contends that a State cannot, consistent with the principle of sovereign immunity this Court has found embodied in the Eleventh Amendment, be compelled to pay an attorney’s fee enhanced to compensate for delay in payment. This question requires us to examine the intersection of two of our precedents, Hutto n. Finney, 437 U. S. 678 (1978), and Library of Congress v. Shaw, 478 U. S. 310 (1986).2 In Hutto v. Finney, the lower courts had awarded attorney’s fees against the State of Arkansas, in part pursuant to § 1988, in connection with litigation over the conditions of confinement in that State’s prisons. The State contended that any such award was subject to the Eleventh Amendment’s constraints on actions for damages payable from a State’s treasury. We relied, in rejecting that contention, on the distinction drawn in our earlier cases between “retroactive monetary relief” and “prospective injunctive relief.” See Edelman v. Jordan, 415 U. S. 651 (1974); Ex parte Young, 209 U. S. 123 (1908). Attorney’s fees, we held, belonged to the latter category, because they constituted reimbursement of “expenses incurred in litigation seeking only prospective relief,” rather than “retroactive liability for prelitigation conduct.” Hutto, 437 U. S., at 695; see also id., at 690. We explained: “Unlike ordinary ‘retroactive’ relief such as damages or restitution, an award of costs does not compensate the plaintiff for the injury that first brought him into court. Instead, the award reimburses him for a portion of the expenses he incurred in seeking prospective relief.” Id., at 695, n. 24. Section 1988, we noted, fit easily into the 2 The holding of the Court of Appeals on this point, 838 F. 2d, at 265-266, is in conflict with the resolution of the same question in Rogers v. Okin, 821 F. 2d 22, 26-28 (CAI 1987), cert, denied sub nom. Commissioner, Massachusetts Dept, of Mental Health v. Rogers, 484 U. S. 1010 (1988). MISSOURI v. JENKINS 279 274 Opinion of the Court longstanding practice of awarding “costs” against States, for the statute imposed the award of attorney’s fees “as part of the costs.” Id., at 695-696, citing Fairmont Creamery Co. v. Minnesota, 275 U. S. 70 (1927). After Hutto, therefore, it must be accepted as settled that an award of attorney’s fees ancillary to prospective relief is not subject to the strictures of the Eleventh Amendment. And if the principle of making such an award is beyond the reach of the Eleventh Amendment, the same must also be true for the question of how a “reasonable attorney’s fee’* is to be calculated. See Hutto, supra, at 696-697. Missouri contends, however, that the principle enunciated in Hutto has been undermined by subsequent decisions of this Court that require Congress to “express its intention to abrogate the Eleventh Amendment in unmistakable language in the statute itself.” Atascadero State Hospital v. Scanlon, 473 U. S. 234, 243 (1985); Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468 (1987). See also Dellmuth v. Muth, ante, p. 223; Pennsylvania v. Union Gas Co., ante, p. 1. The flaw in this argument lies in its misreading of the holding of Hutto. It is true that in Hutto we noted that Congress could, in the exercise of its enforcement power under § 5 of the Fourteenth Amendment, set aside the States’ immunity from retroactive damages, 437 U. S., at 693, citing Fitzpatrick v. Bitzer, 42.1 U. S. 445 (1976), and that Congress intended to do so in enacting § 1988, 437 U. S., at 693-694. But we also made clear that the application of § 1988 to the States did not depend on congressional abrogation of the States’ immunity. We did so in rejecting precisely the “clear statement” argument that Missouri now suggests has undermined Hutto. Arkansas had argued that § 1988 did not plainly abrogate the States’ immunity; citing Employees v. Missouri Dept, of Public Health and Welfare, 411 U. S. 279 (1973), and Edelman v. Jordan, supra, the State contended that “retroactive liability” could not be imposed on the States “in the absence of an extraordinarily ex 280 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. plicit statutory mandate.” Hutto, 437 U. S., at 695. We responded as follows: “[T]hese cases [Employees and Edelman] concern retroactive liability for prelitigation conduct rather than expenses incurred in litigation seeking only prospective relief. The Act imposes attorney’s fees ‘as part of the costs.’ Costs have traditionally been awarded without regard for the States’ Eleventh Amendment immunity.” Ibid. The holding of Hutto, therefore, was not just that Congress had spoken sufficiently clearly to overcome Eleventh Amendment immunity in enacting § 1988, but rather that the Eleventh Amendment did not apply to an award of attorney’s fees ancillary to a grant of prospective relief. See Maine n. Thiboutot, 448 U. S. 1, 9, n. 7 (1980). That holding is unaffected by our subsequent jurisprudence concerning the degree of clarity with which Congress must speak in order to override Eleventh Amendment immunity, and we reaffirm it today. Missouri’s other line of argument is based on our decision in Library of Congress v. Shaw, supra. Shaw involved an application of the longstanding “no-interest rule,” under which interest cannot be awarded against the United States unless it has expressly waived its sovereign immunity. We held that while Congress, in making the Federal Government a potential defendant under Title VII of the Civil Rights Act of 1964, had waived the United States’ immunity from suit and from costs including reasonable attorney’s fees, it had not waived the Federal Government’s traditional immunity from any award of interest. We thus held impermissible a 30 percent increase in the “lodestar” fee to compensate for delay in payment. Because we refused to find in the language of Title VII a waiver of the United States’ immunity from interest, Missouri argues, we should likewise conclude that § 1988 is not sufficiently explicit to constitute an abrogation of the States’ immunity under the Eleventh Amendment in regard to any award of interest. MISSOURI v. JENKINS 281 274 Opinion of the Court The answer to this contention is already clear from what we have said about Hutto v. Finney. Since, as we held in Hutto, the Eleventh Amendment does not bar an award of attorney’s fees ancillary to a grant of prospective relief, our holding in Shaw has no application, even by analogy.3 There is no need in this case to determine whether Congress has spoken sufficiently clearly to meet a “clear statement” requirement, and it is therefore irrelevant whether the Eleventh Amendment standard should be, as Missouri contends, as stringent as the one we applied for purposes of the nointerest rule in Shaw. Rather, the issue here—whether the “reasonable attorney’s fee” provided for in § 1988 should be calculated in such a manner as to include an enhancement, where appropriate, for delay in payment—is a straightfor 3 Our opinion in Shaw does, to be sure, contain some language that, if read in isolation, might suggest a different result in this case. Most significantly, we equated compensation for delay with prejudgment interest, and observed that “[p]rejudgment interest ... is considered as damages, not a component of ‘costs.’ . . . Indeed, the term ‘costs’ has never been understood to include any interest component.” Library of Congress v. Shaw, 478 U. S. 310, 321 (1986). These observations, however, cannot be divorced from the context of the special “no-interest rule” that was at issue in Shaw. That rule, which is applicable to the immunity of the United States and is therefore not at issue here, provides an “added gloss of strictness,” id., at 318, only where the United States’ liability for interest is at issue. Our inclusion of compensation for delay within the definition of prejudgment interest in Shaw must be understood in light of this broad proscription of interest awards against the United States. Shaiv thus does not represent a general-purpose definition of compensation for delay that governs here. Outside the context of the “no-interest rule” of federal immunity, we see no reason why compensation for delay cannot be included within § 1988 attorney’s fee awards, which Hutto held to be “costs” not subject to Eleventh Amendment strictures. We cannot share Justice O’Connor’s view that the two cases she cites, post, at 293, demonstrate the existence of an equivalent rule relating to state immunity that embodies the same ultrastrict rule of construction for interest awards that has grown up around the federal no-interest rule. Cf. Shaw, supra, at 314-317 (discussing historical development of the federal no-interest rule). 282 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ward matter of statutory interpretation. For this question, it is of no relevance whether the party against which fees are awarded is a State. The question is what Congress intended—not whether it manifested “the clear affirmative intent ... to waive the sovereign’s immunity.” Shaw, 478 U. S., at 321.4 This question is not a difficult one. We have previously explained, albeit in dicta, why an enhancement for delay in payment is, where appropriate, part of a “reasonable attorney’s fee.” In Pennsylvania v. Delaware Valley Citizens’ Council, 483 U. S. 711 (1987), we rejected an argument that a prevailing party was entitled to a fee augmentation to compensate for the risk of nonpayment. But we took care to distinguish that risk from the factor of delay: “First is the matter of delay. When plaintiffs’ entitlement to attorney’s fees depends on success, their lawyers are not paid until a favorable decision finally eventuates, which may be years later.... Meanwhile, their expenses of doing business continue and must be met. In setting fees for prevailing counsel, the courts have regularly recognized the delay factor, either by basing the award on current rates or by adjusting the fee based on historical rates to reflect its present value. See, e. g., Sierra Club v. EPA, 248 U. S. App. D. C. 107, 120-121, 769 F. 2d 796, 809-810 (1985); Louisville Black Police Officers Organization, Inc. v. Louisville, 700 F. 2d 268, 276, 281 (CA6 1983). Although delay and the risk of nonpayment are often mentioned in the same breath, adjusting for the former is a distinct issue .... We do not suggest . . . that adjustments for delay are 4 In Shaw, which dealt with the sovereign immunity of the Federal Government, there was of course no prospective-retrospective distinction as there is when, as in Hutto and the present case, it is the Eleventh Amendment immunity of a State that is at issue. MISSOURI v. JENKINS 283 274 Opinion of the Court inconsistent with the typical fee-shifting statute.” Id., at 716. The same conclusion is appropriate under § 1988.5 Our cases have repeatedly stressed that attorney’s fees awarded under this statute are to be based on market rates for the services rendered. See, e. g., Blanchard n. Bergeron, 489 U. S. 87 (1989); Riverside v. Rivera, 477 U. S. 561 (1986); Blum v. Stenson, 465 U. S. 886 (1984). Clearly, compensation received several years after the services were rendered—as it frequently is in complex civil rights litigation—is not equivalent to the same dollar amount received reasonably promptly as the legal services are performed, as would normally be the case with private billings.6 We agree, there 5 Delaware Valley was decided under § 304(d) of the Clean Air Act, 42 U. S. C. § 7604(d). We looked for guidance, however, to § 1988 and our cases construing it. Pennsylvania v. Delaware Valley Citizens’ Council, 483 U. S. 711, 713, n. 1 (1987). 6 This delay, coupled with the fact that, as we recognized in Delaware Valley, the attorney’s expenses are not deferred pending completion of the litigation, can cause considerable hardship. The present case provides an illustration. During a period of nearly three years, the demands of this case precluded attorney Benson from accepting other employment. In order to pay his staff and meet other operating expenses, he was obliged to borrow $633,000. As of January 1987, he had paid over $113,000 in interest on this debt, and was continuing to borrow to meet interest payments. Record 2336-2339; Tr. 130-131. The LDF, for its part, incurred deficits of $700,000 in 1983 and over $1 million in 1984, largely because of this case. Tr. 46. If no compensation were provided for the delay in payment, the prospect of such hardship could well deter otherwise willing attorneys from accepting complex civil rights cases that might offer great benefit to society at large; this result would work to defeat Congress’ purpose in enacting § 1988 of “encouraging] the enforcement of federal law through lawsuits filed by private persons.” Delaware Valley, supra, at 737 (Blackmun, J., dissenting). We note also that we have recognized the availability of interim fee awards under § 1988 when a litigant becomes a prevailing party on one issue in the course of the litigation. Texas State Teachers Assn. v. Garland Independent School Dist., 489 U. S. 782, 791-792 (1989). In eco 284 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. fore, that an appropriate adjustment for delay in payment — whether by the application of current rather than historic hourly rates or otherwise—is within the contemplation of the statute. To summarize: We reaffirm our holding in Hutto v. Finney that the Eleventh Amendment has no application to an award of attorney’s fees, ancillary to a grant of prospective relief, against a State. It follows that the same is true for the calculation of the amount of the fee. An adjustment for delay in payment is, we hold, an appropriate factor in the determination of what constitutes a reasonable attorney’s fee under § 1988. An award against a State of a fee that includes such an enhancement for delay is not, therefore, barred by the Eleventh Amendment. Ill Missouri’s second contention is that the District Court erred in compensating the work of law clerks and paralegals (hereinafter collectively “paralegals”) at the market rates for their services, rather than at their cost to the attorney. While Missouri agrees that compensation for the cost of these personnel should be included in the fee award, it suggests that an hourly rate of $15—which it argued below corresponded to their salaries, benefits, and overhead—would be appropriate, rather than the market rates of $35 to $50. According to Missouri, § 1988 does not authorize billing paralegals’ hours at market rates, and doing so produces a “windfall” for the attorney.7 nomic terms, such an interim award does not differ from an enhancement for delay in payment. 7 The Courts of Appeals have taken a variety of positions on this issue. Most permit separate billing of paralegal time. See, e. g., Save Our Cumberland Mountains, Inc. v. Hodel, 263 U. S. App. D. C. 409, 420, n. 7, 826 F. 2d 43, 54, n. 7 (1987), vacated in part on other grounds, 273 U. S. App. D. C. 78, 857 F. 2d 1516 (1988) (en banc); Jacobs v. Mancuso, 825 F. 2d 559, 563, and n. 6 (CAI 1987) (collecting cases); Spanish Action Committee MISSOURI v. JENKINS 285 274 Opinion of the Court We begin with the statutory language, which provides simply for “a reasonable attorney’s fee as part of the costs.” 42 U. S. C. § 1988. Clearly, a “reasonable attorney’s fee” cannot have been meant to compensate only work performed personally by members of the bar. Rather, the term must refer to a reasonable fee for the work product of an attorney. Thus, the fee must take into account the work not only of attorneys, but also of secretaries, messengers, librarians, janitors, and others whose labor contributes to the work product for which an attorney bills her client; and it must also take account of other expenses and profit. The parties have suggested no reason why the work of paralegals should not be similarly compensated, nor can we think of any. We thus take as our starting point the self-evident proposition that the “reasonable attorney’s fee” provided for by statute should compensate the work of paralegals, as well as that of attorneys. The more difficult question is how the work of paralegals is to be valuated in calculating the overall attorney’s fee. The statute specifies a “reasonable” fee for the attorney’s work product. In determining how other elements of the attorney’s fee are to be calculated, we have consistently looked to the marketplace as our guide to what is “reasonable.” In Blum v. Stenson, 465 U. S. 886 (1984), for example, we rejected an argument that attorney’s fees for nonprofit legal of Chicago v. Chicago, 811 F. 2d 1129, 1138 (CA7 1987); Ramos v. Lamm, 713 F. 2d 546, 558-559 (CAIO 1983); Richardson v. Byrd, 709 F. 2d 1016, 1023 (CA5), cert, denied sub nom. Dallas County Commissioners Court v. Richardson, 464 U. S. 1009 (1983). See also Riverside v. Rivera, 477 U. S. 561, 566, n. 2 (1986) (noting lower court approval of hourly rate for law clerks). Some courts, on the other hand, have considered paralegal work “out-of-pocket expense,” recoverable only at cost to the attorney. See, e. g., Northcross v. Board of Education of Memphis City Schools, 611 F. 2d 624, 639 (CA6 1979), cert, denied, 447 U. S. 911 (1980); Thornberry v. Delta Air Lines, Inc., 676 F. 2d 1240, 1244 (CA9 1982), vacated, 461 U. S. 952 (1983). At least one Court of Appeals has refused to permit any recovery of paralegal expense apart from the attorney’s hourly fee. Abrams v. Baylor College of Medicine, 805 F. 2d 528, 535 (CA5 1986). 286 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. service organizations should be based on cost. We said: “The statute and legislative history establish that ‘reasonable fees’ under § 1988 are to be calculated according to the prevailing market rates in the relevant community . . . .” Id., at 895. See also, e. g., Delaware Valley, 483 U. S., at 732 (O’Connor, J., concurring) (controlling question concerning contingency enhancements is “how the market in a community compensates for contingency”); Rivera, 477 U. S., at 591 (Rehnquist, J., dissenting) (reasonableness of fee must be determined “in light of both the traditional billing practices in the profession, and the fundamental principle that the award of a ‘reasonable’ attorney’s fee under § 1988 means a fee that would have been deemed reasonable if billed to affluent plaintiffs by their own attorneys”). A reasonable attorney’s fee under § 1988 is one calculated on the basis of rates and practices prevailing in the relevant market, i. e., “in line with those [rates] prevailing in the community for similar services by lawyers of reasonably comparable skill, experience, and reputation,” Blum, supra, at 896, n. 11, and one that grants the successful civil rights plaintiff a “fully compensatory fee,” Hensley v. Eckerhart, 461 U. S. 424, 435 (1983), comparable to what “is traditional with attorneys compensated by a feepaying client.” S. Rep. No. 94-1011, p. 6 (1976). If an attorney’s fee awarded under § 1988 is to yield the same level of compensation that would be available from the market, the “increasingly widespread custom of separately billing for the services of paralegals and law students who serve as clerks,” Ramos n. Lamm, 713 F. 2d 546, 558 (CAIO 1983), must be taken into account. All else being equal, the hourly fee charged by an attorney whose rates include paralegal work in her hourly fee, or who bills separately for the work of paralegals at cost, will be higher than the hourly fee charged by an attorney competing in the same market who bills separately for the work of paralegals at “market rates.” In other words, the prevailing “market rate” for attorney time is not independent of the manner in which paralegal MISSOURI v. JENKINS 287 274 Opinion of the Court time is accounted for.8 Thus, if the prevailing practice in a given community were to bill paralegal time separately at market rates, fees awarded the attorney at market rates for attorney time would not be fully compensatory if the court refused to compensate hours billed by paralegals or did so only at “cost.” Similarly, the fee awarded would be too high if the court accepted separate billing for paralegal hours in a market where that was not the custom. We reject the argument that compensation for paralegals at rates above “cost” would yield a “windfall” for the prevailing attorney. Neither petitioners nor anyone else, to our knowledge, has ever suggested that the hourly rate applied to the work of an associate attorney in a law firm creates a windfall for the firm’s partners or is otherwise improper under § 1988, merely because it exceeds the cost of the attorney’s services. If the fees are consistent with market rates and practices, the “windfall” argument has no more force with regard to paralegals than it does for associates. And it would hardly accord with Congress’ intent to provide a “fully compensatory fee” if the prevailing plaintiff’s attorney in a civil rights lawsuit were not permitted to bill separately for paralegals, while the defense attorney in the same litigation was able to take advantage of the prevailing practice and obtain market rates for such work. Yet that is precisely the result sought in this case by the State of Missouri, which appears to have paid its own outside counsel for the work of paralegals at the hourly rate of $35. Record 2696, 2699? 8 The attorney who bills separately for paralegal time is merely distributing her costs and profit margin among the hourly fees of other members of her staff, rather than concentrating them in the fee she sets for her own time. 9 A variant of Missouri’s “windfall” argument is the following: “If paralegal expense is reimbursed at a rate many times the actual cost, will attorneys next try to bill separately—and at a profit—for such items as secretarial time, paper clips, electricity, and other expenses?” Reply Brief for Petitioners 15-16. The answer to this question is, of course, that attorneys seeking fees under § 1988 would have no basis for requesting separate 288 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Nothing in § 1988 requires that the work of paralegals invariably be billed separately. If it is the practice in the relevant market not to do so, or to bill the work of paralegals only at cost, that is all that § 1988 requires. Where, however, the prevailing practice is to bill paralegal work at market rates, treating civil rights lawyers’ fee requests in the same way is not only permitted by § 1988, but also makes economic sense. By encouraging the use of lower cost paralegals rather than attorneys wherever possible, permitting market-rate billing of paralegal hours “encourages cost-effective delivery of legal services and, by reducing the spiraling cost of civil rights litigation, furthers the policies underlying civil rights statutes.” Cameo Convalescent Center, Inc. v. Senn, 738 F. 2d 836, 846 (CA7 1984), cert, denied, 469 U. S. 1106 (1985).10 compensation of such expenses unless this were the prevailing practice in the local community. The safeguard against the billing at a profit of secretarial services and paper clips is the discipline of the market. 10 It has frequently been recognized in the lower courts that paralegals are capable of carrying out many tasks, under the supervision of an attorney, that might otherwise be performed by a lawyer and billed at a higher rate. Such work might include, for example, factual investigation, including locating and interviewing witnesses; assistance with depositions, interrogatories, and document production; compilation of statistical and financial data; checking legal citations; and drafting correspondence. Much such work lies in a gray area of tasks that might appropriately be performed either by an attorney or a paralegal. To the extent that fee applicants under § 1988 are not permitted to bill for the work of paralegals at market rates, it would not be surprising to see a greater amount of such work performed by attorneys themselves, thus increasing the overall cost of litigation. Of course, purely clerical or secretarial tasks should not be billed at a paralegal rate, regardless of who performs them. What the court in Johnson v. Georgia Highway Express, Inc., 488 F. 2d 714, 717 (CA5 1974), said in regard to the work of attorneys is applicable by analogy to paralegals: “It is appropriate to distinguish between legal work, in the strict sense, and investigation, clerical work, compilation of facts and statistics and other work which can often be accomplished by non-lawyers but which a lawyer may do because he has no other help available. Such non-legal MISSOURI v. JENKINS 289 274 Opinion of O’Connor, J. Such separate billing appears to be the practice in most communities today.11 In the present case, Missouri concedes that “the local market typically bills separately for paralegal services,” Tr. of Oral Arg. 14, and the District Court found that the requested hourly rates of $35 for law clerks, $40 for paralegals, and $50 for recent law graduates were the prevailing rates for such services in the Kansas City area. App. to Pet. for Cert. A29, A31, A34. Under these circumstances, the court’s decision to award separate compensation at these rates was fully in accord with § 1988. IV The courts below correctly granted a fee enhancement to compensate for delay in payment and approved compensation of paralegals and law clerks at market rates. The judgment of the Court of Appeals is therefore Affirmed. Justice Marshall took no part in the consideration or decision of this case. Justice O’Connor, with whom Justice Scalia joins, and with whom The Chief Justice joins in part, concurring in part and dissenting in part. I agree with the Court that 42 U. S. C. § 1988 allows compensation for the work of paralegals and law clerks at market rates, and therefore join Parts I and III of its opinion. I do not join Part II, however, for in my view the Eleventh Amendment does not permit enhancement of attorney’s work may command a lesser rate. Its dollar value is not enhanced just because a lawyer does it.” 11 Amicus National Association of Legal Assistants reports that 77 percent of 1,800 legal assistants responding to a survey of the association’s membership stated that their law firms charged clients for paralegal work on an hourly billing basis. Brief for National Association of Legal Assistants as Amicus Curiae 11. 290 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. fees assessed against a State as compensation for delay in payment. The Eleventh Amendment does not, of course, provide a State with across-the-board immunity from all monetary relief. Relief that “serves directly to bring an end to a violation of federal law is not barred by the Eleventh Amendment even though accompanied by a substantial ancillary effect” on a State’s treasury. Papasan n. Allain, 478 U. S. 265, 278 (1986). Thus, in Milliken v. Bradley, 433 U. S. 267, 289-290 (1977), the Court unanimously upheld a decision ordering a State to pay over $5 million to eliminate the effects of de jure segregation in certain school systems. On the other hand, “[r]elief that in essence serves to compensate a party injured in the past,” such as relief “expressly denominated as damages,” or “relief [that] is tantamount to an award of damages for a past violation of federal law, even though styled as something else,” is prohibited by the Eleventh Amendment. Papasan, supra, at 278. The crucial question in this case is whether that portion of respondents’ attorney’s fees based on current hourly rates is properly characterized as retroactive monetary relief. In Library of Congress v. Shaw, 478 U. S. 310 (1986), the Court addressed whether the attorney’s fees provision of Title VII of the Civil Rights Act of 1964, 42 U. S. C. § 2000e-5(k), permits an award of attorney’s fees against the United States to be enhanced in order to compensate for delay in payment. In relevant part, §2000e-5(k) provides: “In any action or proceeding under this subchapter the court, in its discretion, may allow the prevailing party, other than the [Equal Employment Opportunity Commission (EEOC)] or the United States, a reasonable attorney’s fees as part of the costs, and the [EEOC] and the United States shall be liable for costs the same as a private person.” The Court began its analysis in Shaw by holding that “interest is an element of damages separate from damages on the MISSOURI v. JENKINS 291 274 Opinion of O’Connor, J. substantive claim.” 478 U. S., at 314 (citing C. McCormick, Law of Damages § 50, p. 205 (1935)). Given the “no-interest” rule of federal sovereign immunity, under which the United States is not liable for interest absent an express statutory waiver to the contrary, the Court was unwilling to conclude that, by equating the United States’ liability to that of private persons in §2000e-5(k), Congress had waived the United States’ immunity from interest. 478 U. S., at 314-319. The fact that § 2000e-5(k) used the word “reasonable” to modify “attorney’s fees” did not alter this result, for the Court explained that it had “consistently . . . refused to impute an intent to waive immunity from interest into the ambiguous use of a particular word or phrase in a statute.” Id., at 320. The description of attorney’s fees as costs in §2000e-5(k) also did not mandate a contrary conclusion because “[p]rejudgment interest... is considered as damages, not a component of ‘costs,’” and the “term ‘costs’ has never been understood to include any interest component.” Id., at 321 (emphasis added) (citing 10 C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure §§2664, 2666, 2670 (2d ed. 1983); 2 A. Sedgwick & G. Van Nest, Sedgwick on Damages 157-158 (7th ed. 1880)). Finally, the Court rejected the argument that the enhancement was proper because the “no-interest” rule did not prohibit compensation for delay in payment: “Interest and a delay factor share an identical function. They are designed to compensate for the belated receipt of money.” 478 U. S., at 322. As the Court notes, ante, at 281, n. 3, the “no-interest” rule of federal sovereign immunity at issue in Shaw provided an “added gloss of strictness,” 478 U. S., at 318, and may have explained the result reached by the Court in that case, i. e., that §2000e-5(k) did not waive the United States’ immunity against awards of interest. But there is not so much as a hint anywhere in Shaw that the Court’s discussions and definitions of interest and compensation for delay were dictated by, or limited to, the federal “no-interest” rule. As the 292 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. quotations above illustrate, the Court’s opinion in Shaw is filled with broad, unqualified language. The dissenters in Shaw did not disagree with the Court’s sweeping characterization of interest and compensation for delay as damages. Rather, they argued only that §2000e-5(k) had waived the immunity of the United States with respect to awards of interest. See id., at 323-327 (Brennan, J., dissenting). I therefore emphatically disagree with the Court’s statement that “Shaw . . . does not represent a general-purpose definition of compensation for delay that governs here.” Ante, at 281, n. 3. Two general propositions that are relevant here emerge from Shaw. First, interest is considered damages and not costs. Second, compensation for delay, which serves the same function as interest, is also the equivalent of damages. These two propositions make clear that enhancement for delay constitutes retroactive monetary relief barred by the Eleventh Amendment. Given my reading of Shaw, I do not think the Court’s reliance on the cost rationale of § 1988 set forth in Hutto v. Finney, 437 U. S. 678 (1978), is persuasive. Because Shaw teaches that compensation for delay constitutes damages and cannot be considered costs, see 478 U. S., at 321-322, Hutto is not controlling. See Hutto, supra, at 697, n. 27 (“[W]e do not suggest that our analysis would be the same if Congress were to expand the concept of costs beyond the traditional category of litigation expenses”). Furthermore, Hutto does not mean that inclusion of attorney’s fees as costs in a statute forecloses a challenge to the enhancement of fees as compensation for delay in payment. If it did, then Shaw would have been resolved differently, for § 2000e-5(k) lists attorney’s fees as costs. Even if I accepted the narrow interpretation of Shaw proffered by the Court, I would disagree with the result reached by the Court in Part II of its opinion. On its own terms, the Court’s analysis fails. The Court suggests that the definitions of interest and compensation for delay set forth in Shaw MISSOURI v. JENKINS 293 274 Opinion of O’Connor, J. would be triggered only by a rule of sovereign immunity barring awards of interest against the States: “Outside the context of the ‘no-interest rule’ of federal immunity, we see no reason why compensation for delay cannot be included within §1988 attorney’s fee awards.” Ante, at 281, n. 3. But the Court does not inquire about whether such a rule exists. In fact, there is a federal rule barring awards of interest against States. See Virginia v. West Virginia, 238 U. S. 202, 234 (1915) (“Nor can it be deemed in derogation of the sovereignty of the State that she should be charged with interest if her agreement properly construed so provides”) (emphasis added); United States v. North Carolina, 136 U. S. 211, 221 (1890) (“general principle” is that “an obligation of the State to pay interest, whether as interest or as damages, on any debt overdue, cannot arise except by the consent and contract of the State, manifested by statute, or in a form authorized by statute”) (emphasis added). The Court has recently held that the rule of immunity set forth in Virginia and North Carolina is inapplicable in situations where the State does not retain any immunity, see West Virginia v. United States, 479 U. S. 305, 310-312 (1987) (State can be held liable for interest to the United States, against whom it has no sovereign immunity), but the rule has not otherwise been limited, and there is no reason why it should not be relevant in the Eleventh Amendment context presented in this case. As Virginia and North Carolina indicate, a State can waive its immunity against awards of interest. See also Clark v. Barnard, 108 U. S. 436, 447 (1883). The Missouri courts have interpreted Mo. Rev. Stat. §408.020 (1979 and Supp. 1989), providing for prejudgment interest on money that becomes due and payable, and §408.040, providing for prejudgment interest on court judgments and orders, as making the State liable for interest. See Denton Construction Co. v. Missouri State Highway Common, 454 S. W. 2d 44, 59-60 (Mo. 1970) (§408.020); Steppelman n. State Highway Comm’n of Missouri, 650 S. W. 2d 343, 345 (Mo. App. 294 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. 1983) (§408.040). There can be no argument, however, that these Missouri statutes and cases allow interest to be awarded against the State here. A “State’s waiver of sovereign immunity in its own courts is not a waiver of the Eleventh Amendment immunity in the federal courts.” Penn-hurst State School and Hospital v. Halderman, 465 U. S. 89, 99, n. 9 (1984). The fact that a State has immunity from awards of interest is not the end of the matter. In a case such as this one involving school desegregation, interest or compensation for delay (in the guise of current hourly rates) can theoretically be awarded against a State despite the Eleventh Amendment’s bar against retroactive monetary liability. The Court has held that Congress can set aside the States’ Eleventh Amendment immunity in order to enforce the provisions of the Fourteenth Amendment. See City of Rome v. United States, 446 U. S. 156, 179 (1980); Fitzpatrick v. Bitzer, 427 U. S. 445, 456 (1976). Congress must, however, be unequivocal in expressing its intent to abrogate that immunity. See generally Atascadero State Hospital v. Scanlon, 473 U. S. 234, 243 (1985) (“Congress must express its intention to abrogate the Eleventh Amendment in unmistakable language in the statute itself”). In Hutto the Court was able to avoid deciding whether § 1988 met the “clear statement” rule only because attorney’s fees (without any enhancement) are not considered retroactive in nature. See 437 U. S., at 695-697. The Court cannot do the same here, where the attorney’s fees were enhanced to compensate for delay in payment. Cf. Osterneck n. Ernst & Whinney, 489 U. S. 169, 175 (1989) (“[U]nlike attorney’s fees, which at common law were regarded as an element of costs, . . . prejudgment interest traditionally has been considered part of the compensation due [the] plaintiff”). In relevant part, § 1988 provides: “In any action or proceeding to enforce a provision of sections 1981, 1982, 1983, 1985, and 1986 of this title, MISSOURI v. JENKINS 295 274 Rehnquist, C. J., dissenting title IX of Public Law 92-318, or title VI of the Civil Rights Act of 1964, the court, in its discretion, may allow the prevailing party, other than the United States, a reasonable attorney’s fee as part of the costs.” In my view, § 1988 does not meet the “clear statement” rule set forth in Atascadero. It does not mention damages, interest, compensation for delay, or current hourly rates. As one federal court has correctly noted, “Congress has not yet made any statement suggesting that a § 1988 attorney’s fee award should include prejudgment interest. ” Rogers n. Okin, 821 F. 2d 22, 27 (CAI 1987). A comparison of the statute at issue in Shaw also indicates that § 1988, as currently written, is insufficient to allow attorney’s fees assessed against a State to be enhanced to compensate for delay in payment. The language of § 1988 is undoubtedly less expansive than that of §2000e-5(k), for § 1988 does not equate the liability of States with that of private persons. Since § 2000e-5(k) does not allow enhancement of an award of attorney’s fees to compensate for delay, it is logical to conclude that § 1988, a more narrowly worded statute, likewise does not allow interest (through the use of current hourly rates) to be tacked on to an award of attorney’s fees against a State. Compensation for delay in payment was one of the reasons the District Court used current hourly rates in calculating respondents’ attorney’s fees. See App. to Pet. for Cert. A26-A27; 838 F. 2d 260, 263, 265 (CA8 1988). I would reverse the award of attorney’s fees to respondents and remand so that the fees can be calculated without taking compensation for delay into account. Chief Justice Rehnquist, dissenting. I agree with Justice O’Connor that the Eleventh Amendment does not permit an award of attorney’s fees against a State which includes compensation for delay in payment. Unlike Justice O’Connor, however, I do not agree with the 296 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. Court’s approval of the award of law clerk and paralegal fees made here. Title 42 U. S. C. § 1988 gives the district courts discretion to allow the prevailing party in an action under 42 U. S. C. § 1983 “a reasonable attorney’s fee as part of the costs.” The Court reads this language as authorizing recovery of “a ‘reasonable’ fee for the attorney’s work product,” ante, at 285, which, the Court concludes, may include separate compensation for the services of law clerks and paralegals. But the statute itself simply uses the very familiar term “a reasonable attorney’s fee,” which to those untutored in the Court’s linguistic juggling means a fee charged for services rendered by an individual who has been licensed to practice law. Because law clerks and paralegals have not been licensed to practice law in Missouri, it is difficult to see how charges for their services may be separately billed as part of “attorney’s fees.” And since a prudent attorney customarily includes compensation for the cost of law clerk and paralegal services, like any other sort of office overhead—from secretarial staff, janitors, and librarians, to telephone service, stationery, and paper clips—in his own hourly billing rate, allowing the prevailing party to recover separate compensation for law clerk and paralegal services may result in “double recovery.” The Court finds justification for its ruling in the fact that the prevailing practice among attorneys in Kansas City is to bill clients separately for the services of law clerks and paralegals. But I do not think Congress intended the meaning of the statutory term “attorney’s fee” to expand and contract with each and every vagary of local billing practice. Under the Court’s logic, prevailing parties could recover at market rates for the cost of secretaries, private investigators, and other types of lay personnel who assist the attorney in preparing his case, so long as they could show that the prevailing practice in the local market was to bill separately for these services. Such a result would be a sufficiently drastic departure from the traditional concept of “attorney’s fees” that I MISSOURI v. JENKINS 297 274 Rehnquist, C. J., dissenting believe new statutory authorization should be required for it. That permitting separate billing of law clerk and paralegal hours at market rates might “‘reduc[e] the spiraling cost of civil rights litigation’ ” by encouraging attorneys to delegate to these individuals tasks which they would otherwise perform themselves at higher cost, ante, at 288, and n. 10, may be a persuasive reason for Congress to enact such additional legislation. It is not, however, a persuasive reason for us to rewrite the legislation which Congress has in fact enacted. See Badaracco v. Commissioner, 464 U. S. 386, 398 (1984) (“Courts are not authorized to rewrite a statute because they might deem its effects susceptible of improvement”). I also disagree with the State’s suggestion that law clerk and paralegal expenses incurred by a prevailing party, if not recoverable at market rates as “attorney’s fees” under § 1988, are nonetheless recoverable at actual cost under that statute. The language of § 1988 expands the traditional definition of “costs” to include “a reasonable attorney’s fee,” but it cannot fairly be read to authorize the recovery of all other out-of-pocket expenses actually incurred by the prevailing party in the course of litigation. Absent specific statutory authorization for the recovery of such expenses, the prevailing party remains subject to the limitations on cost recovery imposed by Federal Rule of Civil Procedure 54(d) and 28 U. S. C. § 1920, which govern the taxation of costs in federal litigation where a cost-shifting statute is not applicable. Section 1920 gives the district court discretion to tax certain types of costs against the losing party in any federal litigation. The statute specifically enumerates six categories of expenses which may be taxed as costs: fees of the court clerk and marshal; fees of the court reporter; printing fees and witness fees; copying fees; certain docket fees; and fees of court-appointed experts and interpreters. We have held that this list is exclusive. Crawford Fitting Co. n. J. T. Gibbons, Inc., 482 U. S. 437 (1987). Since none of these categories can possibly be construed to include the fees of law clerks and paralegals, 298 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. I would also hold that reimbursement for these expenses may not be separately awarded at actual cost. I would therefore reverse the award of reimbursement for law clerk and paralegal expenses. CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 299 Syllabus CONSOLIDATED RAIL CORPORATION v. RAILWAY LABOR EXECUTIVES’ ASSOCIATION et al. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT No. 88-1. Argued February 28, 1989—Decided June 19, 1989 Since its formation in 1976, petitioner Consolidated Rail Corporation (Con-rail) has required its employees to undergo physical examinations periodically and upon return from leave. Those examinations routinely included a urinalysis for blood sugar and albumin and, in some circumstances, for drugs. In 1987, Conrail announced unilaterally that urinalysis drug screening would be included as part of all periodic and return-from-leave physical examinations. Respondent Railway Labor Executives’ Association opposed this unilateral additional drug testing. The question presented by this case is whether Conrail’s drug-testing program gives rise to a “major” or a “minor” dispute under the Railway Labor Act. Held: 1. Where an employer asserts a contractual right to take a contested action, the ensuing dispute is minor if the action is arguably justified by the terms of the parties’ collective-bargaining agreement. Where, in contrast, the employer’s claims are frivolous or obviously insubstantial, the dispute is major. Pp. 302-307. 2. If an employer asserts a claim that the parties’ agreement gives the employer the discretion to make a particular change in working conditions without prior negotiation, and if that claim is arguably justified by the terms of the parties’ agreement, the employer may make the change and the courts must defer to the arbitral jurisdiction of the Adjustment Board. Pp. 307-311. 3. Conrail’s contractual claim is not obviously insubstantial, and therefore the controversy constitutes a minor dispute that is within the Adjustment Board’s exclusive jurisdiction. Pp. 311-320. 845 F. 2d 1187, reversed. Blackmun, J., delivered the opinion of the Court, in which Rehnquist, C. J., and White, Stevens, O’Connor, Scalia, and Kennedy, JJ., joined. White, J., filed a concurring opinion, post, p. 320. Brennan, J., filed a dissenting opinion, in which Marshall, J., joined, post, p. 321. Dennis J. Morikawa argued the cause for petitioner. With him on the briefs were Harry A. Rissetto, Michael J. 300 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Ossip, Sarah A. Kelly, Bruce B. Wilson, and Jeffrey H. Burton. John O’B. Clarke, Jr., argued the cause for respondents. With him on the brief were Lawrence M. Mann, William G. Mahoney, Laurence Gold, and Cornelius C. O’Brien, Jr* Justice Blackmun delivered the opinion of the Court. In this case, we must examine the concepts of “major” and “minor” disputes in the area of railway labor relations, articulate a standard for differentiating between the two, and apply that standard to a drug-testing dispute. I Since its formation in 1976, petitioner Consolidated Rail Corporation (Conrail), has required its employees to undergo physical examinations periodically and upon return from leave. These examinations include the testing of urine for blood sugar and albumin and, in some circumstances, for drugs. On February 20, 1987, Conrail announced unilaterally that urinalysis drug screening would be included henceforth as part of all periodic and return-from-leave physical examinations. Respondent Railway Labor Executives’ Association (the Union), an unincorporated association of chief executive officers of 19 labor organizations which collectively represent Conrail’s employees, opposes this unilateral drug-testing addition.1 *Briefs of amici curiae urging reversal were filed for the United States by Solicitor General Fried, Assistant Attorney General Bolton, Deputy Solicitor General Merrill, Lawrence S. Robbins, and Leonard Schaitman; and for the National Railway Labor Conference by Richard T. Conway, Ralph J. Moore, Jr., and David P. Lee. Martin C. Seham filed a brief for the Allied Pilots Association as amicus curiae. 1 The Union filed suit against Conrail on May 1, 1986, well before Conrail unilaterally added drug testing to its physical examinations. See App. 3. The Union’s complaint challenged Conrail’s use of drug testing to enforce its disciplinary Rule G and to comply with federal drug-testing regulations affecting the railroad industry. By the time the District Court ruled, how- CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 301 299 Opinion of the Court The parties agree that Conrail’s inclusion of drug testing in all physical examinations has created a labor dispute the resolution of which is governed by the Railway Labor Act (RLA), 44 Stat. 577, as amended, 45 U. S. C. § 151 et seq.2 The question presented by this case is what kind of labor dispute we have before us: whether Conrail’s addition of a drug screen to the urinalysis component of its required periodic and return-to-duty medical examinations gives rise to a “major” or a “minor” dispute under the RLA. The United States District Court for the Eastern District of Pennsylvania agreed with Conrail that this case involves a minor dispute, because Conrail’s policy of conducting physical examinations, which the parties agree is an implied term of their collective-bargaining agreement, arguably gave Conrail the discretion to include drug testing in all physical examinations. The Third Circuit reversed, ruling that “the undisputed terms of the implied agreement governing medical examinations cannot be plausibly interpreted to justify the new testing program.” 845 F. 2d 1187, 1193 (1988). Although we find the question to be a close one, we agree with the District Court, and with those Courts of Appeals that have held, on similar facts, that disputes concerning the addition of a drug-testing component to routine physical examinations are minor disputes. See, e. g., Railway Labor Executives Assn. v. Norfolk & Western R. Co., 833 F. 2d 700, 705-706 (CA7 1987); Brotherhood of Maintenance of Way Employees, Lodge 16 v. Burlington Northern R. Co., 802 F. 2d 1016, 1024 (CA8 1986). ever, the focus of the dispute had shifted to the addition of drug testing to routine physical examinations. That is the question framed by Conrail’s petition for certiorari here. 2Cf. Brotherhood of Locomotive Engineers v. Burlington Northern R. Co., 838 F. 2d 1087, 1089-1090 (CA9 1988) (employer took position that drug testing is not a mandatory subject of bargaining and thus that drugtesting disputes are not “labor disputes” subject to the dispute-resolution processes of the RLA), cert, pending, No. 87-1631. 302 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. II This Court has not articulated an explicit standard for differentiating between major and minor disputes. It adopted the major/minor terminology, drawn from the vocabulary of rail management and rail labor, as a shorthand method of describing two classes of controversy Congress had distinguished in the RLA: major disputes seek to create contractual rights, minor disputes to enforce them. Elgin, J. & E. R. Co. n. Burley, 325 U. S. 711, 723 (1945). The statutory bases for the major dispute category are § 2 Seventh and §6 of the RLA, 48 Stat. 1188, 1197, 45 U. S. C. § 152 Seventh and § 156. The former states that no carrier “shall change the rates of pay, rules, or working conditions of its employees, as a class, as embodied in agreements except in the manner prescribed in such agreements” or through the mediation procedures established in §6. This statutory category “relates to disputes over the formation of collective agreements or efforts to secure them. They arise where there is no such agreement or where it is sought to change the terms of one, and therefore the issue is not whether an existing agreement controls the controversy. They look to the acquisition of rights for the future, not to assertion of rights claimed to have vested in the past.” Burley, 325 U. S., at 723. In the event of a major dispute, the RLA requires the parties to undergo a lengthy process of bargaining and mediation.3 §§5 and 6. Until they have exhausted those procedures, the parties are obligated to maintain the status quo, 3 In addition, the RLA provides for arbitration of a major dispute in the event that mediation fails. Thus, the National Mediation Board is required to “endeavor ... to induce the parties to submit their controversey to arbitration.” §5 First. Participation, however, is voluntary. See Aaron, Voluntary Arbitration of Railroad and Airline Interest Disputes, in The Railway Labor Act at Fifty: Collective Bargaining in the Railroad and Airline Industries 129 (C. Rehmus ed. 1977). CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 303 299 Opinion of the Court and the employer may not implement the contested change in rates of pay, rules, or working conditions. The district courts have subject-matter jurisdiction to enjoin a violation of the status quo pending completion of the required procedures, without the customary showing of irreparable injury. See Detroit & T. S. L. R. Co. v. Transportation Union, 396 U. S. 142 (1969) (upholding status quo injunction without discussing equitable constraints); Division No. 1, Detroit, Brotherhood of Locomotive Engineers v. Consolidated Rail Corp., 844 F. 2d 1218 (CA6 1988). Once this protracted process ends and no agreement has been reached, the parties may resort to the use of economic force. In contrast, the minor dispute category is predicated on § 2 Sixth and § 3 First (i) of the RLA, which set forth conference and compulsory arbitration procedures for a dispute arising or growing “out of grievances or out of the interpretation or application of agreements concerning rates of pay, rules, or working conditions.” This second category of disputes “contemplates the existence of a collective agreement already concluded or, at any rate, a situation in which no effort is made to bring about a formal change in terms or to create a new one. The dispute relates either to the meaning or proper application of a particular provision with reference to a specific situation or to an omitted case. In the latter event the claim is founded upon some incident of the employment relation, or asserted one, independent of those covered by the collective agreement, e. g., claims on account of personal injuries. In either case the claim is to rights accrued, not merely to have new ones created for the future.” Burley, 325 U. S., at 723. A minor dispute in the railroad industry is subject to compulsory and binding arbitration before the National Railroad Adjustment Board, § 3, or before an adjustment board established by the employer and the unions representing the em 304 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ployees. §3 Second.4 The Board (as we shall refer to any adjustment board under the RLA) has exclusive jurisdiction over minor disputes. Judicial review of the arbitral decision is limited. See §3 First (q); Union Pacific R. Co. v. Sheehan, 439 U. S. 89, 93 (1978). Courts may enjoin strikes arising out of minor disputes. Trainmen v. Chicago R. & I. R. Co., 353 U. S. 30 (1957). Although courts in some circumstances may condition the granting of a strike injunction on a requirement that the employer maintain the status quo pending Board resolution of the dispute, see Locomotive Engineers n. Missouri-K.-T. R. Co., 363 U. S. 528, 534 (1960), this Court never has recognized a general statutory obligation on the part of an employer to maintain the status quo pending the Board’s decision. Cf. id., at 531, n. 3 (leaving open the question whether a federal court can require an employer to maintain the status quo during the pendency of a minor dispute at the union’s independent behest, where no strike injunction has been sought by the employer).5 In the airline industry, also covered by the RLA, there is no national adjustment board; a minor dispute is resolved by an adjustment board established by the airline and the unions. 49 Stat. 1189, 45 U. S. C. § 184. See Machinists v. Central Airlines, Inc., 372 U. S. 682 (1963). In both the airline and railroad industries, the National Mediation Board has a limited role to play in resolving a minor dispute: under § 5 Second, the Board may be called upon by a party to interpret “any agreement reached through mediation under the provisions of this Act.” See also 49 Stat. 1189, 45 U. S. C. § 183 (applying § 5 to airlines). 5 See generally Comment, Enjoining Strikes and Maintaining the Status Quo in Railway Labor Disputes, 60 Colum. L. Rev. 381, 386-397 (1960); cf. Air Line Pilots Assn., Int’l v. Eastern Air Lines, Inc., 276 U. S. App. D. C. 199, 202, n. 2, 869 F. 2d 1518, 1520, n. 2 (1989); International Assn, of Machinists and Aerospace Workers v. Northeast Airlines, Inc., 473 F. 2d 549, 555, n. 7 (CAI) (expressing the view that a “union [might] be able to enjoin changes in working conditions if it would be impossible otherwise later to make the workers whole”), cert, denied, 409 U. S. 845 (1972); Division No. 1, Detroit, Brotherhood of Locomotive Engineers v. Consolidated Rail Corp., 844 F. 2d 1218, 1224, n. 10 (CA6 1988) (leaving open the question of injunction based on showing of irreparable harm). As the Union in the present case has not based its claim for injunctive relief on an allegation CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 305 299 Opinion of the Court Although experience in the rail industry suggested to Congress that the second category of disputes involved “comparatively minor” issues that seldom led to strikes, the Court recognized in Burley that this was not invariably the case. See 325 U. S., at 724; see also Trainmen, supra. Thus, the formal demarcation between major and minor disputes does not turn on a case-by-case determination of the importance of the issue presented or the likelihood that it would prompt the exercise of economic self-help. See National Railway Labor Conference v. International Assn, of Machinists and Aerospace Workers, 830 F. 2d 741, 747, n. 5 (CA7 1987). Rather, the line drawn in Burley looks to whether a claim has been made that the terms of an existing agreement either establish or refute the presence of a right to take the disputed action. The distinguishing feature of such a case is that the dispute may be conclusively resolved by interpreting the existing agreement. See Garrison, The National Railroad Adjustment Board: A Unique Administrative Agency, 46 Yale L. J. 567, 568, 576 (1937). To an extent, then, the distinction between major and minor disputes is a matter of pleading. The party who initiates a dispute takes the first step toward categorizing the dispute when it chooses whether to assert an existing contractual right to take or to resist the action in question. But of irreparable injury, we decline to resolve the question whether a status quo injunction based on a claim of irreparable injury would be appropriate. The Union suggests in passing that § 2 First provides a status quo obligation applicable to all minor disputes. See Brief for Respondents 21, 30-31. It relies on Detroit & T. S. L. R. Co. v. Transportation Union, 396 U. S. 142, 151 (1969), but, as we read that case, it does not support the Union’s position. The language upon which the Union relies (a reference to “the implicit status quo requirement in the obligation imposed upon both parties by §2 First, ‘to exert every reasonable effort’ to settle disputes without interruption to interstate commerce”) appears in the context of explaining that the express status quo requirements applicable to a major dispute must be broadly interpreted. It has no direct application to a minor dispute. 306 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the Courts of Appeals early recognized that there is a danger in leaving the characterization of the dispute solely in the hands of one party. In a situation in which the party asserting a contractual basis for its claim is “insincere” in so doing, or its “position [is] founded upon . . . insubstantial grounds,” the result of honoring that party’s characterization would be to undercut “the prohibitions of § 2, Seventh, and § 6 of the Act” against unilateral imposition of new contractual terms. Norfolk & Portsmouth Belt Line R. Co. v. Brotherhood of Railroad Trainmen, Lodge No. 514, 248 F. 2d 34, 43-44, n. 4 (CA4 1957), cert, denied, 355 U. S. 914 (1958); see also United Industrial Workers of Seafarers Inti Union, AFL-CIO v. Board of Trustees of Galveston Wharves, 351 F. 2d 183, 188-189 (CA5 1965). In such circumstances, protection of the proper functioning of the statutory scheme requires the court to substitute its characterization for that of the claimant. To satisfy this need for some degree of judicial control, the Courts of Appeals uniformly have established some variant of the standard employed by the Third Circuit in this case: “‘[I]f the disputed action of one of the parties can “arguably” be justified by the existing agreement or, in somewhat different statement, if the contention that the labor contract sanctions the disputed action is not “obviously insubstantial,” the controversy is a [minor dispute] within the exclusive province of the National Railroad Adjustment Board.’” 845 F. 2d, at 1190, quoting Local 1477 United Transportation Union n. Baker, 482 F. 2d 228, 230 (CA6 1973). Verbal formulations of this standard have differed over time and among the Circuits: phrases such as “not arguably justified,” “obviously insubstantial,” “spurious,” and “frivolous” have been employed.6 See, e. g., Brotherhood of Locomo- fi See, e. g., National Railway Labor Conference v, International Assn, of Machinists and Aerospace Workers, 830 F. 2d 741, 746 (CA7 1987) (not CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 307 299 Opinion of the Court twe Engineers v. Burlington Northern R. Co., 838 F. 2d 1087, 1091 (CA9 1988) (reviewing different formulations used in the Ninth Circuit), cert, pending, No. 87-1631. “These locutions are essentially the same in their result. They illustrate the relatively light burden which the railroad must bear” in establishing exclusive arbitral jurisdiction under the RLA. Brotherhood of Maintenance of Way Employees, Lodge 16 v. Burlington Northern R. Co., 802 F. 2d, at 1022; see also Maine Central R. Co. n. United Transportation Union, 787 F. 2d 780, 783 (CAI) (“The degree of scrutiny, while ill-defined, is clearly light”), cert, denied, 479 U. S. 848 (1986). “To the extent that abstract words can deal with concrete cases, we think that the concept embodied in the language adopted by these . . . Courts of Appeals is correct.” Christiansburg Garment Co. v. EEOC, 434 U. S. 412, 421 (1978). Where an employer asserts a contractual right to take the contested action, the ensuing dispute is minor if the action is arguably justified by the terms of the parties’ collectivebargaining agreement. Where, in contrast, the employer’s claims are frivolous or obviously insubstantial, the dispute is major. Ill In this case, the Union appears to agree that the “arguably justified” standard generally is the appropriate one for distinguishing between major and minor disputes. Brief for Respondents 35, n. 29. But it argues that the dispute in this case, properly viewed, is neither a major dispute nor a minor dispute. According to the Union, where an employer has frivolous or obviously insubstantial); Maine Central R. Co. v. United Transportation Union, 787 F. 2d 780, 782 (CAI) (even arguable), cert, denied, 479 U. S. 848 (1986); International Brotherhood of Electrical Workers v. Washington Terminal Co., 154 U. S. App. D. C. 119, 136, 473 F. 2d 1156, 1173 (1972) (reasonably susceptible), cert, denied, 411 U. S. 906 (1973); Ruby v. Taca International Airlines, S. A., 439 F. 2d 1359, 1363, n. 5 (CA5 1971) (wholly spurious). 308 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. made a clear “change [in] . . . working conditions ... as embodied in agreements,” but asserts that it has made the change “in the manner prescribed in such agreements,” §2 Seventh, because it has a contractual right to make the change, the ensuing dispute is a “hybrid case.” Brief for Respondents 34-35, 40, n. 32. In a hybrid dispute, the Union contends, the employer may ask the Board to determine whether it has the contractual right to make a particular change, but must forgo unilateral implementation of the change until the Board reaches its decision. If the employer makes the change without establishing a clear and patent right to do so, the employer violates its statutory duty not to “change the rates of pay, rules, or working conditions of its employees, as a class, as embodied in agreements except in the manner prescribed in such agreements or in section 6.” §2 Seventh (emphasis added). Stated more simply, the Union’s position is that, while a dispute over the right to make the change would be a minor dispute, the actual making of the change transforms the controversy into a major dispute. This approach unduly constrains the freedom of unions and employers to contract for discretion. Collective-bargaining agreements often incorporate express or implied terms that are designed to give management, or the union, a degree of freedom of action within a specified area of activity. See NLRB v. American National Insurance Co., 343 U. S. 395 (1952); Rutland Railway Corp. v. Brotherhood of Locomotive Engineers, 307 F. 2d 21, 35-36 (CA2 1962), cert, denied, 372 U. S. 954 (1963). Cf. Steelworkers v. Warrior & Gulf Navigation Co., 363 U. S. 574, 580 (1960); see generally Cox & Dunlop, Regulation of Collective Bargaining by the National Labor Relations Board, 63 Harv. L. Rev. 389, 401 (1950). We have held under the National Labor Relations Act (NLRA) that no principle of labor law prohibits “[b]argaining for . . . flexible treatment” and requires instead that, for each working condition, the employer “agre[e] to freeze a CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 309 299 Opinion of the Court standard into a contract.” American National Insurance Co., 343 U. S., at 408. We find no difference between the NLRA and the RLA in this respect. Yet the Union would subject to especially strict scrutiny the bona fides of contractual claims arising out of contract terms that grant management the power to respond flexibly to changing circumstances. The effect of a selectively heightened level of scrutiny (a “clear and patent” rather than an “arguably justified” standard) would be to limit the enforceability of such contract terms, by requiring employers rigidly to maintain the status quo pending arbitration of their right to be flexible. That result is odd in itself, cf. Rutland Railway Corp., 307 F. 2d, at 40 (requiring parties to negotiate over whether they have a duty to negotiate is “a solution sounding a lot like an exercise in theoretical logic”), and has unacceptable implications. To accept the bifurcated standard the Union advocates would, in effect, be impermissibly to “pass upon the desirability of the substantive terms of labor agreements,” American National Insurance Co., 343 U. S., at 408-409, by affording flexible terms a less favored status, cf. International Assn, of Machinists and Aerospace Workers v. Northeast Airlines, Inc., 473 F. 2d 549, 555 (CAI), cert, denied, 409 U. S. 845 (1972).7 ’Even if the Union’s approach had merit in the abstract, it would be unworkable in practice. As discussed below, collective-bargaining agreements often contain implied, as well as express, terms. The Union conceded at oral argument that an employer would have the authority, without engaging in collective bargaining or statutory mediation, to open its locker room 15 minutes later than it had in the past without first establishing its contractual right to do so through a separate arbitration proceeding. Tr. of Oral Arg. 47-48, 50. That acknowledgment stemmed from the assumption that, although a change in opening time was indeed a “change,” and although access to the locker room was a “working condition,” the precise time the locker room opened was not an issue of sufficient significance to have become the subject of an implied contractual agreement, even if the existence of the locker room was itself an implied term of the contract. The Union recognizes, then, that the general framework of a collective-bargaining agreement leaves some play in the joints, permitting 310 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Accordingly, we shall not aggravate the already difficult task of distinguishing between major disputes and minor disputes by adding a third category of hybrid disputes. We hold that if an employer asserts a claim that the parties’ agreement gives the employer the discretion to make a particular change in working conditions without prior negotiation, and if that claim is arguably justified by the terms of the parties’ agreement (i. e., the claim is neither obviously insubstantial or frivolous, nor made in bad faith), the employer may make the change and the courts must defer to the arbitral jurisdiction of the Board. The effect of this ruling, of course, will be to delay collective bargaining in some cases until the arbitration process is exhausted. But we see no inconsistency between that result and the policies of the RLA.8 The core duties imposed upon employers and employees by the RLA, as set forth in §2 First, are to “make and maintain agreements” and to “settle all disputes ... in order to avoid any interruption to commerce.” Referring arbitrable matters to the Board will help to “maintain agreements,” by assuring that collectivebargaining contracts are enforced by arbitrators who are experts in “the common law of [the] particular industry.” management some range of flexibility in responding to changed conditions. The effect of adopting the Union’s “hybrid dispute” proposal would be to require the trial court to make a nonexpert generalized judgment regarding the “importance” of a particular working condition, and to use that judgment as the basis for deciding whether a particular working condition is or is not within the parties’ agreed range of discretion. We decline to put courts to that task. 8 In most cases where the Board determines that the employer’s conduct was not justified by the contract, the Board will be able to fashion an appropriate compensatory remedy which takes account of the delay. See, e. g., Order of Conductors v. Pitney, 326 U. S. 561, 566 (1946); In re Aaxico Airlines, Inc., 47 Lab. Arb. 289, 316 (1966); In re Trans World Airlines, Inc., 34 Lab. Arb. 420, 425 (1959). There may be some circumstances, however, where the delay inherent in permitting the Board to consider the matter in the first instance will lead to remedial difficulties. See generally Comment, 60 Colum. L. Rev., at 394. CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 311 299 Opinion of the Court Steelworkers v. Warrior & Gulf Navigation Co., 363 U. S., at 579. Full utilization of the Board’s procedures also will diminish the risk of interruptions in commerce. Failure of the “virtually endless” process of negotiation and mediation established by the RLA for major disputes, Burlington Northern R. Co. v. Maintenance of Way Employes, 481 U. S. 429, 444 (1987), frees the parties to employ a broad range of economic self-help, which may disturb transportation services throughout the industry and unsettle employer-employee relationships. See TWA, Inc. n. Flight Attendants, 489 U. S. 426 (1989). Delaying the onset of that process until the Board determines on the merits that the employer’s interpretation of the agreement is incorrect will assure that the risks of self-help are not needlessly undertaken and will aid “[t]he peaceable settlement of labor controversies.” Virginian R. Co. v. Railway Employees, 300 U. S. 515, 552 (1937). IV This case, then, turns on whether the inclusion of drug testing in periodic and return-from-leave physical examinations is arguably justified by the parties’ collectivebargaining agreement. Neither party relies on any express provision of the agreement; indeed, the agreement is not part of the record before us. As the parties acknowledge, however, collective-bargaining agreements may include implied, as well as express, terms. See, e. g., Northwest Airlines, Inc. n. Air Line Pilots Assn., Inti, 442 F. 2d 251, 253-254 (CA8), cert, denied, 404 U. S. 871 (1971). Furthermore, it is well established that the parties’ “practice, usage and custom” is of significance in interpreting their agreement. See Transportation Union v. Union Pacific R. Co., 385 U. S. 157, 161 (1966). This Court has observed: “A collective bargaining agreement is not an ordinary contract for the purchase of goods and services, nor is it governed by the same old common-law concepts which control such private contracts. . . [I]t is a generalized code to govern a myriad of cases 312 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. which the draftsmen cannot wholly anticipate. . . . The collective agreement covers the whole employment relationship. It calls into being a new common law—the common law of a particular industry or of a particular plant.’ ” Id., at 160-161 (citation omitted) (quoting Steelworkers v. Warrior & Gulf Navigation Co., 363 U. S., at 578-579). In this case, Conrail’s contractual claim rests solely upon implied contractual terms, as interpreted in light of past practice. Because we agree with Conrail that its contractual claim is neither frivolous nor obviously insubstantial, we conclude that this controversy is properly deemed a minor dispute within the exclusive jurisdiction of the Board. A The essential facts regarding Conrail’s past practices—the facts in support of the positions of both Conrail and the Union—are not disputed.9 Since its founding in 1976, Conrail routinely has required its employees to undergo physical examinations under the supervision of its health services department. The parties agreed in the Court of Appeals, and the District Court found, that Conrail’s authority to conduct physical examinations is an implied term of the collective-bargaining agreement, established by longstanding past practice and acquiesced in by the Union. Conrail conducts physical examinations in three categories of cases. First, it always has required its employees to un- 9 This is not to say that the legal significance of these practices is undisputed. In particular, the parties take different views of how a court is to determine whether a particular past practice has risen to the level of an implied contractual term. Compare Brief for Respondents 42-43 with Brief for Petitioner 19. The precise definition of this standard, however, is of no particular significance to this case. As will become clear, the parties have agreed that Conrail’s power to conduct physical examinations is an implied contractual term. The District Court made no factual findings that Conrail’s specific practices had themselves become implied terms of the contract, and we do not suggest otherwise in the discussion that follows. CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 313 299 Opinion of the Court dergo periodic physical examinations, which have routinely included a urinalysis for blood sugar and albumin. These periodic examinations are conducted every three years for employees up to the age of 50, and every two years thereafter. Second, Conrail has required train and engine employees who have been out of service for at least 30 days due to furlough, leave, suspension, or other similar cause to undergo return-to-duty physical examinations. These also routinely include urinalysis. Conrail employees in other job classifications are required to undergo return-to-duty physical examinations that include urinalysis for blood sugar and albumin, but are required to submit to examinations only after absences of 90 days or more. Third, when justified by the employee’s condition, Conrail has routinely required a follow-up physical examination. For example, such an examination has been required for an employee who has suffered a heart attack, or has been diagnosed as having hypertension or epilepsy. Any employee who undergoes a periodic, return-to-duty, or follow-up physical examination and who fails to meet Con-rail’s established medical standards may be held out of service without pay until the condition is corrected or eliminated. Conrail has implemented medical standards for all three types of physical examination. Over the years, procedures for hearing tests, lung-capacity tests, eye tests, and cardiological tests have been modified to reflect changes in medical science and technology. These changes have been made by Conrail unilaterally, without consulting the Union. Drug testing always has had some place in Conrail’s physical examinations, although its role has changed with time. Conrail has included drug testing by urinalysis as part of periodic physical examinations whenever, in the judgment of the examining physician, the employee may have been using drugs. Drug screens also routinely have been performed as part of the return-to-duty physical examination of any employee who has been taken out of service previously for a drug-related problem; in addition, drug testing is included 314 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. whenever the examining physician thinks the employee may have been using drugs. On April 1, 1984, Conrail issued a Medical Standards Manual stating that a drug screen would be included in all periodic and return-to-duty physicals. For budgetary reasons, however, this policy then was applied only in Conrail’s eastern region and was discontinued after six months. On February 20, 1987, Conrail implemented the Medical Standards Manual in all of its regions, requiring drug testing as part of its periodic and return-to-duty physicals and, in addition, requiring follow-up examinations for all employees returning to duty after disqualification for any reason associated with drug use.10 An employee who tests positive for drugs will not be returned to service unless he provides a negative drug test within 45 days of the date he receives notice of the positive test. An employee whose first test is positive may go to Conrail’s Employee Counseling Service for evaluation. If the evaluation reveals an addiction problem, and the employee agrees to enter an approved treatment program, the employee will be given an extended period of 125 days to provide a negative test. The problem of drug use has been addressed by Conrail not only as a medical concern, but also as a disciplinary one. This Court noted earlier in the present Term that the railroad industry has adopted operating “Rule G,” which governs drug use by employees. Skinner v. Railway Labor Executives’ Assn., 489 U. S. 602, 606-607 (1989). As currently implemented by Conrail, Rule G provides: “The use of intoxicants, narcotics, amphetamines or hallucinogens by employees subject to duty, or their possession or use while on duty, is prohibited. Employees under medication before or while on duty must be certain that such use will not affect the safe 10 The Union suggests that Conrail’s decision to implement its current drug-testing program resulted from a serious Conrail accident in January 1987, in which the engineer and conductor of the train admitted smoking marijuana in the cab just prior to the collision. Brief for Respondents 6. CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 315 299 Opinion of the Court performance of their duties.” See App. 63. At Conrail, as elsewhere in the industry, an employee may be dismissed for violating Rule G. Skinner, 489 U. S., at 607; Tr. of Oral Arg. 43. Conrail has relied chiefly on supervisory observation to enforce Rule G. An employee suspected of drug or alcohol use is encouraged voluntarily to agree to undergo diagnostic tests, but is not required to do so. In addition, Conrail has implemented the Federal Railroad Administration regulations recently upheld in Skinner against a Fourth Amendment challenge. Since March 1986, Conrail has required all employees covered by the Hours of Service Act, 45 U. S. C. §61 et seq., to undergo postaccident drug and alcohol testing, pursuant to 49 CFR §219 et seq. (1987).11 B The dispute between the parties focuses on the meaning of these past practices. Conrail argues that adding urinalysis drug testing to its periodic and retum-to-duty physicals is justified by the parties’ implied agreement regarding physical examinations, as indicated by their longstanding practice of permitting Conrail unilaterally to establish and change fitness-for-duty standards, to revise testing procedures, and to remove from service employees who are deemed unfit for duty under those standards and testing procedures.12 Conrail contends, specifically, that past practice reflects that drug use has been deemed relevant to job fitness, and that Conrail’s physicians have the discretion to utilize drug testing as part of their medical determination of job fitness. The expansion of drug testing in February 1987, Conrail argues, “It was the implementation of the Federal* Railroad Administration regulations that precipitated the instant lawsuit, Brief for Respondents 7, but no issue regarding Conrail’s implementation of those regulations is presently before us. 12 Conrail argued in the District Court that the parties’ implied agreement regarding Rule G enforcement justified its current drug-testing practice, but abandoned that position on appeal. See 845 F. 2d, at 1194. 316 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. represents no more than a diagnostic improvement in its medical procedures, similar to diagnostic improvements Con-rail unilaterally made in the past.13 The Union contends that, even using the “arguably justified” standard, “it is simply not plausible” to conclude that the parties’ agreement contemplated that Conrail had the authority to include drug screens in all routine physical examinations. The Union argues that Conrail has departed materially from the parties’ agreement, as reflected by Conrail’s past medical practice, in several respects. First, the Union states that past practice limited the use of drug testing in physical examinations to circumstances in which there was cause to believe the employee was using drugs; the current program, on the other hand, includes testing without cause. Second, in the Union’s view, Conrail’s general medical policy permits Conrail to remove an employee from active service until the employee’s physical condition improves, but does not permit Conrail to discharge an employee for failure to get well within a specified time; the current drug-testing program includes a fixed time limit, and results in discharge rather than removal from active service. Third, the Union contends that the expansion of drug testing constitutes, for the first time, regulation by Conrail of the private, off-duty conduct of its employees. In addition to pointing to these asserted departures from past practice, the Union argues that the absence of a “meeting of the minds” on the particulars of testing and confidentiality procedures renders untenable Conrail’s claim that the parties tacitly have agreed to Conrail’s current use of drug testing. Finally, the Union presents an alternative view of what Conrail has done: Conrail has expanded the disciplinary use of drug testing to employees not covered by the Federal Railroad Administration regulations, an expansion 13 We note that Conrail does not seek to rely on the 1984 limited implementation of routine drug testing as evidence of a past practice acquiesced in by the Union. See id., at 1193, n. 3. CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 317 299 Opinion of the Court which impermissibly adds drug testing to the list of available means for the enforcement of Rule G. C In the end, the Union’s arguments distinguishing drug testing from other aspects of Conrail’s medical program, and asserting that Conrail’s true motive is disciplinary, conceivably could carry the day in arbitration. But they do not convince us that Conrail’s contractual arguments are frivolous or insubstantial. Conrail’s interpretation of the range of its discretion as extending to drug testing is supported by the general breadth of its freedom of action in the past, and by its practice of including drug testing within routine medical examinations in some circumstances. In the past, the parties have left the establishment and enforcement of medical standards in Conrail’s hands. Conrail long has treated drug use as a matter of medical concern. Cf. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders 163-179 (3d ed. 1980) (substance abuse disorders); BN A Special Report, Alcohol & Drugs in the Workplace: Costs, Controls, and Controversies 1 (1986) (disciplinary and therapeutic approaches to drugs in the workplace); T. Denenberg & R. Denenberg, Alcohol & Drugs: Issues in the Workplace 18 (1983) (drug and alcohol abuse as treatable disorders); cf. Traynor v. Turnage, 485 U. S. 535, 562-564 (1988) (opinion concurring in part and dissenting in part) (alcohol dependence as medical problem). Indeed, although the scope of drug testing within physical examinations has changed over time, drug testing has always played some part (in appropriate circumstances) in Conrail’s medical examinations. In short, there is no established “rule” between the parties that drug use is solely a disciplinary, and never a medical, concern. There need be no “meeting of the minds” between the parties on the details of drug-testing methods or confidentiality standards for Conrail’s current drug-testing program argu 318 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ably to be justified by the parties’ agreement. As we have noted, labor laws do not require all the details of particular practices to be worked out in advance. Conrail’s claim that drug testing is an area in which Conrail retains a degree of discretion finds some support in the fact that the Union never before has intervened in the procedural details of Conrail’s drug testing: such testing has been performed—like other medical tests—according to standards unilaterally promulgated by Conrail. Thus, the absence of a specific agreement between the parties regarding testing procedures and confidentiality does not sufficiently undermine Conrail’s contractual claim to require that this dispute be classified as “major.” Conrail’s well-established recognition of the relevance of drug use to medical fitness substantially weakens the Union’s claim that Conrail now, for the first time, is engaging in medical testing that reveals facts about employees’ private off-duty conduct. Indeed, the fact that medical testing often detects physical problems linked to off-duty behavior makes it difficult to draw a bright line for jurisdictional purposes between testing which does, and that which does not, reflect upon private conduct. As to the relevance of “cause,” we do not doubt that there is a difference between Conrail’s past regime of limiting drug testing to circumstances in which there is cause to believe that the employee has used drugs and Conrail’s present policy of including drug tests in all routine physical examinations. Indeed, the difference between testing with and without cause perhaps could be of significance to arbitrators in deciding the merits of drug-testing disputes. See generally Denenberg & Denenberg, Drug Testing from the Arbitrator’s Perspective, 11 Nova L. Rev. 371, 387-392 (1987); Veglahn, What is a Reasonable Drug Testing Program?: Insight from Arbitration Decisions, 39 Lab. L. J. 688, 689-692 (1988). But under the RLA, it is not the role of the courts to decide the merits of the parties’ dispute. Our role is limited CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 319 299 Opinion of the Court to determining where the “arguably justified” line is to be drawn. For the limited purpose of determining whether Conrail’s claim of contractual right to change its medical testing procedures must be rejected as obviously insubstantial, that line cannot reasonably be drawn between testing for cause and testing without cause. As Conrail pointed out and urged at oral argument, “particularized suspicion” is not an accepted prerequisite for medical testing. Tr. of Oral Arg. 21. A physician’s decision to perform certain diagnostic tests is likely to turn not on the legal concept of “cause” or “individualized suspicion,” but rather on factors such as the expected incidence of the medical condition in the relevant population, the cost, accuracy, and inherent medical risk of the test, and the likely benefits of detection. In designing diagnostic-testing programs, some employers establish a set of basic tests that are to be administered to all employees, see generally M. Rothstein, Medical Screening of Workers 16-19 (1984), regardless of whether there is cause to believe a particular employee will test positive. It is arguably within Conrail’s range of discretion to alter its position on drug testing based on perceived changes in these variables. We turn next to the alleged disciplinary consequences of a positive drug test. It is clear that Conrail is not claiming a right, under its medical policy, to discharge an employee because of a single positive drug test, a right many railroads assert under Rule G. See Skinner, 489 U. S., at 607. Furthermore, an employee has the option of requesting a period of rehabilitative treatment. Thus, it is surely at least arguable that Conrail’s use of drug testing in physical examinations has a medical, rather than a disciplinary, goal. The fact that for drug problems, unlike other medical conditions, Conrail’s standards include a fixed time period in which the employee’s condition must improve does serve to distinguish Conrail’s drug policy from its response to other medical problems. Conrail has argued that it needs, for 320 OCTOBER TERM, 1988 White, J., concurring 491 U. S. medical purposes, to require employees who deny that they are drug dependent to demonstrate that they are capable of producing a drug-free sample at will. Tr. of Oral Arg. 13. In our view, that argument has sufficient merit to satisfy Conrail’s burden of demonstrating that its claim of contractual entitlement to set a time limit for successful recovery from drug problems is not frivolous. V Because we conclude that Conrail’s contractual arguments are not obviously insubstantial, we hold that the case before us constitutes a minor dispute that is within the exclusive jurisdiction of the Board. We make clear, however, that we go no further than to hold that Conrail has met the light burden of persuading this Court that its drug-testing practice is arguably justified by the implied terms of its collectivebargaining agreement. We do not seek to minimize any force in the Union’s arguments that the discretion afforded Conrail by the parties’ implied agreement, as interpreted in light of past practice, cannot be understood to extend this far. Thus, in no way do we suggest that Conrail is or is not entitled to prevail before the Board on the merits of the dispute. The judgment is reversed. It is so ordered. Justice White, concurring. I join the opinion and judgment of the Court. I add these remarks only to emphasize that the parties agree and the courts below held that giving physical examinations is a matter covered by an implied agreement between Conrail and the Union. The company claims that although instituting drug testing is a change in conditions, the implied contract authorizes the change. I agree that this claim has substance and that the dispute is a minor one for the Adjustment Board to resolve. If the Board decides that the company is wrong about its authority under the contract, the CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 321 299 Brennan, J., dissenting result will be that the company has sought a change in the contract without invoking the procedures applicable to major disputes. Justice Brennan, with whom Justice Marshall joins, dissenting. I would affirm the judgment of the Court of Appeals for the reasons stated by that court. The routine medical examinations Conrail relies on as precedent for its drug-testing program could result, at most, in an employee being held out of service until his or her health improved. Conrail would have us believe that, in accepting such medical testing, the Union (arguably) agreed to testing for use of an illegal substance that could result in the employee’s firing. It is unsurprising that the Union agreed to nonpunitive medical testing, and that it acquiesced in the employer making such unilateral changes in testing procedures as it determined were advisable on the basis of current medical technology. But it is inconceivable to me that in so doing the Union was also agreeing to the systematic, suspicionless testing, on such terms and in such manner as the employer alone prescribed, of all employees for evidence of criminal activity that, under the employer’s plan, could result in discharge.* Such a contention, in my view, is not “arguable”—it is frivolous. I agree with the Court of Appeals that “[u]ltimately, Conrail’s argument rests on the premise that testing urine for cannabis metabolites is no different in kind from testing urine for blood sugar. This *The Court rests its holding that the purpose of Conrail’s drug tests is—arguably—medical rather than disciplinary solely on the ground that Conrail will not discharge an employee on the basis of one positive drug test standing alone and that it will permit the employee “a period of rehabilitative treatment” prior to a second test. Ante, at 319. I do not agree that these factors even arguably bring Conrail’s drug-testing program within the realm of the existing medical examinations. Beyond this, however, I note that under the Court’s reasoning the outcome of the case should be different if the employer’s policy were indeed “to discharge an employee because of a single positive drug test.” Ibid. 322 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. ignores considerable differences in what is tested for and the consequences thereof.” 845 F. 2d 1187, 1194 (CA3 1988). It may be helpful to note what the general counsel of the National Labor Relations Board had to say in addressing the somewhat similar question whether, under the National Labor Relations Act, the addition of drug testing to a previously required physical examination constitutes a “substantial change in working conditions”: “In cases where an employer has an existing program of mandatory physical examinations for employees or applicants, an issue arises as to whether the addition of drug testing constitutes a substantial change in the employees’ terms and conditions of employment. In general, we conclude that it does constitute such a change. When conjoined with discipline, up to and including discharge, for refusing to submit to the test or for testing positive, the addition of a drug test substantially changes the nature and fundamental purpose of the existing physical examination. Generally, a physical examination is designed to test physical fitness to perform the work. A drug test is designed to determine whether an employee or applicant uses drugs, irrespective of whether such usage interferes with ability to perform work.” NLRB General Counsel’s Memorandum on Drug and Alcohol Testing, Memorandum GC 87-5 (Sept. 8,1987), reprinted in BNA Daily Labor Report, No. 184, pp. D-l, D-2 (Sept. 24, 1987) (emphasis in original). The general counsel similarly concluded that “a union’s acquiescence in a past practice of requiring applicants and/or current employees to submit to physical examinations that did not include drug testing . . . does not constitute a waiver of the union’s right to bargain over drug testing.” Ibid. Without suggesting that the NLRA question of a “substantial change in working conditions” is precisely the same as the one before us, I do think the general counsel has a better un- CONSOL. RAIL CORP. v. RAILWAY LABOR EXECUTIVES 323 299 Brennan, J., dissenting derstanding than does the Court of the relationship between drug testing and routine physical examinations. I respectfully dissent. 324 OCTOBER TERM, 1988 Syllabus 491 U. S. HEALY et al. V. THE BEER INSTITUTE et al. APPEAL FROM THE UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT No. 88-449. Argued March 28, 1989—Decided June 19, 1989* A Connecticut statute requires out-of-state shippers of beer to affirm that their posted prices for products sold to Connecticut wholesalers are, as of the moment of posting, no higher than the prices at which those products are sold in the bordering States of Massachusetts, New York, and Rhode Island. Appellees, a brewers’ trade association and major producers and importers of beer, filed suit against state officials in the District Court challenging the statute under the Commerce Clause. The court upheld the statute on the basis of Joseph E. Seagram & Sons, Inc. v. Hostetter, 384 U. S. 35. The Court of Appeals reversed, holding that the statute violated the Commerce Clause by controlling the prices at which out-of-state shippers could sell beer in other States, and that appellants’ argument that the statute was a proper exercise of the State’s regulatory authority under the Twenty-first Amendment was foreclosed by Brown-Forman Distillers Corp. v. New York State Liquor Authority, 476 U. S. 573. Held: Connecticut’s beer-price-affirmation statute violates the Commerce Clause. Pp. 335-343. (a) The statute has the impermissible practical effect of controlling commercial activity wholly outside Connecticut. By virtue of its interaction with the regulatory schemes of the border States, the statute requires out-of-state shippers to take account of their Connecticut prices in setting their border-state prices and restricts their ability to offer promotional and volume discounts in the border States, thereby depriving them of whatever competitive advantages they may possess based on the local market conditions in those States. Moreover, the short-circuiting of normal pricing decisions based on local conditions would be carried to a national scale if and when a significant group of States enacted contemporaneous affirmation statutes similar to Connecticut’s that linked instate prices to the lowest price in any State in the country. It is precisely such results that the Commerce Clause was meant to preclude. Brown-Forman, 476 U. S., at 579, 581-583; cf. CTS Corp. v. Dynamics Corp, of America, 481 U. S. 69, 88-89. Pp. 335-340. *Together with No. 88-513, Wine & Spirits Wholesalers of Connecticut, Inc. v. The Beer Institute et al., also on appeal from the same court. HEALY v. THE BEER INSTITUTE 325 324 Syllabus (b) The statute, on its face, also violates the Commerce Clause by discriminating against interstate commerce, since it applies only to brewers and shippers engaged in interstate commerce and not to those engaged solely in Connecticut sales, and since it is not justified by a valid purpose unrelated to economic protectionism. Pp. 340-341. (c) Appellants’ reliance on the Twenty-first Amendment as authorizing the statute regardless of its effect on interstate commerce is foreclosed by Brown-Forman, 476 U. S., at 585, which explicitly held that that Amendment does not immunize state laws from Commerce Clause attack where, as here, their practical effect is to regulate liquor sales in other States. Pp. 341-342. (d) Appellants’ reliance on Seagram, supra, to validate the statute is also foreclosed by Brown-Forman, 476 U. S., at 581-584, and n. 6, which strictly limited Seagram’s scope and removed the underpinnings of its Commerce Clause analysis. To the extent that it held that retrospective affirmation statutes do not facially violate the Commerce Clause, Seagram is no longer good law, since such statutes, like other affirmation statutes, have the inherent practical extraterritorial effect of regulating liquor prices in other States. Pp. 342-343. 849 F. 2d 753, affirmed. Blackmun, J., delivered the opinion of the Court, in which Brennan, White, Marshall, and Kennedy, JJ., joined, and in Parts I and IV of which Scalia, J., joined. Scalia, J., filed an opinion concurring in part and concurring in the judgment, post, p. 344. Rehnquist, C. J., filed a dissenting opinion, in which Stevens and O’Connor, JJ., joined, post, p. 345. Robert F. Vacchelli, Assistant Attorney General of Connecticut, argued the cause for appellants in both cases. With him on the briefs for appellants in No. 88-449 were Joseph I. Lieberman, former Attorney General, Clarine Nardi Riddle, Acting Attorney General, and Richard M. Sheridan, Assistant Attorney General. William A. Wechsler filed briefs for appellant in No. 88-513. Jeffrey I. Glekel argued the cause for appellees in both cases. With him on the brief were Jerome I. Chapman, Wayne C. Holcombe, William H. Allen, Timothy G. Reynolds, and Gary NatemanA tBriefs of amici curiae urging affirmance were filed for the Commonwealth of Pennsylvania by Morey M. Myers and Christopher A. Lewis; and 326 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Justice Blackmun delivered the opinion of the Court. The State of Connecticut requires out-of-state shippers of beer to affirm that their posted prices for products sold to Connecticut wholesalers are, as of the moment of posting, no higher than the prices at which those products are sold in the bordering States of Massachusetts, New York, and Rhode Island. In these appeals, we are called upon to decide whether Connecticut’s beer-price-affirmation statute violates the Commerce Clause.1 I Although appellees challenge Connecticut’s beer-price-affirmation statute as amended in 1984, this litigation has its roots in the 1981 version of Connecticut’s price-affirmation scheme. Having determined that the domestic retail price of beer was consistently higher than the price of beer in the three bordering States, and with the knowledge that, as a result, Connecticut residents living in border areas frequently crossed state lines to purchase beer at lower prices, Connecticut enacted a price-affirmation statute tying Connecticut beer prices to the prices charged in the border States. See United States Brewers Assn., Inc. v. Healy, 532 F. Supp. 1312, 1314, 1316-1317 (Conn. 1982). In an effort to eliminate the price differential between Connecticut and the border States, Connecticut required that brewers and importers (out-of-state shippers)2 post bottle, can, and case prices for for the Distilled Spirits Council of the United States, Inc., by Arnold M. Lerman and Louis R. Cohen. 1 The Commerce Clause states: “The Congress shall have Power ... To regulate Commerce . . . among the several States . . . .” U. S. Const., Art. I, §8, cl. 3. This Court long has recognized that this affirmative grant of authority to Congress also encompasses an implicit or “dormant” limitation on the authority of the States to enact legislation affecting interstate commerce. See, e. g., Hughes v. Oklahoma, 441 U. S. 322, 326, and n. 2 (1979); H. P. Hood & Sons, Inc. v. Du Mond, 336 U. S. 525, 534-535 (1949). 2 The Connecticut beer industry is divided into three marketing levels: (1) brewers and importers, (2) wholesalers, and (3) retailers. Participants HEALY v. THE BEER INSTITUTE 327 324 Opinion of the Court each brand of beer to be sold in Connecticut. Id., at 1317. These posted prices would take effect on the first day of the following month and would continue without change for the rest of that month. Conn. Gen. Stat. Ann. § 30-63(c) (1975 and Supp. 1982). The 1981 statute further required that out-of-state shippers affirm under oath at the time of posting that their posted prices were and would remain no higher than the lowest prices they would charge for each beer product in the border States during the effective period. §30-63b(b), quoted in 532 F. Supp., at 1314, n. 3. Moreover, in calculating the lowest price offered in the border States, the statute deducted from the reported price the value of any rebates, discounts, special promotions, or other inducements that the out-of-state shippers offered in one or more of the border States.3 §30-63c(b), quoted in 532 F. Supp., at 1314, n. 4. To the extent that such inducements lowered border-state prices, the statute thus obligated out-of-state shippers to lower their Connecticut prices as well.4 In 1982, a brewers’ trade association and various beer producers and importers (a subset of the appellees in the instant litigation) filed suit in the United States District Court for the District of Connecticut, challenging the 1981 statute as in each tier of the industry must obtain a license to sell to the tier below, with the retailers selling to the consuming public. While generally each wholesaler carries the products of more than one brewer or importer (because Connecticut currently has no brewery of its own, brewers and importers are referred to. collectively as “out-of-state shippers”), wholesalers may resell these products only to retailers within the geographic area specified in their respective licenses. United States Brewers Assn. v. Healy, 669 F. Supp. 543, 545-546 (Conn. 1987); United States Brewers Assn., Inc. v. Healy, 532 F. Supp. 1312, 1317 (Conn. 1982). 3 The affirmation statute did permit differentials in price based on differing state taxes and transportation costs. Conn. Gen. Stat. § 30-63c(b) (1989). 4 The statute also required out-of-state shippers to offer Connecticut wholesalers every package configuration for each brand of beer offered to wholesalers in the border States. §30-63c(b), quoted in 532 F. Supp., at 1314, n. 4. 328 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. unconstitutional under the Commerce Clause. The District Court, relying primarily on this Court’s decision in Joseph E. Seagram & Sons, Inc. v. Hostetter, 384 U. S. 35 (1966), upheld the 1981 law. United States Brewers Assn., Inc. v. Healy, 532 F. Supp., at 1325-1326. The Court of Appeals, however, reversed. It held that the 1981 Connecticut statute was facially invalid under the Commerce Clause because it had the practical effect of prohibiting out-of-state shippers from selling beer in any neighboring State in a given month at a price below what it had posted in Connecticut at the start of that month. The court explained: “Nothing in the Twenty-first Amendment permits Connecticut to set the minimum prices for the sale of beer in any other state, and well-established Commerce Clause principles prohibit the state from controlling the prices set for sales occurring wholly outside its territory.” United States Brewers Assn., Inc. v. Healy, 692 F. 2d 275, 282 (CA2 1982) (Healy I). This Court summarily affirmed. 464 U. S. 909 (1983). In 1984, the Connecticut Legislature responded to Healy I by amending its beer-price-affirmation statute to its current form. The statute now requires out-of-state shippers to affirm that their posted prices are no higher than prices in the border States only at the time of the Connecticut posting. Conn. Gen. Stat. §30-63b(b) (1989).5 The legislature also 5 As amended by 1984 Conn. Pub. Acts 332, §30-63b(b) provides: “At the time of posting of the bottle, can, keg or barrel and case price required by section 30-63, every holder of a manufacturer or out-of-state shipper’s permit, or the authorized representative of a manufacturer, shall file with the department of liquor control a written affirmation under oath by the manufacturer or out-of-state shipper of each brand of beer posted certifying that, at the time of posting, the bottle, can or case price, or price per keg, barrel or fractional unit thereof, to the wholesaler permittees is no higher than the lowest price at which each such item of beer is sold, offered for sale, shipped, transported or delivered by such manufacturer or out-of-state shipper to any wholesaler in any state bordering this state.” In addition, Connecticut regulations now provide for posting on the sixth day of each month. App. 157. HEALY v. THE BEER INSTITUTE 329 324 Opinion of the Court added §30-63b(e), which provides that nothing in §30-63b prohibits out-of-state shippers from changing their out-of-state prices after the affirmed Connecticut price is posted.6 The legislature, however, did not amend §30-63a(b), which continued to make it unlawful for out-of-state shippers to sell beer in Connecticut at a price higher than the price at which beer is or would be sold in any bordering State during the month covered by the posting.7 In the wake of the 1984 amendments, appellees (a brewers’ trade association and major producers and importers of beer) filed suit in the United States District Court for the District of Connecticut, seeking declaratory and injunctive relief and claiming that the effect of the amended law was not different from that of the law struck down in Healy I.8 See United States Brewers Assn. v. Healy, 669 F. Supp. 543, 544-545 (1987). In response to appellees’ complaint, Connecticut filed a “Declaratory Ruling” by the Department of Liquor Control, interpreting the statute as amended as requiring out-of-state shippers to affirm that their posted prices in Connecticut were no higher than their lowest prices in any 6 As added by 1984 Conn. Pub. Acts 332, § 30-63b(e) provides: “This section shall not prohibit a manufacturer or out-of-state shipper permittee or the authorized representative of a manufacturer from changing prices to any wholesaler in any other state of the United States or in the District of Columbia, or to any state or agency of a state which owns and operates retail liquor outlets at any time during the calendar month covered by such posting.” 7 Conn. Gen. Stat. §30-63a(b) provides in relevant part: “No holder of any manufacturer or out-of-state shipper’s permit shall ship, transport or deliver within this state, or sell or offer for sale to a wholesaler permittee any brand of beer ... at a bottle, can or case price, or price per keg, barrel or fractional unit thereof, higher than the lowest price at which such item is then being sold or offered for sale or shipped, transported or delivered by such manufacturer or out-of-state shipper to any wholesaler in any state bordering this state.” 8 Appellants are the Connecticut officials responsible for enforcing the affirmation statute, and the liquor-wholesalers trade association which entered the case as an intervenor. 330 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. border State only at the time of posting—the sixth day of each month. Id., at 547, and n. 9. After the moment of posting, the ruling stated, the statute imposes no restrictions on the right of out-of-state shippers to raise or lower their border-state prices at will. Ibid. Appellees argued, however, that the Connecticut beeraffirmation statute, even as modified by the declaratory ruling, regulated out-of-state transactions, constituted economic protectionism, and unduly burdened interstate commerce, all in violation of the Commerce Clause. On cross-motions for summary judgment, the District Court upheld the statute as modified by the legislature and construed in the Department of Liquor Control’s declaratory ruling, resting its decision on Seagram, supra, and distinguishing this Court’s subsequent decision in Brown-Forman Distillers Corp. v. New York State Liquor Authority, 476 U. S. 573 (1986), which struck down a statute analogous to Connecticut’s 1981 beer-affirmation statute. The District Court found the 1984 Connecticut law constitutional on its face because, “unlike the version in Healy I and Brown-Forman,” the 1984 law “leaves brewers free to raise or lower prices in the border states before and after posting in Connecticut and does not, therefore, regulate interstate commerce.” 669 F. Supp., at 553. As in Healy I, the Court of Appeals reversed. It held that the 1984 law (even as interpreted by the declaratory ruling), like its predecessor, violated the Commerce Clause by controlling the prices at which out-of-state shippers could sell beer in other States. First, and foremost, the court held that the Connecticut statute’s “purposeful interaction with border-state regulatory schemes” means that shippers cannot, as a practical matter, set prices based on market conditions in a border State without factoring in the effects of those prices on its future Connecticut pricing options. In re Beer Institute, 849 F. 2d 753, 760-761 (CA2 1988) (Healy II). Second, the Court of Appeals found that the 1984 statute unconstitutionally restricted the ability of out-of-state shippers HEALY v. THE BEER INSTITUTE 331 324 Opinion of the Court to offer volume discounts in the border States. Id., at 760. Furthermore, relying on Brown-Forman, supra, the court rejected appellants’ argument that the statute was a proper exercise of its regulatory authority under the Twenty-first Amendment. 849 F. 2d, at 761. We noted probable jurisdiction. 488 U. S. 954 (1988). II In deciding this appeal, we engage in our fourth expedition into the area of price-affirmation statutes. The Court first explored this territory in Seagram, where it upheld against numerous constitutional challenges a New York statute that required liquor-label owners or their agents to affirm that “ ‘the bottle and case price of liquor ... is no higher than the lowest price’ ” at which such liquor was sold “anywhere in the United States during the preceding month.” 384 U. S., at 39-40, quoting the New York law. The Court ruled that the mere fact that the New York statute was geared to appellants’ pricing policies in other States did not violate the Commerce Clause, because under the Twenty-first Amendment’s broad grant of liquor regulatory authority to the States, New York could insist that liquor prices offered to domestic wholesalers and retailers “be as low as prices offered elsewhere in the country.” Id., at 43. Although the appellant brand owners in Seagram had alleged that the New York law created serious discriminatory effects on their business outside New York, the Court considered these injuries too conjectural to support a facial challenge to the statute and suggested that the purported extraterritorial effects could be assessed in a case where they were clearly presented. Ibid. Eighteen years after Seagram, we summarily affirmed the Second Circuit’s judgment in Healy I, and then, another two years later, granted plenary review in Brown-Forman, supra. The New York law at issue in Brown-Forman required every liquor distiller or producer selling to wholesalers within the State to affirm that the prices charged for 332 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. every bottle or case of liquor were no higher than the lowest price at which the same product would be sold in any other State during the month covered by the particular affirmation. 476 U. S., at 576. Appellant Brown-Forman was a liquor distiller that offered “promotional allowances” to wholesalers purchasing Brown-Forman products. The New York Liquor Authority, however, did not allow Brown-Forman to operate its rebate scheme in New York and, moreover, determined for the purposes of the affirmation law that the promotional allowances lowered the effective price charged to wholesalers outside New York. Because other States with affirmation laws similar to New York’s did not deem the promotional allowances to lower the price charged to wholesalers, appellant argued that the New York law offered the company the Hobson’s choice of lowering its New York prices, thereby violating the affirmation laws of other States, or of discontinuing the promotional allowances altogether. This, appellant alleged, amounted to extraterritorial regulation of interstate commerce in violation of the Commerce Clause. Id., at 579-582. This Court agreed, reaffirming and elaborating on our established view that a state law that has the “practical effect” of regulating commerce occurring wholly outside that State’s borders is invalid under the Commerce Clause. We began by reviewing past decisions, starting with Baldwin n. G. A. F. Seelig, Inc., 294 U. S. 511 (1935). The Court in Seelig struck down a New York statute that set minimum prices for milk purchased from producers in New York and other States and banned the resale within New York of milk that had been purchased for a lower price. Because Vermont dairy farmers produced milk at a lower cost than New York dairy farmers, the effect of the statute was to eliminate the competitive economic advantage they enjoyed by equalizing the price of milk from all sources. Writing for the Court, Justice Cardozo pronounced that the Commerce Clause does not permit a State “to establish a wage scale or a scale of prices for use HEALY v. THE BEER INSTITUTE 333 324 Opinion of the Court in other states, and to bar the sale of the products . . . unless the scale has been observed.” Id., at 528. Relying on Seelig, the Court in Brown-Forman concluded: “While a State may seek lower prices for its consumers, it may not insist that producers or consumers in other States surrender whatever competitive advantages they may possess.” 476 U. S., at 580; see also Schwegmann Brothers Giant Super Markets v. Louisiana Milk Common, 365 F. Supp. 1144, 1152-1156 (MD La. 1973), summarily aff’d, 416 U. S. 922 (1974). After drawing upon Seelig, the Brown-Forman Court also discussed Healy I with approval. There, as we have noted, the Court of Appeals struck down an earlier version of Connecticut’s price-affirmation statute, which was essentially identical to the one at issue in Brown-Forman, because the statute “made it impossible for a brewer to lower its price in a bordering State in response to market conditions so long as it had a higher posted price in effect in Connecticut.” 476 U. S., at 581-582.9 9 The Brown-Forman Court cited a third extraterritorial decision, Edgar n. MITE Corp., 457 U. S. 624 (1982), which, though not discussed at length there, significantly illuminates the contours of the constitutional prohibition on extraterritorial legislation. In MITE Corp., the Court struck down the Illinois Business Takeover Act, which required that a takeover offer for a target company having a specified connection to Illinois be registered with the Secretary of State and mandated that such an offer was not to become effective for 20 days, during which time the offer would be subject to administrative evaluation. The statute empowered the Secretary of State to deny registration of the tender offer under certain conditions, such as inequity or fraud. A plurality found the statute to be infirm under the Commerce Clause because it “directly regulates transactions which take place across state lines, even if wholly outside the State of Illinois.” Id., at 641. The plurality observed that, if the target company had sufficient in-state contacts, the Illinois law, unless complied with, could prevent interstate-securities transactions in stock even if not a single one of the target company’s shareholders was a resident of Illinois. Moreover, the plurality noted that if Illinois were free to enact such legislation, others States similarly were so empowered, “and interstate commerce in securities transactions generated by tender offers would be thoroughly stifled.” Id., at 642. Under the Commerce Clause the projection of these extrater 334 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Applying these principles, we concluded that the New York statute had an impermissible extraterritorial effect: “Once a distiller has posted prices in New York, it is not free to change its prices elsewhere in the United States during the relevant month. Forcing a merchant to seek regulatory approval in one State before undertaking a transaction in another directly regulates interstate commerce.” Id., at 582 (footnote omitted). Although New York might regulate the sale of liquor within its borders, and might seek low prices for its residents, it was prohibited by the Commerce Clause from “‘projecting] its legislation into [other States] by regulating the price to be paid’” for liquor in those States. Id., at 583, quoting Seelig, 294 U. S., at 521. Despite the language in Seagram, the Court did not find the prospect of these extraterritorial effects to be speculative. The majority rejected as “Pollyannaish” the dissent’s suggestion that flexible application by the relevant administrative bodies would obviate the problem and noted that the proliferation of affirmation laws after Seagram had greatly multiplied the likelihood that distillers would be subject to blatantly inconsistent obligations.10 The Court squarely rejected New York’s argument that the Twenty-first Amendment, which bans the importation or possession of intoxicating liquors into a State “in violation of the laws thereof,” saved the statute from invalidation under the Commerce Clause. Although the Court acknowledged that the Amendment vested in New York considerable au- ritorial “‘practical effect[s],”’ regardless of the statute’s intention, “‘ex-ceed[ed] the inherent limits of the State’s power.’ ” Id., at 642-643, quoting Shaffer v. Heitner, 433 U. S. 186, 197 (1977). 10 At the time of our decision in Brown-Forman, 39 States, including New York, had adopted affirmation laws. Of these, 18, known as “control” States, each purchased all liquor to be distributed and consumed within its borders. These States subscribed to a standard sales contract that required distillers to guarantee that the price charged the State was no higher than the lowest price offered anywhere in the United States. Twenty States had adopted statutes similar to the New York statute that was under challenge. See 476 U. S., at 576, and n. 1. HEALY v. THE BEER INSTITUTE 335 324 Opinion of the Court thority to regulate the domestic sale of alcohol, the Amendment did not immunize the State from the Commerce Clause’s proscription of state statutes that regulate the sale of alcohol in other States. 476 U. S., at 585. Accordingly, the Court’s conclusion that the New York law regulated out-of-state sales conclusively resolved the Twenty-first Amendment issue against New York. Ibid. The Court acknowledged that its Brown-Forman decision was in considerable tension with Seagram. The statutes at issue in the two cases were, it observed, factually distinguishable: the Seagram statute was retrospective, tying New York prices to out-of-state prices charged during the previous month, while the Brown-Forman statute was prospective, mandating that New York prices could be no higher than out-of-state prices for the following month. But the Court explicitly refused to give this retrospective/prospective distinction any constitutional significance, and even suggested that the effects of the two statutes might well be the same for the purposes of constitutional analysis. Nonetheless, since the Court was not squarely presented with a retrospective statute, it declined to evaluate Seagram’s continued validity. 476 U. S., at 584, n. 6.11 III In light of this history, we now must assess the constitutionality of the Connecticut statute, which is neither prospective nor retrospective, but rather “contemporaneous.” As explained above, the statute requires only that out-of-state shippers affirm that their prices are no higher than the prices being charged in the border States as of the moment of affirmation. The principles guiding this assessment, principles made clear in Brown-Forman and in the cases upon which it relied, reflect the Constitution’s special concern both with the main 11 One Member of the Court concurred separately to advocate that Seagram then be overruled as a “relic of the past.” Id., at 586. 336 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. tenance of a national economic union unfettered by state-imposed limitations on interstate commerce12 and with the autonomy of the individual States within their respective spheres.13 Taken together, our cases concerning the extraterritorial effects of state economic regulation stand at a minimum for the following propositions: First, the “Commerce Clause . . . precludes the application of a state statute to commerce that takes place wholly outside of the State’s borders, whether or not the commerce has effects within the State,” Edgar v. MITE Corp., 457 U. S. 624, 642-643 (1982) (plurality opinion); see also Brown-Forman, 476 U. S., at 581-583; and, specifically, a State may not adopt legislation that has the practical effect of establishing “a scale of prices for use in other states,” Seelig, 294 U. S., at 528. Second, a statute that directly controls commerce occurring wholly outside the boundaries of a State exceeds the inherent limits of the enacting State’s authority and is invalid regardless of whether the statute’s extraterritorial reach was intended by the legislature. The critical inquiry is whether the practical effect of the regulation is to control conduct beyond the boundaries of the State. Brown-Forman, 476 U. S., at 579. Third, the practical effect of the statute must be evaluated not only by considering the consequences of the statute itself, but also by considering how the challenged statute may interact with the legitimate regulatory regimes of other States and what effect would arise if not one, but many or every, State adopted similar legislation. Generally speaking, the 12 The entire Constitution was “framed upon the theory that the peoples of the several states must sink or swim together, and that in the long run prosperity and salvation are in union and not division.” Baldwin n. G. A. F. Seelig, Inc., 294 U. S. 511, 523 (1935). 13 The plurality in Edgar n. MITE Corp, noted: “The limits on a State’s power to enact substantive legislation are similar to the limits on the jurisdiction of state courts. In either case, ‘any attempt “directly” to assert extraterritorial jurisdiction over persons or property would offend sister States and exceed the inherent limits of the State’s power.’ ” 457 U. S., at 643, quoting Shaffer v. Heitner, 433 U. S., at 197. HEALY v. THE BEER INSTITUTE 337 324 Opinion of the Court Commerce Clause protects against inconsistent legislation arising from the projection of one state regulatory regime into the jurisdiction of another State. Cf. CTS Corp. v. Dynamics Corp, of America, 481 U. S. 69, 88-89 (1987). And, specifically, the Commerce Clause dictates that no State may force an out-of-state merchant to seek regulatory approval in one State before undertaking a transaction in another. Brown-Forman, 476 U. S., at 582.14 When these principles are applied to Connecticut’s contemporaneous price-affirmation statute, the result is clear. The Court of Appeals correctly concluded that the Connecticut statute has the undeniable effect of controlling commercial activity occurring wholly outside the boundary of the State. Moreover, the practical effect of this affirmation law, in conjunction with the many other beer-pricing and affirmation laws that have been or might be enacted throughout the country, is to create just the kind of competing and interlocking local economic regulation that the Commerce Clause was meant to preclude. First, as explained by the Court of Appeals, the interaction of the Connecticut affirmation statute with the Massa 14 As a general matter, the Court has adopted a two-tiered approach to analyzing state economic regulation under the Commerce Clause. We summarized in Brown-Forman: “When a state statute directly regulates or discriminates against interstate commerce, or when its effect is to favor in-state economic interests over out-of-state interests, we have generally struck down the statute without further inquiry. When, however, a statute has only indirect effects on interstate commerce and regulates evenhandedly, we have examined whether the State’s interest is legitimate and whether the burden on interstate commerce clearly exceeds the local benefits.” 476 U. S., at 579 (citations omitted). We further recognized in Brown-Forman that the critical consideration in determining whether the extraterritorial reach of a statute violates the Commerce Clause is the overall effect of the statute on both local and interstate commerce. Ibid. Our distillation of principles from prior cases involving extraterritoriality is meant as nothing more than a restatement of those specific concerns that have shaped this inquiry. 338 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. chusetts beer-pricing statute (which does not link domestic prices with out-of-state prices) has the practical effect of controlling Massachusetts prices. See 849 F. 2d, at 759. Massachusetts requires brewers to post their prices on the first day of the month to become effective on the first day of the following month. See Mass. Gen. Laws § 138:25B (1986). Five days later, however, those same brewers, in order to sell beer in Connecticut, must affirm that their Connecticut prices for the following month will be no higher than the lowest price that they are charging in any border State. Accordingly, on January 1, when a brewer posts his February prices for Massachusetts, that brewer must take account of the price he hopes to charge in Connecticut during the month of March. Not only will the January posting in Massachusetts establish a ceiling price for the brewer’s March prices in Connecticut, but also, under the requirements of the Massachusetts law, the brewer will be locked into his Massachusetts price for the entire month of February (absent administrative leave) even though the Connecticut posting will have occurred on February 6. Thus, as a practical matter, Connecticut’s nominally “contemporaneous” affirmation statute “prospectively” precludes the alteration of out-of-state prices after the moment of affirmation. More generally, the end result of the Connecticut statute’s incorporation of out-of-state prices, as the Court of Appeals concluded, is that “[a] brewer can . . . undertake competitive pricing based on the market realities of either Massachusetts or Connecticut, but not both, because the Connecticut statute ties pricing to the regulatory schemes of the border states.” 849 F. 2d, at 759. In other words, the Connecticut statute has the extraterritorial effect, condemned in Brown-Forman, of preventing brewers from undertaking competitive pricing in Massachusetts based on prevailing market conditions. Second, because New York law requires that promotional discounts remain in effect for 180 days, see N. Y. Alco. Bev. Cont. Law §55-b(2) (McKinney 1987), and the Connecticut HEALY v. THE BEER INSTITUTE 339 324 Opinion of the Court statute treats promotional discounts as a reduction in price, the interaction of the New York and Connecticut laws is such that brewers may offer promotional discounts in New York only at the cost of locking in their discounted New York price as the ceiling for their Connecticut prices for the full 180 days of the New York promotional discount. Third, because volume discounts are permitted in Massachusetts, New York, and Rhode Island, but not in Connecticut, the effect of Connecticut’s affirmation scheme is to deter volume discounts in each of these other States, because the lowest of the volume-discounted prices would have to be offered as the regular price for an entire month in Connecticut. See 849 F. 2d, at 760. With respect to both promotional and volume discounts, then, the effect of the Connecticut statute is essentially indistinguishable from the extraterritorial effect found unconstitutional in Brown-Forman. The Connecticut statute, like the New York law struck down in Brown-Forman, requires out-of-state shippers to forgo the implementation of competitive-pricing schemes in out-of-state markets because those pricing decisions are imported by statute into the Connecticut market regardless of local competitive conditions. As we specifically reaffirmed in Brown-Forman, States may not deprive businesses and consumers in other States of “whatever competitive advantages they may possess” based on the conditions of the local market. 476 U. S., at 580. The Connecticut statute does precisely this. The Commerce Clause problem with the Connecticut statute appears in even starker relief when it is recalled that if Connecticut may enact a contemporaneous affirmation statute, so may each of the border States and, indeed, so may every other State in the Nation. Suppose, for example, that the border States each enacted statutes essentially identical to Connecticut’s. Under those circumstances, in January, when a brewer posts his February prices in Connecticut and the border States, he must determine those prices knowing 340 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. that the lowest bottle, can, or case price in any State would become the maximum bottle, can, or case price the brewer would be permitted to charge throughout the region for the month of March. This is true because in February, when the brewer posts his March prices in each State, he will have to affirm that no bottle, can, or case price is higher than the lowest bottle, can, or case price in the region—and these “current” prices would have been determined by the January posting. Put differently, unless a beer supplier declined to sell in one of the States for an entire month, the maximum price in each State would be capped by previous prices in the other State. This maximum price would almost surely be the minimum price as well, since any reduction in either State would permanently lower the ceiling in both. Nor would such “price gridlock” be limited to individual regions. The short-circuiting of normal pricing decisions based on local conditions would be carried to a national scale if a significant group of States enacted contemporaneous affirmation statutes that linked in-state prices to the lowest price in any State in the country. This kind of potential regional and even national regulation of the pricing mechanism for goods is reserved by the Commerce Clause to the Federal Government and may not be accomplished piecemeal through the extraterritorial reach of individual state statutes. IV The Connecticut statute, moreover, violates the Commerce Clause in a second respect: On its face, the statute discriminates against brewers and shippers of beer engaged in interstate commerce. In its previous decisions, this Court has followed a consistent practice of striking down state statutes that clearly discriminate against interstate commerce, see, e. g., New Energy Co. of Indiana n. Limbach, 486 U. S. 269 (1988); Sporhase n. Nebraska ex rel. Douglas, 458 U. S. 941 (1982); Lewis v. BT Investment Managers, Inc., 447 U. S. 27 (1980), unless that discrimination is demonstrably HEALY v. THE BEER INSTITUTE 341 324 Opinion of the Court justified by a valid factor unrelated to economic protectionism, see, e. g., Maine v. Taylor, 477 U. S. 131 (1986). By its plain terms, the Connecticut affirmation statute applies solely to interstate brewers or shippers of beer, that is, either Connecticut brewers who sell both in Connecticut and in at least one border State or out-of-state shippers who sell both in Connecticut and in at least one border State. Under the statute, a manufacturer or shipper of beer is free to charge wholesalers within Connecticut whatever price it might choose so long as that manufacturer or shipper does not sell its beer in a border State. This discriminatory treatment establishes a substantial disincentive for companies doing business in Connecticut to engage in interstate commerce, essentially penalizing Connecticut brewers if they seek border-state markets and out-of-state shippers if they choose to sell both in Connecticut and in a border State. We perceive no neutral justification for this patent discrimination. Connecticut has claimed throughout this litigation that its price-affirmation laws are designed to ensure the lowest possible prices for Connecticut consumers. While this may be a legitimate justification for the statute, it is not advanced by, in effect, exempting brewers and shippers engaging in solely domestic sales from the price regulations imposed on brewers and shippers who engage in sales throughout the region. V A Appellants advance two basic arguments in defense of Connecticut’s statute: first, that the Twenty-first Amendment sanctions Connecticut’s affirmation statute regardless of its effect on interstate commerce; and, second, that the statute is constitutional under this Court’s analysis in Joseph E. Seagram & Sons, Inc. v. Hostetter, 384 U. S. 35 (1966), in which the Court stated that a retrospective affirmation statute does not violate the Commerce Clause merely because it is geared to prices in other States. Appellants’ reliance 342 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. on the Twenty-first Amendment is foreclosed by Brown-Forman, where we explicitly rejected an identical argument. In Brown-Forman, the Court specifically held that the Twenty-first Amendment does not immunize state laws from invalidation under the Commerce Clause when those laws have the practical effect of regulating liquor sales in other States. 476 U. S., at 585. Here, as in Brown-Forman, our finding of unconstitutional extraterritorial effects disposes of the Twenty-first Amendment issue. Appellants’ reliance on Seagram is similarly foreclosed by Brown-Forman. While our decision in Brown-Forman did not overrule Seagram, it strictly limited the scope of that decision to retrospective affirmation statutes. B More important, Brown-Forman removed the legal underpinnings of Seagram’s Commerce Clause analysis. 476 U. S., at 581-584, and n. 6. Seagram rested on the following reasoning: the Twenty-first Amendment gives States wide latitude in the field of liquor regulation; although such state regulation might violate the Commerce Clause in some extreme instances, in particular where a State’s regulations controlled liquor commerce outside the State’s boundaries, the extraterritorial effects of New York’s retrospective affirmation statute were too conjectural to support such a claim. 384 U. S., at 42-43. Brown-Forman, however, holds unequivocally that to the extent that an affirmation statute has the practical effect of regulating out-of-state liquor prices, it cannot stand under the Commerce Clause irrespective of the Twenty-first Amendment. 476 U. S., at 585. In striking down the statute at issue, the Court in Brown-Forman found, in light of 20 years of experience with the affirmation laws that proliferated after Seagram, that prospective affirmation statutes have such extraterritorial effects. Indeed, Brown-Forman leaves Seagram intact only to the extent that the Court in the former case felt no compulsion, in a case not directly raising the question, to address whether HEALY v. THE BEER INSTITUTE 343 324 Opinion of the Court retrospective affirmation shared the extraterritorial effects of prospective affirmation laws. 476 U. S., at 584, n. 6. In the interest of removing any lingering uncertainty about the constitutional validity of affirmation statutes and of avoiding further litigation on the subject of liquor-price affirmation, we recognize today what was all but determined in Brown-Forman: to the extent that Seagram holds that retrospective affirmation statutes do not facially violate the Commerce Clause, it is no longer good law. Retrospective affirmation statutes, like other affirmation statutes, have the inherent practical extraterritorial effect of regulating liquor prices in other States. By tying maximum future prices in one State to the lowest prices in other States as determined at a specified time in the past, retrospective affirmation laws control pricing decisions in nonaffirmation States by requiring that those decisions reflect not only local market conditions, but also market conditions in the affirmation States — market conditions that would be irrelevant absent the binding force of the affirmation statutes. Every pricing decision made in a nonaffirmation State will reflect the certain knowledge that the price chosen will become in the future the maximum permissible price in the States requiring affirmation.15 For the reasons noted today and in Brown-Forman, this extraterritorial effect violates the Commerce Clause. The judgment of the Court of Appeals is affirmed. It is so ordered. 13 Recent economic scholarship confirms: “[B]oth [prospective and retrospective] types of price affirmation burden interstate commerce because they both cause firms to consider jointly their demand and marginal cost curves in more than one state. Accordingly, the impact of an affirmation law adopted by one state will be transmitted to other states, affecting prices charged in those other states in the process.” Pustay & Zardkoohi, An Economic Analysis of Liquor Price Affirmation Laws: Do They Burden Interstate Commerce?, 48 La. L. Rev. 649, 673-674 (1988). 344 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. Justice Scalia, concurring in part and concurring in the judgment. I join the Court’s disposition of this suit and Parts I and IV of its opinion. The Connecticut statute’s invalidity is fully established by its facial discrimination against interstate commerce—through imposition of price restrictions exclusively upon those who sell beer not only in Connecticut but also in the surrounding States—and by Connecticut’s inability to establish that the law’s asserted goal of lower consumer prices cannot be achieved in a nondiscriminatory manner.* See New Energy Co. of Indiana n. Limbach, 486 U. S. 269, 276-277, 279-280 (1988). This is so despite the fact that the law regulates the sale of alcoholic beverages, since its discriminatory character eliminates the immunity afforded by the Twenty-first Amendment. See Bacchus Imports, Ltd. v. Dias, 468 U. S. 263, 275-276 (1984). Since Joseph E. Seagram & Sons, Inc. v. Hostetter, 384 U. S. 35 (1966), upheld a law that operated in like fashion, I agree with the Court that today’s decision requires us to overrule that case. See ante, at 343. *The dissent argues that the facial discrimination inherent in the present statute does not establish its invalidity because no brewer does business solely in Connecticut and because there is no evidence that any shipper sells beer exclusively within that State. Post, at 348. As far as I know we have never required a plaintiff to show that a statute which facially discriminates against out-of-state business in fact benefits a particular in-state business, and we have flatly rejected the kindred contention that the plaintiff could not prevail if the benefit to in-state business was minimal, see New Energy Co. of Indiana v. Limbach, 486 U. S. 269, 276-277 (1988). It would make little sense to require a showing that an instate business in fact exists without also requiring a showing that it is in fact benefited. I see no reason to impose such a burden in order to strike down a statute that is facially discriminatory under the Commerce Clause, any more than we would require the person challenging under the Fourteenth Amendment a state law permitting only Aleuts to vote by mail to show that there are in fact Aleut citizens of the State capable of benefiting from that discrimination. HEALY v. THE BEER INSTITUTE 345 324 Rehnquist, C. J., dissenting I would refrain, however, from applying the more expansive analysis which finds the law unconstitutional because it regulates or controls beer pricing in the surrounding States. See ante, at 335-340. It seems to me this rationale is not only unnecessary but also questionable, resting as it does upon the mere economic reality that the challenged law will require sellers in New York, Massachusetts, and Rhode Island to take account of the price that they must post and charge in Connecticut when setting their prices in those other States. The difficulty with this is that innumerable valid state laws affect pricing decisions in other States—even so rudimentary a law as a maximum price regulation. Suppose, for example, that the Connecticut Legislature had simply provided that beer could not be retailed in Connecticut above $10 a case. Sellers in those portions of New York, Massachusetts, and Rhode Island bordering Connecticut would have to take account of that requirement, just as sellers in those States had to take account of the Connecticut posting requirement here, because prices substantially above the maximum would cause their former in-state purchasers to drive to Connecticut and their former Connecticut purchasers to stay home. The out-of-state impact in this particular example would not be as severe as that in the present cases, but I do not think our Commerce Clause jurisprudence should degenerate into disputes over degree of economic effect. In any case, since this principle is both dubious and unnecessary to decide the present cases, I decline to endorse it. Chief Justice Rehnquist, with whom Justice Stevens and Justice O’Connor join, dissenting. In Baldwin v. G. A. F. Seelig, Inc., 294 U. S. 511 (1935), the Court held that a New York statute setting minimum prices for milk sold in that State violated the Commerce Clause when applied to milk produced more cheaply in Vermont but imported into New York for sale. Today the Court applies the doctrine of that case to invalidate a Connecticut statute which sets a maximum price for beer imported into 346 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. Connecticut from other States. The Court’s analysis seems wrong to me both as a matter of economics and as a matter of law: the maximum prices set by Connecticut in this case have a quite different effect than did the minimum prices set by New York in the Baldwin case, and by reason of the Twenty-first Amendment the States possess greater authority to regulate commerce in beer than they do commerce in milk. The New York statute passed upon in Baldwin provided that no milk could be sold in the New York City metropolitan area unless it had been purchased from the producer for a price at least equal to the minimum specified by law. When this statute was applied to milk produced in Vermont but brought into the New York City metropolitan area for sale, the result was to require Vermont producers to give up the natural advantage which they would otherwise have obtained from the fact that the costs of production of milk in Vermont were lower than the costs of production in New York. The Court rightly held that this sort of a regulation violated the Commerce Clause because it “set a barrier to traffic between one state and another as effective as if customs duties, equal to the price differential, had been laid upon the thing transported.” Id., at 521. In Milk Control Board v. Eisenberg Farm Products, 306 U. S. 346 (1939), decided four years after Baldwin, the Court upheld a different state milk price regulation, and in so doing distinguished Baldwin as a case in which “this Court condemned an enactment aimed solely at interstate commerce attempting to affect and regulate the price to be paid for milk in a sister state.” 306 U. S. at 353. The Connecticut statute here is markedly different from the New York statute condemned in Baldwin. Connecticut has no motive to favor local brewers over out-of-state brewers, because there are no local brewers. Ante, at 327, n. 2. Its motive—unchallenged here—is to obtain from out-of-state brewers prices for Connecticut retailers and Connecticut beer drinkers as low as those charged by the brewers in neighboring States. Connecticut does not seek to erect any HEALY v. THE BEER INSTITUTE 347 324 Rehnquist, C. J., dissenting sort of tariff barrier to exclude out-of-state beer; its residents will drink out-of-state beer if they drink beer at all, and the State simply wishes its inhabitants to be treated as favorably as those of neighboring States by the brewers who sell interstate. There is no “tariff wall” between Connecticut and other States; there is only a maximum price regulation with which the interstate brewer would rather not have to bother. But that is not a sufficient reason for saying that such a regulation violates the Commerce Clause. Neither the parties nor the Court points to any concrete evidence that the Connecticut regulation will have any effect on the beer prices charged in other States, much less a constitutionally impermissible one. It is merely assumed that consumers in the neighboring States possess “competitive advantages” over Connecticut consumers. Ante, at 339. But it is equally possible that Connecticut’s affirmation laws, a response to a history of unusually high beer prices in that State, see United States Brewers Assn., Inc. v. Healy, 692 F. 2d 275, 276 (1982), may be justifiable as a remedy for some market imperfection that permits supracompetitive prices to be charged Connecticut consumers. The Court expresses the view that these regulations will affect the prices of beer in other States and goes on to say that such an effect constitutes “regulating” or “controlling” beer sales beyond its borders. Ante, at 337, 342. But this view is simply the Court’s personal forecast about the business strategies that distributors may use to set their prices in light of regulatory obligations in various States. Certainly a distributor that considers the Connecticut affirmation law when setting its prices in Massachusetts, or offering a discount in New York, is under no legal obligation to do so. And it is quite arbitrary, and inconsistent with other Commerce Clause doctrine, to strike down Connecticut’s affirmation law because together with the laws of neighboring States it might require a brewer to plan its pricing somewhat farther in advance, ante, at 337-338, than it would prefer to do in a totally unregulated economy. 348 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. “[T]he question is not whether what [the State] has done will restrict appellants’ freedom of action outside [the State] by subjecting the exercise of such freedom to financial burdens. The mere fact that state action may have repercussions beyond state lines is of no judicial significance so long as the action is not within that domain which the Constitution forbids.” Osborn v. Ozlin, 310 U. S. 53, 62 (1940). See also Joseph E. Seagram & Sons, Inc. v. Hostetter, 384 U. S. 35, 43 (1966). I am no more convinced by the Court’s alternative rationale, that the Connecticut statute “facially discriminates” against brewers and shippers of beer engaged in interstate commerce in favor of brewers and shippers who do business wholly within Connecticut. Ante, at 340. As the Court acknowledges, there are no Connecticut brewers, ante, at 327, n. 2, and the Court has not pointed to any evidence of shippers doing business in Connecticut but not in its border States. Consequently, the Court strikes down Connecticut’s statute because it facially discriminates in favor of entities that apparently do not exist. But cf. Amerada Hess Corp. v. Director, New Jersey Division of Taxation, 490 U. S. 66, 77-78 (1989) (absence of oil reserves in New Jersey allays concern about a discriminatory motive or effect of a state tax disallowance of a deduction related to oil production). We do not know what actions Connecticut might take to eliminate discriminatory effects if a local brewer began business and a true danger of discrimination in favor of local business appeared. It is not a proper exercise of our constitutional power to invalidate state legislation as facially discriminatory just because it has not taken into account every hypothetical circumstance that might develop in the market. All of the foregoing is based on the assumption that a State has no more freedom to regulate commerce in beer than it does commerce in milk or any other commodity. But the Twenty-first Amendment, as the Court concedes, at least in theory, provides otherwise: HEALY v. THE BEER INSTITUTE 349 324 Rehnquist, C. J., dissenting “The transportation or importation into any State . . . for delivery or use therein of intoxicating liquors, in violation of the laws thereof, is hereby prohibited.” Less than 10 years ago we acknowledged that the Twenty-first Amendment confers on the States “virtually complete control over whether to permit importation or sale of liquor and how to structure the liquor distribution system.” California Retail Liquor Dealers Association v. Midcal Aluminum, Inc., 445 U. S. 97, 110 (1980). And while this “special power” of the States to regulate liquor, id., at 108, must coexist with Congress’ power to regulate commerce, “[t]his Court’s decisions . . . have confirmed that the Amendment primarily created an exception to the normal operation of the Commerce Clause.” Craig v. Boren, 429 U. S. 190, 206 (1976). The Court in the present cases barely pays lipservice to the additional authority of the States to regulate commerce and alcoholic beverages granted by the Twenty-first Amendment. Neglecting to consider that increased authority is especially disturbing here where the perceived proscriptive force of the Commerce Clause does not flow from an affirmative legislative decision and so is at its nadir. Even the most restrictive view of the Twenty-first Amendment should validate Connecticut’s efforts to obtain from interstate brewers prices for its beer drinkers which are as favorable as the prices which those brewers charge in neighboring States. The result reached by the Court in these cases can only be described as perverse. A proper view of the Twenty-first Amendment would require that States have greater latitude under the Commerce Clause to regulate producers of alcoholic beverages than they do producers of milk. But the Court extends to beer producers a degree of Commerce Clause protection that our cases have never extended to milk producers. I would reverse the judgment of the Court of Appeals. 350 OCTOBER TERM, 1988 Syllabus 491 U. S. NEW ORLEANS PUBLIC SERVICE, INC. v. COUNCIL OF THE CITY OF NEW ORLEANS et al. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT No. 88-348. Argued April 25, 1989—Decided June 19, 1989 The Federal Energy Regulatory Commission (FERC) allocated the cost of the Grand Gulf 1 nuclear reactor among several jointly owned companies, including petitioner New Orleans Public Service, Inc. (NOPSI), that had agreed to finance the reactor’s construction and operation. NOPSI, which provides retail electrical service to New Orleans, then sought from respondent New Orleans City Council (Council), the local ratemaking body, a rate increase to cover the increase in its wholesale rates resulting from FERC’s allocation of Grand Gulf costs. Although deferring to FERC’s implicit finding that NOPSI’s decision to participate in the Grand Gulf venture was reasonable, the Council determined that the costs incurred thereby should not be completely reimbursed through a rate increase because NOPSI’s management was negligent in failing, after the risks of nuclear power became apparent, to diversify its supply portfolio by selling a portion of its Grand Gulf power. NOPSI filed a petition in state court for review of the Council’s final rate order. In parallel federal proceedings in the District Court, NOPSI sought declaratory and injunctive relief on the ground that the Council’s order was pre-empted by federal law under Nant aha la Power & Light Co. v. Thornburg, 476 U. S. 953, which held that, for purpose of setting intrastate retail rates, a State may not differ from FERC’s allocations of wholesale power by imposing its own judgment of what would be just and reasonable. The District Court concluded that it should abstain from deciding the suit under Burford v. Sun Oil Co., 319 U. S. 315, and Younger v. Harris, 401 U. S. 37. The Court of Appeals affirmed. Held: The District Court erred in abstaining from exercising jurisdiction. Pp. 358-373. (a) The Burford abstention doctrine—under which federal equity courts must decline to interfere with complex state regulatory schemes in cases involving (1) difficult state-law questions bearing on policy problems of substantial public import, or (2) efforts to establish a coherent state policy regarding a matter of substantial public concern—is not applicable. This case does not involve a state-law claim, nor even an assertion that NOPSI’s federal claims are in any way entangled in a skein of state law that must be unraveled before the federal case can NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 351 350 Syllabus proceed. Because NOPSI’s facial pre-emption claim may be resolved without venturing beyond the four corners of the Council’s rate order, federal adjudication of the claim would not unduly intrude into state governmental process or undermine the State’s ability to maintain desired uniformity in the treatment of essentially local problems. Although NOPSI’s alternative claim—-that the rate order’s nominal emphasis on NOPSI’s failure to diversify its power supply was merely a cover for the determination that the original Grand Gulf investment was itself unwise—cannot be resolved on the face of the order, resolution of that claim does not demand significant familiarity with, and will not disrupt state resolution of, distinctively local facts or policies, since wholesale electricity is not bought and sold within a predominantly local market. Pp. 360-364. (b) Nor is abstention appropriate under Younger, which held that, absent extraordinary circumstances, traditional equity concerns and principles of comity require federal courts to refrain from enjoining pending state criminal prosecutions. This Court has expanded Younger abstention beyond criminal proceedings, and even beyond proceedings in courts, but never to proceedings that are not “judicial in nature.” The Council proceedings at issue here are not judicial in nature, since ratemaking, which establishes a rule for the future, is essentially a legislative act. See, e. g., Prentis v. Atlantic Coast Line Co., 211 U. S. 210, 226-227. Nor can the proceedings in this case be considered a unitary and still-to-be-completed legislative process by virtue of the ongoing state-court review proceedings. There is no contention here that the Louisiana courts’ review involves anything other than a judicial act—that is, the declaration of NOPSI’s rights vis-à-vis the Council on present or past facts under existing law. NOPSI’s pre-emption claim was therefore ripe for federal review when the Council completed the legislative action by entering its final order. Pp. 364-373. 850 F. 2d 1069, reversed and remanded. Scalia, J., delivered the opinion of the Court, in which Brennan, White, Marshall, Stevens, O’Connor, and Kennedy, JJ., joined, and in Parts I and II-B of which Rehnquist, C. J., joined. Brennan, J., filed a concurring opinion, in which Marshall, J., joined, post, p. 373. Rehnquist, C. J., filed an opinion concurring in part and concurring in the judgment, post, p. 373. Blackmun, J., filed an opinion concurring in the judgment, post, p. 374. Rex E. Lee argued the cause for petitioner. With him on the briefs were David W. Carpenter, Thomas O. Lind, Her 352 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. schel L. Abbott, Jr., David G. Radlauer, and Edward H. Bergin. Richard J. Lazarus argued the cause for the United States et al. as amici curiae urging reversal. With him on the brief were Acting Solicitor General Bryson, Deputy Solicitor General Shapiro, Catherine C. Cook, Jerome M. Feit, and Robert H. Solomon. Clinton A. Vince argued the cause for respondents. With him on the brief were Bernhardt K. Wruble, Nancy A. Wodka, and Okla Jones II* Justice Scalia delivered the opinion of the Court. In Nantahala Power & Light Co. v. Thornburg, 476 U. S. 953 (1986), we held that for purposes of setting intrastate retail rates a State may not differ from the Federal Energy Regulatory Commission’s allocations of wholesale power by imposing its own judgment of what would be just and reasonable. Last Term, in Mississippi Power & Light Co. v. Mississippi ex rel. Moore, 487 U. S. 354 (1988), we held that FERC’s allocation of the $3 billion-plus cost of the Grand Gulf 1 nuclear reactor among the operating companies that jointly agreed to finance its construction and operation pre-empted Mississippi’s inquiry into the prudence of a utility retailer’s decision to participate in the joint venture. Today we confront once again a legal issue arising from the question of who must pay for Grand Gulf 1. Here the state ratemaking authority deferred to FERC’s implicit finding that New Orleans Public Service, Inc.’s decision to participate in the Grand Gulf venture was reasonable, but determined that the costs incurred thereby should not be completely reimbursed because, it asserted, the utility’s management was negligent in failing later to diversify its supply portfolio by selling a *Briefs of amici curiae urging affirmance were filed for the National Association of Regulatory Utility Commissioners by William Paul Rodgers, Jr.; for the National League of Cities et al. by Benna Ruth Solomon and Charles Rothfeld; and for the Pennsylvania Public Utility Commission by Lawrence F. Barth and John F. Povilaitis. NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 353 350 Opinion of the Court portion of its Grand Gulf power. Whether the State’s decision to provide less than full reimbursement for the FERC-allocated wholesale costs conflicts with our holdings in Nantahala and Mississippi Power & Light is not at issue in this case. Rather, we address the threshold question whether the District Court, which the utility petitioned for declaratory and injunctive relief from the state ratemaking authority’s order, properly abstained from exercising jurisdiction in deference to the state review process. I Because the abstention questions at stake here have little to do with the intricacies of the factual and procedural history underlying the controversy, we may sketch the background of this case in brief.1 Petitioner New Orleans Public Service, Inc. (NOPSI), a producer, wholesaler, and retailer of electricity that provides retail electrical service to the city of New Orleans, is one of four wholly owned operating subsidiaries of Middle South Utilities, Inc. Middle South operates an integrated “power pool” in which each of the four operating companies transmits produced electricity to a central dispatch center and draws back from the dispatch center the power it needs to meet customer demand. In 1974, NOPSI and its fellow operating companies entered a contract with Middle South Energy, Inc. (MSE), another wholly owned Middle South subsidiary, whereby the operating companies agreed to finance MSE’s construction and operation of two 1250 megawatt nuclear reactors, Grand Gulf 1 and 2, in return for the right to the reactors’ electrical output. The estimated cost of completing the two reactors was $1.2 billion. During the late 1970’s, consumer demand turned out to be far lower than expected, and regulatory delays, enhanced construction requirements, and high inflation led to spiraling 1 For a more in-depth account of the factual and regulatory history of the Grand Gulf nuclear power project, see Mississippi Power & Light Co. v. Mississippi ex rel. Moore, 487 U. S. 354 (1988). 354 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. costs. As a result, construction of Grand Gulf 2 was suspended, and the cost of completing Grand Gulf 1 alone eventually exceeded $3 billion. Not surprisingly, the cost of the electricity produced by the reactor greatly exceeded that of power generated by Middle South’s conventional facilities. Acting pursuant to its exclusive regulatory authority over interstate wholesale power transactions, 49 Stat. 847, as amended, 16 U. S. C. §824 et seq., FERC conducted extensive proceedings to determine “just and reasonable” rates for Grand Gulf 1 power and to prescribe a “just, reasonable, and nondiscriminatory” allocation of Grand Gulf’s costs and output. In June 1985, the Commission issued a final order, Middle South Energy, Inc., 31 FERC 561,305, rehearing denied, 32 FERC 5161,425 (1985), aff’d sub nom. Mississippi Industries v. FERC, 257 U. S. App. D. C. 244, 808 F. 2d 1525, rehearing granted and vacated in part, 262 U. S. App. D. C. 42, 822 F. 2d 1104, cert, denied, 484 U. S. 985 (1987), in which it concluded that, because the planned nuclear reactors had been designed “to meet overall System needs and objectives,” 31 FERC, p. 61,655, the Middle South subsidiaries should pay for the Grand Gulf project “roughly in proportion to each company’s share of System demand,” id., at 61,655-61,656. The Commission allocated 17 percent of Grand Gulf costs (approximately $13 million per month) to NOPSI, rejecting Middle South’s proposal of 29.8 percent as well as the 9 percent figure favored by the respondent here, the New Orleans City Council. “Although it did not expressly discuss the ‘prudence’ of constructing Grand Gulf and bringing it on line, FERC implicitly accepted the uncontroverted testimony of [Middle South] executives who explained why they believed the decisions to construct and to complete Grand Gulf 1 were sound, and approved the finding that ‘continuing construction of Grand Gulf Unit No. 1 was prudent because Middle South’s executives believed Grand NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 355 350 Opinion of the Court Gulf would enable the Middle South system to diversify its base load fuel mix and, it was projected, at the same time, produce power for a total cost (capacity and energy) which would be less than existing alternatives on the system.’ ” Mississippi Power & Light Co. v. Mississippi ex rel. Moore, 487 U. S., at 363, quoting Middle South Energy, Inc., 26 FERC IT 63,044, pp. 65,112-65,113 (1984). When NOPSI sought from the New Orleans City Council (Council)—the local ratemaking body with final authority over the utility’s retail rates, see 16 U. S. C. § 824(b); La. Rev. Stat. Ann. §§33:4405, 33:4495 (West 1988); Home Rule Charter of the City of New Orleans §4-1604 (1986), as amended by Ordinance No. 8264 M. C. S., as amended by Ordinance No. 10340 M. C. S. —a rate increase to cover the increase in wholesale rates resulting from FERC’s allocation of Grand Gulf costs, the Council denied an immediate rate adjustment, explaining that a public hearing was necessary to explore “‘the legality and prudency [sic] of the [contracts relating to Grand Gulf 1, and] the prudency [sic] and reasonableness of the said expenses.’” Brief for United States et al. as Amici Curiae 5, quoting Council Resolution R-85-423. NOPSI responded by filing an action for injunctive and declaratory relief in the United States District Court for the Eastern District of Louisiana, asserting that federal law required the Council to allow it to recover, through an increase in retail rates, its FERC-allocated share of the Grand Gulf expenses. The District Court granted the Council’s motion to dismiss, holding that pursuant to the Johnson Act, 28 U. S. C. § 1342, it had no jurisdiction to entertain the action, and that even if it had jurisdiction it would be compelled by Burford v. Sun Oil Co., 319 U. S. 315 (1943), to abstain. On appeal, the Fifth Circuit initially reversed on both grounds, but later, on its own motion, vacated its earlier opinion in part and held that abstention was proper both under Burford and under 356 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Younger v. Harris, 401 U. S. 37 (1971). New Orleans Pub. Serv., Inc. v. New Orleans, 782 F. 2d 1236, modified, 798 F. 2d 858 (1986), cert, denied, 481 U. S. 1023 (1987) (NOPSI I). By resolution of October 10, 1985, while NOPSI I was still pending before the Fifth Circuit, the Council initiated an investigation into the prudence of NOPSI’s involvement in Grand Gulf 1. Resolution R-85-636 stated the Council’s intention to examine all aspects of NOPSI’s relationship with Grand Gulf, including NOPSI’s “‘efforts to minimize its total cost exposure for the purchase,’” and Grand Gulf’s “‘impact on its other power supply opportunities,’” “‘for the purpose of determining what portion, if any, of NOPSI’s Grand Gulf 1 expense shall be assumed by [NOPSI’s] shareholders.’” App. 113-114. The resolution specifically provided, however, that in setting the appropriate retail rate, the Council would “‘not seek to invalidate any of the agreements surrounding Grand Gulf 1 or to order NOPSI to pay MSE a rate other than that approved by the FERC.’” Id., at 114. In November 1985, NOPSI filed a second suit in the United States District Court for the Eastern District of Louisiana, seeking to preclude the Council from requiring NOPSI or its shareholders to absorb any of NOPSI’s FERC-allocated share of the Grand Gulf costs. The District Court dismissed the suit as unripe, but held in the alternative that abstention was appropriate. On appeal, the Fifth Circuit affirmed the judgment on ripeness grounds. New Orleans Pub. Serv., Inc. v. Council of New Orleans, 833 F. 2d 583 (1987). The Council completed its prudence review on February 4, 1988, and immediately entered a final order disallowing $135 million of the Grand Gulf costs. The order was based on the Council’s determinations that “NOPSI’s . . . oversight and review of its Grand Gulf obligation . . . was uncritical and severely deficient,” App. 24, and that NOPSI acted imprudently in failing to reduce the risk of its Grand Gulf commitment, in the wake of the Three Mile Island nuclear incident in NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 357 350 Opinion of the Court March 1979, “by selling all or part of its share off-system,” id., at 24-25. Upon receipt of the Council’s decree, NOPSI turned once again to the District Court for the Eastern District of Louisiana, seeking declaratory and injunctive relief on the ground that, in light of this Court’s recent decision in Nantahala Power & Light Co. v. Thornburg, 476 U. S. 953 (1986), the Council’s rate order was pre-empted by federal law. Although the District Court expressed considerable doubt as to the merits of the Council’s position on the pre-emption question,2 it concluded that, notwithstanding Nantahala, it should still abstain from deciding the suit. Anticipating that the District Court might again abstain, NOPSI had filed a petition for review of the Council’s order in the Civil District Court for the Parish of Orleans, Louisiana. As filed, NOPSI’s petition raised only state-law claims and federal due process and takings claims, but NOPSI in 2 Adverting to the merits, the District Court commented: “[T]he Council faults NOPSI not for buying a ‘pig in a poke’ but for failing to find a sucker to buy it when the faux-pas became apparent.11 “nP. T. Barnum once said of suckers: ‘There’s one born every minute.’ This court, however, is not ready to assume there are many, if any, such suckers purchasing electricity in the wholesale market today. Indeed, this court is somewhat mystified by the Council’s logic in arriving at the $135 million disallowance in the Rate Order. In the Rate Order, the Council simply concluded that since [NOPSI’s President] said so, savings were actually possible. Then, the Council seemingly pulled from thin air a figure of 8% for the prudence disallowance. However, the Council, and in this case, everyone else knows that the 8% figure was not pulled from thin air but represents the difference between FERC’s 17% allocation and what NOPSI consistently claims as its relative share of the [Middle South] system [and what the Council advocated unsuccessfully in the FERC proceeding], i. e., 9%. Thus, the disallowed costs bear no apparent relationship to the savings NOPSI is said to have foregone [sic]. Must not the ‘savings’ posited as the reason for the disallowance be at least possible in an actual economic market? Furthermore, must not the ultimate disallowance bear some rational relationship to the possible savings which support that disallowance? These questions must be resolved on another day in another court.” App. to Pet. for Cert. 30A-31A, and n. 11. 358 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. formed the state court by letter that it would amend to raise its federal pre-emption claim if the federal court once again dismissed its complaint. When that happened, it did so.3 In the parallel federal proceedings, the Fifth Circuit affirmed the District Court’s dismissal, agreeing that the case was effectively controlled by NOPSI I, i. e., that Burford and Younger abstention applied. 850 F. 2d 1069 (1988). We granted certiorari. 488 U. S. 1003 (1989). II Before proceeding to the merits of the abstention issues, it bears emphasis that the Council does not dispute the District Court’s jurisdiction to decide NOPSI’s pre-emption claim. Our cases have long supported the proposition that federal courts lack the authority to abstain from the exercise of jurisdiction that has been conferred. For example:“We have no more right to decline the exercise of jurisdiction which is given, than to usurp that which is not given. The one or the other would be treason to the Constitution.” Cohens v. Virginia, 6 Wheat. 264, 404 (1821). “ ‘[T]he courts of the United States are bound to proceed to judgment and to afford redress to suitors before them in every case to which their jurisdiction extends. They cannot abdicate their authority or duty in any case in favor of another jurisdiction.’” Chicot County v. Sherwood, 148 U. S. 529, 534 (1893) (citations omitted). “When a Federal court is properly appealed to in a case over which it has by law jurisdiction, it is its duty to 3 NOPSI’s state suit has since been consolidated with a declaratory judgment action filed earlier by the Council, seeking a declaration that the rate order represented a just and reasonable exercise of regulatory power and that NOPSI’s failure to comply with the order would be unlawful, and with a suit filed by a local consumers’ rights organization, the Alliance for Affordable Energy, seeking to force the Council to disallow all or at least a larger proportion of the Grand Gulf costs. That case is still pending. NOPSI v. Council of New Orleans, No. 88-4511; Boissiere v. Cain, No. 88-2503; and Alliance for Affordable Energy, Inc. v. Council of New Orleans, No. 88-2502 (Civ. Dist. Ct., Parish of Orleans, La.). NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 359 350 Opinion of the Court take such jurisdiction .... The right of a party plaintiff to choose a Federal court where there is a choice cannot be properly denied.” Willcox v. Consolidated Gas Co., 212 U. S. 19, 40 (1909) (citations omitted). Underlying these assertions is the undisputed constitutional principle that Congress, and not the Judiciary, defines the scope of federal jurisdiction within the constitutionally permissible bounds. Kline v. Burke Construction Co., 260 U. S. 226, 234 (1922). That principle does not eliminate, however, and the categorical assertions based upon it do not call into question, the federal courts’ discretion in determining whether to grant certain types of relief—a discretion that was part of the common-law background against which the statutes conferring jurisdiction were enacted. See Shapiro, Jurisdiction and Discretion, 60 N. Y. U. L. Rev. 543, 570-577 (1985). Thus, there are some classes of cases in which the withholding of authorized equitable relief because of undue interference with state proceedings is “the normal thing to do,” Younger v. Harris, 401 U. S., at 45. We have carefully defined, however, the areas in which such “abstention” is permissible, and it remains “‘the exception, not the rule.’” Hawaii Housing Authority v. Midkiff, 467 U. S. 229, 236 (1984), quoting Colorado River Water Conservation Dist. v. United States, 424 U. S. 800, 813 (1976). As recently as last Term we described the federal courts’ obligation to adjudicate claims within their jurisdiction as “‘virtually unflagging.’” Deakins v. Monaghan, 484 U. S. 193, 203 (1988) (citation omitted). With these principles in mind, we address the question whether the District Court, relying on Burford v. Sun Oil Co., 319 U. S. 315 (1943), and Younger v. Harris, supra, properly declined to exercise its jurisdiction in the present case. While we acknowledge that “[t]he various types of abstention are not rigid pigeonholes into which federal courts must try to fit cases,” Pennzoil Co. v. Texaco Inc., 481 U. S. 1, 11, n. 9 (1987), the policy considerations supporting Bur 360 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ford and Younger are sufficiently distinct to justify independent analyses. A In Burford v. Sun Oil Co., supra, a Federal District Court sitting in equity was confronted with a Fourteenth Amendment challenge to the reasonableness of the Texas Railroad Commission’s grant of an oil drilling permit. The constitutional challenge was of minimal federal importance, involving solely the question whether the commission had properly applied Texas’ complex oil and gas conservation regulations. Id., at 331, and n. 28. Because of the intricacy and importance of the regulatory scheme, Texas had created a centralized system of judicial review of commission orders, which “permit[ted] the state courts, like the Railroad Commission itself, to acquire a specialized knowledge” of the regulations and industry, id., at 327. We found the state courts’ review of commission decisions “expeditious and adequate,” id., at 334, and, because the exercise of equitable jurisdiction by comparatively unsophisticated Federal District Courts alongside state-court review had repeatedly led to “[d]elay, misunderstanding of local law, and needless federal conflict with the state policy,” id., at 327, we concluded that “a sound respect for the independence of state action requir[ed] the federal equity court to stay its hand,” id., at 334. We applied these same principles in Alabama Pub. Serv. Common n. Southern R. Co., 341 U. S. 341 (1951), where a railroad sought to enjoin enforcement of an order of the Alabama Public Service Commission refusing permission to discontinue unprofitable rail lines. According to the railroad, requiring continued operation of the lines amounted to confiscation of property in violation of federal due process rights. Under Alabama law, a party dissatisfied with a final order of the Public Service Commission had an absolute right of appeal to the Circuit Court of Montgomery County, which was “empowered to set aside any Commission order found to be contrary to the substantial weight of the evidence or errone- NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 361 350 Opinion of the Court ous as a matter of law.” Id., at 348. This right of statutory appeal “concentrated in one circuit court” which exercised “supervisory” powers was, we found, “an integral part of the regulatory process under the Alabama Code.” Ibid. Taking account of the unified nature of the state regulatory process, and emphasizing that “adequate state court review of [the] administrative order [was] available,” id., at 349, and that the success of the railroad’s constitutional challenge depended upon the “predominantly local factor of public need for the service rendered,” id., at 347, we held that the District Court ought to have abstained from exercising its jurisdiction, id., at 350. From these cases, and others on which they relied, we have distilled the principle now commonly referred to as the “Burford doctrine.” Where timely and adequate state-court review is available, a federal court sitting in equity must decline to interfere with the proceedings or orders of state administrative agencies: (1) when there are “difficult questions of state law bearing on policy problems of substantial public import whose importance transcends the result in the case then at bar”; or (2) where the “exercise of federal review of the question in a case and in similar cases would be disruptive of state efforts to establish a coherent policy with respect to a matter of substantial public concern.” Colorado River Water Conservation Dist. v. United States, supra, at 814. The present case does not involve a state-law claim, nor even an assertion that the federal claims are “in any way entangled in a skein of state law that must be untangled before the federal case can proceed,” McNeese n. Board Of Education for Community Unit School Dist. 187, Cahokia, 373 U. S. 668, 674 (1963). The Fifth Circuit acknowledged as much in NOPSI I, but found “the absence of a state law claim • . . not fatal” because, it thought, “[t]he motivating force behind Burford abstention is ... a reluctance to intrude into state proceedings where there exists a complex state regulatory system.” 798 F. 2d, at 861-862. Finding that this case 362 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. involved a complex regulatory scheme of “paramount local concern and a matter which demands local administrative expertise,” id., at 862, it held that the District Court appropriately applied Burford. While Burford is concerned with protecting complex state administrative processes from undue federal interference, it does not require abstention whenever there exists such a process, or even in all cases where there is a “potential for conflict” with state regulatory law or policy. Colorado River Water Conservation Dist., 424 U. S., at 815-816. Here, NOPSI’s primary claim is that the Council is prohibited by federal law from refusing to provide reimbursement for FERC-allocated wholesale costs. Unlike a claim that a state agency has misapplied its lawful authority or has failed to take into consideration or properly weigh relevant state-law factors, federal adjudication of this sort of pre-emption claim would not disrupt the State’s attempt to ensure uniformity in the treatment of an “essentially local problem,” Alabama Pub. Serv. Common, supra, at 347. That Burford abstention is not justified in these circumstances is strongly suggested by our decision in Public Util. Common of Ohio n. United Fuel Gas Co., 317 U. S. 456 (1943), decided just four months prior to Burford, in which a District Court had enjoined on federal pre-emption grounds a State’s attempt to fix interstate gas rates. After determining that the State’s order impinged on the authority Congress had vested solely in the Federal Power Commission, we addressed the State’s contention that the District Court had nonetheless abused its discretion by granting injunctive relief: “It is perhaps unnecessary at this late date to repeat the admonition that the federal courts should be wary of interrupting the proceedings of state administrative tribunals by use of the extraordinary writ of injunction. But this, too, is a rule of equity and not to be applied in blind disregard of fact. And what are the commanding cir- NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 363 350 Opinion of the Court cumstances of the present case? First, and most important, the orders of the state Commission are on their face plainly invalid. No inquiry beyond the orders themselves and the undisputed facts which underlie them is necessary in order to discover that they are in conflict with the federal Act.” 317 U. S., at 468-469 (emphasis added). Similarly in the case at bar, no inquiry beyond the four corners of the Council’s retail rate order is needed to determine whether it is facially pre-empted by FERC’s allocative decree and relevant provisions of the Federal Power Act. Such an inquiry would not unduly intrude into the processes of state government or undermine the State’s ability to maintain desired uniformity. It may, of course, result in an injunction against enforcement of the rate order, but “there is . . . no doctrine requiring abstention merely because resolution of a federal question may result in the overturning of a state policy.” Zablocki n. Redhail, 434 U. S. 374, 380, n. 5 (1978). It is true that in its initial complaint, NOPSI asserted, as an alternative to its facial pre-emption challenge, that the rate order’s nominal emphasis on NOPSI’s failure in 1979-1980 to diversify its power supply by selling off a portion of its Grand Gulf allocation was merely a cover for the determination that the original Grand Gulf investment was itself unwise. Unlike the facial challenge, this claim cannot be resolved on the face of the rate order, because it hinges largely on the plausibility of the Council’s finding that NOPSI should have, and could have, diversified its supply portfolio and thereby lowered its average wholesale costs. See n. 2, supra. Analysis of this pretext claim requires an inquiry into industry practice, wholesale rates, and power availability during the relevant time period, an endeavor that demands some level of industry-specific expertise. But since, as the facts of this case amply demonstrate, wholesale electricity is not bought and sold within a predominantly local 364 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. market, it does not demand significant familiarity with, and will not disrupt state resolution of, distinctively local regulatory facts or policies. The principles underlying Burford are therefore not implicated. B In Younger n. Harris, 401 U. S. 37 (1971), which involved a facial First Amendment-based challenge to the California Criminal Syndicalism Act, we held that absent extraordinary circumstances federal courts should not enjoin pending state criminal prosecutions. That far-from-novel holding was based partly on traditional principles of equity, id., at 43-44, but rested primarily on the ‘‘even more vital consideration” of comity, id., at 44. As we explained, this includes “a proper respect for state functions, a recognition of the fact that the entire country is made up of a Union of separate state governments, and a continuance of the belief that the National Government will fare best if the States and their institutions are left free to perform their separate functions in their separate ways.” Ibid. The state-court proceeding at issue here is not a criminal prosecution, and one of the issues in the present case is whether the principle of Younger can properly be extended to this type of suit. NOPSI argues that that issue does not have to be reached, however, for several reasons. First, NOPSI argues that Younger does not require abstention in the face of a substantial claim that the challenged state action is completely pre-empted by federal law. Such a claim, NOPSI contends, calls into question the prerequisite of Younger abstention that the State have a legitimate, substantial interest in its pending proceedings, Middlesex County Ethics Comm. v. Garden State Bar Assn., 457 U. S. 423, 432 (1982). Thus, it contends, a district court presented with a pre-emption-based request for equitable relief should take a quick look at the merits; and if upon that look the claim appears substantial, the court should endeavor to resolve it. NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 365 350 Opinion of the Court We disagree. There is no greater federal interest in enforcing the supremacy of federal statutes than in enforcing the supremacy of explicit constitutional guarantees, and constitutional challenges to state action, no less than preemption-based challenges, call into question the legitimacy of the State’s interest in its proceedings reviewing or enforcing that action. Yet it is clear that the mere assertion of a substantial constitutional challenge to state action will not alone compel the exercise of federal jurisdiction. See Younger, 401 U. S., at 53. That is so because when we inquire into the substantiality of the State’s interest in its proceedings we do not look narrowly to its interest in the outcome of the particular case—which could arguably be offset by a substantial federal interest in the opposite outcome. Rather, what we look to is the importance of the generic proceedings to the State. In Younger, for example, we did not consult California’s interest in prohibiting John Harris from distributing handbills, but rather its interest in “carrying out the important and necessary task” of enforcing its criminal laws. Id., at 51-52. Similarly, in Ohio Civil Rights Common v. Dayton Christian Schools, Inc., 477 U. S. 619 (1986), we looked not to Ohio’s specific concern with Dayton Christian Schools’ firing of Linda Hoskinson, but to its more general interest in preventing employers from engaging in sex discrimination. Id., at 628. Because pre-emption-based challenges merit a similar focus, the appropriate question here is not whether Louisiana has a substantial, legitimate interest in reducing NOPSI’s retail rate below that necessary to recover its wholesale costs, but whether it has a substantial, legitimate interest in regulating intrastate retail rates. It clearly does. “[T]he regulation of utilities is one of the most important of the functions traditionally associated with the police power of the States.” Arkansas Electric Cooperative Corp. v. Arkansas Pub. Serv. Comm’n, 461 U. S. 375, 377 (1983). Accord, Pacific Gas & Electric Co. v. State Energy Resources Conservation and Development Common, 461 U. S. 190, 205- 366 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. 206 (1983); Central Hudson Gas & Electric Corp. v. Public Serv. Comm’n of New York, 447 U. S. 557, 569 (1980). NOPSI attempts to avoid this conclusion by stressing that it challenges not only the result of the Council’s deliberations, but the very right of the Council to conduct those deliberations. (This argument assumes, of course, that enjoining the Louisiana state courts can be equated with enjoining the Council proceedings, a point we shall address in due course.) But that is simply not true, if the reference to “the Council’s deliberations” is as generic as it should be. NOPSI does not deny that the State has an interest affirmatively protected by federal law in conducting proceedings to set intrastate retail electricity rates; rather, it contends that under the particular facts of the present case its FERC-allocated wholesale costs are not a proper subject for such proceedings. That is no different from the contention in Younger that the defendant’s violation of the particular (allegedly unconstitutional) state statute was not a proper subject of prosecution. In other words, this argument of NOPSI ultimately reduces once again to insistence upon too narrow an analytical focus. NOPSI’s second argument to the effect that abstention is improper even assuming the state proceedings here are the sort to which Younger applies rests upon the principle that abstention is not appropriate if the federal plaintiff will “suffer irreparable injury” absent equitable relief. Younger, 401 U. S., at 43-44; see also id., at 48. Irreparable injury may possibly be established, Younger suggested, by a showing that the challenged state statute is “ ‘flagrantly and patently violative of express constitutional prohibitions . . . /” id., at 53-54, quoting Watson n. Buck, 313 U. S. 387, 402 (1941). Relying on Public Util. Comm’n of Ohio v. United Fuel Gas Co., 317 U. S. 456 (1943), where we upheld the order of a District Court enjoining the State Public Utilities Commission from attempting directly to regulate interstate gas prices because such actions were “on their face plainly invalid,” id., at 469 (emphasis added), NOPSI asserts that Younger’s pos- NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 367 350 Opinion of the Court ited exception for state statutes “flagrantly and patently violative of express constitutional prohibitions” ought to apply equally to state proceedings and orders flagrantly and patently violative of federal pre-emption (which is unlawful only because it violates the express constitutional prescription of the Supremacy Clause). Thus, NOPSI argues, even if a substantial claim of federal pre-emption is not sufficient to render abstention inappropriate, at least a facially conclusive claim is. Perhaps so. But we do not have to decide the matter here, since the proceeding and order at issue do not meet that description. The Council has not sought directly to regulate interstate wholesale rates; nor has it questioned the validity of the FERC-prescribed allocation of power within the Grand Gulf system, or the FERC-prescribed wholesale rates; nor has it reexamined the prudence of NOPSI’s agreement to participate in Grand Gulf 1 in the first place. Rather, the Council maintains that it has examined the prudence of NOPSI’s failure, after the risks of nuclear power became apparent, to diversify its supply portfolio, and that finding that failure negligent, it has taken the normal ratemaking step of making NOPSI’s shareholders rather than the ratepayers bear the consequences. Nothing in this is directly or even indirectly foreclosed by the federal statute, the regulations implementing it, or the case law applying it. There may well be reason to doubt the Council’s necessary factual finding that NOPSI would have saved money had it diversified. See n. 2, supra. But we cannot conclusively say it is wrong without further factual inquiry—and what requires further factual inquiry can hardly be deemed “flagrantly” unlawful for purposes of a threshold abstention determination. We conclude, therefore, that NOPSI’s challenge must stand or fall upon the answer to the question whether the Louisiana court action is the type of proceeding to which Younger applies. Viewed in isolation, it plainly is not. Although our concern for comity and federalism has led us to 368 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. expand the protection of Younger beyond state criminal prosecutions, to civil enforcement proceedings, Huffman v. Pursue, Ltd., 420 U. S. 592, 604 (1975); Trainor v. Hernandez, 431 U. S. 434, 444 (1977); Moore v. Sims, 442 U. S. 415, 423 (1979), and even to civil proceedings involving certain orders that are uniquely in furtherance of the state courts’ ability to perform their judicial functions, see Juidice v. Vail, 430 U. S. 327, 336, n. 12 (1977) (civil contempt order); Pennzoil Co. v. Texaco Inc., 481 U. S. 1, 13 (1987) (requirement for the posting of bond pending appeal), it has never been suggested that Younger requires abstention in deference to a state judicial proceeding reviewing legislative or executive action. Such a broad abstention requirement would make a mockery of the rule that only exceptional circumstances justify a federal court’s refusal to decide a case in deference to the States. Colorado River Water Conservation Dist. v. United States, 424 U. S., at 817; Moses H. Cone Memorial Hospital v. Mercury Construction Corp., 460 U. S. 1, 25 (1983); cf. Moore v. Sims, supra, at 423, n. 8 (“[W]e do not remotely suggest ‘that every pending proceeding between a State and a federal plaintiff justifies abstention unless one of the exceptions to Younger applies’” (citation omitted)). In asserting that Younger is applicable, however, respondents focus not upon the Louisiana court action in isolation, but upon that action as a mere continuation of the Council proceeding. Their contention is that “[t]he Council’s own ratemaking and prudence inquiry, even though complete, constitutes an ‘ongoing proceeding’ because it is subject to state judicial review.” Brief for Respondents 31. The proper question, they contend, is whether the Council proceeding qualified for Younger treatment—because if it did, the proceeding is not complete until judicial review is concluded. Respondents argue by analogy to the treatment of court proceedings, for Younger purposes, as an uninterruptible whole. When, in a proceeding to which Younger applies, a state trial court has entered judgment, the losing NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 369 350 Opinion of the Court party cannot, of course, pursue equitable remedies in federal district court while concurrently challenging the trial court’s judgment on appeal. For Younger purposes, the State’s trial-and-appeals process is treated as a unitary system, and for a federal court to disrupt its integrity by intervening in midprocess would demonstrate a lack of respect for the State as sovereign. For the same reason, a party may not procure federal intervention by terminating the state judicial process prematurely—forgoing the state appeal to attack the trial court’s judgment in federal court. “[A] necessary concomitant of Younger is that a party [wishing to contest in federal court the judgment of a state judicial tribunal] must exhaust his state appellate remedies before seeking relief in the District Court.” Huffman v. Pursue, Ltd., supra, at 608. Respondents urge that these principles apply equally where the initial adjudicatory tribunal is an agency—i. e., that the litigation, from agency through courts, is to be viewed as a unitary process that should not be disrupted, so that federal intervention is no more permitted at the conclusion of the administrative stage than during it. We will assume, without deciding, that this is correct.4 Respondents’ case for abstention still requires, however, that the Council proceeding be the sort of proceeding entitled to Younger treatment. We think it is not. While we have ex 4In Ohio Civil Rights Comm’n v. Dayton Christian Schools, Inc., 477 U. S. 619 (1986), we held that the Younger doctrine prevented an injunction against an ongoing sex discrimination proceeding before the Ohio Civil Rights Commission. The only other decision of ours arguably applying Younger to an administrative proceeding, Middlesex County Ethics Comm. v. Garden State Bar Assn., 457 U. S. 423 (1982), similarly involved a situation in which the proceeding was not yet at an end. The fact that Dayton Christian Schools relied, as an alternative argument, upon the fact that the federal challenge could be made upon appeal to the state courts, see 477 U. S., at 629, suggests, perhaps, that an administrative proceeding to which Younger applies cannot be challenged in federal court even after the administrative action has become final. But we have never squarely faced the question. 370 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. panded Younger beyond criminal proceedings, and even beyond proceedings in courts, we have never extended it to proceedings that are not “judicial in nature.” See Middlesex County Ethics Comm. v. Garden State Bar Assn., 457 U. S., at 433-434 (“It is clear beyond doubt that the New Jersey Supreme Court considers its bar disciplinary proceedings as ‘judicial in nature.’ As such, the proceedings are of a character to warrant federal-court deference”). See also Ohio Civil Rights Comm’n v. Dayton Christian Schools, Inc., 477 U. S., at 627 (“Because we found that the administrative proceedings in Middlesex were ‘judicial in nature’ from the outset, ... it was not essential to the decision that they had progressed to state-court review by the time we heard the federal injunction case”). The Council’s proceedings in the present case were not judicial in nature. In Prentis n. Atlantic Coast Line Co., 211 U. S. 210 (1908), several railroads requested a Federal Circuit Court “to enjoin . . . the Virginia State Corporation Commission from publishing or taking any steps to enforce a certain order fixing passenger rates,” on the ground that the proposed rates were confiscatory. Id., at 223. To decide whether the federal court was at liberty to issue the requested injunction, we examined first the nature of the challenged agency action. Under Virginia law the commission was invested with both legislative and judicial powers, and we assumed, without deciding, that “if it were proceeding against [a railroad] to enforce [the rate] order or to punish [the railroad] for a breach, “it then would be sitting as a court and would be protected from interference on the part of courts of the United States,” id., at 226. But, upon analysis, we found the proceedings in the case at hand to be legislative. Justice Holmes, writing for the Court, explained as follows: “A judicial inquiry investigates, declares and enforces liabilities as they stand on present or past facts and under laws supposed already to exist. That is its purpose and end. Legislation on the other hand looks to the future NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 371 350 Opinion of the Court and changes existing conditions by making a new rule to be applied thereafter to all or some part of those subject to its power. The establishment of a rate is the making of a rule for the future, and therefore is an act legislative and not judicial in kind . . . Ibid. He then considered and rejected the notion that the nature of the agency’s proceedings might depend on their form: “[The proper characterization of an agency’s actions] depends not upon the character of the body but upon the character of the proceedings. . . . And it does not matter what inquiries may have been made as a preliminary to the legislative act. Most legislation is preceded by hearings and investigations. But the effect of the inquiry, and of the decision upon it, is determined by the nature of the act to which the inquiry and decision lead up. . . . The nature of the final act determines the nature of the previous inquiry. As the judge is bound to declare the law he must know or discover the facts that establish the law. So when the final act is legislative the decision which induces it cannot be judicial in the practical sense, although the questions considered might be the same that would arise in the trial of a case.” Id., at 226-227 (citations omitted). We have since reaffirmed both the general mode of analysis of Prentis, see District of Columbia Court of Appeals v. Feldman, 460 U. S. 462, 476-479 (1983), and its specific holding that ratemaking is an essentially legislative act, Colorado Interstate Gas Co. n. FPC, 324 U. S. 581, 589 (1945). Thus, the Council’s proceedings here were plainly legislative. That characterization does not, however, end the inquiry. In Prentis, while we found the challenged agency proceeding legislative in character, we nonetheless held equitable intervention inappropriate because, we determined, the attack on the rate order was premature. Although we made clear that those challenging the rates “were not bound to wait for pro 372 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ceedings brought to enforce the rate and to punish them for departing from it,” 211 U. S., at 228, because Virginia provided for legislative review of commission rates by appeal to the state courts, we concluded that the challengers “should make sure that the State in its final legislative action would not respect what they think their rights to be, before resorting to the courts of the United States.” Id., at 230. We were as concerned, in other words, to preserve the integrity of a unitary and still-to-be-completed legislative process as we were, under Huffman v. Pursue, Ltd., 420 U. S. 592 (1975), to preserve the integrity of judicial proceedings. Similarly in the present case, if the Louisiana courts’ review of Council ratemaking was legislative in nature, NOPSI’s challenge to the Council’s order should have been dismissed as unripe. There is no contention here that the Louisiana courts’ review involves anything other than a judicial act—that is, not “the making of a rule for the future,” but the declaration of NOPSI’s rights vis-à-vis the Council “on present or past facts and under laws supposed already to exist,” Prentis, supra, at 226. Nor does there seem to be room for such a contention. See State ex rei. Guste v. Council of New Orleans, 309 So. 2d 290, 294-296 (La. 1975). Since the state-court review is not an extension of the legislative process, NOPSI’s pre-emption claim was ripe for federal review when the Council’s order was entered. See Lane v. Wilson, 307 U. S. 268, 274-275 (1939); Bacon n. Rutland R. Co., 232 U. S. 134, 138 (1914). As a challenge to completed legislative action, NOPSI’s suit represents neither the interference with ongoing judicial proceedings against which Younger was directed, nor the interference with an ongoing legislative process against which our ripeness holding in Prentis was directed. It is, insofar as our policies of federal comity are concerned, no different in substance from a facial challenge to an allegedly unconstitutional statute or zoning ordinance—which we would assuredly not require to be brought in state courts. See Wooley v. NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 373 350 Opinion of Rehnquist, C. J. Maynard, 430 U. S. 705, 711 (1977). It is true, of course, that the federal court’s disposition of such a case may well affect, or for practical purposes pre-empt, a future—or, as in the present circumstances, even a pending—state-court action. But there is no doctrine that the availability or even the pendency of state judicial proceedings excludes the federal courts. Viewed, as it should be, as no more than a statecourt challenge to completed legislative action, the Louisiana suit comes within none of the exceptions that Younger and later cases have established. For the reasons stated, the judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion. So ordered. Justice Brennan, with whom Justice Marshall joins, concurring. I join the Court’s opinion. I continue to adhere to my view, however, that the abstention doctrine of Younger v. Harris, 401 U. S. 37 (1971), is in general inapplicable to civil proceedings. See Pennzoil Co. v. Texaco Inc., 481 U. S. 1, 19 (1987) (Brennan, J., concurring in judgment); Trainor v. Hernandez, 431 U. S. 434, 450 (1977) (Brennan, J., dissenting); Juidice v. Vail, 430 U. S. 327, 341 (1977) (Brennan, J., dissenting); Huffman v. Pursue, Ltd., 420 U. S. 592, 613 (1975) (Brennan, J., dissenting). Chief Justice Rehnquist, concurring in Parts I and II-B and concurring in the judgment. I agree with the Court that our prior cases extending Younger beyond criminal prosecutions to civil proceedings have limited its application to proceedings which are “judicial in nature,” and that, under our longstanding characterization of the distinction between “judicial” and “legislative” proceedings, see Prentis v. Atlantic Coast Line Co., 211 U. S. 210, 226 (1908), the Council’s ratemaking proceedings at issue here were not judicial in nature. Under these circum- 374 OCTOBER TERM, 1988 Blackmun, J., concurring in judgment 491 U. S. stances, I agree that Younger abstention is inappropriate, despite the pendency of state-court review of the Council’s ratemaking order. Nothing in the Court’s opinion curtails our prior application of Younger to certain administrative proceedings which are “judicial in nature,” see Ohio Civil Rights Common v. Dayton Christian Schools, Inc., 477 U. S. 619 (1986); Middlesex County Ethics Committee v. Garden State Bar Assn., 457 U. S. 423 (1982); nor does it alter our prior case law indicating that such proceedings should be regarded as “ongoing” for the purposes of Younger abstention until state appellate review is completed, see Dayton Christian Schools, supra, at 629. With this understanding, I join the portion of the Court’s opinion holding that Younger abstention is inappropriate here. I agree with the Court’s conclusion that Burford abstention is inappropriate on the facts of this case. But I would not foreclose the possibility of Burford abstention in a case like this had the State consolidated review of the orders of local ratemaking bodies in a specialized state court with power to hear a federal pre-emption claim. Accordingly, I concur only in the judgment as to Burford abstention. Justice Blackmun, concurring in the judgment. I concur in the judgment in this case. I also agree with what I take to be the core of the majority’s reasoning: in the posture of this case, a legislative proceeding ended when the Council entered its ratemaking order; after that point, adjudication in the District Court would not have interfered with any ongoing proceeding, be it judicial, quasi-legislative, or legislative. Ante, at 372. I find, however, that the majority’s understanding of Burford abstention is much narrower than my own in respects not relevant to the disposition of this case, and that there is considerable tension between its discussion of the nature of the State’s interests in the Burford context and its discussion of the State’s interests in the Younger context. Compare ante, at 362-363, with ante, at 366-367. Furthermore, I am not entirely persuaded NEW ORLEANS PUB. SERV., INC. v. NEW ORLEANS 375 350 Blackmun, J., concurring in judgment that this Court’s decisions applying Younger abstention to administrative proceedings that are judicial in nature leave open the question whether abstention must continue through the judicial review process. Ante, at 369, and n. 4. In my view, the majority’s observations on these questions are not necessary to the result or to the legal standard the majority has adopted. 376 OCTOBER TERM, 1988 Syllabus 491 U. S. JONES, SUPERINTENDENT, MISSOURI TRAINING CENTER FOR MEN AT MOBERLY v. THOMAS CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE EIGHTH CIRCUIT No. 88-420. Argued April 26, 1989—Decided June 19, 1989 Respondent Thomas was convicted of both attempted robbery and first-degree felony murder arising out of the same incident and was sentenced to consecutive terms of 15 years for the attempted robbery and life imprisonment for the felony murder, with the 15-year sentence to run first. This conviction was affirmed on appeal. While Thomas’ motion for postconviction relief was pending in Missouri trial court, the Governor commuted his 15-year sentence to time served. After the Missouri Supreme Court, in unrelated cases, held that the state legislature had not intended to allow separate punishments for both felony murder and the underlying felony, the trial court vacated the attempted robbery conviction and the corresponding sentence. The court left the felonymurder conviction in place, but credited the time served under the attempted robbery conviction against the life sentence. The State Court of Appeals affirmed the trial court’s order and rejected Thomas’ argument that, since he had completed his commuted sentence, his continued confinement under the longer sentence violated the double jeopardy prohibition against multiple sentences for the same offense. Thomas then sought a writ of habeas corpus in the Federal District Court. The court denied relief, ruling that Thomas had not suffered a double jeopardy violation because he had not been subjected to a greater punishment than intended by the legislature. The Court of Appeals reversed, holding that under this Court’s decisions in Ex parte Lange, 18 Wall. 163, and In re Bradley, 318 U. S. 50, once Thomas had satisfied one of the two sentences that could have been imposed by law, he could not be required to serve the other. It held further that Morris v. Mathews, 475 U. S. 237—which held that an unlawful conviction of felony murder and the underlying felony could be remedied by resentencing on a lesser included offense of nonfelony murder—was inapposite, since the prisoner in that case had not completed either of his sentences. Held: The state-court remedy fully vindicated Thomas’ double jeopardy rights. In the multiple punishments context, the Double Jeopardy Clause does no more than prevent the sentencing court from prescribing greater punishment than the legislature intended. Missouri v. Hunter, 459 U. S. 359, 366. As a result of the state trial court’s ruling, Thomas now stands convicted of felony murder alone and his confinement under JONES v. THOMAS 377 376 Syllabus the single sentence imposed for that crime with credit for time already served is not double jeopardy. Thomas’ reliance on Lange, supra, and Bradley, supra, is misplaced. Both cases involved alternative punishments that were prescribed by the legislature for a single criminal act, whereas the issue here involves separate sentences imposed for what the sentencing court thought to be separately punishable offenses, one far more serious than the other. Bradley also involved alternative sentences of two different types, fine and imprisonment. While it would not have been possible to “credit” a fine against time in prison, crediting time served under one sentence against the term of another has long been an accepted practice. Moreover, in a true alternative sentences case, it is difficult to say that the legislature intended one punishment over the other, for the legislature viewed each alternative as appropriate for some cases. Here, however, the legislature plainly intended that the person who committed murder during a felony would be convicted of felony murder or separately of the felony and nonfelony murder. It did not intend that an attempted robbery conviction would suffice as an alternative sanction for murder. Extension of Bradley beyond its facts would also lead to anomalous results since, had Thomas been sentenced to life imprisonment first, he would not have had a double jeopardy claim; and since he concedes that the unlawful imposition of concurrent sentences can be cured by vacating the shorter of the two even where it has been completed. Sentencing is not a game where a wrong move by a judge means immunity for the prisoner. Bozza v. United States, 330 U. S. 160, 166-167. Pp. 380-387. 844 F. 2d 1337, reversed and remanded. Kennedy, J., delivered the opinion of the Court, in which Rehnquist, C. J., and White, Blackmun, and O’Connor, JJ., joined. Brennan, J., filed a dissenting opinion, in which Marshall, J., joined, post, p. 387. Scalia, J., filed a dissenting opinion, in which Stevens, J., joined, and in which Brennan and Marshall, JJ., joined, except as to the footnote, post, p. 388. Stephen D. Hawke, Assistant Attorney General of Missouri, argued the cause for petitioner. With him on the briefs were William L. Webster, Attorney General, and John M. Morris III, Assistant Attorney General. Springfield Baldwin, by appointment of the Court, 489 U. S. 1006, argued the cause and filed a brief for respondent.* * Briefs of amici curiae urging reversal were filed for the United States by Acting Solicitor General Bryson, Assistant Attorney General Dennis, 378 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Justice Kennedy delivered the opinion of the Court. After it became apparent that two consecutive sentences had been imposed where state law permitted but one, a Missouri court vacated the shorter of the two and credited the time already served against the remaining sentence. At the time the court entered its order, the prisoner had completed serving the shorter sentence. The question presented is whether the longer sentence can remain in force, consistent with double jeopardy principles. I Respondent Larry Thomas attempted to rob a St. Louis, Missouri, auto parts store in 1972. Inside the store, respondent drew a gun and announced a holdup. One of the store’s customers was armed, and he tried to thwart the robbery. Respondent shot and killed him in an exchange of gunfire. Respondent was convicted in 1973 by a St. Louis Circuit Court jury both of attempted robbery and of first-degree felony murder for killing during the commission of a felony. The trial court sentenced respondent to consecutive terms of 15 years for the attempted robbery and life imprisonment for the felony murder, with the 15-year sentence to run first. The Missouri Court of Appeals affirmed respondent’s conviction on direct appeal. State v. Thomas, 522 S. W. 2d 74 (Mo. App. 1975). In 1977, respondent sought state postconviction relief, arguing that it was improper for the trial court to impose separate sentences for felony murder and the underlying felony. While respondent’s case was pending, the Missouri Supreme Court accepted this argument in unrelated cases, holding that the Missouri Legislature had not intended to allow separate punishments under the felony-murder statute. and Brian J. Martin; and for the Criminal Justice Legal Foundation by Kent S. Scheidegger and Charles L. Hobson. JONES v. THOMAS 379 376 Opinion of the Court See State v. Morgan, 612 S. W. 2d 1 (1981) (en banc); State v. Olds, 603 S. W. 2d 501 (1980) (en banc).1 In June 1981, with respondent’s postconviction motion still pending, the Governor of Missouri commuted his 15-year sentence for attempted robbery to “a term ending June 16, 1981.” Respondent remained in prison under the murder sentence. In 1982, the state trial court vacated respondent’s attempted robbery conviction and 15-year sentence, holding under Olds, supra, that respondent could not be required to serve both sentences. The Missouri Court of Appeals affirmed the order vacating the sentence, but rejected respondent’s argument that he was entitled to immediate release. Respondent had argued that because he had completed the shorter, commuted sentence, his continued confinement under the longer sentence constituted double jeopardy. The Missouri Court noted that respondent was in no way prejudiced by the trial court’s ruling, as his entire time of incarceration was credited against the life sentence. Thomas n. State, 665 S. W. 2d 621 (1983). Respondent then sought a writ of habeas corpus in federal court. The United States District Court for the Eastern District of Missouri denied relief, holding that respondent had not suffered a double jeopardy violation because he had not been subjected to greater punishment than intended by the legislature. A three-judge panel of the Eighth Circuit reversed and remanded. 816 F. 2d 364 (1987). The majority opinion noted that as a result of the Governor’s commutation, respondent had legally satisfied the 15-year sentence. See State v. Cerny, 248 S. W. 2d 844 (Mo. 1952). It further held that under this Court’s decisions in Ex parte Lange, 18 Wall. 163 (1874), and In re Bradley, 318 U. S. 50 (1943), once 1 After the Missouri Supreme Court decided Morgan and Olds, the Missouri Legislature amended the felony murder statute. The statute now provides that punishment may be imposed for both felony murder (now defined as second-degree murder) and the underlying felony. See Mo. Rev. Stat. §565.021(2) (1986). 380 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. respondent completed one of the two sentences that could have been imposed by law, he could not be required to serve any part of the other. The majority went on, however, to hold that the double jeopardy violation could be cured under this Court’s decision in Morris n. Mathews, 475 U. S. 237 (1986), which held that an unlawful conviction of both felony murder and the underlying felony could be remedied by resentencing on a lesser included offense of nonfelony murder. The panel therefore granted a conditional writ, so that respondent could be resentenced for the non-jeopardy-barred offense of nonfelony murder or released. Judge McMillian concurred in part and dissented in part. He agreed that respondent’s double jeopardy rights were violated, but stated that he would not allow resentencing because he preferred the analysis of Justice Brennan’s dissenting opinion in Mathews. 816 F. 2d, at 371. Judge Bowman dissented, concluding that the double jeopardy prohibition against multiple punishments was not violated because respondent would serve time only under the life sentence, which was a single valid punishment intended by the legislature. Judge Bowman joined Judge Hanson, however, in holding that respondent could be resentenced under Mathews. The Eighth Circuit granted rehearing en banc and ordered respondent’s unconditional release. 844 F. 2d 1337 (1988). The court held that under Lange, supra, and Bradley, supra, respondent could not be punished further once he had satisfied the sentence for attempted robbery. The court further held that Mathews, supra, was inapplicable because the prisoner in that case had not completed either of his sentences. Four judges dissented. We granted certiorari, 488 U. S. 1003 (1989), and now reverse. II The Double Jeopardy Clause of the Fifth Amendment provides that no person shall be “subject for the same offence to be twice put in jeopardy of life or limb.” The Clause affords JONES v. THOMAS 381 376 Opinion of the Court three protections to the criminal defendant. The first two, which are the most familiar, protect against a second prosecution for the same offense after acquittal, and against a second prosecution for the same offense after conviction. See, e. g., Ohio v. Johnson, 467 U. S. 493, 498 (1984). Neither of these protections against successive prosecutions is involved here. Rather, respondent’s initial conviction and sentence for both felony murder and the underlying felony violated the third aspect of the Double Jeopardy Clause, the protection against “multiple punishments for the same offense” imposed in a single proceeding. See North Carolina v. Pearce, 395 U. S. 711, 717 (1969). The constitutional question in this case is what remedy is required to cure the admitted violation. The answer turns on the interest that the Double Jeopardy Clause seeks to protect. Our cases establish that in the multiple punishments context, that interest is “limited to ensuring that the total punishment did not exceed that authorized by the legislature.” United States v. Halper, 490 U. S. 435, 450 (1989); see Johnson, supra, at 499; Missouri v. Hunter, 459 U. S. 359, 366-367 (1983). The purpose is to ensure that sentencing courts do not exceed, by the device of multiple punishments, the limits prescribed by the legislative branch of government, in which lies the substantive power to define crimes and prescribe punishments. See, e. g., Johnson, supra, at 499. In this case, respondent’s conviction of both felony murder and attempted robbery gave rise to a double jeopardy claim only because the Missouri Legislature did not intend to allow conviction and punishment for both felony murder and the underlying felony. E. g., Hunter, supra, at 368; see also Morgan, supra, at 1; Olds, supra, at 510 (construing Missouri statute). Given that, in its application to the case before us, “the Double Jeopardy Clause does no more than prevent the sentencing court from prescribing greater punishment than the legislature intended,” Hunter, supra, at 366, the state-court 382 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. remedy fully vindicated respondent’s double jeopardy rights. The Missouri court vacated the attempted robbery conviction and sentence and credited the time that respondent had served under that conviction against the remaining sentence for felony murder. This remedy of crediting time already served against the sentence that remained in place is consistent with our approach to multiple punishments problems in other contexts. See Pearce, supra, at 718-719 (credit for time served applied on resentencing at second trial following appeal). Respondent now stands convicted of felony murder alone, and his continued confinement under the single sentence imposed for that crime is not double jeopardy.2 Respondent, as did the Court of Appeals below, relies on this Court’s opinions in Lange, supra, and Bradley, supra, for the proposition that the Double Jeopardy Clause requires immediate release for the prisoner who has satisfied the shorter of two consecutive sentences that could not both lawfully be imposed. We think this approach depends on an overly broad reading of those precedents. Lange and Bradley do contain language to the effect that once a defendant “had fully suffered one of the alternative punishments to which alone the law subjected him, the power of the court to punish further was gone.” 18 Wall., at 176. But application of this language to the facts presented here is neither compelled by precedent nor supported by any double jeopardy principle. In Ex parte Lange, the defendant had been convicted of stealing mail bags, a federal offense punishable by either a $200 fine or a 1-year prison term. The trial court, how 2 Even if the Double Jeopardy Clause provided an absolute bar to multiple punishments in a single trial regardless of legislative intent, see Missouri v. Hunter, 459 U. S. 359, 369 (1983) (Marshall, J., dissenting), the fact would remain that respondent is now serving only a single sentence for a single offense. Under any view of the substantive content of the double jeopardy bar against multiple punishments, respondent has had every benefit the Clause affords. JONES v. THOMAS 383 376 Opinion of the Court ever, sentenced Lange to a $200 fine and one year in prison. Lange paid the fine and spent five days in prison before seeking a writ of habeas corpus from the trial court. The trial judge then vacated the earlier judgment and sentenced Lange to one year’s imprisonment from that date. Lange sought a writ of habeas corpus in this Court, which held that he was entitled to be released. The Court noted that Lange’s fine had already passed into the Treasury and could not be returned to him. If the second sentence were enforced, Lange would therefore have paid a $200 fine and spent a year plus five days in prison. See id., at 175. This punishment would obviously have exceeded that authorized by the legislature. Lange therefore stands for the uncontested proposition that the Double Jeopardy Clause prohibits punishment in excess of that authorized by the legislature, see United States v. DiFrancesco, 449 U. S. 117, 139 (1980), and not for the broader rule suggested by its dictum. In re Bradley, 318 U. S. 50 (1943), provides a closer analogy to this case. The defendant in Bradley was sentenced for contempt to a $500 fine and six months’ imprisonment under a statute that provided only for fine or imprisonment. Bradley was taken to prison, and two days later paid the fine. The trial court then realized its mistake, amended its sentencing order by omitting the fine and retaining only the 6-month prison sentence, and instructed the Clerk to return the fine to Bradley’s attorney, who refused to accept it. This Court, in a brief opinion citing Lange, held that Bradley was entitled to be released, stating that where ‘‘one valid alternative provision of the original sentence has been satisfied, the petitioner is entitled to be freed of further restraint.” 318 U. S., at 52. Strict application of Bradley would support respondent here. Under this view, satisfaction of one of two alternatives that could lawfully be imposed (e. g., the fine in Bradley and the commuted sentence here) is dispositive, and any attempt to correct the erroneous sentence by repaying the fine 384 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. or crediting time served would be futile. We think this approach ignores important differences between this case and Bradley. Bradley and Lange both involved alternative punishments that were prescribed by the legislature for a single criminal act. The issue presented here, however, involves separate sentences imposed for what the sentencing court thought to be separately punishable offenses, one far more serious than the other. The alternative sentences in Bradley, moreover, were of a different type, fine and imprisonment. While it would not have been possible to “credit” a fine against time in prison, crediting time served under one sentence against the term of another has long been an accepted practice. See, e. g., North Carolina v. Pearce, 395 U. S. 711 (1969). In a true alternative sentences case such as Bradley, it would be difficult to say that one punishment or the other was intended by the legislature, for the legislature viewed each alternative as appropriate for some cases. But here the legislature plainly intended one of two results for persons who committed murder in the commission of a felony: Either they were to be convicted of felony murder, or they were to be convicted separately of the felony and of nonfelony murder.3 It cannot be suggested seriously that the legislature 3 The Court of Appeals’ conclusion that the state court could not cure the double jeopardy violation through the alternative procedure approved in Morris v. Mathews, 475 U. S. 237 (1986), is therefore difficult to understand. In Mathews, we held that a violation of the double jeopardy rule against multiple punishments for the same offense in successive trials could be cured by resentencing to a lesser included offense that was not jeopardy barred. In that case, Mathews was first convicted of aggravated robbery. In a separate trial, he was then convicted of felony murder based on the robbery. The second conviction violated the Double Jeopardy Clause. See, e. g., Harris v. Oklahoma, 433 U. S. 682 (1977) (per curiam) (successive prosecutions for felony murder and the underlying felony a double jeopardy violation). Yet Mathews’ conviction of felony murder necessarily entailed a jury finding that he was guilty of the lesser included offense of nonfelony murder. Because nonfelony murder is not the “same offense” as aggravated robbery, there was no double jeopardy bar to a successive JONES v. THOMAS 385 376 Opinion of the Court intended an attempted robbery conviction to suffice as an alternative sanction for murder. The suggestion of Justice Scalia’s dissent, that the same analysis of legislative intent applies to the $200 fine imposed in Lange, post, at 390, is difficult to understand. By the terms of the statute itself, the legislature in Lange plainly did intend that in some cases the sentencing judge would impose "a mere $200 fine for the gravity of offense at issue there.” Ibid. Justice Scalia observes that the Double Jeopardy Clause protects not only against punishment in excess of legislative intent, but also against additions to a sentence in a subsequent proceeding that upset a defendant’s legitimate expectation of finality. Post, at 393-394. But this case does not present the situation posited by the dissent where a judge imposes only a 15-year sentence under a statute that permitted 15 years to life, has second thoughts after the defendant serves the sentence, and calls him back to impose another 10 years. Post, at 392. Here we must determine whether prosecution for that offense. We therefore held that the violation could be cured by resentencing respondent for nonfelony murder, unless Mathews could show prejudice from the admission of evidence on the felony-murder charge that would not have been admissible as to nonfelony murder, in which case he would be entitled to a new trial. The Court of Appeals concluded that Mathews was not applicable to this case because the prisoner in Mathews had not completed his sentence for robbery prior to the resentencing for nonfelony murder, while here Thomas satisfied the attempted robbery sentence. 844 F. 2d 1337, 1342 (CA8 1988). This distinction has no legal significance. Because nonfelony murder is not the same offense as attempted robbery, see, e. g., Blockburger v. United States, 284 U. S. 299 (1932) (defining “same offense”), there would be no double jeopardy bar to punishing Thomas for that offense, even through a second full trial. The rule of Morris v. Mathews merely allows entry of judgment without the need for a new trial where the jury’s verdict of guilt as to felony murder in the first trial necessarily included a determination that the defendant committed nonfelony murder. Under the Missouri felony-murder statute that applied to Thomas, the jury did make this determination, and there is no reason that Mathews could not have applied here if the state court had chosen that course. 386 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the resentencing of respondent was indeed the imposition of an additional sentence, or a valid remedy for improper “cumulative sentences imposed in a single trial.” Hunter, 459 U. S., at 366. There can be no doubt it was the latter. Justice Scalia’s discussion of the defendant’s expectation of finality makes no independent contribution to the inquiry, for in the end the dissent’s argument boils down to Bradley. Respondent plainly had no expectation of serving only an attempted robbery sentence when he was convicted by the Missouri trial court. Indeed, since Morgan and Olds had not been decided when respondent was sentenced, his expectation at that point was to serve both consecutive sentences. Once it was established that Missouri law would not allow imposition of both sentences, respondent had an expectation in serving “either 15 years (on the one sentence) or life (on the other sentence).” Post, at 395. The dissent rejects our conclusion that the Missouri court’s remedy fulfilled that expectation as “ruled out by Bradley. ” Ibid. But as discussed above, we do not think the law compels application of Bradley beyond its facts. Instead, we believe that the intent of the legislature, which this aspect of the Double Jeopardy Clause serves to protect, provides the standard for evaluating the Missouri court’s remedy for the Clause’s violation. Extension of Bradley to these facts would also lead to anomalous results. Under respondent’s theory, for example, everything depends on the order in which the consecutive sentences were originally imposed. Had respondent been sentenced to the life sentence first, he would be serving the very same term, but could advance no double jeopardy claim. There is no indication that the order of the sentences was of the slightest importance to the sentencing judge, and there is no reason constitutional adjudication should turn on such fortuities. Respondent also concedes that where concurrent sentences are imposed, unlawful imposition of two sentences may be cured by vacating the shorter of the two sentences even where it has been completed. See Hardy n. JONES v. THOMAS 387 376 Brennan, J., dissenting United States, 292 F. 2d 192 (CA8 1961); United States v. Leather, 271 F. 2d 80 (CA7 1959), cert, denied, 363 U. S. 831 (1960). Ironically, respondent’s argument for immediate release thus depends on the fact that he was given consecutive terms, which are typically reserved for more culpable offenders. We have previously observed that “[t]he Constitution does not require that sentencing should be a game in which a wrong move by the judge means immunity for the prisoner.” Bozza v. United States, 330 U. S. 160, 166-167 (1947). We will not depart from that principle today, and we decline to extend Bradley beyond its facts. Ill Double jeopardy is an area of the law filled with technical rules, and the protections it affords defendants might at times be perceived as technicalities. This is irrelevant where the ancient and important principles embodied in the Double Jeopardy Clause are implicated. “Violations of the Double Jeopardy Clause are no less serious than violations of other constitutional protections.” Mathews, 475 U. S., at 255 (Blackmun, J., concurring in judgment). But neither the Double Jeopardy Clause nor any other constitutional provision exists to provide unjustified windfalls. The Missouri court’s alteration of respondent’s sentence to a single term for felony murder with credit for time served provided suitable protection of his double jeopardy rights. The decision of the Court of Appeals is reversed, and the case is remanded for dismissal of respondent’s petition. It is so ordered. Justice Brennan, with whom Justice Marshall joins, dissenting. I join in Justice Scalia’s dissenting opinion, with the exception of its closing footnote. I adhere to my view that the Double Jeopardy Clause requires, except in very limited circumstances, that all charges against a defendant growing out 388 OCTOBER TERM, 1988 Scalia, J., dissenting 491 U. S. of a single criminal transaction be tried in one proceeding. See Ashe v. Swenson, 397 U. S. 436, 448-460 (1970) (Brennan, J., concurring); Morris n. Mathews, 475 U. S. 237, 257-258 (1986) (Brennan, J., dissenting). For this reason I do not agree that the State is free to retry respondent for a nonjeopardy-barred lesser included offense. Justice Scalia, with whom Justice Stevens joins, and with whom Justice Brennan and Justice Marshall join as to all but the footnote, dissenting. This is not the first time we have been called upon to consider whether a criminal defendant’s satisfaction of one of two alternative penalties prevents a court from imposing (or reimposing) the second penalty in a subsequent proceeding. In Ex parte Lange, 18 Wall. 163 (1874), the first case to recognize the Double Jeopardy Clause’s protection against multiple punishment, petitioner was convicted of stealing mailbags from the Post Office, under a statute carrying a punishment of either imprisonment for up to one year or a fine of up to $200. The presiding judge erroneously imposed the maximum of both punishments. After petitioner had paid his fine (which was remitted by the Clerk of Court to the United States Treasury) and had spent five days in prison, the judge realized his mistake and entered an order vacating the former judgment and resentencing petitioner to one year in prison. This Court stated that because petitioner had “fully performed, completed, and endured one of the alternative punishments which the law prescribed for that offence,” id., at 176, the court’s “power to punish for that offence was at an end, ” ibid, (emphasis added). Holding that the judge’s second order violated petitioner’s rights under the Double Jeopardy Clause, the Court ordered that petitioner be freed. More recently, in In re Bradley, 318 U. S. 50 (1943), a District Judge found petitioner guilty of contempt and sentenced him to six months in prison and a $500 fine. Petitioner began serving his prison sentence, and his attorney JONES v. THOMAS 389 376 Scalia, J., dissenting paid the fine to the Clerk of the Court three days later. The fine was not paid into the Treasury. Later that day, having discovered that the relevant statute permitted imprisonment or fine, but not both, the court issued a new order amending the sentence to omit the fine and instructed the Clerk to return the $500 to petitioner. Petitioner refused to accept the money. We held that order to be “a nullity.” Id., at 52. “When, on October 1, the fine was paid to the clerk and receipted for by him, the petitioner had complied with a portion of the sentence which could lawfully have been imposed. As the judgment of the court was thus executed so as to be a full satisfaction of one of the two alternative penalties of the law, the power of the court was at an end.” Ibid. The present case is indistinguishable from Lange and Bradley. Here, as there, only one of two available punishments could lawfully be imposed for the conduct in question; and here, as there, the defendant fully satisfied one of the two. Under the law of the State of Missouri, respondent’s actions in the Reid Auto Parts store on November 8, 1972, allowed the State to convict him of attempted armed robbery, with a maximum penalty of 15 years in prison, or of felony murder, with a maximum penalty of life imprisonment. The State could not convict him or punish him for both offenses. Therefore, once respondent “fully suffered one of the alternative punishments to which alone the law subjected him, the power of the court to punish further was gone.” Ex parte Lange, supra, at 176. In the present case, as in Bradley, the State attempted in a second proceeding to “give back” the detriment respondent had suffered as a result of the fully satisfied alternative—by crediting the 15-year sentence for attempted armed robbery that he had already served against the second (life) sentence that had been imposed. But I see no more reason to allow a crediting here than there was to allow a refund in Bradley. Does this produce, as the Court 390 OCTOBER TERM, 1988 Scalia, J., dissenting 491 U. S. alleges, an “anomalous resul[t],” ante, at 386, and an “unjustified windfall],” ante, at 387? Undoubtedly. Just as it did in Bradley. And just as the Double Jeopardy Clause often does (to an even greater degree) in other contexts—where, for example, a prosecutorial error after the jury has been impaneled permits the defendant to go off scot free. E. g., Downum v. United States, 372 U. S. 734, 737-738 (1963). The Court candidly recognizes that a “[s]trict application of Bradley,” ante, at 383, compels the conclusion that requiring respondent to serve the life sentence after completion of the 15-year sentence violates the Double Jeopardy Clause. It advances three related arguments, however, to explain why “strict application” can be avoided. I find none of them persuasive. Most readily answered is the contention that “Bradley and Lange both involved alternative punishments that were prescribed by the legislature for a single criminal act.” Ante, at 384. This in no way distinguishes those cases, since it describes the facts of this case just as well. Although the sentencing court undoubtedly thought attempted armed robbery and felony murder “to be separately punishable offenses,” ibid., that court, we now know, was wrong. Under the correct view of Missouri law, the 15-year sentence and the life sentence were “alternative punishments . . . prescribed by the legislature for a single criminal act,” ibid. The Court states that “[i]t cannot be suggested seriously that the legislature intended an attempted robbery conviction to suffice as an alternative sanction for murder,” ante, at 384-385. Perhaps not, but it might also have been said in Lange that the legislature did not intend a mere $200 fine for the gravity of offense at issue there. Just as the judge in that case frustrated the probable legislative intent by inadvertently imposing the lesser penalty that was available, unaware that it would preclude the greater, so the judge in the present case frustrated the probable legislative intent by inadvertently entering the lesser conviction and sentence, unaware that it would preclude the greater. But that is beside the point. JONES v. THOMAS 391 376 Scalia, J., dissenting The Double Jeopardy Clause is not a device designed to assure effectuation of legislative intent—but to the contrary is often the means of frustrating it. The relevant question pertaining to legislative intent is not whether the Missouri Legislature intended an attempted armed robbery sentence for the crime of murder, but whether it intended that both a felony-murder sentence and an attempted armed robbery sentence could be imposed for the same crime. The Missouri Supreme Court has said not. See State v. Morgan, 612 S. W. 2d 1 (1981); State v. Olds, 603 S. W. 2d 501, 510 (1980). That being so, if respondent has served one of the two alternative sentences that could lawfully be imposed, he cannot be required to serve the other as well. Second, the Court distinguishes Bradley on the ground that there “[t]he alternative sentences . . . were of a different type, fine and imprisonment,” ante, at 384, so that it would not have been possible to credit the satisfied fine against the as-yet-unserved sentence. It is difficult to imagine, however, why the difference between a credit and a refund (which could have been made in Bradley) should be of constitutional dimensions insofar as the Double Jeopardy Clause is concerned. Bradley, of course, did not rely upon any difference in the nature of the two punishments, but upon the mere fact that one of them had been completely executed. “As the judgment of the court was thus executed so as to be a full satisfaction of one of the alternative punishments of the law, the power of the court was at an end.” 318 U. S., at 52. Likewise Lange: “[I]n that very case, and for that very offence, the prisoner had fully performed, completed, and endured one of the alternative punishments which the law prescribed .... [T]hus . . . [the court’s] power to punish for that offence was at an end. . . . [T]he authority of the court to punish the prisoner was gone. The power was exhausted; its further exercise was prohibited.” 18 Wall., at 176. 392 OCTOBER TERM, 1988 Scalia, J., dissenting 491 U. S. Finally, the Court states that in the multiple punishments context, “‘the Double Jeopardy Clause does no more than prevent the sentencing court from prescribing greater punishment than the legislature intended.’” Ante, at 381, quoting Missouri n. Hunter, 459 U. S. 359, 366 (1983). If that were true it would certainly permit proceedings quite foreign to our criminal-law tradition. If, for example, a judge imposed only a 15-year sentence under a statute that permitted 15 years to life, he could—as far as the Court’s understanding of the Double Jeopardy Clause is concerned—have second thoughts after the defendant has served that time, and add on another 10 years. I am sure that cannot be done, because the Double Jeopardy Clause is a statute of repose for sentences as well as for proceedings. Done is done. The Court is able to quote Hunter for this unusual result only because its quotation is incomplete. What we said in that case, and have subsequently repeated in other cases, is that “[w]ith respect to cumulative sentences imposed in a single trial, the Double Jeopardy Clause does no more than prevent the sentencing court from prescribing greater punishment than the legislature intended.” Ibid. See also id., at 368 (The Double Jeopardy Clause does not “preclud[e] the imposition, in a single trial, of cumulative punishments pursuant to those statutes”) (emphasis added); id., at 368-369 (“Where . . . a legislature specifically authorizes cumulative punishment under two statutes . . . the prosecutor may seek and the trial court or jury may impose cumulative punishment under such statutes in a single trial”) (emphasis added). In both of the cases in which we have applied the Court’s “legislative intent” formulation of the Double Jeopardy Clause to uphold the imposition of multiple penalties, the penalties had been imposed (or would have been imposed) in a single proceeding. See Missouri v. Hunter, supra (defendant convicted of both armed criminal action and the underlying felony of armed robbery in single trial); Ohio v. Johnson, 467 U. S. 493 (1984) (defendant pleaded guilty to two lesser JONES v. THOMAS 393 376 Scalia, J., dissenting offenses and trial court dismissed three greater offenses, stating that prosecution would be barred under Double Jeopardy Clause). But when the added punishment, even though authorized by the legislature, was imposed in a later proceeding, we held that the Double Jeopardy Clause was a bar. In United States v. Halper, 490 U. S. 435, 451, n. 10 (1989), we said: “That the Government seeks the civil penalty in a second proceeding is critical in triggering the protections of the Double Jeopardy Clause. Since a legislature may authorize cumulative punishment under two statutes for a single course of conduct, the multiple-punishment inquiry in the context of a single proceeding focuses on whether the legislature actually authorized the cumulative punishment. See Ohio v. Johnson, 467 U. S. 493, 499-500 (1984). On the other hand, when the Government has already imposed a criminal penalty and seeks to impose additional punishment in a second proceeding, the Double Jeopardy Clause protects against the possibility that the Government is seeking the second punishment because it is dissatisfied with the sanction obtained in the first proceeding.” See also id., at 450 (“In a single proceeding the multiple punishment issue would be limited to ensuring that the total punishment did not exceed that authorized by the legislature”) (emphasis added); ibid. (“Nor does the decision [in Halper] prevent the Government from seeking and obtaining both the full civil penalty and the full range of statutorily authorized civil penalties in the same proceeding") (emphasis added). In the present case, of course, it was not the same proceeding but a second proceeding that added time to the 15-year sentence the defendant had already satisfied for his crime. In those circumstances, our cases establish that the relevant double jeopardy criterion is not only whether the total 394 OCTOBER TERM, 1988 Scalia, J., dissenting 491 U. S. punishment authorized by the legislature has been exceeded, but also whether the addition upsets the defendant’s legitimate “expectation of finality in the original sentence,” United States v. DiFrancesco, 449 U. S. 117, 139 (1980). In the latter case we upheld against a double jeopardy challenge a statute that allowed the Government to appeal as inadequate a District Court’s sentence for a “dangerous special offender.” We did so because, by reason of the appeal provision itself, the defendant had no legitimate expectation of finality in the original sentence. See id., at 136-137. We applied the same rule in Pennsylvania v. Goldhammer, 474 U. S. 28 (1985) (per curiam). There the defendant was convicted of 56 counts of forgery and 56 counts of theft. The trial court sentenced him to a term of imprisonment on one theft count and a term of probation on one forgery count, and suspended sentence on the remaining counts. On appeal, the Supreme Court of Pennsylvania held that the theft count on which the defendant had been sentenced was barred by the applicable statute of limitations, and denied, on double jeopardy grounds, the State’s request that the case be remanded for resentencing on the nonbarred theft counts. We did not reverse that disposition outright, but remanded so that the Supreme Court of Pennsylvania might consider, pursuant to DiFrancesco, “whether the Pennsylvania laws in effect at the time allowed the State to obtain review of the sentences on the counts for which the sentence had been suspended.” 474 U. S., at 30. It is clear from DiFrancesco and Gold-hammer that when a sentence is increased in a second proceeding “the application of the double jeopardy clause . . . turns on the extent and legitimacy of a defendant’s expectation of finality in that sentence. If a defendant has a legitimate expectation of finality, then an increase in that sentence is prohibited . . . .” United States v. Fogel, 264 U. S. App. D. C. 292, 302, 829 F. 2d 77, 87 (1987) (Bork, J.). The principle enunciated in DiFrancesco also explains our decision in Bozza n. United States, 330 U. S. 160 (1947). JONES v. THOMAS 395 376 Scalia, J., dissenting There the defendant was convicted of operating an illegal still, a crime which carried a mandatory sentence of a $100 fine and a term in prison. The trial court originally sentenced the defendant only to the term of imprisonment. When the court realized its mistake five hours later, it recalled the defendant for resentencing and imposed the $100 fine as well. We held that the resentencing did not violate the defendant’s rights under the Double Jeopardy Clause. There, as in DiFrancesco, the defendant could not argue that his legitimate expectation of finality in the original sentence had been violated, because he was charged with knowledge that the court lacked statutory authority to impose the subminimum sentence in the first instance. See 330 U. S., at 166, 167. See also United States v. Arrellano-Rios, 799 F. 2d 520, 524 (CA9 1986) (stating that defendant can have no legitimate expectation of finality in an illegal sentence); United States v. Edmondson, 792 F. 2d 1492, 1496, n. 4 (CA9 1986) (same). Applying DiFrancesco and Bozza here, it seems to me respondent must prevail. There is no doubt that the court had authority to impose the 15-year sentence, and respondent therefore had a legitimate expectation of its finality. There are only two grounds on which that could possibly be contested: (1) that the court had authority to impose a 15-year sentence, but not both a 15-year sentence and life, or (2) that his legitimate expectation was not necessarily 15 years, but rather either 15 years (on the one sentence) or life (on the other sentence). But at least where, as here, the one sentence has been fully served, these alternative approaches to defining his legitimate expectation are ruled out by Bradley. There also it could have been said that the court had no authority to impose both the $500 fine and the six months’ imprisonment; and there also it could have been said that the defendant’s legitimate expectation was not necessarily a $500 fine, but either a $500 fine or six months’ imprisonment. But we in effect rejected those approaches, holding that once the fine had been paid a subsequent proceeding could not re 396 OCTOBER TERM, 1988 Scalia, J., dissenting 491 U. S. place it with the alternative penalty. There is simply no basis for departing from that holding here. The Double Jeopardy Clause is and has always been, not a provision designed to assure reason and justice in the particular case, but the embodiment of technical, prophylactic rules that require the Government to turn square corners. Whenever it is applied to release a criminal deserving of punishment it frustrates justice in the particular case, but for the greater purpose of assuring repose in the totality of criminal prosecutions and sentences. There are many ways in which these technical rules might be designed. We chose one approach in Bradley—undoubtedly not the only possible approach, but also not one that can be said to be clearly wrong. (The fact that it produces a “windfall” separates it not at all from other applications of the double jeopardy guarantee.) With technical rules, above all others, it is imperative that we adhere strictly to what we have stated the rules to be. A technical rule with equitable exceptions is no rule at all. Three strikes is out. The State broke the rules here, and must abide by the result. For these reasons, I believe the Court of Appeals was correct to set aside respondent’s life sentence. I would therefore affirm the judgment of the Court of Appeals, and respectfully dissent from the Court’s disposition of this case.* *1 agree with the Court, ante, at 384-385, n. 3, that the Court of Appeals erred in saying that the State could not resentence or retry respondent for a non-jeopardy-barred lesser included offense, see Morris n. Mathews, 475 U. S. 237 (1986). Since it is undisputed, however, that the State has made no attempt to do that, that portion of the Court of Appeals’ opinion was the purest dictum, and no basis for reversal of its judgment. TEXAS v. JOHNSON 397 Syllabus TEXAS v. JOHNSON CERTIORARI TO THE COURT OF CRIMINAL APPEALS OF TEXAS No. 88-155. Argued March 21, 1989—Decided June 21, 1989 During the 1984 Republican National Convention in Dallas, Texas, respondent Johnson participated in a political demonstration to protest the policies of the Reagan administration and some Dallas-based corporations. After a march through the city streets, Johnson burned an American flag while protesters chanted. No one was physically injured or threatened with injury, although several witnesses were seriously offended by the flag burning. Johnson was convicted of desecration of a venerated object in violation of a Texas statute, and a State Court of Appeals affirmed. However, the Texas Court of Criminal Appeals reversed, holding that the State, consistent with the First Amendment, could not punish Johnson for burning the flag in these circumstances. The court first found that Johnson’s burning of the flag was expressive conduct protected by the First Amendment. The court concluded that the State could not criminally sanction flag desecration in order to preserve the flag as a symbol of national unity. It also held that the statute did not meet the State’s goal of preventing breaches of the peace, since it was not drawn narrowly enough to encompass only those flag burnings that would likely result in a serious disturbance, and since the flag burning in this case did not threaten such a reaction. Further, it stressed that another Texas statute prohibited breaches of the peace and could be used to prevent disturbances without punishing this flag desecration. Held: Johnson’s conviction for flag desecration is inconsistent with the First Amendment. Pp. 402-420. (a) Under the circumstances, Johnson’s burning of the flag constituted expressive conduct, permitting him to invoke the First Amendment. The State conceded that the conduct was expressive. Occurring as it did at the end of a demonstration coinciding with the Republican National Convention, the expressive, overtly political nature of the conduct was both intentional and overwhelmingly apparent. Pp. 402-406. (b) Texas has not asserted an interest in support of Johnson’s conviction that is unrelated to the suppression of expression and would therefore permit application of the test set forth in United States v. O’Brien, 391 U. S. 367, whereby an important governmental interest in regulating nonspeech can justify incidental limitations on First Amendment freedoms when speech and nonspeech elements are combined in the same course of conduct. An interest in preventing breaches of the peace is not implicated on this record. Expression may not be prohib 398 OCTOBER TERM, 1988 Syllabus 491 U. S. ited on the basis that an audience that takes serious offense to the expression may disturb the peace, since the government cannot assume that every expression of a provocative idea will incite a riot but must look to the actual circumstances surrounding the expression. Johnson’s expression of dissatisfaction with the Federal Government’s policies also does not fall within the class of “fighting words” likely to be seen as a direct personal insult or an invitation to exchange fisticuffs. This Court’s holding does not forbid a State to prevent “imminent lawless action” and, in fact, Texas has a law specifically prohibiting breaches of the peace. Texas’ interest in preserving the flag as a symbol of nationhood and national unity is related to expression in this case and, thus, falls outside the O'Brien test. Pp. 406-410. (c) The latter interest does not justify Johnson’s conviction. The restriction on Johnson’s political expression is content based, since the Texas statute is not aimed at protecting the physical integrity of the flag in all circumstances, but is designed to protect it from intentional and knowing abuse that causes serious offense to others. It is therefore subject to “the most exacting scrutiny.” Boos v. Barry, 485 U. S. 312. The government may not prohibit the verbal or nonverbal expression of an idea merely because society finds the idea offensive or disagreeable, even where our flag is involved. Nor may a State foster its own view of the flag by prohibiting expressive conduct relating to it, since the government may not permit designated symbols to be used to communicate a limited set of messages. Moreover, this Court will not create an exception to these principles protected by the First Amendment for the American flag alone. Pp. 410-422. 755 S. W. 2d 92, affirmed. Brennan, J., delivered the opinion of the Court, in which Marshall, Blackmun, Scalia, and Kennedy, JJ., joined. Kennedy, J., filed a concurring opinion, post, p. 420. Rehnquist, C. J., filed a dissenting opinion, in which White and O’Connor, JJ., joined, post, p. 421. Stevens, J., filed a dissenting opinion, post, p. 436. Kathi Alyce Drew argued the cause for petitioner. With her on the briefs were John Vance and Dolena T. Westergard. William M. Kunstler argued the cause for respondent. With him on the brief was David D. Cole.* *Briefs of amici curiae urging reversal were filed for the Legal Affairs Council by Wyatt B. Durrette, Jr., and Bradley B. Cavedo; and for the Washington Legal Foundation by Daniel J. Popeo and Paul D. Kamenar. Briefs of amici curiae urging affirmance were filed for the American Civil Liberties Union et al. by Peter Linzer, James C. Harrington, and TEXAS V. JOHNSON 399 397 Opinion of the Court Justice Brennan delivered the opinion of the Court. After publicly burning an American flag as a means of political protest, Gregory Lee Johnson was convicted of desecrating a flag in violation of Texas law. This case presents the question whether his conviction is consistent with the First Amendment. We hold that it is not. I While the Republican National Convention was taking place in Dallas in 1984, respondent Johnson participated in a political demonstration dubbed the “Republican War Chest Tour.” As explained in literature distributed by the demonstrators and in speeches made by them, the purpose of this event was to protest the policies of the Reagan administration and of certain Dallas-based corporations. The demonstrators marched through the Dallas streets, chanting political slogans and stopping at several corporate locations to stage “die-ins” intended to dramatize the consequences of nuclear war. On several occasions they spray-painted the walls of buildings and overturned potted plants, but Johnson himself took no part in such activities. He did, however, accept an American flag handed to him by a fellow protestor who had taken it from a flagpole outside one of the targeted buildings. The demonstration ended in front of Dallas City Hall, where Johnson unfurled the American flag, doused it with kerosene, and set it on fire. While the flag burned, the protestors chanted: “America, the red, white, and blue, we spit on you.” After the demonstrators dispersed, a witness to the flag burning collected the flag’s remains and buried them in his backyard. No one was physically injured or threatened with injury, though several witnesses testified that they had been seriously offended by the flag burning. Steven R. Shapiro; for the Christie Institute et al. by James C. Goodale; and for Jasper Johns et al. by Robert G. Sugarman and Gloria C. Phares. 400 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Of the approximately 100 demonstrators, Johnson alone was charged with a crime. The only criminal offense with which he was charged was the desecration of a venerated object in violation of Tex. Penal Code Ann. § 42.09(a)(3) (1989).1 After a trial, he was convicted, sentenced to one year in prison, and fined $2,000. The Court of Appeals for the Fifth District of Texas at Dallas affirmed Johnson’s conviction, 706 S. W. 2d 120 (1986), but the Texas Court of Criminal Appeals reversed, 755 S. W. 2d 92 (1988), holding that the State could not, consistent with the First Amendment, punish Johnson for burning the flag in these circumstances. The Court of Criminal Appeals began by recognizing that Johnson’s conduct was symbolic speech protected by the First Amendment: “Given the context of an organized demonstration, speeches, slogans, and the distribution of literature, anyone who observed appellant’s act would have understood the message that appellant intended to convey. The act for which appellant was convicted was clearly ‘speech’ contemplated by the First Amendment.” Id., at 95. To justify Johnson’s conviction for engaging in symbolic speech, the State asserted two interests: preserving the flag as a symbol of national unity and preventing breaches of the peace. The Court of Criminal Appeals held that neither interest supported his conviction. ‘Texas Penal Code Ann. §42.09 (1989) provides in full: “§ 42.09. Desecration of Venerated Object “(a) A person commits an offense if he intentionally or knowingly desecrates: “(1) a public monument; “(2) a place of worship or burial; or “(3) a state or national flag. “(b) For purposes of this section, ‘desecrate’ means deface, damage, or otherwise physically mistreat in a way that the actor knows will seriously offend one or more persons likely to observe or discover his action. “(c) An offense under this section is a Class A misdemeanor.” TEXAS V. JOHNSON 401 397 Opinion of the Court Acknowledging that this Court had not yet decided whether the Government may criminally sanction flag desecration in order to preserve the flag’s symbolic value, the Texas court nevertheless concluded that our decision in West Virginia Board of Education n. Barnette, 319 U. S. 624 (1943), suggested that furthering this interest by curtailing speech was impermissible. “Recognizing that the right to differ is the center piece of our First Amendment freedoms,” the court explained, “a government cannot mandate by fiat a feeling of unity in its citizens. Therefore, that very same government cannot carve out a symbol of unity and prescribe a set of approved messages to be associated with that symbol when it cannot mandate the status or feeling the symbol purports to represent.” 755 S. W. 2d, at 97. Noting that the State had not shown that the flag was in “grave and immediate danger,” Barnette, supra, at 639, of being stripped of its symbolic value, the Texas court also decided that the flag’s special status was not endangered by Johnson’s conduct. 755 S. W. 2d, at 97. As to the State’s goal of preventing breaches of the peace, the court concluded that the flag-desecration statute was not drawn narrowly enough to encompass only those flag burnings that were likely to result in a serious disturbance of the peace. And in fact, the court emphasized, the flag burning in this particular case did not threaten such a reaction. “‘Serious offense’ occurred,” the court admitted, “but there was no breach of peace nor does the record reflect that the situation was potentially explosive. One cannot equate ‘seri-ous’offense’ with incitement to breach the peace.” Id., at 96. The court also stressed that another Texas statute, Tex. Penal Code Ann. §42.01 (1989), prohibited breaches of the peace. Citing Boos v. Barry, 485 U. S. 312 (1988), the court decided that §42.01 demonstrated Texas’ ability to prevent disturbances of the peace without punishing this flag desecration. 755 S. W. 2d, at 96. 402 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Because it reversed Johnson’s conviction on the ground that §42.09 was unconstitutional as applied to him, the state court did not address Johnson’s argument that the statute was, on its face, unconstitutionally vague and overbroad. We granted certiorari, 488 U. S. 907 (1988), and now affirm. II Johnson was convicted of flag desecration for burning the flag rather than for uttering insulting words.2 This fact 2 Because the prosecutor’s closing argument observed that Johnson had led the protestors in chants denouncing the flag while it burned, Johnson suggests that he may have been convicted for uttering critical words rather than for burning the flag. Brief for Respondent 33-34. He relies on Street v. New York, 394 U. S. 576, 578 (1969), in which we reversed a conviction obtained under a New York statute that prohibited publicly defying or casting contempt on the flag “either by words or act” because we were persuaded that the defendant may have been convicted for his words alone. Unlike the law we faced in Street, however, the Texas flag-desecration statute does not on its face permit conviction for remarks critical of the flag, as Johnson himself admits. See Brief for Respondent 34. Nor was the jury in this case told that it could convict Johnson of flag desecration if it found only that he had uttered words critical of the flag and its referents. Johnson emphasizes, though, that the jury was instructed—according to Texas’ law of parties—that “‘a person is criminally responsible for an offense committed by the conduct of another if acting with intent to promote or assist the commission of the offense, he solicits, encourages, directs, aids, or attempts to aid the other person to commit the offense.’” Id., at 2, n. 2, quoting 1 Record 49. The State offered this instruction because Johnson’s defense was that he was not the person who had burned the flag. Johnson did not object to this instruction at trial, and although he challenged it on direct appeal, he did so only on the ground that there was insufficient evidence to support it. 706 S. W. 2d 120, 124 (Tex. App. 1986). It is only in this Court that Johnson has argued that the law-of-parties instruction might have led the jury to convict him for his words alone. Even if we were to find that this argument is properly raised here, however, we would conclude that it has no merit in these circumstances. The instruction would not have permitted a conviction merely for the pejorative nature of Johnson’s words, and those words themselves did not encourage the burning of the flag as the instruction seems to require. Given the additional fact that “the bulk of the State’s TEXAS v. JOHNSON 403 397 Opinion of the Court somewhat complicates our consideration of his conviction under the First Amendment. We must first determine whether Johnson’s burning of the flag constituted expressive conduct, permitting him to invoke the First Amendment in challenging his conviction. See, e. g., Spence n. Washington, 418 U. S. 405, 409-411 (1974). If his conduct was expressive, we next decide whether the State’s regulation is related to the suppression of free expression. See, e. g., United States v. O’Brien, 391 U. S. 367, 377 (1968); Spence, supra, at 414, n. 8. If the State’s regulation is not related to expression, then the less stringent standard we announced in United States v. O’Brien for regulations of noncommuni-cative conduct controls. See O’Brien, supra, at 377. If it is, then we are outside of O’Brien’s test, and we must ask whether this interest justifies Johnson’s conviction under a more demanding standard.3 See Spence, supra, at 411. A argument was premised on Johnson’s culpability as a sole actor,” ibid., we find it too unlikely that the jury convicted Johnson on the basis of this alternative theory to consider reversing his conviction on this ground. 3 Although Johnson has raised a facial challenge to Texas’ flag-desecration statute, we choose to resolve this case on the basis of his claim that the statute as applied to him violates the First Amendment. Section 42.09 regulates only physical conduct with respect to the flag, not the written or spoken word, and although one violates the statute only if one “knows” that one’s physical treatment of the flag “will seriously offend one or more persons likely to observe or discover his action,” Tex. Penal Code Ann. § 42.09(b) (1989), this fact does not necessarily mean that the statute applies only to expressive conduct protected by the First Amendment. Cf. Smith v. Goguen, 415 U. S. 566, 588 (1974) (White, J., concurring in judgment) (statute prohibiting “contemptuous” treatment of flag encompasses only expressive conduct). A tired person might, for example, drag a flag through the mud, knowing that this conduct is likely to offend others, and yet have no thought of expressing any idea; neither the language nor the Texas courts’ interpretations of the statute precludes the possibility that such a person would be prosecuted for flag desecration. Because the prosecution of a person who had not engaged in expressive conduct would pose a different case, and because this case may be disposed of on narrower grounds, we address only Johnson’s claim that § 42.09 as applied to political expression like his violates the First Amendment. 404 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. third possibility is that the State’s asserted interest is simply not implicated on these facts, and in that event the interest drops out of the picture. See 418 U. S., at 414, n. 8. The First Amendment literally forbids the abridgment only of “speech,” but we have long recognized that its protection does not end at the spoken or written word. While we have rejected “the view that an apparently limitless variety of conduct can be labeled ‘speech’ whenever the person engaging in the conduct intends thereby to express an idea,” United States v. O’Brien, supra, at 376, we have acknowledged that conduct may be “sufficiently imbued with elements of communication to fall within the scope of the First and Fourteenth Amendments,” Spence, supra, at 409. In deciding whether particular conduct possesses sufficient communicative elements to bring the First Amendment into play, we have asked whether “[a]n intent to convey a particularized message was present, and [whether] the likelihood was great that the message would be understood by those who viewed it.” 418 U. S., at 410-411. Hence, we have recognized the expressive nature of students’ wearing of black armbands to protest American military involvement in Vietnam, Tinker v. Des Moines Independent Community School Dist., 393 U. S. 503, 505 (1969); of a sit-in by blacks in a “whites only” area to protest segregation, Brown v. Louisiana, 383 U. S. 131, 141-142 (1966); of the wearing of American military uniforms in a dramatic presentation criticizing American involvement in Vietnam, Schacht v. United States, 398 U. S. 58 (1970); and of picketing about a wide variety of causes, see, e. g., Food Employees n, Logan Valley Plaza, Inc., 391 U. S. 308, 313-314 (1968); United States v. Grace, 461 U. S. 171, 176 (1983). Especially pertinent to this case are our decisions recognizing the communicative nature of conduct relating to flags. Attaching a peace sign to the flag, Spence, supra, at 409-410; refusing to salute the flag, Barnette, 319 U. S., at 632; and displaying a red flag, Stromberg n. California, 283 U. S. 359, TEXAS V. JOHNSON 405 397 Opinion of the Court 368-369 (1931), we have held, all may find shelter under the First Amendment. See also Smith v. Goguen, 415 U. S. 566, 588 (1974) (White, J., concurring in judgment) (treating flag “contemptuously” by wearing pants with small flag sewn into their seat is expressive conduct). That we have had little difficulty identifying an expressive element in conduct relating to flags should not be surprising. The very purpose of a national flag is to serve as a symbol of our country; it is, one might say, “the one visible manifestation of two hundred years of nationhood.” Id., at 603 (Rehnquist, J., dissenting). Thus, we have observed: “[T]he flag salute is a form of utterance. Symbolism is a primitive but effective way of communicating ideas. The use of an emblem or flag to symbolize some system, idea, institution, or personality, is a short cut from mind to mind. Causes and nations, political parties, lodges and ecclesiastical groups seek to knit the loyalty of their followings to a flag or banner, a color or design.” Barnette, supra, at 632. Pregnant with expressive content, the flag as readily signifies this Nation as does the combination of letters found in “America.” We have not automatically concluded, however, that any action taken with respect to our flag is expressive. Instead, in characterizing such action for First Amendment purposes, we have considered the context in which it occurred. In Spence, for example, we emphasized that Spence’s taping of a peace sign to his flag was “roughly simultaneous with and concededly triggered by the Cambodian incursion and the Kent State tragedy.” 418 U. S., at 410. The State of Washington had conceded, in fact, that Spence’s conduct was a form of communication, and we stated that “the State’s concession is inevitable on this record.” Id., at 409. The State of Texas conceded for purposes of its oral argument in this case that Johnson’s conduct was expressive conduct, Tr. of Oral Arg. 4, and this concession seems to us as 406 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. prudent as was Washington’s in Spence. Johnson burned an American flag as part—indeed, as the culmination—of a political demonstration that coincided with the convening of the Republican Party and its renomination of Ronald Reagan for President. The expressive, overtly political nature of this conduct was both intentional and overwhelmingly apparent. At his trial, Johnson explained his reasons for burning the flag as follows: “The American Flag was burned as Ronald Reagan was being renominated as President. And a more powerful statement of symbolic speech, whether you agree with it or not, couldn’t have been made at that time. It’s quite a just position [juxtaposition]. We had new patriotism and no patriotism.” 5 Record 656. In these circumstances, Johnson’s burning of the flag was conduct “sufficiently imbued with elements of communication,” Spence, 418 U. S., at 409, to implicate the First Amendment. Ill The government generally has a freer hand in restricting expressive conduct than it has in restricting the written or spoken word. See O’Brien, 391 U. S. at 376-377; Clark v. Community for Creative Non-Violence, 468 U. S. 288, 293 (1984); Dallas v. Stanglin, 490 U. S. 19, 25 (1989). It may not, however, proscribe particular conduct because it has expressive elements. “[W]hat might be termed the more generalized guarantee of freedom of expression makes the communicative nature of conduct an inadequate basis for singling out that conduct for proscription. A law directed at the communicative nature of conduct must, like a law directed at speech itself, be justified by the substantial showing of need that the First Amendment requires.” Community for Creative Non-Violence n. Watt, 227 U. S. App. D. C. 19, 55-56, 703 F. 2d 586, 622-623 (1983) (Scalia, J., dissenting) (emphasis in original), rev’d sub nom. Clark v. Community for Creative Non-Violence, supra. It is, in short, not simply the verbal or nonverbal nature of the expression, but the govern TEXAS v. JOHNSON 407 397 Opinion of the Court mental interest at stake, that helps to determine whether a restriction on that expression is valid. Thus, although we have recognized that where “‘speech’ and ‘nonspeech’ elements are combined in the same course of conduct, a sufficiently important governmental interest in regulating the nonspeech element can justify incidental limitations on First Amendment freedoms,” O’Brien, supra, at 376, we have limited the applicability of O’Brien’s relatively lenient standard to those cases in which “the governmental interest is unrelated to the suppression of free expression.” Id., at 377; see also Spence, supra, at 414, n. 8. In stating, moreover, that O’Brien’s test “in the last analysis is little, if any, different from the standard applied to time, place, or manner restrictions,” Clark, supra, at 298, we have highlighted the requirement that the governmental interest in question be unconnected to expression in order to come under O’Brien’s less demanding rule. In order to decide whether O’Brien’s test applies here, therefore, we must decide whether Texas has asserted an interest in support of Johnson’s conviction that is unrelated to the suppression of expression. If we find that an interest asserted by the State is simply not implicated on the facts before us, we need not ask whether O’Brien’s test applies. See Spence, supra, at 414, n. 8. The State offers two separate interests to justify this conviction: preventing breaches of the peace and preserving the flag as a symbol of nationhood and national unity. We hold that the first interest is not implicated on this record and that the second is related to the suppression of expression. A Texas claims that its interest in preventing breaches of the peace justifies Johnson’s conviction for flag desecration.4 4 Relying on our decision in Boos v. Barry, 485 U. S. 312 (1988), Johnson argues that this state interest is related to the suppression of free expression within the meaning of United States v. O’Brien, 391 U. S. 367 (1968). He reasons that the violent reaction to flag burnings feared by 408 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. However, no disturbance of the peace actually occurred or threatened to occur because of Johnson’s burning of the flag. Although the State stresses the disruptive behavior of the protestors during their march toward City Hall, Brief for Petitioner 34-36, it admits that “no actual breach of the peace occurred at the time of the flagburning or in response to the flagburning.” Id., at 34. The State’s emphasis on the protestors’ disorderly actions prior to arriving at City Hall is not only somewhat surprising given that no charges were brought on the basis of this conduct, but it also fails to show that a disturbance of the peace was a likely reaction to Johnson’s conduct. The only evidence offered by the State at trial to show the reaction to Johnson’s actions was the testimony of several persons who had been seriously offended by the flag burning. Id., at 6-7. The State’s position, therefore, amounts to a claim that an audience that takes serious offense at particular expression is necessarily likely to disturb the peace and that the expression may be prohibited on this basis.5 Our precedents do not countenance such a presumption. On the contrary, they recognize that a principal “function of free speech under our system of government is to invite dispute. It may indeed best serve its high purpose when it induces a condition of unrest, creates dissatisfaction with conditions as they are, or Texas would be the result of the message conveyed by them, and that this fact connects the State’s interest to the suppression of expression. Brief for Respondent 12, n. 11. This view has found some favor in the lower courts. See Monroe v. State Court of Fulton County, 739 F. 2d 568, 574-575 (CA11 1984). Johnson’s theory may overread Boos insofar as it suggests that a desire to prevent a violent audience reaction is “related to expression” in the same way that a desire to prevent an audience from being offended is “related to expression.” Because we find that the State’s interest in preventing breaches of the peace is not implicated on these facts, however, we need not venture further into this area. 5 There is, of course, a tension between this argument and the State’s claim that one need not actually cause serious offense in order to violate § 42.09. See Brief for Petitioner 44. TEXAS v. JOHNSON 409 397 Opinion of the Court even stirs people to anger.” Terminiello v. Chicago, 337 U. S. 1, 4 (1949). See also Cox v. Louisiana, 379 U. S. 536, 551 (1965); Tinker v. Des Moines Independent Community School Dist. 393 U. S., at 508-509; Coates v. Cincinnati, 402 U. S. 611, 615 (1971); Hustler Magazine, Inc. v. Falwell, 485 U. S. 46, 55-56 (1988). It would be odd indeed to conclude both that “if it is the speaker’s opinion that gives offense, that consequence is a reason for according it constitutional protection,” FCC v. Pacifica Foundation, 438 U. S. 726, 745 (1978) (opinion of Stevens, J.), and that the government may ban the expression of certain disagreeable ideas on the unsupported presumption that their very disagreeableness will provoke violence. Thus, we have not permitted the government to assume that every expression of a provocative idea will incite a riot, but have instead required careful consideration of the actual circumstances surrounding such expression, asking whether the expression “is directed to inciting or producing imminent lawless action and is likely to incite or produce such action.” Brandenburg v. Ohio, 395 U. S. 444, 447 (1969) (reviewing circumstances surrounding rally and speeches by Ku Klux Klan). To accept Texas’ arguments that it need only demonstrate “the potential for a breach of the peace,” Brief for Petitioner 37, and that every flag burning necessarily possesses that potential, would be to eviscerate our holding in Brandenburg. This we decline to do. Nor does Johnson’s expressive conduct fall within that small class of “fighting words” that are “likely to provoke the average person to retaliation, and thereby cause a breach of the peace.” Chaplinsky v. New Hampshire, 315 U. S. 568, 574 (1942). No reasonable onlooker would have regarded Johnson’s generalized expression of dissatisfaction with the policies of the Federal Government as a direct personal insult or an invitation to exchange fisticuffs. See id., at 572-573; Cantwell v. Connecticut, 310 U. S. 296, 309 (1940); FCC n. Pacifica Foundation, supra, at 745 (opinion of Stevens, J.). 410 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. We thus conclude that the State’s interest in maintaining order is not implicated on these facts. The State need not worry that our holding will disable it from preserving the peace. We do not suggest that the First Amendment forbids a State to prevent “imminent lawless action.” Brandenburg, supra, at 447. And, in fact, Texas already has a statute specifically prohibiting breaches of the peace, Tex. Penal Code Ann. §42.01 (1989), which tends to confirm that Texas need not punish this flag desecration in order to keep the peace. See Boos v. Barry, 485 U. S., at 327-329. B The State also asserts an interest in preserving the flag as a symbol of nationhood and national unity. In Spence, we acknowledged that the government’s interest in preserving the flag’s special symbolic value “is directly related to expression in the context of activity” such as affixing a peace symbol to a flag. 418 U. S., at 414, n. 8. We are equally persuaded that this interest is related to expression in the case of Johnson’s burning of the flag. The State, apparently, is concerned that such conduct will lead people to believe either that the flag does not stand for nationhood and national unity, but instead reflects other, less positive concepts, or that the concepts reflected in the flag do not in fact exist, that is, that we do not enjoy unity as a Nation. These concerns blossom only when a person’s treatment of the flag communicates some message, and thus are related “to the suppression of free expression” within the meaning of O’Brien. We are thus outside of O’Brien’s, test altogether. IV It remains to consider whether the State’s interest in preserving the flag as a symbol of nationhood and national unity justifies Johnson’s conviction. As in Spence, “[w]e are confronted with a case of prosecution for the expression of an idea through activity,” and “[a]c-cordingly, we must examine with particular care the inter TEXAS v. JOHNSON 411 397 Opinion of the Court ests advanced by [petitioner] to support its prosecution.” 418 U. S., at 411. Johnson was not, we add, prosecuted for the expression of just any idea; he was prosecuted for his expression of dissatisfaction with the policies of this country, expression situated at the core of our First Amendment values. See, e. g., Boos v. Barry, supra, at 318; Frisby v. Schultz, 487 U. S. 474, 479 (1988). Moreover, Johnson was prosecuted because he knew that his politically charged expression would cause “serious offense.” If he had burned the flag as a means of disposing of it because it was dirty or torn, he would not have been convicted of flag desecration under this Texas law: federal law designates burning as the preferred means of disposing of a flag “when it is in such condition that it is no longer a fitting emblem for display,” 36 U. S. C. § 176(k), and Texas has no quarrel with this means of disposal. Brief for Petitioner 45. The Texas law is thus not aimed at protecting the physical integrity of the flag in all circumstances, but is designed instead to protect it only against impairments that would cause serious offense to others.6 Texas concedes as much: “Section 42.09(b) reaches only those severe acts of physical abuse of the flag carried out in a way likely to be offensive. The statute mandates intentional or knowing abuse, that is, the land of mistreatment that is not innocent, but rather is intentionally designed to seriously offend other individuals.” Id., at 44. Whether Johnson’s treatment of the flag violated Texas law thus depended on the likely communicative impact of his expressive conduct.7 Our decision in Boos v. Barry, supra, 6Cf. Smith v. Goguen, 415 U. S., at 590-591 (Blackmun, J., dissenting) (emphasizing that lower court appeared to have construed state statute so as to protect physical integrity of the flag in all circumstances); id., at 597-598 (Rehnquist, J., dissenting) (same). 7 Texas suggests that Johnson’s conviction did not depend on the onlookers’ reaction to the flag burning because § 42.09 is violated only when a person physically mistreats the flag in a way that he “knows will seriously offend one or more persons likely to observe or discover his action.” Tex. 412 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. tells us that this restriction on Johnson’s expression is content based. In Boos, we considered the constitutionality of a law prohibiting “the display of any sign within 500 feet of a foreign embassy if that sign tends to bring that foreign government into ‘public odium’ or ‘public disrepute.’” Id., at 315. Rejecting the argument that the law was content neutral because it was justified by “our international law obligation to shield diplomats from speech that offends their dignity,” id., at 320, we held that “[t]he emotive impact of speech on its audience is not a ‘secondary effect’ ” unrelated to the content of the expression itself. Id., at 321 (plurality opinion); see also id., at 334 (Brennan, J., concurring in part and concurring in judgment). According to the principles announced in Boos, Johnson’s political expression was restricted because of the content of the message he conveyed. We must therefore subject the State’s asserted interest in preserving the special symbolic character of the flag to “the most exacting scrutiny.” Boos v. Barry, supra, at 321.8 Penal Code Ann. § 42.09(b) (1989) (emphasis added). “The ‘serious offense’ language of the statute,” Texas argues, “refers to an individual’s intent and to the manner in which the conduct is effectuated, not to the reaction of the crowd.” Brief for Petitioner 44. If the statute were aimed only at the actor’s intent and not at the communicative impact of his actions, however, there would be little reason for the law to be triggered only when an audience is “likely” to be present. At Johnson’s trial, indeed, the State itself seems not to have seen the distinction between knowledge and actual communicative impact that it now stresses; it proved the element of knowledge by offering the testimony of persons who had in fact been seriously offended by Johnson’s conduct. Id., at 6-7. In any event, we find the distinction between Texas’ statute and one dependent on actual audience reaction too precious to be of constitutional significance. Both kinds of statutes clearly are aimed at protecting onlookers from being offended by the ideas expressed by the prohibited activity. 8 Our inquiry is, of course, bounded by the particular facts of this case and by the statute under which Johnson was convicted. There was no evidence that Johnson himself stole the flag he burned, Tr. of Oral Arg. 17, nor did the prosecution or the arguments urged in support of it depend on TEXAS V. JOHNSON 413 397 Opinion of the Court Texas argues that its interest in preserving the flag as a symbol of nationhood and national unity survives this close analysis. Quoting extensively from the writings of this Court chronicling the flag’s historic and symbolic role in our society, the State emphasizes the “‘special place’” reserved for the flag in our Nation. Brief for Petitioner 22, quoting Smith v. Goguen, 415 U. S., at 601 (Rehnquist, J., dissenting). The State’s argument is not that it has an interest simply in maintaining the flag as a symbol of something, no matter what it symbolizes; indeed, if that were the State’s position, it would be difficult to see how that interest is endangered by highly symbolic conduct such as Johnson’s. Rather, the State’s claim is that it has an interest in preserving the flag as a symbol of nationhood and national unity, a symbol with a determinate range of meanings. Brief for Petitioner 20-24. According to Texas, if one physically treats the flag in a way that would tend to cast doubt on either the idea that nationhood and national unity are the flag’s referents or that national unity actually exists, the message conveyed thereby is a harmful one and therefore may be prohibited.9 the theory that the flag was stolen. Ibid. Thus, our analysis does not rely on the way in which the flag was acquired, and nothing in our opinion should be taken to suggest that one is free to steal a flag so long as one later uses it to communicate an idea. We also emphasize that Johnson was prosecuted only for flag desecration—not for trespass, disorderly conduct, or arson. ’Texas claims that “Texas is not endorsing, protecting, avowing or prohibiting any particular philosophy.” Brief for Petitioner 29. If Texas means to suggest that its asserted interest does not prefer Democrats over Socialists, or Republicans over Democrats, for example, then it is beside the point, for Johnson does not rely on such an argument. He argues instead that the State’s desire to maintain the flag as a symbol of nationhood and national unity assumes that there is only one proper view of the flag. Thus, if Texas means to argue that its interest does not prefer any viewpoint over another, it is mistaken; surely one’s attitude toward the flag and its referents is a viewpoint. 414 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. If there is a bedrock principle underlying the First Amendment, it is that the government may not prohibit the expression of an idea simply because society finds the idea itself offensive or disagreeable. See, e. g., Hustler Magazine, Inc. v. Falwell, 485 U. S., at 55-56; City Council of Los Angeles v. Taxpayers for Vincent, 466 U. S. 789, 804 (1984); Bolger v. Youngs Drug Products Corp., 463 U. S. 60, 65, 72 (1983); Carey v. Brown, 447 U. S. 455, 462-463 (1980); FCC n. Pacifica Foundation, 438 U. S., at 745-746; Young v. American Mini Theatres, Inc., 427 U. S. 50, 63-65, 67-68 (1976) (plurality opinion); Buckley n. Valeo, 424 U. S. 1, 16-17 (1976); Grayned v. Rockford, 408 U. S. 104, 115 (1972); Police Dept, of Chicago v. Mosley, 408 U. S. 92, 95 (1972); Bachellar v. Maryland, 397 U. S. 564, 567 (1970); O’Brien, 391 U. S., at 382; Brown v. Louisiana, 383 U. S., at 142-143; Stromberg v. California, 283 U. S., at 368-369. We have not recognized an exception to this principle even where our flag has been involved. In Street v. New York, 394 U. S. 576 (1969), we held that a State may not criminally punish a person for uttering words critical of the flag. Rejecting the argument that the conviction could be sustained on the ground that Street had “failed to show the respect for our national symbol which may properly be demanded of every citizen,” we concluded that “the constitutionally guaranteed ‘freedom to be intellectually . . . diverse or even contrary,’ and the ‘right to differ as to things that touch the heart of the existing order,’ encompass the freedom to express publicly one’s opinions about our flag, including those opinions which are defiant or contemptuous.” Id., at 593, quoting Barnette, 319 U. S., at 642. Nor may the government, we have held, compel conduct that would evince respect for the flag. “To sustain the compulsory flag salute we are required to say that a Bill of Rights which guards the individual’s right to speak his own mind, left it open to public authorities to compel him to utter what is not in his mind.” Id., at 634. TEXAS v. JOHNSON 415 397 Opinion of the Court In holding in Barnette that the Constitution did not leave this course open to the government, Justice Jackson described one of our society’s defining principles in words deserving of their frequent repetition: “If there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion or force citizens to confess by word or act their faith therein.” Id., at 642. In Spence, we held that the same interest asserted by Texas here was insufficient to support a criminal conviction under a flag-misuse statute for the taping of a peace sign to an American flag. “Given the protected character of [Spence’s] expression and in light of the fact that no interest the State may have in preserving the physical integrity of a privately owned flag was significantly impaired on these facts,” we held, “the conviction must be invalidated.” 418 U. S., at 415. See also Goguen, supra, at 588 (White, J., concurring in judgment) (to convict person who had sewn a flag onto the seat of his pants for “contemptuous” treatment of the flag would be “[t]o convict not to protect the physical integrity or to protect against acts interfering with the proper use of the flag, but to punish for communicating ideas unacceptable to the controlling majority in the legislature”). In short, nothing in our precedents suggests that a State may foster its own view of the flag by prohibiting expressive conduct relating to it.10 To bring its argument outside our 10 Our decision in Halter v. Nebraska, 205 U. S. 34 (1907), addressing the validity of a state law prohibiting certain commercial uses of the flag, is not to the contrary. That case was decided “nearly 20 years before the Court concluded that the First Amendment applies to the States by virtue of the Fourteenth Amendment.” Spence v. Washington, 418 U. S. 405, 413, n. 7 (1974). More important, as we continually emphasized in Halter itself, that case involved purely commercial rather than political speech. 205 U. S., at 38, 41, 42, 45. Nor does San Francisco Arts & Athletics, Inc. v. United States Olympic Committee, 483 U. S. 522, 524 (1987), addressing the validity of Congress’ decision to “authoriz[e] the United States Olympic Committee to prohibit 416 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. precedents, Texas attempts to convince us that even if its interest in preserving the flag’s symbolic role does not allow it to prohibit words or some expressive conduct critical of the flag, it does permit it to forbid the outright destruction of the flag. The State’s argument cannot depend here on the distinction between written or spoken words and nonverbal conduct. That distinction, we have shown, is of no moment where the nonverbal conduct is expressive, as it is here, and where the regulation of that conduct is related to expression, as it is here. See supra, at 402-403. In addition, both Barnette and Spence involved expressive conduct, not only verbal communication, and both found that conduct protected. Texas’ focus on the precise nature of Johnson’s expression, moreover, misses the point of our prior decisions: their enduring lesson, that the government may not prohibit expression simply because it disagrees with its message, is not dependent on the particular mode in which one chooses to express an idea.11 If we were to hold that a State may forbid flag burning wherever it is likely to endanger the flag’s symbolic role, but allow it wherever burning a flag promotes that role—as where, for example, a person ceremoniously bums a dirty flag—we would be saying that when it comes to impairing the flag’s physical integrity, the flag itself may be used as certain commercial and promotional uses of the word ‘Olympic,’” relied upon by The Chief Justice’s dissent, post, at 429, even begin to tell us whether the government may criminally punish physical conduct towards the flag engaged in as a means of political protest. “The Chief Justice’s dissent appears to believe that Johnson’s conduct may be prohibited and, indeed, criminally sanctioned, because “his act . . . conveyed nothing that could not have been conveyed and was not conveyed just as forcefully in a dozen different ways.” Post, at 431. Not only does this assertion sit uneasily next to the dissent’s quite correct reminder that the flag occupies a unique position in our society—which demonstrates that messages conveyed without use of the flag are not “just as forcefu[l]” as those conveyed with it—but it also ignores the fact that, in Spence, supra, we “rejected summarily” this very claim. See 418 U. S., at 411, n. 4. TEXAS V. JOHNSON 417 397 Opinion of the Court a symbol—as a substitute for the written or spoken word or a “short cut from mind to mind”—only in one direction. We would be permitting a State to “prescribe what shall be orthodox” by saying that one may burn the flag to convey one’s attitude toward it and its referents only if one does not endanger the flag’s representation of nationhood and national unity. We never before have held that the Government may ensure that a symbol be used to express only one view of that symbol or its referents. Indeed, in Schacht v. United States, we invalidated a federal statute permitting an actor portraying a member of one of our Armed Forces to “ ‘wear the uniform of that armed force if the portrayal does not tend to discredit that armed force.’” 398 U. S., at 60, quoting 10 U. S. C. § 772(f). This proviso, we held, “which leaves Americans free to praise the war in Vietnam but can send persons like Schacht to prison for opposing it, cannot survive in a country which has the First Amendment.” Id., at 63. We perceive no basis on which to hold that the principle underlying our decision in Schacht does not apply to this case. To conclude that the government may permit designated symbols to be used to communicate only a limited set of messages would be to enter territory having no discernible or defensible boundaries. Could the government, on this theory, prohibit the burning of state flags? Of copies of the Presidential seal? Of the Constitution? In evaluating these choices under the First Amendment, how would we decide which symbols were sufficiently special to warrant this unique status? To do so, we would be forced to consult our own political preferences, and impose them on the citizenry, in the very way that the First Amendment forbids us to do. See Carey v. Brown, 447 U. S., at 466-467. There is, moreover, no indication—either in the text of the Constitution or in our cases interpreting it—that a separate juridical category exists for the American flag alone. Indeed, we would not be surprised to learn that the persons 418 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. who framed our Constitution and wrote the Amendment that we now construe were not known for their reverence for the Union Jack. The First Amendment does not guarantee that other concepts virtually sacred to our Nation as a whole— such as the principle that discrimination on the basis of race is odious and destructive—will go unquestioned in the marketplace of ideas. See Brandenburg v. Ohio, 395 U. S. 444 (1969). We decline, therefore, to create for the flag an exception to the joust of principles protected by the First Amendment. It is not the State’s ends, but its means, to which we object. It cannot be gainsaid that there is a special place reserved for the flag in this Nation, and thus we do not doubt that the government has a legitimate interest in making efforts to “preserv[e] the national flag as an unalloyed symbol of our country.” Spence, 418 U. S., at 412. We reject the suggestion, urged at oral argument by counsel for Johnson, that the government lacks “any state interest whatsoever” in regulating the manner in which the flag may be displayed. Tr. of Oral Arg. 38. Congress has, for example, enacted precatory regulations describing the proper treatment of the flag, see 36 U. S. C. §§ 173-177, and we cast no doubt on the legitimacy of its interest in making such recommendations. To say that the government has an interest in encouraging proper treatment of the flag, however, is not to say that it may criminally punish a person for burning a flag as a means of political protest. “National unity as an end which officials may foster by persuasion and example is not in question. The problem is whether under our Constitution compulsion as here employed is a permissible means for its achievement.” Barnette, 319 U. S., at 640. We are fortified in today’s conclusion by our conviction that forbidding criminal punishment for conduct such as Johnson s will not endanger the special role played by our flag or the feelings it inspires. To paraphrase Justice Holmes, we submit that nobody can suppose that this one gesture of an un TEXAS V. JOHNSON 419 397 Opinion of the Court known man will change our Nation’s attitude towards its flag. See Abrams v. United States, 250 U. S. 616, 628 (1919) (Holmes, J., dissenting). Indeed, Texas’ argument that the burning of an American flag “ ‘is an act having a high likelihood to cause a breach of the peace,’” Brief for Petitioner 31, quoting Sutherland v. DeWulf, 323 F. Supp. 740, 745 (SD Ill. 1971) (citation omitted), and its statute’s implicit assumption that physical mistreatment of the flag will lead to “serious offense,” tend to confirm that the flag’s special role is not in danger; if it were, no one would riot or take offense because a flag had been burned. We are tempted to say, in fact, that the flag’s deservedly cherished place in our community will be strengthened, not weakened, by our holding today. Our decision is a reaffirmation of the principles of freedom and inclusiveness that the flag best reflects, and of the conviction that our toleration of criticism such as Johnson’s is a sign and source of our strength. Indeed, one of the proudest images of our flag, the one immortalized in our own national anthem, is of the bombardment it survived at Fort McHenry. It is the Nation’s resilience, not its rigidity, that Texas sees reflected in the flag—and it is that resilience that we reassert today. The way to preserve the flag’s special role is not to punish those who feel differently about these matters. It is to persuade them that they are wrong. “To courageous, self-reliant men, with confidence in the power of free and fearless reasoning applied through the processes of popular government, no danger flowing from speech can be deemed clear and present, unless the incidence of the evil apprehended is so imminent that it may befall before there is opportunity for full discussion. If there be time to expose through discussion the falsehood and fallacies, to avert the evil by the processes of education, the remedy to be applied is more speech, not enforced silence.” Whitney v. California, 274 U. S. 357, 377 (1927) (Brandeis, J., concurring). And, precisely because it is our flag that is involved, one’s response to the flag 420 OCTOBER TERM, 1988 Kennedy, J., concurring 491 U. S. burner may exploit the uniquely persuasive power of the flag itself. We can imagine no more appropriate response to burning a flag than waving one’s own, no better way to counter a flag burner’s message than by saluting the flag that burns, no surer means of preserving the dignity even of the flag that burned than by—as one witness here did—according its remains a respectful burial. We do not consecrate the flag by punishing its desecration, for in doing so we dilute the freedom that this cherished emblem represents. V Johnson was convicted for engaging in expressive conduct. The State’s interest in preventing breaches of the peace does not support his conviction because Johnson’s conduct did not threaten to disturb the peace. Nor does the State’s interest in preserving the flag as a symbol of nationhood and national unity justify his criminal conviction for engaging in political expression. The judgment of the Texas Court of Criminal Appeals is therefore Affirmed. Justice Kennedy, concurring. I write not to qualify the words Justice Brennan chooses so well, for he says with power all that is necessary to explain our ruling. I join his opinion without reservation, but with a keen sense that this case, like others before us from time to time, exacts its personal toll. This prompts me to add to our pages these few remarks. The case before us illustrates better than most that the judicial power is often difficult in its exercise. We cannot here ask another Branch to share responsibility, as when the argument is made that a statute is flawed or incomplete. For we are presented with a clear and simple statute to be judged against a pure command of the Constitution. The outcome can be laid at no door but ours. The hard fact is that sometimes we must make decisions we do not like. We make them because they are right, right TEXAS V. JOHNSON 421 397 Rehnquist, C. J., dissenting in the sense that the law and the Constitution, as we see them, compel the result. And so great is our commitment to the process that, except in the rare case, we do not pause to express distaste for the result, perhaps for fear of undermining a valued principle that dictates the decision. This is one of those rare cases. Our colleagues in dissent advance powerful arguments why respondent may be convicted for his expression, reminding us that among those who will be dismayed by our holding will be some who have had the singular honor of carrying the flag in battle. And I agree that the flag holds a lonely place of honor in an age when absolutes are distrusted and simple truths are burdened by unneeded apologetics. With all respect to those views, I do not believe the Constitution gives us the right to rule as the dissenting Members of the Court urge, however painful this judgment is to announce. Though symbols often are what we ourselves make of them, the flag is constant in expressing beliefs Americans share, beliefs in law and peace and that freedom which sustains the human spirit. The case here today forces recognition of the costs to which those beliefs commit us. It is poignant but fundamental that the flag protects those who hold it in contempt. For all the record shows, this respondent was not a philosopher and perhaps did not even possess the ability to comprehend how repellent his statements must be to the Republic itself. But whether or not he could appreciate the enormity of the offense he gave, the fact remains that his acts were speech, in both the technical and the fundamental meaning of the Constitution. So I agree with the Court that he must go free. Chief Justice Rehnquist, with whom Justice White and Justice O’Connor join, dissenting. In holding this Texas statute unconstitutional, the Court ignores Justice Holmes’ familiar aphorism that “a page of history is worth a volume of logic.” New York Trust Co. v. 422 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. Eisner, 256 U. S. 345, 349 (1921). For more than 200 years, the American flag has occupied a unique position as the symbol of our Nation, a uniqueness that justifies a governmental prohibition against flag burning in the way respondent Johnson did here. At the time of the American Revolution, the flag served to unify the Thirteen Colonies at home, while obtaining recognition of national sovereignty abroad. Ralph Waldo Emerson’s “Concord Hymn” describes the first skirmishes of the Revolutionary War in these lines: “By the rude bridge that arched the flood Their flag to April’s breeze unfurled, Here once the embattled farmers stood And fired the shot heard round the world.” During that time, there were many colonial and regimental flags, adorned with such symbols as pine trees, beavers, anchors, and rattlesnakes, bearing slogans such as “Liberty or Death,” “Hope,” “An Appeal to Heaven,” and “Don’t Tread on Me.” The first distinctive flag of the Colonies was the “Grand Union Flag”—with 13 stripes and a British flag in the left corner—which was flown for the first time on January 2, 1776, by troops of the Continental Army around Boston. By June 14, 1777, after we declared our independence from England, the Continental Congress resolved: “That the flag of the thirteen United States be thirteen stripes, alternate red and white: that the union be thirteen stars, white in a blue field, representing a new constellation.” 8 Journal of the Continental Congress 1774-1789, p. 464 (W. Ford ed. 1907). One immediate result of the flag’s adoption was that American vessels harassing British shipping sailed under an authorized national flag. Without such a flag, the British could treat captured seamen as pirates and hang them summarily; with a national flag, such seamen were treated as prisoners of war. TEXAS V. JOHNSON 423 397 Rehnquist, C. J., dissenting During the War of 1812, British naval forces sailed up Chesapeake Bay and marched overland to sack and burn the city of Washington. They then sailed up the Patapsco River to invest the city of Baltimore, but to do so it was first necessary to reduce Fort McHenry in Baltimore Harbor. Francis Scott Key, a Washington lawyer, had been granted permission by the British to board one of their warships to negotiate the release of an American who had been taken prisoner. That night, waiting anxiously on the British ship, Key watched the British fleet firing on Fort McHenry. Finally, at daybreak, he saw the fort’s American flag still flying; the British attack had failed. Intensely moved, he began to scribble on the back of an envelope the poem that became our national anthem: “0 say can you see by the dawn’s early light What so proudly we hail’d at the twilight’s last gleaming, Whose broad stripes & bright stars through the perilous fight O’er the ramparts we watch’d, were so gallantly streaming? And the rocket’s red glare, the bomb bursting in air, Gave proof through the night that our flag was still there, 0 say does that star-spangled banner yet wave O’er the land of the free & the home of the brave?” The American flag played a central role in our Nation’s most tragic conflict, when the North fought against the South. The lowering of the American flag at Fort Sumter was viewed as the start of the war. G. Preble, History of the Flag of the United States of America 453 (1880). The Southern States, to formalize their separation from the Union, adopted the “Stars and Bars” of the Confederacy. The Union troops marched to the sound of “Yes We’ll Rally Round The Flag Boys, We’ll Rally Once Again.” President Abraham Lincoln refused proposals to remove from the 424 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. American flag the stars representing the rebel States, because he considered the conflict not a war between two nations but an attack by 11 States against the National Government. Id., at 411. By war’s end, the American flag again flew over “an indestructible union, composed of indestructible states.” Texas v. White, 7 Wall. 700, 725 (1869). One of the great stories of the Civil War is told in John Greenleaf Whittier’s poem, “Barbara Frietchie”: “Up from the meadows rich with corn, Clear in the cool September morn, The clustered spires of Frederick stand Green-walled by the hills of Maryland. Round about them orchards sweep, Apple- and peach-tree fruited deep, Fair as a garden of the Lord To the eyes of the famished rebel horde, On that pleasant mom of the early fall When Lee marched over the mountain wall, — Over the mountains winding down, Horse and foot, into Frederick town. Forty flags with their silver stars, Forty flags with their crimson bars, Flapped in the morning wind: the sun Of noon looked down, and saw not one. Up rose old Barbara Frietchie then, Bowed with her fourscore years and ten; Bravest of all in Frederick town, She took up the flag the men hauled down; In her attic-window the staff she set, To show that one heart was loyal yet. Up the street came the rebel tread, Stonewall Jackson riding ahead. Under his slouched hat left and right He glanced: the old flag met his sight. ‘Halt!’—the dust-brown ranks stood fast. ‘Fire!’—out blazed the rifle-blast. TEXAS V. JOHNSON 425 397 Rehnquist, C. J., dissenting It shivered the window, pane and sash; It rent the banner with seam and gash. Quick, as it fell, from the broken staff Dame Barbara snatched the silken scarf; She leaned far out on the window-sill, And shook it forth with a royal will. ‘Shoot, if you must, this old gray head, But spare your country’s flag,’ she said. A shade of sadness, a blush of shame, Over the face of the leader came; The nobler nature within him stirred To life at that woman’s deed and word: ‘Who touches a hair of yon gray head Dies like a dog! March on! ’ he said. All day long through Frederick street Sounded the tread of marching feet: All day long that free flag tost Over the heads of the rebel host. Ever its torn folds rose and fell On the loyal winds that loved it well; And through the hill-gaps sunset light Shone over it with a warm good-night. Barbara Frietchie’s work is o’er, And the Rebel rides on his raids no more. Honor to her! and let a tear Fall, for her sake, on Stonewall’s bier. Over Barbara Frietchie’s grave, Flag of Freedom and Union, wave! Peace and order and beauty draw Round thy symbol of light and law; And ever the stars above look down On thy stars below in Frederick town!” In the First and Second World Wars, thousands of our countrymen died on foreign soil fighting for the American cause. At Iwo Jima in the Second World War, United States Marines fought hand to hand against thousands of 426 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. Japanese. By the time the Marines reached the top of Mount Suribachi, they raised a piece of pipe upright and from one end fluttered a flag. That ascent had cost nearly 6,000 American lives. The Iwo Jima Memorial in Arlington National Cemetery memorializes that event. President Franklin Roosevelt authorized the use of the flag on labels, packages, cartons, and containers intended for export as lend-lease aid, in order to inform people in other countries of the United States’ assistance. Presidential Proclamation No. 2605, 58 Stat. 1126. During the Korean war, the successful amphibious landing of American troops at Inchon was marked by the raising of an American flag within an hour of the event. Impetus for the enactment of the Federal Flag Desecration Statute in 1967 came from the impact of flag burnings in the United States on troop morale in Vietnam. Representative L. Mendel Rivers, then Chairman of the House Armed Services Committee, testified that “[t]he burning of the flag . . . has caused my mail to increase 100 percent from the boys in Vietnam, writing me and asking me what is going on in America.” Desecration of the Flag, Hearings on H. R. 271 before Subcommittee No. 4 of the House Committee on the Judiciary, 90th Cong., 1st Sess., 189 (1967). Representative Charles Wiggins stated: “The public act of desecration of our flag tends to undermine the morale of American troops. That this finding is true can be attested by many Members who have received correspondence from servicemen expressing their shock and disgust of such conduct.” 113 Cong. Rec. 16459 (1967). The flag symbolizes the Nation in peace as well as in war. It signifies our national presence on battleships, airplanes, military installations, and public buildings from the United States Capitol to the thousands of county courthouses and city halls throughout the country. Two flags are prominently placed in our courtroom. Countless flags are placed by the graves of loved ones each year on what was first called TEXAS V. JOHNSON 427 397 Rehnquist, C. J., dissenting Decoration Day, and is now called Memorial Day. The flag is traditionally placed on the casket of deceased members of the Armed Forces, and it is later given to the deceased’s family. 10 U. S. C. §§ 1481, 1482. Congress has provided that the flag be flown at half-staff upon the death of the President, Vice President, and other government officials “as a mark of respect to their memory.” 36 U. S. C. § 175(m). The flag identifies United States merchant ships, 22 U. S. C. §454, and “[t]he laws of the Union protect our commerce wherever the flag of the country may float.” United States v. Guthrie, 17 How. 284, 309 (1855). No other American symbol has been as universally honored as the flag. In 1931, Congress declared “The Star-Spangled Banner” to be our national anthem. 36 U. S. C. § 170. In 1949, Congress declared June 14th to be Flag Day. § 157. In 1987, John Philip Sousa’s “The Stars and Stripes Forever” was designated as the national march. Pub. L. 101-186, 101 Stat. 1286. Congress has also established “The Pledge of Allegiance to the Flag” and the manner of its deliverance. 36 U. S. C. § 172. The flag has appeared as the principal symbol on approximately 33 United States postal stamps and in the design of at least 43 more, more times than any other symbol. United States Postal Service, Definitive Mint Set 15 (1988). Both Congress and the States have enacted numerous laws regulating misuse of the American flag. Until 1967, Congress left the regulation of misuse of the flag up to the States. Now, however, 18 U. S. C. § 700(a) provides that: “Whoever knowingly casts contempt upon any flag of the United States by publicly mutilating, defacing, defiling, burning, or trampling upon it shall be fined not more than $1,000 or imprisoned for not more than one year, or both.” Congress has also prescribed, inter alia, detailed rules for the design of the flag, 4 U. S. C. § 1, the time and occasion of flag’s display, 36 U. S. C. § 174, the position and manner of 428 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. its display, § 175, respect for the flag, § 176, and conduct during hoisting, lowering, and passing of the flag, § 177. With the exception of Alaska and Wyoming, all of the States now have statutes prohibiting the burning of the flag.1 Most of the state statutes are patterned after the Uniform Flag Act of 1917, which in §3 provides: “No person shall publicly mutilate, deface, defile, defy, trample upon, or by word or act cast contempt upon any such flag, standard, color, ensign or shield.” Proceedings of National Conference of Commissioners on Uniform State Laws 323-324 (1917). Most were passed by the States at about the time of World War I. Rosenblatt, Flag Desecration Statutes: History and Analysis, 1972 Wash. U. L. Q. 193, 197. 1 See Ala. Code § 13A-11-12 (1982); Ariz. Rev. Stat. Ann. § 13-3703 (1978); Ark. Code Ann. § 5-51-207 (1987); Cal. Mil. & Vet. Code Ann. § 614 (West 1988); Colo. Rev. Stat. §18-11-204 (1986); Conn. Gen. Stat. §53-258a (1985); Del. Code Ann., Tit. 11, §1331 (1987); Fla. Stat. §§256.05-256.051, 876.52 (1987); Ga. Code Ann. §50-3-9 (1986); Haw. Rev. Stat. §711-1107 (1988); Idaho Code §18-3401 (1987); Ill. Rev. Stat., ch. 1, Uli 3307, 3351 (1980); Ind. Code §35-45-1-4 (1986); Iowa Code §32.1 (1978 and Supp. 1989); Kan. Stat. Ann. §21-4114 (1988); Ky. Rev. Stat. Ann. §525.110 (Michie Supp. 1988); La. Rev. Stat. Ann. §14:116 (West 1986); Me. Rev. Stat. Ann., Tit. 1, §254 (1979); Md. Ann. Code, Art. 27, §83 (1988); Mass. Gen. Laws §§264, 265 (1987); Mich. Comp. Laws §750.246 (1968); Minn. Stat. §609.40 (1987); Miss. Code Ann. §97-7-39 (1973); Mo. Rev. Stat. §578.095 (Supp. 1989); Mont. Code Ann. §45-8-215 (1987); Neb. Rev. Stat. §28-928 (1985); Nev. Rev. Stat. §201.290 (1986); N. H. Rev. Stat. Ann. §646.1 (1986); N. J. Stat. Ann. §2C:33-9 (West 1982); N. M. Stat. Ann. § 30-21-4 (1984); N. Y. Gen. Bus. Law § 136 (McKinney 1988); N. C. Gen. Stat. §14-381 (1986); N. D. Cent. Code §12.1-07-02 (1985); Ohio Rev. Code Ann. §2927.11 (1987); Okla. Stat., Tit. 21, §372 (1983); Ore. Rev. Stat. §166.075 (1987); 18 Pa. Cons. Stat. §2102 (1983); R. I. Gen. Laws § 11-15-2 (1981); S. C. Code §§ 16-17-220, 16-17-230 (1985 and Supp. 1988); S. D. Codified Laws §22-9-1 (1988); Tenn. Code Ann. §§39-5-843, 39-5-847 (1982); Tex. Penal Code Ann. §42.09 (1974); Utah Code Ann. §76-9-601 (1978); Vt. Stat. Ann., Tit. 13, § 1903 (1974); Va. Code §18.2-488 (1988); Wash. Rev. Code §9.86.030 (1988); W. Va. Code § 61-1-8 (1989); Wis. Stat. § 946.05 (1985-1986). TEXAS v. JOHNSON 429 397 Rehnquist, C. J., dissenting The American flag, then, throughout more than 200 years of our history, has come to be the visible symbol embodying our Nation. It does not represent the views of any particular political party, and it does not represent any particular political philosophy. The flag is not simply another “idea” or “point of view” competing for recognition in the marketplace of ideas. Millions and millions of Americans regard it with an almost mystical reverence regardless of what sort of social, political, or philosophical beliefs they may have. I cannot agree that the First Amendment invalidates the Act of Congress, and the laws of 48 of the 50 States, which make criminal the public burning of the flag. More than 80 years ago in Halter v. Nebraska, 205 U. S. 34 (1907), this Court upheld the constitutionality of a Nebraska statute that forbade the use of representations of the American flag for advertising purposes upon articles of merchandise. The Court there said: “For that flag every true American has not simply an appreciation but a deep affection. . . . Hence, it has often occurred that insults to a flag have been the cause of war, and indignities put upon it, in the presence of those who revere it, have often been resented and sometimes punished on the spot.” Id., at 41. Only two Terms ago, in San Francisco Arts & Athletics, Inc. v. United States Olympic Committee, 483 U. S. 522 (1987), the Court held that Congress could grant exclusive use of the word “Olympic” to the United States Olympic Committee. The Court thought that this “restrictio[n] on expressive speech properly [was] characterized as incidental to the primary congressional purpose of encouraging and rewarding the USOC’s activities.” Id., at 536. As the Court stated, “when a word [or symbol] acquires value ‘as the result of organization and the expenditure of labor, skill, and money’ by an entity, that entity constitutionally may obtain a limited property right in the word [or symbol].” Id., at 532, quoting International News Service v. Associated Press, 248 430 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. U. S. 215, 239 (1918). Surely Congress or the States may recognize a similar interest in the flag. But the Court insists that the Texas statute prohibiting the public burning of the American flag infringes on respondent Johnson’s freedom of expression. Such freedom, of course, is not absolute. See Schenck v. United States, 249 U. S. 47 (1919). In Chaplinsky v. New Hampshire, 315 U. S. 568 (1942), a unanimous Court said: “Allowing the broadest scope to the language and purpose of the Fourteenth Amendment, it is well understood that the right of free speech is not absolute at all times and under all circumstances. There are certain well-defined and narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem. These include the lewd and obscene, the profane, the libelous, and the insulting or ‘fighting’ words—those which by their very utterance inflict injury or tend to incite an immediate breach of the peace. It has been well observed that such utterances are no essential part of any exposition of ideas, and are of such slight social value as a step to truth that any benefit that may be derived from them is clearly outweighed by the social interest in order and morality.” Id., at 571-572 (footnotes omitted). The Court upheld Chaplinsky’s conviction under a state statute that made it unlawful to “address any offensive, derisive or annoying word to any person who is lawfully in any street or other public place.” Id., at 569. Chaplinsky had told a local marshal, “‘“You are a God damned racketeer” and a “damned Fascist and the whole government of Rochester are Fascists or agents of Fascists.”’” Ibid. Here it may equally well be said that the public burning of the American flag by Johnson was no essential part of any exposition of ideas, and at the same time it had a tendency to incite a breach of the peace. Johnson was free to make any verbal denunciation of the flag that he wished; indeed, he was TEXAS v. JOHNSON 431 397 Rehnquist, C. J., dissenting free to burn the flag in private. He could publicly burn other symbols of the Government or effigies of political leaders. He did lead a march through the streets of Dallas, and conducted a rally in front of the Dallas City Hall. He engaged in a “die-in” to protest nuclear weapons. He shouted out various slogans during the march, including: “Reagan, Mondale which will it be? Either one means World War III”; “Ronald Reagan, killer of the hour, Perfect example of U. S. power”; and “red, white and blue, we spit on you, you stand for plunder, you will go under.” Brief for Respondent 3. For none of these acts was he arrested or prosecuted; it was only when he proceeded to burn publicly an American flag stolen from its rightful owner that he violated the Texas statute. The Court could not, and did not, say that Chaplinsky’s utterances were not expressive phrases — they clearly and succinctly conveyed an extremely low opinion of the addressee. The same may be said of Johnson’s public burning of the flag in this case; it obviously did convey Johnson’s bitter dislike of his country. But his act, like Chaplinsky’s provocative words, conveyed nothing that could not have been conveyed and was not conveyed just as forcefully in a dozen different ways. As with “fighting words,” so with flag burning, for purposes of the First Amendment: It is “no essential part of any exposition of ideas, and [is] of such slight social value as a step to truth that any benefit that may be derived from [it] is clearly outweighed” by the public interest in avoiding a probable breach of the peace. The highest courts of several States have upheld state statutes prohibiting the public burning of the flag on the grounds that it is so inherently inflammatory that it may cause a breach of public order. See, e. g., State v. Royal, 113 N. H. 224, 229, 305 A. 2d 676, 680 (1973); State v. Waterman, 190 N. W. 2d 809, 811-812 (Iowa 1971); see also State v. Mitchell, 32 Ohio App. 2d 16, 30, 288 N. E. 2d 216, 226 (1972). 432 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. The result of the Texas statute is obviously to deny one in Johnson’s frame of mind one of many means of “symbolic speech.” Far from being a case of “one picture being worth a thousand words,” flag burning is the equivalent of an inarticulate grunt or roar that, it seems fair to say, is most likely to be indulged in not to express any particular idea, but to antagonize others. Only five years ago we said in City Council of Los Angeles v. Taxpayers for Vincent, 466 U. S. 789, 812 (1984), that “the First Amendment does not guarantee the right to employ every conceivable method of communication at all times and in all places.” The Texas statute deprived Johnson of only one rather inarticulate symbolic form of protest—a form of protest that was profoundly offensive to many—and left him with a full panoply of other symbols and every conceivable form of verbal expression to express his deep disapproval of national policy. Thus, in no way can it be said that Texas is punishing him because his hearers—or any other group of people—were profoundly opposed to the message that he sought to convey. Such opposition is no proper basis for restricting speech or expression under the First Amendment. It was Johnson’s use of this particular symbol, and not the idea that he sought to convey by it or by his many other expressions, for which he was punished. Our prior cases dealing with flag desecration statutes have left open the question that the Court resolves today. In Street v. New York, 394 U. S. 576, 579 (1969), the defendant burned a flag in the street, shouting “We don’t need no damned flag” and “[i]f they let that happen to Meredith we don’t need an American flag.” The Court ruled that since the defendant might have been convicted solely on the basis of his words, the conviction could not stand, but it expressly reserved the question whether a defendant could constitutionally be convicted for burning the flag. Id., at 581. Chief Justice Warren, in dissent, stated: “I believe that the States and Federal Government do have the power to protect the flag from acts of desecration and disgrace. . . . [I]t is dif TEXAS V. JOHNSON 433 397 Rehnquist, C. J., dissenting ficult for me to imagine that, had the Court faced this issue, it would have concluded otherwise.” Id., at 605. Justices Black and Fortas also expressed their personal view that a prohibition on flag burning did not violate the Constitution. See id., at 610 (Black, J., dissenting) (“It passes my belief that anything in the Federal Constitution bars a State from making the deliberate burning of the American Flag an offense”); id., at 615-617 (Fortas, J., dissenting) (“[T]he States and the Federal Government have the power to protect the flag from acts of desecration committed in public. . . . [T]he flag is a special kind of personality. Its use is traditionally and universally subject to special rules and regulation. ... A person may ‘own’ a flag, but ownership is subject to special burdens and responsibilities. A flag may be property, in a sense; but it is property burdened with peculiar obligations and restrictions. Certainly . . . these special conditions are not per se arbitrary or beyond governmental power under our Constitution”). In Spence v. Washington, 418 U. S. 405 (1974), the Court reversed the conviction of a college student who displayed the flag with a peace symbol affixed to it by means of removable black tape from the window of his apartment. Unlike the instant case, there was no risk of a breach of the peace, no one other than the arresting officers saw the flag, and the defendant owned the flag in question. The Court concluded that the student’s conduct was protected under the First Amendment, because “no interest the State may have in preserving the physical integrity of a privately owned flag was significantly impaired on these facts.” Id., at 415. The Court was careful to note, however, that the defendant “was not charged under the desecration statute, nor did he permanently disfigure the flag or destroy it.” Ibid. In another related case, Smith v. Goguen, 415 U. S. 566 (1974), the appellee, who wore a small flag on the seat of his trousers, was convicted under a Massachusetts flag-misuse statute that subjected to criminal liability anyone who 434 OCTOBER TERM, 1988 Rehnquist, C. J., dissenting 491 U. S. “publicly . . . treats contemptuously the flag of the United States.” Id., at 568-569. The Court affirmed the lower court’s reversal of appellee’s conviction, because the phrase “treats contemptuously” was unconstitutionally broad and vague. Id., at 576. The Court was again careful to point out that “[c]ertainly nothing prevents a legislature from defining with substantial specificity what constitutes forbidden treatment of United States flags.” Id., at 581-582. See also id., at 587 (White, J., concurring in judgment) (“The flag is a national property, and the Nation may regulate those who would make, imitate, sell, possess, or use it. I would not question those statutes which proscribe mutilation, defacement, or burning of the flag or which otherwise protect its physical integrity, without regard to whether such conduct might provoke violence. . . . There would seem to be little question about the power of Congress to forbid the mutilation of the Lincoln Memorial. . . . The flag is itself a monument, subject to similar protection”); id., at 591 (Black-mun, J., dissenting) (“Goguen’s punishment was constitutionally permissible for harming the physical integrity of the flag by wearing it affixed to the seat of his pants”). But the Court today will have none of this. The uniquely deep awe and respect for our flag felt by virtually all of us are bundled off under the rubric of “designated symbols,” ante, at 417, that the First Amendment prohibits the government from “establishing.” But the government has not “established” this feeling; 200 years of history have done that. The government is simply recognizing as a fact the profound regard for the American flag created by that history when it enacts statutes prohibiting the disrespectful public burning of the flag. The Court concludes its opinion with a regrettably patronizing civics lecture, presumably addressed to the Members of both Houses of Congress, the members of the 48 state legislatures that enacted prohibitions against flag burning, and the troops fighting under that flag in Vietnam who objected to its TEXAS v. JOHNSON 435 397 Rehnquist, C. J., dissenting being burned: “The way to preserve the flag’s special role is not to punish those who feel differently about these matters. It is to persuade them that they are wrong.” Ante, at 419. The Court’s role as the final expositor of the Constitution is well established, but its role as a Platonic guardian admonishing those responsible to public opinion as if they were truant schoolchildren has no similar place in our system of government. The cry of “no taxation without representation” animated those who revolted against the English Crown to found our Nation—the idea that those who submitted to government should have some say as to what kind of laws would be passed. Surely one of the high purposes of a democratic society is to legislate against conduct that is regarded as evil and profoundly offensive to the majority of people—whether it be murder, embezzlement, pollution, or flag burning. Our Constitution wisely places limits on powers of legislative majorities to act, but the declaration of such limits by this Court “is, at all times, a question of much delicacy, which ought seldom, if ever, to be decided in the affirmative, in a doubtful case.” Fletcher v. Peck, 6 Cranch 87, 128 (1810) (Marshall, C. J.). Uncritical extension of constitutional protection to the burning of the flag risks the frustration of the very purpose for which organized governments are instituted. The Court decides that the American flag is just another symbol, about which not only must opinions pro and con be tolerated, but for which the most minimal public respect may not be enjoined. The government may conscript men into the Armed Forces where they must fight and perhaps die for the flag, but the government may not prohibit the public burning of the banner under which they fight. I would uphold the Texas statute as applied in this case.2 2 In holding that the Texas statute as applied to Johnson violates the First Amendment, the Court does not consider Johnson’s claims that the statute is unconstitutionally vague or overbroad. Brief for Respondent 24-30. I think those claims are without merit. In New York State Club Assn. v. City of New York, 487 U. S. 1, 11 (1988), we stated that a facial 436 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. Justice Stevens, dissenting. As the Court analyzes this case, it presents the question whether the State of Texas, or indeed the Federal Government, has the power to prohibit the public desecration of the American flag. The question is unique. In my judgment rules that apply to a host of other symbols, such as state flags, armbands, or various privately promoted emblems of political or commercial identity, are not necessarily controlling. Even if flag burning could be considered just another species of symbolic speech under the logical application of the rules that the Court has developed in its interpretation of the First Amendment in other contexts, this case has an intangible dimension that makes those rules inapplicable. A country’s flag is a symbol of more than “nationhood and national unity.” Ante, at 407, 410, 413, and n. 9, 417, 420. It also signifies the ideas that characterize the society that has chosen that emblem as well as the special history that has animated the growth and power of those ideas. The fleurs-de-lis and the tricolor both symbolized “nationhood and national unity,” but they had vastly different meanings. The message conveyed by some flags—the swastika, for example-may survive long after it has outlived its usefulness as a symbol of regimented unity in a particular nation. challenge is only proper under the First Amendment when a statute can never be applied in a permissible manner or when, even if it may be validly applied to a particular defendant, it is so broad as to reach the protected speech of third parties. While Tex. Penal Code Ann. §42.09 (1989) “may not satisfy those intent on finding fault at any cost, [it is] set out in terms that the ordinary person exercising ordinary common sense can sufficiently understand and comply with.” CSC v. Letter Carriers, 413 U. S. 548, 579 (1973). By defining “desecrate” as “deface,” “damage” or otherwise “physically mistreat” in a manner that the actor knows will “seriously offend’ others, §42.09 only prohibits flagrant acts of physical abuse and destruction of the flag of the sort at issue here—soaking a flag with lighter fluid and igniting it in public—and not any of the examples of improper flag etiquette cited in respondent’s brief. TEXAS V. JOHNSON 437 397 Stevens, J., dissenting So it is with the American flag. It is more than a proud symbol of the courage, the determination, and the gifts of nature that transformed 13 fledgling Colonies into a world power. It is a symbol of freedom, of equal opportunity, of religious tolerance, and of good will for other peoples who share our aspirations. The symbol carries its message to dissidents both at home and abroad who may have no interest at all in our national unity or survival. The value of the flag as a symbol cannot be measured. Even so, I have no doubt that the interest in preserving that value for the future is both significant and legitimate. Conceivably that value will be enhanced by the Court’s conclusion that our national commitment to free expression is so strong that even the United States as ultimate guarantor of that freedom is without power to prohibit the desecration of its unique symbol. But I am unpersuaded. The creation of a federal right to post bulletin boards and graffiti on the Washington Monument might enlarge the market for free expression, but at a cost I would not pay. Similarly, in my considered judgment, sanctioning the public desecration of the flag will tarnish its value—both for those who cherish the ideas for which it waves and for those who desire to don the robes of martyrdom by burning it. That tarnish is not justified by the trivial burden on free expression occasioned by requiring that an available, alternative mode of expression— including uttering words critical of the flag, see Street v. New York, 394 U. S. 576 (1969)-be employed. It is appropriate to emphasize certain propositions that are not implicated by this case. The statutory prohibition of flag desecration does not “prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion or force citizens to confess by word or act their faith therein.” West Virginia Board of Education v. Barnette, 319 U. S. 624, 642 (1943). The statute does not compel any conduct or any profession of respect for any idea or any symbol. 438 OCTOBER TERM, 1988 Stevens, J., dissenting 491 U. S. Nor does the statute violate “the government’s paramount obligation of neutrality in its regulation of protected communication.” Young v. American Mini Theatres, Inc., 427 U. S. 50, 70 (1976) (plurality opinion). The content of respondent’s message has no relevance whatsoever to the case. The concept of “desecration” does not turn on the substance of the message the actor intends to convey, but rather on whether those who view the act will take serious offense. Accordingly, one intending to convey a message of respect for the flag by burning it in a public square might nonetheless be guilty of desecration if he knows that others—perhaps simply because they misperceive the intended message—will be seriously offended. Indeed, even if the actor knows that all possible witnesses will understand that he intends to send a message of respect, he might still be guilty of desecration if he also knows that this understanding does not lessen the offense taken by some of those witnesses. Thus, this is not a case in which the fact that “it is the speaker’s opinion that gives offense” provides a special “reason for according it constitutional protection,” FCC v. Pacifica Foundation, 438 U. S. 726, 745 (1978) (plurality opinion). The case has nothing to do with “disagreeable ideas,” see ante, at 409. It involves disagreeable conduct that, in my opinion, diminishes the value of an important national asset. The Court is therefore quite wrong in blandly asserting that respondent “was prosecuted for his expression of dissatisfaction with the policies of this country, expression situated at the core of our First Amendment values.” Ante, at 411. Respondent was prosecuted because of the method he chose to express his dissatisfaction with those policies. Had he chosen to spray-paint—or perhaps convey with a motion picture projector—his message of dissatisfaction on the facade of the Lincoln Memorial, there would be no question about the power of the Government to prohibit his means of expression. The prohibition would be supported by the legitimate interest in preserving the quality of an important TEXAS V. JOHNSON 439 397 Stevens, J., dissenting national asset. Though the asset at stake in this case is intangible, given its unique value, the same interest supports a prohibition on the desecration of the American flag. * The ideas of liberty and equality have been an irresistible force in motivating leaders like Patrick Henry, Susan B. Anthony, and Abraham Lincoln, schoolteachers like Nathan Hale and Booker T. Washington, the Philippine Scouts who fought at Bataan, and the soldiers who scaled the bluff at Omaha Beach. If those ideas are worth fighting for—and our history demonstrates that they are—it cannot be true that the flag that uniquely symbolizes their power is not itself worthy of protection from unnecessary desecration. I respectfully dissent. *The Court suggests that a prohibition against flag desecration is not content neutral because this form of symbolic speech is only used by persons who are critical of the flag or the ideas it represents. In making this suggestion the Court does not pause to consider the far-reaching consequences of its introduction of disparate-impact analysis into our First Amendment jurisprudence. It seems obvious that a prohibition against the desecration of a gravesite is content neutral even if it denies some protesters the right to make a symbolic statement by extinguishing the flame in Arlington Cemetery where John F. Kennedy is buried while permitting others to salute the flame by bowing their heads. Few would doubt that a protester who extinguishes the flame has desecrated the gravesite, regardless of whether he prefaces that act with a speech explaining that his purpose is to express deep admiration or unmitigated scorn for the late President. Likewise, few would claim that the protester who bows his head has desecrated the gravesite, even if he makes clear that his purpose is to show disrespect. In such a case, as in a flag burning case, the prohibition against desecration has absolutely nothing to do with the content of the message that the symbolic speech is intended to convey. 440 OCTOBER TERM, 1988 Syllabus 491 U. S. PUBLIC CITIZEN v. UNITED STATES DEPARTMENT OF JUSTICE ET AL. APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA No. 88-429. Argued April 17, 1989—Decided June 21, 1989* To aid the President in fulfilling his constitutional duty to appoint federal judges, the Department of Justice regularly seeks advice from the Standing Committee on Federal Judiciary of the American Bar Association (ABA) regarding potential nominees for judgeships. The ABA Committee’s investigations, reports, and votes on potential nominees are kept confidential, although its rating of a particular candidate is made public if he or she is in fact nominated. Appellant Washington Legal Foundation (WLF) filed suit against the Justice Department after the ABA Committee refused WLF’s request for the names of potential nominees it was considering and for its reports and minutes of its meetings. The action was brought under the Federal Advisory Committee Act (FACA), which, among other things, defines an “advisory committee” as any group “established or utilized” by the President or an agency to give advice on public questions, and requires a covered group to file a charter, afford notice of its meetings, open those meetings to the public, and make its minutes, records, and reports available to the public. Joined by appellant Public Citizen, WLF asked the District Court to declare the Committee an “advisory group” subject to FACA’s requirements and to enjoin the Department from utilizing the ABA Committee until it complied with those requirements. The court dismissed the complaint, holding that the Department’s use of the ABA Committee is subject to FACA’s strictures, but ruling that applying FACA to the ABA Committee would unconstitutionally infringe on the President’s Article II power to nominate federal judges and violate the doctrine of separation of powers. Held: 1. Appellants have standing to bring this suit. The refusal to permit them to scrutinize the ABA Committee’s activities to the extent FACA allows constitutes a sufficiently distinct injury to provide standing, and the fact that other groups or citizens might make the same complaint as appellants does not lessen that injury. Moreover, although the stat- *Together with No. 88-494, Washington Legal Foundation v. United States Department of Justice et al., also on appeal from the same court. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 441 440 Syllabus ute’s disclosure exemptions might bar public access to many of the meetings appellants seek to attend and many of the documents they wish to view, the exemptions probably would not deny access to all meetings and documents, particularly discussions and documents regarding the ABA Committee’s overall functioning, and would not excuse the ABA Committee’s noncompliance with FACA’s other provisions, such as those requiring a covered organization to file a charter and give notice of its meetings. Thus, appellants may gain significant and genuine relief if they prevail in their suit, and such potential gains are sufficient to give them standing. Pp. 448-451. 2. FAC A does not apply to the Justice Department’s solicitation of the ABA Committee’s views on prospective judicial nominees. Pp. 451-467. (a) Whether the ABA Committee is an “advisory committee” under FAC A depends upon whether it is “utilized” by the President or the Department within the statute’s meaning. Read unqualifiedly, that verb would extend FACA’s coverage to the ABA Committee. However, since FAC A was enacted to cure specific ills—particularly the wasteful expenditure of public funds for worthless committee meetings and biased proposals by special interest groups —it is unlikely that Congress intended the statute to cover every formal and informal consultation between the President or an Executive agency and a group rendering advice. When the literal reading of a statutory term compels an odd result, this Court searches beyond the bare text for other evidence of congressional intent. Pp. 451-455. (b) Although the question is a close one, a careful review of the regulatory scheme prior to FACA’s enactment and that statute’s legislative history strongly suggests that Congress did not intend that the term “utilized” apply to the Justice Department’s use of the ABA Committee. FACA’s regulatory predecessor, Executive Order No. 11007, applied to advisory committees formed by a governmental unit and to those not so formed when “being utilized by [the Government] in the same manner as a Government-formed . . . committee.” That the ABA Committee was never deemed to be “utilized” in the relevant sense is evidenced by the fact that no President operating under the Order or any Justice Department official ever applied the Order to the ABA Committee, despite its highly visible role in advising the Department as to potential nominees. That is not surprising, since the ABA Committee—which was formed privately, rather than at the Government’s prompting, to assist the President in performing a constitutionally specified function, and which receives no federal funds and is not amenable to the strict management by agency officials envisaged by the Order—cannot easily be said to have been “utilized” in the same manner as a Government-formed committee. Moreover, FACA adopted many of the Order’s provisions, and there is 442 OCTOBER TERM, 1988 Syllabus 491 U. S. considerable evidence in the statute’s legislative history that Congress sought only to achieve compliance with FACA’s more stringent requirements by advisory committees already covered by the Order and by Presidential advisory committees, and that the statute’s “or utilized” phrase was intended to clarify that FACA applies to committees “established . . . by” the Government in a generous sense of that term, encompassing groups formed indirectly by quasi-public organizations “for” public agencies as well as “by” such agencies themselves. Read in this way, the word “utilized” does not describe the Justice Department’s use of the ABA Committee. Pp. 455-465. (c) Construing FACA to apply to the Justice Department’s consultations with the ABA Committee would present formidable constitutional difficulties. Where, as here, a plausible alternative construction exists that will allow the Court to avoid such problems, the Court will adopt that construction. See, e. g., Crowell v. Benson, 285 U. S. 22, 62. Pp. 465-467. 691 F. Supp. 483, affirmed. Brennan, J., delivered the opinion of the Court, in which White, Marshall, Blackmun, and Stevens, JJ., joined. Kennedy, J., filed an opinion concurring in the judgment, in which Rehnquist, C. J., and O’Connor, J., joined, post, p. 467. Scalia, J., took no part in the consideration or decision of the cases. Eric R. Glitzenstein argued the cause for appellant in No. 88-429. With him on the briefs were Patti A. Goldman and Alan B. Morrison. Paul D. Kamenar argued the cause for appellant in No. 88-494. With him on the briefs was Daniel J. Popeo. Deputy Solicitor General Shapiro argued the cause for appellees in both cases. With him on the brief were Acting Solicitor General Wallace, Assistant Attorney General Bolton, Paul J. Larkin, Jr., and Douglas Letter. Rex E. Lee, Ronald S. Flagg, Carter G. Phillips, Mark D. Hopson, H. Blair White, David T. Pritikin, and Darryl L. DePriest filed a brief for appellee American Bar Association.! tBriefs of amici curiae urging affirmance were filed for the American Federation of Labor and Congress of Industrial Organizations by Robert M. Weinberg, Walter A. Kamiat, and Laurence Gold; and for the People for the American Way Action Fund et al. by Timothy B. Dyk, Thomas F. Connell, and William L. Taylor. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 443 440 Opinion of the Court Justice Brennan delivered the opinion of the Court. The Department of Justice regularly seeks advice from the American Bar Association’s Standing Committee on Federal Judiciary regarding potential nominees for federal judgeships. The question before us is whether the Federal Advisory Committee Act (FACA), 86 Stat. 770, as amended, 5 U. S. C. App. § 1 et seq. (1982 ed. and Supp. V), applies to these consultations and, if it does, whether its application interferes unconstitutionally with the President’s prerogative under Article II to nominate and appoint officers of the United States; violates the doctrine of separation of powers; or unduly infringes the First Amendment right of members of the American Bar Association to freedom of association and expression. We hold that FACA does not apply to this special advisory relationship. We therefore do not reach the constitutional questions presented. I A The Constitution provides that the President “shall nominate, and by and with the Advice and Consent of the Senate, shall appoint” Supreme Court Justices and, as established by Congress, other federal judges. Art. II, §2, cl. 2. Since 1952 the President, through the Department of Justice, has requested advice from the American Bar Association’s Standing Committee on Federal Judiciary (ABA Committee) in making such nominations. The American Bar Association is a private voluntary professional association of approximately 343,000 attorneys. It has several working committees, among them the advisory body whose work is at issue here. The ABA Committee consists of 14 persons belonging to, and chosen by, the American Bar Association. Each of the 12 federal judicial Circuits (not including the Federal Circuit) has one representative on the ABA Committee, except for the Ninth Circuit, which has 444 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. two; in addition, one member is chosen at large. The ABA Committee receives no federal funds. It does not recommend persons for appointment to the federal bench of its own initiative. Prior to announcing the names of nominees for judgeships on the courts of appeals, the district courts, or the Court of International Trade, the President, acting through the Department of Justice, routinely requests a potential nominee to complete a questionnaire drawn up by the ABA Committee and to submit it to the Assistant Attorney General for the Office of Legal Policy, to the chair of the ABA Committee, and to the committee member (usually the representative of the relevant judicial Circuit) charged with investigating the nominee. See American Bar Association Standing Committee on Federal Judiciary, What It Is and How It Works (1983), reprinted in App. 43-49; Brief for Federal Appellee 2.1 The potential nominee’s answers and the referral of his or her name to the ABA Committee are kept confidential. The committee member conducting the investigation then reviews the legal writings of the potential nominee, interviews judges, legal scholars, and other attorneys regarding the potential nominee’s qualifications, and discusses the matter confidentially with representatives of various professional organizations and other groups. The committee member also interviews the potential nominee, sometimes with other committee members in attendance. Following the initial investigation, the committee representative prepares for the chair an informal written report describing the potential nominee’s background, summarizing all interviews, assessing the candidate’s qualifications, and recommending one of four possible ratings: “exceptionally well qualified,” “well qualified,” “qualified,” or “not quali- ’The Justice Department does not ordinarily furnish the names of potential Supreme Court nominees to the ABA Committee for evaluation prior to their nomination, although in some instances the President has done so. See Brief for Federal Appellee 4-5. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 445 440 Opinion of the Court fied.”2 The chair then makes a confidential informal report to the Attorney General’s Office. The chair’s report discloses the substance of the committee representative’s report to the chair, without revealing the identity of persons who were interviewed, and indicates the evaluation the potential nominee is likely to receive if the Department of Justice requests a formal report. If the Justice Department does request a formal report, the committee representative prepares a draft and sends copies to other members of the ABA Committee, together with relevant materials. A vote is then taken and a final report approved. The ABA Committee conveys its rating— though not its final report—in confidence to the Department of Justice, accompanied by a statement whether its rating was supported by all committee members, or whether it only commanded a majority or substantial majority of the ABA Committee. After considering the rating and other information the President and his advisers have assembled, including a report by the Federal Bureau of Investigation and additional interviews conducted by the President’s judicial selection committee, the President then decides whether to nominate the candidate. If the candidate is in fact nominated, the ABA Committee’s rating, but not its report, is made public at the request of the Senate Judiciary Committee.3 B FAC A was born of a desire to assess the need for the “numerous committees, boards, commissions, councils, and simi 2 The ratings now used in connection with Supreme Court nominees are “well qualified,” “not opposed,” and “not qualified.” See American Bar Association Standing Committee on Federal Judiciary, What It Is and How It Works (1983), reprinted in App. 50. The Senate regularly requests the ABA Committee to rate Supreme Court nominees if the Justice Department has not already sought the ABA Committee’s opinion. As with nominees for other federal judgeships, the ABA Committee’s rating is made public at confirmation hearings before the Senate Judiciary Committee. 446 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. lar groups which have been established to advise officers and agencies in the executive branch of the Federal Government.” §2(a), as set forth in 5 U. S. C. App. §2(a).4 Its purpose was to ensure that new advisory committees be established only when essential and that their number be minimized; that they be terminated when they have outlived their usefulness; that their creation, operation, and duration be subject to uniform standards and procedures; that Congress and the public remain apprised of their existence, activities, and cost; and that their work be exclusively advisory in nature. § 2(b). To attain these objectives, FAC A directs the Director of the Office of Management and Budget and agency heads to establish various administrative guidelines and management controls for advisory committees. It also imposes a number of requirements on advisory groups. For example, FACA requires that each advisory committee file a charter, §9(c), and keep detailed minutes of its meetings. § 10(c). Those meetings must be chaired or attended by an officer or employee of the Federal Government who is authorized to adjourn any meeting when he or she deems its adjournment in the public interest. § 10(e). FACA also requires advisory committees to provide advance notice of their meetings and to open them to the public, § 10(a), unless the President or the agency head to which an advisory committee reports determines that it may be closed to the public in accordance with the Government in the Sunshine Act, 5 U. S. C. § 552b(c). § 10(d). In addition, FACA stipulates that advisory committee minutes, records, and reports be made avail- 4 Federal advisory committees are legion. During fiscal year 1988, 58 federal departments sponsored 1,020 advisory committees. General Services Administration, Seventeenth Annual Report of the President on Federal Advisory Committees 1 (1988). Over 3,500 meetings were held, and close to 1,000 reports were issued. Ibid. Costs for fiscal year 1988 totaled over $92 million, roughly half of which was spent on federal staff support. Id., at 3. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 447 440 Opinion of the Court able to the public, provided they do not fall within one of the Freedom of Information Act’s exemptions, see 5 U. S. C. § 552, and the Government does not choose to withhold them. § 10(b). Advisory committees established by legislation or created by the President or other federal officials must also be “fairly balanced in terms of the points of view represented and the functions” they perform. §§ 5(b)(2), (c). Their existence is limited to two years, unless specifically exempted by the entity establishing them. § 14(a)(1). C In October 1986, appellant Washington Legal Foundation (WLF) brought suit against the Department of Justice after the ABA Committee refused WLF’s request for the names of potential judicial nominees it was considering and for the ABA Committee’s reports and minutes of its meetings.5 WLF asked the District Court for the District of Columbia to declare the ABA Committee an “advisory committee” as FAC A defines that term. WLF further sought an injunction ordering the Justice Department to cease utilizing the ABA Committee as an advisory committee until it complied with FACA. In particular, WLF contended that the ABA Committee must file a charter, afford notice of its meetings, open those meetings to the public, and make its minutes, records, and reports available for public inspection and copying. See WLF Complaint, App. 5-11. The Justice Department moved to dismiss, arguing that the ABA Committee did not fall within FACA’s definition of “advisory committee” WLF originally sued the ABA Committee, its members, and the American Bar Association, but not the Department of Justice. The District Court dismissed that complaint on the ground that the Justice Department was the proper defendant. Washington Legal Foundation v. American Bar Assn. Standing Comm, on Federal Judiciary, 648 F. Supp. 1353 (DC 1986). WLF’s appeal on the issue whether a committee can be sued directly for noncompliance with FACA is pending before the Court of Appeals. See Brief for Appellant in No. 88-494, p. 10, n. 9. 448 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. and that, if it did, FACA would violate the constitutional doctrine of separation of powers. Appellant Public Citizen then moved successfully to intervene as a party plaintiff. Like WLF, Public Citizen requested a declaration that the Justice Department’s utilization of the ABA Committee is covered by FACA and an order enjoining the Justice Department to comply with FACA’s requirements. The District Court dismissed the action following oral argument. 691 F. Supp. 483 (1988). The court held that the Justice Department’s use of the ABA Committee is subject to FACA’s strictures, but that “FACA cannot constitutionally be applied to the ABA Committee because to do so would violate the express separation of nomination and consent powers set forth in Article II of the Constitution and because no overriding congressional interest in applying FACA to the ABA Committee has been demonstrated. ” Id., at 486. Congress’ role in choosing judges “is limited to the Senate’s advice and consent function,” the court concluded; “the purposes of FACA are served through the public confirmation process and any need for applying FACA to the ABA Committee is outweighed by the President’s interest in preserving confidentiality and freedom of consultation in selecting judicial nominees.” Id., at 496. We noted probable jurisdiction, 488 U. S. 979 (1988), and now affirm on statutory grounds, making consideration of the relevant constitutional issues unnecessary. II As a preliminary matter, appellee American Bar Association contests appellants’ standing to bring this suit.6 Appellee’s challenge is twofold. First, it contends that neither appellant has alleged injury sufficiently concrete and specific to confer standing; rather, appellee maintains, they have fiThe American Bar Association was not a party below, but intervened for purposes of this appeal after the District Court rendered judgment. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 449 440 Opinion of the Court advanced a general grievance shared in substantially equal measure by all or a large class of citizens, and thus lack standing under our precedents. Brief for Appellee ABA 12-15. Second, appellee argues that even if appellants have asserted a sufficiently discrete injury, they have not demonstrated that a decision in their favor would likely redress the alleged harm, because the meetings they seek to attend and the minutes and records they wish to review would probably be closed to them under FAC A. Hence, the American Bar Association submits, Article III bars their suit. Id., at 15-17. We reject these arguments. Appellee does not, and cannot, dispute that appellants are attempting to compel the Justice Department and the ABA Committee to comply with FACA’s charter and notice requirements, and that they seek access to the ABA Committee’s meetings and records in order to monitor its workings and participate more effectively in the judicial selection process. Appellant WLF has specifically requested, and been refused, the names of candidates under consideration by the ABA Committee, reports and minutes of the Committee’s meetings, and advance notice of future meetings. WLF Complaint, App. 8. As when an agency denies requests for information under the Freedom of Information Act, refusal to permit appellants to scrutinize the ABA Committee’s activities to the extent FACA allows constitutes a sufficiently distinct injury to provide standing to sue. Our decisions interpreting the Freedom of Information Act have never suggested that those requesting information under it need show more than that they sought and were denied specific agency records. See, e. g., Department of Justice v. Reporters Comm, for Freedom of Press, 489 U. S. 749 (1989); Department of Justice v. Julian, 486 U. S. 1 (1988); United States v. Weber Aircraft Corp., 465 U. S. 792 (1984); FBI v. Abramson, 456 U. S. 615 (1982); Department of Air Force v. Rose, 425 U. S. 352 (1976). There is no reason for a different rule here. The fact that other citizens 450 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. or groups of citizens might make the same complaint after unsuccessfully demanding disclosure under FACA does not lessen appellants’ asserted injury, any more than the fact that numerous citizens might request the same information under the Freedom of Information Act entails that those who have been denied access do not possess a sufficient basis to sue. We likewise find untenable the American Bar Association’s claim that appellants lack standing because a ruling in their favor would not provide genuine relief as a result of FACA’s exceptions to disclosure. Appellants acknowledge that many meetings of the ABA Committee might legitimately be closed to the public under FACA and that many documents might properly be shielded from public view. But they by no means concede that FACA licenses denying them access to all meetings and papers, or that it excuses noncompliance with FACA’s other provisions. As Public Citizen contends, if FACA applies to the Justice Department’s use of the ABA Committee without violating the Constitution, the ABA Committee will at least have to file a charter and give notice of its meetings. In addition, discussions and documents regarding the overall functioning of the ABA Committee, including its investigative, evaluative, and voting procedures, could well fall outside FACA’s exemptions. See Reply Brief for Appellant in No. 88-429, pp. 5-6, and n. 3. Indeed, it is difficult to square appellee’s assertion that appellants cannot hope to gain noteworthy relief with its contention that “even more significant interference [than participation of Government officials in the ABA Committee’s affairs] would result from the potential application of the ‘public inspection’ provisions of Section 10 of the Act.” Brief for Appellee ABA 36. The American Bar Association explains: “Disclosure and public access are the rule under FACA; the exemptions generally are construed narrowly. In fact, the Govemment-in-the-Sunshine Act has no deliberative process privilege under which ABA Committee meet- PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 451 440 Opinion of the Court ings could be closed.” Id., at 38-39 (citations omitted). Appellee therefore concludes: “At bottom, there can be no question that application of FACA will impair the sensitive and necessarily confidential process of gathering information to assess accurately the qualifications and character of prospective judicial nominees.” Id., at 39. Whatever the merits of these claims and whatever their relevance to appellee’s constitutional objections to FACA’s applicability, they certainly show, as appellants contend, that appellants might gain significant relief if they prevail in their suit. Appellants’ potential gains are undoubtedly sufficient to give them standing.7 Ill Section 3(2) of FACA, as set forth in 5 U. S. C. App. §3(2), defines “advisory committee” as follows: “For the purpose of this Act — “(2) The term ‘advisory committee’ means any committee, board, commission, council, conference, panel, task force, or other similar group, or any subcommittee or other subgroup thereof (hereafter in this paragraph referred to as ‘committee’), which is — “(A) established by statute or reorganization plan, or “(B) established or utilized by the President, or “(C) established or utilized by one or more agencies, in the interest of obtaining advice or recommendations for the President or one or more agencies or officers of the Federal Government, except that such term ex 7 The Justice Department concedes that appellants have standing to challenge the application of at least some of FACA’s provisions to the Justice Department’s consultations with the ABA Committee. See Brief for Federal Appellee 11-16. Because those challenges present the threshold question whether the ABA Committee constitutes an advisory committee for purposes of FACA, and because we hold that it does not, we need not address the Department’s claim that appellants lack standing to contest the application of certain other provisions. 452 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. eludes (i) the Advisory Commission on Intergovernmental Relations, (ii) the Commission on Government Procurement, and (iii) any committee which is composed wholly of full-time officers or employees of the Federal Government.” Appellants agree that the ABA Committee was not “established” by the President or the Justice Department. See Brief for Appellant in No. 88-429, p. 16; Brief for Appellant in No. 88-494, pp. 13, 15-16, 21. Equally plainly, the ABA Committee is a committee that furnishes “advice or recommendations” to the President via the Justice Department. Whether the ABA Committee constitutes an “advisory committee” for purposes of FACA therefore depends upon whether it is “utilized” by the President or the Justice Department as Congress intended that term to be understood. A There is no doubt that the Executive makes use of the ABA Committee, and thus “utilizes” it in one common sense of the term. As the District Court recognized, however, “reliance on the plain language of FACA alone is not entirely satisfactory.” 691 F. Supp., at 488. “Utilize” is a woolly verb, its contours left undefined by the statute itself. Read unqualifiedly, it would extend FACA’s requirements to any group of two or more persons, or at least any formal organization, from which the President or an Executive agency seeks advice.8 We are convinced that Congress did not intend that result. A nodding acquaintance with FACA’s pur- 8 FACA provides exceptions for advisory committees established or utilized by the Central Intelligence Agency or the Federal Reserve System, § 4(b), as well as for “any local civic group whose primary function is that of rendering a public service with respect to a Federal program, or any State or local committee, council, board, commission, or similar group established to advise or make recommendations to State or local officials or agencies.” § 4(c). The presence of these exceptions does little to curtail the almost unfettered breadth of a dictionary reading of FACA’s definition of “advisory committee.” PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 453 440 Opinion of the Court poses, as manifested by its legislative history and as recited in § 2 of the Act, reveals that it cannot have been Congress’ intention, for example, to require the filing of a charter, the presence of a controlling federal official, and detailed minutes any time the President seeks the views of the National Association for the Advancement of Colored People (NAACP) before nominating Commissioners to the Equal Employment Opportunity Commission, or asks the leaders of an American Legion Post he is visiting for the organization’s opinion on some aspect of military policy. Nor can Congress have meant—as a straightforward reading of “utilize” would appear to require—that all of FACA’s restrictions apply if a President consults with his own political party before picking his Cabinet. It was unmistakably not Congress’ intention to intrude on a political party’s freedom to conduct its affairs as it chooses, cf. Eu v. San Francisco County Democratic Central Comm., 489 U. S. 214, 230 (1989), or its ability to advise elected officials who belong to that party, by placing a federal employee in charge of each advisory group meeting and making its minutes public property. FAC A was enacted to cure specific ills, above all the wasteful expenditure of public funds for worthless committee meetings and biased proposals; although its reach is extensive, we cannot believe that it was intended to cover every formal and informal consultation between the President or an Executive agency and a group rendering advice.9 As we ’Justice Kennedy agrees with our conclusion that an unreflective reading of the term “utilize” would include the President’s occasional consultations with groups such as the NAACP and committees of the President’s own political party. See post, at 472. Having concluded that groups such as these are covered by the statute when they render advice, however, Justice Kennedy refuses to consult FACA’s legislative history—which he later denounces, with surprising hyperbole, as “unauthori-tative materials,” post, at 473, although countless opinions of this Court, including many written by the concurring Justices, have rested on just such materials—because this result would not, in his estimation, be “absurd,” post, at 472. Although this Court has never adopted so strict a 454 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. said in Church of the Holy Trinity v. United States, 143 U. S. 457, 459 (1892): “[Frequently words of general meaning are used in a statute, words broad enough to include an act in question, and yet a consideration of the whole legislation, or of the circumstances surrounding its enactment, or of the absurd results which follow from giving such broad meaning to the words, makes it unreasonable to believe that the legislator intended to include the particular act.” Where the literal reading of a statutory term would “compel an odd result,” Green v. Bock Laundry Machine Co., 490 U. S. 504, 509 (1989), we must search for other evidence of congressional intent to lend the term its proper scope. See also, e. g., Church of the Holy Trinity, supra, at 472; FDIC v. Philadelphia Gear Corp., 476 U. S. 426, 432 (1986). “The circumstances of the enactment of particular legislation,” for example, “may persuade a court that Congress did not intend words of common meaning to have their literal effect.” Watt v. Alaska, 451 U. S. 259, 266 (1981). Even though, as Judge Learned Hand said, “the words used, even in their literal sense, are the primary, and ordinarily the most reliable, source of interpreting the meaning of any writing,” nevertheless “it is one of the surest indexes of a mature and developed jurisprudence not to make a fortress out of the dictionary; standard for reviewing committee reports, floor debates, and other non-statutory indications of congressional intent, and we explicitly reject that standard today, see also infra, at 455, even if “absurdity” were the test, one would think it was met here. The idea that Members of Congress would vote for a bill subjecting their own political parties to bureaucratic intrusion and public oversight when a President or Cabinet officer consults with party committees concerning political appointments is outlandish. Nor does it strike us as in any way “unhealthy,” post, at 470, or undemocratic, post, at 473, to use all available materials in ascertaining the intent of our elected representatives, rather than read their enactments as requiring what may seem a disturbingly unlikely result, provided only that the result is not “absurd.” Indeed, the sounder and more democratic course, the course that strives for allegiance to Congress’ desires in all cases, not just those where Congress’ statutory directive is plainly sensible or borders on the lunatic, is the traditional approach we reaffirm today. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 455 440 Opinion of the Court but to remember that statutes always have some purpose or object to accomplish, whose sympathetic and imaginative discovery is the surest guide to their meaning.” Cabell v. Markham, 148 F. 2d 737, 739 (CA2), aff’d, 326 U. S. 404 (1945). Looking beyond the naked text for guidance is perfectly proper when the result it apparently decrees is difficult to fathom or where it seems inconsistent with Congress’ intention, since the plain-meaning rule is “rather an axiom of experience than a rule of law, and does not preclude consideration of persuasive evidence if it exists.” Boston Sand & Gravel Co. v. United States, 278 U. S. 41, 48 (1928) (Holmes, J.). See also United States v. American Trucking Assns., Inc., 310 U. S. 534, 543-544 (1940) (“When aid to construction of the meaning of words, as used in the statute, is available, there certainly can be no ‘rule of law’ which forbids its use, however clear the words may appear on ‘superficial examination’ ”) (citations omitted). Consideration of FACA’s purposes and origins in determining whether the term “utilized” was meant to apply to the Justice Department’s use of the ABA Committee is particularly appropriate here, given the importance we have consistently attached to interpreting statutes to avoid deciding difficult constitutional questions where the text fairly admits of a less problematic construction. See infra, at 465-467. It is therefore imperative that we consider indicators of congressional intent in addition to the statutory language before concluding that FAC A was meant to cover the ABA Committee’s provision of advice to the Justice Department in connection with judicial nominations. B Close attention to FACA’s history is helpful, for FAC A did not flare on the legislative scene with the suddenness of a meteor. Similar attempts to regulate the Federal Government’s use of advisory committees were common during the 20 years preceding FACA’s enactment. See Note, The Fed- 456 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. eral Advisory Committee Act, 10 Harv. J. Legis. 217, 219— 221 (1973). An understanding of those efforts is essential to ascertain the intended scope of the term “utilize.” In 1950, the Justice Department issued guidelines for the operation of federal advisory committees in order to forestall their facilitation of anticompetitive behavior by bringing industry leaders together with Government approval. See Hearings on WOC’s [Without Compensation Government employees] and Government Advisory Groups before the Antitrust Subcommittee of the House Committee on the Judiciary, 84th Cong., 1st Sess., pt. 1, pp. 586-587 (1955) (reprinting guidelines). Several years later, after the House Committee on Government Operations found that the Justice Department’s guidelines were frequently ignored, Representative Fascell sponsored a bill that would have accorded the guidelines legal status. H. R. 7390, 85th Cong., 1st Sess. (1957). Although the bill would have required agencies to report to Congress on their use of advisory committees and would have subjected advisory committees to various controls, it apparently would not have imposed any requirements on private groups, not established by the Federal Government, whose advice was sought by the Executive. See H. R. Rep. No. 576, 85th Cong., 1st Sess., 5-7 (1957); 103 Cong. Rec. 11252 (1957) (remarks of Rep. Fascell and Rep. Vorys). Despite Congress’ failure to enact the bill, the Bureau of the Budget issued a directive in 1962 incorporating the bulk of the guidelines. See Perritt & Wilkinson, Open Advisory Committees and the Political Process: The Federal Advisory Committee Act After Two Years, 63 Geo. L. J. 725, 731 (1975). Later that year, President Kennedy issued Executive Order No. 11007, 3 CFR 573 (1959-1963 Comp.), which governed the functioning of advisory committees until FACA’s passage. Executive Order No. 11007 is the probable source of the term “utilize” as later employed in FACA. The Order applied to advisory committees “formed by a PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 457 440 Opinion of the Court department or agency of the Government in the interest of obtaining advice or recommendations,” or “not formed by a department or agency, but only during any period when it is being utilized by a department or agency in the same manner as a Government-formed advisory committee.” §2(a) (emphasis added). To a large extent, FAC A adopted wholesale the provisions of Executive Order No. 11007. For example, like FACA, Executive Order No. 11007 stipulated that no advisory committee be formed or utilized unless authorized by law or determined as a matter of formal record by an agency head to be in the public interest, § 3; that all advisory committee meetings be held in the presence of a Government employee empowered to adjourn the meetings whenever he or she considered adjournment to be in the public interest, § 6(b); that meetings only occur at the call of, or with the advance approval of, a federal employee, § 6(a); that minutes be kept of the meetings, §§ 6(c), (d); and that committees terminate after two years unless a statute or an agency head decreed otherwise, § 8. There is no indication, however, that Executive Order No. 11007 was intended to apply to the Justice Department’s consultations with the ABA Committee. Neither President Kennedy, who issued the Order, nor President Johnson, nor President Nixon apparently deemed the ABA Committee to be “utilized” by the Department of Justice in the relevant sense of that term. Notwithstanding the ABA Committee’s highly visible role in advising the Justice Department regarding potential judicial nominees, and notwithstanding the fact that the Order’s requirements were established by the Executive itself rather than Congress, no President or Justice Department official applied them to the ABA Commitfee. As an entity formed privately, rather than at the Federal Government’s prompting, to render confidential advice with respect to the President’s constitutionally specified power to nominate federal judges—-an entity in receipt of no federal funds and not amenable to the strict management by 458 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. agency officials envisaged by Executive Order No. 11007— the ABA Committee cannot easily be said to have been “utilized by a department or agency in the same manner as a Government-formed advisory committee.” That the Executive apparently did not consider the ABA Committee’s activity within the terms of its own Executive Order is therefore unsurprising. Although FACA’s legislative history evinces an intent to widen the scope of Executive Order No. 11007’s definition of “advisory committee” by including “Presidential advisory committees,” which lay beyond the reach of Executive Order No. 11007,10 see H. R. Rep. No. 91-1731, pp. 9-10 (1970); H. R. Rep. No. 92-1017, p. 4 (1972); S. Rep. No. 92-1098, pp. 3-5, 7 (1972), as well as to augment the restrictions ap- 10 Neither Public Citizen nor WLF contends that the ABA Committee is a Presidential advisory committee as Congress understood that term. Nor does it appear to be one. In a House Report on the effectiveness of federal advisory committees, which provided the impetus for legislative proposals that eventually produced FACA, the Committee on Government Operations noted that Presidential committees were a special concern because they often consumed large amounts of federal money and were subject to no controls. The House Committee, however, defined “Presidential committee” narrowly, “as a group with either one or all of its members appointed by the President with a function of advising or making recommendations to him.” H. R. Rep. No. 91-1731, p. 10 (1970). None of the ABA Committee’s members are appointed by the President, nor does the ABA Committee report directly to him. The House and Senate Reports accompanying early versions of FACA likewise referred to advisory committees “formed” or “established” or “organized” by the President, or to committees created by an Act of Congress to advise the President—categories into which the ABA Committee cannot readily be fitted. See H. R. Rep. No. 92-1017, pp. 4-5 (1972); S. Rep. No. 92-1098, p. 7 (1972). Although FACA itself provides a more open-ended definition of “Presidential advisory committee,” applying it to “an advisory committee which advises the President,” § 3(4), as set forth in 5 U. S. C. § 3(4), that category is a species of “advisory committee,” and does not purport to cover committees advising the President that were not “established or utilized” by him. As FACA’s legislative history reveals, the Presidential advisory committees Congress intended FACA to reach do not include the ABA Committee. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 459 440 Opinion of the Court plicable to advisory committees covered by the statute, there is scant reason to believe that Congress desired to bring the ABA Committee within FACA’s net. FACA’s principal purpose was to enhance the public accountability of advisory committees established by the Executive Branch and to reduce wasteful expenditures on them. That purpose could be accomplished, however, without expanding the coverage of Executive Order No. 11007 to include privately organized committees that received no federal funds. Indeed, there is considerable evidence that Congress sought nothing more than stricter compliance with reporting and other requirements—which were made more stringent—by advisory committees already covered by the Order and similar treatment of a small class of publicly funded groups created by the President. The House bill which in its amended form became FACA applied exclusively to advisory committees “established” by statute or by the Executive, whether by a federal agency or by the President himself. H. R. 4383, 92d Cong., 2d Sess. § 3(2) (1972). Although the House Committee Report stated that the class of advisory committees was to include “committees which may have been organized before their advice was sought by the President or any agency, but which are used by the President or any agency in the same way as an advisory committee formed by the President himself or the agency itself,” H. R. Rep. No. 92-1017, supra, at 4, it is questionable whether the Report’s authors believed that the Justice Department used the ABA Committee in the same way as it used advisory committees it established. The phrase “used • • . in the same way” is reminiscent of Executive Order No. H007’s reference to advisory committees “utilized ... in the same manner” as a committee established by the Federal Government, and the practice of three administrations demonstrates that Executive Order No. 11007 did not encompass the ABA Committee. 460 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. This inference draws support from the earlier House Report which instigated the legislative efforts that culminated in FACA. That Report complained that committees “utilized” by an agency—as opposed to those established directly by an agency—rarely complied with the requirements of Executive Order No. 11007. See H. R. Rep. No. 91-1731, supra, at 15. But it did not cite the ABA Committee or similar advisory committees as willful evaders of the Order. Rather, the Report’s paradigmatic example of a committee “utilized” by an agency for purposes of Executive Order No. 11007 was an advisory committee established by a quasipublic organization in receipt of public funds, such as the National Academy of Sciences.11 There is no indication in the Report that a purely private group like the ABA Committee that was not formed by the Executive, accepted no public funds, and assisted the Executive in performing a constitutionally specified task committed to the Executive was within the terms of Executive Order No. 11007 or was the type of advisory entity that legislation was urgently needed to address. 11 The relevant paragraph of H. R. Rep. No. 91-1731, supra, at 15 (footnotes omitted), reads in full: “The definition, further, states ‘the term also includes any committee, board, . . . that is not formed by a department or agency, when it is being utilized by a department or agency in the same manner as a Government-formed advisory committee.’ Rarely were such committees reported. A great number of the approximately 500 advisory committees of the National Academy of Sciences (NAS) and its affiliates possibly should be added to the above 1800 advisory committees as the NAS committees fall within the intent and literal definition of advisory committees under Executive Order 11007. The National Academy of Sciences was created by Congress as a semi-private organization for the explicit purpose of furnishing advice to the Government. This is done by the use of advisory committees. The Government meets the expense of investigations and reports prepared by the Academy committees at the request of the Government. Yet, very few of the Academy committees were reported by the agencies and departments of the Government.” PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 461 440 Opinion of the Court Paralleling the initial House bill, the Senate bill that grew into FACA defined “advisory committee” as one “established or organized” by statute, the President, or an Executive agency. S. 3529, 92d Cong., 2d Sess. §§3(1), (2) (1972). Like the House Report, the accompanying Senate Report stated that the phrase “established or organized” was to be understood in its “most liberal sense, so that when an officer brings together a group by formal or informal means, by contract or other arrangement, and whether or not Federal money is expended, to obtain advice and information, such group is covered by the provisions of this bill.” S. Rep. No. 92-1098, supra, at 8. While the Report manifested a clear intent not to restrict FACA’s coverage to advisory committees funded by the Federal Government, it did not indicate any desire to bring all private advisory committees within FACA’s terms. Indeed, the examples the Senate Report offers—“the Advisory Council on Federal Reports, the National Industrial Pollution Control Council, the National Petroleum Council, advisory councils to the National Institutes of Health, and committees of the national academies where they are utilized and officially recognized as advisory to the President, to an agency, or to a Government official,” ibid.— are limited to groups organized by, or closely tied to, the Federal Government, and thus enjoying quasi-public status. Given the prominence of the ABA Committee’s role and its familiarity to Members of Congress, its omission from the list of groups formed and maintained by private initiative to offer advice with respect to the President’s nomination of Government officials is telling. If the examples offered by the Senate Committee on Government Operations are representative, as seems fair to surmise, then there is little reason to think that there was any support, at least at the committee stage, for going beyond the terms of Executive Order No. 11007 to regulate comprehensively the workings of the ABA Committee. It is true that the final version of FACA approved by both Houses employed the phrase “established or utilized,” 462 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. and that this phrase is more capacious than the word “established” or the phrase “established or organized.” But its genesis suggests that it was not intended to go much beyond those narrower formulations. The words “or utilized” were added by the Conference Committee to the definition included in the House bill. See H. R. Conf. Rep. No. 92-1403, p. 2 (1972). The Joint Explanatory Statement, however, said simply that the definition contained in the House bill was adopted “with modification.” Id., at 9. The Conference Report offered no indication that the modification was significant, let alone that it would substantially broaden FACA’s application by sweeping within its terms a vast number of private groups, such as the Republican National Committee, not formed at the behest of the Executive or by quasi-public organizations whose opinions the Federal Government sometimes solicits. Indeed, it appears that the House bill’s initial restricted focus on advisory committees established by the Federal Government, in an expanded sense of the word “established,” was retained rather than enlarged by the Conference Committee. In the section dealing with FACA’s range of application, the Conference Report stated: “The Act does not apply to persons or organizations which have contractual relationships with Federal agencies nor to advisory committees not directly established by or for such agencies” Id., at 10 (emphasis added). The phrase “or utilized” therefore appears to have been added simply to clarify that FACA applies to advisory committees established by the Federal Government in a generous sense of that term, encompassing groups formed indirectly by quasi-public organizations such as the National Academy of Sciences “for” public agencies as well as “by” such agencies themselves. Read in this way, the term “utilized” would meet the concerns of the authors of House Report No. 91-1731 that advisory committees covered by Executive Order No. 11007, because they were “utilized by a department or agency in the same manner as a Government-formed advisory commit- PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 463 440 Opinion of the Court tee”—such as the groups organized by the National Academy of Sciences and its affiliates which the Report discussed— would be subject to FACA’s requirements. And it comports well with the initial House and Senate bills’ limited extension to advisory groups “established,” on a broad understanding of that word, by the Federal Government, whether those groups were established by the Executive Branch or by statute or whether they were the offspring of some organization created or permeated by the Federal Government. Read in this way, however, the word “utilized” does not describe the Justice Department’s use of the ABA Committee. Consultations between the Justice Department and the ABA Committee were not within the purview of Executive Order No. 11007, nor can the ABA Committee be said to have been formed by the Justice Department or by some semiprivate entity the Federal Government helped bring into being. In sum, a literalistic reading of §3(2) would bring the Justice Department’s advisory relationship with the ABA Committee within FACA’s terms, particularly given FACA’s objective of opening many advisory relationships to public scrutiny except in certain narrowly defined situations.12 A 12 Appellants note as well that regulations of the General Services Administration (GSA), the agency responsible for administering FACA, define a “utilized” advisory committee as “a committee or other group composed in whole or in part of other than full-time officers or employees of the Federal Government with an established existence outside the agency seeking its advice which the President or agency official(s) adopts, such as through institutional arrangements, as a preferred source from which to obtain advice or recommendations . . . in the same manner as that individual would obtain advice or recommendations from an established advisory committee.” 41 CFR § 101-6.1003 (1988). Appellants argue that the ABA Committee comes within the terms of this regulatory definition, because it exists outside the Justice Department and because it serves as a “preferred source” of advice, inasmuch as the ABA Committee’s recommendations regarding potential judicial nominees are unfailingly requested and accorded considerably more weight than 464 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. literalistic reading, however, would catch far more groups and consulting arrangements than Congress could conceivably have intended. And the careful review which this interpretive difficulty warrants of earlier efforts to regulate those advanced by other groups. See Brief for Appellant in No. 88-429, pp. 17-18; Brief for Appellant in No. 88-494, pp. 18-20. This argument is not without force. For several reasons, however, we do not think it conclusive, either alone or together with appellants’ arguments from FACA’s text and legislative history. The first is that the regulation, like FACA’s definition of “advisory committee,” appears too sweeping to be read without qualification unless further investigation of congressional intent confirms that reading. And our review of FACA’s legislative history and purposes demonstrates that the Justice Department, assisting the Executive’s exercise of a constitutional power specifically assigned to the Executive alone, does not use the ABA Committee in what is obviously the “same manner” as federal agencies use other advisory committees established by them or by some other creature of the Federal Government. Second, appellants’ claim that the regulation applies to the ABA Committee is questionable. GSA publishes an annual report listing advisory committees covered by FAC A. Although 17 reports have thus far been issued, not once has the ABA Committee been included in that list. The agency’s own interpretation of its regulation thus appears to contradict the expansive construction appellants ask us to give it—a fact which, though not depriving the regulation’s language of independent force, see post, at 479, nevertheless weakens the claim that the regulation applies to the Justice Department’s use of the ABA Committee. Third, even if the ABA Committee were covered by the regulation, appellants’ case would not be appreciably bolstered. Deference to the agency’s expertise in interpreting FACA is less appropriate here than it would be were the regulatory definition a contemporaneous construction of the statute, since the current definition was first promulgated in 1983, see 48 Fed. Reg. 19327 (1983), and did not become final until 1987, see 52 Fed. Reg. 45930 (1987)—more than a decade after FACA’s passage. See, e. g., Aluminum Co. of America v. Central Lincoln Peoples’ Utility Dist., 467 U. S. 380, 390 (1984); Zenith Radio Corp. v. United States, 437 U. S. 443, 450 (1978); General Electric Co. v. Gilbert, 429 U. S. 125, 142 (1976) (discounting significance of agency interpretive guideline promulgated eight years after statute’s enactment, although fact that guideline contradicted agency’s earlier position deemed “more importan[t]”); Udall v. Tailman, 380 U. S. 1, 16 (1965); Power Reactor Co. v. Electricians, 367 U. S. 396, PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 465 440 Opinion of the Court federal advisory committees and the circumstances surrounding FACA’s adoption strongly suggests that FACA’s definition of “advisory committee” was not meant to encompass the ABA Committee’s relationship with the Justice Department. That relationship seems not to have been within the contemplation of Executive Order No. 11007. And FACA’s legislative history does not display an intent to widen the Order’s application to encircle it. Weighing the deliberately inclusive statutory language against other evidence of congressional intent, it seems to us a close question whether FACA should be construed to apply to the ABA Committee, although on the whole we are fairly confident it should not. There is, however, one additional consideration which, in our view, tips the balance decisively against FACA’s application. C “When the validity of an act of the Congress is drawn in question, and even if a serious doubt of constitutionality is raised, it is a cardinal principle that this Court will first ascertain whether a construction of the statute is fairly possible 408 (1961); Norwegian Nitrogen Products Co. v. United States, 288 U. S. 294, 315 (1933). In addition, we owe GSA’s regulation diminished deference for a reason independent of its not having been issued contemporaneously with FACA’s passage. In General Electric Co. v. Gilbert, supra, we held that an agency’s interpretive regulations not promulgated pursuant to express statutory authority should be accorded less weight than “administrative regulations which Congress has declared shall have the force of law, or to regulations which under the enabling statute may themselves supply the basis for imposition of liability.” Id., at 141 (citations omitted). GSA’s regulatory definition falls into neither category. Section 7(c), as set forth m 5 U. S. C. App. § 7(c), authorizes the Administrator to “prescribe administrative guidelines and management controls applicable to advisory committees, and, to the maximum extent feasible, provide advice, assistance, and guidance to advisory committees to improve their performance.” It does not empower the agency to issue, in addition to these guidelines, a regulatory definition of “advisory committee” carrying the force of law. Justice Kennedy’s assertion that GSA’s interpretation of FACA’s provisions is “binding,” post, at 478, 480, confuses wish with reality. 466 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. by which the question may be avoided.” Crowell v. Benson, 285 U. S. 22, 62 (1932) (footnote collecting citations omitted). It has long been an axiom of statutory interpretation that “where an otherwise acceptable construction of a statute would raise serious constitutional problems, the Court will construe the statute to avoid such problems unless such construction is plainly contrary to the intent of Congress.” Edward J. DeBartolo Corp. n. Florida Gulf Coast Building & Construction Trades Council, 485 U. S. 568, 575 (1988). See also St. Martin Evangelical Lutheran Church v. South Dakota, 451 U. S. 772, 780 (1981); NLRB v. Catholic Bishop of Chicago, 440 U. S. 490, 500-501 (1979); Machinists v. Street, 367 U. S. 740, 749-750 (1961). This approach, we said recently, “not only reflects the prudential concern that constitutional issues not be needlessly confronted, but also recognizes that Congress, like this Court, is bound by and swears an oath to uphold the Constitution.” Edward J. DeBartolo Corp., supra, at 575. Our reluctance to decide constitutional issues is especially great where, as here, they concern the relative powers of coordinate branches of government. See American Foreign Service Assn. v. Garfinkel, 490 U. S. 153, 161 (1989) (per curiam). Hence, we are loath to conclude that Congress intended to press ahead into dangerous constitutional thickets in the absence of firm evidence that it courted those perils. That construing FACA to apply to the Justice Department’s consultations with the ABA Committee would present formidable constitutional difficulties is undeniable. The District Court declared FACA unconstitutional insofar as it applied to those consultations, because it concluded that FACA, so applied, infringed unduly on the President’s Article II power to nominate federal judges and violated the doctrine of separation of powers.13 Whether or not the court’s conclu- 13 In addition, appellee American Bar Association contends that application of FACA to the ABA Committee would impermissibly interfere with the associational and expressive rights guaranteed its members by the PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 467 440 Kennedy, J., concurring in judgment sion was correct, there is no gainsaying the seriousness of these constitutional challenges. To be sure, “[w]e cannot press statutory construction ‘to the point of disingenuous evasion’ even to avoid a constitutional question.” United States v. Locke, 471 U. S. 84, 96 (1985), quoting Moore Ice Cream Co. v. Rose, 289 U. S. 373, 379 (1933). But unlike in Locke, where “nothing in the legislative history remotely suggest[ed] a congressional intent contrary to Congress’ chosen words,” 471 U. S., at 96, our review of the regulatory scheme prior to FACA’s enactment and the likely origin of the phrase “or utilized” in FACA’s definition of “advisory committee” reveals that Congress probably did not intend to subject the ABA Committee to FACA’s requirements when the ABA Committee offers confidential advice regarding Presidential appointments to the federal bench. Where the competing arguments based on FACA’s text and legislative history, though both plausible, tend to show that Congress did not desire FACA to apply to the Justice Department’s confidential solicitation of the ABA Committee’s views on prospective judicial nominees, sound sense counsels adherence to our rule of caution. Our unwillingness to resolve important constitutional questions unnecessarily thus solidifies our conviction that FACA is inapplicable. The judgment of the District Court is Affirmed. Justice Scalia took no part in the consideration or decision of these cases. Justice Kennedy, with whom The Chief Justice and Justice O’Connor join, concurring in the judgment. “In a government, where the liberties of the people are to be preserved . . . , the executive, legislative and judicial, should ever be separate and distinct, and consist First Amendment. See Brief for Appellee ABA 40-48; Brief for People for the American Way Action Fund and Alliance for Justice as Amicus Curiae 22-29. 468 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. of parts, mutually forming a check upon each other.” C. Pinckney, Observations on the Plan of Government Submitted to the Federal Convention of May 28, 1787, reprinted in 3 M. Farrand, Records of the Federal Convention of 1787, p. 108 (rev. ed. 1966). The Framers of our Government knew that the most precious of liberties could remain secure only if they created a structure of Government based on a permanent separation of powers. See, e. g., The Federalist Nos. 47-51 (J. Madison). Indeed, the Framers devoted almost the whole of their attention at the Constitutional Convention to the creation of a secure and enduring structure for the new Government. It remains one of the most vital functions of this Court to police with care the separation of the governing powers. That is so even when, as is the case here, no immediate threat to liberty is apparent. When structure fails, liberty is always in peril. As Justice Frankfurter stated: “The accretion of dangerous power does not come in a day. It does come, however slowly, from the generative force of unchecked disregard of the restrictions that fence in even the most disinterested assertion of authority.” Youngstown Sheet & Tube Co. v. Sawyer, 343 U. S. 579, 594 (1952) (concurring opinion). Although one is perhaps more obvious than the other, this suit presents two distinct issues of the separation of powers. The first concerns the rules this Court must follow in interpreting a statute passed by Congress and signed by the President. On this subject, I cannot join the Court’s conclusion that the Federal Advisory Committee Act (FACA), 85 Stat. 770, as amended, 5 U. S. C. App. § 1 et seq. (1982 ed. and Supp. V), does not cover the activities of the American Bar Association’s Standing Committee on Federal Judiciary in advising the Department of Justice regarding potential nominees for federal judgeships. The result seems sensible in the abstract; but I cannot accept the method by which the Court PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 469 440 Kennedy, J., concurring in judgment arrives at its interpretation of FACA, which does not accord proper respect to the finality and binding effect of legislative enactments. The second question in the case is the extent to which Congress may interfere with the President’s constitutional prerogative to nominate federal judges. On this issue, which the Court does not reach because of its conclusion on the statutory question, I think it quite plain that the application of FACA to the Government’s use of the ABA Committee is unconstitutional. I The statutory question in this suit is simple enough to formulate. FACA applies to “any committee” that is “established or utilized” by the President or one or more agencies, and which furnishes “advice or recommendations” to the President or one or more agencies. 5 U. S. C. App. § 3(2). All concede that the ABA Committee furnishes advice and recommendations to the Department of Justice and through it to the President. Ante, at 452. The only question we face, therefore, is whether the ABA Committee is “utilized” by the Department of Justice or the President. See ibid. There is a ready starting point, which ought to serve also as a sufficient stopping point, for this kind of analysis: the plain language of the statute. Yet the Court is unwilling to rest on this foundation, for several reasons. One is an evident unwillingness to define the application of the statute in terms of the ordinary meaning of its language. We are told that “utilize” is “a woolly verb,” ibid., and therefore we cannot be content to rely on what is described, with varying levels of animus, as a “literal reading,” ante, at 454, a “literalistic reading,” ante, at 463, 464, and “a dictionary reading” of this word, ante, at 452, n. 8. We also are told in no uncertain terms that we cannot rely on (what I happen to regard as a more accurate description) “a straightforward reading of ‘utilize.’” Ante, at 453. Reluctance to working with the basic meaning of words in a normal manner undermines the legal process. These cases demonstrate that reluctance of this 470 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. sort leads instead to woolly judicial construction that mars the plain face of legislative enactments. The Court concedes that the Executive Branch “utilizes” the ABA Committee in the common sense of that word. Ibid. Indeed, this point cannot be contested. As the Court’s own recitation of the facts makes clear, the Department of Justice has, over the last four decades, made regular use of the ABA Committee to investigate the background of potential nominees and to make critical recommendations regarding their qualifications. See ante, at 443-445. This should end the matter. The Court nevertheless goes through several more steps to conclude that, although “it seems to us a close question,” ante, at 465, Congress did not intend that FACA would apply to the ABA Committee. Although I believe the Court’s result is quite sensible, I cannot go along with the unhealthy process of amending the statute by judicial interpretation. Where the language of a statute is clear in its application, the normal rule is that we are bound by it. There is, of course, a legitimate exception to this rule, which the Court invokes, see ante, at 453-454, citing Church of the Holy Trinity v. United States, 143 U. S. 457, 459 (1892), and with which I have no quarrel. Where the plain language of the statute would lead to “patently absurd consequences,” United States v. Brown, 333 U. S. 18, 27 (1948), that “Congress could not possibly have intended,” FBI v. Abramson, 456 U. S. 615, 640 (1982) (O’Connor, J., dissenting) (emphasis added), we need not apply the language in such a fashion. When used in a proper manner, this narrow exception to our normal rule of statutory construction does not intrude upon the lawmaking powers of Congress, but rather demonstrates a respect for the coequal Legislative Branch, which we assume would not act in an absurd way. This exception remains a legitimate tool of the Judiciary, however, only as long as the Court acts with self-discipline by limiting the exception to situations where the result of applying the plain language would be, in a genuine sense, absurd, PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 471 440 Kennedy, J., concurring in judgment i. e., where it is quite impossible that Congress could have intended the result, see ibid., and where the alleged absurdity is so clear as to be obvious to most anyone. A few examples of true absurdity are given in the Holy Trinity decision cited by the Court, ante, at 454, such as where a sheriff was prosecuted for obstructing the mails even though he was executing a warrant to arrest the mail carrier for murder, or where a medieval law against drawing blood in the streets was to be applied against a physician who came to the aid of a man who had fallen down in a fit. See 143 U. S., at 460-461. In today’s opinion, however, the Court disregards the plain language of the statute not because its application would be patently absurd, but rather because, on the basis of its view of the legislative history, the Court is “fairly confident” that “FACA should [not] be construed to apply to the ABA Committee.” Ante, at 465. I believe the Court’s loose invocation of the “absurd result” canon of statutory construction creates too great a risk that the Court is exercising its own “will instead of judgment,” with the consequence of “substituti[ng] [its own] pleasure to that of the legislative body.” The Federalist No. 78, p. 469 (C. Rossiter ed. 1961) (A. Hamilton). The Court makes only a passing effort to show that it would be absurd to apply the term “utilize” to the ABA Committee according to its commonsense meaning. It offers three examples that we can assume are meant to demonstrate this point: the application of FACA to an American Legion Post should the President visit that organization and happen to ask its opinion on some aspect of military policy; the application of FACA to the meetings of the National Association for the Advancement of Colored People (NAACP) should the President seek its views in nominating Commissioners to the Equal Employment Opportunity Commission; and the application of FACA to the national committee of the President’s political party should he consult it for advice and 472 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. recommendations before picking his Cabinet. See ante, at 452-453. None of these examples demonstrate the kind of absurd consequences that would justify departure from the plain language of the statute. A commonsense interpretation of the term “utilize” would not necessarily reach the kind of ad hoc contact with a private group that is contemplated by the Court’s American Legion hypothetical. Such an interpretation would be consistent, moreover, with the regulation of the General Services Administration (GSA) interpreting the word “utilize,” which the Court in effect ignores. See infra, at 477. As for the more regular use contemplated by the Court’s examples concerning the NAACP and the national committee of the President’s political party, it would not be at all absurd to say that, under the Court’s hypothetical, these groups would be “utilized” by the President to obtain “advice or recommendations” on appointments, and therefore would fall within the coverage of the statute. Rather, what is troublesome about these examples is that they raise the very same serious constitutional questions that confront us here (and perhaps others as well).1 The Court confuses the two points. The fact that a particular application of the clear terms of a statute might be unconstitutional does not, in and of itself, render a straightforward application of the language absurd, so as to allow us to conclude that the statute does not apply. See infra, at 481. Unable to show that an application of FAC A according the plain meaning of its terms would be absurd, the Court turns instead to the task of demonstrating that a straightforward reading of the statute would be inconsistent with the congressional purposes that lay behind its passage. To the student of statutory construction, this move is a familiar one. It is, as the Court identifies it, the classic Holy Trinity argument. “[A] thing may be within the letter of the statute and 11 do not address here any possible problems under the First Amendment with the application of FACA to the ABA Committee. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 473 440 Kennedy, J., concurring in judgment yet not within the statute, because not within its spirit, nor within the intention of its makers.” Holy Trinity, supra, at 459. I cannot embrace this principle. Where it is clear that the unambiguous language of a statute embraces certain conduct, and it would not be patently absurd to apply the statute to such conduct, it does not foster a democratic exegesis for this Court to rummage through unauthoritative materials to consult the spirit of the legislation in order to discover an alternative interpretation of the statute with which the Court is more comfortable. It comes as a surprise to no one that the result of the Court’s lengthy journey through the legislative history is the discovery of a congressional intent not to include the activities of the ABA Committee within the coverage of FACA. The problem with spirits is that they tend to reflect less the views of the world whence they come than the views of those who seek their advice. Lest anyone think that my objection to the use of the Holy Trinity doctrine is a mere point of interpretive purity divorced from more practical considerations, I should pause for a moment to recall the unhappy genesis of that doctrine and its unwelcome potential. In Holy Trinity, the Court was faced with the interpretation of a statute which made it unlawful for “any person, company, partnership, or corporation, in any manner whatsoever, to prepay the transportation, or in any way assist or encourage the importation or migration of any alien or aliens, any foreigner or foreigners, into the United States . . . , under contract or agreement . . . made previous to the importation or migration of such alien or aliens, foreigner or foreigners, to perform labor or service of any kind in the United States.” 143 U. S., at 458. The Church of the Holy Trinity entered into a contract with an alien residing in England to come to the United States to serve as the director and pastor of the church. Notwithstanding the fact that this agreement fell within the plain lan 474 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. guage of the statute, which was conceded to be the case, see ibid., the Court overrode the plain language, drawing instead on the background and purposes of the statute to conclude that Congress did not intend its broad prohibition to cover the importation of Christian ministers. The central support for the Court’s ultimate conclusion that Congress did not intend the law to cover Christian ministers is its lengthy review of the “mass of organic utterances” establishing that “this is a Christian nation,” and which were taken to prove that it could not “be believed that a Congress of the United States intended to make it a misdemeanor for a church of this country to contract for the services of a Christian minister residing in another nation.” Id., at 471. I should think the potential of this doctrine to allow judges to substitute their personal predelictions for the will of the Congress is so self-evident from the case which spawned it as to require no further discussion of its susceptibility to abuse. Even if I were inclined to disregard the unambiguous language of FACA, I could not join the Court’s conclusions with regard to Congress’ purposes. I find the Court’s treatment of the legislative history one sided and offer a few observations on the difficulties of perceiving the true contours of a spirit. The first problem with the Court’s use of legislative history is the questionable relevance of its detailed account of Executive practice before the enactment of FACA. This background is interesting but not instructive, for as the Court acknowledges, even the legislative history as presented by the Court “evinces an intent to widen the scope of” the coverage of prior Executive Orders, ante, at 458, and in any event the language of the statute is “more capacious” than any of the previous “narrower formulations,” ante, at 462. Indeed, Congress would have had little reason to legislate at all in this area if it had intended FACA to be nothing more than a reflection of the provisions of Executive Order No. 11007, 3 CFR 573 (1959-1963 Comp.), which was already the settled PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 475 440 Kennedy, J., concurring in judgment and governing law at the time this bill was introduced, considered, and enacted. In other words, the background to FACA cannot be taken to illuminate its breadth precisely because FACA altered the landscape to address the many concerns Congress had about the increasing growth and use of advisory committees. Another problem with the Court’s approach lies in its narrow preoccupation with the ABA Committee against the background of a bill that was intended to provide comprehensive legislation covering a widespread problem in the organization and operation of the Federal Government. The Court’s discussion takes portentous note of the fact that Congress did not mention or discuss the ABA Committee by name in the materials that preceded the enactment of FACA. But that is hardly a remarkable fact. The legislation was passed at a time when somewhere between 1,800 and 3,200 target committees were thought to be in existence, see S. Rep. No. 92-1098, pp. 3, 4 (1972), and the congressional Reports mentioned few committees by name. More to the point, its argument reflects an incorrect understanding of the kinds of laws Congress passes: it usually does not legislate by specifying examples, but by identifying broad and general principles that must be applied to particular factual instances. And that is true of FACA. Finally, though the stated objective of the Court’s inquiry into legislative history is the identification of Congress’ purposes in passing FACA, the inquiry does not focus on the most obvious place for finding those purposes, which is the section of the Conference Committee Report entitled “Findings and Purposes.” That section lists six findings and purposes that underlie FACA: “(1) the need for many existing advisory committees has not been adequately reviewed; “(2) new advisory committees should be established only when they are determined to be essential and their number should be kept to the minimum necessary; 476 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. “(3) advisory committees should be terminated when they are no longer carrying out the purposes for which they were established; “(4) standards and uniform procedures should govern the establishment, operation, administration, and duration of advisory committees; “(5) the Congress and the public should be kept informed with respect to the number, purpose, membership, activities, and cost of advisory committees; and “(6) the function of advisory committees should be advisory only, and that all matters under their consideration should be determined, in accordance with law, by the official, agency, or officer involved.” H. R. Conf. Rep. No. 92-1403, pp. 1-2 (1972). The most pertinent conclusion to be drawn from this list of purposes is that all of them are implicated by the Justice Department’s use of the ABA Committee. In addition, it shows that Congress’ stated purposes for addressing the use of advisory committees went well beyond the amount of public funds devoted to their operations, which in any event is not the sole component in the cost of their use; thus the Court errs in focusing on this point. It is most striking that this section of the Conference Committee Report, which contains Congress’ own explicit statement of its purposes in adopting F AC A, receives no mention by the Court on its amble through the legislative history. The one statement the Court does quote from this Report, that F AC A does not apply “ ‘to advisory committees not directly established by or for [federal] agencies,’” ante, at 462, quoting H. R. Conf. Rep. 92-1403, supra, at 10 (emphasis deleted), is of uncertain value. It is not clear that this passage would exclude the ABA Committee, which was established in 1946 and began almost at once to advise the Government on judicial nominees. It also is not clear why the reasons a committee was formed should determine whether and how they are “utilized by” the Government, or how this consideration PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 477 440 Kennedy, J., concurring in judgment can be squared with the plain language of the statute. The Court professes puzzlement because the Report says only that the Conference Committee modified the definition of “advisory committee” to include the phrase “or utilized,” but does not explain the extent of the modification in any detail. Ante, at 461-462. One would have thought at least that the Court would have been led to consider how the specific purposes Congress identified for this legislation might shed light on the reasons for the change. Not only does the Court’s decision today give inadequate respect to the statute passed by Congress, it also gives inadequate deference to the GSA’s regulations interpreting FAC A. I have already mentioned that, under the GSA’s interpretation of FACA, the Court’s hypothetical applications of the Act to groups such as the American Legion are impossible. More important, however, it is plain that, under the GSA’s regulations, the ABA Committee is covered by the Act. The GSA defines a “utilized” advisory committee as “a committee or other group composed in whole or in part of other than full-time officers or employees of the Federal Government with an established existence outside the agency seeking its advice which the President or agency official(s) adopts, such as through institutional arrangements, as a preferred source from which to obtain advice or recommendations on a specific issue or policy within the scope of his or her responsibilities in the same manner as that individual would obtain advice or recommendations from an established advisory committee.” 41 CFR § 101-6.1003 (1988). I cannot imagine a better description of the function of the ABA Committee. First, the ABA Committee is “composed in whole or in part of other than full-time officers or employees of the Federal Government. ” Second, the committee has an established existence outside the agency seeking its advice.” Third, the committee has been adopted by the Department of Justice “as a preferred source from which to ob 478 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. tain advice or recommendations of a specific issue or policy.” Indeed, the committee performs no other significant function beyond advising the Government on judicial appointments. Fourth, the relation is carried out through what cannot in fairness be denied, after four decades, to be an “institutional arrangement.” The committee’s views are sought on a regular and frequent basis, are given careful consideration, and are usually followed by the Department. Fifth, the committee is used to obtain advice and recommendations on judicial appointments “in the same manner as ... an established advisory committee.” In this regard, it is pertinent that the Department discloses to the committee the names of the candidates and other confidential Government information. This unusual privilege is normally accorded only to other parts of the Government. The Court concedes that the regulations present difficulties for its conclusion that FACA does not apply to the ABA Committee. Ante, at 464, n. 12. It nevertheless relegates its entire discussion of this controlling point to a footnote appended as a ragged afterthought to its extensive discussion of the legislative history. See ante, at 463-465, n. 12. The Court offers four reasons for slighting the agency’s interpretation in favor of its own. First, we are told that the language of the GSA regulations, like the statute itself, “appears too sweeping” to be read according to its terms. Of course, once again the Court does not mean either that the agency regulation is not a reasonable interpretation of the plain language of the statute, or that the agency interpretation itself would produce absurd consequences. Rather, what the Court means is that the agency regulation is not entirely consistent with the “spirit” of the Act which it professes to have divined from the legislative history. I do not think this a sound reason for ignoring the binding interpretation of the statute rendered by the implementing agency. Second, the Court tells us that it “is questionable” whether the GSA regulations apply to the ABA Committee. This is PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 479 440 Kennedy, J., concurring in judgment quite wrong. The Court does not deny that the committee falls squarely within the terms of the regulations. The Court’s doubts on this issue stem entirely from the fact that the GSA’s annual report does not list the ABA Committee as one of the advisory committees covered by FACA. But it seems to me to be without relevance one way or the other whether the GSA is aware that the regulations cover the committee. What matters is that the regulations the GSA adopted, which contain a very reasonable interpretation of the statute, plainly cover the committee. If the Court’s interpretive approach on this issue were accepted, then the text of the agency’s regulations, for which notice was afforded and upon which comment was received, would be of no independent force. Third, the Court notes that the agency’s interpretation was not promulgated until 1983 and not made final until 1987, whereas FACA was passed in 1972. I cannot imagine why it is a sensible principle that an agency regulation which is promulgated a decade after the initial passage of a statute should be given less deference because of the mere passage of time. I would not draw any such distinction one way or the other, but if anything one would think that the GSA’s regulation should be entitled to more deference than a regulation promulgated immediately after the passage of a bill, for at least in the situation we have here, we can have some assurance that GSA thought long and hard, based upon considerable experience and the benefits of extensive notice and comment, before it promulgated an administrative rule that has the binding force of law. The primary case cited in support of the Court’s view, see ante, at 464-465, n. 12, citing General Electric Co. v. Gilbert, 429 U. S. 125 (1976), is not at all pertinent. Although in Gilbert the Court mentioned the passage of time in its discussion of the regulations, it made nothing of this point on its own but instead refused to defer to the regulations because they flatly contradict[ed] the position which the agency had enun 480 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. ciated at an earlier date, closer to the enactment of the governing statute.” Id., at 142. Here, however, the GSA’s regulations are consistent with a memorandum prepared by the Office of Management and Budget and distributed to all Government agencies immediately after FAC A was enacted. See 38 Fed. Reg. 2307 (1973) (the “utilized by” language of FAC A would apply, for example, “to an already existing organization of scholars enlisted by an agency to provide advice on a continuing basis”).2 The fourth justification the Court offers for ignoring the agency’s interpretation is that the GSA lacks statutory authority to issue a binding regulatory interpretation of the term “advisory committee.” In Gilbert, for example, the agency which adopted the regulations at issue did not act pursuant to explicit statutory authority to promulgate regulations, and thus its regulations were at most of persuasive rather than controlling force. 429 U. S., at 141-142. But the Court errs in suggesting that the GSA’s regulations are mere nonbinding administrative guidelines. The GSA is conceded to be the agency “charged with the administration of [FACA],” Blum v. Bacon, 457 U. S. 132, 141 (1982); see ante, at 463, n. 12; it possesses statutory authority to implement the law by promulgating regulations and performing various other specific tasks that have binding effect on other Government agencies and all advisory committees, see FACA, 5 U. S. C. App. §§4(a), 7(a)-7(e), 10(a)(2), 10(a)(3) (1982 ed. and Supp. V); see also 40 U. S. C. § 486(c) (granting statutory authority for the GSA to promulgate regulations 2 Although the Court cites six cases to support the view that a noncontemporaneous agency interpretation of the governing statute is entitled to less deference from a reviewing court, five of the cases do not stand for that proposition, but only quote one another on the general issue. In fact, in those cases the Court did defer to agency regulations because they were promulgated pursuant to statutory authority, constituted reasonable interpretations and practical applications of the statutory language, and reflected a consistent agency position of long standing. See ante, at 464-465, n. 12 (citing cases). All those points are true in the cases before us. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 481 440 Kennedy, J., concurring in judgment necessary to implement the Federal Property and Administrative Services Act of 1949), and it issued its regulations pursuant to that authority, see 41 CFR §§101-6.1001 to 101-6.1035 (1988). In sum, it is quite desirable not to apply FACA to the ABA Committee. I cannot, however, reach this conclusion as a matter of fair statutory construction. The plain and ordinary meaning of the language passed by Congress governs, and its application does not lead to any absurd results. An unnecessary recourse to the legislative history only confirms this conclusion. And the reasonable and controlling interpretation of the statute adopted by the agency charged with its implementation is also in accord. The Court’s final step is to summon up the traditional principle that statutes should be construed to avoid constitutional questions. Although I agree that we should “first ascertain whether a construction of the statute is fairly possible by which the [constitutional] question may be avoided,” Crowell v. Benson, 285 U. S. 22, 62 (1932), this principle cannot be stretched beyond the point at which such a construction remains “fairly possible.” And it should not be given too broad a scope lest a whole new range of Government action be proscribed by interpretive shadows cast by constitutional provisions that might or might not invalidate it. The fact that a particular application of the clear terms of a statute might be unconstitutional does not provide us with a justification for ignoring the plain meaning of the statute. If that were permissible, then the power of judicial review of legislation could be made unnecessary, for whenever the application of a statute would have potential inconsistency with the Constitution, we could merely opine that the statute did not cover the conduct in question because it would be discomforting or even absurd to think that Congress intended to act in an unconstitutional manner. The utter circularity of this approach explains why it has never been our rule. 482 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. The Court’s ultimate interpretation of FACA is never clearly stated, except for the conclusion that the ABA Committee is not covered. It seems to read the “utilized by” portion of the statute as encompassing only a committee “established by a quasi-public organization in receipt of public funds,” ante, at 460, or encompassing “groups formed indirectly by quasi-public organizations such as the National Academy of Sciences,” ante, at 462. This is not a “fairly possible” construction of the statutory language even to a generous reader. I would find the ABA Committee to be covered by FACA. It is, therefore, necessary for me to reach and decide the constitutional issue presented. II Although I disagree with the Court’s conclusion that FACA does not cover the Justice Department’s use of the ABA Committee, I concur in the judgment of the Court because, in my view, the application of FACA in this context would be a plain violation of the Appointments Clause of the Constitution. The essential feature of the separation-of-powers issue in this suit, and the one that dictates the result, is that this application of the statute encroaches upon a power that the text of the Constitution commits in explicit terms to the President. Article II, §2, cl. 2, of the Constitution provides as follows: “[The President] shall nominate, and by and with the Advice and Consent of the Senate, shall appoint Ambassadors, other public Ministers and Consuls, Judges of he supreme Court, and all other Officers of the United States, whose Appointments are not herein otherwise provided for, and which shall be established by Law: but the Congress may by Law vest the Appointment of such inferior Officers, as they think proper, in the Presi- PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 483 440 Kennedy, J., concurring in judgment dent alone, in the Courts of Law, or in the Heads of Departments.”3 By its terms, the Clause divides the appointment power into two separate spheres: the President’s power to “nominate,” and the Senate’s power to give or withhold its “Advice and Consent.” No role whatsoever is given either to the Senate or to Congress as a whole in the process of choosing the person who will be nominated for appointment. As Hamilton emphasized: “In the act of nomination, [the President’s] judgment alone would be exercised; and as it would be his sole duty to point out the man who, with the approbation of the Senate, should fill an office, his responsibility would be as complete as if he were to make the final appointment.” The Federalist No. 76, 456-457 (C. Rossiter ed. 1961) (emphasis added). And again: “It will be the office of the President to nominate, and, with the advice and consent of the Senate, to appoint. There will, of course, be no exertion of choice on the part of the Senate. They may defeat one choice of the Executive, and oblige him to make another; but they cannot themselves choose—they can only ratify or reject the choice he may have made.” Id., No. 66, at 405 (emphasis in original).4 3 No issue has been raised in this suit with respect to the Congress’ power to vest the appointment of “inferior” officers in anyone other than the President. Cf. Morrison v. Olson, 487 U. S. 654, 673-677 (1988). 4 Hamilton also explained why it is that the President was given the sole prerogative of nominating principal officers: “The sole and undivided responsibility of one man will naturally beget a livelier sense of duty and a more exact regard to reputation. He will, on this account, feel himself under stronger obligations, and more interested to investigate with care the qualities requisite to the stations to be filled, and to prefer with impartiality the persons who may have the fairest pretensions to them.” The Federalist No. 76, at 455-456. 484 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. Indeed, the sole limitation on the President’s power to nominate these officials is found in the Incompatability Clause, which provides that “[n]o Senator or Representative shall, during the Time for which he was elected, be appointed to any civil Office under the Authority of the United States, which shall have been created, or the Emoluments whereof shall have been increased during such time.” U. S. Const., Art. I, §6, cl. 2, In some of our more recent cases involving the powers and prerogatives of the President, we have employed something of a balancing approach, asking whether the statute at issue prevents the President “‘from accomplishing [his] constitutionally assigned functions.’” Morrison v. Olson, 487 U. S. 654, 695 (1988), quoting Nixon v. Administrator of General Services, 433 U. S. 425, 443 (1977), and whether the extent of the intrusion on the President’s powers “is justified by an overriding need to promote objectives within the constitutional authority of Congress.” Ibid. In each of these cases, the power at issue was not explicitly assigned by the text of the Constitution to be within the sole province of the President, but rather was thought to be encompassed within the general grant to the President of the “executive Power.” U. S. Const., Art. II, § 1, cl. 1. Thus, for example, the relevant aspect of our decision in Morrison involved the President’s power to remove Executive officers, a power we had recognized is not conferred by any explicit provision in the text of the Constitution (as is the appointment power), but rather is inferred to be a necessary part of the grant of the “executive Power.” See Myers n. United States, 272 U. S. 52, 115-116 (1926). Similarly, in Administrator of General Services, supra, we were confronted with the question of the Executive Branch’s power to control the disposition of Presidential materials, a matter which, though vital to the President’s ability to perform his assigned functions, is not given to exclusive Presidential control by any explicit provision in the Constitution itself. We said there that “the proper in- PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 485 440 Kennedy, J., concurring in judgment quiry focuses on the extent to which [the congressional restriction] prevents the Executive Branch from accomplishing its constitutionally assigned functions,” and that we would invalidate the statute only if the potential for disruption of the President’s constitutional functions were present and if “that impact [were not] justified by an overriding need to promote objectives within the constitutional authority of Congress.” 433 U. S., at 443. See also United States v. Nixon, 418 U. S. 683, 703-707 (1974) (Executive privilege). In a line of cases of equal weight and authority, however, where the Constitution by explicit text commits the power at issue to the exclusive control of the President, we have refused to tolerate any intrusion by the Legislative Branch. For example, the Constitution confers upon the President the “Power to grant Reprieves and Pardons for Offenses against the United States, except in Cases of Impeachment.” U. S. Const., Art. II, §2, cl. 1. In United States v. Klein, 13 Wall. 128 (1872), the Court considered a federal statute that allowed citizens who had remained loyal to the Union during the Civil War to recover compensation for property abandoned to Union troops during the War. At issue was the validity of a provision in the statute that barred the admission of a Presidential pardon in such actions as proof of loyalty. Although this provision did not impose direct restrictions on the President’s power to pardon, the Court held that the Congress could not in any manner limit the full legal effect of the President’s power. As we said there: “[I]t is clear that the legislature cannot change the effect of ... a pardon any more than the executive can change a law.” Id., at 148. More than a century later, in Schick v. Reed, 419 U. S. 256 (1974), we reiterated in most direct terms the principle that Congress cannot interfere in any way with the President’s power to pardon. The pardon power “flows from the Constitution alone . . . and . . . cannot be modified, abridged, or diminished by the Congress.” Id., at 266. See also Ex parte Garland, 4 Wall. 333, 380 (1867). 486 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. INS v. Chadha, 462 U. S. 919 (1983), is another example of the Court’s refusal to apply a balancing test to assess the validity of an enactment which interferes with a power that the Constitution, in express terms, vests within the exclusive control of the President. In Chadha, the Court struck down a legislative veto provision in the Immigration and Nationality Act on the ground, inter alia, that it violated the explicit constitutional requirement that all legislation be presented to the President for his signature before becoming law. Id., at 946-948, 957-959. In so holding, the Court did not ask whether the “overriding need to promote objectives within the constitutional authority of Congress” justified this intrusion upon the Executive’s prerogative, but rather stated that the lawmaking process must adhere in strict fashion to the “[e]xplicit and unambiguous provisions of the Constitution [which] prescribe and define the respective functions of the Congress and of the Executive in the legislative process.” Id., at 945.5 The justification for our refusal to apply a balancing test in these cases, though not always made explicit, is clear enough. Where a power has been committed to a particular Branch of the Government in the text of the Constitution, the balance already has been struck by the Constitution itself. It is improper for this Court to arrogate to itself the power to adjust a balance settled by the explicit terms of the Constitution. To take an obvious example, it would be improper for us to hold that, although the Constitution sets 35 as the age below which one cannot be President, age 30 would in fact be a permissible construction of this term. See U. S. Const., Art. II, § 1. And it would be equally improper for us to determine that the level of importance at which a jury trial in a 5 Our decision in Chadha might also be read for the more general principle that where an enactment transgresses the explicit distribution of power in the text of the Constitution, then regardless of whether it implicates the Legislative, the Judicial, or the Executive power, a balancing inquiry is not appropriate. I need not address the broader principle in this case. PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 487 440 Kennedy, J., concurring in judgment common-law suit becomes available is $1,000 instead of $20, as the Constitution provides. See U. S. Const., Arndt. 7. These minor adjustments might be seen as desirable attempts to modernize the original constitutional provisions, but where the Constitution draws a clear line, we may not engage in such tinkering. However improper would be these slight adjustments to the explicit and unambiguous balances that are struck in various provisions of the Constitution, all the more improper would it be for this Court, which is, after all, one of the three coequal Branches of the Federal Government, to rewrite the particular balance of power that the Constitution specifies among the Executive, Legislative, and Judicial Departments. This is not to say that each of the three Branches must be entirely separate and distinct, for that is not the governmental structure of checks and balances established by the Framers. See Mistretta v. United States, 488 U. S. 361, 380-381 (1989); Humphrey’s Executor v. United States, 295 U. S. 602, 629 (1935). But as to the particular divisions of power that the Constitution does in fact draw, we are without authority to alter them, and indeed we are empowered to act in particular cases to prevent any other Branch from undertaking to alter them. These considerations are decisive of the suit before us. The President’s power to nominate principal officers falls within the line of cases in which a balancing approach is inapplicable. The Appointments Clause sets out the respective powers of the Executive and Legislative Branches with admirable clarity. The President has the sole responsibility for nominating these officials, and the Senate has the sole responsibility of consenting to the President’s choice. See supra, at 483. We have, in effect, already recognized as much in Buckley n. Valeo, 424 U. S. 1 (1976). In Buckley, the Court held that the appointment of Federal Election Commissioners through procedures that were inconsistent with those set forth in the Appointments Clause was uncon 488 OCTOBER TERM, 1988 Kennedy, J., concurring in judgment 491 U. S. stitutional. In doing so, it rejected outright the arguments advanced by the Federal Election Commission and various amici that because the Constitution gave Congress “explicit and plenary authority to regulate [the] field of activity” at issue (federal elections), and because Congress “had good reason[s] for not [creating] a commission composed wholly of Presidential appointees,” that Congress could allow these officials to be appointed to their positions without complying with the strict letter of the Appointments Clause. As we stated there: “While one cannot dispute the basis for [Congress’ concern that an election commission exist not in whole of presidential appointees] as a practical matter, it would seem that those who sought to challenge incumbent Congressmen might have equally good reason to fear a Commission which was unduly responsive to members of Congress whom they were seeking to unseat. But such fears, however rational, do not by themselves warrant a distortion of the Framers’ work.” Id., at 134 (emphasis added). It is also plain that the application of FACA would constitute a direct and real interference with the President’s exclusive responsibility to nominate federal judges. The District Court found, “at minimum, that the application of FACA to the ABA Committee would potentially inhibit the President’s freedom to investigate, to be informed, to evaluate, and to consult during the nomination process,” and that these consequences create an “obvious and significant potential for ‘disruption’ of the President’s constitutional prerogative during the nomination process,” 691 F. Supp. 483, 493 (DC 1988), and these findings are not contested here. As we said in the context of the pardon power, “[t]he simplest statement is the best.” United States v. Klein, 13 Wall., at 148. The mere fact that FACA would regulate so as to interfere with the manner in which the President obtains information necessary to discharge his duty assigned under the Constitution to PUBLIC CITIZEN v. DEPARTMENT OF JUSTICE 489 440 Kennedy, J., concurring in judgment nominate federal judges is enough to invalidate the Act. “We think it unnecessary to enlarge.” Ibid. For these reasons, I concur in the judgment affirming the District Court. 490 OCTOBER TERM, 1988 Syllabus 491 U. S. PITTSBURGH & LAKE ERIE RAILROAD CO. v. RAILWAY LABOR EXECUTIVES’ ASSOCIATION CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT No. 87-1589. Argued March 29, 1989—Decided June 21, 1989* Faced with heavy losses, petitioner Pittsburgh and Lake Erie Railroad Co. (P&LE) agreed to sell its assets to P&LE Rail Co., Inc. (Railco), a newly formed subsidiary of another railroad. Railco intended to operate the railroad as P&LE had except that it would not assume P&LE’s various collective-bargaining contracts and would need only 250 of the 750 employees then working for P&LE. P&LE rejected the assertion of the unions representing its employees that the proposed sale could not be implemented until P&LE complied with the Railway Labor Act (RLA) provisions prohibiting carriers from changing rates of pay, rules, or working conditions embodied in agreements, 45 U. S. C. § 152 Seventh, unless they first give at least 30 days’ written notice of, and proceed to bargain over, the intended agreement change, § 156. Section 156 also requires that the working conditions in question remain in status quo until the controversy is resolved. Most of the unions then filed § 156 notices proposing extensive changes in existing agreements to ameliorate the proposed sale’s adverse impact on employees. P&LE again refused to bargain, asserting that the sale was within the exclusive jurisdiction of the Interstate Commerce Commission (ICC) under the Interstate Commerce Act (ICA), which requires that noncarriers such as Railco seeking to acquire a rail line first obtain the ICC’s approval, 49 U. S. C. § 10901. Respondent Railway Labor Executives’ Association (RLE A) filed suit on behalf of the unions in the District Court, seeking a declaratory judgment with respect to P&LE’s RLA obligations and an injunction against the sale pending completion of RLA bargaining. The unions then went on strike, and the District Court denied P&LE’s request for a restraining order on the ground that the Norris-LaGuardia Act (NLGA) prohibited it from enjoining a work stoppage growing out of a labor dispute. Subsequently, Railco filed an application for exemption from § 10901’s requirements pursuant to the ICC’s Ex Parte No. 392 Class Exemption (Ex Parte 392), which provides abbreviated proce- *Together with No. 87-1888, Pittsburgh & Lake Erie Railroad Co. v. Railway Labor Executives’ Association et al., also on certiorari to the same court. PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 491 490 Syllabus dures for seeking approval for acquisitions by noncarriers of an operating railroad or its assets, and the ICC allowed the exemption to become effective on the date required by Ex Parte 392. Although § 10901 and Ex Parte 392 allow the ICC to require the acquiring company to provide protection for railroad employees affected by a sale, none of the unions requested protective provisions. The District Court then granted P&LE’s reapplication for an order restraining the strike, ruling that the ICC’s approval of the sale negated any duty that P&LE had to bargain over the sale’s effects on its employees, and that the NLGA did not forbid issuance of an injunction under such circumstances. However, the Court of Appeals summarily reversed, holding that the ICA did not require accommodation of the NLGA’s restrictions on the District Court’s powers, and remanding for a determination whether the sale or strike violated the RLA. The District Court then held that although P&LE did not have a duty to bargain over its decision to sell, the RLA required it to bargain over the sale’s effects on its employees, and that § 156’s status quo provision required that P&LE’s RLA bargaining obligations be satisfied before the sale could be consummated despite the ICC’s approval of the transaction. The court therefore granted RLE A’s request for an injunction against the sale, and the Court of Appeals affirmed. This Court then granted P&LE’s petitions for certiorari challenging, in No. 87-1888, the Court of Appeals’ affirmance of the injunction against the sale and seeking, in No. 87-1589, reversal of the Court of Appeals’ judgment setting aside the strike injunction. Held: 1. The RLA did not require or authorize an injunction against the sale of P&LE’s assets to Railco. Pp. 502-512. (a) The RLA did not require P&LE itself to give notice of its decision to sell and thereafter to bargain about the effects of the sale. Section 156 requires those actions only when the carrier is proposing a “change in agreements.” In holding that the loss of jobs resulting from the proposed sale clearly would require such a “change,” the Court of Appeals did not point out how the sale would alter any specific provision of any agreement; did not suggest that any such agreement dealt with the possibility of a sale, sought to confer any rights on employees in the event of a sale, or guaranteed that jobs would be available indefinitely; and did not explain how P&LE’s decision to remove itself from the railroad business and terminate its position as a railroad employer would violate or require changing any agreement. Nor did the court purport to find an implied agreement that P&LE would not go out of business, would not sell its assets, or would protect its employees from the adverse consequences of a sale. Pp. 503-504. 492 OCTOBER TERM, 1988 Syllabus 491 U. S. (b) Although the unions’ § 156 notices may have proposed far-reaching changes in the existing agreements, the giving of those notices did not obligate P&LE to maintain the status quo and to postpone its sale to Railco beyond the time the sale was approved by the ICC and was scheduled to be consummated. Detroit & Toledo Shore Line Railroad Co. v. Transportation Union, 396 U. S. 142—which held that, when a rail union files a § 156 notice, the status quo provision forbids the railroad to make any change in pre-existing “objective working conditions” even if those conditions are not contained in express or implied agreements — does not control here, primarily because that case did not involve a carrier’s decision to quit the railroad business, sell its assets, and cease to be a railroad employer. Textile Workers v. Darlington Mfg. Co., 380 U. S. 263, established that the decision to close down a business entirely is so much a management prerogative that only an unmistakable expression of congressional intent will require a ruling to the contrary. The RLA contains no expression of intent to render an employer’s decision to go out of business and consequently to reduce to zero the number of available jobs a change in the conditions of employment forbidden by § 156’s status quo provision. Where, as here, the collective-bargaining agreement is silent concerning the sale of a railroad’s assets and the railroad has proceeded in accordance with the ICA to obtain the ICC’s approval of the sale, a union cannot delay the immediate consummation of the sale by filing a § 156 notice. This construction of the RLA satisfies the Court’s obligation to avoid conflicts between overlapping statutory regimes, since it maintains the ICC’s plenary authority over rail acquisitions by noncarriers. In contrast, enjoining the sale’s consummation based on the unions’ § 156 notices would likely result in the sale’s cancellation and the frustration of Congress’ intent through ICA amendments to deregulate the rail industry and to assist small rail lines with financial problems. Pp. 504-511. (c) However, P&LE did have a limited duty to bargain in response to the unions’ § 156 notices. Its decision to sell, as such, was not a bargainable subject. Nevertheless, to the extent that it could satisfy the unions’ proposals for changes, it was required to bargain about the effects that the sale would or might have upon its employees. That obligation ceased on the date for closing the sale after the Ex Parte 392 exemption became effective. P. 512. 2. The record is insufficient to allow this Court to determine whether the Court of Appeals correctly set aside the injunction against the strike. Although the NLGA’s general limitation on district courts’ power to issue injunctions in labor disputes must give way when necessary to enforce a duty specifically imposed by another statute, the Court of Appeals correctly ruled that the ICA creates no such duty. While § 10901 PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 493 490 Syllabus does authorize the ICC to impose labor protective provisions and gives unions the right to seek such protection and the right to judicial review if dissatisfied, no applicable IC A provision requires unions to participate in ICC proceedings or to seek protection rather than striking. Furthermore, labor protection provisions run against the acquiring railroad rather than the seller, and nothing in the ICA relieved P&LE of its duty to bargain with the unions until its transaction was completed or empowered the ICC to intrude into the relationship between P&LE and the unions. However, the NLGA’s limitation on injunctions must also be accommodated to the specific provisions of the RLA, such that district courts have jurisdiction and power to issue injunctive orders against strikes violative of those provisions. Since neither of the lower courts ever ruled on whether the RLA creates a duty not to strike while its dispute resolution mechanisms are underway, the matter must be resolved on remand. Pp. 513-515. No. 87-1589, 831 F. 2d 1231, vacated and remanded; No. 87-1888, 845 F. 2d 420, reversed and remanded. White, J., delivered the opinion for a unanimous Court with respect to Parts I and III, and the opinion of the Court with respect to Parts II and IV, in which Rehnquist, C. J., and O’Connor, Scalia, and Kennedy, JJ., joined. Stevens, J., filed an opinion concurring in part and dissenting in part, in which Brennan, Marshall, and Blackmun, JJ., joined, post, p. 515. Richard L. Wyatt, Jr., argued the cause for petitioner. With him on the briefs were Ronald M. Johnson, Charles L. Warren, Eric D. Witkin, and G. Edward Yurcon. Jeffrey P. Minear argued the cause for the United States as amicus curiae urging affirmance in both cases. With him on the brief were Solicitor General Fried and Deputy Solicitor General Merrill. John O’B. Clarke, Jr., argued the cause for respondents. With him on the brief for respondent Railway Labor Executives’ Association was William G. Mahoney. Robert S. Burk, Henri F. Rush, and John J. McCarthy, Jr., filed a brief for the Interstate Commerce Commission, respondent in No. 87-1888. t tBriefs of amici curiae urging reversal in both cases were filed for the State of South Dakota by Richard A. Allen and Julie A. Tigges; and for the National Railway Labor Conference by Richard T. Conway, Ralph J. 494 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Justice White delivered the opinion of the Court. These cases involve the interaction of three federal statutes with respect to the proposed sale of the rail line of the Pittsburgh and Lake Erie Railroad Co. (P&LE). The statutes are the Railway Labor Act (RLA), 44 Stat. 577, as amended, 45 U. S. C. § 151 et seq.; the Interstate Commerce Act (ICA), 49 U. S. C. §10101 et seq. (1982 ed. and Supp. V); and the Norris-LaGuardia Act (NLGA), 47 Stat. 70, 29 U. S. C. § 101 et seq. I Petitioner, P&LE, is a small rail carrier owning and operating 182 miles of rail line serving points in Ohio and western Pennsylvania and possessing trackage rights over other lines extending into New York. P&LE has experienced financial problems of increasing severity, having lost $60 million during the five years preceding the onset of these cases. After other efforts to improve its condition failed, notably work force reductions, concessions from its employees, and market expansion, P&LE decided that in order to recoup for its owners any part of their investments it must sell its assets.1 On July 8, 1987, P&LE agreed to sell its assets for Moore, Jr., D. Eugenia Langan, and David P. Lee. Briefs of amici curiae urging reversal in No. 87-1888 were filed for the Airline Industrial Relations Conference by Harry A. Rissetto and Thomas E. Reinert, Jr.; for Chicago & North Western Transportation Co. et al. by Ralph J. Moore, Jr., D. Eugenia Langan, James P. Daley, Stuart F. Gassner, and Robert J. Corber; and for Guilford Transportation Industries, Inc., et al. by Ralph J. Moore, Jr., and D. Eugenia Langan. David Silberman and Laurence Gold filed a brief for the American Federation of Labor and Congress of Industrial Organizations as amicus curiae urging affirmance in No. 87-1589. Mark M. Levin filed a brief for the Regional Railroads of America et al. as amici curiae in both cases. 1 Attempts to interest major rail lines in the property were unavailing because of the high cost of labor protection that would have been mandatory under the section of the ICA applicable to purchases by an existing carrier. 49 U. S. C. § 11347 (1982 ed., Supp. V), which is set forth in n. 7, infra. PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 495 490 Opinion of the Court approximately $70 million to a newly formed subsidiary, P&LE Rail Co., Inc. (Railco), of Chicago West Pullman Transportation Corporation (CWP).2 Railco intended to operate the railroad as P&LE had except that Railco would not assume P&LE’s collective-bargaining contracts with its various unions and would need only about 250 employees rather than the 750 then working for P&LE.3 When the unions representing P&LE’s employees were notified of the proposed sale, they asserted that the sale would have an effect on the working conditions of the carrier’s employees and therefore was subject to the requirements of the RLA, 45 U. S. C. §§ 152 Seventh and 156, which provide: “§ 152 . . . Seventh. Change in pay, rules, or working conditions contrary to agreement or to section 156 forbidden “No carrier, its officers, or agents shall change the rates of pay, rules, or working conditions of its employees, as a class, as embodied in agreements except in the manner prescribed in such agreements or in section 156 of this title.” “§ 156. Procedure in changing rates of pay, rules, and working conditions “Carriers and representatives of the employees shall give at least thirty days’ written notice of an intended change in agreements affecting rates of pay, rules, or working conditions, and the time and place for the beginning of conference between the representatives of the parties interested in such intended changes shall be agreed upon within ten days after the receipt of said notice, and said time shall be within the thirty days provided in the notice. In every case where such notice of intended change has been given, or conferences are 2 P&LE would keep certain real estate and some 6,000 railcars. 3 CWP anticipated inviting all P&LE employees to submit applications and intended to give preference to them in hiring. CWP also expected to bargain for new contracts with the existing unions. 496 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. being held with reference thereto, or the services of the Mediation Board have been requested by either party, or said Board has proffered its services, rates of pay, rules, or working conditions shall not be altered by the carrier until the controversy has been finally acted upon, as required by section 155 of this title, by the Mediation Board, unless a period of ten days has elapsed after termination of conferences without request for or proffer of the services of the Mediation Board.”4 The unions advised that they stood ready to negotiate all aspects of the matter, including the decision to sell the railroad assets. P&LE responded that it was willing to discuss the matter but that § 156 notice and bargaining were not required since the transaction was subject to the jurisdiction of the Interstate Commerce Commission (ICC or Commission) 4 Disputes about proposals to change rates of pay, rules, or working conditions are known as major disputes. Minor disputes are those involving the interpretation or application of existing contracts. The latter are subject to compulsory arbitration. The former are subject to the procedures set out in §§ 156 and 155, which specify the functions of the Mediation Board. In Trainmen v. Jacksonville Terminal Co., 394 U. S. 369, 378 (1969), we described the procedures applicable to major disputes: “The Act provides a detailed framework to facilitate the voluntary settlement of major disputes. A party desiring to effect a change of rates, pay, rules, or working conditions must give advance written notice. § 6. The parties must confer, § 2 Second, and if conference fails to resolve the dispute, either or both may invoke the services of the National Mediation Board, which may also proffer its services sua sponte if it finds a labor emergency to exist. § 5 First. If mediation fails, the Board must endeavor to induce the parties to submit the controversy to binding arbitration, which can take place, however, only if both consent. §§ 5 First, 7. If arbitration is rejected and the dispute threatens ‘substantially to interrupt interstate commerce to a degree such as to deprive any section of the country of essential transportation service, the Mediation Board shall notify the President,’ who may create an emergency board to investigate and report on the dispute. § 10. While the dispute is working its way through these stages, neither party may unilaterally alter the status quo. §§ 2 Seventh, 5 First, 6, 10.” PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 497 490 Opinion of the Court under the ICA and since the requirements of §§ 155 and 156 would intrude on that regime as well as upon management’s prerogatives to conduct the affairs of the company with respect to the sales transaction. Most of the unions then responded by themselves filing §156 notices proposing changes in existing agreements to ameliorate the adverse impacts of the proposed sale upon P&LE’s employees. The unions sought guarantees that the sale would not cause any employee to be deprived of employment or to be placed in any worse position with respect to pay or working conditions and that P&LE would require that the purchaser of its rail line assume P&LE’s collectivebargaining agreements.5 P&LE again declined to bargain, asserting that the transaction was within the exclusive jurisdiction of the ICC. On August 19, respondent, Railway Labor Executives’ Association (RLE A), on behalf of P&LE’s unions, filed suit in the United States District Court for the Western District of Pennsylvania, seeking a declaratory judgment with respect to P&LE’s obligations under the RLA °The unions’ proposals were essentially these: “1. No employee of the P&LE Railroad Company who [was actively employed or on authorized leave of absence] between August 1, 1986 and August 1, 1987 . . . shall be deprived of employment or placed in a worse position with respect to compensation or working conditions for any reason except resignation, retirement, death or dismissal for justifiable cause. . . . The formulae for the protective allowances, with a separation option, shall be comparable to those established in the New York Dock conditions. “2. If an employee is placed in a worse position with respect to compensation or working conditions, that employee shall receive, in addition to a make-whole-remedy, penalty pay equal to three times the lost pay, fringe benefits and consequential damages suffered by such employee. ‘3. P&LE agrees to obtain binding commitments from any purchaser of its vail line operating properties and assets to assume all [of P&LE’s] collective bargaining agreements ... to hire P&LE employees in seniority order without physicals, and to negotiate with the P&LE and this Organization an agreement to apply this Agreement to the sale transaction and to select the forces to perform the work over the lines being acquired.” App. 38, 42, 46, 50, 54, 58, 62, 66, 122, 126. 498 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. and an injunction against the sale pending completion of RLA bargaining obligations. On September 15, 1987, the unions went on strike. P&LE’s request for a restraining order against the strike was denied by the District Court on the ground that the NLGA forbade such an order.6 The proposed sale of assets could not be carried out without compliance with the terms of the ICA, 49 U. S. C. § 10901, which requires that noncarriers seeking to acquire a rail line first obtain a certificate of public convenience and necessity from the ICC. Section 10901(e) specifies the procedures for this purpose and provides that the ICC “may” require the acquiring company “to provide a fair and equitable arrangement for the protection of railroad employees who may be affected thereby no less protective of and beneficial to the interests of such employees than those established pursuant to section 11347 of this title.”7 Section 10505, however, 6 Section 4 of the NLGA, as set forth in 29 U. S. C. § 104, provides in part: “§ 104. Enumeration of specific acts not subject to restraining orders or injunctions “No court of the United States shall have jurisdiction to issue any restraining order or temporary or permanent injunction in any case involving or growing out of any labor dispute to prohibit any person or persons participating or interested in such dispute (as these terms are herein defined) from doing, whether singly or in concert, any of the following acts: “(a) Ceasing or refusing to perform any work or to remain in any relation of employment . . . .” Section 8 of that Act, as set forth in 29 U. S. C. § 108, is also relevant here: “§ 108. Noncompliance with obligations involved in labor disputes or failure to settle by negotiation or arbitration as preventing injunctive relief “No restraining order or injunctive relief shall be granted to any complainant who has failed to comply with any obligation imposed by law which is involved in the labor dispute in question, or who has failed to make every reasonable effort to settle such dispute either by negotiation or with the aid of any available governmental machinery of mediation or voluntary arbitration.” 7 Title 49 U. S. C. § 11347 (1982 ed., Supp. V) requires labor protective provisions when a rail carrier is involved in certain transactions such as mergers or consolidations: PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 499 490 Opinion of the Court authorizes the Commission to grant exemptions from the requirements of the Act when not necessary to carry out the national transportation policy.8 Based on its experience with acquisitions under § 10901, the ICC had issued what is known as the Ex Parte No. 392 Class Exemption, see Ex Parte No. 392 (Sub. No. 1), Class Exemption for the Acquisition and Operation of Rail Lines Under 49 U. S. C. 10901, “§11347. Employee protective arrangements in transactions involving rail carriers “When a rail carrier is involved in a transaction for which approval is sought under sections 11344 and 11345 or section 11346 of this title, the Interstate Commerce Commission shall require the carrier to provide a fair arrangement at least as protective of the interests of employees who are affected by the transaction as the terms imposed under this section before February 5, 1976, and the terms established under section 405 of the Rail Passenger Service Act (45 U. S. C. 565). Notwithstanding this subtitle, the arrangement may be made by the rail carrier and the authorized representative of its employees. The arrangement and the order approving the transaction must require that the employees of the affected rail carrier will not be in a worse position related to their employment as a result of the transaction during the 4 years following the effective date of the final action of the Commission (or if an employee was employed for a lesser period of time by the carrier before the action became effective, for that lesser period).” 8 Section 10505 provides in part: “§ 10505. Authority to exempt rail carrier transportation “(a) In a matter related to a rail carrier providing transportation subject to the jurisdiction of the Interstate Commerce Commission under this subchapter, the Commission shall exempt a person, class of persons, or a transaction or service when the Commission finds that the application of a provision of this subtitle— (1) is not necessary to carry out the transportation policy of section 10101a of this title; and (2) either (A) the transaction or service is of limited scope, or (B) the application of a provision of this subtitle is not needed to protect shippers from the abuse of market power. “(g) The Commission may not exercise its authority under this section • • • (2) to relieve a carrier of its obligation to protect the interests of employees as required by this subtitle.” 500 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. 1 I. C. C. 2d 810 (1985) (Ex Parte 392), review denied sub nom. Illinois Commerce Comm’n v. ICC, 260 U. S. App. D. C. 38, 817 F. 2d 145 (1987),9 which provides abbreviated procedures for seeking approval for acquisitions by noncarriers such as Railco of an operating railroad or its assets. The regulatory procedure, see 49 CFR § 1150.32(b) (1987), involved the filing of an application for exemption which would become effective seven days after filing absent contrary notice from the Commission.10 An interested party could op- 9 The Commission’s brief in this Court provides this background: “In the years just after the partial deregulation of the railroad industry occasioned by the passage of the Staggers Rail Act of 1980, Pub. L. No. 96-448, 94 Stat. 194-45, numerous new short lines and regional rail lines were created, pursuant to 49 U. S. C. 10901, through the sale of marginally profitable and unprofitable rail lines to new entities eager to provide rail service. In considering and approving these sales, the Commission became convinced that the expense imposed on such sales by the imposition of labor protective conditions was hampering the development of short line railroads and, indeed, was forcing the selling carriers to abandon these marginal lines pursuant to 49 U. S. C. 10903 of the ICA. “In order to foster the development of short line railroads to preserve rail facilities, service and employment that would otherwise be lost through abandonments, the Commission began withholding labor protections in individual sales. After considering over five years many such applications, the Commission determined that the formation of new rail carriers should be encouraged. In order to aid rail formations, the Commission promulgated the procedures in Ex Parte No. 392. In Ex Parte 392 the Commission exempted rail line sales to new carriers from full compliance with Commission procedures while retaining authority, under its revocation power, to review the transaction and correct any problem arising out of the transaction.” Brief for Interstate Commerce Commission 3-4 (footnote and citations omitted). In the Ex Parte 392 proceedings, the RLEA demanded that the Commission impose labor conditions in all § 10901 sale transactions. The Commission, however, ruled that labor protective provisions would be imposed in individual cases only upon a showing of exceptional circumstances. 1 I. C. C. 2d, at 815. 10 The ICC modified the Ex Parte 392 procedure in 1988 to extend the waiting period from 7 to 35 days. See 53 Fed. Reg. 5981-5982 (1988). PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 501 490 Opinion of the Court pose the exemption by filing a petition to revoke at any time, after consideration of which the ICC could revoke the exemption in whole or in part or impose labor protective provisions. The ICC had indicated, however, that only in exceptional situations would such protective provisions be imposed. Accordingly, Railco on September 19, 1987, filed a notice of exemption pursuant to Ex Parte 392. After denying various requests by the unions to reject the notice of exemption and stay the sale, the Commission allowed the exemption to become effective on September 26. A petition to revoke filed by RLEA on October 2 is still pending before the Commission. At no time did RLEA request imposition of labor protective provisions pursuant to the Commission’s authority under § 10901.11 On October 5, 1987, P&LE reapplied to the District Court for an order restraining the strike. The District Court granted the request on October 8, ruling that the authorization of the sale by the ICC negated any duty that P&LE had to bargain over the effects of the sale on its employees, and that the NLGA did not forbid issuance of an injunction under such circumstances.12 On October 26, however, the Court of Appeals summarily reversed, holding that the ICA did not require accommodation of the NLGA’s restrictions on the District Court’s powers. 831 F. 2d 1231 (CA3 1987). A remand was ordered to determine whether the sale or strike violated the RLA. The unions did not resume their strike when the Court of Appeals reversed the District Court’s injunction, but threatened to do so if P&LE attempted to consummate the sale to Railco.13 11 See n. 7, supra. 12 The order was to remain in effect until the District Court ruled on the preliminary injunction sought by P&LE. It was this order that was reviewed by the Court of Appeals. 13 The strike and the decisions of the Court of Appeals effectively terminated the proposed sale to Railco. Efforts to find another buyer were unsuccessful, but since P&LE is still interested in selling its assets and the 502 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. The case in the District Court then went forward. Addressing the unions’ request for an injunction, the District Court held that although P&LE did not have a duty to bargain over its decision to sell, P&LE was required by the RLA to bargain over the effects of the sale on employees, and that the status quo provision of § 156 required that its bargaining obligations under the RLA must be satisfied before the sale could be consummated despite approval of the transaction by the ICC acting pursuant to the ICA. 677 F. Supp. 830 (WD Pa. 1987). A divided Court of Appeals affirmed the judgment of the District Court. 845 F. 2d 420 (CA3 1988). We granted P&LE’s petition in No. 87-1888, challenging the Court of Appeals’ affirmance of the injunction against the sale issued by the District Court, as well as P&LE’s petition in No. 87-1589, asking for reversal of the judgment of the Court of Appeals setting aside the strike injunction issued by the District Court. 488 U. S. 965 (1988). II In No. 87-1888, the issue is whether the RLA, properly construed, required or authorized an injunction against closing the sale of P&LE’s assets to Railco because of an unsatisfied duty to bargain about the effects of the sale on P&LE’s employees. We first address whether the RLA required P&LE to give notice of its decision to sell and to bargain about the effects of the sale. We then consider whether the unions’ own notices and the status quo provision of § 156 justified the injunction. issues in these cases have a bearing on those efforts, the cases, as the Court of Appeals recognized and the parties agree, are not moot. Also, in late September, P&LE and its unions had informal exchanges about the effects of the sale. On October 14, one of the unions invoked the services of the Mediation Board. After the April 8, 1988, Court of Appeals decision, 845 F. 2d 420 (CA3), effects bargaining proceeded, and as these cases indicate, the parties have not resolved their differences. PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 503 490 Opinion of the Court A P&LE submits that neither its decision to sell nor the impact that sale of the company might have had on its employees was a “change in agreements affecting rates of pay, rules, or working conditions” (emphasis added) within the meaning of the RLA, 45 U. S. C. § 156, and that P&LE therefore had no duty to give notice or to bargain with respect to these matters. The Court of Appeals rejected this submission, focusing on the effects the sale would have on employees and concluding that the “loss of jobs by possibly two-thirds of the employees clearly would require a ‘change in agreements affecting rates of pay, rules, or working conditions.’” 845 F. 2d, at 428. The court did not point out how the proposed sale would require changing any specific provision of any of P&LE’s collective-bargaining agreements. It did not suggest that any of those agreements dealt with the possibility of the sale of the company, sought to confer any rights on P&LE’s employees in the event of the sale, or guaranteed that jobs would continue to be available indefinitely.14 What P&LE proposed to do would remove it from the railroad business and terminate its position as a railroad employer; and like the Court of Appeals, RLEA does not explain how such action would violate or require changing any of the provisions of the unions’ written contracts with P&LE. Of course, not all working conditions to which parties may have agreed are to be found in written contracts. Detroit & Toledo Shore Line R. Co. v. Transportation Union, 396 U. S. 142, 154-155 (1969) (Shore Line). It may be that 14 Indeed, the Court of Appeals stated that “P&LE’s agreements with its unions, however, do not appear to contemplate this type of transaction & e., sale of the rail lines], and thus neither expressly permit nor prohibit the sale.” 845 F. 2d, at 428, n. 9. RLEA asserts that P&LE had granted job security guarantees to some of its employees, see Brief for Respondent RLEA 3, but the record does not contain the collective-bargaining contracts, and if there were such guarantees, there is no claim that they would survive the sale of the rail line. 504 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. “in the context of the relationship between the principals, taken as a whole, there is a basis for implying an understanding on the particular practice involved.” Id., at 160 (Harlan, J., dissenting). But the Court of Appeals did not purport to find an implied agreement that P&LE would not go out of business, would not sell its assets, or if it did, would protect its employees from the adverse consequences of such action. Neither does RLEA. We therefore see no basis for holding that P&LE should have given a § 156 notice of a proposed “change” in its express or implied agreements with the unions when it contracted to sell its assets to Railco. Nor was it, based on its own decision to sell, obligated to bargain about the impending sale or to delay its implementation. We find RLEA’s arguments to the contrary quite unconvincing. B There is more substance to the Court of Appeals’ holding, and to RLEA’s submission, that the unions’ § 156 notices proposed far-reaching changes in the existing agreements over which P&LE was required to bargain and that the status quo provision of § 156 prohibited P&LE from going forward with the sale pending completion of the “purposely long and drawn out” procedures which the Act requires to be followed in order to settle a “major” dispute. Railway Clerks v. Florida East Coast R. Co., 384 U. S. 238, 246 (1966). Section 156 provides that when a notice of change in agreements has been given, “rates of pay, rules, or working conditions shall not be altered by the carrier until the controversy has been finally acted upon, as required by section 155.” Relying on Shore Line, RLEA argues, and the Court of Appeals held, that when a rail labor union files a § 156 notice to change the terms of an agreement, the “working conditions” that the carrier may not change pending conclusion of the bargaining process are not limited to those contained in express or implied agreements but include, as Shore Line held, “those actual, objective working conditions and practices, broadly con- PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 505 490 Opinion of the Court ceived, which were in effect prior to the time the pending dispute arose and which are involved in or related to that dispute.” 396 U. S., at 153. RLE A submits that the relationship of employer-employee and the state of being employed are among those working conditions that may not be changed until the RLA procedures are satisfied. We are unconvinced, for several reasons, that this is the case. The facts of Shore Line, briefly stated, were these: Shore Line operated 50 miles of rail line between Lang Yard in Toledo, Ohio, and Dearoad Yard near Detroit, Michigan. For many years, all train and engine crews reported for duty and finished the day at Lang Yard. When it was necessary to perform switching and other operations at other points, crews were transported at railroad expense to those outlying points. The company proposed to establish outlying work assignments at Trenton, Michigan, some 35 miles north of Lang Yard. Crews assigned there would have to report there. The proposed change was not forbidden by, and would not have violated, the parties’ collective-bargaining agreement. The union filed a § 156 notice seeking to amend the agreement to forbid the railroad to make outlying assignments. The issue was not settled by the parties and the union called for mediation. While the Mediation Board proceedings were pending, the railroad posted a bulletin creating the disputed assignment at Trenton. The union threatened a strike, the company sued to restrain the strike, and the union counterclaimed for an injunction relying on the status quo provision of § 156. The District Court and the Court °f Appeals held for the union, and we affirmed over a dissent by Justice Harlan, joined by Chief Justice Burger. We held that even though Shore Line did not propose to change any of its agreements, the status quo provision of § 156—“rates of pay, rules, or working conditions shall not be altered” pending exhaustion of the required procedure—forbade any change by Shore Line in the “objective working conditions” then existing. 396 U. S., at 153. We noted that had it been 506 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the practice to make outlying work assignments, the company would have been within its rights to make the Trenton assignment; but the prior practice, the objective working condition, was to have crews report for work and come back to Lang Yard. That working condition could not be changed pending resolution of the dispute without violating the status quo provision of § 156 even though there was nothing in the agreement between the parties to prevent outlying assignments. Id., at 153-154. Shore Line, in our view, does not control these cases. In the first place, our conclusion in that case that the status quo provision required adherence not only to working conditions contained in express or implied agreements between the railroad and its union but also to conditions “objectively” in existence when the union’s notice was served, and that otherwise could be changed without violating any agreement, extended the relevant language of § 156 to its outer limits, and we should proceed with care before applying that decision to the facts of these cases.15 Second, reporting at Lang Yard, we thought, had been the unquestioned practice for many years, and we considered it reasonable for employees to deem it sufficiently established that it would not be changed without bargaining and compliance with the status quo provisions of the RLA. 15 Section 156 deals with bargaining and settlement procedures with respect to changes in agreements affecting rates of pay, rules, or working conditions. There must be notice of such intended changes, as well as bargaining and mediation if requested or proffered. And in every case involving such notice, i. e., of intended changes in agreements, rates of pay, rules or working conditions shall not be changed by the carrier until the specified procedures are satisfied. Because § 156 concerns changes in agreements, it is surely arguable that it is open to a construction that would not require the status quo with respect to working conditions that have never been the subject of an agreement, expressed or implied, and that, if no notice of changes had been served by the union, could be changed by the carrier without any bargaining whatsoever. Shore Line rejected that construction, but as indicated in the text, we are not inclined to apply Shore Line to the decision of P&LE to go out of business. PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 507 490 Opinion of the Court Third, and more fundamentally, the decision did not involve a proposal by the railroad to terminate its business. Here, it may be said that the working condition existing prior to the § 156 notice was that P&LE was operating a railroad through the agency of its employees, but there was no reason to expect, simply from the railroad’s long existence, that it would stay in business, especially in view of its losses, or that rail labor would have a substantial role in the decision to sell or in negotiating the terms of the sale. Whatever else Shore Line might reach, it did not involve the decision of a carrier to quit the railroad business, sell its assets, and cease to be a railroad employer at all, a decision that we think should have been accorded more legal significance than it received in the courts below. Our cases indicate as much. In Textile Workers v. Darlington Mfg. Co., 380 U. S. 263 (1965), an employer closed its textile mill when a union won a representation election. The National Labor Relations Board concluded that this action was an unfair labor practice under §§ 8(a)(1) and (3) of the National Labor Relations Act (NLRA). The Court of Appeals disagreed, holding that the complete or partial liquidation of an employer’s business even though motivated by antiunion animus was not an unfair practice. We affirmed in part,16 ruling that insofar as the NLRA is concerned, an employer “has an absolute right to terminate his entire business for any reason he pleases. . . .” 380 U. S., at 268. Whatever may be the limits of § 8(a)(1), we said, an employer’s decision to terminate its business is one of those decisions “so peculiarly matters of management prerogative that they would never constitute violations” of that section. Id., at 269. Neither would ceasing business and refusing to bargain about it violate § 8(a)(3) or § 8(a)(5) even if done with antiunion animus. Id., at 267, n. 5, 269-274. “A proposition that a single businessman cannot choose to go out of business if he wants to would represent 16 We thought that a partial liquidation might present a different case and remanded for further findings. See 380 U. S., at 268, 276-277. 508 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. such a startling innovation that it should not be entertained without the clearest manifestation of legislative intent or unequivocal judicial precedent so construing the Labor Relations Act.” Id., at 270. We found neither.17 17 In First National Maintenance Corp. v. NLRB, 452 U. S. 666 (1981), which, like Textile Workers v. Darlington Mfg. Co., arose under the NLRA, we concluded that “the harm likely to be done to an employer’s need to operate freely in deciding whether to shut down part of its business purely for economic reasons outweighs the incremental benefit that might be gained through the union’s participation in making the decision.” 452 U. S., at 686. Further, we held that the employer’s decision to close down a segment of its business “is not part of § 8 (d)’s ‘terms and conditions,’ . . . over which Congress has mandated bargaining.” Ibid. In so holding, we did not feel constrained by the Court’s decision in Railroad Telegraphers v. Chicago & N. W. R. Co., 362 U. S. 330 (1960). Indeed, we rejected the argument that Telegraphers compelled us to find bargaining over this decision mandatory. Although we pointed in First National Maintenance to the important distinctions between the RLA and the NLRA, there are other reasons why Telegraphers does not dictate the result in these cases. In Telegraphers, the Court held that the District Court was without jurisdiction to grant injunctive relief against a labor strike under the provisions of the NLGA. A closely divided Court reasoned that a railroad’s proposal to abandon certain single-agent stations and hence abolish some jobs was a bargainable issue. In Darlington and First National Maintenance, we concluded that the analysis in Telegraphers, which rested on an “expansive” reading of the RLA and the NLGA, did not govern a decision under the NLRA. 452 U. S., at 687, n. 23. In this case, we examine Telegraphers once again in the context of the RLA. In Telegraphers a railroad was seeking simply to eliminate or consolidate some of its little-used local stations. The railroad here, by contrast, sought to sell all its lines and go out of business. There is nothing in Telegraphers that forces us to reach the result, in this extreme case, that P&LE was prohibited from terminating its operations without first bargaining with the unions. Notwithstanding the policy considerations prompting the enlarged scope of mandatory bargaining under the RLA, m light of Darlington, which First National Maintenance reaffirmed, we are not inclined to extend Telegraphers to a case in which the railroad decides to retire from the railroad business. The dissent, post, at 518-520, seems to assert that Shore Line and Telegraphers dealt with a railroad’s freedom to leave the market. But as we point out, that is precisely what those cases did not involve. We are PITTSBURGH & L. E. R. CO. v. RAIUNXY EXECUTIVES 509 490 Opinion of the Court Although Darlington arose under the NLRA, we are convinced that we should be guided by the admonition in that case that the decision to close down a business entirely is so much a management prerogative that only an unmistakable expression of congressional intent will suffice to require the employer to postpone a sale of its assets pending the fulfillment of any duty it may have to bargain over the subject matter of union notices such as were served in this litigation. Absent statutory direction to the contrary, the decision of a railroad employer to go out of business and consequently to reduce to zero the number of available jobs is not a change in the conditions of employment forbidden by the status quo provision of § 156. In these cases, P&LE concluded that it must sell its assets, and its agreement to sell to Railco, if implemented, would have removed it from the railroad business; no longer would it be a railroad employer. No longer would it need the services of members of the rail unions. The RLE A concedes that had the collective-bargaining agreements expressly waived bargaining concerning sale of P&LE’s assets, the unions’ § 156 notices to change the agreements could not trump the terms of the agreements and could not delay the sale. Brief for Respondent RLEA 44. We think the same result follows where the agreement is silent on the matter and the railroad employer has proceeded in accordance with the ICA. In these circumstances, there is little or no basis for the unions to expect that a § 156 notice would be effective to delay the company’s departure from the railroad business. Congress clearly requires that sales transactions like P&LE’s proposal must satisfy the requirements of the ICA, but we find nothing in the RLA to prevent the immediate consummation of P&LE’s contract to sell. When the ICC approved the sale by permitting the Ex Parte 392 exemption to become effective, P&LE was free to close the transaction and should not have been enjoined from doing so. plainly at odds with the dissent with respect to the significance of P&LE’s decision to leave the railroad business. 510 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. This construction of the RLA also responds to our obligation to avoid conflicts between two statutory regimes, namely, the RLA and ICA, that in some respects overlap. As the Court has said, we “are not at liberty to pick and choose among congressional enactments, and when two statutes are capable of co-existence, it is the duty of the courts, absent a clearly expressed congressional intention to the contrary, to regard each as effective.” Morton v. Mancari, 417 U. S. 535, 551 (1974). We should read federal statutes “to give effect to each if we can do so while preserving their sense and purpose.” Watt v. Alaska, 451 U. S. 259, 267 (1981); see also United States v. Fausto, 484 U. S. 439, 453 (1988). We act accordingly in this litigation. Congress has exercised its Commerce Clause authority to regulate rail transportation for over a century. See Act to regulate commerce of 1887 (the ICA), ch. 104, 24 Stat. 379. In doing so, Congress has assigned to the ICC plenary authority over rail transactions, ranging from line extensions, consolidations, and abandonments, to acquisitions. In particular, the ICA in 49 U. S. C. § 10901(a) permits noncarriers to acquire a rail line only if the ICC determines that “the present or future public convenience and necessity require or permit” the rail acquisition and operation. The ICC may approve certification on satisfaction of various conditions. Specifically, it has authority to impose labor protection provisions though it is not obligated to do so. § 10901(e). Acting pursuant to § 10505, the ICC, in its Ex Parte 392 exemption proceedings, declared all noncarrier acquisitions presumptively exempt from § 10901 regulation. Such transactions would be deemed approved seven days after a notice filed by the acquiring entities. 49 CFR § 1150.32(b) (1987). And absent a showing of exceptional circumstances, which rail labor was entitled to demonstrate, labor protection provisions would not be imposed. The Ex Parte 392 procedures, and the ICA, § 10505 exemption authority generally, like amendments to ICA in the last two PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 511 490 Opinion of the Court decades, see, e. g., the Railroad Revitalization and Regulatory Reform Act of 1976, Pub. L. 94-210, 90 Stat. 31; the Staggers Rail Act of 1980, Pub. L. 96-448, 94 Stat. 1895, aimed at reversing the rail industry’s decline through de-regulatory efforts, above all by streamlining procedures to effectuate economically efficient transactions. Here P&LE agreed to sell its assets to Railco. The transaction was presented to the ICC and an Ex Parte 392 exemption was requested. The ICC rejected the unions’ applications to stay or reject the exemption, which became effective seven days after it was requested. The unions then successfully sought an injunction delaying the closing of the transaction based on their § 156 notices. The Court of Appeals several times noted the tension between the two regimes, but concluded that the provisions of the RLA left no room for a construction easing those tensions. This was the case even though the injunction that was affirmed would likely result in cancellation of P&LE’s sale and the frustration of Congress’ intent through ICA amendments to deregulate the rail and air industries generally and more specifically to assist small rail lines with financial problems. We disagree with that conclusion, for as we have said, we are confident that the RLA is reasonably subject to a construction that would, at least to a degree, harmonize the two statutes.18 The injunction, which effectively prevented the sale from going forward, should not have been granted. lfiP&LE argues that the RLA injunction was an impermissible collateral attack on the ICC order approving the sale. But the ICA, 49 U. S. C. § 10901, and the RLA, 45 U. S. C. § 156, as we construe them, are complementary regimes. Here, the ICC simply granted an exemption from the strictures of § 10901, which permitted, but did not order, the consummation of the sale. It made no finding that would prevent enforcement of §156. The dissent, post, at 515, asserts that we ignore the principle that P&LE, a regulated utility, may not enter or leave the market without agency approval. Of course, we do not, for we set out the law that requires ICC consent to the sale, which was obtained. 512 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. c Our holding in these cases, which rests on our construction of the RLA and not on the pre-emptive force of the ICA, is that petitioner was not obligated to serve its own § 156 notice on the unions in connection with the proposed sale. We also conclude that the unions’ notices did not obligate P&LE to maintain the status quo and postpone the sale beyond the time the sale was approved by the Commission and was scheduled to be consummated. We do not hold, however, that P&LE had no duty at all to bargain in response to the unions’ § 156 motions. The courts below held, and RLEA agrees, that P&LE’s decision to sell, as such, was not a bargainable subject. The disputed issue is whether P&LE was required to bargain about the effects that the sale would or might have upon its employees. P&LE, in our view, was not entirely free to disregard the unions’ demand that it bargain about such effects. When the unions’ notices were served, however, the terms of P&LE’s agreement with Railco were more or less settled, and P&LE’s decision to sell on those terms had been made. To the extent that the unions’ demands could be satisfied only by the assent of the buyers, they sought to change or dictate the terms of the sale, and in effect challenged the decision to sell itself. At that time, P&LE was under no obligation to bargain about the terms it had already negotiated. To the extent that the unions’ proposals could be satisfied by P&LE itself, those matters were bargainable but only until the date for closing the sale arrived, which, of course, could not occur until the Ex Parte 392 exemption became effective.19 We are therefore constrained to reverse the Court of Appeals in No. 87-1888. 19 We address the duty to bargain about the effects of the sale only in the context of the facts existing when the unions’ notices were served. We do not deal with a railroad employer’s duty to bargain in response to a union s § 156 notice proposing labor protection provisions in the event that a sale, not yet contemplated, should take place. PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 513 490 Opinion of the Court III In No. 87-1589, the issue is whether the Court of Appeals was correct in setting aside the injunction against the strike issued on October 8, 1987. At that time, the Ex Parte 392 exemption had become effective, and the District Court held that because the ICC had in effect authorized the sale and had ruled that delay would be prejudicial to the parties and the public interests, the NLGA prohibition against issuing injunctions in labor dispute cases must be accommodated to the ICC’s decision that the sale of assets should go forward. It was this decision, based on the legal significance of the ICA and its impact on the NLGA, that the Court of Appeals summarily reversed. We agree with that decision. We have held that the NLGA § 4 general limitation on district courts’ power to issue injunctions in labor disputes must be accommodated to the more specific provisions of the RLA: “[T]he District Court has jurisdiction and power to issue necessary injunctive orders” to enforce compliance with the requirements of the RLA “notwithstanding the provisions of the Norris-LaGuardia Act.” Trainmen v. Howard, 343 U. S. 768, 774 (1952). Thus, a union may be enjoined from striking when the dispute concerns the interpretation or application of its contract and is therefore subject to compulsory arbitration. Trainmen v. Chicago River & Indiana R. Co., 353 U. S. 30 (1957). “[T]he specific provisions of the Railway Labor Act take precedence over the more general provisions of the Norris-LaGuardia Act.” Id., at 41-42. The same accommodation of the NLGA to the specific provisions of the NLRA must be made. A union that has agreed to arbitrate contractual disputes and is subject to a no-strike clause may be enjoined from striking despite the NLGA. Boys Markets, Inc. n. Retail Clerks, 398 U. S. 235 (1970). Petitioner contends that the NLGA must likewise be accommodated to the procedures mandated by Congress in 49 U. S. C. §10901 specifically the authority of the ICC to impose labor protective provisions, the right of rail 514 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. labor to seek such provisions from the ICC, and its right to judicial review if dissatisfied. It is urged that the ICA provides a comprehensive scheme for the resolution of labor protection issues arising out of ICC-regulated transactions and that rail labor must take advantage of those procedures rather than strike. We are unpersuaded that this is the case. The prohibition of the NLGA must give way when necessary to enforce a duty specifically imposed by another statute. But no applicable provision has been called to our attention that imposes any duty on rail unions to participate in ICC proceedings and to seek ICC protections with which they must be satisfied. Furthermore, labor protection provisions run against the acquiring railroad rather than the seller. Yet here it is with the seller, P&LE, that the unions wanted to bargain, seeking to ease the adverse consequences of the sale. To that end, the unions served § 156 notices, which at least to some extent obligated P&LE to bargain until its transaction was closed. We find nothing in the ICA that relieved P&LE of that duty, nor anything in that Act that empowers the ICC to intrude into the relationship between the selling carrier and its railroad unions. We are thus quite sure that the NLGA forbade an injunction against that strike unless the strike was contrary to the unions’ duties under the RLA. As to that issue, the Court of Appeals stated: “We intimate no view as to whether the provisions of the Railway Labor Act are applicable to this dispute so that the district court would be entitled to enjoin the strike while that Act’s dispute resolution mechanisms are underway. RLEA’s complaint seeking a declaration that the Railway Labor Act is applicable to this dispute is the merits issue before the district court.” 831 F. 2d, at 1237. On remand, the District Court held that the RLA was indeed applicable to the dispute and on that basis issued an injunction against P&LE. It did not, however, ever address the question whether the unions PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 515 490 Opinion of Stevens, J. strike, which occurred after their suit was filed, was enjoin-able under the RLA. Neither did the Court of Appeals deal with that issue in affirming the District Court. P&LE perfunctorily asserts in its briefs in this Court that the strike injunction was proper because the unions were obligated to bargain rather than strike after their § 156 notices were served. RLEA did not respond to this assertion. With the case in this position, we shall not pursue the issue. Instead, we vacate the judgment of the Court of Appeals, and leave the matter, if it is a live issue, to be dealt with on remand. IV The judgment of the Court of Appeals in No. 87-1888 is reversed and the judgment in No. 87-1589 is vacated, and the cases are remanded for further proceedings consistent with this opinion. So ordered. Justice Stevens, with whom Justice Brennan, Justice Marshall, and Justice Blackmun join, concurring in part and dissenting in part. Regulated utilities do not have the same freedom to respond to market pressures that unregulated firms have.1 They may not raise rates or cut services, for example, without permission from a regulatory agency. Most significantly for these cases, they may neither enter nor leave the market without agency approval. Ignoring this principle, the Court in Part II of its opinion arrives at a result that, while perhaps preferable as a matter of policy, contradicts our previous interpretations of the relevant statute.2 *Æ. g., D. Hjelmfelt, Antitrust and Regulated Industries §1.3 (1985). 2 Because I agree with the Court’s factual description of these cases and its resolution of No. 87-1589, I join Parts I and III of its opinion. 516 OCTOBER TERM, 1988 Opinion of Stevens, J. 491 U. S. The railroad industry long has been the subject of governmental regulation.3 A year after this Court held that individual States were powerless to regulate rail lines extending beyond their boundaries, Wabash, S. L. & P. R. Co. v. Illinois, 118 U. S. 557 (1886), Congress established the Interstate Commerce Commission (ICC) to regulate economic aspects of the rail industry. Interstate Commerce Act, 49 U. S. C. § 10101 et seq. (1982 ed. and Supp. V). Regulation of employment relationships within the rail industry followed,4 and in 1926, Congress enacted the Railway Labor Act (RLA), 45 U. S. C. § 151 et seq. The intervening six decades were marked by relatively peaceful coexistence between the two statutes. During the course of the employment relationship, the RLA provided the means for resolving disputes. See ante, at 496, n. 4; Consolidated Rail Corporation v. Railway Labor Executives9 Assn., ante, at 302-304. If a railroad sought to end that relationship by sale, consolidation, or abandonment, the ICC routinely conditioned approval on the railroad’s acceptance of either job protection or some form of severance pay for employees who would be affected by the change. See United States v. Lowden, 308 U. S. 225 (1939).5 Cf. ante, at 498. 3 See W. Jones, Cases and Materials on Regulated Industries 7-11, 24-55 (2d ed. 1976); M. Glaeser, Public Utilities in American Capitalism 57-71 (1957); I. Sharfman, The Interstate Commerce Commission, pt. 1, pp. 11-35 (1931). 4 See, e. g., Hours of Service Act of 1907, as amended, 45 U. S. C. §§ 61-66; Employers’ Liability Act of 1908, as amended, 45 U. S. C. §§ 51-60. See also Sharfman 180-182; L. Lecht, Experience under Railway Labor Legislation 14-46 (1955). 5 Labor protective provisions approved in Lowden included salary-level maintenance, preservation of seniority rights, and severance and relocation payments. 308 U. S., at 228. In concluding that the ICC had the power to impose such conditions, Justice Stone wrote for the Court: “One must disregard the entire history of railroad labor relations in the United States to be able to say that the just and reasonable treatment of railroad employees in mitigation of the hardship imposed on them in carrying out the national policy of railway consolidation, has no bearing on the PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 517 490 Opinion of Stevens, J. This symbiosis ended in 1985, when the ICC announced that it no longer would impose labor protective conditions on sales of short-line railroads unless exceptional circumstances were shown. Ex Parte No. 392 (Sub. No. 1), Class Exemption for the Acquisition and Operation of Rail Lines Under 49 U. S. C. 10901, 1 I. C. C. 2d 810, 815 (1985), review denied sub nom. Illinois Commerce Common v. ICC, 260 U. S. App. D. C. 38, 817 F. 2d 145 (1987); see ante, at 498-501. Suddenly it became important for railroad unions to obtain such labor protections through collective bargaining. Unlike other employment contracts, however, rail labor agreements are altered not by periodic renegotiation but by notification, pursuant to § 6 of the RLA, 45 U. S. C. § 156, of a desire to change terms in the agreements. See Tr. of Oral Arg. 66-67. Thus it is not surprising that the unions in this litigation did not seek labor protective provisions until—just 18 months after the ICC abdicated its traditional protective role—plans to sell the railroad surfaced.6 There is no disagreement that labor protective provisions related to the effects of an abandonment or sale may be the subject of collective bargaining. It follows, I believe, that when railway labor unions request the inclusion of such pro successful prosecution of that policy and no relationship to the maintenance of an adequate and efficient transportation system.” Id., at 234. See also ICC v. Railway Labor Executives’ Assn., 315 U. S. 373 (1942). 6 The railroad might have had a greater duty to bargain, the Court suggests, had the unions served notice before sale negotiations had commenced. See ante, at 512, n. 19. Yet in the two opinions that I believe should control these cases, we did not fault the unions for filing § 6 notices in reaction to—rather than in anticipation of—the railroads’ initiatives. Compare Railroad Telegraphers n. Chicago & N. W. R. Co., 362 U. S. 330, 332 (1960), with id., at 349 (Whittaker, J., dissenting) (majority rejects railroad’s argument that § 6 notices were improper because filed after railroad petitioned state regulatory commissions for permission to abolish jobs). See also Detroit & Toledo Shore Line R. Co. v. Transportation Union, 396 U. S. 142, 146 (1969). In light of the ICC’s abrupt halt to its practice of requiring labor protections, moreover, the Court’s distinction unfairly penalizes the unions in this litigation. 518 OCTOBER TERM, 1988 Opinion of Stevens, J. 491 U. S. visions in their collective-bargaining agreements by proper statutory notice, see ante, at 496-497, and n. 5, the employer must maintain the status quo during the statutorily mandated negotiating process or risk a strike as a consequence of its breach of that duty. See § § 2 First, Seventh of the RLA, 45 U. S. C. §§ 152 First, Seventh. The Court admits the force of this proposition and acknowledges that an employer has some duty to bargain when a sale is announced. Ante, at 504, 512. Nevertheless, it indicates that this particular dispute did not obligate the railroad to preserve the status quo, for the Court would prohibit any bargaining that “in effect challenged the decision to sell,” and would allow negotiations to cease as soon as the sale is closed. Ante, at 512.7 This diminution of the employer’s duty contravenes two of our decisions interpreting the RLA. In Railroad Telegraphers n. Chicago & N. W. R. Co., 362 U. S. 330 (1960), a railroad had decided, with the approval of state regulatory commissions, to abandon a large number of its local stations and thus remove several hundred station attendants from the payroll. This Court held that because the RLA “commandfs] that employees as well as railroads exert every reasonable effort to settle all disputes ‘concerning rates of pay, rules, and working conditions,”’ the union had a right to strike to prevent the railroad from implementing the partial abandonment without bargaining over effects. Id., at 339 (quoting §2 First of the RLA, 45 U. S. C. § 152 First). The Court continued: 7 The Court neglects to mention that a sale may be closed within a matter of months, whereas resort to RLA procedures may entail “virtually endless ‘negotiation, mediation, voluntary arbitration, and conciliation.’” Burlington Northern R. Co. v. Maintenance of Way Employes, 481 U. S. 429, 444 (1987) (quoting Shore Line, 396 U. S., at 148-149). If the railroad knows its obligations will end when the sale is consummated, it will have no incentive to expedite bargaining. Thus the Court’s imposition of a minimal bargaining duty affords employees scarcely more protection than they would have absent any duty. PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 519 490 Opinion of Stevens, J. “It would stretch credulity too far to say that the Railway Labor Act, designed to protect railroad workers, was somehow violated by the union acting precisely in accordance with that Act’s purpose to obtain stability and permanence in employment for workers. There is no express provision of law, and certainly we can infer none from the Interstate Commerce Act, making it unlawful for unions to want to discuss with railroads actions that may vitally and adversely affect the security, seniority and stability of railroad jobs.” 362 U. S., at 339-340. Telegraphers thus holds that if management decides to abandon a significant part of a railroad’s business, the impact of that decision on employees’ job security is a proper subject for bargaining under the RLA. Detroit & Toledo Shore Line R. Co. v. Transportation Union, 396 U. S. 142 (1969) (Shore Line), concerned a railroad’s proposal to make new work assignments, a change neither authorized nor prohibited by the collective-bargaining agreement. The Court held that once the union had served notice of its desire to bargain, the railroad was obligated to maintain the status quo until completion of the RLA’s “ ‘purposely long and drawn out’ ” bargaining process. Id., at 149 (quoting Railway Clerks v. Florida East Coast R. Co., 384 U. S. 238, 246 (1966)). It further rejected the railroad’s argument that the “status quo” encompassed only working conditions expressed in an agreement between the parties: “[T]he language of § 6 simply does not say what the railroad would have it say. Instead, the section speaks plainly of ‘rates of pay, rules, or working conditions’ without any limitation to those obligations already embodied in collective agreements. More important, we are persuaded that the railroad’s interpretation of this section is sharply at variance with the overall design and purpose of the Railway Labor Act.” 396 U. S., at 148. 520 OCTOBER TERM, 1988 Opinion of Stevens, J. 491 U. S. The Court therefore construed “status quo” to mean “those actual, objective working conditions and practices, broadly conceived, which were in effect prior to the time the pending dispute arose and which are involved in or related to that dispute.” Id., at 153. Today the Court proffers three reasons why Shore Line does not control these cases. First, it asserts that the Shore Line holding that “status quo” includes “conditions ‘objectively’ in existence when the union’s notice was served” stretched the language of the statute “to its outer limits,” ante, at 506. I am not at all sure that is true; even if it is, the holding is unambiguous and has the force of law. Second, the Court suggests that the fact that the work assignment changed in Shore Line had been in effect for many years justified an expectation “that it would not be changed without bargaining and compliance with the status quo provisions of the RLA.” Ante, at 506. This effectively restates Justice Harlan’s argument in dissent that while not limited to the terms of written agreements, the status quo obligation is limited to a change in settled practice. See Shore Line, 396 U. S., at 159-160. The Court’s emphasis on those dissenting remarks avails it nothing, because the instant controversy also arose out of a change in established procedure. By either a subjective or objective measure, therefore, it is reasonable to conclude that these employees’ jobs are among the “working conditions” that must be preserved throughout the bargaining process. Third, and most importantly, the Court points out that in contrast with these cases, the railroad in Shore Line had not proposed “to quit the railroad business, sell its assets, and cease to be a railroad employer at all, ” ante, at 507. The simple reply is that, in spite of claims of “ ‘managerial prerogative’” much like those advanced here,8 the Court in Tele- 8 Compare Telegraphers, 362 U. S., at 336 (“We cannot agree with the Court of Appeals that the union’s effort to negotiate about the job security of its members ‘represents an attempt to usurp legitimate managerial pre- PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 521 490 Opinion of Stevens, J. graphers held that the effects of a railroad’s decision to terminate a part of its business constituted a proper subject of bargaining. There is no relevant difference between the partial abandonment in Telegraphers and the transfer of ownership proposed in these cases: in both, rail service would continue as before, but many employees would lose their jobs. Management’s motive in Telegraphers, to cut costs by eliminating a large number of dispensable jobs, was of course perfectly reasonable. Thus when the Court held that the RLA required the railroad to bargain over the effects of the change, Justice Clark wrote: “Today the Court tells the railroad that it must bargain with the union or suffer a strike. The latter would be the death knell of the railroad. Hence, for all practical purposes, the Court is telling the railroad that it must secure the union’s approval before severing the hundreds of surplus employees now carried on its payroll. Everyone knows what the answer of the union will be. It is like the suitor who, when seeking the hand of a young lady, was told by her to ‘go to father.’ But, as the parody goes, ‘She knew that he knew that her father was dead; she knew that he knew what a life he had led; and she knew that he knew what she meant when she said “go to father.”’” 362 U. S., at 343-344 (dissenting opinion). Had the sale in these cases proceeded, the railroad would have operated the same service with a work force of 250 as compared to 750 employees. Ante, at 495. The economic benefits of that reduction are as obvious as those that would have been achieved by closing obsolete stations on the railroad system in Telegraphers. It is just as obvious, I believe, that rogative in the exercise of business judgment with respect to the most economical and efficient conduct of its operations’ ”), with Brief for Petitioner 25 (“A decision to go out of business is the quintessential managerial prerogative”). 522 OCTOBER TERM, 1988 Opinion of Stevens, J. 491 U. S. the RLA again commands bargaining. As Judge Becker noted in his opinion for the Court of Appeals: “We are fully aware of the unfortunate ramifications and irony of our decision. A bargaining order, and a status quo injunction, designed to foster conciliation, promote labor peace, and ultimately keep the rails running, may ultimately have the perverse effect of destroying the only chance P & LE has for survival and perhaps even the very jobs that the unions are now trying to protect. Although we are not happy with this result, we feel constrained to reach it, because the Supreme Court has appropriately admonished the judiciary not to apply its own brand of ‘common sense’ in the face of a contrary statutory mandate.” 845 F. 2d 420, 446 (CA3 1988) (citing TVA v. Hill, 437 U. S. 153, 193-195 (1978)). To evade the natural result of adherence to Shore Line and Telegraphers, the Court relies on two later opinions declaring that “an employer has the absolute right to terminate his entire business for any reason he pleases,” Textile Workers n. Darlington Mfg. Co., 380 U. S. 263, 268 (1965), and that the consequences of a partial closure are not a mandatory subject of bargaining, First National Maintenance Corp. v. NLRB, 452 U. S. 666 (1981). See ante, at 507-509, and n. 17. But those opinions interpreted the strictures that the National Labor Relations Act places on an unregulated industry. As we noted in First National Maintenance Corp., that is a situation far different from the RLA’s governance of a regulated industry.9 9 We stressed that the decision in Telegraphers “rested on the particular aims of the Railway Labor Act and national transportation policy. See 362 U. S., at 336-338. The mandatory scope of bargaining under the Railway Labor Act. . . [is] not coextensive with the National Labor Relations Act and the [National Labor Relations] Board’s jurisdiction over unfair labor practices. See Chicago & N. W. R. Co. v. Transportation Union, 402 U. S. 570, 579, n. 11 (1971) (‘parallels between the duty to bargain in good faith and the duty to exert every reasonable PITTSBURGH & L. E. R. CO. v. RAILWAY EXECUTIVES 523 490 Opinion of Stevens, J. At issue today is the RLA’s regulation of a railroad’s freedom to leave the market. Perhaps the RLA’s restrictions on that freedom, as interpreted in Telegraphers and Shore Line, do not best serve national transportation interests. But since Congress has not overruled those interpretations, it is, as Judge Becker observed, inappropriate for judges to undertake to fill the perceived policy void. For these reasons, I would affirm the judgment of the Court of Appeals in No. 87-1888. effort, like all parallels between the NLRA and the Railway Labor Act, should be drawn with the utmost care and with full awareness of the differences between the statutory schemes’).” First National Maintenance Corp. v. NLRB, 452 U. S. 666, 686-687, n. 23 (1981). 524 OCTOBER TERM, 1988 Syllabus 491 U. S. THE FLORIDA STAR v. B. J. F. APPEAL FROM THE DISTRICT COURT OF APPEAL OF FLORIDA, FIRST DISTRICT No. 87-329. Argued March 21, 1989—Decided June 21, 1989 Appellant, The Florida Star, is a newspaper which publishes a “Police Reports” section containing brief articles describing local criminal incidents under police investigation. After appellee B. J. F. reported to the Sheriff’s Department (Department) that she had been robbed and sexually assaulted, the Department prepared a report, which identified B. J. F. by her full name, and placed it in the Department’s pressroom. The Department does not restrict access to the room or to the reports available there. A Star reporter-trainee sent to the pressroom copied the police report verbatim, including B. J. F.’s full name. Consequently, her name was included in a “Police Reports” story in the paper, in violation of the Star’s internal policy. Florida Stat. § 794.03 makes it unlawful to “print, publish, or broadcast ... in any instrument of mass communication” the name of the victim of a sexual offense. B. J. F. filed suit in a Florida court alleging, inter alia, that the Star had negligently violated § 794.03. The trial court denied the Star’s motion to dismiss, which claimed, among other things, that imposing civil sanctions on the newspaper pursuant to § 794.03 violated the First Amendment. However, it granted B. J. F.’s motion for a directed verdict on the issue of negligence, finding the Star per se negligent based on its violation of § 794.03. The jury then awarded B. J. F. both compensatory and punitive damages. The verdict was upheld on appeal. Held: Imposing damages on the Star for publishing B. J. F.’s name violates the First Amendment. Pp. 530-541. (a) The sensitivity and significance of the interests presented in clashes between First Amendment and privacy rights counsels the Court to rely on limited principles that sweep no more broadly than the appropriate context of the instant case, rather than to accept invitations to hold broadly that truthful publication may never be punished consistent with the First Amendment or that publication of a rape victim’s name never enjoys constitutional protection. One such principle is that “if a newspaper lawfully obtains truthful information about a matter of public significance then state officials may not constitutionally punish publication of the information, absent a need to further a state interest of the highest order.” Smith v. Daily Mail Publishing Co., 443 U. S. 97, 103. THE FLORIDA STAR v. B. J. F. 525 524 Syllabus Applied to the instant case, the Daily Mail principle commands reversal. Pp. 530-536. (b) The Star “lawfully obtain[ed] truthful information.” The actual news article was accurate, and the Star lawfully obtained B. J. F.’s name from the government. The fact that state officials are not required to disclose such reports or that the Sheriff’s Department apparently failed to fulfill its § 794.03 obligation not to cause or allow B. J. F.’s name to be published does not make it unlawful for the Star to have received the information, and Florida has taken no steps to proscribe such receipt. The government has ample means to safeguard the information that are less drastic than punishing truthful publication. Furthermore, it is clear that the news article generally, as opposed to the specific identity contained in it, involved “a matter of public significance”: the commission, and investigation, of a violent crime that had been reported to authorities. Pp. 536-537. (c) Imposing liability on the Star does not serve “a need to further a state interest of the highest order.” Although the interests in protecting the privacy and safety of sexual assault victims and in encouraging them to report offenses without fear of exposure are highly significant, imposing liability on the Star in this case is too precipitous a means of advancing those interests. Since the Star obtained the information because the Sheriff’s Department failed to abide by § 794.03’s policy, the imposition of damages can hardly be said to be a narrowly tailored means of safeguarding anonymity. Self-censorship is especially likely to result from imposition of liability when a newspaper gains access to the information from a government news release. Moreover, the negligence per se standard adopted by the courts below does not permit case-by-case findings that the disclosure was one a reasonable person would find offensive and does not have a scienter requirement of any kind. In addition, § 794.03’s facial under inclusiveness—which prohibits publication only by an “instrument of mass communication” and does not prohibit the spread of victims’ names by other means—raises serious doubts about whether Florida is serving the interests specified by B. J. F. A State must demonstrate its commitment to the extraordinary measure of punishing truthful publication in the name of privacy by applying its prohibition evenhandedly to both the smalltime disseminator and the media giant. Pp. 537-541. 499 So. 2d 883, reversed. Marshall, J., delivered the opinion of the Court, in which Brennan, Blackmun, Stevens, and Kennedy, JJ., joined. Scalia, J., filed an opinion concurring in part and concurring in the judgment, post, p. 541. 526 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. White, J., filed a dissenting opinion, in which Rehnquist, C. J., and O’Connor, J., joined, post, p. 542. George K. Rahdert argued the cause and filed briefs for appellant. Joel D. Eaton argued the cause and filed a brief for appellee.* Justice Marshall delivered the opinion of the Court. Florida Stat. §794.03 (1987) makes it unlawful to “print, publish, or broadcast... in any instrument of mass communication” the name of the victim of a sexual offense.1 Pursuant to this statute, appellant The Florida Star was found civilly liable for publishing the name of a rape victim which it had obtained from a publicly released police report. The issue presented here is whether this result comports with the First Amendment. We hold that it does not. I The Florida Star is a weekly newspaper which serves the community of Jacksonville, Florida, and which has an average circulation of approximately 18,000 copies. A regular feature of the newspaper is its “Police Reports” section. * Briefs of amici curiae urging reversal were filed for the American Newspaper Publishers Association et al. by Richard J. Ovelmen, W. Terry Maguire, Gary B. Pruitt, Paul J. Levine, Laura Besvinick, and Gregg D. Thomas; and for the Reporters Committee for Freedom of the Press et al. by Jane E. Kirtley, Robert J. Brinkmann, and J. Laurent Scharff. Ronald A. Zumbrun and Anthony T. Caso filed a brief for the Pacific Legal Foundation as amicus curiae urging affirmance. 1 The statute provides in its entirety: “Unlawful to publish or broadcast information identifying sexual offense victim.—No person shall print, publish, or broadcast, or cause or allow to be printed, published, or broadcast, in any instrument of mass communication the name, address, or other identifying fact or information of the victim of any sexual offense within this chapter. An offense under this section shall constitute a misdemeanor of the second degree, punishable as provided in § 775.082, § 775.083, or § 775.084.” Fla. Stat. § 794.03 (1987). THE FLORIDA STAR v. B. J. F. 527 524 Opinion of the Court That section, typically two to three pages in length, contains brief articles describing local criminal incidents under police investigation. On October 20, 1983, appellee B. J. F.2 reported to the Duval County, Florida, Sheriff’s Department (Department) that she had been robbed and sexually assaulted by an unknown assailant. The Department prepared a report on the incident which identified B. J. F. by her full name. The Department then placed the report in its pressroom. The Department does not restrict access either to the pressroom or to the reports made available therein. A Florida Star reporter-trainee sent to the pressroom copied the police report verbatim, including B. J. F.’s full name, on a blank duplicate of the Department’s forms. A Florida Star reporter then prepared a one-paragraph article about the crime, derived entirely from the trainee’s copy of the police report. The article included B. J. F.’s full name. It appeared in the “Robberies” subsection of the “Police Reports” section on October 29, 1983, one of 54 police blotter stories in that day’s edition. The article read: “[B. J. F.] reported on Thursday, October 20, she was crossing Brentwood Park, which is in the 500 block of Golfair Boulevard, enroute to her bus stop, when an unknown black man ran up behind the lady and placed a knife to her neck and told her not to yell. The suspect then undressed the lady and had sexual intercourse with her before fleeing the scene with her 60 cents, Timex watch and gold necklace. Patrol efforts have been suspended concerning this incident because of a lack of evidence.” 2 In filing this lawsuit, appellee used her full name in the caption of the case. On appeal, the Florida District Court of Appeal sua sponte revised the caption, stating that it would refer to the appellee by her initials, “in order to preserve [her] privacy interests.” 499 So. 2d 883, 883, n. (1986). Respecting those interests, we, too, refer to appellee by her initials, both in the caption and in our discussion. 528 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. In printing B. J. F.’s full name, The Florida Star violated its internal policy of not publishing the names of sexual offense victims. On September 26, 1984, B. J. F. filed suit in the Circuit Court of Duval County against the Department and The Florida Star, alleging that these parties negligently violated §794.03. See n. 1, supra. Before trial, the Department settled with B. J. F. for $2,500. The Florida Star moved to dismiss, claiming, inter alia, that imposing civil sanctions on the newspaper pursuant to §794.03 violated the First Amendment. The trial judge rejected the motion. App. 4. At the ensuing daylong trial, B. J. F. testified that she had suffered emotional distress from the publication of her name. She stated that she had heard about the article from fellow workers and acquaintances; that her mother had received several threatening phone calls from a man who stated that he would rape B. J. F. again; and that these events had forced B. J. F. to change her phone number and residence, to seek police protection, and to obtain mental health counseling. In defense, The Florida Star put forth evidence indicating that the newspaper had learned B. J. F.’s name from the incident report released by the Department, and that the newspaper’s violation of its internal rule against publishing the names of sexual offense victims was inadvertent. At the close of B. J. F.’s case, and again at the close of its defense, The Florida Star moved for a directed verdict. On both occasions, the trial judge denied these motions. He ruled from the bench that § 794.03 was constitutional because it reflected a proper balance between the First Amendment and privacy rights, as it applied only to a narrow set of “rather sensitive . . . criminal offenses.” App. 18-19 (rejecting first motion); see id., at 32-33 (rejecting second motion). At the close of the newspaper’s defense, the judge granted B. J. F.’s motion for a directed verdict on the issue of negligence, finding the newspaper per se negligent based upon its THE FLORIDA STAR v. B. J. F. 529 524 Opinion of the Court violation of § 794.03. Id., at 33. This ruling left the jury to consider only the questions of causation and damages. The judge instructed the jury that it could award B. J. F. punitive damages if it found that the newspaper had “acted with reckless indifference to the rights of others.” Id., at 35. The jury awarded B. J. F. $75,000 in compensatory damages and $25,000 in punitive damages. Against the actual damages award, the judge set off B. J. F.’s settlement with the Department. The First District Court of Appeal affirmed in a three-paragraph per curiam opinion. 499 So. 2d 883 (1986). In the paragraph devoted to The Florida Star’s First Amendment claim, the court stated that the directed verdict for B. J. F. had been properly entered because, under §794.03, a rape victim’s name is “of a private nature and not to be published as a matter of law.” Id., at 884, citing Doe v. Sarasota-Bradenton Florida Television Co., 436 So. 2d 328, 330 (Fla. App. 1983) (footnote omitted).3 The Supreme Court of Florida denied discretionary review. The Florida Star appealed to this Court.4 We noted probable jurisdiction, 488 U. S. 887 (1988), and now reverse. 3 In Doe v. Sarasota-Bradenton Florida Television Co., 436 So. 2d, at 329, the Second District Court of Appeal upheld the dismissal on First Amendment grounds of a rape victim’s damages claim against a Florida television station which had broadcast portions of her testimony at her assailant’s trial. The court reasoned that, as in Cox Broadcasting Corp. n. Cohn, 420 U. S. 469 (1975), the information in question “was readily available to the public, through the vehicle of a public trial.” 436 So. 2d, at 330. The court stated, however, that § 794.03 could constitutionally be applied to punish publication of a sexual offense victim’s name or other identifying information where it had not yet become “part of an open public record” by virtue of being revealed in “open, public judicial proceedings.” Ibid., citing Fla. Op. Atty. Gen. 075-203 (1975). 4 Before noting probable jurisdiction, we certified to the Florida Supreme Court the question whether it had possessed jurisdiction when it declined to hear the newspaper’s case. 484 U. S. 984 (1987). The State Supreme Court answered in the affirmative. 530 So. 2d 286, 287 (1988). 530 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. II The tension between the right which the First Amendment accords to a free press, on the one hand, and the protections which various statutes and common-law doctrines accord to personal privacy against the publication of truthful information, on the other, is a subject we have addressed several times in recent years. Our decisions in cases involving government attempts to sanction the accurate dissemination of information as invasive of privacy, have not, however, exhaustively considered this conflict. On the contrary, although our decisions have without exception upheld the press’ right to publish, we have emphasized each time that we were resolving this conflict only as it arose in a discrete factual context.5 The parties to this case frame their contentions in light of a trilogy of cases which have presented, in different contexts, the conflict between truthful reporting and state-protected privacy interests. In Cox Broadcasting Corp. v. Cohn, 420 U. S. 469 (1975), we found unconstitutional a civil damages award entered against a television station for broadcasting the name of a rape-murder victim which the station had obtained from courthouse records. In Oklahoma Publishing 5 The somewhat uncharted state of the law in this area thus contrasts markedly with the well-mapped area of defamatory falsehoods, where a long line of decisions has produced relatively detailed legal standards governing the multifarious situations in which individuals aggrieved by the dissemination of damaging untruths seek redress. See, e. g., New York Times Co. v. Sullivan, 376 U. S. 254 (1964); Garrison v. Louisiana, 379 U. S. 64 (1964); Henry v. Collins, 380 U. S. 356 (1965); Rosenblatt v. Baer, 383 U. S. 75 (1966); Time, Inc. v. Hill, 385 U. S. 374 (1967); Greenbelt Cooperative Publishing Assn., Inc. v. Bresler, 398 U. S. 6 (1970); Monitor Patriot Co. v. Roy, 401 U. S. 265 (1971); Time, Inc. v. Pape, 401 U. S. 279 (1971); Rosenbloom v. Metromedia, Inc., 403 U. S. 29 (1971); Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974); Herbert v. Lando, 441 U. S. 153 (1979); Hutchinson n. Proxmire, 443 U. S. Ill (1979); Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U. S. 749 (1985); Philadelphia Newspapers, Inc. v. Hepps, 475 U. S. 767 (1986); Anderson v. Liberty Lobby, Inc., 477 U. S. 242 (1986). THE FLORIDA STAR v. B. J. F. 531 524 Opinion of the Court Co. v. Oklahoma County District Court, 430 U. S. 308 (1977), we found unconstitutional a state court’s pretrial order enjoining the media from publishing the name or photograph of an 11-year-old boy in connection with a juvenile proceeding involving that child which reporters had attended. Finally, in Smith v. Daily Mail Publishing Co., 443 U. S. 97 (1979), we found unconstitutional the indictment of two newspapers for violating a state statute forbidding newspapers to publish, without written approval of the juvenile court, the name of any youth charged as a juvenile offender. The papers had learned about a shooting by monitoring a police band radio frequency and had obtained the name of the alleged juvenile assailant from witnesses, the police, and a local prosecutor. Appellant takes the position that this case is indistinguishable from Cox Broadcasting. Brief for Appellant 8. Alternatively, it urges that our decisions in the above trilogy, and in other cases in which we have held that the right of the press to publish truth overcame asserted interests other than personal privacy,6 can be distilled to yield a broader First Amendment principle that the press may never be punished, civilly or criminally, for publishing the truth. Id., at 19. Appellee counters that the privacy trilogy is inapposite, because in each case the private information already appeared on a “public record,” Brief for Appellee 12, 24, 25, and because the privacy interests at stake were far less profound than in the present case. See, e. g., id., at 34. In the alternative, appellee urges that Cox Broadcasting be overruled and replaced with a categorical rule that publication of the 6 See, e. g., Landmark Communications, Inc. v. Virginia, 435 U. S. 829 (1978) (interest in confidentiality of judicial disciplinary proceedings); Bates v. State Bar of Arizona, 433 U. S. 350 (1977) (interest in maintaining professionalism of attorneys); Nebraska Press Assn. v. Stuart, 427 U. S. 539 (1976) (interest in accused’s right to fair trial); Virginia Pharmacy Bd. v. Virginia Citizens Consumer Council, Inc., 425 U. S. 748 (1976) (interest in maintaining professionalism of licensed pharmacists); New York Times Co. v. United States, 403 U. S. 713 (1971) (interest in national secu-rity); Garrison, supra (interest in public figure’s reputation). 532 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. name of a rape victim never enjoys constitutional protection. Tr. of Oral Arg. 44. We conclude that imposing damages on appellant for publishing B. J. F.’s name violates the First Amendment, although not for either of the reasons appellant urges. Despite the strong resemblance this case bears to Cox Broadcasting, that case cannot fairly be read as controlling here. The name of the rape victim in that case was obtained from courthouse records that were open to public inspection, a fact which Justice White’s opinion for the Court repeatedly noted. 420 U. S., at 492 (noting “special protected nature of accurate reports of judicial proceedings”) (emphasis added); see also id., at 493, 496. Significantly, one of the reasons we gave in Cox Broadcasting for invalidating the challenged damages award was the important role the press plays in subjecting trials to public scrutiny and thereby helping guarantee their fairness. Id., at 492-493.7 That role is not directly compromised where, as here, the information in question comes from a police report prepared and disseminated at a time at which not only had no adversarial criminal proceedings begun, but no suspect had been identified. Nor need we accept appellant’s invitation to hold broadly that truthful publication may never be punished consistent with the First Amendment. Our cases have carefully eschewed reaching this ultimate question, mindful that the future may bring scenarios which prudence counsels our not resolving anticipatorily. See, e. g., Near v. Minnesota ex rel. Olson, 283 U. S. 697, 716 (1931) (hypothesizing “publication of the sailing dates of transports or the number and location of troops”); see also Garrison v. Louisiana, 379 U. S. 64, 72, 7 We also recognized that privacy interests fade once information already appears on the public record, 420 U. S., at 494-495, and that making public records generally available to the media while allowing their publication to be punished if offensive would invite “self-censorship and very likely lead to the suppression of many items that. . . should be made available to the public.” Id., at 496. THE FLORIDA STAR v. B. J. F. 533 524 Opinion of the Court n. 8, 74 (1964) (endorsing absolute defense of truth “where discussion of public affairs is concerned,” but leaving unsettled the constitutional implications of truthfulness “in the discrete area of purely private libels”); Landmark Communications, Inc. v. Virginia, 435 U. S. 829, 838 (1978); Time, Inc. v. Hill, 385 U. S. 374, 383, n. 7 (1967). Indeed, in Cox Broadcasting, we pointedly refused to answer even the less sweeping question “whether truthful publications may ever be subjected to civil or criminal liability” for invading “an area of privacy” defined by the State. 420 U. S., at 491. Respecting the fact that press freedom and privacy rights are both “plainly rooted in the traditions and significant concerns of our society,” we instead focused on the less sweeping issue “whether the State may impose sanctions on the accurate publication of the name of a rape victim obtained from public records—more specifically, from judicial records which are maintained in connection with a public prosecution and which themselves are open to public inspection.” Ibid. We continue to believe that the sensitivity and significance of the interests presented in clashes between First Amendment and privacy rights counsel relying on limited principles that sweep no more broadly than the appropriate context of the instant case. In our view, this case is appropriately analyzed with reference to such a limited First Amendment principle. It is the one, in fact, which we articulated in Daily Mail in our synthesis of prior cases involving attempts to punish truthful publication: “[I]f a newspaper lawfully obtains truthful information about a matter of public significance then state officials may not constitutionally punish publication of the information, absent a need to further a state interest of the highest order.” 443 U. S., at 103. According the press the ample protection provided by that principle is supported by at least three separate considerations, in addition to, of course, the overarching “‘public interest, secured by the Constitution, in the dissemination of truth.’” Cox Broad 534 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. casting, supra, at 491, quoting Garrison, supra, at 73 (footnote omitted). The cases on which the Daily Mail synthesis relied demonstrate these considerations. First, because the Daily Mail formulation only protects the publication of information which a newspaper has “lawfully obtain[ed],” 443 U. S., at 103, the government retains ample means of safeguarding significant interests upon which publication may impinge, including protecting a rape victim’s anonymity. To the extent sensitive information rests in private hands, the government may under some circumstances forbid its nonconsensual acquisition, thereby bringing outside of the Daily Mail principle the publication of any information so acquired. To the extent sensitive information is in the government’s custody, it has even greater power to forestall or mitigate the injury caused by its release. The government may classify certain information, establish and enforce procedures ensuring its redacted release, and extend a damages remedy against the government or its officials where the government’s mishandling of sensitive information leads to its dissemination. Where information is entrusted to the government, a less drastic means than punishing truthful publication almost always exists for guarding against the dissemination of private facts. See, e. g., Landmark Communications, supra, at 845 (“[M]uch of the risk [from disclosure of sensitive information regarding judicial disciplinary proceedings] can be eliminated through careful internal procedures to protect the confidentiality of Commission proceedings”); Oklahoma Publishing, 430 U. S., at 311 (noting trial judge’s failure to avail himself of the opportunity, provided by a state statute, to close juvenile hearing to the public, including members of the press, who later broadcast juvenile defendant’s name); Cox Broadcasting, supra, at 496 ("If there are privacy interests to be protected in judicial proceedings, the States must respond by means which THE FLORIDA STAR v. B. J. F. 535 524 Opinion of the Court avoid public documentation or other exposure of private information”).8 A second consideration undergirding the Daily Mail principle is the fact that punishing the press for its dissemination of information which is already publicly available is relatively unlikely to advance the interests in the service of which the State seeks to act. It is not, of course, always the case that information lawfully acquired by the press is known, or accessible, to others. But where the government has made certain information publicly available, it is highly anomalous to sanction persons other than the source of its release. We noted this anomaly in Cox Broadcasting: “By placing the information in the public domain on official court records, the State must be presumed to have concluded that the public interest was thereby being served.” 420 U. S., at 495. The Daily Mail formulation reflects the fact that it is a limited set of cases indeed where, despite the accessibility of the public to certain information, a meaningful public interest is served by restricting its further release by other entities, like the press. As Daily Mail observed in its summary of Oklahoma Publishing, “once the truthful information was ‘publicly revealed’ or ‘in the public domain’ the court could not constitutionally restrain its dissemination.” 443 U. S., at 103. A third and final consideration is the “timidity and selfcensorship” which may result from allowing the media to be punished for publishing certain truthful information. Cox Broadcasting, supra, at 496. Cox Broadcasting noted this concern with overdeterrence in the context of information made public through official court records, but the fear of ex 8 The Daily Mail principle does not settle the issue whether, in cases where information has been acquired unlawfully by a newspaper or by a source, government may ever punish not only the unlawful acquisition, but the ensuing publication as well. This issue was raised but not definitively resolved in New York Times Co. v. United States, 403 U. S. 713 (1971), and reserved in Landmark Communications, 435 U. S., at 837. We have no occasion to address it here. 536 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. cessive media self-suppression is applicable as well to other information released, without qualification, by the government. A contrary rule, depriving protection to those who rely on the government’s implied representations of the lawfulness of dissemination, would force upon the media the onerous obligation of sifting through government press releases, reports, and pronouncements to prune out material arguably unlawful for publication. This situation could inhere even where the newspaper’s sole object was to reproduce, with no substantial change, the government’s rendition of the event in question. Applied to the instant case, the Daily Mail principle clearly commands reversal. The first inquiry is whether the newspaper “lawfully obtain[ed] truthful information about a matter of public significance.” 443 U. S., at 103. It is undisputed that the news article describing the assault on B. J. F. was accurate. In addition, appellant lawfully obtained B. J. F.’s name. Appellee’s argument to the contrary is based on the fact that under Florida law, police reports which reveal the identity of the victim of a sexual offense are not among the matters of “public record” which the public, by law, is entitled to inspect. Brief for Appellee 17-18, citing Fla. Stat. § 119.07(3)(h) (1983). But the fact that state officials are not required to disclose such reports does not make it unlawful for a newspaper to receive them when furnished by the government. Nor does the fact that the Department apparently failed to fulfill its obligation under § 794.03 not to “cause or allow to be . . . published” the name of a sexual offense victim make the newspaper’s ensuing receipt of this information unlawful. Even assuming the Constitution permitted a State to proscribe receipt of information, Florida has not taken this step. It is, clear, furthermore, that the news article concerned “a matter of public significance,” 443 U. S., at 103, in the sense in which the Daily Mail synthesis of prior cases used that term. That is, the article generally, as opposed to the specific identity contained within it, involved a THE FLORIDA STAR v. B. J. F. 537 524 Opinion of the Court matter of paramount public import: the commission, and investigation, of a violent crime which had been reported to authorities. See Cox Broadcasting, supra (article identifying victim of rape-murder); Oklahoma Publishing Co. n. Oklahoma County District Court, 430 U. S. 308 (1977) (article identifying juvenile alleged to have committed murder); Daily Mail, supra (same); cf. Landmark Communications, Inc. v. Virginia, 435 U. S. 829 (1978) (article identifying judges whose conduct was being investigated). The second inquiry is whether imposing liability on appellant pursuant to § 794.03 serves “a need to further a state interest of the highest order.” Daily Mail, 443 U. S., at 103. Appellee argues that a rule punishing publication furthers three closely related interests: the privacy of victims of sexual offenses; the physical safety of such victims, who may be targeted for retaliation if their names become known to their assailants; and the goal of encouraging victims of such crimes to report these offenses without fear of exposure. Brief for Appellee 29-30. At a time in which we are daily reminded of the tragic reality of rape, it is undeniable that these are highly significant interests, a fact underscored by the Florida Legislature’s explicit attempt to protect these interests by enacting a criminal statute prohibiting much dissemination of victim identities. We accordingly do not rule out the possibility that, in a proper case, imposing civil sanctions for publication of the name of a rape victim might be so overwhelmingly necessary to advance these interests as to satisfy the Daily Mail standard. For three independent reasons, however, imposing liability for publication under the circumstances of this case is too precipitous a means of advancing these interests to convince us that there is a “need” within the meaning of the Daily Mail formulation for Florida to take this extreme step. Cf. Landmark Communications, supra (invalidating penalty °n publication despite State’s expressed interest in non 538 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. dissemination, reflected in statute prohibiting unauthorized divulging of names of judges under investigation). First is the manner in which appellant obtained the identifying information in question. As we have noted, where the government itself provides information to the media, it is most appropriate to assume that the government had, but failed to utilize, far more limited means of guarding against dissemination than the extreme step of punishing truthful speech. That assumption is richly borne out in this case. B. J. F.’s identity would never have come to light were it not for the erroneous, if inadvertent, inclusion by the Department of her full name in an incident report made available in a pressroom open to the public. Florida’s policy against disclosure of rape victims’ identities, reflected in §794.03, was undercut by the Department’s failure to abide by this policy. Where, as here, the government has failed to police itself in disseminating information, it is clear under Cox Broadcasting, Oklahoma Publishing, and Landmark Communications that the imposition of damages against the press for its subsequent publication can hardly be said to be a narrowly tailored means of safeguarding anonymity. See supra, at 534-535. Once the government has placed such information in the public domain, “reliance must rest upon the judgment of those who decide what to publish or broadcast,” Cox Broadcasting, 420 U. S., at 496, and hopes for restitution must rest upon the willingness of the government to compensate victims for their loss of privacy and to protect them from the other consequences of its mishandling of the information which these victims provided in confidence. That appellant gained access to the information in question through a government news release makes it especially likely that, if liability were to be imposed, self-censorship would result. Reliance on a news release is a paradigmatically “routine newspaper reporting techniqu[e].” Daily Mail, supra, at 103. The government’s issuance of such a release, without qualification, can only convey to recipients that the THE FLORIDA STAR v. B. J. F. 539 524 Opinion of the Court government considered dissemination lawful, and indeed expected the recipients to disseminate the information further. Had appellant merely reproduced the news release prepared and released by the Department, imposing civil damages would surely violate the First Amendment. The fact that appellant converted the police report into a news story by adding the linguistic connecting tissue necessary to transform the report’s facts into full sentences cannot change this result. A second problem with Florida’s imposition of liability for publication is the broad sweep of the negligence per se standard applied under the civil cause of action implied from §794.03. Unlike claims based on the common-law tort of invasion of privacy, see Restatement (Second) of Torts § 652D (1977), civil actions based on §794.03 require no case-by-case findings that the disclosure of a fact about a person’s private life was one that a reasonable person would find highly offensive. On the contrary, under the per se theory of negligence adopted by the courts below, liability follows automatically from publication. This is so regardless of whether the identity of the victim is already known throughout the community; whether the victim has voluntarily called public attention to the offense; or whether the identity of the victim has otherwise become a reasonable subject of public concern—because, perhaps, questions have arisen whether the victim fabricated an assault by a particular person. Nor is there a scienter requirement of any kind under § 794.03, engendering the perverse result that truthful publications challenged pursuant to this cause of action are less protected by the First Amendment than even the least protected defamatory falsehoods: those involving purely private figures, where liability is evaluated under a standard, usually applied by a jury, of ordinary negligence. See Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974). We have previously noted the impermissibility of categorical prohibitions upon media access where important First Amendment interests are at stake. See Globe 540 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Newspaper Co. n. Superior Court of Norfolk County, 457 U. S. 596, 608 (1982) (invalidating state statute providing for the categorical exclusion of the public from trials of sexual offenses involving juvenile victims). More individualized adjudication is no less indispensable where the State, seeking to safeguard the anonymity of crime victims, sets its face against publication of their names. Third, and finally, the facial underinclusiveness of § 794.03 raises serious doubts about whether Florida is, in fact, serving, with this statute, the significant interests which appellee invokes in support of affirmance. Section 794.03 prohibits the publication of identifying information only if this information appears in an “instrument of mass communication,” a term the statute does not define. Section 794.03 does not prohibit the spread by other means of the identities of victims of sexual offenses. An individual who maliciously spreads word of the identity of a rape victim is thus not covered, despite the fact that the communication of such information to persons who live near, or work with, the victim may have consequences as devastating as the exposure of her name to large numbers of strangers. See Tr. of Oral Arg. 49-50 (appellee acknowledges that §794.03 would not apply to “the backyard gossip who tells 50 people that don’t have to know”). When a State attempts the extraordinary measure of punishing truthful publication in the name of privacy, it must demonstrate its commitment to advancing this interest by applying its prohibition evenhandedly, to the smalltime disseminator as well as the media giant. Where important First Amendment interests are at stake, the mass scope of disclosure is not an acceptable surrogate for injury. A ban on disclosures effected by “instrument[s] of mass communication simply cannot be defended on the ground that partial prohibitions may effect partial relief. See Daily Mail, 443 U. S., at 104-105 (statute is insufficiently tailored to interest in protecting anonymity where it restricted only newspapers, not THE FLORIDA STAR v. B. J. F. 541 524 Opinion of Scalia, J. the electronic media or other forms of publication, from identifying juvenile defendants); id., at 110 (Rehnquist, J., concurring in judgment) (same); cf. Arkansas Writers’ Project, Inc. v. Ragland, 481 U. S. 221, 229 (1987); Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U. S. 575, 585 (1983). Without more careful and inclusive precautions against alternative forms of dissemination, we cannot conclude that Florida’s selective ban on publication by the mass media satisfactorily accomplishes its stated purpose.9 Ill Our holding today is limited. We do not hold that truthful publication is automatically constitutionally protected, or that there is no zone of personal privacy within which the State may protect the individual from intrusion by the press, or even that a State may never punish publication of the name of a victim of a sexual offense. We hold only that where a newspaper publishes truthful information which it has lawfully obtained, punishment may lawfully be imposed, if at all, only when narrowly tailored to a state interest of the highest order, and that no such interest is satisfactorily served by imposing liability under § 794.03 to appellant under the facts of this case. The decision below is therefore Reversed. Justice Scalia, concurring in part and concurring in the judgment. I think it sufficient to decide this case to rely upon the third ground set forth in the Court’s opinion, ante, at 540 and this page: that a law cannot be regarded as protecting an in 9 Having concluded that imposing liability on appellant pursuant to § 794.03 violates the First Amendment, we have no occasion to address appellant’s subsidiary arguments that the imposition of punitive damages for publication independently violated the First Amendment, or that § 794.03 functions as an impermissible prior restraint. See Smith v. Daily Mail Publishing Co., 443 U. S. 97, 101-102 (1979). 542 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. terest “of the highest order,” Smith v. Daily Mail Publishing Co., 443 U. S. 97, 103 (1979), and thus as justifying a restriction upon truthful speech, when it leaves appreciable damage to that supposedly vital interest unprohibited. In the present case, I would anticipate that the rape victim’s discomfort at the dissemination of news of her misfortune among friends and acquaintances would be at least as great as her discomfort at its publication by the media to people to whom she is only a name. Yet the law in question does not prohibit the former in either oral or written form. Nor is it at all clear, as I think it must be to validate this statute, that Florida’s general privacy law would prohibit such gossip. Nor, finally, is it credible that the interest meant to be served by the statute is the protection of the victim against a rapist still at large—an interest that arguably would extend only to mass publication. There would be little reason to limit a statute with that objective to rape alone; or to extend it to all rapes, whether or not the felon has been apprehended and confined. In any case, the instructions here did not require the jury to find that the rapist was at large. This law has every appearance of a prohibition that society is prepared to impose upon the press but not upon itself. Such a prohibition does not protect an interest “of the highest order.” For that reason, I agree that the judgment of the court below must be reversed. Justice White, with whom The Chief Justice and Justice O’Connor join, dissenting. “Short of homicide, [rape] is the ‘ultimate violation of self.’” Coker v. Georgia, 433 U. S. 584, 597 (1977) (opinion of White, J.). For B. J. F., however, the violation she suffered at a rapist’s knifepoint marked only the beginning of her ordeal. A week later, while her assailant was still at large, an account of this assault—identifying by name B. J. F. as the victim—was published by The Florida Star. As a result, B. J. F. received harassing phone calls, required mental health counseling, was forced to move from THE FLORIDA STAR v. B. J. F. 543 524 White, J., dissenting her home, and was even threatened with being raped again. Yet today, the Court holds that a jury award of $75,000 to compensate B. J. F. for the harm she suffered due to the Star’s negligence is at odds with the First Amendment. I do not accept this result. The Court reaches its conclusion based on an analysis of three of our precedents and a concern with three particular aspects of the judgment against appellant. I consider each of these points in turn, and then consider some of the larger issues implicated by today’s decision. I The Court finds its result compelled, or at least supported in varying degrees, by three of our prior cases: Cox Broadcasting Corp. v. Cohn, 420 U. S. 469 (1975); Oklahoma Publishing Co. v. Oklahoma County District Court, 430 U. S. 308 (1977); and Smith v. Daily Mail Publishing Co., 443 U. S. 97 (1979). I disagree. None of these cases requires the harsh outcome reached today. Cox Broadcasting reversed a damages award entered against a television station, which had obtained a rape victim’s name from public records maintained in connection with the judicial proceedings brought against her assailants. While there are similarities, critical aspects of that case make it wholly distinguishable from this one. First, in Cox Broadcasting, the victim’s name had been disclosed in the hearing where her assailants pleaded guilty; and, as we recognized, judicial records have always been considered public information in this country. See Cox Broadcasting, supra, at 492-493. In fact, even the earliest notion of privacy rights exempted the information contained in judicial records from its protections. See Warren & Brandeis, The Right to Privacy, 4 Harv. L. Rev. 193, 216-217 (1890). Second, unlike the incident report at issue here, which was meant by state law to be withheld from public release, the judicial pro 544 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. ceedings at issue in Cox Broadcasting were open as a matter of state law. Thus, in Cox Broadcasting, the state-law scheme made public disclosure of the victim’s name almost inevitable; here, Florida law forbids such disclosure. See Fla. Stat. 794.03 (1987). These facts—that the disclosure came in judicial proceedings, which were open to the public—were critical to our analysis in Cox Broadcasting. The distinction between that case and this one is made obvious by the penultimate paragraph of Cox Broadcasting: “We are reluctant to embark on a course that would make public records generally available to the media but would forbid their publication if offensive .... [T]he First and Fourteenth Amendments will not allow exposing the press to liability for truthfully publishing information released to the public in official court records. If there are privacy interests to be protected in judicial proceedings, the States must respond by means which avoid public documentation or other exposure of private information. . . . Once true information is disclosed in public court documents open to public inspection, the press cannot be sanctioned for publishing it.” Cox Broadcasting, supra, at 496 (emphasis added). Cox Broadcasting stands for the proposition that the State cannot make the press its first line of defense in withholding private information from the public—it cannot ask the press to secrete private facts that the State makes no effort to safeguard in the first place. In this case, however, the State has undertaken “means which avoid [but obviously, not altogether prevent] public documentation or other exposure of private information.” No doubt this is why the Court frankly admits that “Cox Broadcasting . . . cannot fairly be read as controlling here.” Ante, at 532. Finding Cox Broadcasting inadequate to support its result, the Court relies on Smith v. Daily Mail Publishing Co. as its THE FLORIDA STAR v. B. J. F. 545 524 White, J., dissenting principal authority.1 But the flat rule from Daily Mail on which the Court places so much reliance—“[I]f a newspaper lawfully obtains truthful information . . . then state officials may not constitutionally punish publication of the information, absent a need to further a state interest of the highest order”—was introduced in Daily Mail with the cautious qualifier that such a rule was “suggest[ed]” by our prior cases, “[n]one of [which] . . . directly controlled]” in Daily Mail. See Daily Mail, 443 U. S., at 103. The rule the Court takes as a given was thus offered only as a hypothesis in Daily Mail: it should not be so uncritically accepted as constitutional dogma. More importantly, at issue in Daily Mail was the disclosure of the name of the perpetrator of an infamous murder of a 15-year-old student. Id., at 99. Surely the rights of those accused of crimes and those who are their victims must differ with respect to privacy concerns. That is, whatever rights alleged criminals have to maintain their anonymity pending an adjudication of guilt—and after Daily Mail, those rights would seem to be minimal—the rights of crime victims to stay shielded from public view must be infinitely more substantial. Daily Mail was careful to state that the “holding in this case is narrow .... there is no issue here of privacy.” Id., at 105 (emphasis added). But in this case, there is an issue of privacy—indeed, that is the principal issue—and therefore, this case falls outside of Daily Mail’s “rule” ‘The second case in the “trilogy” which the Court cites is Oklahoma Publishing Co. v. Oklahoma County District Court, 430 U. S. 308 (1977). See ante, at 530-531. But not much reliance is placed on that case, and I do not discuss it with the degree of attention devoted to Cox Broadcasting or Daily Mail. As for the support Oklahoma Publishing allegedly provides for the Court’s result here, the reasons that distinguish Cox Broadcasting and Daily Mail from this case are even more apt in the case of Oklahoma Publishing. Probably that is why the Court places so little weight on this middle leg of the three. 546 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. (which, as I suggest above, was perhaps not even meant as a rule in the first place). Consequently, I cannot agree that Cox Broadcasting, or Oklahoma Publishing, or Daily Mail requires—or even substantially supports—the result reached by the Court today. II We are left, then, to wonder whether the three “independent reasons” the Court cites for reversing the judgment for B. J. F. support its result. See ante, at 537-541. The first of these reasons relied on by the Court is the fact “appellant gained access to [B. J. F.’s name] through a government news release.” Ante, at 538. “The government’s issuance of such a release, without qualification, can only convey to recipients that the government considered dissemination lawful,” the Court suggests. Ante, at 538-539. So described, this case begins to look like the situation in Oklahoma Publishing, where a judge invited reporters into his courtroom, but then tried to prohibit them from reporting on the proceedings they observed. But this case is profoundly different. Here, the “release” of information provided by the government was not, as the Court says, “without qualification.” As the Star’s own reporter conceded at trial, the crime incident report that inadvertently included B. J. F.’s name was posted in a room that contained signs making it clear that the names of rape victims were not matters of public record, and were not to be published. See 2 Record 113, 115, 117. The Star’s reporter indicated that she understood that she “[was not] allowed to take down that information” (i. e., B. J. F.’s name) and that she “[was] not supposed to take the information from the police department.” Id., at 117. Thus, by her own admission the posting of the incident report did not convey to the Star’s reporter the idea that “the government considered dissemination lawful”; the Courts suggestion to the contrary is inapt. THE FLORIDA STAR v. B. J. F. 547 524 White, J., dissenting Instead, Florida has done precisely what we suggested, in Cox Broadcasting, that States wishing to protect the privacy rights of rape victims might do: “respond [to the challenge] by means which avoid public documentation or other exposure of private information.” 420 U. S., at 496 (emphasis added). By amending its public records statute to exempt rape victims names from disclosure, Fla. Stat. § 119.07(3)(h) (1983), and forbidding its officials to release such information, Fla. Stat. § 794.03 (1983), the State has taken virtually every step imaginable to prevent what happened here. This case presents a far cry, then, from Cox Broadcasting or Oklahoma Publishing, where the State asked the news media not to publish information it had made generally available to the public: here, the State is not asking the media to do the State’s job in the first instance. Unfortunately, as this case illustrates, mistakes happen: even when States take measures to “avoid” disclosure, sometimes rape victims’ names are found out. As I see it, it is not too much to ask the press, in instances such as this, to respect simple standards of decency and refrain from publishing a victim’s name, address, and/or phone number.2 2 The Court’s concern for a free press is appropriate, but such concerns should be balanced against rival interests in a civilized and humane society. An absolutist view of the former leads to insensitivity as to the latter. This was evidenced at trial, when the Florida Star’s lawyer explained why the paper was not to blame for any anguish caused B. J. F. by a phone call she received, the day after the Star’s story was published, from a man threatening to rape B. J. F. again. Noting that the phone call was received at B. J. F.’s home by her mother (who was babysitting B. J. F.’s children while B. J. F. was in the hospital), who relayed the threat to J. F., the Star’s counsel suggested: [I]n reference to the [threatening] phone call, it is sort of blunted by the fact that [B. J. F.] didn’t receive the phone call. Her mother did. And if there is any pain and suffering in connection with the phone call, it has to lay in her mother’s hands. I mean, my God, she called [B. J. F.] up at the hospital to tell her [of the threat] —you know, I think that is tragic, but I 548 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. Second, the Court complains that appellant was judged here under too strict a liability standard. The Court contends that a newspaper might be found liable under the Florida courts’ negligence per se theory without regard to a newspaper’s scienter or degree of fault. Ante, at 539-540. The short answer to this complaint is that whatever merit the Court’s argument might have, it is wholly inapposite here, where the jury found that appellant acted with “reckless indifference towards the rights of others,” 2 Record 170, a standard far higher than the Gertz standard the Court urges as a constitutional minimum today. Ante, at 539-540. B. J. F. proved the Star’s negligence at trial—and, actually, far more than simple negligence; the Court’s concerns about damages resting on a strict liability or mere causation basis are irrelevant to the validity of the judgment for appellee. But even taking the Court’s concerns in the abstract, they miss the mark. Permitting liability under a negligence per se theory does not mean that defendants will be held liable without a showing of negligence, but rather, that the standard of care has been set by the legislature, instead of the courts. The Court says that negligence per se permits a plaintiff to hold a defendant liable without a showing that the disclosure was “of a fact about a person’s private life . .. that a reasonable person would find highly offensive.” Ante, at 539. But the point here is that the legislature—reflecting popular sentiment—has determined that disclosure of the fact that a person was raped is categorically a revelation that reasonable people find offensive. And as for the Court’s suggestion that the Florida courts’ theory permits liability without regard for whether the victim’s identity is already don’t think that is something you can blame the Florida Star for.” 2 Record 154-155. While I would not want to live in a society where freedom of the press was unduly limited, I also find regrettable an interpretation of the First Amendment that fosters such a degree of irresponsibility on the part of the news media. THE FLORIDA STAR v. B. J. F. 549 524 White, J., dissenting known, or whether she herself has made it known—these are facts that would surely enter into the calculation of damages in such a case. In any event, none of these mitigating factors was present here; whatever the force of these arguments generally, they do not justify the Court’s ruling against B. J. F. in this case. Third, the Court faults the Florida criminal statute for being under inclusive: §794.03 covers disclosure of rape victim’s names in “‘instruments] of mass communication,’” but not other means of distribution, the Court observes. Ante, at 540. But our cases which have struck down laws that limit or burden the press due to their underinclusiveness have involved situations where a legislature has singled out one segment of the news media or press for adverse treatment, see, e. g., Daily Mail (restricting newspapers and not radio or television), or singled out the press for adverse treatment when compared to other similarly situated enterprises, see, e. g., Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U. S. 575, 578 (1983). Here, the Florida law evenhandedly covers all “instrument[s] of mass communication” no matter their form, media, content, nature, or purpose. It excludes neighborhood gossips, cf. ante, at 540, because presumably the Florida Legislature has determined that neighborhood gossips do not pose the danger and intrusion to rape victims that “instrument[s] of mass communication” do. Simply put: Florida wanted to prevent the widespread distribution of rape victims’ names, and therefore enacted a statute tailored almost as precisely as possible to achieving that end. Moreover, the Court’s “underinclusiveness” analysis itself is “underinclusive.” After all, the lawsuit against the Star which is at issue here is not an action for violating the statute which the Court deems under inclusive, but is, more accurately, for the negligent publication of appellee’s name. See App. to Juris. Statement A10. The scheme which the Court should review, then, is not only §794.03 (which, as 550 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. noted above, merely provided the standard of care in this litigation), but rather, the whole of Florida privacy tort law. As to the latter, Florida does recognize a tort of publication of private facts.3 Thus, it is quite possible that the neighborhood gossip whom the Court so fears being left scot free to spread news of a rape victim’s identity would be subjected to the same (or similar) liability regime under which appellant was taxed. The Court’s myopic focus on §794.03 ignores the probability that Florida law is more comprehensive than the Court gives it credit for being. Consequently, neither the State’s “dissemination” of B. J. F.’s name, nor the standard of liability imposed here, nor the underinclusiveness of Florida tort law requires setting aside the verdict for B. J. F. And as noted above, such a result is not compelled by our cases. I turn, therefore, to the more general principles at issue here to see if they recommend the Court’s result. Ill At issue in this case is whether there is any information about people, which—though true—may not be published in the press. By holding that only “a state interest of the highest order” permits the State to penalize the publication of truthful information, and by holding that protecting a rape victim’s right to privacy is not among those state interests of the highest order, the Court accepts appellant’s invitation, see Tr. of Oral Arg. 10-11, to obliterate one of the most noteworthy legal inventions of the 20th century: the tort of the publication of private facts. W. Prosser, J. Wade, & V. Schwartz, Torts 951-952 (8th ed. 1988). Even if the Court’s opinion does not say as much today, such obliteration will follow inevitably from the Court’s conclusion here. If the First Amendment prohibits wholly private per 3 See, e. g., Cape Publications, Inc. v. Hitchner, 514 So. 2d 1136, 1137-1138 (Fla. App. 1987); Loft n. Fuller, 408 So. 2d 619, 622 (Fla. App. 1981). THE FLORIDA STAR v. B. J. F. 551 524 White, J., dissenting sons (such as B. J. F.) from recovering for the publication of the fact that she was raped, I doubt that there remain any “private facts” which persons may assume will not be published in the newspapers or broadcast on television.4 Of course, the right to privacy is not absolute. Even the article widely relied upon in cases vindicating privacy rights, Warren & Brandeis, The Right to Privacy, 4 Harv. L. Rev. 193 (1890), recognized that this right inevitably conflicts with the public’s right to know about matters of general concern— and that sometimes, the latter must trump the former. Id., at 214-215. Resolving this conflict is a difficult matter, and I fault the Court not for attempting to strike an appropriate balance between the two, but rather, fault it for according too little weight to B. J. F.’s side of equation, and too much on the other. 4 The consequences of the Court’s ruling—that a State cannot prevent the publication of private facts about its citizens which the State inadvertently discloses—is particularly troubling when one considers the extensive powers of the State to collect information. One recent example illustrates this point. In Boettger v. Loverro, 521 Pa. 366, 555 A. 2d 1234 (1989), police officers had lawfully “tapped” the telephone of a man suspected of bookmaking. Under Pennsylvania law transcripts of the conversations intercepted this way may not be disclosed. 18 Pa. Cons. Stat. § 5703 (1988). Another statute imposes civil liability on any person who “discloses” the content of tapped conversations. § 5725. Nonetheless, in a preliminary court hearing, a prosecutor inadvertently attached a transcript of the phone conversations to a document filed with the court. A reporter obtained a copy of the transcript due to this error, and his paper published a version of the remarks disclosed by the telephone tap. On appeal, the Supreme Court of Pennsylvania upheld a civil liability award of $1,000 against the paper for its unlawful disclosure of the contents of the phone conversations, concluding that individuals’ rights to privacy outweighed the interest in public disclosure of such private telephone communications. Boettger, supra, at 376-377, 555 A. 2d, at 1239-1240. The Court’s decision today suggests that this ruling by the Pennsylvania court was erroneous. In light of the substantial privacy interest in such communications, though, cf. Katz v. United States, 389 U. S. 347 (1967), I would strike the balance as the Pennsylvania Supreme Court did. 552 OCTOBER TERM, 1988 White, J., dissenting 491 U. S. I would strike the balance rather differently. Writing for the Ninth Circuit, Judge Merrill put this view eloquently: “Does the spirit of the Bill of Rights require that individuals be free to pry into the unnewsworthy private affairs of their fellowmen? In our view it does not. In our view, fairly defined areas of privacy must have the protection of law if the quality of life is to continue to be reasonably acceptable. The public’s right to know is, then, subject to reasonable limitations so far as concerns the private facts of its individual members.” Virgil v. Time, Inc., 527 F. 2d 1122, 1128 (1975), cert, denied, 425 U. S. 998 (1976). Ironically, this Court, too, had occasion to consider this same balance just a few weeks ago, in United States Department of Justice n. Reporters Committee for Freedom of Press, 489 U. S. 749 (1989). There, we were faced with a press request, under the Freedom of Information Act, for a “rap sheet” on a person accused of bribing a Congressman— presumably, a person whose privacy rights would be far less than B. J. F.’s. Yet this Court rejected the media’s request for disclosure of the “rap sheet,” saying: “The privacy interest in maintaining the practical obscurity of rap-sheet information will always be high. When the subject of such a rap sheet is a private citizen and when the information is in the Government’s control as a compilation, rather than as a record of ‘what the government is up to,’ the privacy interest... is ... at its apex while the . . . public interest in disclosure is at its nadir.’ Id., at 780. The Court went on to conclude that disclosure of rap sheets “categorical[ly]” constitutes an “unwarranted” invasion of privacy. Ibid. The same surely must be true—indeed, much more so—for the disclosure of a rape victim’s name. I do not suggest that the Court’s decision today is a radical departure from a previously charted course. The Courts THE FLORIDA STAR v. B. J. F. 553 524 White, J., dissenting ruling has been foreshadowed. In Time, Inc. v. Hill, 385 U. S. 374, 383-384, n. 7 (1967), we observed that—after a brief period early in this century where Brandeis’ view was ascendant—the trend in “modern” jurisprudence has been to eclipse an individual’s right to maintain private any truthful information that the press wished to publish. More recently, in Cox Broadcasting, 420 U. S. at 491, we acknowledged the possibility that the First Amendment may prevent a State from ever subjecting the publication of truthful but private information to civil liability. Today, we hit the bottom of the slippery slope. I would find a place to draw the line higher on the hillside: a spot high enough to protect B. J. F.’s desire for privacy and peace-of-mind in the wake of a horrible personal tragedy. There is no public interest in publishing the names, addresses, and phone numbers of persons who are the victims of crime—and no public interest in immunizing the press from liability in the rare cases where a State’s efforts to protect a victim’s privacy have failed. Consequently, I respectfully dissent.5 ’The Court does not address the distinct constitutional questions raised by the award of punitive damages in this case. Ante, at 541, n. 9. Consequently, I do not do so either. That award is more troublesome than the compensatory award discussed above. Cf. Note, Punitive Damages and Libel Law, 98 Harv. L. Rev. 847 (1985). 554 OCTOBER TERM, 1988 Syllabus 491 U. S. UNITED STATES v. ZOLIN et al. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT No. 88-40. Argued March 20, 1989—Decided June 21, 1989 The Internal Revenue Service (IRS), as part of its investigation of the tax returns of L. Ron Hubbard, founder of the Church of Scientology (the Church), filed in the Federal District Court a petition to enforce a summons it had served upon the Clerk of the Los Angeles County Superior Court demanding that he produce documents, including two tapes, in his possession in conjunction with a pending suit. The Church and Mary Sue Hubbard, intervenors in the state-court action and respondents here, intervened to oppose production of the materials. They claimed, inter alia, that the IRS was not seeking the materials in good faith and that the attorney-client privilege barred the tapes’ disclosure. The IRS argued, among other things, that the tapes fell within the exception to the attorney-client privilege for communications in furtherance of future illegal conduct—the so-called “crime-fraud” exception—and urged the District Court to listen to the tapes in making its privilege determination. In addition, the IRS submitted a declaration by a special agent which had included partial tape transcripts the IRS lawfully had obtained. The court rejected respondents’ bad-faith claim and ordered production of five of the requested documents, but it conditioned its enforcement order by placing restrictions upon IRS dissemination of the documents. The court also ruled that the tapes need not be produced since they contained privileged attorney-client communications to which, the quoted excerpts revealed, the crime-fraud exception did not apply-The court rejected the request that it listen to the tapes, on the ground that that request had been abandoned in favor of using the agent’s declaration as the basis for determining the privilege question. The Court of Appeals affirmed the conditional-enforcement order. As to the privilege issue, it agreed with respondents that the District Court would have been without power to grant the IRS’ demand for in camera review of the tapes, because the Government’s evidence of crime or fraud must come from sources independent of the attorney-client communications on the tapes. Reviewing the independent evidence (a review that excluded the partial transcripts), the court affirmed the District Court’s determination as to the inapplicability of the crime-fraud exception. UNITED STATES v. ZOLIN 555 554 Syllabus Held: 1. Insofar as it upheld the District Court’s conditional-enforcement order, the Court of Appeals’ judgment is affirmed by an equally divided Court. P. 561. 2. In appropriate circumstances, in camera review of allegedly privileged attorney-client communications may be used to determine whether the communications fall within the crime-fraud exception. Pp. 562-575. (a) Federal Rule of Evidence 104(a), which provides that a court is bound by the rules of evidence with respect to privileges when determining the existence of a privilege, does not prohibit the use of in camera review. Pp. 565-570. (b) However, before a district court may engage in in camera review at the request of the party opposing the privilege, that party must present evidence sufficient to support a reasonable belief that such review may reveal evidence that establishes the exception’s applicability. Once this threshold showing is made, the decision whether to engage in in camera review rests in the sound discretion of the court. Pp. 570-572. (c) The party opposing the privilege may use any relevant nonprivileged evidence, lawfully obtained, to meet the threshold showing, even if its evidence is not “independent” of the contested communications as the Court of Appeals uses that term. Pp. 573-574. (d) On remand, the Court of Appeals should consider whether the District Court’s refusal to listen to the tapes in toto was justified by the manner in which the IRS presented and preserved its in camera review request. If its demand was properly preserved, that court, or the District Court on remand, should determine whether the IRS has presented a sufficient evidentiary basis for in camera review and whether it is appropriate for the District Court, in its discretion, to grant the request. Pp. 574-575. 809 F. 2d 1411, 842 F. 2d 1135, and 850 F. 2d 610, affirmed in part, vacated in part, and remanded. Blackmun, J., delivered the opinion of the Court, in which all other Members joined, except Brennan, J., who took no part in the consideration or decision of the case. Alan I. Horowitz argued the cause for the United States. With him on the briefs were Solicitor General Fried, Acting Solicitor General Bryson, Assistant Attorney General Rose, Deputy Solicitor General Wallace, Charles E. Brookhart, and John A. Dudeck, Jr. 556 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Michael Lee Hertzberg argued the cause for respondents. With him on the brief were Eric M. Lieberman and David Golove. * Justice Blackmun delivered the opinion of the Court. This case arises out of the efforts of the Criminal Investigation Division of the Internal Revenue Service (IRS) to investigate the tax returns of L. Ron Hubbard, founder of the Church of Scientology (the Church), for the calendar years 1979 through 1983. We granted certiorari, 488 U. S. 907 (1988), to consider two issues that have divided the Courts of Appeals. The first is whether, when a district court enforces an IRS summons, see 26 U. S. C. §7604, the court may condition its enforcement order by placing restrictions on the disclosure of the summoned information.1 The Court of Appeals in this case upheld the restrictions. We affirm its judgment on that issue by an equally divided Court. The second issue concerns the testimonial privilege for attorney-client communications and, more particularly, the generally recognized exception to that privilege for communications in furtherance of future illegal conduct—the so-called “crime-fraud” exception. The specific question presented is whether the applicability of the crime-fraud exception must be established by “independent evidence” (i. e., without reference to the content of the contested communications themselves), or, alternatively, whether the applicability of that exception can be resolved by an in camera inspection of the allegedly privileged material.2 We reject the “independent evidence” approach and hold that the district court, under *Edward D. Urquhart, Silvia T. Hassell, and Charles J. Escher filed a brief for Bernard M. Barrett, Jr., M. D., as amicus curiae. 'Compare United States v. Author Services, Inc., 804 F. 2d 1520, 1525— 1526 (CA9 1986), opinion amended, 811 F. 2d 1264 (1987), with United States v. Barrett, 837 F. 2d 1341 (CA5 1988) (en banc), cert, pending, No. 87-1705. 2 Compare United States v. Shewfelt, 455 F. 2d 836 (CA9), cert, denied, 406 U. S. 944 (1972), with In re Berkley & Co., 629 F. 2d 548 (CA8 1980). UNITED STATES v. ZOLIN 557 554 Opinion of the Court circumstances we explore below, and at the behest of the party opposing the claim of privilege, may conduct an in camera review of the materials in question. Because the Court of Appeals considered only “independent evidence,” we vacate its judgment on this issue and remand the case for further proceedings.3 I In the course of its investigation, the IRS sought access to 51 documents that had been filed with the Clerk of the Los Angeles County Superior Court in connection with a case entitled Church of Scientology of California v. Armstrong, No. C420 153. The Armstrong litigation involved, among other things, a charge by the Church that one of its former members, Gerald Armstrong, had obtained by unlawful means documentary materials relating to Church activities, including two tapes. Some of the documents sought by the IRS had been filed under seal. The IRS, by its Special Agent Steven Petersell, served a summons upon the Clerk on October 24, 1984, pursuant to 26 U. S. C. §7603, demanding that he produce the 51 documents.4 The tapes were among those listed. App. 33-38. On November 21, IRS agents were permitted to inspect and copy some of the summoned materials, including the tapes. On November 27, the Church and Mary Sue Hubbard, who had intervened in Armstrong, secured a temporary restrain 3 Respondents suggest that this case is now moot, because L. Ron Hubbard died January 24, 1986, thus foreclosing any further criminal investigation of him, and because the IRS civil audit of Mr. Hubbard for the relevant tax years was terminated as a “‘closed case.’” Brief in Opposition 8-10. The IRS disagrees, largely because the civil tax audit has not been terminated, and its result could affect the liability of Mr. Hubbard’s estate. We are satisfied that a live controversy remains. 4 The current Clerk of the Superior Court, Frank S. Zolin, is a named respondent in this case, but did not participate in briefing or argument before the Court of Appeals or before this Court. We use the term “respondents” to refer to Mary Sue Hubbard and the Church, the only active respondents in this Court. 558 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ing order from the United States District Court for the Central District of California. The order required the IRS to file with the District Court all materials acquired on November 21 and all reproductions and notes related thereto, pending disposition of the intervenors’ motion for a preliminary injunction to bar IRS use of these materials. Exh. 2 to Petition to Enforce Internal Revenue Summons. By order dated December 10, the District Court returned to the IRS all materials except the tapes and the IRS’ notes reflecting their contents. See App. 30. On January 18, 1985, the IRS filed in the District Court a petition to enforce its summons. In addition to the tapes, the IRS sought 12 sealed documents the Clerk had refused to produce in response to the IRS summons. The Church and Mary Sue Hubbard intervened to oppose production of the tapes and the sealed documents. Respondents claimed that IRS was not seeking the documents in good faith, and objected on grounds of lack of relevance and attorney-client privilege. Respondents asserted the privilege as a bar to disclosure of the tapes. The IRS argued, among other things, however, that the tapes fell within the crime-fraud exception to the attorney-client privilege, and urged the District Court to listen to the tapes in the course of making its privilege determination. In addition, the IRS submitted to the court two declarations by Agent Petersell. In the first, Petersell stated his grounds for believing that the tapes were relevant to the investigation. See Declaration in No. CV85-0440-HLH, 113 (March 8, 1985). In the second, Petersell offered a description of the tapes’ contents, based on information he received during several interviews. Appended to this declaration—over respondents’ objection—were partial transcripts of the tapes, which the IRS lawfully had obtained from a confidential source. See March 15, 1985, declaration UNITED STATES v. ZOLIN 559 554 Opinion of the Court (filed under seal).5 In subsequent briefing, the IRS reiterated its request that the District Court listen to the tapes in camera before making its privilege ruling. After oral argument and an evidentiary hearing, the District Court rejected respondents’ claim of bad faith. App. to Pet. for Cert. 27a. The court ordered production of 5 of the 12 documents, id., at 28a, and specified: “The documents delivered hereunder shall not be delivered to any other government agency by the IRS unless criminal tax prosecution is sought or an Order of Court is obtained.” Id., at 29a. Turning to the tapes, the District Court ruled that respondents had demonstrated that they contain confidential attorney-client communications, that the privilege had not been waived, and that “[t]he ‘fraud-crime’ exception to the attorney-client privilege does not apply. The quoted excerpts tend to show or admit past fraud but there is no clear indication that future fraud or crime is being planned.” Id., at 28a. On this basis, the court held that the Clerk “need not produce its copy of the tapes pursuant to the summons.” Id., at 29a. The District Court denied the IRS’ motion for reconsideration, rejecting the IRS’ renewed request that the court listen to the tapes in toto. “While this was at one time discussed with counsel, thereafter Mr. Petersell’s declaration was submitted, and no one suggested that this ’The IRS denied that the transcripts were made using tapes obtained from the Superior Court or from any other illicit source. Agent Petersell declared: “The partial transcripts were not prepared by the United States from the tapes in the custody of the Superior Court for Los Angeles County, California, nor from copies of the tape now in the custody of the Clerk of this Court. The transcripts were obtained from a confidential source by another Special Agent prior to the issuance of this summons. The source was not a party to Church of Scientology v. Armstrong, No. 410153, nor an attorney for any party in that proceeding.” See Declaration of Agent Petersell in No. CV85-0440-HLH (Tx) (March 21,1985). As the District Court made no finding of illegality, we assume for present purposes that the transcripts were legally obtained. 560 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. was an inadequate basis on which to determine the attorneyclient privilege question.” Id., at 25a-26a. Respondents appealed to the Court of Appeals for the Ninth Circuit, and the IRS cross-appealed on two relevant grounds. First, the IRS claimed that the District Court abused its discretion by placing conditions on the IRS’ future use of the subpoenaed information. The Court of Appeals disagreed, holding: “A district court may, when appropriate, condition enforcement of a summons on the IRS’ agreeing to abide by disclosure restrictions.” 809 F. 2d 1411, 1417 (1987). Second, the IRS contended that the District Court erred in rejecting the application of the crime-fraud exception to the tapes. In particular, the IRS argued that the District Court incorrectly held that the IRS had abandoned its request for in camera review of the tapes, and that the court should have listened to the tapes before ruling that the crime-fraud exception was inapplicable. Answering Brief for United States as Appellee in No. 85-6065, and Opening Brief for United States as Cross-Appellant in No. 85-6105 (CA9), pp. 48-49 (filed under seal). Respondents contended, in contrast, that the District Court erred in the opposite direction: they argued that it was error for the court to rely on the partial transcripts, because “[i]n this Circuit, a party cannot rely on the communications themselves—whether by listening to the tapes or reviewing excerpts or transcripts of them—to bear its burden to invoke the exception but must bear the burden by independent evidence. This is the clear and unambiguous holding of United States v. Shewfelt, 455 F. 2d 836 (9th Cir.), cert, denied, 406 U. S. 944 (1972).” (Emphasis added.) Answering Brief for Church of Scientology of California and Mary Sue Hubbard as Cross-Appellees in No. 85-6065, and Reply Brief as Appellants in No. 85-6105 (CA9), p. 24 (filed under seal). The panel of the Court of Appeals agreed with respondents that, under Shewfelt, “the Government’s evidence of crime or UNITED STATES v. ZOLIN 561 554 Opinion of the Court fraud must come from sources independent of the attorneyclient communications recorded on the tapes,” 809 F. 2d, at 1418, thereby implicitly holding that even if the IRS had properly preserved its demand for in camera review, the District Court would have been without power to grant it. The Court of Appeals then reviewed “the Government’s independent evidence.” Id., at 1418-1419. That review appears to have excluded the partial transcripts, and thus the Court of Appeals implicitly agreed with respondents that it was improper for the District Court to have considered even the partial transcripts. See Brief for United States 7. On the basis of its review of the “independent evidence,” the Court of Appeals affirmed the District Court’s determination that the IRS had failed to establish the applicability of the crime-fraud exception. 809 F. 2d, at 1419. The full Court of Appeals vacated the panel opinion and ordered en banc review, on the basis of a perceived conflict between Shewfelt and United States v. Friedman, 445 F. 2d 1076 (CA9), cert, denied sub nom. Jacobs v. United States, 404 U. S. 958 (1971). 832 F. 2d 127 (1987). Upon consideration, a majority of the limited en banc court, see Ninth Circuit Rule 35-3, determined that the intracircuit conflict was illusory; it agreed with respondents that Friedman did not address the independent-evidence rule. 842 F. 2d 1135, 1136, amended by 850 F. 2d 610 (1988). The limited en banc court vacated the order for rehearing en banc as improvi-dently granted and reinstated the panel opinion in relevant part. Ibid. II This Court is evenly divided with respect to the issue of the power of a district court to place restrictions upon the dissemination by the IRS of information obtained through a § 7604 subpoena-enforcement action. We therefore affirm the judgment of the Court of Appeals insofar as it upheld the District Court’s conditional-enforcement order. 562 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. HI Questions of privilege that arise in the course of the adjudication of federal rights are “governed by the principles of the common law as they may be interpreted by the courts of the United States in the light of reason and experience.” Fed. Rule Evid. 501. We have recognized the attorneyclient privilege under federal law, as “the oldest of the privileges for confidential communications known to the common law.” Upjohn Co. n. United States, 449 U. S. 383, 389 (1981). Although the underlying rationale for the privilege has changed over time, see 8 J. Wigmore, Evidence §2290 (McNaughton rev. 1961),6 courts long have viewed its central concern as one “to encourage full and frank communication between attorneys and their clients and thereby promote broader public interests in the observance of law and administration of justice.” Upjohn, 449 U. S., at 389. That purpose, of course, requires that clients be free to “make full disclosure to their attorneys” of past wrongdoings, Fisher v. United States, 425 U. S. 391, 403 (1976), in order that the client may obtain “the aid of persons having knowledge of the law and skilled in its practice,” Hunt v. Blackburn, 128 U. S. 464, 470 (1888). The attorney-client privilege is not without its costs. Cf. Trammel v. United States, 445 U. S. 40, 50 (1980). “[S]ince the privilege has the effect of withholding relevant information from the factfinder, it applies only where necessary to achieve its purpose.” Fisher, 425 U. S., at 403. The attorney-client privilege must necessarily protect the confidences of wrongdoers, but the reason for that protection—the centrality of open client and attorney communication to the proper functioning of our adversary system of justice—“ceas[es] to operate at a certain point, namely, where the desired advice refers not to prior wrongdoing, but fiSee also Hazard, An Historical Perspective on the Attorney-Client Privilege, 66 Calif. L. Rev. 1061 (1978); Developments in the Law—Privileged Communications, 98 Harv. L. Rev. 1450, 1455-1458 (1985). UNITED STATES v. ZOLIN 563 554 Opinion of the Court to future wrongdoing.” 8 Wigmore, §2298, p. 573 (emphasis in original); see also Clark v. United States, 289 U. S. 1, 15 (1933). It is the purpose of the crime-fraud exception to the attorney-client privilege to assure that the “seal of secrecy,” ibid., between lawyer and client does not extend to communications “made for the purpose of getting advice for the commission of a fraud” or crime. O'Rourke n. Darbishire, [1920] A. C. 581, 604 (P. C.). The District Court and the Court of Appeals found that the tapes at issue in this case recorded attorney-client communications and that the privilege had not been waived when the tapes were inadvertently given to Armstrong. 809 F. 2d, at 1417 (noting that Armstrong had acquired the tapes from L. Ron Hubbard’s personal secretary, who was under the mistaken impression that the tapes were blank). These findings are not at issue here. Thus, the remaining obstacle to respondents’ successful assertion of the privilege is the Government’s contention that the recorded attorney-client communications were made in furtherance of a future crime or fraud. A variety of questions may arise when a party raises the crime-fraud exception. The parties to this case have not been in complete agreement as to which of these questions are presented here. In an effort to clarify the matter, we observe, first, that we need not decide the quantum of proof necessary ultimately to establish the applicability of the crime-fraud exception. Cf. Clark, 289 U. S., at 15, quoting O'Rourke; S. Stone & R. Liebman, Testimonial Privileges §1.65, p. 107 (1983).7 Rather, we are concerned here with ' We note, however, that this Court’s use in Clark v. United States, 289 U. S. 1, 14 (1933), of the phrase “prima facie case” to describe the showing needed to defeat the privilege has caused some confusion. See Gardner, The Crime or Fraud Exception to the Attorney-Client Privilege, 47 A. B. A. J. 708, 710-711 (1961); Note, 51 Brooklyn L. Rev. 913, 918-919 (1985) (“The prima facie standard is commonly used by courts in civil litigation to shift the burden of proof from one party to the other. In the context of the fraud exception, however, the standard is used to dispel the privilege altogether without affording the client an opportunity to rebut 564 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the type of evidence that may be used to make that ultimate showing. Within that general area of inquiry, the initial question in this case is whether a district court, at the request of the party opposing the privilege, may review the allegedly privileged communications in camera to determine whether the crime-fraud exception applies.8 If such in camera review is permitted, the second question we must consider is whether some threshold evidentiary showing is needed before the district court may undertake the requested the prima facie showing” (emphasis in original)). See also In re Grand Jury Subpoena Duces Tecum Dated September 15, 1983, 731 F. 2d 1032, 1039 (CA2 1984). In using the phrase in Clark, the Court was aware of scholarly controversy concerning the role of the judge in the decision of such preliminary questions of fact. See 289 U. S., at 14, n. The quantum of proof needed to establish admissibility was then, and remains, subject to question. See, e. g., Maguire & Epstein, Preliminary Questions of Fact in Determining the Admissibility of Evidence, 40 Harv. L. Rev. 392, 400 (criticizing courts insofar as they “have allowed themselves to be led into holding that only a superficial, one-sided showing is allowable on any admissibility controversy”), 414-424 (exploring alternative rules) (1927); 21 C. Wright & K. Graham, Federal Practice and Procedure: Evidence § 5052, p. 248 (1977) (suggesting, with respect to the process of proving preliminary questions of fact, that “[p]erhaps it is a task, like riding a bicycle, that is easier to do if you do not think too much about what you are doing”). In light of the narrow question presented here for review, this case is not the proper occasion to visit these questions. 8 In addition, the facts of this case also suggest the question whether the partial transcripts the IRS possessed may be used by it in meeting its ultimate burden. It is by no means clear that the Government has presented that question for this Court’s review. The Government noted in its petition for certiorari that the Court of Appeals had not considered the partial transcripts in making its determination that the IRS had failed to establish the applicability of the crime-fraud exception. See Pet. for Cert. 7-8. The question presented for review, however, relates solely to in camera review, as does the relevant discussion in the petition. See id., at 20-23. The question whether the partial transcripts may be used in meeting the IRS’ ultimate burden of demonstrating the applicability of the crime-fraud exception is fairly included within the question presented, however, and we therefore address it. See this Court’s Rule 21.1(a). The answer to the question would follow inexorably from our discussion in any event. UNITED STATES v. ZOLIN 565 554 Opinion of the Court review. Finally, if a threshold showing is required, we must consider the type of evidence the opposing party may use to meet it: i. e., in this case, whether the partial transcripts the IRS possessed may be used for that purpose. A We consider first the question whether a district court may ever honor the request of the party opposing the privilege to conduct an in camera review of allegedly privileged communications to determine whether those communications fall within the crime-fraud exception. We conclude that no express provision of the Federal Rules of Evidence bars such use of in camera review, and that it would be unwise to prohibit it in all instances as a matter of federal common law.9 (1) At first blush, two provisions of the Federal Rules of Evidence would appear to be relevant. Rule 104(a) provides: “Preliminary questions concerning the qualification of a person to be a witness, the existence of a privilege, or the admissibility of evidence shall be determined by the court .... In making its determination it is not bound by the rules of evidence except those with respect to privileges. ” (Emphasis added.) Rule 1101(c) provides: “The rule with respect to 9 There is some ambiguity as to whether the Court of Appeals squarely barred all use of in camera review for these purposes, although that is the fairest reading of the court’s opinion. Respondents at times appear to advocate that position, see Brief in Opposition 19-21, but at times suggest otherwise, see Brief for Respondents 13; see also Reply Brief for United States 15. The ambiguity in respondents’ position is perhaps due to the fact that they accept the premise that in camera review is permitted under Circuit precedent in different circumstances from those at issue in this case—i. e., where the proponent of the privilege seeks in camera review to demonstrate the applicability of the privilege in the first instance, see Brief for Respondents 14, or when the proponent requests in camera review to ensure that an order requiring production of some materials held not to be privileged does not inadvertently yield privileged information, see id., at 20-21. 566 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. privileges applies at all stages of all actions, cases, and proceedings.” Taken together, these Rules might be read to establish that in a summons-enforcement proceeding, attorneyclient communications cannot be considered by the district court in making its crime-fraud ruling: to do otherwise, under this view, would be to make the crime-fraud determination without due regard to the existence of the privilege. Even those scholars who support this reading of Rule 104(a) acknowledge that it leads to an absurd result. “Because the judge must honor claims of privilege made during his preliminary fact determinations, many exceptions to the rules of privilege will become ‘dead letters,’ since the preliminary facts that give rise to these exceptions can never be proved. For example, an exception to the attorney-client privilege provides that there is no privilege if the communication was made to enable anyone to commit a crime or fraud. There is virtually no way in which the exception can ever be proved, save by compelling disclosure of the contents of the communication; Rule 104(a) provides that this cannot be done.” 21 C. Wright & K. Graham, Federal Practice & Procedure: Evidence §5055, p. 276 (1977) (footnote omitted). We find this Draconian interpretation of Rule 104(a) inconsistent with the Rule’s plain language. The Rule does not provide by its terms that all materials as to which a “clai[m] of privilege” is made must be excluded from consideration. In that critical respect, the language of Rule 104(a) is markedly different from the comparable California evidence rule, which provides that “the presiding officer may not require disclosure of information claimed to be privileged under this division in order to rule on the claim of privilege. Cal. Evid. Code Ann. § 915(a) (West Supp. 1989) (emphasis UNITED STATES v. ZOLIN 567 554 Opinion of the Court added).10 There is no reason to read Rule 104(a) as if its text were identical to that of the California rule. Nor does it make sense to us to assume, as respondents have throughout this litigation, that once the attorney-client nature of the contested communications is established, those communications must be treated as presumptively privileged for evidentiary purposes until the privilege is “defeated” or “stripped away” by proof that the communications took place in the course of planning future crime or fraud. See Brief for Respondents 15 (asserting that respondents had “established their entitlement to the privilege,” and that the communications had been “determined to be privileged,” before the crime-fraud question was resolved). Although some language in Clark might be read as supporting this view, see 289 U. S., at 15, respondents acknowledged at oral argument that no prior holding of this Court requires the imposition of a strict progression of proof in crime-fraud cases. See Tr. of Oral Arg. 33-35. 10 A good example of the effect of the California rule is provided by the record in this case. While the disputed matters were being briefed in Federal District Court, the State Superior Court held a hearing on a motion by Government attorneys seeking access to materials in the Armstrong case for ongoing litigation in Washington, D. C. The transcript of the hearing was made part of the record before the District Court in this case. Regarding the tapes, the Government argued to the Superior Court that the attorney-client conversations on the tapes reflect the planning or commission of a crime or fraud. Tr. of Hearing of February 11, 1985, in No. C420 153 (Super. Ct. Cal.), p. 52. That claim was supported by several declarations and other extrinsic evidence. The Government noted, however, that the tape recordings themselves would ... be the best evidence of exactly what was going on.” Id., at 53. The intervenors stressed that, as a matter of California law, “you can’t show the tapes are not privileged by the contents.” Id., at 58; see also id., at 68. The Superior Court acknowledged the premise that “you can’t look at the conversation itself to make [the crime-fraud] determination,” id., at 74, and concluded that the extrinsic evidence was not sufficient to make out a prima facie case that the crime-fraud exception applies, id., at 75-76. 568 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. We see no basis for holding that the tapes in this case must be deemed privileged under Rule 104(a) while the question of crime or fraud remains open. Indeed, respondents concede that “if the proponent of the privilege is able to sustain its burden only by submitting the communications to the court” for in camera review, Brief for Respondents 14-15 (emphasis in original), the court is not required to avert its eyes (or close its ears) once it concludes that the communication would be privileged, if the court found the crime-fraud exception inapplicable. Rather, respondents acknowledge that the court may “then consider the same communications to determine if the opponent of the privilege has established that the crime-fraud exception applies.” Id., at 15. Were the tapes truly deemed privileged under Rule 104(a) at the moment the trial court concludes they contain potentially privileged attorneyclient communications, district courts would be required to draw precisely the counterintuitive distinction that respondents wisely reject. We thus shall not adopt a reading of Rule 104(a) that would treat the contested communications as “privileged” for purposes of the Rule, and we shall not interpret Rule 104(a) as categorically prohibiting the party opposing the privilege on crime-fraud grounds from relying on the results of an in camera review of the communications. (2) Having determined that Rule 104(a) does not prohibit the in camera review sought by the 1RS, we must address the question as a matter of the federal common law of privileges. See Rule 501. We conclude that a complete prohibition against opponents’ use of in camera review to establish the applicability of the crime-fraud exception is inconsistent with the policies underlying the privilege. We begin our analysis by recognizing that disclosure of allegedly privileged materials to the district court for purposes of determining the merits of a claim of privilege does not have the legal effect of terminating the privilege. Indeed, this UNITED STATES v. ZOLIN 569 554 Opinion of the Court Court has approved the practice of requiring parties who seek to avoid disclosure of documents to make the documents available for in camera inspection, see Kerr v. United States District Court for Northern District of Cal., 426 U. S. 394, 404-405 (1976), and the practice is well established in the federal courts. See, e. g., In re Antitrust Grand Jury, 805 F. 2d 155, 168 (CA6 1986); In re Vargas, 723 F. 2d 1461, 1467 (CAIO 1983); United States v. Lawless, 709 F. 2d 485, 486, 488 (CA7 1983); In re Grand Jury Witness, 695 F. 2d 359, 362 (CA9 1982). Respondents do not dispute this point: they acknowledge that they would have been free to request in camera review to establish the fact that the tapes involved attorney-client communications, had they been unable to muster independent evidence to serve that purpose. Brief for Respondents 14-15. Once it is clear that in camera review does not destroy the privileged nature of the contested communications, the question of the propriety of that review turns on whether the policies underlying the privilege and its exceptions are better fostered by permitting such review or by prohibiting it. In our view, the costs of imposing an absolute bar to consideration of the communications in camera for purpose of establishing the crime-fraud exception are intolerably high. “No matter how light the burden of proof which confronts the party claiming the exception, there are many blatant abuses of privilege which cannot be substantiated by extrinsic evidence. This is particularly true . . . of . . . situations in which an alleged illegal proposal is made in the context of a relationship which has an apparent legitimate end.” Note, The Future Crime or Tort Exception to Communications Privileges, 77 Harv. L. Rev. 730, 737 (1964). A per se rule that the communications in question may never be considered creates, we feel, too great an impediment to the proper functioning of the adversary process. See generally 2 D. Louisell & C. Mueller, Federal Evidence § 213, pp. 828-829 (1985); 2 J. Weinstein & M. Berger, Weinstein’s Evi 570 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. dence 51503(d)(l)[01], p. 503-71 (1988). This view is consistent with current trends in the law. Compare National Conference of Commissioners on Uniform State Laws, Uniform Rules of Evidence, Rule 26(2)(a) (1953 ed.) (“Such privileges shall not extend ... to a communication if the judge finds that sufficient evidence, aside from the communication, has been introduced to warrant a finding that the legal service was sought or obtained in order to enable or aid the client to commit or plan to commit a crime or a tort” (emphasis added)), reprinted in 1 J. Bailey & 0. Trelles, The Federal Rules of Evidence: Legislative Histories and Related Documents (1980), with Uniform Rule of Evidence 502 (adopted 1974), 13A U. L. A. 256 (1986) (omitting explicit independent evidence requirement). B We turn to the question whether in camera review at the behest of the party asserting the crime-fraud exception is always permissible, or, in contrast, whether the party seeking in camera review must make some threshold showing that such review is appropriate. In addressing this question, we attend to the detrimental effect, if any, of in camera review on the policies underlying the privilege and on the orderly administration of justice in our courts. We conclude that some such showing must be made. Our endorsement of the practice of testing proponents’ privilege claims through in camera review of the allegedly privileged documents has not been without reservation. This Court noted in United States v. Reynolds, 345 U. S. 1 (1953), a case which presented a delicate question concerning the disclosure of military secrets, that “examination of the evidence, even by the judge alone, in chambers” might in some cases “jeopardize the security which the privilege is meant to protect.” Id., at 10. Analogizing to claims of Fifth Amendment privilege, it observed more generally: “Too much judicial inquiry into the claim of privilege would force disclosure of the thing the privilege was meant to pro UNITED STATES v. ZOLIN 571 554 Opinion of the Court tect, while a complete abandonment of judicial control would lead to intolerable abuses.” Id., at 8. The Court in Reynolds recognized that some compromise must be reached. See also United States v. Weisman, 111 F. 2d 260, 261-262 (CA2 1940). In Reynolds, it declined to “go so far as to say that the court may automatically require a complete disclosure to the judge before the claim of privilege will be accepted in any case. ” 345 U. S., at 10 (emphasis added). We think that much the same result is in order here. A blanket rule allowing in camera review as a tool for determining the applicability of the crime-fraud exception, as Reynolds suggests, would place the policy of protecting open and legitimate disclosure between attorneys and clients at undue risk. There is also reason to be concerned about the possible due process implications of routine use of in camera proceedings. See, e. g., In re John Doe Corp., 675 F. 2d 482, 489-490 (CA2 1982); In re Special September 1978 Grand Jury, 640 F. 2d 49, 56-58 (CA7 1980). Finally, we cannot ignore the burdens in camera review places upon the district courts, which may well be required to evaluate large evidentiary records without open adversarial guidance by the parties. There is no reason to permit opponents of the privilege to engage in groundless fishing expeditions, with the district courts as their unwitting (and perhaps unwilling) agents. Courts of Appeals have suggested that in camera review is available to evaluate claims of crime or fraud only “when justified,” In re John Doe Corp., 675 F. 2d, at 490, or “[i]n appropriate cases,” In re Sealed Case, 219 U. S. App. D. C. 195, 217, 676 F. 2d 793, 815 (1982) (opinion of Wright, J.). Indeed, the Government conceded at oral argument (albeit reluctantly) that a district court would be mistaken if it reviewed documents in camera solely because “the government beg[ged it]” to do so, “with no reason to suspect crime or fraud.” Tr. of Oral Arg. 26; see also id., at 60. We agree. 572 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. In fashioning a standard for determining when in camera review is appropriate, we begin with the observation that “in camera inspection ... is a smaller intrusion upon the confidentiality of the attorney-client relationship than is public disclosure.” Fried, Too High a Price for Truth: The Exception to the Attorney-Client Privilege for Contemplated Crimes and Frauds, 64 N. C. L. Rev. 443, 467 (1986). We therefore conclude that a lesser evidentiary showing is needed to trigger in camera review than is required ultimately to overcome the privilege. Ibid. The threshold we set, in other words, need not be a stringent one. We think that the following standard strikes the correct balance. Before engaging in in camera review to determine the applicability of the crime-fraud exception, “the judge should require a showing of a factual basis adequate to support a good faith belief by a reasonable person,” Caldwell v. District Court, 644 P. 2d 26, 33 (Colo. 1982), that in camera review of the materials may reveal evidence to establish the claim that the crime-fraud exception applies. Once that showing is made, the decision whether to engage in in camera review rests in the sound discretion of the district court. The court should make that decision in light of the facts and circumstances of the particular case, including, among other things, the volume of materials the district court has been asked to review, the relative importance to the case of the alleged privileged information, and the likelihood that the evidence produced through in camera review, together with other available evidence then before the court, will establish that the crime-fraud exception does apply. The district court is also free to defer its in camera review if it concludes that additional evidence in support of the crime-fraud exception may be available that is not allegedly privileged, and that production of the additional evidence will not unduly disrupt or delay the proceedings. UNITED STATES v. ZOLIN 573 554 Opinion of the Court c The question remains as to what kind of evidence a district court may consider in determining whether it has the discretion to undertake an in camera review of an allegedly privileged communication at the behest of the party opposing the privilege. Here, the issue is whether the partial transcripts may be used by the IRS in support of its request for in camera review of the tapes. The answer to that question, in the first instance, must be found in Rule 104(a), which establishes that materials that have been determined to be privileged may not be considered in making the preliminary determination of the existence of a privilege. Neither the District Court nor the Court of Appeals made factual findings as to the privileged nature of the partial transcripts,11 so we cannot determine on this record whether Rule 104(a) would bar their consideration. Assuming for the moment, however, that no rule of privilege bars the IRS’ use of the partial transcripts, we fail to see what purpose would be served by excluding the transcripts from the District Court’s consideration. There can be little doubt that partial transcripts, or other evidence directly but incompletely reflecting the content of the contested communications, generally will be strong evidence of the subject matter of the communications themselves. Permitting district courts to consider this type of evidence would aid them substantially in rapidly and reliably determining whether in camera review is appropriate. 11 There are no findings as to whether respondents themselves would be privileged to resist a demand that they produce the partial transcripts. Nor has there been any legal and factual exploration of whether respondents may claim privilege as a bar to the IRS’ use of the copy of the transcripts it lawfully obtained from a third party. See, e. g., Developments in the Law—Privileged Communications, 98 Harv. L. Rev., at 1648, 1660-1661 (discussing controversy concerning the legal effect of an inadvertent disclosure which does not constitute a waiver of the privilege, and citing cases); 8 Wigmore § 2326. 574 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Respondents suggest only one serious countervailing consideration. In their view, a rule that would allow an opponent of the privilege to rely on such material would encourage litigants to elicit confidential information from disaffected employees or others who have access to the information. Tr. of Oral Arg. 40-41. We think that deterring the aggressive pursuit of relevant information from third-party sources is not sufficiently central to the policies of the attorney-client privilege to require us to adopt the exclusionary rule urged by respondents. We conclude that the party opposing the privilege may use any nonprivileged evidence in support of its request for in camera review, even if its evidence is not “independent” of the contested communications as the Court of Appeals uses that term.12 D In sum, we conclude that a rigid independent evidence requirement does not comport with “reason and experience,” Fed. Rule Evid. 501, and we decline to adopt it as part of the developing federal common law of evidentiary privileges. We hold that in camera review may be used to determine whether allegedly privileged attorney-client communications fall within the crime-fraud exception. We further hold, however, that before a district court may engage in in camera review at the request of the party opposing the privilege, that party must present evidence sufficient to support a reasonable belief that in camera review may yield evidence that 12 In addition, we conclude that evidence that is not “independent” of the contents of allegedly privileged communications—like the partial transcripts in this case—may be used not only in the pursuit of in camera review, but also may provide the evidentiary basis for the ultimate showing that the crime-fraud exception applies. We see little to distinguish these two uses: in both circumstances, if the evidence has not itself been determined to be privileged, its exclusion does not serve the policies which underlie the attorney-client privilege. See generally Note, The Future Crime or Tort Exception to Communications Privileges, 77 Harv. L. Rev. 730, 737 (1964). UNITED STATES v. ZOLIN 575 554 Opinion of the Court establishes the exception’s applicability. Finally, we hold that the threshold showing to obtain in camera review may be met by using any relevant evidence, lawfully obtained, that has not been adjudicated to be privileged. Because the Court of Appeals employed a rigid independent-evidence requirement which categorically excluded the partial transcripts and the tapes themselves from consideration, we vacate its judgment on this issue and remand the case for further proceedings consistent with this opinion. On remand, the Court of Appeals should consider whether the District Court’s refusal to listen to the tapes in toto was justified by the manner in which the IRS presented and preserved its request for in camera review.13 In the event the Court of Appeals holds that the IRS’ demand for review was properly preserved, the Court of Appeals should then determine, or remand the case to the District Court to determine in the first instance, whether the IRS has presented a sufficient evidentiary basis for in camera review, and whether, if so, it is appropriate for the District Court, in its discretion, to grant such review. It is so ordered. Justice Brennan took no part in the consideration or decision of this case. 13 The Court of Appeals also will have the opportunity to review the partial transcripts, and to determine whether, even without in camera review of the tapes, the IRS presented sufficient evidence to establish that the tapes are within the crime-fraud exception. 576 OCTOBER TERM, 1988 Syllabus 491 U. S. MASSACHUSETTS v. OAKES CERTIORARI TO THE SUPREME JUDICIAL COURT OF MASSACHUSETTS No. 87-1651. Argued January 17, 1989—Decided June 21, 1989 In 1984, respondent Oakes took color photographs of his partially nude and physically mature 14-year-old stepdaughter, L. S. He was indicted, tried, and convicted of violating a Massachusetts statute (§ 29A) prohibiting adults from posing or exhibiting minors “in a state of nudity” for purposes of visual representation or reproduction in any publication, motion picture, photograph, or picture. The Massachusetts Supreme Judicial Court reversed the conviction. After holding that Oakes’ posing of L. S. was speech for First Amendment purposes, the court struck down the statute as substantially overbroad under the First Amendment without addressing whether § 29A could be constitutionally applied to Oakes. It concluded that § 29A criminalized conduct that virtually every person would regard as lawful, such as the taking of family photographs of nude infants. Subsequently, § 29A was amended to add a “lascivious intent” requirement to the “nudity” portion of the statute and to eliminate exemptions contained in the prior version. Held: The judgment is vacated, and the case is remanded. 401 Mass. 602, 518 N. E. 2d 836, vacated and remanded. Justice O’Connor, joined by The Chief Justice, Justice White, and Justice Kennedy, concluded that: 1. As a practical matter, the intervening amendment of the statute moots the overbreadth question in this case. Thus, overbreadth analysis is inappropriate under Bigelow v. Virginia, 421 U. S. 809. The overbreadth doctrine—an exception to the general rule that a person to whom a statute may be constitutionally applied can no longer challenge the statute on the ground that it may be unconstitutionally applied to others—is designed to prevent the chilling of protected expression, which the former version of § 29A cannot do since it has been repealed. That overbreadth was discussed and rejected as a mode of analysis in Bigelow—where there was no need to comment on that issue since the defendant’s conviction was reversed on the narrower and alternative ground that the statute was unconstitutional as applied—is evidence that the application of Bigelow does not depend on whether other questions presented will be answered adversely to the defendant. It is not constitutionally offensive to decline to reach Oakes’ challenge, since an overbroad statute is not void ab initio but merely voidable. Since the special MASSACHUSETTS v. OAKES 577 576 Syllabus concern that animates the overbreadth doctrine is no longer present, the doctrine’s benefits need not be extended to a defendant whose conduct is not protected. Moreover, the amendment of a state statute pending appeal to eliminate overbreadth is not different, in terms of applying the new law to past conduct, from a state appellate court adopting a limiting construction of a statute to cure overbreadth. This Court has long held in the latter situation that the statute, as construed, may be applied to conduct occurring before the limiting construction. Pp. 581-584. 2. Since the sole issue before this Court has become moot, and a live dispute remains as to whether the former version of § 29A can constitutionally be applied to Oakes, this case is remanded for a determination of that remaining live issue. Pp. 584-585. Justice Scalia, joined by Justice Brennan, Justice Marshall, Justice Blackmun, and Justice Stevens, concluded that the subsequent amendment of §29A to eliminate the basis for the overbreadth challenge does not eliminate the overbreadth defense. The overbreadth doctrine serves to protect constitutionally legitimate speech not only after an offending statute is enacted, but also when a legislature is contemplating what sort of statute to enact. If no conviction of constitutionally proscribable conduct would be lost, so long as the offending statute was narrowed before the final appeal, legislatures would have significantly reduced incentive to stay within constitutional bounds in the first place. Moreover, while this Court has the power to adopt a rule of law which says that the defendant’s acts were lawful because the statute that sought to prohibit them was overbroad and therefore invalid, it does not have the power to pursue the policy underlying that rule by conditioning the defendant’s criminal liability on whether, by the time his last appeal is exhausted, letting him challenge the statute might serve to eliminate any First Amendment “chill.” Pp. 585-588. Justice Scalia, joined by Justice Blackmun, also concluded that the case should be remanded for the court below to dispose of the as-applied challenge, since the statute is not impermissibly overbroad. The scope of this statute has already been validated except as to nonpornographic depictions, New York v. Ferber, 458 U. S. 747, and has been narrowed further by statutory exemptions, and any possibly unconstitutional application of it—for example, to artistic depictions not otherwise exempt or to family photographs—is insubstantial judged in relation to the statute’s plainly legitimate sweep. Pp. 588-590. O’Connor, J., announced the judgment of the Court and delivered an opinion, in which Rehnquist, C. J., and White and Kennedy, JJ., joined. Scalia, J., filed an opinion concurring in the judgment in part and dissenting in part, in which Blackmun, J., joined, and in which Brennan, 578 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. Marshall, and Stevens, JJ., joined as to Part I, post, p. 585. Brennan, J., filed a dissenting opinion, in which Marshall and Stevens, JJ., joined, post, p. 590. James M. Shannon, Attorney General of Massachusetts, argued the cause for petitioner. With him on the briefs were Phyllis N. Segal and A. John Pappalardo, Deputy Attorneys General, and Madelyn F. Wessel, Judy G. Zeprun, and H. Reed Witherby, Assistant Attorneys General. Richard J. Vita argued the cause and filed a brief for respondent. * Justice O’Connor announced the judgment of the Court and delivered an opinion, in which The Chief Justice, Justice White, and Justice Kennedy join. This case involves an overbreadth challenge to a Massachusetts criminal statute generally prohibiting adults from posing or exhibiting nude minors for purposes of visual representation or reproduction in any book, magazine, pamphlet, motion picture, photograph, or picture. I The statute at issue in this case, Mass. Gen. Laws § 272:29A (1986), was enacted in 1982.1 It provides as follows: “Whoever with knowledge that a person is a child under eighteen years of age, or whoever while in posses *Briefs of amici curiae urging reversal were filed for the State of Indiana by Linley E. Pearson, Attorney General, and William E. Daily, Deputy Attorney General; for the District Attorney for the Middle District of the Commonwealth of Massachusetts by John J. Conte, pro se, and Daniel F. Toomey; for Citizens for Decency Through Law, Inc., by Bruce A. Taylor; for Covenant House et al. by Gregory A. Loken; and for the Massachusetts Society for Prevention of Cruelty to Children et al. by Elizabeth K. Spahn. Briefs of amici curiae urging affirmance were filed for the American Sunbathing Association, Inc., by Robert T. Page; and for the Law and Humanities Institute by Edward de Grazia. 1 For background on the enactment of §29A, see Boston Globe, June 14, 1982, p. 17, col. 1; Boston Globe, July 21, 1982, p. 17, col. 2. MASSACHUSETTS v. OAKES 579 576 Opinion of O’Connor, J. sion of such facts that he should have reason to know that such person is a child under eighteen years of age, hires, coerces, solicits or entices, employs, procures, uses, causes, encourages, or knowingly permits such child to pose or be exhibited in a state of nudity or to participate or engage in any live performance or in any act that depicts, describes or represents sexual conduct for purpose of visual representation or reproduction in any book, magazine, pamphlet, motion picture film, photograph, or picture shall be punished by imprisonment in the state prison for a term of not less than ten nor more than twenty years, or by a fine of not less than ten thousand dollars nor more than fifty thousand dollars, or by both such a fine and imprisonment. "It shall be a defense in any prosecution pursuant to this section that such visual representation or reproduction of any posture or exhibition in a state of nudity was produced, processed, published, printed or manufactured for a bona fide scientific or medical purpose, or for an educational or cultural purpose for a bona fide school, museum or library. “As used in this section, the term ‘performance’ shall mean any play, dance or exhibit shown or presented to an audience of one or more persons.” Another statute, Mass. Gen. Laws §272:31 (1986), defines “nudity” as “uncovered or less than opaquely covered post-pubertal human genitals, pubic areas, the post-pubertal human female breast below a point immediately above the top of the areola, or the covered male genitals in a discemibly turgid state. For purposes of this definition, a female breast is considered uncovered if the nipple or the nipple or areola only are covered. In the case of pre-pubertal persons nudity shall mean uncovered or less than opaquely covered pre-pubertal human genitals or pubic area.” 580 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. In 1984, respondent Douglas Oakes took approximately 10 color photographs of his partially nude and physically mature 14-year-old stepdaughter, L. S., who at the time was attending modeling school. Tr. 22-30. The photographs depict L. S. sitting, lying, and reclining on top of a bar, clad only in a red and white striped bikini panty and a red scarf. The scarf does not cover L. S.’s breasts, which are fully exposed in all the photographs. The dissent below described the photographs as “sexually provocative photographs of the type frequently found in magazines displayed by storekeepers in sealed cellophane wrappers.” 401 Mass. 602, 606, 518 N. E. 2d 836, 838 (1988). See also Brief for Law and Humanities Institute as Amicus Curiae 47 (referring to the photographs as “pin-up” art). Oakes was indicted and tried for violating § 29A. The jury returned a general verdict of guilty, and Oakes was sentenced to 10 years’ imprisonment. Because the jury was not instructed on the “sexual conduct” portion of §29A, Tr. 101-104, its verdict rested on a finding that Oakes “hire[d], coerce[d], solicit[ed] or entice[d], employ[ed], procure[d], use[d], cause[d], encourage[d], or knowingly permit[ted]” L. S. to “pose or be exhibited in a state of nudity.” The acts proscribed by § 29A are listed disjunctively, so it is impossible to ascertain which of those acts the jury concluded Oakes had committed. The jury was instructed on the exemptions set forth in § 29A, Tr. 104, but its guilty verdict indicates that the exemptions were found to be inapplicable. A divided Massachusetts Supreme Judicial Court reversed Oakes’ conviction. The majority first held that Oakes’ posing of L. S. was speech for First Amendment purposes because it could not “fairly be isolated” from the “expressive process of taking her picture.” 401 Mass., at 604, 518 N. E. 2d, at 837. Without addressing whether §29A could be constitutionally applied to Oakes, the majority struck down the statute as substantially overbroad under the First Amendment. The majority concluded that §29A “criminalize[d] MASSACHUSETTS v. OAKES 581 576 Opinion of O’Connor, J. conduct that virtually every person would regard as lawful,” and would make “a criminal of a parent who takes a frontal view picture of his or her naked one-year-old running on a beach or romping in a wading pool.” Id., at 605, 518 N. E. 2d, at 838. The dissent argued that Oakes’ conduct did not constitute speech for First Amendment purposes: “Soliciting, causing, or encouraging, or permitting a minor to pose for photographs is no more speech than is setting a house afire in order to photograph a burning house.” Id., at 610, 518 N. E. 2d, at 841. The dissent also argued that even if the “nudity” portion of § 29A was overbroad, that portion should have been severed from the remainder of the statute. Id., at 611, n. 4, 518 N. E. 2d, at 841, n. 4. We granted certiorari to review the decision of the Massachusetts Supreme Judicial Court, 486 U. S. 1022 (1988), and now vacate and remand. II The First Amendment doctrine of substantial overbreadth is an exception to the general rule that a person to whom a statute may be constitutionally applied cannot challenge the statute on the ground that it may be unconstitutionally applied to others. Board of Airport Commers of Los Angeles v. Jews for Jesus, Inc., 482 U. S. 569, 574 (1987); Brockett n. Spokane Arcades, Inc., 472 U. S. 491, 503-504 (1985). See generally Monaghan, Overbreadth, 1981 S. Ct. Rev. 1. The doctrine is predicated on the danger that an overly broad statute, if left in place, may cause persons whose expression is constitutionally protected to refrain from exercising their rights for fear of criminal sanctions. Schaumburg v. Citizens for a Better Environment, 444 U. S. 620, 634 (1980). Overbreadth doctrine has wide-ranging effects, for a statute found to be substantially overbroad is subject to facial invalidation. We have therefore referred to overbreadth as “manifestly] strong medicine” that is employed “sparingly, and only as a last resort.” Broadrick v. Oklahoma, 413 U. S. 601, 613 (1973). 582 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. We have addressed overbreadth only where its effect might be salutary. In Bigelow v. Virginia, 421 U. S. 809 (1975), the defendant argued that the criminal statute under which he was convicted was overbroad. After the defendant was convicted, the statute was amended. The amendment eliminated any possibility that the statute’s former version would “be applied again to [the defendant] or [would] chill the rights of others.” Id., at 817-818. Because, “[a]s a practical matter,” the question of the statute’s “overbreadth ha[d] become moot for the future,” we declined to “rest our decision on overbreadth,” choosing instead to consider whether the former version of the statute had been constitutionally applied to the defendant. Id., at 818. In our view, Bigelow stands for the proposition that overbreadth analysis is inappropriate if the statute being challenged has been amended or repealed. The statute in Bigelow was challenged on both overbreadth and as-applied grounds. There was no need for any comment on the overbreadth challenge, as the defendant’s conviction could have been—and indeed was—reversed on a narrower and alternative ground, i. e., that the statute was unconstitutional as applied. See id., at 829. That overbreadth was discussed and rejected as a mode of analysis is, we think, evidence that application of Bigelow does not depend on whether other questions presented will be answered adversely to the defendant. Indeed, the Bigelow overbreadth analysis appears to have been based on the argument made by the State that the amendment of the statute being challenged eliminated the “justification for the application of the overbreadth doctrine.” Brief for Appellee in Bigelow n. Virginia, 0. T. 1974, No. 73-1309, p. 19, n. 10. The procedural posture of the overbreadth question in this case is indistinguishable from that in Bigelow. After we granted certiorari, §29 A was amended. See 1988 Mass. Acts, ch. 226. The current version of §29A, which is set MASSACHUSETTS v. OAKES 583 576 Opinion of O’Connor, J. forth in the margin,2 adds a “lascivious intent” requirement to the “nudity” portion, but not the “sexual conduct” portion, of the former version of § 29A. In addition, the current version of §29A contains no exemptions. Because it has been repealed, the former version of §29A cannot chill protected expression in the future. Thus, as in Bigelow, the over 2 The current version of §29A, codified at Mass. Gen. Laws §272:29A (Supp. 1988), provides: “(a) Whoever, either with knowledge that a person is a child under eighteen years or while in possession of such facts that he should have reason to know that such person is a child under eighteen years of age, and with lascivious intent, hires, coerces, solicits, or entices, employs, procures, uses, causes, encourages, or knowingly permits such child to pose or be exhibited in a state of nudity, for the purpose of representation or reproduction in any visual material, shall be punished by imprisonment in the state prison for a term of not less than ten nor more than twenty years, or by a fine of not less than ten thousand nor more than fifty thousand dollars, or by both such fine and imprisonment. “(b) Whoever, either with knowledge that a person is a child under eighteen years of age or while in possession of such facts that he should have reason to know that such person is a child under eighteen years of age, hires, coerces, solicits or entices, employs, procures, uses, causes, encourages, or knowingly permits such child to participate or engage in any act that depicts, describes, or represents sexual conduct for the purpose of representation or reproduction in any visual material, or to engage in any live performance involving sexual conduct, shall be punished by imprisonment in the state prison for a term of not less than ten nor more than twenty years, or by a fine of not less than ten thousand nor more than fifty thousand dollars, or by both such fine and imprisonment. “(c) In a prosecution under this section, a minor shall be deemed incapable of consenting to any conduct of the defendant for which said defendant is being prosecuted. “(d) For purposes of this section, the determination whether the person m any visual material prohibited hereunder is under eighteen years of age may be made by the personal testimony of such person, by the testimony of a person who produced, processed, published, printed or manufactured such visual material that the child therein was known to him to be under eighteen years of age, or by expert testimony as to the age of the person based upon the person’s physical appearance, by inspection of the visual material, or by any other method authorized by any general or special law or by any applicable rule of evidence.” 584 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. breadth question in this case has become moot as a practical matter, and we do not address it. There is nothing constitutionally offensive about declining to reach Oakes’ overbreadth challenge. Overbreadth is a judicially created doctrine designed to prevent the chilling of protected expression. An overbroad statute is not void ab initio, but rather voidable, subject to invalidation notwithstanding the defendant’s unprotected conduct out of solicitude to the First Amendment rights of parties not before the court. Because the special concern that animates the overbreadth doctrine is no longer present after the amendment or repeal of the challenged statute, we need not extend the benefits of the doctrine to a defendant whose conduct is not protected. See Pope n. Illinois, 481 U. S. 497, 501-502 (1987) (“Facial invalidation” of a repealed statute “would not serve the purpose of preventing future prosecutions under a constitutionally defective standard”). Cf. Upper Midwest Booksellers Assn. v. Minneapolis, 602 F. Supp. 1361, 1369 (Minn.) (amendment of ordinance rendered overbreadth challenge moot, but no conviction involved), aff’d, 780 F. 2d 1389 (CA8 1985). We also note that the amendment of a statute pending appeal to eliminate overbreadth is not different, in terms of applying the new law to past conduct, from a state appellate court adopting a limiting construction of a statute to cure overbreadth. We have long held that in such situations the statute, as construed, “may be applied to conduct occurring prior to the construction, provided such application affords fair warning to the defendants.” Dombrowski v. Pfister, 380 U. S. 479, 491, n. 7 (1965) (citations omitted). See also Broadrick v. Oklahoma, 413 U. S., at 613 (“Facial overbreadth has not been invoked when a limiting construction has been or could be placed on the challenged statute”). Ill Massachusetts has not asked us to consider Oakes’ as-applied challenge to the former version of §29A in its peti MASSACHUSETTS v. OAKES 585 576 Opinion of Scalia, J. tion for certiorari, and we took the case to decide the overbreadth question alone. When the sole question on which we granted certiorari has become moot, our usual course, in cases coming to us from state courts when part of the dispute remains alive, is to vacate the judgment below and remand for further proceedings. See DeFunis v. Odegaard, 416 U. S. 312 (1974). We have dismissed state court cases rather than vacate and remand them, but only in situations where no state or federal claim remained once the particular claim before us became moot, thereby making a remand unnecessary. See Attorney General of New Jersey v. First Family Mortgage Corp, of Florida, 487 U. S. 1213 (1988) (underlying mortgage foreclosure dispute ended because debt was satisfied); Michigan v. Shabaz, 478 U. S. 1017 (1986) (respondent died); Tiverton Board of License Commers n. Pastore, 469 U. S. 238 (1985) (respondent went out of business and no longer had any claim to press); Aikens v. California, 406 U. S. 813 (1972) (petitioner obtained complete relief under state constitution before federal constitutional claim was decided); Ditson n. California, 372 U. S. 933 (1963) (petitioner executed before petition for certiorari was acted upon). Here, a live dispute remains as to whether the former version of § 29A can constitutionally be applied to Oakes. Thus, we vacate the judgment below and remand for further proceedings. Vacated and remanded. Justice Scalia, with whom Justice Blackmun joins, and with whom Justice Brennan, Justice Marshall, and Justice Stevens join as to Part I, concurring in the judgment in part and dissenting in part. I I do not agree with Justice O’Connor’s conclusion that the overbreadth defense is unavailable when the statute alleged to run afoul of that doctrine has been amended to eliminate the 586 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. basis for the overbreadth challenge. It seems to me strange judicial theory that a conviction initially invalid can be resuscitated by postconviction alteration of the statute under which it was obtained. Indeed, I would even think it strange judicial theory that an act which is lawful when committed (because the statute that proscribes it is overbroad) can become retroactively unlawful if the statute is amended preindictment. Of course the reason we are tempted to create such curiosities is that the overbreadth doctrine allows a defendant to attack a statute because of its effect on conduct other than the conduct for which the defendant is being punished, thus protecting the right to engage in conduct not directly before the court. See Brockett n. Spokane Arcades, Inc., 472 U. S. 491, 503 (1985). And the argument is made that it is senseless to apply this doctrine when the protection of other conduct can no longer be achieved, which is the case when the statute has already been amended to eliminate any unconstitutional “chilling” of First Amendment rights. Even as a policy argument, this analysis fails. The overbreadth doctrine serves to protect constitutionally legitimate speech not merely ex post, that is, after the offending statute is enacted, but also ex ante, that is, when the legislature is contemplating what sort of statute to enact. If the promulgation of overbroad laws affecting speech was cost free, as Justice O’Connor’s new doctrine would make it—that is, if no conviction of constitutionally proscribable conduct would be lost, so long as the offending statute was narrowed before the final appeal—then legislatures would have significantly reduced incentive to stay within constitutional bounds in the first place. When one takes account of those overbroad statutes that are never challenged, and of the time that elapses before the ones that are challenged are amended to come within constitutional bounds, a substantial amount of legitimate speech would be “chilled” as a consequence of the rule Justice O’Connor would adopt. MASSACHUSETTS v. OAKES 587 576 Opinion of Scalia, J. More fundamentally, however, even if Justice O’Connor’s policy analysis were correct, it seems to me that we are only free to pursue policy objectives through the modes of action traditionally followed by the courts and by the law. In my view we have the power to adopt a rule of law which says that the defendant’s acts were lawful because the statute that sought to prohibit them was overbroad and therefore invalid. I do not think we have the power to pursue the policy underlying that rule of law more directly and precisely, saying that we will hold the defendant criminally liable or not, depending upon whether, by the time his last appeal is exhausted, letting him off would serve to eliminate any First Amendment “chill.” Even if one were of the view that some of the uses of the overbreadth doctrine have been excessive, this would not be a legitimate manner in which to rein it in.1 Justice O’Connor 1 Bigelow v. Virginia, 421 U. S. 809 (1975), is not to the contrary. In that case, which similarly involved both a facial and an as-applied challenge to a statute that had been amended postconviction, the Court said: “In view of the statute’s amendment since Bigelow’s conviction in such a way as ‘effectively to repeal’ its prior application, there is no possibility now that the statute’s pre-1972 form will be applied again to appellant or will chill the rights of others. As a practical matter, the issue of its overbreadth has become moot for the future. We therefore decline to rest our decision on overbreadth and we pass on to the further inquiry, of greater moment not only for Bigelow but for others, whether the statute as applied to appellant infringed constitutionally protected speech.” Id., at 817-818. Although the dissent in Bigelow characterized this as a statement that “Virginia’s statute cannot properly be invalidated on grounds of overbreadth,” id., at 830 (Rehnquist, J., dissenting), I do not think it says that. Whether the statute is invalid because of overbreadth and whether the issue of overbreadth should be reached are two quite different questions, and it is only the latter that the Court addressed. The Court simply decided that since the question whether the statute was overbroad was no longer of general interest (“ha[d] become moot for the future”), whereas the issues involved in the as-applied challenge were of continuing importance, the Court would more profitably expend its time on the latter. Moreover, as the Court held Bigelow’s conviction unconstitutional on as-applied grounds, it was unnecessary to decide the merits of the overbreadth issue in that case. 588 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. seeks to cloak its extravagant constitutional doctrine in conservative garb borrowed from an entirely different area of the law, saying that “[a]n overbroad statute is not void ab initio, but rather voidable.” Ante, at 584. I have heard of a voidable contract, but never of a voidable law. The notion is bizarre. II Since I find that the subsequent amendment of the statute under which Oakes acted and was convicted does not eliminate the defense of overbreadth, I reach the question whether the statute is impermissibly overbroad. I do not believe that it is. Because the Court as a whole does not reach the question, I sketch my views on it only in brief. In order to be invalidated under our overbreadth doctrine, a statute’s unconstitutional application must be substantial, not just in an absolute sense, but “judged in relation to the statute’s plainly legitimate sweep.” Broadrick v. Oklahoma, 413 U. S. 601, 615 (1973). We held in New York v. Ferber, 458 U. S. 747, 756-757 (1982), that the State has a “compelling” interest in “safeguarding the physical and psychological well-being of . . . minor[s]” against harm of the sort at issue here. That case upheld against First Amendment attack a law directed against the use of children in pornographic (including nonobscene) materials. (Although the prohibition related to the distribution of pictures rather than the making of them, the former would seem to be even closer to the core of the First Amendment.) Thus, the scope of this statute has already been validated except as to nonpornographic depiction of preadolescent genitals, and postadoles-cent genitals and female breasts. On that basis alone, given the known extent of the so-called kiddie-porn industry, Act of May 21, 1984, 98 Stat. 204, and of pornographic magazines that use young female models (to one of which the defendant here apparently intended to send his stepdaughter’s photograph), I would estimate that the legitimate scope vastly exceeds the illegitimate. MASSACHUSETTS v. OAKES 589 576 Opinion of Scalia, J. But the statute is narrowed further still, since it excludes material “produced, processed, published, printed or manufactured for a bona fide scientific or medical purpose, or for an educational or cultural purpose for a bona fide school, museum or library.” The only significant body of material that would remain, I estimate, consists of artistic depictions not “produced, processed, published, printed or manufactured ... for a bona fide school, museum or library,” and (the example posited by the Massachusetts court) family snapshots. As to the former: Even assuming that proscribing artistic depictions of preadolescent genitals and postadolescent breasts is impermissible,2 the body of material that would be covered is, as far as I am aware, insignificant compared with the lawful scope of the statute. That leaves the family photos. The Supreme Judicial Court interpreted the statute to cover “a parent who takes a frontal view picture of his or her naked one-year-old running on a beach or romping in a wading pool.” 401 Mass. 602, 605, 518 N. E. 2d 836, 838 (1988). Assuming that it is unconstitutional (as opposed to merely foolish) to prohibit such photography, I do not think it so common as to make the statute substantially overbroad. We can deal with such a situation in the unlikely event some prosecutor brings an indictment. Cf. Ferber, supra, at 773-774, quoting Broadrick, supra, at 615-616. Perhaps I am wrong in my estimation of how frequently the posings prohibited by this law are done for artistic purposes, or for family photographs—or in some other legitimate 2 Justice Brennan evidently believes that the State cannot bar the use of children for nude modeling without reference to “the adult’s intentions or the sexually explicit nature of the minor’s conduct.” Post, at 597. That is not unquestionably true. Most adults, I expect, would not hire themselves out as nude models, whatever the intention of the photographer or artist, and however unerotic the pose. There is no cause to think children are less sensitive. It is not unreasonable, therefore, for a State to regard parents’ using (or permitting the use) of their children as nude models, or other adults’ use of consenting minors, as a form of child exploitation. 590 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. and constitutionally protected context I have not envisioned. My perception differs, for example, from Justice Brennan’s belief that there is an “abundance of baby and child photographs taken every day” depicting genitals, post, at 598. But it is the burden of the person whose conduct is legitimately proscribable, and who seeks to invalidate the entire law because of its application to someone else, to “demonstrate from the text of [the law] and from actual fact” that substantial overbreadth exists. New York State Club Assn. v. New York City, 487 U. S. 1, 14 (1988) (emphasis added). That has not been done here. Having found the ground upon which the Supreme Judicial Court of Massachusetts relied to be in error, I would reverse and remand the case to permit that court to dispose of the as-applied challenge. Justice Brennan, with whom Justice Marshall and Justice Stevens join, dissenting. The proper framework for analyzing respondent’s claims is not in doubt. First, we must determine whether the Massachusetts statute criminalizes expression protected by the First Amendment. If it does, then we must decide whether Massachusetts has a compelling interest in regulating that expression. To the extent that the Commonwealth’s interest does not justify the suppression of all protected conduct prohibited by the statute, we must further ask whether the law’s overbreadth is “not only . . . real, but substantial as well, judged in relation to the statute’s plainly legitimate sweep,” Broadrick n. Oklahoma, 413 U. S. 601, 615 (1973), once we have adopted any available narrowing constructions or severed offensive portions insofar as it lies within our power to do so. If the statute is excessively overbroad, we have no choice but to strike it down on its face, notwithstanding its laudable objectives and its numerous permissible applications; if it is not, then Oakes and others charged under MASSACHUSETTS v. OAKES 591 576 Brennan, J., dissenting it may argue only that their actions, though forbidden by the statute, may not constitutionally be proscribed.1 With the possible exception of the final step in this analysis, the resolution of these questions is straightforward. Photography, painting, and other two-dimensional forms of artistic reproduction described in Mass. Gen. Laws § 272:29A (1986) are plainly expressive activities that ordinarily qualify for First Amendment protection. See, e. g., Miller v. California, 413 U. S. 15 (1973) (works which, taken as a whole, possess serious artistic value are protected). And modeling, both independently and by virtue of its close association with those activities, enjoys like shelter under the First Amendment. Cf. Schad v. Mount Ephraim, 452 U. S. 61, 66 (1981) (“[N]ude dancing is not without its First Amendment protections from official regulation”). Visual depictions of children engaged in live sexual performances or lewdly exhibiting their genitals cannot, of course, claim protected status, even though those depictions are not obscene. See New York v. Ferber, 458 U. S. 747 (1982). But other nonobscene representations of minors, including some that are pornographic, are shielded by the Constitution’s guarantee of free speech. Id., at 764-765. In particular, “nudity, without more is protected expression.” Id., at 765, n. 18, citing Erznoznik v. City of Jacksonville, 422 U. S. 205, 213 (1975). Because ‘I agree with Justice Scalia that a State cannot salvage a criminal conviction under a law found to be overbroad, or safeguard its right to prosecute under a law challenged as overbroad, by curing the statute’s adjudicated or alleged infirmity prior to review of that conviction or ruling of statutory invalidation by the highest reviewing court. The deterrent effect of the overbreadth doctrine would be significantly impaired if this avenue were open to the States, for oftentimes the strongest and earliest attacks on overbroad laws are, not surprisingly, brought by criminal defendants. Accordingly, I join Part I of Justice Scalia’s opinion holding that a defendant’s overbreadth challenge cannot be rendered moot by narrowing the statute after the conduct for which he has been indicted occurred—the only proposition to which five Members of the Court have subscribed in this case. 592 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. § 29A’s prohibition extends to posing or exhibiting children “in a state of nudity,” rather than merely to their participation in live or simulated sexual conduct, the statute clearly restrains expression within the ambit of the First Amendment. It is equally evident that the Commonwealth’s asserted interest in preventing the sexual exploitation and abuse of minors is “of surpassing importance.” Ferber, supra, at 757. See also Ginsberg v. New York, 390 U. S. 629, 639-641 (1968). The coercive enlistment, both overt and subtle, of children in the production of pornography is a grave and widespread evil which the States are amply justified in seeking to eradicate. Massachusetts’ interest in ending such conduct undoubtedly suffices to sustain the statute’s ban on encouraging, causing, or permitting persons one has reason to know are under 18 years of age to engage in any live sexual performance or any act that represents sexual conduct, for the purpose of visual representation or reproduction. The Commonwealth lacks an overriding interest, however, in prohibiting adults from allowing minors to appear naked in photographs, films, and pictures with their genitals or, in the case of adolescent girls, their breasts less than opaquely covered under all circumstances except the production of such works “for a bona fide scientific or medical purpose, or for an educational or cultural purpose for a bona fide school, museum or library.” §29A. One situation where the Commonwealth’s interest falls glaringly short was cited by the Massachusetts Supreme Judicial Court: parents might want to photograph their infant children or toddlers in the bath or romping naked on the beach, yet § 29A threatens them with a prison term of between 10 and 20 years or a minimum fine of $10,000 for doing so. And §29A imposes those penalties even though parents have the same First Amendment interest in taking those photographs as they do in keeping a diary or boasting of their children’s antics, and even though their children would not thereby be harmed. Amicus American Sunbathing Association, a nudist organization with 30,000 MASSACHUSETTS v. OAKES 593 576 Brennan, J., dissenting members in the United States and Canada, further notes that family photographs taken by its members would subject them to possible prosecution, notwithstanding the protected character of their activity and their denial of any intrinsic connection between public nudity and shame. Massachusetts likewise lacks a compelling interest in forbidding nonexploitative films or photographs of topless adolescents—for instance, the poolside shots that are the norm rather than the exception along the Mediterranean seaboard, and that occur with some frequency on this side of the Atlantic as well—or in barring acting or professional modeling by teenagers that does not involve sexually explicit conduct. In my view, the First Amendment also blocks the prohibition of nude posing by minors in connection with the production of works of art not depicting lewd behavior and not specifically prepared, in accordance with §29A’s exclusion, for museums or libraries. Many of the world’s great artists—Degas, Renoir, Donatello,2 to name but a few—have worked from models under 18 years of age, and many acclaimed photographs and films have included nude or partially clad minors.3 The First Amendment rights of models, actors, artists, photographers, and filmmakers are surely not overborne by the Commonwealth’s interest in protecting minors from the risk of sexual abuse and exploitation, especially in view of the comprehensive set of laws targeted at those evils.4 2 See, e.g., R. Thomson, Degas, The Nudes 40-53, 119-125 (1988); K. Clark, The Nude 48-49, 154-161 (1956). 3 Numerous contemporary examples of nonpornographic photographs, films, and paintings that would invite prosecution under § 29A if produced in Massachusetts but which almost certainly caused no harm to those depicted in them are collected in App. to Brief for Law & Humanities Institute as Amicus Curiae. 4 The utility of §29A in preventing sexual abuse and exploitation appears dubious when assessed against the backdrop of other statutes designed to achieve the same end. Massachusetts already has laws prohibiting assault and battery, Mass. Gen. Laws §265:13A (1986); indecent 594 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. Given that §29A is demonstrably overbroad, the next question is whether it fairly admits of a narrowing construction or whether offending portions of the statute might be severèd, leaving its legitimate core prohibition intact. The answer to this question is that a restrictive reading of the statute or its partial invalidation is beyond our power. When we sit to review a decision resting on a state court’s construction of a state statute, that construction is binding on us, regardless of whether in its absence we would have read the statute in the same way or would have pruned it back before passing judgment. Ferber, 458 U. S., at 769, n. 24; Erznoznik, 422 U. S., at 216. “[W]e will not rewrite a state law to conform it to constitutional requirements.” Virginia v. American Booksellers Assn., Inc., 484 U. S. 383, assault and battery of a child under 14, § 265:13B; indecent assault and battery of a child 14 or older, § 265:13H; rape, § 265:22; forcible rape of a child under 16, § 265:22A; rape and abuse of a child under 16, § 265:23; assault with intent to rape, § 265:24; drugging persons to commit unlawful sexual intercourse, §272:3; open and gross lewdness, §272:16; dissemination of matter harmful to minors, § 272:28; dissemination or possession with intent to disseminate visual material of child in state of nudity or engaged in sexual conduct, §272:29B; and unnatural and lascivious acts with a child under 16, § 272:35A. Virtually every prosecution under § 29A has also involved charges under several of these other statutes. See App. to Brief for Petitioner. The marginal deterrent effect of § 29A may therefore be slight, thereby reducing the Commonwealth’s interest as against the First Amendment interests in conflict with § 29A. Of course, the penalties for violating § 29A are high; in fact, however odd the underlying scale of values or predictions of deterrence may appear, the punishment for allowing a child to be photographed nude exceeds that for dissemination of matter harmful to minors, § 272:28, and unnatural and lascivious acts with a child under 16, § 272:35A, and includes a maximum prison term in excess of that for indecent assault and battery of a child under 14, §265:13B (20 years under §29A versus 10 years under §265:13B). It is questionable, however, what marginal difference the unusual stiffness of these penalties makes in forestalling the production of pornography or the sexual abuse of children, which are often punishable under separate statutes. Section 29A’s most significant deterrent effect may well be on constitutionally protected conduct. MASSACHUSETTS v. OAKES 595 576 Brennan, J., dissenting 397 (1988) (certifying interpretive questions to Virginia Supreme Court before ruling on First Amendment facial attack). In this case, § 29A’s prohibition on causing or allowing a minor to pose naked is unambiguous, and the Massachusetts Supreme Judicial Court expressly held that it forbids the various forms of constitutionally protected conduct just described. 401 Mass. 602, 605, 518 N. E. 2d 836, 838 (1988). In addition, although the phrase “to pose or be exhibited in a state of nudity” might easily have been excised, the court refused to sever and delete it, over the protest of three dissenters. Id., at 611, n. 4, 518 N. E. 2d, at 841, n. 4 (O’Connor, J., dissenting). We have no choice but to accept these authoritative pronouncements in adjudging the validity of §29A. The test we employ is familiar. Because “conduct and not merely speech is involved, . . . the overbreadth of a statute must not only be real, but substantial as well, judged in relation to the statute’s plainly legitimate sweep.” Broadrick, 413 U. S., at 615. See also, e. g., Board of Airport Comm’rs of Los Angeles v. Jews for Jesus, Inc., 482 U. S. 569, 574 (1987); Houston n. Hill, 482 U. S. 451, 458 (1987); Ferber, supra, at 769. We will not topple a statute merely because we can conceive of a few impermissible applications. See City Council of Los Angeles v. Taxpayers for Vincent, 466 U. S. 789, 800, and n. 19 (1984). The possibility of a substantial number of realistic applications in contravention of the First Amendment, however, suffices to overturn a statute on its face. In this regard, it bears emphasizing that “the penalty to be imposed is relevant in determining whether demonstrable overbreadth is substantial.” Ferber, 458 U. S., at 773. Although “the fact that a criminal prohibition is involved does not obviate the need for the inquiry or a priori warrant a finding of substantial overbreadth,” ibid., it does appreciably shrink the amount of overbreadth we will find constitutionally tolerable, particularly when the penalty is se 596 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. vere. See also Houston v. Hill, supra, at 459 (“Criminal statutes must be scrutinized with particular care”).5 In this case, there is no gainsaying the gravity of the penalties meted out for violations of §29A. Infractions carry a fine of between $10,000 and $50,000, a prison term of between 10 and 20 years, or both. Respondent was himself sentenced to 10 years’ imprisonment for taking fewer than a dozen snapshots of his stepdaughter, which he apparently showed no one except the complainant. The severity of these sanctions significantly reduces the degree of overbreadth that the Constitution permits. One can also readily adduce actual examples of protected conduct within §29A’s compass. Parents photograph their children without abusing them sexually in Massachusetts as elsewhere. The arts flourish there. Four nudist clubs affiliated with the American Sunbathing Association alone have been established in the Commonwealth, Brief for American Sunbathing Association as Amicus Curiae 2, and there may well be others. The only question that might give one pause is whether the statute’s overbreadth is substantial. Unhappily, our precedents provide limited guidance in resolving this issue, because substantiality cannot be defined with exactitude and 5 In considering a facial challenge of this kind, we have no reason to decide, of course, whether respondent’s own conduct may legitimately be proscribed. Nor is it for us to say what exactly Oakes did when the evidence is sharply conflicting, particularly when we are remanding the case for further consideration of his as-applied challenge. Justice Scalia’s statement that “the defendant here apparently intended to send his stepdaughter’s photograph” to one of the “pornographic magazines that use young female models,” ante, at 588, therefore seems to me inappropriate. The only record support for this assertion of which I am aware is the complainant’s testimony at trial, ambiguous with respect to Oakes’ intentions regarding the photographs at issue here, that “[h]e wanted to make me big for Playboy Magazine.” App. 25a. In any event, nothing this Court says should be taken to constrain the power of the Massachusetts courts to determine facts on remand. MASSACHUSETTS v. OAKES 597 576 Brennan, J., dissenting little overlap exists between the factual situations presented in our previous overbreadth cases and the circumstances Confronting us here. But several considerations that have led us to strike down laws by reason of overbreadth tug with equal force in this case, strongly suggesting that § 29A cannot stand as it was written at the time respondent photographed his stepdaughter. In Houston v. Hill, supra, at 464-466, we asked whether the sweeping nature of an ordinance making it a criminal offense to oppose, abuse, or interrupt a policeman in the performance of his duties was essential to achieve its ends, or whether a more narrowly tailored law could have attained the same objectives without abridging First Amendment freedoms to the same extent. Our finding that the law could have been drafted more tightly without sacrificing the achievement of its legitimate purposes impelled us to pronounce it fatally overbroad. Section 29A suffers from the same flaw. Its blanket prohibition on permitting minors to pose nude or employing nude models, without regard to the adult’s intentions or the sexually explicit nature of the minor’s conduct, nets a considerable amount of protected conduct. The statute can, moreover, easily be truncated. As the plurality describes, ante, at 582-583, and n. 2, Massachusetts itself has recently amended § 29A to lessen its threat to protected conduct by requiring that an adult act with “lascivious intent” to come within the statute’s prohibition.6 Mass. Gen. Laws §272:29A(a) (Supp. 1988). Alternatively, the Commonwealth could have followed the advice offered by the Justice Department in 1977. In considering legislation designed to combat the sexual exploitation of children in photographs and films, the House of Representatives initially considered banning the interstate dissemination, or the taking of photographs with intent or reason to know that they will be transported in interstate commerce, of children in a state of ” I venture no views as to the constitutionality of § 29A as amended. 598 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. “ ‘nudity. . . depicted for the purpose of sexual stimulation or gratification of any individual who may view such depiction.’” See H. R. Rep. No. 95-696, p. 21 (1977) (quoting H. R. 4571, 95th Cong., 1st Sess. (1977)). The Justice Department opposed the inclusion of this provision on the ground that “it would be difficult to determine by what standard the ‘sexual stimulation or gratification’ could be assessed.” H. R. Rep. No. 95-696, at 21 (statement of Deputy Assistant Attorney General Keeney). The Justice Department suggested that “lewd exhibition of the genitals” be used in its place, ibid., and the House heeded that recommendation. Massachusetts could have followed the same course and modified § 29A’s reference to simple nudity, thereby aligning the law with the New York statute we upheld in Ferber. The availability of such simple correctives renders the statute’s overbreadth less acceptable. Together with the stern sanctions § 29A imposes, the ease with which its unconstitutional applications might be eliminated lowers the hurdle respondent must clear in proving substantial overbreadth. By the standards set in our earlier decisions, that proof has in my judgment been made. The abundance of baby and child photographs taken every day without full frontal covering, not to mention the work of artists and filmmakers and nudist family snapshots, allows one to say, as the Court said in Houston v. Hill, 482 U. S., at 466-467 (citation omitted), that “[t]he ordinance’s plain language is admittedly violated scores of times daily, yet only some individuals—those chosen by the police in their unguided discretion—are arrested.” Indeed, even if I were less confident that the statute was routinely violated by protected conduct—and the test, of course, is the relative frequency of such violations, not what we believe is the likelihood that such violations will in fact be prosecuted—I would reach the same conclusion. In Erznoz-nik v. City of Jacksonville, 422 U. S. 205 (1975), we struck down for overbreadth a statute making it a public nuisance to MASSACHUSETTS v. OAKES 599 576 Brennan, J., dissenting show films at a drive-in theater displaying bare buttocks, pubic areas, or female breasts, if the screen was visible from a public area. By way of justification we said: “[The statute] would bar a film containing a picture of a baby’s buttocks, the nude body of a war victim, or scenes from a culture in which nudity is indigenous. The ordinance also might prohibit newsreel scenes of the opening of an art exhibit as well as shots of bathers on a beach. Clearly all nudity cannot be deemed obscene even as to minors.” Id., at 213. We saw no reason to inquire into the frequency with which such scenes appeared at drive-in movies in Jacksonville; the fact that they might be shown, and sometimes were shown, was enough. The amount of protected conduct that occurs and quite plainly is covered by § 29A is undoubtedly far greater than the speculative occurrences we found sufficient to establish substantial overbreadth in Erznoznik, where, in addition, the attendant penalties were puny by comparison. Thus, even granted a stingy estimate of the extent of § 29A’s overbreadth, the statute must fall. I would affirm the decision of the Massachusetts Supreme Judicial Court. 600 OCTOBER TERM, 1988 Syllabus 491 U. S. UNITED STATES v. MONSANTO CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT No. 88-454. Argued March 21, 1989—Decided June 22, 1989 Respondent, who allegedly directed a large-scale heroin distribution enterprise, was indicted for alleged violations of racketeering laws, creation of a continuing criminal enterprise, and tax and firearm offenses. The indictment also alleged that respondent had accumulated three specified assets as a result of his narcotics trafficking, which were subject to forfeiture under the Comprehensive Forfeiture Act of 1984, 21 U. S. C. § 853. After the indictment was unsealed, the District Court granted the Government’s ex parte motion under § 853(e)(1)(A) for a restraining order freezing the assets pending trial. Respondent, raising various statutory arguments and claiming that the order interfered with his Sixth Amendment right to counsel of his choice, moved to vacate the order to permit him to use frozen assets to retain an attorney. He also sought a declaration that if the assets were used to pay attorney’s fees, §853(c)’s third-party transfer provision would not be used to reclaim such payments if respondent was convicted and his assets forfeited. The District Court denied the motion. However, the Court of Appeals ultimately ordered that the restraining order be modified to permit the restrained assets to be used to pay attorney’s fees. Held: 1. There is no exemption from § 853’s forfeiture or pretrial restraining order provisions for assets that a defendant wishes to use to retain an attorney. Pp. 606-614. (a) Section 853’s language is plain and unambiguous. Congress could not have chosen stronger words to express its intent that forfeiture be mandatory than §853(a)’s language that upon conviction a person “shall forfeit . . . any property” and that the sentencing court “shall order” a forfeiture. Likewise, the statute provides a broad definition of property which does not even hint at the idea that assets used for attorney’s fees are not included. Every Court of Appeals that has finally passed on this argument has agreed with this view. Neither the Act’s legislative history nor legislators’ postenactment statements support respondent’s argument that an exception should be created because the statute does not expressly include property to be used for attorney’s fees or because Congress simply did not consider the prospect that forfeiture UNITED STATES v. MONSANTO 601 600 Syllabus would reach such property. To the contrary, in the Victims of Crime Act—which requires forfeiture of a convicted defendant’s collateral profits derived from his crimes and which was enacted simultaneously with the statute in question—Congress adopted expressly the precise exemption from forfeiture which respondent is seeking to have implied in § 853. Moreover, respondent’s admonition that courts should construe statutes to avoid decision as to their constitutionality is not license for the judiciary to rewrite statutory language. Pp. 606-611. (b) Respondent’s reading of § 853(e)(1)(A)—which provides that a district court “may enter a restraining order or injunction ... or take any other action to preserve the availability of property ... for forfeiture”—misapprehends the nature of §853 by giving a district court equitable discretion to determine whether to exempt assets from pretrial restraint and by concluding that if such assets are used for attorney’s fees, they may not subsequently be seized for forfeiture to the Government under § 853(c). Section 853(e)(1)(A) plainly is aimed at implementing § 853(a)’s commands and cannot sensibly be construed to give the district court discretion to permit the dissipation of the very property it requires be forfeited upon conviction, since this would nullify § 853(a)’s strong language as well as § 853(c)’s powerful “relation-back” provision. Pp. 611-614. 2. The restraining order did not violate respondent’s right to counsel of choice as protected by the Sixth Amendment or the Due Process Clause of the Fifth Amendment. For the reasons stated in Caplin & Drysdale, Chartered v. United States, post, p. 617, neither the Fifth nor the Sixth Amendment requires Congress to permit a defendant to use assets adjudged to be forfeitable to pay the defendant’s legal fees. Moreover, a defendant’s assets may be frozen before conviction based on a finding of probable cause to believe the assets are forfeitable. See, e. g., United States v. $8,850, 461 U. S. 555; Calero-Toledo v. Pearson Yacht Leasing Co., 416 U. S. 663. Indeed, concluding that the Government could not restrain such property would be odd considering that, under appropriate circumstances, the Government may restrain persons accused of a serious offense on a probable-cause finding. See United States v. Salerno, 481 U. S. 739. Pp. 614-616. 852 F. 2d 1400, reversed and remanded. White, J., delivered the opinion of the Court, in which Rehnquist, C. J., and O’Connor, Scalia, and Kennedy, JJ., joined. Blackmun, J., filed a dissenting opinion, in which Brennan, Marshall, and Stevens, JJ., joined, post, p. 635. 602 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Acting Solicitor General Bryson argued the cause for the United States. With him on the briefs were Assistant Attorney General Dennis, Edwin S. Kneedler, and Sara Criscitelli. Edward M. Chikofsky argued the cause and filed a brief for respondent. * Justice White delivered the opinion of the Court. The questions presented here are whether the federal drug forfeiture statute authorizes a district court to enter a pretrial order freezing assets in a defendant’s possession, even where the defendant seeks to use those assets to pay an attorney; if so, we must decide whether such an order is permissible under the Constitution. We answer both of these questions in the affirmative. I In July 1987, an indictment was entered, alleging that respondent had directed a large-scale heroin distribution enterprise. The multicount indictment alleged violations of racketeering laws, creation of a continuing criminal enterprise (CCE), and tax and firearm offenses. The indictment also alleged that three specific assets—a home, an apartment, and $35,000 in cash—had been accumulated by respondent as a result of his narcotics trafficking. These assets, the indict- *Briefs of amici curiae urging reversal were filed for the State of California by John K. Van de Kamp, Attorney General, Steve White, Chief Assistant Attorney General, John A. Gordnier, Senior Assistant Attorney General, and Gary W. Schons, Deputy Attorney General; and for Eugene R. Anderson, pro se. Briefs of amici curiae urging affirmance were filed for the Committees on Criminal Advocacy and Criminal Law of the Association of the Bar of the City of New York et al. by Arthur L. Liman; and for the National Association of Criminal Defense Lawyers et al. by Joseph Beeler and Bruce J. Winick. Briefs of amici curiae were filed for the American Bar Association by Robert D. Raven, Charles G. Cole, Antonia B. lanniello, and Terrance G. Reed; and for the Appellate Committee of the California District Attorneys Association by Ira Reiner, Harry B. Sondheim, and Arnold T. Guminski. UNITED STATES v. MONSANTO 603 600 Opinion of the Court ment alleged, were subject to forfeiture under the Comprehensive Forfeiture Act of 1984 (CFA), 98 Stat. 2044, as amended, 21 U. S. C. § 853(a) (1982 ed., Supp. V), because they were “property constituting, or derived from . . . proceeds . . . obtained” from drug-law violations.1 On the same day that the indictment was unsealed, the District Court granted the Government’s ex parte motion, pursuant to § 853(e)(1)(A),2 for a restraining order freezing 1 The CFA added or amended forfeiture provisions for two classes of violations under federal law, racketeering offenses and CCE offenses, see 98 Stat. 2040-2053, as amended. The CCE forfeiture statute at issue here, now provides: “§ 853. Criminal forfeitures “(a) Property subject to criminal forfeiture “Any person convicted of a violation of this subchapter or subchapter II of this chapter punishable by imprisonment for more than one year shall forfeit to the United States, irrespective of any provision of State law— “(1) any property constituting, or derived from, any proceeds the person obtained, directly or indirectly, as the result of such violation; “(2) any of the person’s property used, or intended to be used, in any manner or part, to commit, or to facilitate the commission of, such violation; and “(3) in the case of a person convicted of engaging in a continuing criminal enterprise in violation of section 848 of this title, the person shall forfeit, in addition to any property described in paragraph (1) or (2), any of his interest in, claims against, and property or contractual rights affording a source of control over, the continuing criminal enterprise. “The court, in imposing sentence on such person, shall order, in addition to any other sentence imposed pursuant to this subchapter or subchapter II of this chapter, that the person forfeit to the United States all property described in this subsection. In lieu of a fine otherwise authorized by this part, a defendant who derives profits or other proceeds from an offense may be fined not more than twice the gross profits or other proceeds.” 2 This statutory provision, the principal focus of this petition, says that: “Upon application of the United States, the court may enter a restraining order or injunction ... or take any other action to preserve the availability of property described in subsection (a) of [§ 853] for forfeiture under this section— “(A) upon the filing of an indictment or information charging a violation • . . for which criminal forfeiture may be ordered under [§ 853] and alleging 604 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the above-mentioned assets pending trial. Shortly thereafter, respondent moved to vacate this restraining order, to permit him to use the frozen assets to retain an attorney. Respondent’s motion further sought a declaration that if these assets were used to pay an attorney’s fees, §853(c)’s third-party transfers provision would not subsequently be used to reclaim such payments if respondent was convicted and his assets forfeited.3 Respondent raised various statutory challenges to the restraining order, and claimed that it interfered with his Sixth Amendment right to counsel of choice. The District Court denied the motion to vacate. that the property with respect to which the order is sought would, in the event of conviction, be subject to forfeiture under this section.” 3 Section 853(c), the third-party transfer provision, states that: “All right, title, and interest in property described in [§ 853] vests in the United States upon the commission of the act giving rise to forfeiture under this section. Any such property that is subsequently transferred to a person other than the defendant may be the subject of a special verdict of forfeiture and thereafter shall be ordered forfeited to the United States, unless the transferee [establishes his entitlement to such property pursuant to § 853(n)].” As noted in the quotation of § 853(c), a person making a claim for forfeited assets must file a petition with the court pursuant to § 853(n)(6): “If, after [a] hearing [on the petition], the court determines that the petitioner has established . . . that — “(A) the petitioner has a legal right, title, or interest in the property. . • [that predates] commission of the acts which gave rise to the forfeiture of the property under [§ 853]; or “(B) the petitioner is a bona fide purchaser for value of the . . . property and was at the time of purchase reasonably without cause to believe that the property was subject to forfeiture under this section; “the court shall amend the order of forfeiture in accordance with its determination.” An attorney seeking a payment of fees from forfeited assets under §853(n)(6) would presumably rest his petition on subsection (B) quoted above, though (for reasons we explain in Caplin & Drysdale, Chartered v. United States, post, at 632, n. 10) it is highly doubtful that one who defends a client in a criminal case that results in forfeiture could prove that he was “without cause to believe that the property was subject to forfeiture.” Cf. 852 F. 2d, 1400, 1410 (CA2 1988) (Winter, J., concurring). UNITED STATES v. MONSANTO 605 600 Opinion of the Court On appeal, the Second Circuit concluded that respondent’s statutory and Sixth Amendment challenges were lacking, but remanded the case to the District Court for an adversarial hearing “at which the government ha[d] the burden to demonstrate the likelihood that the assets are forfeitable”; if the Government failed its burden at such a hearing, the Court of Appeals held, any fees paid to an attorney would be exempt from forfeiture irrespective of the final outcome at respondent’s trial. 836 F. 2d 74, 84 (1987). Pursuant to this mandate, on remand, the District Court held a 4-day hearing on whether continuing the restraining order was proper. At the end of the hearing, the District Court ruled that it would continue the restraining order because the Government had “overwhelmingly established a likelihood” that the property in question would be forfeited at the end of trial. App. to Pet. for Cert. 86a. Ultimately, respondent’s criminal case proceeded to trial, where he was represented by a Criminal Justice Act-appointed attorney.4 In the meantime, the Second Circuit vacated its earlier opinion and heard respondent’s appeal en banc.5 The en 4 At the end of the trial, respondent was convicted of the charges against him, and the jury returned a special verdict finding the assets in question to be forfeitable beyond a reasonable doubt. Accordingly, the District Court entered a judgment of conviction and declared the assets forfeited. We do not believe that these subsequent proceedings render the dispute over the pretrial restraining order moot. The restraining order remains in effect pending the appeal of respondent’s conviction, see App. to Pet. for Cert. 77a-78a, which has not yet been decided. Consequently, the dispute before us concerning the District Court’s order remains a live one. “Respondent’s trial had commenced on February 16, 1988, after the Court of Appeals had agreed to hear the case en banc, but before it rendered its ruling. Consequently, respondent’s assets remained frozen, and respondent was defended by appointed counsel. In the midst of respondent’s trial—on July 1, 1988—the en banc Court of Appeals rendered its decision for respondent. At a hearing held four days later, the District Court offered to permit respondent to use the frozen assets to hire private counsel. Respondent rejected this offer, coming as 606 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. banc court, by an 8-to-4 vote, ordered that the District Court’s restraining order be modified to permit the restrained assets to be used to pay attorney’s fees. 852 F. 2d 1400 (1988). The Court was sharply divided as to its rationale. Three of the judges found that the order violated the Sixth Amendment, while three others questioned it on statutory grounds; two judges found § 853 suspect under the Due Process Clause for its failure to include a statutory provision requiring the sort of hearing that the panel had ordered in the first place. The four dissenting judges would have upheld the restraining order. We granted certiorari, 488 U. S. 941 (1988), because the Second Circuit’s decision created a conflict among the Courts of Appeals over the statutory and constitutional questions presented.6 We now reverse. II We first must address the question whether § 853 requires, upon conviction, forfeiture of assets that an accused intends to use to pay his attorneys. A “In determining the scope of a statute, we look first to its language.” United States v. Turkette, 452 U. S. 576, 580 (1981). In the case before us, the language of §853 is plain and unambiguous: all assets falling within its scope are to be forfeited upon conviction, with no exception existing for the assets used to pay attorney’s fees—or anything else, for that matter. summations were about to get underway at the end of a 4 V2-month trial, and instead continued with his appointed attorney. Three weeks later, on July 25, 1988, the jury returned a guilty verdict. 6 See, e. g., United States v. Moya-Gomez, 860 F. 2d 706 (CA7 1988); United States v. Nichols, 841 F. 2d 1485 (CAIO 1988); United States v. Jones, 837 F. 2d 1332 (CA5), rehearing granted, 844 F. 2d 215 (1988); In re Forfeiture Hearing as to Caplin & Drysdale, Chartered, 837 F. 2d 637 (CA4 1988) (en banc), aff’d sub nom. Caplin & Drysdale, Chartered v. United States, post, p. 617. UNITED STATES v. MONSANTO 607 600 Opinion of the Court As observed above, § 853(a) provides that a person convicted of the offenses charged in respondent’s indictment “shall forfeit . . . any property” that was derived from the commission of these offenses. After setting out this rule, § 853(a) repeats later in its text that upon conviction a sentencing court “shall order” forfeiture of all property described in § 853(a). Congress could not have chosen stronger words to express its intent that forfeiture be mandatory in cases where the statute applied, or broader words to define the scope of what was to be forfeited. Likewise, the statute provides a broad definition of “property” when describing what types of assets are within the section’s scope: “real property . . . tangible and intangible personal property, including rights, privileges, interests, claims, and securities.” 21 U. S. C. § 853(b) (1982 ed., Supp. V). Nothing in this all-inclusive listing even hints at the idea that assets to be used to pay an attorney are not “property” within the statute’s meaning. Nor are we alone in concluding that the statute is unambiguous in failing to exclude assets that could be used to pay an attorney from its definition of forfeitable property. This argument, advanced by respondent here, see Brief for Respondent 12-19, has been unanimously rejected by every Court of Appeals that has finally passed on it,7 as it was by the Second Circuit panel below, see 836 F. 2d, at 78-80; id., at 85-86 (Oakes, J., dissenting); even the judges who concurred on statutory grounds in the en banc decision did not accept this position, see 852 F. 2d, at 1405-1410 (Winter, J., concurring). We note also that the Brief for American Bar 7 See United States v. Bissell, 866 F. 2d 1343, 1348-1350 (CA11 1989); United States v. Moya-Gomez, supra, at 722-723; United States v. Nichols, 841 F. 2d, at 1491-1496; id., at 1509 (Logan, J., dissenting); In re Forfeiture Hearing as to Caplin & Drysdale, Chartered, 837 F. 2d, at 641-642 (en banc); id., at 651 (Phillips, J., dissenting). Only one Court of Appeals—the Fifth Circuit—has issued any decisions providing support for this reading of the statute, see, e. g., United States v. Jones, supra, but this ruling is currently being reconsidered en banc, 844 F. 2d 215 (1988). 608 OCTOBER TERM, 1988 Opinion of the Court 491 U. S, Association as Amicus Curiae 6 frankly admits that the statute “on [its] face, broadly co ver [s] all property derived from alleged criminal activity and contain[s] no specific exemption for property used to pay bona fide attorneys’ fees.” Respondent urges us, nonetheless, to interpret the statute to exclude such property for several reasons. Principally, respondent contends that we should create such an exemption because the statute does not expressly include property to be used for attorneys’ fees, and/or because Congress simply did not consider the prospect that forfeiture would reach assets that could be used to pay for an attorney. In support, respondent observes that the legislative history is “silent” on this question, and that the House and Senate debates fail to discuss this prospect.8 But this proves nothing: the legisla- * Respondent is correct that, by and large, the relevant House and Senate Reports make no mention of the attorney’s fees question. However, in discussing the background motivating the adoption of the CFA, the House Judiciary Committee discussed the failure of previous, more lax forfeiture statutes: “One highly publicized case ... is illustrative of the problem. That case was United States v. Meinster .... In this prosecution ... a Florida based criminal organization had . . . grossed about $300 million over a 16-month period. The Federal Government completed a successful prosecution in which the three primary defendants were convicted and this major drug operation was aborted. However, forfeiture was attempted on only two [residences] worth $750,000 .... “Of the $750,000 for the residences, $175,000 was returned to the wife of one of the defendants, and $559,000 was used to pay the defendant's attorneys. . . . “The Government wound up with $16,000. . . . “It is against this background that present Federal forfeiture procedures are tested and found wanting.” H. R. Rep. No. 98-845, pt. 1, p. 3 (1984) (emphasis added). This passage suggests, at the very least, congressional frustration with the diversion of large amounts of forfeitable assets to pay attorney’s fees. It certainly does not suggest an intent on Congress’ part to exempt from forfeiture such fees. Respondent claims support from only one piece of preenactment legislative history: a footnote in the same House Report quoted above, which dis UNITED STATES v. MONSANTO 609 600 Opinion of the Court tive history and congressional debates are similarly silent on the use of forfeitable assets to pay stockbroker’s fees, laundry bills, or country club memberships; no one could credibly argue that, as a result, assets to be used for these purposes are similarly exempt from the statute’s definition of forfeitable property. The fact that the forfeiture provision reaches assets that could be used to pay attorney’s fees, even though it contains no express provisions to this effect, “‘does not demonstrate ambiguity’” in the statute: ‘“It demonstrates breadth.’” Sedima, S. P. R. L. v. Imrex Co., 473 U. S. 479, 499 (1985) (quoting Haroco, Inc. v. American Nat. Bank & Trust Co. of Chicago, 747 F. 2d 384, 398 (CA7 1984)). The statutory provision at issue here is broad and unambiguous, and Congress’ failure to supplement § 853(a)’s comprehensive phrase—“any property”—with an exclamatory “and we even mean assets to be used to pay an attorney” does not lessen the force of the statute’s plain language. cussed the newly proposed provision for pretrial restraint on forfeitable assets. The footnote stated that: “Nothing in this section is intended to interfere with a person’s Sixth Amendment right to counsel. The Committee, therefore, does not resolve the conflict in District Court opinions on the use of restraining orders that impinge on a person’s right to retain counsel in a criminal case.” Id., at 19, n. 1. Respondent argues that the Committee’s disclaimer of any interest in resolving the conflict among the District Courts indicates the Committee’s understanding that the statute would not be employed to freeze assets that might be used to pay legitimate attorney’s fees. See Brief for Respondent 14, and n. 8. This ambiguous passage however, can be read for the opposite proposition as well, as the Report expressly refrained from disapproving of cases where pretrial restraining orders similar to the one issued here were imposed. See H. R. Rep. No. 98-845, supra, at 19, n. 1 (citing United States v. Bello, 470 F. Supp. 723, 724-725 (SD Cal. 1979)). Moreover, the Committee’s statement that the statute should not be applied in a manner contrary to the Sixth Amendment appears to be nothing more than an exhortation for the courts to tread carefully in this delicate area. 610 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. We also find unavailing respondent’s reliance on the comments of several legislators—made following enactment—to the effect that Congress did not anticipate the use of the forfeiture law to seize assets that would be used to pay attorneys. See Brief for Respondent 15-16, and n. 9 (citing comments of Sen. Leahy and Reps. Hughes and Shaw). As we have noted before, such postenactment views “form a hazardous basis for inferring the intent” behind a statute, United States v. Price, 361 U. S. 304, 313 (1960); instead, Congress’ intent is “best determined by [looking to] the statutory language that it chooses,” Sedima, S. P. R. L., supra, at 495, n. 13. Moreover, we observe that these comments are further subject to question because Congress has refused to act on repeated suggestions by the defense bar for the sort of exemption respondent urges here,9 even though it has amended § 853 in other respects since these entreaties were first heard. See Pub. L. 99-570, §§ 1153(b), 1864, 100 Stat. 3207-13, 3207-54. In addition, we observe that in the very same law by which Congress adopted the CFA—Pub. L. 98-473, 98 Stat. 1837— Congress also adopted a provision for the special forfeiture of collateral profits (e. g., profits from books, movies, etc.) that a convicted defendant derives from his crimes. See Victims of Crime Act of 1984, 98 Stat. 2175-2176 (now codified at 18 U. S. C. §§3681-3682 (1982 ed., Supp. V)). That forfeiture provision expressly exempts “pay[ments] for legal representation of the defendant in matters arising from the offense for which such defendant has been convicted, but no more than 20 percent of the total [forfeited collateral profits] may be so used.” § 3681(c)(l)(B)(ii). Thus, Congress adopted expressly—in a statute enacted simultaneously with the one under review in this case—the precise exemption from for 9 See, e. g., Attorneys’ Fees Forfeiture: Hearing before the Senate Committee on the Judiciary, 99th Cong., 2d Sess., 148-213 (1986); Forfeiture Issues: Hearing before the Subcommittee on Crime of the House Committee on the Judiciary, 99th Cong., 1st Sess., 187-242 (1985). UNITED STATES v. MONSANTO 611 600 Opinion of the Court feiture which respondent asks us to imply into §853. The express exemption from forfeiture of assets that could be used to pay attorney’s fees in Chapter XIV of Pub. L. 98-473 indicates to us that Congress understood what it was doing in omitting such an exemption from Chapter III of that enactment. Finally, respondent urges us, see Brief for Respondent 20-29, to invoke a variety of general canons of statutory construction, as well as several prudential doctrines of this Court, to create the statutory exemption he advances; among these doctrines is our admonition that courts should construe statutes to avoid decision as to their constitutionality. See, e. g., Edward J. DeBartolo Corp. v. Florida Gulf Coast Building & Constr. Trades Council, 485 U. S. 568, 575 (1988); NLRB. v. Catholic Bishop of Chicago, 440 U. S. 490, 500 (1979). We respect these canons, and they are quite often useful in close cases, or when statutory language is ambiguous. But we have observed before that such “interpretative canon[s are] not a license for the judiciary to rewrite language enacted by the legislature.” United States v. Albertini, 472 U. S. 675, 680 (1985). Here, the language is clear and the statute comprehensive: §853 does not exempt assets to be used for attorney’s fees from its forfeiture provisions. In sum, whatever force there might be to respondent’s claim for an exemption from forfeiture under § 853(a) of assets necessary to pay attorney’s fees—based on his theories about the statute’s purpose, or the implications of interpretative canons, or the understandings of individual Members of Congress about the statute’s scope—“[t]he short answer is that Congress did not write the statute that way.” United States v. Naftalin, 441 U. S. 768, 773 (1979). B Although § 853(a) recognizes no general exception for assets used to pay an attorney, we are urged that the provision 612 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. in § 853(e)(1)(A) for pretrial restraining orders on assets in a defendant’s possession should be interpreted to include such an exemption. It was on this ground that Judge Winter concurred below. 852 F. 2d, at 1405-1411. The restraining order subsection provides that, on the Government’s application, a district court “may enter a restraining order or injunction ... or take any other action to preserve the availability of property . . . for forfeiture under this section.” 21 U. S. C. § 853(e)(1) (1982 ed., Supp. V). Judge Winter read the permissive quality of the subsection (i. e., “may enter”) to authorize a district court to employ “traditional principles of equity” before restraining a defendant’s use of forfeitable assets; a balancing of hardships, he concluded, generally weighed against restraining a defendant’s use of forfeitable assets to pay for an attorney. 852 F. 2d, at 1406. Judge Winter further concluded that assets not subjected to pretrial restraint under § 853(e), if used to pay an attorney, may not be subsequently seized for forfeiture to the Government, notwithstanding the authorization found in § 853(c) for recoupment of forfeitable assets transferred to third parties. This reading seriously misapprehends the nature of the provisions in question. As we have said, § 853(a) is categorical: it contains no reference at all to § 853(e) or § 853(c), let alone any reference indicating that its reach is limited by those sections. Perhaps some limit could be implied if these provisions were necessarily inconsistent with § 853(a). But that is not the case. Under § 853(e)(1), the trial court “may” enter a restraining order if the United States requests it, but not otherwise, and it is not required to enter such an order if a bond or some other means to “preserve the availability of property described in subsection (a) of this section for forfeiture” is employed. Thus, § 853(e)(1)(A) is plainly aimed at implementing the commands of § 853(a) and cannot sensibly be construed to give the district court discretion to permit UNITED STATES v. MONSANTO 613 600 Opinion of the Court the dissipation of the very property that § 853(a) requires be forfeited upon conviction. We note that the “equitable discretion” that is given to the judge under § 853(e)(1)(A) turns out to be no discretion at all as far as the issue before us here is concerned: Judge Winter concludes that assets necessary to pay attorney’s fees must be excluded from any restraining order. See 852 F. 2d, at 1407-1409. For that purpose, the word “may” becomes “may not.” The discretion found in § 853(e) becomes a command to use that subsection (and § 853(c)) to frustrate the attainment of §853(a)’s ends. This construction is improvident. Whatever discretion Congress gave the district courts in §§ 853(e) and 853(c), that discretion must be cabined by the purposes for which Congress created it: “to preserve the availability of property . . . for forfeiture.” We cannot believe that Congress intended to permit the effectiveness of the powerful “relation-back” provision of § 853(c), and the comprehensive “any property . . . any proceeds” language of § 853(a), to be nullified by any other construction of the statute. This result may seem harsh, but we have little doubt that it is the one that the statute mandates. Section 853(c) states that “[a]ll right, title, and interest in [forfeitable] property . . . vests in the United States upon the commission of the act giving rise to forfeiture.” Permitting a defendant to use assets for his private purposes that, under this provision, will become the property of the United States if a conviction occurs cannot be sanctioned. Moreover, this view is supported by the relevant legislative history, which states that “[t]he sole purpose of [§ 853’s] restraining order provision ... is to preserve the status quo, i. e., to assure the availability of the property pending disposition of the criminal case.” S. Rep. No. 98-225, p. 204 (1983). If, instead, the statutory interpretation adopted by Judge Winter’s concurrence were applied, this purpose would not be achieved. 614 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. We conclude that there is no exemption from § 853’s forfeiture or pretrial restraining order provisions for assets which a defendant wishes to use to retain an attorney. In enacting §853, Congress decided to give force to the old adage that “crime does not pay.” We find no evidence that Congress intended to modify that nostrum to read, “crime does not pay, except for attorney’s fees.” If, as respondent and supporting amici so vigorously assert, we are mistaken as to Congress’ intent, that body can amend this statute to otherwise provide. But the statute, as presently written, cannot be read any other way. Ill Having concluded that the statute authorized the restraining order entered by the District Court, we reach the question whether the order violated respondent’s right to counsel of choice as protected by the Sixth Amendment or the Due Process Clause of the Fifth Amendment. A Respondent’s most sweeping constitutional claims are that, as a general matter, operation of the forfeiture statute interferes with a defendant’s Sixth Amendment right to counsel of choice, and the guarantee afforded by the Fifth Amendment’s Due Process Clause of a “balance of forces” between the accused and the Government. In this regard, respondent contends, the mere prospect of post-trial forfeiture is enough to deter a defendant’s counsel of choice from representing him. In another decision we announce today, Caplin & Drysdale, Chartered n. United States, post, p. 617, we hold that neither the Fifth nor the Sixth Amendment to the Constitution requires Congress to permit a defendant to use assets adjudged to be forfeitable to pay that defendant’s legal fees. We rely on our conclusion in that case to dispose of the similar constitutional claims raised by respondent here. UNITED STATES v. MONSANTO 615 600 Opinion of the Court B In addition to the constitutional issues raised in Caplin & Drysdale, respondent contends that freezing the assets in question before he is convicted—and before they are finally adjudged to be forfeitable—raises distinct constitutional concerns. We conclude, however, that assets in a defendant’s possession may be restrained in the way they were here based on a finding of probable cause to believe that the assets are forfeit able.10 We have previously permitted the Government to seize property based on a finding of probable cause to believe that the property will ultimately be proved forfeitable. See, e. g., United States v. $8,850, 461 U. S. 555 (1983); Calero-Toledo v. Pearson Yacht Leasing Co., 416 U. S. 663 (1974). Here, where respondent was not ousted from his property, but merely restrained from disposing of it, the governmental intrusion was even less severe than those permitted by our prior decisions. Indeed, it would be odd to conclude that the Government may not restrain property, such as the home and apartment in respondent’s possession, based on a finding of probable cause, when we have held that (under appropriate circumstances), the Government may restrain persons where there 10 We do not consider today, however, whether the Due Process Clause requires a hearing before a pretrial restraining order can be imposed. As noted above, in its initial consideration of this case, a panel of the Second Circuit ordered that such a hearing be held before permitting the entry of a restraining order; on remand, the District Court held an extensive, 4-day hearing on the question of probable cause. Though the United States petitioned for review of the Second Circuit’s holding that such a hearing was required, see Pet. for Cert. I, given that the Government prevailed in the District Court notwithstanding the hearing, it would be pointless for us now to consider whether a hearing was required by the Due Process Clause. Furthermore, because the Court of Appeals, in its en banc decision, did not address the procedural due process issue, we also do not inquire whether the hearing—if a hearing was required at all—was an adequate one. 616 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. is a finding of probable cause to believe that the accused has committed a serious offense. See United States v. Salerno, 481 U. S. 739 (1987). Given the gravity of the offenses charged in the indictment, respondent himself could have been subjected to pretrial restraint if deemed necessary to “reasonably assure [his] appearance [at trial] and the safety of. . . the community,” 18 U. S. C. § 3142(e) (1982 ed., Supp. V); we find no constitutional infirmity in § 853(e)’s authorization of a similar restraint on respondent’s property to protect its “appearance” at trial and protect the community’s interest in full recovery of any ill-gotten gains. Respondent contends that both the nature of the Government’s property right in forfeitable assets, and the nature of the use to which he would have put these assets (i. e., retaining an attorney), require some departure from our established rule of permitting pretrial restraint of assets based on probable cause. We disagree. In Caplin & Drysdale, we conclude that a weighing of these very interests suggests that the Government may—without offending the Fifth or Sixth Amendment—obtain forfeiture of property that a defendant might have wished to use to pay his attorney. Post, p. 617. Given this holding, we find that a pretrial restraining order does not “arbitrarily” interfere with a defendant’s “fair opportunity” to retain counsel. Cf. Powell n. Alabama, 287 U. S. 45, 69, 53 (1932). Put another way: if the Government may, post-trial, forbid the use of forfeited assets to pay an attorney, then surely no constitutional violation occurs when, after probable cause is adequately established, the Government obtains an order barring a defendant from frustrating that end by dissipating his assets prior to trial. IV For the reasons given above, the judgment of the Second Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 617 Syllabus CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT No. 87-1729. Argued March 21, 1989—Decided June 22, 1989 Christopher Reckmeyer was charged with running a massive drug importation and distribution scheme alleged to be a continuing criminal enterprise (CCE) in violation of 21 U. S. C. § 848. Relying on a portion of the CCE statute that authorizes forfeiture to the Government of property acquired as a result of drug-law violations, § 853, the indictment sought forfeiture of specified assets in Reckmeyer’s possession. The District Court, acting pursuant to § 853(e)(1)(A), entered a restraining order forbidding Reckmeyer from transferring any of the potentially forfeitable assets. Nonetheless, he transferred $25,000 to petitioner, a law firm, for preindictment legal services. Petitioner continued to represent Reckmeyer after his indictment. Reckmeyer moved to modify the District Court’s order to permit him to use some of the restrained assets to pay petitioner’s fees and to exempt such assets from postconviction forfeiture. However, before the court ruled on his motion, Reckmeyer entered a plea agreement with the Government in which, inter alia, he agreed to forfeit all of the specified assets. The court then denied Reckmeyer’s motion and, subsequently, entered an order forfeiting virtually all of his assets to the Government. Petitioner—arguing that assets used to pay an attorney are exempt from forfeiture under § 853 and, if they are not, that the statute’s failure to provide such an exemption renders it unconstitutional—filed a petition under § 853(n) seeking an adjudication of its third-party interest in the forfeited assets. The District Court granted the relief sought. However, the Court of Appeals reversed, finding that the statute acknowledged no exception to its forfeiture requirement and that the statutory scheme is constitutional. Held: 1. For the reasons stated in United States v. Monsanto, ante, at 611-614, whatever discretion § 853(e) does provide district court judges to refuse to issue pretrial restraining orders on potentially forfeitable assets, it does not grant them equitable discretion to allow a defendant to withhold assets to pay bona fide attorney’s fees. Nor does the exercise of judges’ § 853(e) discretion “immunize” nonrestrained assets used for attorney’s fees from subsequent forfeiture under § 853(c), which provides for recapture of forfeitable assets transferred to third parties. Pp. 622-623. 618 OCTOBER TERM, 1988 Syllabus 491 U. S. 2. The forfeiture statute does not impermissibly burden a defendant’s Sixth Amendment right to retain counsel of his choice. A defendant has no Sixth Amendment right to spend another person’s money for services rendered by an attorney even if those funds are the only way that that defendant will be able to retain the attorney of his choice. Such money, though in his possession, is not rightfully his. Petitioner’s contention that, since the Government’s claim to forfeitable assets rests on a penal statute that is merely a mechanism for preventing fraudulent conveyances of the assets and is not a device for determining true title to property, the burden the statute places on a defendant’s rights greatly outweighs the Government’s interest in forfeiture is unsound. Section 853(c) reflects the application of the long-recognized and lawful practice of vesting title to any forfeitable assets in the hands of the Government at the time of the criminal act giving rise to forfeiture. Moreover, there is a strong governmental interest in obtaining full recovery of the assets, since the assets are deposited in a fund that supports law-enforcement efforts, since the statute allows property to be recovered by its rightful owners, and since a major purpose behind forfeiture provisions such as the CCE’s is to lessen the economic power of organized crime and drug enterprises, including the use of such power to retain private counsel. Pp. 624-633. 3. The forfeiture statute does not upset the balance of power between the Government and the accused in a manner contrary to the Due Process Clause of the Fifth Amendment. The Constitution does not forbid the imposition of an otherwise permissible criminal sanction, such as forfeiture, merely because in some cases prosecutors may abuse the processes available to them. Such due process claims are cognizable only in specific cases of prosecutorial misconduct, which has not been alleged here. Pp. 633-635. 837 F. 2d 637, affirmed. White, J., delivered the opinion of the Court, in which Rehnquist, C. J., and O’Connor, Scalia, and Kennedy, JJ., joined. Blackmun, J., filed a dissenting opinion, in which Brennan, Marshall, and Stevens, JJ., joined, post, p. 635. Peter Van N. Lockwood argued the cause for petitioner. With him on the briefs were Graeme W. Bush, Albert G. Lauber, Jr., Julia L. Porter, and Robert L. Cohen. Acting Solicitor General Bryson argued the cause for the United States. With him on the briefs were Assistant CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 619 617 Opinion of the Court Attorney General Dennis, Edwin S. Kneedler, and Sara Criscitelli* Justice White delivered the opinion of the Court. We are called on to determine whether the federal drug forfeiture statute includes an exemption for assets that a defendant wishes to use to pay an attorney who conducted his defense in the criminal case where forfeiture was sought. Because we determine that no such exemption exists, we must decide whether that statute, so interpreted, is consistent with the Fifth and Sixth Amendments. We hold that it is. I In January 1985, Christopher Reckmeyer was charged in a multicount indictment with running a massive drug importation and distribution scheme. The scheme was alleged to be a continuing criminal enterprise (CCE), in violation of 84 Stat. 1265, as amended, 21 U. S. C. §848 (1982 ed., Supp. V). Relying on a portion of the CCE statute that authorizes forfeiture to the Government of “property constituting, or derived from . . . proceeds . . . obtained” from drug-law * Joseph Beeler and Bruce J. Winick filed a brief for the National Association of Criminal Defense Lawyers et al. as amici curiae urging reversal. Briefs of amici curiae urging affirmance were filed for the State of California by John K. Van de Kamp, Attorney General, Steve White, Chief Assistant Attorney General, John A. Gordnier, Senior Assistant Attorney General, and Gary W. Schons, Deputy Attorney General; and for Eugene R. Anderson, pro se. Briefs of amici curiae were filed for the American Bar Association by Robert D. Raven, Charles G. Cole, Antonia B. lanniello, and Terrance G. Reed; and for the Appellate Committee of the California District Attorneys Association by Ira Reiner, Harry B. Sondheim, and Arnold T. Guminski. 620 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. violations, § 853(a),1 the indictment sought forfeiture of specified assets in Reckmeyer’s possession. App. 33-40. At this time, the District Court, acting pursuant to § 853(e)(1) (A),2 entered a restraining order forbidding Reckmeyer to transfer any of the listed assets that were potentially forfeitable. Sometime earlier, Reckmeyer had retained petitioner, a law firm, to represent him in the ongoing grand jury investigation which resulted in the January 1985 indictments. Notwithstanding the restraining order, Reckmeyer paid the firm $25,000 for preindictment legal services a few days after the indictment was handed down; this sum was placed by petitioner in an escrow account. Petitioner continued to represent Reckmeyer following the indictment. 1 The forfeiture statute provides, in relevant part, that any person convicted of a particular class of criminal offenses “shall forfeit to the United States, irrespective of any provision of State law— “(1) any property constituting, or derived from, any proceeds the person obtained, directly or indirectly, as the result of such violation; “The court, in imposing sentence on such person, shall order, in addition to any other sentence imposed . . . , that the person forfeit to the United States all property described in this subsection.” 21 U. S. C. § 853(a) (1982 ed., Supp. V). There is no question here that the offenses Reckmeyer was accused of in the indictment fell within the class of crimes triggering this forfeiture provision. 2 The pretrial restraining order provision states that “[u]pon application of the United States, the court may enter a restraining order or injunction ... or take any other action to preserve the availability of property described in subsection (a) of [§ 853] for forfeiture under this section— “(A) upon the filing of an indictment or information charging a violation ... for which criminal forfeiture may be ordered under [§853] and alleging that the property with respect to which the order is sought would, in the event of conviction, be subject to forfeiture under this section. § 853(e)(1). CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 621 617 Opinion of the Court On March 7, 1985, Reckmeyer moved to modify the District Court’s earlier restraining order to permit him to use some of the restrained assets to pay petitioner’s fees; Reckmeyer also sought to exempt from any postconviction forfeiture order the assets that he intended to use to pay petitioner. However, one week later, before the District Court could conduct a hearing on this motion, Reckmeyer entered a plea agreement with the Government. Under the agreement, Reckmeyer pleaded guilty to the drug-related CCE charge, and agreed to forfeit all of the specified assets listed in the indictment. The day after the Reckmeyer’s plea was entered, the District Court denied his earlier motion to modify the restraining order, concluding that the plea and forfeiture agreement rendered irrelevant any further consideration of the propriety of the court’s pretrial restraints. App. 54-55. Subsequently, an order forfeiting virtually all of the assets in Reckmeyer’s possession was entered by the District Court in conjunction with his sentencing. Id., at 57-65. After this order was entered, petitioner filed a petition under § 853(n), which permits third parties with an interest in forfeited property to ask the sentencing court for an adjudication of their rights to that property; specifically, §853(n) (6)(B) gives a third party who entered into a bona fide transaction with a defendant a right to make claims against forfeited property, if that third party was “at the time of [the transaction] reasonably without cause to believe that the [defendant’s assets were] subject to forfeiture.” See also §853 (c). Petitioner claimed an interest in $170,000 of Reckmeyer’s assets, for services it had provided Reckmeyer in conducting his defense; petitioner also sought the $25,000 being held in the escrow account, as payment for preindictment legal services. Petitioner argued alternatively that assets used to pay an attorney were exempt from forfeiture under § 853, and if not, the failure of the statute to provide such an exemption rendered it unconstitutional. The District Court granted petitioner’s claim for a share of the forfeited assets. 622 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. A panel of the Fourth Circuit affirmed, finding that—while §853 contained no statutory provision authorizing the payment of attorney’s fees out of forfeited assets—the statute’s failure to do so impermissibly infringed a defendant’s Sixth Amendment right to the counsel of his choice. United States v. Harvey, 814 F. 2d 905 (1987). The Court of Appeals agreed to hear the case en banc and reversed. Sub nom. In re Forfeiture Hearing as to Caplin & Drysdale, Chartered, 837 F. 2d 637 (1988). All the judges of the Fourth Circuit agreed that the language of the CCE statute acknowledged no exception to its forfeiture requirement that would recognize petitioner’s claim to the forfeited assets. A majority found this statutory scheme constitutional, id., at 642-648; four dissenting judges, however, agreed with the panel’s view that the statute so construed violated the Sixth Amendment, id., at 651-653 (Phillips, J., dissenting). Petitioner sought review of the statutory and constitutional issues raised by the Court of Appeals’ holding. We granted certiorari, 488 U. S. 940 (1988), and now affirm. II Petitioner’s first submission is that the statutory provision that authorizes pretrial restraining orders on potentially forfeitable assets in a defendant’s possession, 21 U. S. C. § 853 (e) (1982 ed., Supp. V), grants district courts equitable discretion to determine when such orders should be imposed. This discretion should be exercised under “traditional equitable standards,” petitioner urges, including a “weighting] of the equities and competing hardships on the parties”; under this approach, a court “must invariably strike the balance so as to allow a defendant [to pay] . . . for bona fide attorneys fees,” petitioner argues. Brief for Petitioner 8. Petitioner further submits that once a district court so exercises its discretion, and fails to freeze assets that a defendant then uses to pay an attorney, the statute’s provision for recapture of CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 623 617 Opinion of the Court forfeitable assets transferred to third parties, § 853(c), may not operate on such sums. Petitioner’s argument, as it acknowledges, is based on the view of the statute expounded by Judge Winter of the Second Circuit in his concurring opinion in that Court of Appeals’ en banc decision, United States v. Monsanto, 852 F. 2d 1400, 1405-1411 (1988). We reject this interpretation of the statute today in our decision in United States v. Monsanto, ante, p. 600, which reverses the Second Circuit’s holding in that case. As we explain in our Monsanto decision, ante, at 611-614, whatever discretion § 853(e) provides district court judges to refuse to enter pretrial restraining orders, it does not extend as far as petitioner urges—nor does the exercise of that discretion “immunize” nonrestrained assets from subsequent forfeiture under § 853(c), if they are transferred to an attorney to pay legal fees. Thus, for the reasons provided in our opinion in Monsanto, we reject petitioner’s statutory claim. Ill We therefore address petitioner’s constitutional challenges to the forfeiture law.3 Petitioner contends that the statute 8 The United States argues that petitioner lacks jus tertii standing to advance Reckmeyer’s Sixth Amendment rights. See Brief for United States 35, and n. 17. Though the argument is not without force, we conclude that petitioner has the requisite standing. When a person or entity seeks standing to advance the constitutional rights of others, we ask two questions: first, has the litigant suffered some mjury-in-fact, adequate to satisfy Article Ill’s case-or-controversy requirement; and second, do prudential considerations which we have identified in our prior cases point to permitting the litigant to advance the claim? See Singleton v. Wulff, 428 U. S. 106, 112 (1976). As to the first inquiry, there can be little doubt that petitioner’s stake in $170,000 of the forfeited assets—which it would almost certainly receive if the Sixth Amendment claim it advances here were vindicated—is adequate injury-in-fact to meet the constitutional minimum of Article III standing. The second inquiry—the prudential one—is more difficult. To answer this question, our cases have looked at three factors: the relationship of the litigant to the person whose rights are being asserted; the ability of the 624 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. infringes on criminal defendants’ Sixth Amendment right to counsel of choice, and upsets the “balance of power” between the Government and the accused in a manner contrary to the Due Process Clause of the Fifth Amendment. We consider these contentions in turn. A Petitioner’s first claim is that the forfeiture law makes impossible, or at least impermissibly burdens, a defendant’s right “to select and be represented by one’s preferred attorney.” Wheat v. United States, 486 U. S. 153, 159 (1988). Petitioner does not, nor could it defensibly do so, assert that impecunious defendants have a Sixth Amendment right to choose their counsel. The Amendment guarantees defendants in criminal cases the right to adequate representation, but those who do not have the means to hire their own lawyers have no cognizable complaint so long as they are adequately represented by attorneys appointed by the courts. “[A] defendant may not insist on representation by an attorney he cannot afford.” Wheat, supra, at 159. Petitioner does not dispute these propositions. Nor does the Government deny that the Sixth Amendment guarantees a defendant the right to be represented by an otherwise qualified attorney whom that defendant can afford to hire, or who is willing to represent the defendant even though he is without person to advance his own rights; and the impact of the litigation on third-party interests. See, e. g., Craig v. Boren, 429 U. S. 190, 196 (1976); Singleton v. Wulff, supra, at 113-118; Eisenstadt v. Baird, 405 U. S. 438, 443-446 (1972). The second of these three factors counsels against review here: as Monsanto, ante, p. 600, illustrates, a criminal defendant suffers none of the obstacles discussed in Wulff, supra, at 116-117, to advancing his own constitutional claim. We think that the first and third factors, however, clearly weigh in petitioner’s favor. The attorney-client relationship between petitioner and Reckmeyer, like the doctor-patient relationship in Baird, is one of special consequence; and like Baird, it is credibly alleged that the statute at issue here may “materially impair the ability of” third persons in Reckmeyer’s position to exercise their constitutional rights. See Baird, supra, at 445. Petitioner therefore satisfies our requirements for jus tertii standing. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 625 617 Opinion of the Court funds. Applying these principles to the statute in question here, we observe that nothing in § 853 prevents a defendant from hiring the attorney of his choice, or disqualifies any attorney from serving as a defendant’s counsel. Thus, unlike Wheat, this case does not involve a situation where the Government has asked a court to prevent a defendant’s chosen counsel from representing the accused. Instead, petitioner urges that a violation of the Sixth Amendment arises here because of the forfeiture, at the instance of the Government, of assets that defendants intend to use to pay their attorneys. Even in this sense, of course, the burden the forfeiture law imposes on a criminal defendant is limited. The forfeiture statute does not prevent a defendant who has nonforfeitable assets from retaining any attorney of his choosing. Nor is it necessarily the case that a defendant who possesses nothing but assets the Government seeks to have forfeited will be prevented from retaining counsel of choice. Defendants like Reckmeyer may be able to find lawyers willing to represent them, hoping that their fees will be paid in the event of acquittal, or via some other means that a defendant might come by in the future. The burden placed on defendants by the forfeiture law is therefore a limited one. Nonetheless, there will be cases where a defendant will be unable to retain the attorney of his choice, when that defendant would have been able to hire that lawyer if he had access to forfeitable assets, and if there was no risk that fees paid by the defendant to his counsel would later be recouped under § 853(c).4 It is in these cases, petitioner argues, that the Sixth Amendment puts limits on the forfeiture statute. 4 That section of the statute, which includes the so-called “relation back” provision, states: “All right, title, and interest in property described in [§ 853] vests in the United States upon the commission of the act giving rise to forfeiture under this section. Any such property that is subsequently transferred to a person other than the defendant may be the subject of a special verdict of forfeiture and thereafter shall be forfeited to the United States, unless the transferee establishes” his entitlement to such property pursuant to §853(n), discussed supra. 21 U. S. C. § 853(c) (1982 ed., Supp. V). 626 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. This submission is untenable. Whatever the full extent of the Sixth Amendment’s protection of one’s right to retain counsel of his choosing, that protection does not go beyond “the individual’s right to spend his own money to obtain the advice and assistance of . . . counsel.” Walters n. National Assn, of Radiation Survivors, 473 U. S. 305, 370 (1985) (Stevens, J., dissenting). A defendant has no Sixth Amendment right to spend another person’s money for services rendered by an attorney, even if those funds are the only way that that defendant will be able to retain the attorney of his choice. A robbery suspect, for example, has no Sixth Amendment right to use funds he has stolen from a bank to retain an attorney to defend him if he is apprehended. The money, though in his possession, is not rightfully his; the Government does not violate the Sixth Amendment if it seizes the robbery proceeds and refuses to permit the defendant to use them to pay for his defense. “[N]o lawyer, in any case, . . . has the right to . . . accept stolen property, or. . . ransom money, in payment of a fee.. . . The privilege to practice law is not a license to steal.” Laska v. United States, 82 F. 2d 672, 677 (CAIO 1936). Petitioner appears to concede as much, see Brief for Petitioner 40, n. 25, as respondent in Monsanto clearly does, see Brief for Respondent in No. 88-454, pp. 36-37. Petitioner seeks to distinguish such cases for Sixth Amendment purposes by arguing that the bank’s claim to robbery proceeds rests on “pre-existing property rights,” while the Government’s claim to forfeitable assets rests on a “penal statute” which embodies the “Active property-law concept of . . . relation-back” and is merely “a mechanism for preventing fraudulent conveyances of the defendant’s assets, not. . . a device for determining true title to property.” Brief for Petitioner 40-41. In light of this, petitioner contends, the burden placed on defendant’s Sixth Amendment rights by the forfeiture statute outweighs the Government’s interest in forfeiture. Ibid. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 627 617 Opinion of the Court The premises of petitioner’s constitutional analysis are unsound in several respects. First, the property rights given the Government by virtue of the forfeiture statute are more substantial than petitioner acknowledges. In § 853(c), the so-called “relation-back” provision, Congress dictated that “[a]ll right, title and interest in property” obtained by criminals via the illicit means described in the statute “vests in the United States upon the commission of the act giving rise to forfeiture.” 21 U. S. C. §853(c) (1982 ed., Supp. V). As Congress observed when the provision was adopted, this approach, known as the “taint theory,” is one that “has long been recognized in forfeiture cases,” including the decision in United States v. Stowell, 133 U. S. 1 (1890). See S. Rep. No. 98-225, p. 200, and n. 27 (1983). In Stowell, the Court explained the operation of a similar forfeiture provision (for violations of the Internal Revenue Code) as follows: “As soon as [the possessor of the forfeitable asset committed the violation] of the internal revenue laws, the forfeiture under those laws took effect, and (though needing judicial condemnation to perfect it) operated from that time as a statutory conveyance to the United States of all the right, title and interest then remaining in the [possessor]; and was as valid and effectual, against all the world, as a recorded deed. The right so vested in the United States could not be defeated or impaired by any subsequent dealings of the . . . [possessor].” Stowell, supra, at 19. In sum, § 853(c) reflects the application of the long-recognized and lawful practice of vesting title to any forfeitable assets, in the United States, at the time of the criminal act giving rise to forfeiture. Concluding that Reckmeyer cannot give good title to such property to petitioner because he did not hold good title is neither extraordinary or novel. Nor does petitioner claim, as a general proposition that the relation-back provision is unconstitutional, or that Congress cannot, as a general matter, vest title to assets derived from the crime in 628 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the Government, as of the date of the criminal act in question. Petitioner’s claim is that whatever part of the assets that is necessary to pay attorney’s fees cannot be subjected to forfeiture. But given the Government’s title to Reckme-yer’s assets upon conviction, to hold that the Sixth Amendment creates some right in Reckmeyer to alienate such assets, or creates a right on petitioner’s part to receive these assets, would be peculiar. There is no constitutional principle that gives one person the right to give another’s property to a third party, even where the person seeking to complete the exchange wishes to do so in order to exercise a constitutionally protected right. While petitioner and its supporting amici attempt to distinguish between the expenditure of forfeitable assets to exercise one’s Sixth Amendment rights, and expenditures in the pursuit of other constitutionally protected freedoms, see, e. g., Brief for American Bar Association as Amicus Curiae 6, there is no such distinction between, or hierarchy among, constitutional rights. If defendants have a right to spend forfeitable assets on attorney’s fees, why not on exercises of the right to speak, practice one’s religion, or travel? The full exercise of these rights, too, depends in part on one’s financial wherewithal; and forfeiture, or even the threat of forfeiture, may similarly prevent a defendant from enjoying these rights as fully as he might otherwise. Nonetheless, we are not about to recognize an antiforfeiture exception for the exercise of each such right; nor does one exist for the exercise of Sixth Amendment rights.5 6 It would be particularly odd to recognize the Sixth Amendment as a defense to forfeiture, because forfeiture is a substantive charge in the indictment against a defendant. Thus, petitioner asks us to take the Sixth Amendment’s guarantee of counsel “for his defense” and make that guarantee petitioner’s defense to the indictment. We doubt that the Amendment’s guarantees, which are procedural in nature, cf. Faretta v. California, 422 U. S. 806, 818 (1975), provide such a substantive defense to charges against an accused. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 629 617 Opinion of the Court Petitioner’s “balancing analysis” to the contrary rests substantially on the view that the Government has only a modest interest in forfeitable assets that may be used to retain an attorney. Petitioner takes the position that, in large part, once assets have been paid over from client to attorney, the principal ends of forfeiture have been achieved: dispossessing a drug dealer or racketeer of the proceeds of his wrongdoing. See Brief for Petitioner 39; see also 814 F. 2d, at 924-925. We think that this view misses the mark for three reasons. First, the Government has a pecuniary interest in forfeiture that goes beyond merely separating a criminal from his ill-gotten gains; that legitimate interest extends to recovering all forfeitable assets, for such assets are deposited in a Fund that supports law-enforcement efforts in a variety of important and useful ways. See 28 U. S. C. § 524(c), which establishes the Department of Justice Assets Forfeiture Fund. The sums of money that can be raised for law-enforcement activities this way are substantial,6 and the Government’s interest in using the profits of crime to fund these activities should not be discounted. Second, the statute permits “rightful owners” of forfeited assets to make claims for forfeited assets before they are retained by the Government. See 21 U. S. C. § 853(n)(6)(A). The Government’s interest in winning undiminished forfeiture thus includes the objective of returning property, in full, to those wrongfully deprived or defrauded of it. Where the Government pursues this restitutionary end, the Government’s interest in forfeiture is virtually indistinguishable from its interest in returning to a bank the proceeds of a bank robbery; and a forfeiture-defendant’s claim of right to use 6 For example, just one of the assets which Reckmeyer agreed to forfeit, a parcel of land known as “Shelburne Glebe,” see App. 57 (forfeiture order), was recently sold by federal authorities for $5.3 million. Washington Post, May 10, 1989, p. DI, cols. 1-4. The proceeds of the sale will fund federal, state, and local law-enforcement activities. Ibid. 630 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. such assets to hire an attorney, instead of having them returned to their rightful owners, is no more persuasive than a bank robber’s similar claim. Finally, as we have recognized previously, a major purpose motivating congressional adoption and continued refinement of the racketeer influenced and corrupt organizations (RICO) and CCE forfeiture provisions has been the desire to lessen the economic power of organized crime and drug enterprises. See Russello v. United States, 464 U. S. 16, 27-28 (1983). This includes the use of such economic power to retain private counsel. As the Court of Appeals put it: “Congress has already underscored the compelling public interest in stripping criminals such as Reckmeyer of their undeserved economic power, and part of that undeserved power may be the ability to command high-priced legal talent.” 837 F. 2d, at 649. The notion that the Government has a legitimate interest in depriving criminals of economic power, even insofar as that power is used to retain counsel of choice, may be somewhat unsettling. See, e. g., Tr. of Oral Arg. 50-52. But when a defendant claims that he has suffered some substantial impairment of his Sixth Amendment rights by virtue of the seizure or forfeiture of assets in his possession, such a complaint is no more than the reflection of “the harsh reality that the quality of a criminal defendant’s representation frequently may turn on his ability to retain the best counsel money can buy.” Morris v. Slappy, 461 U. S. 1, 23 (1983) (Brennan, J., concurring in result). Again, the Court of Appeals put it aptly: “The modern day Jean Valjean must be satisfied with appointed counsel. Yet the drug merchant claims that his possession of huge sums of money . . . entitles him to something more. We reject this contention, and any notion of a constitutional right to use the proceeds of crime to finance an expensive defense.” 837 F. 2d, at 649.7 7 We also reject the contention, advanced by amici, see, e. g., Brief for American Bar Association as Amicus Curiae 20-22, and accepted by some courts considering claims like petitioner’s, see, e. g., United States v. Rogers, 602 F. Supp. 1332, 1349-1350 (Colo. 1985), that a type of “per se” in- CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 631 617 Opinion of the Court It is our view that there is a strong governmental interest in obtaining full recovery of all forfeitable assets, an interest that overrides any Sixth Amendment interest in permitting criminals to use assets adjudged forfeitable to pay for their defense. Otherwise, there would be an interference with a defendant’s Sixth Amendment rights whenever the Government freezes or takes some property in a defendant’s possession before, during, or after a criminal trial. So-called “jeopardy assessments”—Internal Revenue Service (IRS) seizures of assets to secure potential tax liabilities, see 26 U. S. C. § 6861—may impair a defendant’s ability to retain counsel in a way similar to that complained of here. Yet these assessments have been upheld against constitutional attack,8 and we note that the respondent in Monsanto concedes their constitutionality, see Brief for Respondent in No. 88-454, p. 37, n. 20. Moreover, petitioner’s claim to a share of the forfeited assets postconviction would suggest that the Government could never impose a burden on assets within a defendant’s control that could be used to pay a lawyer.9 Criminal defendants, however, are not exempted effective assistance of counsel results—due to the particular complexity of RICO or drug-enterprise cases—when a defendant is not permitted to use assets in his possession to retain counsel of choice, and instead must rely on appointed counsel. If such an argument were accepted, it would bar the trial of indigents charged with such offenses, because those persons would have to rely on appointed counsel—which this view considers per se ineffective. If appointed counsel is ineffective in a particular case, a defendant has resort to the remedies discussed in Strickland v. Washington, 466 U. S. 668 (1984). But we cannot say that the Sixth Amendment’s guarantee of effective assistance of counsel is a guarantee of a privately retained counsel in every complex case, irrespective of a defendant’s ability to pay. 8 See, e. g., Avco Delta Corporation Canada Ltd. v. United States, 484 F. 2d 692 (CA7 1973); Summers v. United States, 250 F. 2d 132, 133-135 (CA9 1957); United States v. Brodson, 241 F. 2d 107, 109-111 (CA7 1957) (en banc). 9 A myriad of other law-enforcement mechanisms operate in a manner similar to IRS jeopardy assessments, and might also be subjected to Sixth 632 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. from federal, state, and local taxation simply because these financial levies may deprive them of resources that could be used to hire an attorney. We therefore reject petitioner’s claim of a Sixth Amendment right of criminal defendants to use assets that are the Government’s—assets adjudged forfeitable, as Reckmeyer’s were—to pay attorney’s fees, merely because those assets are in their possession.10 See also Monsanto, ante, at 613, Amendment invalidation if petitioner’s claim were accepted. See Brickey, Attorneys’ Fee Forfeitures, 36 Emory L. J. 761, 770-772 (1987). 10 Petitioner advances three additional reasons for invalidating the forfeiture statute, all of which concern possible ethical conflicts created for lawyers defending persons facing forfeiture of assets in their possession. See Brief for Petitioner 35-37; see also Brief for American Bar Association as Amicus Curiae 17-22. Petitioner first notes the statute’s exemption from forfeiture of property transferred to a bona fide purchaser who was “reasonably without cause to believe that the property was subject to forfeiture.” 21 U. S. C. § 853(n)(6)(B). This provision, it is said, might give an attorney an incentive not to investigate a defendant’s case as fully as possible, so that the lawyer can invoke it to protect from forfeiture any fees he has received. Yet given the requirement that any assets which the Government wishes to have forfeited must be specified in the indictment, see Fed. Rule Crim. Proc. 7(c)(2), the only way a lawyer could be a beneficiary of § 853(n)(6)(B) would be to fail to read the indictment of his client. In this light, the prospect that a lawyer might find himself in conflict with his client, by seeking to take advantage of § 853(n)(6)(B), amounts to very little. Petitioner itself concedes that such a conflict will, as a practical matter, never arise: a defendarit’s “lawyer . . . could not demonstrate that he was ‘reasonably without cause to believe that the property was subject to forfeiture,’ ” petitioner concludes at one point. Brief for Petitioner 31. The second possible conflict arises in plea bargaining: petitioner posits that a lawyer may advise a client to accept an agreement entailing a more harsh prison sentence but no forfeiture—even where contrary to the client’s interests—in an effort to preserve the lawyer’s fee. Following such a strategy, however, would surely constitute ineffective assistance of counsel. We see no reason why our cases such as Strickland v. Washington, 466 U. S. 668 (1984), are inadequate to deal with any such ineffectiveness where it arises. In any event, there is no claim that such conduct occurred here, nor could there be, as Reckmeyer’s plea agreement included forfeit- CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 633 617 Opinion of the Court which rejects a similar claim with respect to pretrial orders and assets not yet judged forfeitable. B Petitioner’s second constitutional claim is that the forfeiture statute is invalid under the Due Process Clause of the Fifth Amendment because it permits the Government to upset the “balance of forces between the accused and his accuser.” Wardius v. Oregon, 412 U. S. 470, 474 (1973). We are not sure that this contention adds anything to petitioner’s Sixth Amendment claim, because, while “[t]he Constitution guarantees a fair trial through the Due Process Clauses . . . it defines the basic elements of a fair trial largely through the several provisions of the Sixth Amendment,” Strickland v. Washington, 466 U. S. 668, 684-685 (1984). We have concluded above that the Sixth Amendment is not offended by the forfeiture provisions at issue here. Even if, however, the Fifth Amendment provides some added protection not encompassed in the Sixth Amendment’s more specific provisions, we find petitioner’s claim based on the Fifth Amendment unavailing. ure of virtually every asset in his possession. Moreover, we rejected a claim similar to this one in Evans v. Jeff D., 475 U. S. 717, 727-728 (1986). Finally, petitioner argues that the forfeiture statute, in operation, will create a system akin to “contingency fees” for defense lawyers: only a defense lawyer who wins acquittal for his client will be able to collect his fees, and contingent fees in criminal cases are generally considered unethical. See ABA Model Rule of Professional Conduct 1.5(d)(2) (1983); ABA Model Code of Professional Responsibility DR 2-106(0 (1979). But there is no indication here that petitioner, or any other firm, has actually sought to charge a defendant on a contingency basis; rather the claim is that a law firm’s prospect of collecting its fee may turn on the outcome at trial. This, however, may often be the case in criminal defense work. Nor is it clear why permitting contingent fees in criminal cases—if that is what the forfeiture statute does—violates a criminal defendant’s Sixth Amendment rights. The fact that a federal statutory scheme authorizing contingency fees—again, if that is what Congress has created in §853 (a premise we doubt)—is at odds with model disciplinary rules or state disciplinary codes hardly renders the federal statute invalid. 634 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Forfeiture provisions are powerful weapons in the war on crime; like any such weapons, their impact can be devastating when used unjustly. But due process claims alleging such abuses are cognizable only in specific cases of prosecutorial misconduct (and petitioner has made no such allegation here) or when directed to a rule that is inherently unconstitutional. “The fact that the . . . Act might operate unconstitutionally under some conceivable set of circumstances is insufficient to render it. . . invalid,” United States v. Salerno, 481 U. S. 739, 745 (1987). Petitioner’s claim—that the power available to prosecutors under the statute could be abused— proves too much, for many tools available to prosecutors can be misused in a way that violates the rights of innocent persons. As the Court of Appeals put it, in rejecting this claim when advanced below: “Every criminal law carries with it the potential for abuse, but a potential for abuse does not require a finding of facial invalidity.” 837 F. 2d, at 648. We rejected a claim similar to petitioner’s last Term, in Wheat v. United States, 486 U. S. 153 (1988). In Wheat, the petitioner argued that permitting a court to disqualify a defendant’s chosen counsel because of conflicts of interest— over that defendant’s objection to the disqualification—would encourage the Government to “manufacture” such conflicts to deprive a defendant of his chosen attorney. Id., at 163. While acknowledging that this was possible, we declined to fashion the per se constitutional rule petitioner sought in Wheat, instead observing that “trial courts are undoubtedly aware of [the] possibility” of abuse, and would have to “take it into consideration,” when dealing with disqualification motions. A similar approach should be taken here. The Constitution does not forbid the imposition of an otherwise permissible criminal sanction, such as forfeiture, merely because in some cases prosecutors may abuse the processes available to them, e. g., by attempting to impose them on persons who should not be subjected to that punishment. Cf. Brady v. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 635 617 Blackmun, J., dissenting United States, 397 U. S. 742, 751, and n. 8 (1970). Cases involving particular abuses can be dealt with individually by the lower courts, when (and if) any such cases arise. IV For the reasons given above, we find that petitioner’s statutory and constitutional challenges to the forfeiture imposed here are without merit. The judgment of the Court of Appeals is therefore Affirmed. Justice Blackmun, with whom Justice Brennan, Justice Marshall, and Justice Stevens join, dissenting.* Those jurists who have held forth against the result the majority reaches in these cases have been guided by one core insight: that it is unseemly and unjust for the Government to beggar those it prosecutes in order to disable their defense at trial. The majority trivializes “the burden the forfeiture law imposes on a criminal defendant.” Caplin & Drysdale, Chartered v. United States, ante, at 625. Instead, it should heed the warnings of our District Court judges, whose day-to-day exposure to the criminal-trial process enables them to understand, perhaps far better than we, the devastating consequences of attorney’s fee forfeiture for the integrity of our adversarial system of justice.1 * [This opinion applies also to No. 88-454, United States v. Monsanto, ante, p. 600.] 'See, e. g., United States v. Rogers, 602 F. Supp. 1332 (Colo. 1985); United States v. Badalamenti, 614 F. Supp. 194 (SDNY 1985); United States v. Reckmeyer, 631 F. Supp. 1191, 1197 (ED Va. 1986), aff’d on other grounds sub nom. United States v. Harvey, 814 F. 2d 905 (CA4 1987), rev’d sub nom. In re Forfeiture Hearing as to Caplin & Drysdale, Chartered, 837 F. 2d 637 (CA4 1988) (en banc); United States v. Bassett, 632 F. Supp. 1308,1317 (Md. 1986), aff’d on other grounds sub nom. United States v. Harvey, 814 F. 2d 905 (CA4 1987); United States v. lanniello, 644 F. Supp. 452 (SDNY 1985); United States v. Estevez, 645 F. Supp. 869 (ED Wis. 1986), app. dism’d for untimeliness, 852 F. 2d 239 (CA7 1988). 636 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. The criminal-forfeiture statute we consider today could have been interpreted to avoid depriving defendants of the ability to retain private counsel—and should have been so interpreted, given the grave “constitutional and ethical problems” raised by the forfeiture of funds used to pay legitimate counsel fees. United States v. Badalamenti, 614 F. Supp. 194, 196 (SDNY 1985). But even if Congress in fact required this substantial incursion on the defendant’s choice of counsel, the Court should have recognized that the Framers stripped Congress of the power to do so when they added the Sixth Amendment to our Constitution. I The majority acknowledges, as it must, that no language in the Comprehensive Forfeiture Act of 1984 (Act), ch. 3, 98 Stat. 2040, as amended, codified in relevant part at 21 U. S. C. §853 et seq. (1982 ed., Supp. V), expressly provides for the forfeiture of attorney’s fees, and that the legislative history contains no substantive discussion of the question. United States v. Monsanto, ante, at 608-609, and n. 8.2 The fact that “the legislative history and congressional debates are similarly silent on the use of forfeitable assets to pay stockbroker’s fees, laundry bills, or country club memberships,” ante, at 608-609, means nothing, for one cannot believe that Congress was unaware that interference with the payment of attorney’s fees, unlike interference with these other expenditures, would raise Sixth Amendment concerns. See Edward J. DeBartolo Corp. v. Florida Gulf Coast Building & Construction Trades Council, 485 U. S. 568, 575 (1988). 2 Indeed, the strongest statement on the question is the comment in the House Report: “Nothing in this section is intended to interfere with a person’s Sixth Amendment right to counsel.” H. R. Rep. No. 98-845, pt. 1, p. 19, n. 1 (1984). Even if the majority were correct that this statement is “nothing more than an exhortation for the courts to tread carefully in this delicate area,” United States v. Monsanto, ante, at 609, n. 8, the majority does not explain why it proceeds to ignore Congress’ exhortation to construe the statute to avoid implicating Sixth Amendment concerns. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 637 617 Blackmun, J., dissenting Despite the absence of any indication that Congress intended to use the forfeiture weapon against legitimate attorney’s fees, the majority—all the while purporting to “respect” the established practice of construing a statute to avoid constitutional problems, Monsanto, ante, at 611—contends that it is constrained to conclude that the Act reaches attorney’s fees. The Court cannot follow its usual practice here, we are told, because this is not a “close cas[e]” in which “statutory language is ambiguous.” Ibid. The majority finds unambiguous language in 21 U. S. C. § 853(a), which provides that when a defendant is convicted of certain crimes, the defendant “shall forfeit to the United States” any property derived from proceeds of the crime or used to facilitate the crime. I agree that § 853(a) is broad in language and is cast in mandatory terms.3 But I do not agree with the majority’s conclusion that the lack of an express exemption for attorney’s fees in § 853(a) makes the Act as a whole unambiguous. The majority succeeds in portraying the Act as “unambiguous” by making light of its most relevant provisions. As Judge Winter observed, the broad mandatory language of § 853(a) applies by its terms only to “ ‘any person convicted’ of the referenced crimes.” United States v. Monsanto, 852 F. 2d 1400, 1410 (CA2 1988). Because third parties to whom assets have been transferred in return for services rendered are not “person[s] convicted,” however, forfeiture of property in their possession is controlled by § 853(c) rather than by § 853(a). Section 853(c) provides: “Any such property that is subsequently transferred to a person other than the defendant may be the subject of a special verdict of forfeiture and thereafter shall be ordered forfeited to the United States” (emphasis added) if the third party fails to satisfy cer 3 As the majority acknowledges, so did Judge Winter, whose interpretation of the Act Caplin & Drysdale and Monsanto adopt in their briefs to this Court. See Monsanto, ante, at 607; United States v. Monsanto, 852 F. 2d 1400, 1409-1410 (CA2 1988) (en banc) (Winter, J., concurring). 638 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. tain requirements for exemption. Thus, § 853(c) does not, like § 853(a), provide that all property defined as forfeitable under §853 “must” or “shall” be forfeited:4 forfeitable property held by a third party presumptively “shall be ordered forfeited” only if it is included in the special verdict, and its inclusion in the verdict is discretionary.5 There is also considerable room for discretion in the language of § 853(e)(1), which controls the Government’s use of postindictment protective orders to prevent the preconviction transfer of potentially forfeitable assets to third parties. That section provides: “Upon application of the United States, the court may enter a restraining order or injunction ... or take any 4 This language differs from the language in Federal Rule of Criminal Procedure 31(e), which was promulgated in 1972 to provide procedural rules for Congress’ earlier forays into criminal forfeiture. The Rule provides: “If the indictment or the information alleges that an interest or property is subject to criminal forfeiture, a special verdict shall be returned as to the extent of the interest or property subject to forfeiture, if any.” (Emphasis added.) Congress’ decision to depart from mandatory language in § 853(c), where it fashioned a special verdict provision for assets transferred to third parties, is significant. 5 That the Act is mandatory in its treatment of forfeiture of property in the defendant’s hands, but not in its treatment of property transferred to third parties, is consistent with the distinction between civil forfeiture and criminal forfeiture. The theory (or, more properly, the fiction) underlying civil forfeiture is that the property subject to forfeiture is itself tainted by having been used in an unlawful manner. The right of the Government to take possession does not depend on the Government’s ultimately convicting the person who used the property in an unlawful way, nor is it diminished by the innocence or bona fides of the party into whose hands the property falls. See United States v. Stowell, 133 U. S. 1 (1890). Criminal forfeiture, in contrast, is penal in nature: it is predicated on the adjudicated guilt of the defendant, and has punishment of the defendant as its express purpose. See generally Cloud, Forfeiting Defense Attorneys’ Fees: Applying an Institutional Role Theory to Define Individual Constitutional Rights, 1987 Wis. L. Rev. 1, 18-19. Where the purpose of forfeiture is to punish the defendant, the Government’s penal interests are weakest when the punishment also burdens third parties. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 639 617 Blackmun, J., dissenting other action to preserve the availability of property . . . for forfeiture under this section . . . upon the filing of an indictment or information charging a violation . . . for which criminal forfeiture may be ordered . . . and alleging that the property with respect to which the order is sought would, in the event of conviction, be subject to forfeiture under this section” (emphasis added). The Senate Report makes clear that a district court may hold a hearing to “consider factors bearing on the reasonableness of the order sought.” S. Rep. No. 98-225, p. 202 (1983). Even if the court chooses to enter an order ex parte at the Government’s request, it may “modify the order” if it later proves to be unreasonable. Id., at 203. In the course of this process, the court may also consider the circumstances of any third party whose interests are implicated by the restraining order. Id., at 206, n. 42. Thus, the Government does not have an absolute right to an order preserving the availability of property by barring its transfer to third parties. Preconviction injunctive relief is available, but at the discretion of the district court. The majority does not deny that §§ 853(c) and 853(e)(1) contain discretionary language. It argues, however, that the exercise of discretion must be “cabined by the purposes” of the Act. Monsanto, ante, at 613. That proposition, of course, is unassailable: I agree that discretion created by the Act cannot be used to defeat the purposes of the Act. The majority errs, however, in taking an overly broad view of the Act’s purposes. Under the majority’s view, the Act aims to preserve the availability of all potentially forfeitable property during the preconviction period, and to achieve the forfeiture of all such property upon conviction. Ibid. This view of the Act’s purposes effectively writes all discretion out of §§ 853(c) and 853(e)(1), because any exercise of discretion will diminish the Government’s postconviction “take.” But a review of the legislative history of the Act demonstrates that 640 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. the Act does not seek forfeiture of property for its own sake merely to maximize the amount of money the Government collects.6 The central purposes of the Act, properly understood, are fully served by an approach to forfeiture that leaves ample room for the exercise of statutory discretion. Congress’ most systematic goal for criminal forfeiture was to prevent the profits of criminal activity from being poured into future such activity, for “it is through economic power that [criminal activity] is sustained and grows.” Senate Report, at 191. “Congress recognized in its enactment of statutes specifically addressing organized crime and illegal drugs that the conviction of individual racketeers and drug dealers would be of only limited effectiveness if the economic power bases of criminal organizations or enterprises were left intact, and so included forfeiture authority designed to strip these offenders and organizations of their economic power.” Ibid.; see also H. R. Rep. No. 98-845, pt. 1, p. 6 (1984) (criminal forfeiture statutes are “a bold attempt to attack the economic base of the criminal activity”).7 6 In adopting this view of the Act, the majority ignores the Government’s concession at oral argument before the en banc Court of Appeals for the Second Circuit that the Act was not enacted as a revenue-raising measure. See United States v. Monsanto, 852 F. 2d, at 1407, and n. 1 (Winter, J., concurring). Thus, although the Government’s interest in “using the profits of crime to fund [law-enforcement] activities” should perhaps not be “discounted,” Caplin & Drysdale, ante, at 629, it is not dispositive. Nor does Congress’ willingness to return forfeited funds to victims of crime instead of using them for law-enforcement purposes indicate that restitution is a primary goal of the Act. See ante, at 629-630. Restitution, in any event, is not a likely result in the typical case for which the Act was designed: one in which the property forfeited consists of derivative proceeds of illegal activity, rather than of stolen property that is readily traceable to a particular victim. See Cloud, 1987 Wis. L. Rev., at 20. 7 The majority contends that “the desire to lessen the economic power of organized crime and drug enterprises . . . includes the use of such economic power to retain private counsel.” Caplin & Drysdale, ante, at 630. “The notion that the Government has a legitimate interest in depriving criminals”—before they are convicted—“of economic power, even insofar as CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 641 617 Blackmun, J., dissenting Congress also had a more traditional punitive goal in mind: to strip convicted criminals of all assets purchased with the proceeds of their criminal activities. Particularly in the area of drug trafficking, Congress concluded that crime had become too lucrative for criminals to be deterred by conventional punishments. “Drug dealers have been able to accumulate huge fortunes as a result of their illegal activities. The sad truth is that the financial penalties for drug dealing are frequently only seen by dealers as a cost of doing business.” Id., at 2. The image of convicted drug dealers returning home from their prison terms to all the comforts their criminal activity can buy is one Congress could not abide.8 Finally, Congress was acutely aware that defendants, if unhindered, routinely would defeat the purposes of the Act by sheltering their assets in order to preserve them for their own future use and for the continued use of their criminal organizations. The purpose of § 853(c) is to “to permit the voiding of certain pre-conviction transfers and so close a potential loophole in current law whereby the criminal forfeiture sanction could be avoided by transfers that were not ‘arms’ length’ transactions.” Senate Report, at 200-201. With these purposes in mind, it becomes clear that a district court acts within the bounds of its statutory discretion that power is used to retain counsel of choice” is more than just “somewhat unsettling,” as the majority suggests. Ibid. That notion is constitutionally suspect, and—equally important for present purposes—completely foreign to Congress’ stated goals. The purpose of the relation-back provision is to assure that assets proved at trial to be the product of criminal activity cannot be channeled into further criminal activity—not to strip defendants of their assets on no more than a showing of probable cause that they are “tainted.” See United States v. Bassett, 632 F. Supp., at 1316; Comment, 61 N. Y. U. L. Rev. 124, 139 (1986). For its contrary view, the majority relies on nothing more than the rhetoric of the en banc Court °f Appeals’ majority opinion in Caplin & Drysdale. 8 Congress’ desire to maximize punishment, however, cannot be viewed as a blanket authorization of Government action that punishes the defendant before he is proved guilty. 642 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. when it exempts from preconviction restraint and postconviction forfeiture those assets a defendant needs to retain private counsel for his criminal trial. Assets used to retain counsel by definition will be unavailable to the defendant or his criminal organization after trial, even if the defendant is eventually acquitted. See Cloud, Government Intrusions Into the Attorney-Client Relationship: The Impact of Fee Forfeitures on the Balance of Power in the Adversary System of Criminal Justice, 36 Emory L. J. 817, 832 (1987). Thus, no important and legitimate purpose is served by employing § 853(c) to require postconviction forfeiture of funds used for legitimate attorney’s fees, or by employing § 853(e)(1) to bar preconviction payment of fees. The Government’s interests are adequately protected so long as the district court supervises transfers to the attorney to make sure they are made in good faith.9 See Comment, 61 N. Y. U. L. Rev. 124, 138-139 (1986). All that is lost is the Government’s power to punish the defendant before he is convicted. That power is not one the Act intended to grant.10 9 Judge Winter noted that the same logic suggests that the forfeiture of assets the defendant uses to support himself and his family is unduly harsh and is not necessary to achieve the goals of the Act. United States v. Monsanto, 852 F. 2d, at 1405. The majority chides Judge Winter for suggesting that, once it is established that there is discretion to exclude assets used to pay attorney’s fees and normal living expenses from forfeiture, the necessary result is that such assets must be excluded. Monsanto, ante, at 612-613. I find it exceedingly unlikely that a district court, instructed that it had the discretion to permit a defendant to retain counsel, would ever choose not to do so. Normal equitable considerations, combined with a proper regard for Sixth Amendment interests, would weigh so strongly in favor of that result that any “slippage” from permissive to mandatory language on Judge Winter’s part seems to me entirely accurate as a predictive matter. 10 The majority states in Monsanto, ante, at 610-611, that another forfeiture statute contemporaneous with the Act contains “the precise exemption from forfeiture which respondent asks us to imply into § 853,” and suggests that this is evidence that “Congress understood what it was doing in CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 643 617 Blackmun, J., dissenting A careful analysis of the language of the Act and its legislative history thus proves that “a construction of the statute is fairly possible by which the [constitutional] question may be avoided.” Crowell v. Benson, 285 U. S. 22, 62 (1932).11 Indeed, the prudentially preferable construction is also the only one that gives full effect to the discretionary language in §§ 853(c) and 853(e)(1). Thus, “if anything remains of the canon that statutes capable of differing interpretations should be construed to avoid constitutional issues... it surely applies here.” United States v. Monsanto, 852 F. 2d, at 1409. omitting such an exemption” from the Act. This argument is makeweight. The express exemption to which the majority refers involves the use of proceeds from publications and other accounts of a crime to: “(i) satisfy a money judgment rendered in any court in favor of a victim of any offense for which such defendant has been convicted, or a legal representative of such victim; and “(ii) pay for legal representation of the defendant in matters arising from the offense for which such defendant has been convicted, but no more than 20 percent of the total proceeds may be so used.” Pub. L. 98-473, § 1406(c)(1)(B), 98 Stat. 2175, codified as 18 U. S. C. § 3681(c)(1)(B) (1982 ed., Supp. V) (emphasis added). When this provision is read in context, it is clear that it concerns payment of attorney’s fees related to postconviction civil suits brought against convicted defendants by their victims. It does not, therefore, constitute the “precise exemption” sought in these cases. Indeed, the provision cuts against the result the majority reaches. In light of Congress’ decision to permit a convicted criminal to use wealth he has obtained by publicizing his crime to hire counsel to resist his victim’s damages claims, it would be bizarre to think that Congress intended to be more punitive when it comes to a defendant’s need for counsel prior to conviction, when the defendant’s own liberty is at stake. u For this reason, I need not rely on NLRB v. Catholic Bishop of Chicago, 440 U. S. 490, 500 (1979), in which the Court held that even the broadest statutory language may be interpreted as excluding cases that would raise serious constitutional questions, absent a clear expression of an affirmative intention of Congress to include those cases. See also Edward J. De Bartolo Corp. v. Florida Gulf Coast Building & Construction Trades Council, 485 U. S. 568 (1988). Under the Catholic Bishop approach, however, there could be no doubt that “the required ‘clearest indication in the 644 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. II The majority has decided otherwise, however, and for that reason is compelled to reach the constitutional issue it could have avoided. But the majority pauses hardly long enough to acknowledge “the Sixth Amendment’s protection of one’s right to retain counsel of his choosing,” let alone to explore its “full extent.” Caplin & Drysdale, ante, at 626. Instead, ante, at 624, it moves rapidly from the observation that “ ‘ [a] defendant may not insist on representation by an attorney he cannot afford,’” quoting Wheat v. United States, 486 U. S. 153, 159 (1988), to the conclusion that the Government is free to deem the defendant indigent by declaring his assets “tainted” by criminal activity the Government has yet to prove. That the majority implicitly finds the Sixth Amendment right to counsel of choice so insubstantial that it can be outweighed by a legal fiction demonstrates, still once again, its “ ‘apparent unawareness of the function of the independent lawyer as a guardian of our freedom.’” See id., at 172 (Stevens, J., dissenting), quoting Walters v. National Assn, of Radiation Survivors, 473 U. S. 305, 371 (1985) (Stevens, J., dissenting). A Over 50 years ago, this Court observed: “It is hardly necessary to say that the right to counsel being conceded, a defendant should be afforded a fair opportunity to secure counsel of his own choice.” Powell v. Alabama, 287 U. S. 45, 53 (1932). For years, that proposition was settled; the controversial question was whether the defendant’s right to use his own funds to retain his chosen counsel was the outer limit of the right protected by the Sixth Amendment. See, e. g., Chandler v. Fretag, 348 U. S. 3, 9 (1954). The Court’s subsequent decisions have made clear that an indigent defendant has the right to appointed counsel, see, e. g., Gideon v. legislative history’” or statutory language is absent here. 485 U. S., at 578. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 645 617 Blackmun, J., dissenting Wainwright, 372 U. S. 335 (1963), and that the Sixth Amendment guarantees at least minimally effective assistance of counsel, see, e. g., Strickland v. Washington, 466 U. S. 668 (1984). But while court appointment of effective counsel plays a crucial role in safeguarding the fairness of criminal trials, it has never defined the outer limits of the Sixth Amendment’s demands. The majority’s decision in Caplin & Drysdale reveals that it has lost track of the distinct role of the right to counsel of choice in protecting the integrity of the judicial process, a role that makes “the right to be represented by privately retained counsel . . . the primary, preferred component of the basic right” protected by the Sixth Amendment. United States v. Harvey, 814 F. 2d 905, 923 (CA4 1987), rev’d sub nom. In re Forfeiture Hearing as to Caplin & Drysdale, Chartered, 837 F. 2d 637 (CA4 1988) (en banc). The right to retain private counsel serves to foster the trust between attorney and client that is necessary for the attorney to be a truly effective advocate. See ABA Standards for Criminal Justice 4-3.1, p. 4-29 (commentary) (2d ed. 1980). Not only are decisions crucial to the defendant’s liberty placed in counsel’s hands, see Faretta v. California, 422 U. S. 806 (1975), but the defendant’s perception of the fairness of the process, and his willingness to acquiesce in its results, depend upon his confidence in his counsel’s dedication, loyalty, and ability. Cf. Joint Anti-Fascist Refugee Committee v. McGrath, 341 U. S. 123, 171-172 (1951) (Frankfurter, J., concurring). When the Government insists upon the right to choose the defendant’s counsel for him, that relationship of trust is undermined: counsel is too readily perceived as the Government’s agent rather than his own. Indeed, when the Court in Faretta held that the Sixth Amendment prohibits a court from imposing appointed counsel on a defendant who prefers to represent himself, its decision was predicated on the insight that “[t]o force a lawyer 646 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. on a defendant can only lead him to believe that the law contrives against him.” 422 U. S., at 834. The right to retain private counsel also serves to assure some modicum of equality between the Government and those it chooses to prosecute. The Government can be expected to “spend vast sums of money ... to try defendants accused of crime,” Gideon v. Wainwright, 372 U. S., at 344, and of course will devote greater resources to complex cases in which the punitive stakes are high. Precisely for this reason, “there are few defendants charged with crime, few indeed, who fail to hire the best lawyers they can get to prepare and present their defenses.” Ibid. But when the Government provides for appointed counsel, there is no guarantee that levels of compensation and staffing will be even average.12 Where cases are complex, trials long, and stakes high, that problem is exacerbated. “Despite the legal profession’s commitment to pro bono work,” United States v. Bassett, 632 F. Supp. 1308, 1316 (Md. 1986), aff’d on other grounds sub nom. United States v. Harvey, 814 F. 2d 905 (CA4 1987), even the best intentioned of attorneys may have no choice but to decline the task of representing defendants in cases for which they will not receive adequate compensation. See, e. g., United States v. Rogers, 602 F. Supp. 1332, 1349 (Colo. 1985). Over the long haul, the result of lowered compensation levels will be that talented attorneys will “decline to enter criminal practice. . . . This exodus of talented 12 “Even in the federal courts under the Criminal Justice Act of 1964,18 U. S. C. § 3006A, which provides one of the most generous compensation plans, the rates for appointed counsel. . . are low by American standards. Consequently, the majority of persons willing to accept appointments are the young and inexperienced.” Argersinger v. Hamlin, 407 U. S. 25, 57, n. 21 (1972) (Powell, J., concurring in result). Indeed, there is evidence that “Congress did not design [the Criminal Justice Act] to be compensatory, but merely to reduce financial burdens on assigned counsel.” See Winick, Forfeiture of Attorneys’ Fees under RICO and CCE and the Right to Counsel of Choice: The Constitutional Dilemma and How to Avoid It, 43 U. Miami L. Rev. 765, 773, and n. 40 (1989). CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 647 617 Blackmun, J., dissenting attorneys could devastate the criminal defense bar. ” Winick, Forfeiture of Attorneys’ Fees under RICO and CCE and the Right to Counsel of Choice: The Constitutional Dilemma and How to Avoid It, 43 U. Miami L. Rev. 765, 781 (1989). Without the defendant’s right to retain private counsel, the Government too readily could defeat its adversaries simply by outspending them.13 The right to privately chosen and compensated counsel also serves broader institutional interests. The “virtual socialization of criminal defense work in this country” that would be the result of a widespread abandonment of the right to retain chosen counsel, Brief for Committees on Criminal Advocacy and Criminal Law of the Association of the Bar of the City of New York et al. as Amici Curiae in No. 88-454, p. 9, too readily would standardize the provision of criminaldefense services and diminish defense counsel’s independence. There is a place in our system of criminal justice for the maverick and the risk taker and for approaches that might not fit into the structured environment of a public defender’s office, or that might displease a judge whose preference for nonconfrontational styles of advocacy might influence the judge’s appointment decisions. See Bazelon, The Defective Assistance of Counsel, 42 U. Cin. L. Rev. 1, 6-7 (1973); S. Kadish, S. Schulhofer, & M. Paulsen, Criminal Law and its Processes 32 (4th ed. 1983); cf. Sacher v. United States, 343 U. S. 1, 8-9 (1952) (“The nature of the proceedings presupposes, or at least stimulates, zeal in the opposing lawyers”). There is also a place for the employment of “specialized defense counsel” for technical and complex cases, see United States v. Thier, 801 F. 2d 1463, 1476 (CA5 1986) (concurring opinion), modification not relevant here, 809 F. 2d 249 (1987). The choice of counsel is the primary means for the defendant to establish the kind of defense he will put forward. See 13 That the Government has this power when the defendant is indigent is unfortunate, but “[i]t is an irrelevancy once recognized.” United States v. Harvey, 814 F. 2d, at 923. 648 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. United States v. Laura, 607 F. 2d 52, 56 (CA3 1979). Only a healthy, independent defense bar can be expected to meet the demands of the varied circumstances faced by criminal defendants, and assure that the interests of the individual defendant are not unduly “subordinat[ed] ... to the needs of the system.” Bazelon, 42 U. Cin. L. Rev., at 7. In sum, our chosen system of criminal justice is built upon a truly equal and adversarial presentation of the case, and upon the trust that can exist only when counsel is independent of the Government. Without the right, reasonably exercised, to counsel of choice, the effectiveness of that system is imperiled. B Had it been Congress’ express aim to undermine the adversary system as we know it, it could hardly have found a better engine of destruction than attomey’s-fee forfeiture. The main effect of forfeitures under the Act, of course, will be to deny the defendant the right to retain counsel, and therefore the right to have his defense designed and presented by an attorney he has chosen and trusts.14 If the Government restrains the defendant’s assets before trial, private counsel will be unwilling to continue, or to take on, the defense. Even if no restraining order is entered, the possibility of forfeiture after conviction will itself substantially 14 There is reason to fear that, in addition to depriving a defendant of counsel of choice, there will be circumstances in which the threat of forfeiture will deprive the defendant of any counsel. If the Government chooses not to restrain transfers by employing § 853(e)(1), it is likely that the defendant will not qualify as “indigent” under the Criminal Justice Act. Potential private counsel will be aware of the threat of forfeiture, and, as a result, will likely refuse to take the case. Although it is to be hoped that a solution will be developed for a defendant who “falls between the cracks” in this manner, there is no guarantee that accommodation will be made in an orderly fashion, and that trial preparation will not be substantially delayed because of the difficulties in securing counsel. For discussions of this problem, see United States v. lanniello, 644 F. Supp., at 456-457; United States v. Badalamenti, 614 F. Supp., at 197. CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 649 617 Blackmun, J., dissenting diminish the likelihood that private counsel will agree to take the case. The “message [to private counsel] is ‘Do not represent this defendant or you will lose your fee.’ That being the kind of message lawyers are likely to take seriously, the defendant will find it difficult or impossible to secure representation.’” United States v. Badalamenti, 614 F. Supp., at 196. The resulting relationship between the defendant and his court-appointed counsel will likely begin in distrust, and be exacerbated to the extent that the defendant perceives his new-found “indigency” as a form of punishment imposed by the Government in order to weaken his defense. If the defendant had been represented by private counsel earlier in the proceedings, the defendant’s sense that the Government has stripped him of his defenses will be sharpened by the concreteness of his loss. Appointed counsel may be inexperienced and undercompensated and, for that reason, may not have adequate opportunity or resources to deal with the special problems presented by what is likely to be a complex trial. The already scarce resources of a public defender’s office will be stretched to the limit. Facing a lengthy trial against a better armed adversary, the temptation to recommend a guilty plea will be great. The result, if the defendant is convicted, will be a sense, often well grounded, that justice was not done. Even if the defendant finds a private attorney who is “so foolish, ignorant, beholden or idealistic as to take the business,” ibid., the attorney-client relationship will be undermined by the forfeiture statute. Perhaps the attorney will be willing to violate ethical norms by working on a contingentfee basis in a criminal case. See Caplin & Drysdale, ante, at 633, n. 10. But if he is not—and we should question the integrity of any criminal-defense attorney who would violate the ethical norms of the profession by doing so—the attorney’s own interests will dictate that he remain ignorant of the source of the assets from which he is paid. Under § 853(c), a 650 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. third-party transferee may keep assets if “the transferee establishes . . . that he is a bona fide purchaser for value of such property who at the time of purchase was reasonably without cause to believe that the property was subject to forfeiture under this section.” The less an attorney knows, the greater the likelihood that he can claim to have been an “innocent” third party. The attorney’s interest in knowing nothing is directly adverse to his client’s interest in full disclosure. The result of the conflict may be a less vigorous investigation of the defendant’s circumstances, leading in turn to a failure to recognize or pursue avenues of inquiry necessary to the defense. Other conflicts of interest are also likely to develop. The attorney who fears for his fee will be tempted to make the Government’s waiver of fee forfeiture the sine qua non for any plea agreement, a position which conflicts with his client’s best interests. See United States v. Badalamenti, 614 F. Supp., at 196-197; United States v. Bassett, 632 F. Supp., at 1316, n. 5. Perhaps most troubling is the fact that forfeiture statutes place the Government in the position to exercise an intolerable degree of power over any private attorney who takes on the task of representing a defendant in a forfeiture case. The decision whether to seek a restraining order rests with the prosecution, as does the decision whether to waive forfeiture upon a plea of guilty or a conviction at trial. The Government will be ever tempted to use the forfeiture weapon against a defense attorney who is particularly talented or aggressive on the client’s behalf—the attorney who is better than what, in the Government’s view, the defendant deserves. The specter of the Government’s selectively excluding only the most talented defense counsel is a serious threat to the equality of forces necessary for the adversarial system to perform at its best. See United States v. Monsanto, 852 F. 2d, at 1404 (concurring opinion); United States v. Rogers, 602 F. Supp., at 1347, 1350; Cloud, 36 Emory L. J., at 829. An attorney whose fees are potentially CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 651 617 Blackmun, J., dissenting subject to forfeiture will be forced to operate in an environment in which the Government is not only the defendant’s adversary, but also his own. The long-term effects of the fee-forfeiture practice will be to decimate the private criminal-defense bar. As the use of the forfeiture mechanism expands to new categories of federal crimes and spreads to the States, only one class of defendants will be free routinely to retain private counsel: the affluent defendant accused of a crime that generates no economic gain. As the number of private clients diminishes, only the most idealistic and the least skilled of young lawyers will be attracted to the field, while the remainder seek greener pastures elsewhere. See Winick, 43 U. Miami L. Rev., at 781-782. In short, attorney’s-fee forfeiture substantially undermines every interest served by the Sixth Amendment right to chosen counsel, on the individual and institutional levels, over the short term and the long haul. C We have recognized that although there is a “presumption in favor of [the defendant’s] counsel of choice,” Wheat v. United States, 486 U. S., at 158, 160, the right to counsel of choice is not absolute. Some substantial and legitimate governmental interests may require the courts to disturb the defendant’s choice of counsel, as “[w]hen a defendant’s selection of counsel, under the particular facts and circumstances of a case, gravely imperils the prospect of a fair trial,” id., at 166 (Marshall, J., dissenting), or threatens to undermine the orderly disposition of the case, see Ungar n. Sarafite, 376 U. S. 575, 589 (1964). But never before today has the Court suggested that the Government’s naked desire to deprive a defendant of “‘the best counsel money can buy,’” Caplin & Drysdale, ante, at 630, quoting Morris n. Slappy, 461 U. S. 1, 23 (1983) (Brennan, J., opinion concurring in result), is itself a legitimate Government interest that can justify the 652 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. Government’s interference with the defendant’s right to chosen counsel—and for good reason. “[W]eakening the ability of an accused to defend himself at trial is an advantage for the government. But it is not a legitimate government interest that can be used to justify invasion of a constitutional right.” United States v. Monsanto, 852 F. 2d, at 1403 (Feinberg, C. J., concurring). And the legitimate interests the Government asserts are extremely weak, far too weak to justify the Act’s substantial erosion of the defendant’s Sixth Amendment rights. The Government claims a property interest in forfeitable assets, predicated on the relation-back provision, § 853(c), which employs a legal fiction to grant the Government title in all forfeitable property as of the date of the crime. The majority states: “Permitting a defendant to use assets for his private purposes that, under this provision, will become the property of the United States if conviction occurs, cannot be sanctioned. ” Monsanto, ante, at 613. But the Government’s insistence that it has a paramount interest in the defendant’s resources “simply begs the constitutional question rather than answering it. Indeed, the ultimate constitutional issue might well be framed precisely as whether Congress may use this wholly Active device of property law to cut off this fundamental right of the accused in a criminal case. If the right must yield here to countervailing governmental interests, the relation-back device undoubtedly could be used to implement the governmental interests, but surely it cannot serve as a substitute for them.” In re Forfeiture Hearing as to Caplin & Drysdale, Chartered, 837 F. 2d, at 652 (dissenting opinion). Furthermore, the relation-back fiction gives the Government no property interest whatsoever in the defendant’s assets before the defendant is convicted. In most instances, the assets the Government attempts to reach by using the forfeiture provisions of the Act are derivative proceeds of crime, property that was not itself acquired illegally, but was purchased with the profits of criminal activity. Prior to con- CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 653 617 Blackmun, J., dissenting viction, sole title to such assets—not merely possession, as is the case in the majority’s bank robbery example, Caplin & Drysdale, ante, at 626—rests in the defendant; no other party has any present legal claim to them.15 Yet it is in the preconviction period that the forfeiture threat (or the force of a § 853(e)(1) restraining order) deprives the defendant of use of the assets to retain counsel. The Government’s interest in the assets at the time of their restraint is no more than an interest in safeguarding Active property rights, one which hardly weighs at all against the defendant’s formidable Sixth Amendment right to retain counsel for his defense. The majority contends, of course, that assets are only restrained upon a finding of probable cause to believe that the property ultimately will be proved forfeitable, and that because “the Government may restrain persons where there is a finding of probable cause that the accused has committed a serious offense,” the Government necessarily has the right to 15 Other analogies the majority and the Government have drawn are also inapt. We do not deal with contraband, which the Government is free to seize because the law recognizes no right to possess it. See One 1958 Plymouth Sedan v. Pennsylvania, 380 U. S. 693, 699 (1965). Nor do we deal with instrumentalities of crime, which may have evidentiary value, and may also traditionally be seized by the Government and retained even if the defendant is not proved guilty, unless a party with a rightful claim to the property comes forward to refute the Government’s contention that the property was put to an unlawful use. See Calero-Toledo v. Pearson Yacht Leasing Co., 416 U. S. 663, 679 (1974); Comment, 48 U. Chi. L. Rev. 960, 963-964 (1981). As to the analogy to “jeopardy assessments” under the Internal Revenue Code, the Internal Revenue Service in that situation has a legal claim to the sums at issue at the time of the assessment, based upon substantive provisions of the Code. Here, in contrast, the Government’s claim will not arise until after conviction. In addition, even if a jeopardy assessment were to deprive a taxpayer of the funds necessary to file a challenge to the assessment in the Tax Court, the proceeding in that court is civil, and the Sixth Amendment therefore does not apply. I agree with Judge Phillips when he observes that the constitutionality of a jeopardy assessment that deprived the defendant of the funds necessary to hire counsel to ward off a criminal challenge is not to be assumed. See United States v. Harvey, 814 F. 2d, at 926. 654 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. restrain property the defendant seeks to use to retain counsel on a showing of probable cause as well. Monsanto, ante, at 615-616, citing United States v. Salerno, 481 U. S. 739 (1987). Neither the majority’s premise nor its conclusion is well founded. Although obtaining a restraining order requires a showing of probable cause, the practical effects of the threat of forfeiture are felt long before the indictment stage. Any attorney who is asked to represent the target of a drug or racketeering investigation—or even a routine tax investigation, as the facts of Caplin & Drysdale demonstrate—must think ahead to the possibility that the defendant’s assets will turn out to be forfeitable. While the defendant is not formally restrained from using his assets to pay counsel during this period, the reluctance of any attorney to represent the defendant in the face of the forfeiture threat effectively strips the defendant of the right to retain counsel. The threat of forfeiture does its damage long before the Government must come forward with a showing of probable cause. But even if the majority were correct that no defendant is ever deprived of the right to retain counsel without a showing of probable cause, the majority’s analogy to permissible pretrial restraints would fail. The Act gives the Government the right to seek a restraining order solely on the basis of the indictment, which signifies that there has been a finding of probable cause to believe that the assets are tainted. When a defendant otherwise is incarcerated before trial, in contrast, the restraint cannot be justified by the fact of the indictment alone. In addition, there must be a showing that other alternatives will not “reasonably assure the appearance of the person [for trial] and the safety of any other person and the community.” 18 U. S. C. § 3142(e) (1982 ed., Supp. V). No equivalent individualized showing that the defendant will likely dissipate his assets or fraudulently transfer them to third parties is necessary under the majority’s reading of § 853(e)(1). Furthermore, the potential danger resulting CAPLIN & DRYSDALE, CHARTERED v. UNITED STATES 655 617 Blackmun, J., dissenting from the failure to restrain assets differs in kind and severity from the danger faced by the public when a defendant who is believed to be violent remains at large before trial. Finally, even if the Government’s asserted interests were entitled to some weight, the manner in which the Government has chosen to protect them undercuts its position. Under § 853(c), a third-party transferee may keep assets if he was “reasonably without cause to believe that the property was subject to forfeiture.” Most legitimate providers of services will meet the requirements for this statutory exemption. The exception is the defendant’s attorney, who cannot do his job (or at least cannot do his job well) without asking questions that will reveal the source of the defendant’s assets. It is difficult to put great weight on the Government’s interest in increasing the amount of property available for forfeiture when the means chosen are so starkly underinclu-sive, and the burdens fall almost exclusively upon the exercise of a constitutional right.16 Interests as ephemeral as these should not be permitted to defeat the defendant’s right to the assistance of his chosen counsel. Ill In my view, the Act as interpreted by the majority is inconsistent with the intent of Congress, and seriously under- 16 Certainly criminal defendants “are not exempted from federal, state, and local taxation simply because these financial levies may deprive them of resources that could be used to hire an attorney.” Caplin & Drysdale, ante, at 631-632. The Government’s interest in raising revenue need not stand aside merely because the individual being taxed would rather spend the money by participating in a constitutionally protected activity. But I doubt that we would hesitate to reject as an undue burden on the exercise of a constitutional right a system that generally exempted personal-service transactions from taxation, but taxed payments to criminal-defense attorneys. In such circumstances, a clear-headed analysis of the Government’s action would likely reveal that burdening the exercise of the defendant’s Sixth Amendment right was not the unfortunate consequence of the Government’s action, but its very purpose. 656 491 U. S. OCTOBER TERM, 1988 Blackmun, J., dissenting mines the basic fairness of our criminal-justice system. That a majority of this Court has upheld the constitutionality of the Act as so interpreted will not deter Congress, I hope, from amending the Act to make clear that Congress did not intend this result. This Court has the power to declare the Act constitutional, but it cannot thereby make it wise. I dissent. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 657 Syllabus HARTE-HANKS COMMUNICATIONS, INC. v. CONNAUGHTON CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT No. 88-10. Argued March 20, 1989—Decided June 22, 1989 Respondent was the unsuccessful challenger for the position of Municipal Judge of Hamilton, Ohio, in an election conducted on November 8, 1983. A local newspaper, the Journal News, published by petitioner supported the reelection of the incumbent. A little over a month before the election, the incumbent’s Director of Court Services resigned and was arrested on bribery charges, and a grand jury investigation of those charges was in progress on November 1, 1983. On that day, the Journal News ran a front-page story quoting a grand jury witness (Thompson) as stating that respondent had used “dirty tricks” and offered her and her sister jobs and a trip to Florida “in appreciation” for their help in the investigation. Respondent filed a diversity action against petitioner for libel in Federal District Court, alleging that the story was false, had damaged his personal and professional reputation, and had been published with actual malice. After listening to six days of testimony and three taped interviews—one conducted by respondent and two by Journal News reporters—and reviewing the contents of 56 exhibits, the jury was given instructions defining the elements of public figure libel and directed to answer three special verdicts. It found by a preponderance of the evidence that the story in question was defamatory and false, and by clear and convincing proof that the story was published with actual malice, and awarded respondent $5,000 in compensatory damages and $195,000 in punitive damages. The Court of Appeals affirmed. It separately considered the evidence supporting each of the jury’s special verdicts, concluding that neither the finding that the story was defamatory nor the finding that it was false was clearly erroneous. In considering the actual malice issue, but without attempting to make an independent evaluation of the credibility of conflicting oral testimony concerning the subsidiary facts underlying the jury’s finding of actual malice, the court identified 11 subsidiary facts that the jury “could have” found and held that such findings would not have been clearly erroneous, and, based on its independent review, that when considered cumulatively they provided clear and convincing evidence of actual malice. 658 OCTOBER TERM, 1988 Syllabus 491 U. S. Held: 1. A showing of “‘highly unreasonable conduct constituting an extreme departure from the standards of investigation and reporting ordinarily adhered to by responsible publishers’ ” cannot alone support a verdict in favor of a public figure plaintiff in a libel action. Rather, such a plaintiff must prove by clear and convincing evidence that the defendant published the false and defamatory material with actual malice, i. e., with knowledge of falsity or with a reckless disregard for the truth. Although there is language in the Court of Appeals’ opinion suggesting that it applied the less severe professional standards rule, when read as a whole, it is clear that this language is merely supportive of the court’s ultimate conclusion that the Journal News acted with actual malice. Pp. 663-668. 2. A reviewing court in a public figure libel case must “exercise independent judgment and determine whether the record establishes actual malice with convincing clarity” to ensure that the verdict is consistent with the constitutional standard set out in New York Times Co. v. Sullivan, 376 U. S. 254, and subsequent decisions. See Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485. Based on this Court’s review of the entire record, the Court of Appeals properly held that the evidence did in fact support a finding of actual malice, but it should have taken a somewhat different approach in reaching that result. While the jury may have found each of the 11 subsidiary facts, the case should have been decided on a less speculative ground. Given the trial court’s instructions, the jury’s answers to the three special interrogatories, and an understanding of those facts not in dispute, it is evident that the jury must have rejected (1) the testimony of petitioner’s witnesses that Thompson’s sister, the most important witness to the bribery charges against the Director of Court Services, was not contacted simply because respondent failed to place her in touch with the newspaper; (2) the testimony of the editorial director of the Journal News that he did not listen to the taped interviews simply because he thought that they would provide him with no new information; and (3) the testimony of Journal News employees who asserted that they believed Thompson’s allegations were substantially true. When those findings are considered alongside the undisputed evidence, the conclusion that the newspaper acted with actual malice inexorably follows. The evidence in the record in this case, when reviewed in its entirety, is “unmistakably” sufficient to support a finding of actual malice. Pp. 685-693. 842 F. 2d 825, affirmed. Stevens, J., delivered the opinion of the Court, in which Rehnquist, C. J., and Brennan, White, Marshall, Blackmun, O’Connor, and Kennedy, JJ., joined. White, J., filed a concurring opinion, in which HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 659 657 Opinion of the Court Rehnquist, C. J., joined, post, p. 694. Blackmun, J., post, p. 694, and Kennedy, J., post, p. 696, filed concurring opinions. Scalia, J., filed an opinion concurring in the judgment, post, p. 696. Lee Levine argued the cause for petitioner. With him on the briefs were Richard L. Creighton, Jr., Kevin E. Irwin, Michael D. Sullivan, and James E. Grossberg. John A. Lloyd, Jr., argued the cause for respondent. With him on the brief were Sallie Conley Lux and Jeanette H. Rost.* Justice Stevens delivered the opinion of the Court. A public figure may not recover damages for a defamatory falsehood without clear and convincing proof that the false “statement was made with ‘actual malice’—that is, with knowledge that it was false or with reckless disregard of whether it was false or not.” New York Times Co. v. Sullivan, 376 U. S. 254, 279-280 (1964). See Curtis Publishing Co. v. Butts, 388 U. S. 130, 162 (1967) (opinion of Warren, C. J.). In Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485 (1984), we held that judges in such cases have a constitutional duty to “exercise independent judgment and determine whether the record establishes actual malice with convincing clarity.” Id., at 514. In this case the Court of Appeals affirmed a libel judgment against a newspaper without attempting to make an independent evaluation of the credibility of conflicting oral testimony concerning the subsidiary facts underlying the jury’s finding of actual malice. We granted certiorari to consider whether the Court of Appeals’ analysis was consistent with our holding in Bose. 488 U. S. 907 (1988). *A brief of amici curiae urging reversal was filed for the Associated Press et al. by P. Cameron DeVore, Daniel M. Waggoner, Douglas P. Jacobs, Alice N. Lucan, Mark L. Tuft, Harvey L. Lipton, Jeffrey N. Paule, Lois J. Schiffer, Robert D. Sack, E. Susan Garsh, William A. Niese, Deborah R. Linfield, Samuel E. Klein, W. Terry Maguire, Rene P. Milam, Richard M. Schmidt, Roslyn A. Mazer, Lawrence Gunnels, Steven R. Shapiro, Robert J. Brinkmann, J. Laurent Scharff, Jane Kirtley, and Bruce W. Sanford. 660 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. I Respondent, Daniel Connaughton, was the unsuccessful candidate for the office of Municipal Judge of Hamilton, Ohio, in an election conducted on November 8, 1983. Petitioner is the publisher of the Journal News, a local newspaper that supported the reelection of the incumbent, James Dolan. A little over a month before the election, the incumbent’s Director of Court Services resigned and was arrested on bribery charges. A grand jury investigation of those charges was in progress on November 1, 1983. On that date, the Journal News ran a front-page story quoting Alice Thompson, a grand jury witness, as stating that Connaughton had used “dirty tricks” and offered her and her sister jobs and a trip to Florida “in appreciation” for their help in the investigation. Invoking the federal court’s diversity jurisdiction, Connaughton filed an action for damages, alleging that the article was false, that it had damaged his personal and professional reputation, and that it had been published with actual malice. After discovery, petitioner filed a motion for summary judgment relying in part on an argument that even if Thompson’s statements were false, the First Amendment protects the accurate and disinterested reporting of serious charges against a public figure. The District Court denied the motion, noting that the evidence raised an issue of fact as to the newspaper’s interest in objective reporting and that the “neutral reportage doctrine” did not apply to Thompson’s statements.1 The case accordingly proceeded to trial. ’The District Court explained that the neutral reportage doctrine, as defined by the Ohio Court of Appeals, see J. V. Peters & Co. v. Knight Ridder Co., 10 Media L. Rptr. 1576 (1984), and the United States Court of Appeals for the Second Circuit, see Edwards v. National Audubon Society, Inc., 556 F. 2d 113, cert, denied, 434 U. S. 1002 (1977), “immunizes from liability the accurate and disinterested reporting of serious charges made against a public figure by a responsible, prominent organization.” App. to Pet. for Cert. 78a. Because the court was convinced that Thompson did not qualify as “a responsible, prominent organization on a par with HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 661 657 Opinion of the Court After listening to six days of testimony and three taped interviews—one conducted by Connaughton and two by Journal News reporters—and reviewing the contents of 56 exhibits, the jury was given succinct instructions accurately defining the elements of public figure libel and directed to answer three special verdicts.2 It unanimously found by a preponderance of the evidence that the November 1 story was defamatory and that it was false. It also found by clear and convincing proof that the story was published with actual malice. After a separate hearing on damages, the jury awarded Connaughton $5,000 in compensatory damages and $195,000 in punitive damages. Thereafter, the District Court denied a motion for judgment notwithstanding the verdict, App. to Pet. for Cert. 83a, and petitioner appealed. the State Attorney General’s Office in J. V. Peters or the National Audubon Society in Edwards,” it concluded that the defense was unavailable. Ibid. Petitioner did not argue in its petition for a writ of certiorari, and does not now argue, that the neutral reportage doctrine immunized its coverage of Thompson’s allegations. Accordingly, we do not review this aspect of the District Court’s judgment. 2 The jury was asked: 1. “Do you unanimously find by a preponderance of the evidence that the publication in question was defamatory toward the plaintiff?” 2. “Do you unanimously find by a preponderance of the evidence that the publication in question was false?” 3. “Do you unanimously find by clear and convincing proof that the publication in question was published with actual malice?” App. 201. There is some debate as to whether the element of falsity must be established by clear and convincing evidence or by a preponderance of the evidence. Compare Firestone v. Time, Inc., 460 F. 2d 712, 722-723 (CA5) (Bell, J., specially concurring), cert, denied, 409 U. S. 875 (1972), with Goldwater v. Ginzburg, 414 F. 2d 324, 341 (CA2 1969), cert, denied, 396 U. S. 1049 (1970). See also Tavoulareas v. Piro, 260 U. S. App. D. C. 39, 63-64, n. 33, 817 F. 2d 762, 786, n. 33 (en banc), cert, denied, 484 U. S. 870 (1987); Franklin & Bussel, The Plaintiff’s Burden in Defamation: Awareness and Falsity, 25 Wm. & Mary L. Rev. 825, 863-865 (1984). We express no view on this issue. 662 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. The Court of Appeals affirmed. 842 F. 2d 825 (CA6 1988). In a lengthy opinion, the majority detailed why its “independent examination of the entire record” had demonstrated that “the judgment does not pose a forbidden intrusion into the First Amendment rights of free expression.” Id., at 828. The opinion identified the “core issue” as “simply one of credibility to be attached to the witnesses appearing on behalf of the respective parties and the reasonableness and probability assigned to their testimony.” Id., at 839-840. It separately considered the evidence supporting each of the jury’s special verdicts, concluding that neither the finding that the article was defamatory3 nor the finding that it was false4 was clearly erroneous. The Court of Appeals’ review of the actual malice determination involved four steps. It first noted the wide disparity between the respective parties’ versions of the critical evidence, pointing out that if the jury had credited petitioner’s evidence it “could have easily concluded that Thompson’s 3 The Court of Appeals observed that “the article was defamatory in its implication that Connaughton was an unethical lawyer and an undesirable candidate for the Hamilton Municipal judgeship who was capable of extortion, who was a liar and an opportunist not fit to hold public office, particularly a judgeship.” 842 F. 2d, at 840-841. 4 As to the finding of falsity, the Court of Appeals wrote: “Equally apparent from the jury’s answer to the second special interrogatory is that it considered the published Thompson charges to be false. Its finding is understandable in light of the plaintiff’s proof which disclosed that the Journals effort to verify her credibility ended in an avalanche of denials by knowledgeable individuals; [and] its inability to produce a single person who supported Thompson’s accusations .... “Moreover, the jury obviously refused to credit the Journals construction of Connaughton’s interview of October 31. It accepted Connaughton’s express denials of each Thompson charge and considered the significant language interpreted by the Journal to constitute his admissions of those charges, when read in context, as nothing more than conjecture elicited by structured questions calculated to evoke speculation. Thus, upon reviewing the record in its entirety, this court concludes that the jury’s determinations of the operational facts bearing upon the falsity of the article in issue were not clearly erroneous.” Id., at 841. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 663 657 Opinion of the Court charges were true and/or that the Journal’s conduct in determining Thompson’s credibility was not a highly unreasonable departure from the standards of investigation and reporting ordinarily adhered to by reasonable publishers.” Id., at 840. Second, it inferred from the jury’s answers to the three special interrogatories that “it obviously elected to assign greater credibility to the plaintiff’s witnesses and proof [and that] the jury simply did not believe the defendants’ witnesses, its evidentiary presentations or its arguments.” Ibid. Third, having considered what it regarded as the “subsidiary or operative facts” that constituted the plaintiff’s theory of the case, it concluded that the jury’s findings concerning those operative facts were not clearly erroneous. Id., at 843-844. Fourth, “in the exercise of its independent judgment” based on its evaluation of the “cumulative impact of the subsidiary facts,” the court concluded that “Connaughton proved, by clear and convincing evidence, that the Journal demonstrated its actual malice when it published the November 1, 1983, article despite the existence of serious doubt which attached to Thompson’s veracity and the accuracy of her reports.” Id., at 846. Judge Guy dissented. In his opinion the admissions made by Connaughton in his interview with Journal News reporters the day before the story was published sufficiently corroborated Thompson’s charges to preclude a finding of actual malice. Id., at 853-854. He was satisfied, as a matter of law, that respondent had failed to prove actual malice by clear and convincing evidence, regardless of whether determinations of credibility made by the jury are subject to a de novo standard of review. Id., at 855. II Petitioner contends that the Court of Appeals made two basic errors. First, while correctly stating the actual malice standard announced in New York Times Co. v. Sullivan, 376 U. S. 254 (1964), the court actually applied a less severe 664 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. standard that merely required a showing of “ ‘highly unreasonable conduct constituting an extreme departure from the standards of investigation and reporting ordinarily adhered to by responsible publishers.’” 842 F. 2d, at 845 (quoting Curtis Publishing Co. v. Butts, 388 U. S., at 155 (opinion of Harlan, J.)). Second, the court failed to make an independent de novo review of the entire record and therefore incorrectly relied on subsidiary facts implicitly established by the jury’s verdict instead of drawing its own inferences from the evidence. There is language in the Court of Appeals’ opinion that supports petitioner’s first contention. For example, the Court of Appeals did expressly state that the Journal News’ decision to publish Alice Thompson’s allegations constituted an extreme departure from professional standards.5 Moreover, the opinion attributes considerable weight to the evidence that the Journal News was motivated by its interest in the reelection of the candidate it supported and its economic interest in gaining a competitive advantage over the Cincin- 5 The Court of Appeals wrote: “In Curtis Publishing Co. v. Butts, the Supreme Court accorded public figures as well as public officials recovery of damages for the publication of ‘defamatory falsehood whose substance makes substantial danger to reputation apparent, on a showing of highly unreasonable conduct constituting an extreme departure from the standards of investigation and reporting ordinarily adhered to by responsible publishers.’ 388 U. S. at 155.” Id., at 845. At another point, the court wrote: “Accordingly, this court concludes that the Journal’s decision to rely on Thompson’s highly questionable and condemning allegations without first verifying those accusations through her sister, [Stephens], and without independent supporting evidence constituted an extreme departure from the standards of investigation and reporting ordinarily adhered to by responsible publishers which demonstrated a reckless disregard as to the truth or falsity of Thompson’s allegations and thus provided clear and convincing proof of ‘actual malice’ as found by the jury. Butts, 388 U. S. at 153.” Id., at 847 (emphasis supplied). See also id., at 840. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 665 657 Opinion of the Court nati Enquirer, its bitter rival in the local market.6 Petitioner is plainly correct in recognizing that a public figure plaintiff must prove more than an extreme departure from professional standards and that a newspaper’s motive in publishing a story—whether to promote an opponent’s candidacy or to increase its circulation—cannot provide a sufficient basis for finding actual malice. The language in the Court of Appeals’ opinion discussing professional standards is taken from Justice Harlan’s plurality opinion in Curtis Publishing Co. v. Butts, supra, at 155. In that case, Justice Harlan had opined that the New York Times actual malice standard should be reserved for cases brought by public officials. The New York Times decision, in his view, was primarily driven by the repugnance of seditious libel and a concern that public official libel “lay close” to 6 As to the newspaper’s motives, the Court of Appeals asserted: “A review of the entire record of the instant case discloses substantial probative evidence from which a jury could have concluded (1) that the Journal was singularly biased in favor of [the incumbent] and prejudiced against Connaughton as evidenced by the confidential personal relationship that existed between [the incumbent] and Blount, the Journal Editorial Director, and the unqualified, consistently favorable editorial and daily news coverage received by [the incumbent] from the Journal as compared with the equally consistent unfavorable news coverage afforded Connaughton; (2) that the Journal was engaged in a bitter rivalry with the Cincinnati Enquirer for domination of the greater Hamilton circulation market as evidenced by Blount’s vituperous public statements and criticism of the Enquirer; (3) that the Enquirer’s initial expose of the questionable operation of the [incumbent’s] court was a high profile news attraction of great public interest and notoriety that had ‘scooped’ the Journal and by Blount’s own admission was the most significant story impacting the . . . cam-paign[;] (4) that by discrediting Connaughton the Journal was effectively impugning the Enquirer thereby undermining its market share of the Hamilton area.” Id., at 843. Later in the opinion, the court again stressed that “the evidence adduced at trial demonstrated that the Journal was motivated to publicize Thompson’s allegations, not only by a desire to establish its preeminence in the reporting of Hamilton political news, but also by a desire to aid the [incumbent’s] campaign.” Id., at 846. 666 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. this universally renounced, and long-defunct, doctrine. 388 U. S., at 153. In place of the actual malice standard, Justice Harlan suggested that a public figure need only make “a showing of highly unreasonable conduct constituting an extreme departure from the standards of investigation and reporting ordinarily adhered to by responsible publishers.” Id., at 155. This proposed standard, however, was emphatically rejected by a majority of the Court in favor of the stricter New York Times actual malice rule. See 388 U. S., at 162 (opinion of Warren, C. J.); id., at 170 (Black, J., dissenting); id., at 172 (Brennan, J., dissenting). Moreover, just four years later, Justice Harlan acquiesced in application of the actual malice standard in public figure cases, see Rosenbloom v. Metromedia, Inc., 403 U. S. 29, 69-70 (1971) (dissenting opinion), and by the time of the Court’s decision in Gertz n. Robert Welch, Inc., 418 U. S. 323 (1974), the Court was apparently unanimously of this view. Today, there is no question that public figure libel cases are controlled by the New York Times standard and not by the professional standards rule, which never commanded a majority of this Court. It also is worth emphasizing that the actual malice standard is not satisfied merely through a showing of ill will or “malice” in the ordinary sense of the term.7 See Beck- 7 The trial judge correctly instructed the jury that “[a]ctual malice may not be inferred alone from evidence of personal spite, ill will or intention to injure on the part of the writer.” App. 199. The phrase “actual malice” is unfortunately confusing in that it has nothing to do with bad motive or ill will. See Rosenbloom v. Metromedia, Inc., 403 U. S. 29, 52, n. 18 (1971) (opinion of Brennan, J.). By instructing the jury “in plain English” at appropriate times during the course of the trial concerning the not-so-plain meaning of this phrase, the trial judge can help ensure that the New York Times standard is properly applied. Tavoulareas, 260 U. S. App. D. C., at 84, 817 F. 2d, at 807 (R. B. Ginsburg, J., concurring). See also Westmoreland v. CBS Inc., 596 F. Supp. 1170, 1172-1173, n. 1 (SDNY 1984) (suggesting that jury confusion can be minimized if a less confusing phrase, such as “state-of-mind,” “delib HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 667 657 Opinion of the Court ley Newspapers Corp. v. Hanks, 389 U. S. 81 (1967) (per curiam); Henry v. Collins, 380 U. S. 356 (1965) (per curiam). Indeed, just last Term we unanimously held that a public figure “may not recover for the tort of intentional infliction of emotional distress . . . without showing . . . that the publication contains a false statement of fact which was made . . . with knowledge that the statement was false or with reckless disregard as to whether or not it was true.” Hustler Magazine, Inc. n. Falwell, 485 U. S. 46, 56 (1988). Nor can the fact that the defendant published the defamatory material in order to increase its profits suffice to prove actual malice. The allegedly defamatory statements at issue in the New York Times case were themselves published as part of a paid advertisement. 376 U. S., at 265-266. If a profit motive could somehow strip communications of the otherwise available constitutional protection, our cases from New York Times to Hustler Magazine would be little more than empty vessels. Actual malice, instead, requires at a minimum that the statements were made with a reckless disregard for the truth. And although the concept of “reckless disregard” “cannot be fully encompassed in one infallible definition,” St. Amant v. Thompson, 390 U. S. 727, 730 (1968), we have made clear that the defendant must have made the false publication with a “high degree of awareness of. . . probable falsity,” Garrison v. Louisiana, 379 U. S. 64, 74 (1964), or must have “entertained serious doubts as to the truth of his publication,” St. Amant, supra, at 731. Certain statements in the Court of Appeals’ opinion, when read in isolation, appear to indicate that the court at times substituted the professional standards rule for the actual malice requirement and at other times inferred actual malice from the newspaper’s motive in publishing Thompson’s story. Nevertheless, when the opinion is read as a whole, it is clear that the conclusion concerning the newspaper’s departure erate or reckless falsity,” or “constitutional limitation” is used in the jury’s presence). 668 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. from accepted standards and the evidence of motive were merely supportive of the court’s ultimate conclusion that the record “demonstrated a reckless disregard as to the truth or falsity of Thompson’s allegations and thus provided clear and convincing proof of ‘actual malice’ as found by the jury.” 842 F. 2d, at 847. Although courts must be careful not to place too much reliance on such factors, a plaintiff is entitled to prove the defendant’s state of mind through circumstantial evidence, see Herbert v. Lando, 441 U. S. 153, 160 (1979); Tavoulareas n. Piro, 260 U. S. App. D. C. 39, 66, 817 F. 2d 762, 789 (en banc), cert, denied, 484 U. S. 870 (1987), and it cannot be said that evidence concerning motive or care never bears any relation to the actual malice inquiry. Thus, we are satisfied that the Court of Appeals judged the case by the correct substantive standard. The question whether the Court of Appeals gave undue weight to the jury’s findings—whether it failed to conduct the kind of independent review mandated by our opinion in Bose—requires more careful consideration. A proper answer to that question must be prefaced by additional comment on some of the important conflicts in the evidence. Ill The most important witness to the bribery charges against the Director of Court Services was Patsy Stephens, Alice Thompson’s older sister. In a tape-recorded interview conducted in Connaughton’s home between 12:30 and 4:30 a.m. on September 17, 1983, Stephens explained how, on 40 or 50 occasions, she had visited with the Court Administrator, Billy Joe New, in his office and made cash payments to dispose of “DUI” and other minor criminal charges against her former husband and various other relatives and acquaintances.8 On September 22, pursuant to an arrangement 8 Early in September Connaughton’s wife Martha was advised that Patsy Stephens was willing to disclose important information about the special treatment her former husband had received in the Hamilton Munic- HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 669 657 Opinion of the Court made by Connaughton at the suggestion of the county prosecutor, Stephens took a lie detector test. After learning that she had passed the test, Connaughton filed a written complaint against New. In due course, New was arrested, indicted, and convicted. Alice Thompson was one of the eight persons present at the tape-recorded interview on September 17.9 One of the cases Patsy Stephens described was a shoplifting charge against her sister. Thompson volunteered some comments about the incident, but otherwise had little to say during the long interview with Stephens. Thompson was also present on the 22d, when Stephens took the polygraph test, but Thompson declined to submit to such a test. App. 301. On that day, the two sisters spent several hours in the company of Connaughton, his wife, and two of his supporters. They discussed a number of subjects, including the fact that Billy Joe New had just resigned, the question whether there was reason to be concerned about the safety of the two sisters, the fact that Martha Connaughton might open an ice cream parlor sometime in the future, the possibility that the two sisters might be employed there as waitresses, and a vacation in Florida planned by the Connaughtons for after the election. ipal Court. The source of this advice was the president of the local chapter of Mothers Against Drunk Driving. Martha Connaughton and her brother then visited with Patsy Stephens in her mother’s home for about 30 minutes and arranged for a later interview with Connaughton. Alice Thompson was present at a part of that meeting, as well as at the subsequent interview. Shortly before midnight on September 16, after Patsy Stephens and Alice Thompson had returned home from work, two of Connaughton’s supporters (Berry and Cox) picked the two sisters up and drove them to Connaughton’s home where they remained until about 4:30 a.m. on September 17. 9 The other seven were: Patsy Stephens, Dan and Martha Connaughton, Martha Connaughton’s brother Dave Berry, Connaughton’s campaign manager, Joe Cox, and two of Connaughton’s neighbors, Jeanette and Ernest Barnes. 670 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Late in October, New’s lawyer, Henry Masana, met with Jim Blount, the editorial director of the Journal News, and Joe Cocozzo, the newspaper’s publisher, to arrange a meeting with Alice Thompson. Masana explained that Thompson wanted to be interviewed about the “dirty tricks” Connaughton was using in his campaign. Thereafter, on October 27, Blount and Pam Long, a Journal News reporter, met with Thompson in the lawyer’s office and tape-recorded the first of the two interviews that provided the basis for the story that Long wrote and the Journal News published on November 1. The tape of Alice Thompson’s interview is 1 hour and 20 minutes long. Significant portions of it are inaudible or incoherent. It is clear, however, that Thompson made these specific charges: —that Connaughton had stated that his purpose in taping the interview with Patsy Stephens was to get evidence with which he could confront New and Judge Dolan and “scare them into resigning” without making any public use of the tapes;10 10 “A. They started asking me a bunch of questions so I asked Dan Connaughton .... I said why are you doing this .... And of course, he turned off the tape recorder. And he said, I’ll tell you the truth. He said, all I want is to get enough evidence on Billy, he said, and have Billy resign. And he said, of course, if Billy resigns, Dolan will resign, and he said, then I can just step up on the bench. . . . But he said right out of his own mouth, all I want to do is to get a story in evidence on them, to meet them face to face, and show them what evidence he had against him, or whatever, to get them to resign, and no more would be said about it. “Q. Okay. So in other words, based on what he said to you, you believed him? “A. Blackmail. I mean, you know, the way he phrased it, the way he said it, you know. He said all he wanted to do was get enough evidence on Billy, and he also used Dolan’s name, which I don’t know what he was going to get on Dolan—to scare them into resigning. I said what happens when they resign? Nothing more will be said about anything. He said when I take the bench nothing will be said.” App. 291-292. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 671 657 Opinion of the Court —that he would pay the expenses for a 3-week vacation in Florida for the two sisters;11 —that he would buy a restaurant for the two sisters’ parents to operate;12 11 “A. ... I asked them what I was going to get out of it. “Q. What did they promise you? Or what did they say when you asked them? “A. They said my help would be deeply appreciated. And they went on to talk about the three weeks vacation they was planning on taking when the election was . . . “Q. He was planning to take three weeks vacation? “A. Yes, the family—Dave Berry and Martha, and Dan. “MR. BLOUNT: They wanted you to go along? “A. Me and my sister would be welcome to go along with Dave . . . “(By Mrs. Long) “Q. Did they say they would pay your expenses? “A. Yeah. I made it clear to them that I couldn’t afford a trip to Florida. “MR. BLOUNT: Was the tape recorder on at that time? “A. Oh, no. “(By Mrs. Long) “Q. Now where were they going to go? “A. Three weeks in Florida. “Q. And they added Disneyworld? “A. (Inaudible) a three weeks trip to Florida. And they had a friend in Florida that wouldn’t be home at the time, that we could stay at their condominium.” Id., at 293-294. 12 “A. . . . [Connaughton] said he was thinking about putting a restaurant in there, and he was wanting to know if my mother and father [Zella and Brownie Breedlove] would run it for him. And I said oh yeah, my mother would love to get back into the restaurant business. He said good, when the lease is up, he said, we’ll tear the inside out and put a restaurant in there, and he said, your mother and father can run it, and he said that way, he said you girls can help run it too, and put your sisters in there working too; he said just... he even made up a name—Breedlove’s Lunch or something like that. Ma Breedlove’s Cooking, you know. He had the names figured out and everything. He offered to buy us a restaurant, you know, and put us in that building. “Q. Okay. So it would just be your parents being a manager, they wouldn’t have to buy—did you understand him that they wouldn’t have to . . . 672 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. —that he would provide jobs for both Patsy Stephens and Alice Thompson;13 —that he would take them out to a victory dinner at an expensive French restaurant after the election;14 and —that Connaughton would not allow knowledge of the sisters’ involvement to become public.15 “A. Oh, they was going to do everything, you know. They was just going to put us in there to work, or to run it. They wanted my mother to run the business for them.” Id., at 307. 18 “Q. Did he promise you to find a job? “A. Yeah. “Q. Why did he offer to find you a job? “A. Because the day at the house, going back to the first time I met them, Martha was asking me did I work, or anything, and I was telling her I was looking for work. I had been out of a job. Evidently she must have talked to her husband about it, and that night over at his home, he said are you employed now, you know, . . . and I said no. So he said, we’ll see if we can’t do something about that. I told him I wanted away from bar-tending and stuff; he said we’ll see if we can’t do something about it. You know, a decent job.” Id., at 295-296. “MR. MASANA: I’m going to interject. What about the job you were promised? “A. Oh, when they promised me, you know, the secure job and everything, they also promised—they promised Patsy a job too. (By Mrs. Long) “Q. That she would be in with Breedlove’s Lunch, or cafe? “A. No, they promised Patsy a decent job, you know. “Q. That she would be (inaudible). “A. That she would be good up in Court. That come out of his own mouth. That come out of Dan’s mouth; he said we need somebody like you up at the courthouse. Municipal Court.” Id., at 309. 14 “A. . . . And he said he wanted me and Pat to definitely be there, and for a victory dinner he wanted to take me and Patsy to dinner at the Maisonette. “Q. This would be after he wins the election? “A. Ummm-hmmm.” Id., at 306. 16 “A. . . . But as far as anybody else, the public, or anything like that— or it going to Court, we wouldn’t have to worry about it; we wouldn’t have to go to Court and our names wouldn’t be on there.” Id., at 296. “A. [T]hey had already promised that our names wouldn’t be mentioned that nobody would know about us . . . .” Id., at 302. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 673 657 Opinion of the Court During the course of the interview, Thompson indicated that she had told her story to the Cincinnati Enquirer, which declined to print it, id., at 284, and that the local police, likewise, were not interested, id., at 310.16 Thompson indicated that she was “against” Connaughton becoming a judge. Id., at 311. She also asserted that since Connaughton had made public that she and her sister had provided evidence against New, friends had accused her “of being a snitch and a rat”—epithets to which she took great offense—and that one reason she came to the Journal News was “to get that cleared up.”17 In her description of the interview in Connaughton’s home on September 17, Thompson stated that Connaughton had frequently turned off the tape recorder,18 that his voice would not be heard 16 The transcript of the interview quotes Thompson as saying: “I explained to them the whole story, how it got off to this, or that, you know. They was embarrassed evidently.” Id., at 310. However, the tape recording of the interview, which the jury heard, makes clear that Thompson actually stated: “I explained to them the whole story, how I got offered this and that, you know. They wasn’t interested in this evidently.” Defendant’s Exh. J. 17 “A. . . . Can’t get any worse than what Dan (inaudible). Makes it sound like I’m the bad guy. “Q. Have you had any repercussions from this? “A. I’ve been under a lot of (inaudible) strain. I guess. “Q. Other people calling you besides the Enquirer? “A. Yeah. I’ve had people that I thought were my friends call me and accuse me of being a snitch and a rat. I don’t like to carry that name, and that’s what a lot of people is thinking. That knows me. “MR. BLOUNT: They were just mad, they didn’t threaten you? “A. (inaudible) a snitch. You name it, and I’m that. I just want to get that cleared up.” App. 320. 18 “A. ... I said, what’s the whole deal? And of course, he turned off the tape recorder. . . . “MR. BLOUNT: Was being questioned by the Connaughtons tougher than going to Court? “A. Ummm-hmmm. They turned that tape recorder on and off so many times, you know, left out what they wanted to. 674 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. on the tape,19 and, somewhat inconsistently (and in response to a leading question), that most other comments had been made in response to leading questions by Connaughton.20 Toward the end of the interview, Blount made two significant comments. He announced that “Pam will, of course, write the story,” id., at 314, and he asked “[w]hat would happen if we called your sister,” id., at 316. In response to the first comment, Thompson volunteered a somewhat improbable explanation for her motivation in seeking the interview,21 “MR. BLOUNT: Was the tape recorder on at that time? “A. Oh, no.” Id., at 291-293. 19 “MR. BLOUNT: They had it on when you were talking and off when they were talking? “A. I don’t think Dan Connaughton’s voice is on it.” Id., at 293. “MR. BLOUNT: Was it Dan Connaughton himself who talked about the trip? “A. Yeah. He did most of the talking in the living room. Like I said though the tape recorde[r] was off when Dan spoke.” Id., at 295. 20 “MR. MASANA: Off the record—you were saying something about Dan was encouraging you to say things in a certain way? “A. Oh, yeah. He was leading me in questions, you know. (By Mrs. Long) “Q. Can you give us an example? “A. Well, he kept on trying to get me to say that Dolan had something to do with this, you know? “Q. Would he phrase it in a question? Like, did Judge Dolan have anything to do with it? “MR. BLOUNT: Wasn’t it true that Judge Dolan did this, or something? “A. Yeah, you know, and so on. But like I say, if you listen to the tapes you’re not going to hear it, because his voice ain’t on the tape. . . . “Q. Sure. So it was a yes, no, situation for you in that he’d phrase it a certain way and all you had to do was yes or no? “A. Ummm-hmmm. And then, you know, he’d say to repeat that. Id., at 296-298. 21 “A. I just want people to know. Because they shouldn’t vote for a man that is this dirty, you know, because I call it blackmail, what he was trying to do.” Id., at 314. There is some tension between this civic interest in fair procedure and Thompson’s reluctant participation in the exposure of the corrupt pro HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 675 657 Opinion of the Court and in response to the second she gave an equivocal answer,22 even though she had previously assured Blount that Stephens would confirm everything she had said.23 On Sunday, October 30, an editorial appeared in the Journal News under the headline “Municipal Court Race will have More than One Loser.”24 App. to Pet. for Cert. 45a. In the column, Blount observed that the campaign “battle has been all it was expected to be and more,” and predicted that “[a] lot could still happen in the next eight to nine days.” Ibid. He went on to discuss the charges pending against New, stating that the “array of charges and counter charges probably has taken some votes from Dolan.” Ibid. He cautioned, however, that the race was still wide open and quoted an unidentified voter as saying, “I resent voting for a person who I later find has been deceitful or dishonest cedures at the Municipal Court, her assertion that although she realized that Connaughton’s offers were improper, she would have accepted them if her name had never been mentioned because “that’s the way [the system] works,” id., at 315, and her displeasure at being called a “snitch and a rat,” id., at 320. 22 “A. I think she’s scared right now to talk to anyone, because the Cincinnati Enquirer has been trying to get her to talk to them. She’s getting scared now since this is all reality. My sister is . . . she’s kind of weak-minded when it comes to anything like that. She won’t do nothing for nobody unless she thinks she’s benefiting from it. And she honestly thought she was a getting a job out of this, and would make something of herself out of this. And the Connaughtons just used her all the way. And now since she’s seeing that it’s coming down to where she ain’t going to get nothing out of it, she’s brought up in the middle of all this and everything, she’s scared.” Id., at 316. 23 “MR. BLOUNT: Obviously, we can’t quote your sister from you (inaudible). What’s your sister’s position in this, would she support you or would she support him? In other words, if somebody said to her, who’s telling the truth here? “A. She’ll tell you about the trips, the dinner at the Maisonette, the jobs and everything. She’ll tell you that’s the truth, because they was offered to her too.” Id., at 313. 24 The full text of this editorial is reprinted as Appendix A to the Court of Appeals’ opinion. 842 F. 2d, at 848-849. 676 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. in campaigning.” Id., at 46a. Significantly, this unidentified person did not express indignation at dishonesty in the administration of the Municipal Court—a concern one would think the arrest of New might have prompted—but rather, a distaste for dishonesty in campaigning—a concern that the then-uninvestigated and unwritten November 1 story would soon engender. After questioning the Cincinnati Enquirer’s coverage of a story critical of Dolan and suggesting that “the Connaughton forces have a wealthy, influential link to Enquirer decisionmakers,” the column indicated that the Journal News had not yet decided which candidate it favored, but implied that an endorsement was forthcoming. Id., at 48a. On October 31, a reporter for the Journal News telephoned Connaughton and asked him to attend a meeting with Jim Blount, stating “that the endorsement may hang in the balance.” Tr. 457 (Aug. 9, 1985). Connaughton met with the reporter, Blount, and Cocozzo that afternoon and discussed a variety of subjects. One of the subjects was the rumor that Connaughton had an influential link to the Cincinnati Enquirer. Connaughton asserted that he had “no extraordinary pull or any inside track to anybody down there,” and that any rumor to the contrary was “a lie.” Id., at 458. Another subject was Connaughton’s participation in the investigation of Billy Joe New. Connaughton provided a chronology of the events that led to his filing of the complaint against New and explained that he believed that he had an obligation “as an attorney and officer of the court to report [New’s] crimes.” Id., at 458-459. No mention was made of Thompson’s interview or her charges against Connaughton. Id., at 460. After about an hour, Jim Blount received a telephone call and then told Connaughton that a reporter wanted to interview him. Id., at 462. Connaughton then went to another office where Blount and Long advised him that they had interviewed Alice Thompson HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 677 657 Opinion of the Court and were “trying to find out. . . how much of her statement was true.” App. 256. The ensuing tape-recorded interview lasted 55 minutes. Connaughton acknowledged that the meetings that Thompson described had taken place and that there had been some speculative discussion about each of the subjects that Thompson mentioned. He stated, however, that Thompson’s account of their meetings was “obviously shaded and bizarre,” id., at 276, and that there was “absolutely” no “quid pro quo for information.”25 Thus, while categorically denying that he intended to confront New and Judge Dolan with the tape of the Stephens interview to scare them into resigning, Connaughton admitted that he might well have speculated about what they would say or do if they heard the tapes.26 Similarly, while denying 25 The transcript of the Connaughton interview states: “MR. CONNAUGHTON: No, and it had nothing to do with (inaudible) for information or something, i[f ] that’s what the point of this question is. That’s absolutely no, if that’s that question. Well, the tape will speak for itself.” App. 265. The tape recording of this interview makes clear that Connaughton said, “No, and it had nothing to do with a quid pro quo for information . . . .” Defendant’s Exh. I. 26 “A. ... I think it would be fair to say, sometime during those three or four hours that they were there, that I probably made a remark along the lines that I just can’t believe what I’m hearing, and, you know, I would think if they could hear what we’re hearing, they would probably resign. I mean, I thought the allegation was that serious. But to tell her that —to answer that—and if she’s saying that was my announced purpose of what I had them there for and what we were going to do with the information, my answer would be no. “MR. BLOUNT: You didn’t tell her you were going to take the tapes to him? And play them for them? “A. No. No. What I might have said is, boy, I’d sure like to let them hear these tapes and see what they’ve got to say for themselves, you know, in a fashion such as that. “MR. BLOUNT: In an expression of shock. “MR. CONNAUGHTON: Yeah. Yeah, as I almost fell off of the fireplace. Right.” App. 262-263. 678 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. that he had promised Stephens and Thompson anonymity, he agreed that he had told them that he had hoped that they could remain anonymous.27 He also categorically denied that he had promised Thompson a job as a waitress, promised Stephens a job at the Municipal Court, or promised to set their parents up in a restaurant, although he did acknowledge a general conversation in which his wife had discussed the possibility that if her dream of opening “a gourmet ice cream shop” should materialize, the sisters might work there.28 There were similar acknowledgments of references 27 “Q. Did you ever promise Alice Thompson anonymity? “A. That question was discussed, and I was hoping to her, and I told her it would be my intention and hope that she could remain anonymous, yes. But did I promise her anonymity, the answer would be no. Did we discuss it, we sure did, and I expressed to her my desire as well as her desire that she could remain anonymous.” Id., at 264. 28 “Q. Did you ever talk to Alice about getting a job for her in appreciation for her help with your investigation of New and Dolan? “A. No. “Q. Not a waitress job? “A. No. “Q. Did you promise a Municipal Court job for her sister Patsy Stephens? “A. No. “Q. Did you offer to have ‘the sisters go on a post election trip to Florida with you and your family to stay in a condominium?’ “A. No. “Q. Did you offer to set up Thompson’s parents, the Breedloves, in what is now Walt’s Chambers, which you own and lease? “A. Absolutely not. “Q. Why would she say this to us? “A. What was discussed in an off-handed way, the people who own that bar, who we’re not very pleased with, their lease expires next September. My wife has the idea that she wants to open an ice cream type shop like Graeters, or some such thing as that, and I heard her discussing with them that maybe, since Patty had run this Homette Restaurant or something of that nature, that maybe she would help out and participate in the operation of this—whatever you want to call it— HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 679 657 Opinion of the Court to a possible Florida trip and postelection victory dinner, but denials of any promises.29 At the end of the interview, Long went back—stressing that Thompson’s charge was a “hefty” deli shop or gourmet ice cream shop. Yes, and I was present when that took place. “Q. And when was that? “A. Well, I don’t think it was that night. As I recall, this was a later time that we had seen them. “Q. But that would only be for Patty (unclear)? “A. I guess Alice was there, and the offer may have been extended to her in that fashion, that she could work there or something—I wouldn’t be surprised if that was said.” Id., at 264-265. 29 “Q. What about this post election trip to Florida? . . . “MR. BLOUNT: Did you talk about anything like that? “A. Ummm-hmmm. After getting over the initial shock it became a little clearer to me of—kind of how scary this thing was with the information they gave to us, as far as, if their personal safety was at stake .... I do remember in an off-handed way it being discussed . . . they could go down to Hilton Head or Florida, or something like that, or maybe hide out or something like that, I don’t know. But I own no property and have nothing to offer them. “Q. But there was talk about a friend that had a condominium that would be vacant and it was in terms of a full blown trip, you know, you, the Berrys, the whole group going down to Florida and they were welcome to go along. . . . “A. No. The only conversation I remember along those lines was in connection with, if their personal safety might be in question because of going out on the line and making these serious allegations. ...” Id., at 266. “Q. One last statement. At lunch Thompson said that you promised to take her and her sister out to a post election victory dinner at the Maisonette? “A. I promised to take them to the Maisonette? Hell, I haven’t been to the Maisonette for years. “MR. BLOUNT: Was it discussed? . . . “A. It may have been. It may have been. I won’t deny that some loose discussion in a kidding way was . . . “A. ... If she says that I made a firm statement that we were going to definitely plan a party at the Maisonette, that’s not true. . . .” Id., at 272-273. 680 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. one—and asked for a second time whether Connaughton had promised Stephens a job at the Municipal Court if he was elected. He once again unequivocally denied the allegation.30 The following day the lead story in the Journal News — under the headline “Bribery case witness claims jobs, trip offered”—reported that “[a] woman called to testify before the . . . Grand Jury in the Billy Joe New bribery case claims Dan Connaughton, candidate for Hamilton Municipal Judge, offered her and her sister jobs and a trip to Florida in appreciation’ for their help.”31 Id., at 329. The article, which carried Pam Long’s byline, stated that Thompson accused Connaughton of using “ ‘dirty tricks’ ” to gain her cooperation in investigating New and that Connaughton, although admitting that he did meet with Thompson, “denied any wrongdoing.” Ibid. Each of Thompson’s allegations was accurately reported, including her claims that Connaughton had promised to “protect her anonymity,” id., at 330, that he had promised Stephens “a municipal court job” and Thompson some other sort of work, that he had invited both sisters on “a post-election trip to Florida,” and that he had offered “to set up Thompson’s parents ... in the restaurant business,” id., at 333. The article conveyed Thompson’s allegation that “the tapes were turned off and on during a session [that] lasted until 5:30 a.m.,” and that these promises were 30 “Q. So her sister Patty, again getting back and going over the promises —pardon me for going back to them but that seems to be a hefty charge against you. “A. That’s alright. “Q. Her sister Patty is not going to get a job in the Municipal Court if you’re elected? “A. Not that I know of. “Q. And she’s not going to be disappointed to find that out, right? “[A. She’s not going to be disappointed at that. Right.]” Id., at 277. The bracketed response does not appear in the written transcript, but can be heard on the tape recording. Defendant’s Exh. I. 31 The full text of this article is reprinted as Appendix B to Judge Guy’s dissenting opinion. 842 F. 2d, at 858-859. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 681 657 Opinion of the Court made “[w]hen the tape was turned off.” Ibid. In addition, Long wrote, “Thompson claimed Connaughton had told her the tapes he made of her . . . statement. . . were to be presented to Dolan” with the hope that Dolan might resign, thereby allowing Connaughton to assume the municipal judgeship. Id., at 335. Connaughton’s contrary version of the events was also accurately reported. As the Court of Appeals correctly noted, there was evidence in the record—both in the Thompson tape and in the Connaughton tape—that would have supported the conclusion that Thompson was telling the truth and that Connaughton was dissembling. See 842 F. 2d, at 840. On the other hand, notwithstanding the partial confirmation of Thompson’s charges in the Connaughton tape, there remained a sharp conflict between their respective versions of the critical events. There was unquestionably ample evidence in the record to support a finding that Thompson’s principal charges were false, either because she misinterpreted remarks by Connaughton and his wife, or because Thompson was deliberately lying. The jury listened to the tape recordings of the two conflicting interviews and also observed the demeanor of the two witnesses as they testified in open court. They found that Connaughton was telling the truth and that Thompson’s charges were false. The fact that an impartial jury unanimously reached that conclusion does not, however, demonstrate that the Journal News acted with actual malice. Unlike a newspaper, a jury is often required to decide which of two plausible stories is correct. Difference of opinion as to the truth of a matter—even a difference of 11 to 1—does not alone constitute clear and convincing evidence that the defendant acted with a knowledge of falsity or with a “high degree of awareness of . . . probable falsity,” Garrison, 379 U. S., at 74. The jury’s verdict in this case, however, derived additional support from several critical pieces of information that strongly support the inference that the Jour 682 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. nal News acted with actual malice in printing Thompson’s false and defamatory statements. IV On October 27, after the interview with Alice Thompson, the managing editor of the Journal News assembled a group of reporters and instructed them to interview all of the witnesses to the conversation between Connaughton and Thompson with one exception—Patsy Stephens. No one was asked to interview her and no one made any attempt to do so. See App. 56-57, 61, 83-85. This omission is hard to explain in light of Blount’s and Long’s repeated questions during the Connaughton and Thompson interviews concerning whether Stephens would confirm Thompson’s allegations. See id., at 277, 313, 316. It is utterly bewildering in light of the fact that the Journal News committed substantial resources to investigating Thompson’s claims, yet chose not to interview the one witness who was most likely to confirm Thompson’s account of the events. However, if the Journal News had serious doubts concerning the truth of Thompson’s remarks, but was committed to running the story, there was good reason not to interview Stephens — while denials coming from Connaughton’s supporters might be explained as motivated by a desire to assist Connaughton, a denial coming from Stephens would quickly put an end to the story. The remaining six witnesses, including Connaughton, were all interviewed separately on October 31. Each of them denied Alice Thompson’s charges and corroborated Connaughton’s version of the events. Thus, one Journal News reporter testified at trial that Jeanette and Ernest Barnes denied that any promises, offers, or inducements were made and that he had known the Barneses for several years and considered them both credible. Id., at 89-90. Another reporter testified that she interviewed Dave Berry and that Berry stated that absolutely no promises or offers were made. Id., at 91-92. By the time the November 1 story ap- HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 683 657 Opinion of the Court peared, six witnesses had consistently and categorically denied Thompson’s allegations, yet the newspaper chose not to interview the one witness that both Thompson and Connaughton claimed would verify their conflicting accounts of the relevant events. The newspaper’s decision not to listen to the tapes of the Stephens interview in Connaughton’s home also supports the finding of actual malice. During the Connaughton interview, Long and Blount asked if they could hear the tapes. Id., at 259. Connaughton agreed, ibid., and later made the tapes available, id., at 48, 142. Much of what Thompson had said about the interview could easily have been verified or disproved by listening to the tapes. Listening to the tapes, for example, would have revealed whether Thompson accurately reported that the tape recorders were selectively turned on and off and that Connaughton was careful not to speak while the recorders were running. Similarly, the tapes presented a simple means of determining whether Stephens and Thompson had been asked leading questions, as Thompson claimed. Furthermore, if Blount was truly in equipoise about the question whether to endorse the incumbent judge for reelection—as he indicated in the column that he published on Sunday, October 30—it is difficult to understand his lack of interest in a detailed description of the corrupt disposition of 40 to 50 cases in Judge Dolan’s court. Even though he may have correctly assumed that the account did not reflect on the integrity of the judge himself, surely the question whether administrative shortcomings might be revealed by the tapes would be a matter in which an editor in the process of determining which candidate to endorse would normally have an interest.32 Although simply one piece of 32 Blount testified at trial as follows: “Q. . . . Did you listen to any of the tapes of the interview conducted by Dan Connaughton with Miss Stephens and Miss Thompson on the 17th of September? Did you listen to any of those tapes before you approved and 684 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. evidence in a much larger picture, one might reasonably infer in light of this broader context that the decision not to listen to the tapes was motivated by a concern that they would raise additional doubts concerning Thompson’s veracity. Moreover, although also just a small part of the larger picture, Blount’s October 30 editorial can be read to set the stage for the November 1 article. Significantly, this editorial appeared before Connaughton or any of the other witnesses were interviewed. Its prediction that further information concerning the integrity of the candidates might surface in the last few days of the campaign can be taken to indicate that Blount had already decided to publish Thompson’s allegations, regardless of how the evidence developed and regardless of whether or not Thompson’s story was credible upon ultimate reflection. Finally, discrepancies in the testimony of Journal News witnesses may have given the jury the impression that the published the article about Dan Connaughton on the figures of November 18, 1983? “A. No, because we had from several sources what was on the tape, there was several sources including Mr. Connaughton, that there was no mention of things we were exploring at this time[.] “Q. You were, I presume, concerned that you were dealing with a credible person in Alice Thompson, were you not? “A. Correct. “Q. Wouldn’t one of the simplest ways to determine her credibility be to play the tape to see whether her statement that Dan’s voice is not on it is true? “A. No, because we had been told from other sources that this matter, as I previously said, saying it was not on the tape. This was not discussed on the tape. We had been told by other persons that the tape was junk as far as evidence. “Q. The tape was what? “A. Junk.” App. 30-31. Blount further testified that by the time of trial, almost two years after he received the tapes, he had only listened to 15 minutes of the 2V2 hours of tape. Id., at 33. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 685 657 Opinion of the Court failure to conduct a complete investigation involved a deliberate effort to avoid the truth. Thus, for example, Blount’s superiors testified that they understood that Blount had directed reporter Tom Grant to ask the police whether Thompson had repeated her charges against Connaughton to them and whether they considered her a credible witness. Id., at 86-87 (Walker), 95 (Cocozzo). Blount also so testified. Id., at 37-38. Grant, however, denied that he had been given such an assignment. Id., at 88. Similarly, at the early stages of the proceeding, there was testimony that on October 31 Pam Long had tried to arrange a meeting with Patsy Stephens over the telephone, id., at 94, that Blount was standing at her desk during the conversation and overheard Long talking to Stephens, id., at 36-37, and that Connaughton had volunteered that he would have Stephens get in touch with them, id., at 57. Connaughton categorically denied that the issue of getting in touch with Stephens was even discussed, id., at 142, and ultimately Blount and Long agreed that there was no contact—and no attempt to make contact—with Stephens on the 31st or at any other time before the story was published, id., at 48-49 (Blount), 56-57 (Long). V The question whether the evidence in the record in a defamation case is sufficient to support a finding of actual malice is a question of law. Bose Corp. v. Consumers Union of United States, Inc., 466 U. S., at 510-511. This rule is not simply premised on common-law tradition,33 but on the 33 The following cases are illustrative of this tradition: Bose, 466 U. S., at 510-511 (actual malice); Jenkins v. Georgia, 418 U. S. 153, 161 (1974) (obscenity); Hess v. Indiana, 414 U. S. 105, 108-109 (1973) (per curiam) (incitement); Miller v. California, 413 U. S. 15, 25 (1973) (obscenity); Time, Inc. v. Pape, 401 U. S. 279, 284 (1971) (actual malice); Greenbelt Cooperative Publishing Assn., Inc. v. Bresler, 398 U. S. 6, 11 (1970) (defamation); Street v. New York, 394 U. S. 576, 589, 592 (1969) (fighting words); Beckley Newspapers Corp. v. Hanks, 389 U. S. 81, 83 (1967) (per curiam) (actual malice); New York Times Co. v. Sullivan, 376 U. S. 254, 285 (1964) 686 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. unique character of the interest protected by the actual malice standard. Our profound national commitment to the free exchange of ideas, as enshrined in the First Amendment, demands that the law of libel carve out an area of “ ‘breathing space’ ” so that protected speech is not discouraged. Gertz, 418 U. S., at 342 (quoting NAACP v. Button, 371 U. S. 415, 433 (1963)); New York Times Co., 376 U. S., at 272 (same). The meaning of terms such as “actual malice”—and, more particularly, “reckless disregard”—however, is not readily captured in “one infallible definition.” St. Amant v. Thompson, 390 U. S., at 730. Rather, only through the course of case-by-case adjudication can we give content to these otherwise elusive constitutional standards. Bose, supra, at 503. Moreover, such elucidation is particularly important in the area of free speech for precisely the same reason that the actual malice standard is itself necessary. Uncertainty as to the scope of the constitutional protection can only dissuade protected speech—the more elusive the standard, the less protection it affords. Most fundamentally, the rule is premised on the recognition that “[j Judges, as expositors of the Constitution,” have a duty to “independently decide whether the evidence in the record is sufficient to cross the constitutional threshold that bars the entry of any judgment that is not supported by clear and convincing proof of ‘actual malice.’” Bose, supra, at 511. There is little doubt that “public discussion of the qualifications of a candidate for elective office presents what is probably the strongest possible case for application of the New York Times rule,” Ocala Star-Banner Co. v. Damron, 401 U. S. 295, 300 (1971), and the strongest possible case for in- (actual malice); Edwards v. South Carolina, 372 U. S. 229, 235 (1963) (peaceful assembly); Niemotko v. Maryland, 340 U. S. 268, 271 (1951) (failure to issue license for religious meeting in public park); Pennekamp v. Florida, 328 U. S. 331, 335 (1946) (clear and present danger to integrity of court). HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 687 657 Opinion of the Court dependent review. As Madison observed in 1800, just nine years after ratification of the First Amendment: “Let it be recollected, lastly, that the right of electing the members of the government constitutes more particularly the essence of a free and responsible government. The value and efficacy of this right depends on the knowledge of the comparative merits and demerits of the candidates for public trust, and on the equal freedom, consequently, of examining and discussing these merits and demerits of the candidates respectively.” 4 J. Elliot, Debates on the Federal Constitution 575 (1861). This value must be protected with special vigilance. When a candidate enters the political arena, he or she “must expect that the debate will sometimes be rough and personal,” Oilman n. Evans, 242 U. S. App. D. C. 301, 333, 750 F. 2d 970, 1002 (1984) (en banc) (Bork, J., concurring), cert, denied, 471 U. S. 1127 (1985), and cannot “‘cry Foul!’ when an opponent or an industrious reporter attempts to demonstrate” that he or she lacks the “sterling integrity” trumpeted in campaign literature and speeches, Monitor Patriot Co. v. Roy, 401 U. S. 265, 274 (1971). Vigorous reportage of political campaigns is necessary for the optimal functioning of democratic institutions and central to our history of individual liberty.34 34 Of course, the protection of “calculated falsehoods” does not promote self-determination. As we observed in Garrison v. Louisiana, 379 U. S. 64 (1964): “At the time the First Amendment was adopted, as today, there were those unscrupulous enough and skillful enough to use the deliberate or reckless falsehood as an effective political tool to unseat the public servant or even topple an administration. Cf. Riesman, Democracy and Defamation: Fair Game and Fair Comment I, 42 Col. L. Rev. 1085, 1088-1111 (1942). That speech is used as a tool for political ends does not automatically bring it under the protective mantle of the Constitution. For the use of the known lie as a tool is at once at odds with the premises of democratic government and with the orderly manner in which economic, social, or political change is to be effected.” Id., at 75. 688 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. We have not gone so far, however, as to accord the press absolute immunity in its coverage of public figures or elections. If a false and defamatory statement is published with knowledge of falsity or a reckless disregard for the truth, the public figure may prevail. See Curtis Publishing Co. v. Butts, 388 U. S., at 162 (opinion of Warren, C. J.). A “reckless disregard” for the truth, however, requires more than a departure from reasonably prudent conduct. “There must be sufficient evidence to permit the conclusion that the defendant in fact entertained serious doubts as to the truth of his publication.” St. Amant, 390 U. S., at 731. The standard is a subjective one—there must be sufficient evidence to permit the conclusion that the defendant actually had a “high degree of awareness of . . . probable falsity.” Garrison v. Louisiana, 379 U. S., at 74. As a result, failure to investigate before publishing, even when a reasonably prudent person would have done so, is not sufficient to establish reckless disregard. See St. Amant, supra, at 731, 733. See also Hunt v. Liberty Lobby, 720 F. 2d 631, 642 (CA11 1983); Schultz v. Newsweek, Inc., 668 F. 2d 911, 918 (CA6 1982). In a case such as this involving the reporting of a third party’s allegations, “recklessness may be found where there are obvious reasons to doubt the veracity of the informant or the accuracy of his reports.” St. Amant, supra, at 732. In determining whether the constitutional standard has been satisfied, the reviewing court must consider the factual record in full. Although credibility determinations are reviewed under the clearly-erroneous standard because the trier of fact has had the “opportunity to observe the demeanor of the witnesses,” Bose, 466 U. S., at 499-500, the reviewing court must “ ‘examine for [itself] the statements in issue and the circumstances under which they were made to see . . . whether they are of a character which the principles of the First Amendment. . . protect,’” New York Times Co., 376 U. S., at 285 (quoting Pennekamp v. Florida, 328 U. S. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 689 657 Opinion of the Court 331, 335 (1946)).35 Based on our review of the entire record, we agree with the Court of Appeals that the evidence did in fact support a finding of actual malice. Our approach, however, differs somewhat from that taken by the Court of Appeals. In considering the actual malice issue, the Court of Appeals identified 11 subsidiary facts that the jury “could have” found.36 842 F. 2d, at 843-844. The court held that such 85 Petitioner concedes that “when conducting the independent review mandated by New York Times and Bose, a reviewing court should properly hesitate to disregard a jury’s opportunity to observe live testimony and assess witness credibility.” Brief for Petitioner 36, n. 45. It contends, however, that this Court did reject the trial court’s credibility determination in Bose. We disagree with this reading of Bose. In Bose we accepted the trial court’s determination that the author of the report at issue did not provide credible testimony concerning the reason for his choice of words and his understanding of the meaning of the word “about.” 466 U. S., at 511-512. Unlike the District Court, however, we were unwilling to infer actual malice from the finding that the witness “refused to admit [his mistake] and steadfastly attempted to maintain that no mistake had been made—that the inaccurate was accurate.” Id., at 512. 36 The Court of Appeals asserted: “A review of the entire record of the instant case disclosed substantial probative evidence from which a jury could have concluded (1) that the Journal was singularly biased in favor of Dolan and prejudiced against Connaughton as evidenced by the confidential personal relationship that existed between Dolan and Blount, the Journal Editorial Director, and the unqualified, consistently favorable editorial and daily news coverage received by Dolan from the Journal as compared with the equally consistently unfavorable news coverage afforded Connaughton; (2) that the Journal was engaged in a bitter rivalry with the Cincinnati Enquirer for domination of the greater Hamilton circulation market as evidenced by Blount’s vituperous public statements and criticism of the Enquirer; (3) that the Enquirer's, initial expose of the questionable operation of the Dolan court was a high profile news attraction of great public interest and notoriety that had ‘scooped’ the Journal and by Blount’s own admission was the most significant story impacting the Connaughton-Dolan campaign[;] (4) that by discrediting Connaughton the Journal was effectively impugning the Enquirer thereby undermining its market share of the Hamilton area; (5) that Thompson’s emotional instability coupled with her obviously vin- 690 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. findings would not have been not clearly erroneous, id., at 844, and, based on its independent review, that when considered cumulatively they provide clear and convincing evidence of actual malice, id., at 847. We agree that the jury may have found each of those facts, but conclude that the case should be decided on a less speculative ground. Given the trial court’s instructions, the jury’s answers to the three special interrogatories, and an understanding of those facts not in dispute, it is evident that the jury must have rejected (1) the testimony of petitioner’s witnesses that Stephens was not contacted simply because Connaughton failed to place her in touch with the newspaper; (2) the testimony of Blount that he did not listen to the tapes simply because he thought they would provide him with no new information; and (3) the testimony of those Journal News employees who asserted that they believed Thompson’s allegations were substantially true. When these findings are considered alongside the undisputed dictive and antagonistic attitudes toward Connaughton as displayed during an interview on October 27, 1983, arranged by Billy New’s defense attorney, afforded the Journal an ideal vehicle to accomplish its objectives; (6) that the Journal was aware of Thompson’s prior criminal convictions and reported psychological infirmities and the treatment she had received for her mental condition; (7) that every witness interviewed by Journal reporters discredited Thompson’s accusations; (8) that the Journal intentionally avoided interviewing Stephens between October 27, 1983, the date of its initial meeting with Thompson, and November 1, 1983 when it printed its first story even though it knew that Stephens could either credit or discredit Thompson’s statements; (9) that the Journal knew that publication of Thompson’s allegations charging Connaughton with unethical conduct and criminal extortion and her other equally damaging statements would completely discredit and irreparably damage Connaughton personally, professionally and politically; (10) that its prepublication legal review was a sham; (11) that the Journal timed the release of the initial story so as to accommodate follow-up stories and editorial comments in a manner calculated to peak immediately before the election in an effort to maximize the effect of its campaign to discredit Connaughton and the Enquirer. ” 842 F. 2d, at 843-844. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 691 657 Opinion of the Court evidence, the conclusion that the newspaper acted with actual malice inexorably follows. There is no dispute that Thompson’s charges had been denied not only by Connaughton, but also by five other witnesses before the story was published. Thompson’s most serious charge—that Connaughton intended to confront the incumbent judge with the tapes to scare him into resigning and otherwise not to disclose the existence of the tapes — was not only highly improbable, but inconsistent with the fact that Connaughton had actually arranged a lie detector test for Stephens and then delivered the tapes to the police. These facts were well known to the Journal News before the story was published. Moreover, because the newspaper’s interviews of Thompson and Connaughton were captured on tape, there can be no dispute as to what was communicated, nor how it was said. The hesitant, inaudible, and sometimes unresponsive and improbable tone of Thompson’s answers to various leading questions raise obvious doubts about her veracity. Moreover, contrary to petitioner’s contention that the prepublication interview with Connaughton confirmed the factual basis of Thompson’s statements, Brief for Petitioner 47, review of the tapes makes clear that Connaughton unambiguously denied each allegation of wrongful conduct. Connaughton’s acknowledgment, for instance, that his wife may have discussed with Stephens and Thompson the possibility of working at an ice cream store that she might someday open, hardly confirms the allegations that Connaughton had promised to buy a restaurant for the sister’s parents to operate, that he would provide Stephens with a job at the Municipal Court, or even that he would provide Thompson with suitable work.37 It is extraordinarily unlikely that the 87 Nor can petitioner claim immunity from suit because portions of Thompson’s account of the relevant events were confirmed by Connaughton. “[T]he defamer may be [all] the more successful when he baits the hook with truth.” Afro-American Publishing Co. v. Jaffe, 125 U. S. App. D. C. 70, 76, 366 F. 2d 649, 655 (1966) (en banc). See also Tavoulareas, 692 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. reporters missed Connaughton’s denials simply because he confirmed certain aspects of Thompson’s story. It is also undisputed that Connaughton made the tapes of the Stephens interview available to the Journal News and that no one at the newspaper took the time to listen to them. Similarly, there is no question that the Journal News was aware that Patsy Stephens was a key witness and that they failed to make any effort to interview her. Accepting the jury’s determination that petitioner’s explanations for these omissions were not credible, it is likely that the newspaper’s inaction was a product of a deliberate decision not to acquire knowledge of facts that might confirm the probable falsity of Thompson’s charges. Although failure to investigate will not alone support a finding of actual malice, see St. Amant, 390 U. S., at 731, 733, the purposeful avoidance of the truth is in a different category. There is a remarkable similarity between this case—and in particular, the newspaper’s failure to interview Stephens and failure to listen to the tape recording of the September 17 interview at Connaughton’s home—and the facts that supported the Court’s judgment in Curtis Publishing Co. v. Butts, 388 U. S. 130 (1967). In Butts the evidence showed that the Saturday Evening Post had published an accurate account of an unreliable informant’s false description of the Georgia athletic director’s purported agreement to “fix” a college football game. Although there was reason to question the informant’s veracity, just as there was reason to doubt Thompson’s story, the editors did not interview a witness who had the same access to the facts as the informant and did not look at films that revealed what actually hap- 260 U. S. App. D. C., at 64, 817 F. 2d, at 787. Of course, the press need not accept “denials, however vehement; such denials are so commonplace in the world of polemical charge and countercharge that, in themselves, they hardly alert the conscientious reporter to the likelihood of error.” Edwards v. National Audubon Society, Inc., 556 F. 2d, at 121. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 693 657 Opinion of the Court pened at the game in question.38 This evidence of an intent to avoid the truth was not only sufficient to convince the plurality that there had been an extreme departure from professional publishing standards, but it was also sufficient to satisfy the more demanding New York Times standard applied by Chief Justice Warren,39 Justice Brennan, and Justice White.40 As in Butts, the evidence in the record in this case, when reviewed in its entirety, is “unmistakably” sufficient to support a finding of actual malice. The judgment of the Court of Appeals is accordingly Affirmed. 38 As Justice Harlan observed in Butts: “Burnett’s notes were not even viewed by any of the magazine’s personnel prior to publication. John Carmichael who was supposed to have been with Burnett when the phone call was overheard was not interviewed. No attempt was made to screen the films of the game to see if Burnett’s information was accurate, and no attempt was made to find out whether Alabama had adjusted its plans after the alleged divulgence of information.” 388 U. S., at 157. In this passage, “Stephens” might easily be substituted for “Carmichael,” “Thompson” for “Burnett,” and “the tapes” for “Burnett’s notes” and “the films of the game.” 39 Chief Justice Warren wrote: “The slipshod and sketchy investigatory techniques employed to check the veracity of the source and the inferences to be drawn from the few facts believed to be true are detailed at length in the opinion of Mr. Justice Harlan. Suffice it to say that little investigative effort was expended initially, and no additional inquiries were made even after the editors were notified by respondent and his daughter that the account to be published was absolutely untrue. Instead, the Saturday Evening Post proceeded on its reckless course with full knowledge of the harm that would likely result from publication of the article.” Id., at 169-170. 40 Although concluding that the case should be remanded for a new trial, Justice Brennan, joined by Justice White, agreed with Chief Justice Warren that the evidence presented at the original trial “unmistakably would support a judgment for Butts under the New York Times standard.” Id., at 172. 694 OCTOBER TERM, 1988 Blackmun, J., concurring 491 U. S. Justice White, with whom The Chief Justice joins, concurring. In my view, in cases like this the historical facts—e. g., who did what to whom and when—are reviewable only under the clearly-erroneous standard mandated by Federal Rule of Civil Procedure 52. Credibility determinations fall in this category, as does the issue of knowledge of falsity. But as I observed in dissent in Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485, 515 (1984), the reckless disregard component of the New York Times Co. v. Sullivan “actual malice” standard is not a question of historical fact. A trial court’s determination of that issue therefore is to be reviewed independently by the appellate court. As I read it, the Court’s opinion is consistent with these views, and—as Justice Kennedy observes—is consistent with the views expressed by Justice Scalia in his concurrence. Based on these premises, I join the Court’s opinion. Justice Blackmun, concurring. I agree with the majority’s analysis and with the result it reaches. I write separately, however, to stress two points. First, the case reaches us in an odd posture, one which stands in the way of giving full consideration to aspects of the content of the article under attack that perhaps are of constitutional significance. Petitioner has abandoned the defense of truth, see Philadelphia Newspapers, Inc. v. Hepps, 475 U. S. 767 (1986), despite the fact that there might be some support for that defense. We therefore must presume that the jury correctly found that the article was false, see ante, at 681, and decide whether petitioner acted with knowledge or reckless disregard of its falsity. In addition, petitioner has eschewed any reliance on the “neutral reportage” defense. Cf. Edwards v. National Audubon Society, Inc., 556 F. 2d 113, 120 (CA2), cert, denied, 434 U. S. 1002 (1977). This strategic decision appears to have been unwise in light of the facts of this case. The article accurately reported HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 695 657 Blackmun, J., concurring newsworthy allegations that Daniel Connaughton, a political candidate, had used “dirty tricks” to elicit information from Alice Thompson and her sister, information that had become central to the political campaign, and also accurately reported Connaughton’s response, which confirmed the existence of discussions with Thompson that touched upon the subject matter of her allegations but claimed that Thompson’s version of these discussions was incorrect. Were this Court to adopt the neutral reportage theory, the facts of this case arguably might fit within it. That question, however, has also not been squarely presented. Second, I wish to emphasize that the form and content of the story are relevant not only to the falsity and neutral reportage questions, but also to the question of actual malice. In the past, this Court’s decisions dealing with actual malice have placed considerable emphasis on the manner in which the allegedly false content was presented by the publisher. See Greenbelt Cooperative Publishing Assn., Inc. v. Bresler, 398 U. S. 6, 12-13 (1970) (truthful and accurate reporting of what was said at public meeting on issues of public importance not actionable); Time, Inc. v. Pape, 401 U. S. 279, 290-292 (1971) (erroneous interpretation of Government report not “actual malice”). Under our precedents, I find significant the fact that the article in this case accurately portrayed Thompson’s allegations as allegations, and also printed Connaughton’s partial denial of their truth. The form of the story in this case is markedly different from the form of the story in Curtis Publishing Co. v. Butts, 388 U. S. 130 (1967), where the informant’s description of the events was presented as truth rather than .as contested allegations. These differences in presentation are relevant to the question whether the publisher acted in reckless disregard of the truth: presenting the content of Thompson’s allegations as though they were established fact would have shown markedly less regard of their possible falsity. 696 OCTOBER TERM, 1988 Scalia, J., concurring in judgment 491 U. S. Several aspects of the majority’s opinion in this case might be interpreted as breaking with our practice of considering the form and content of the article in making malice determinations. The majority notes the form of the story, see ante, at 680-681, but its account of the evidence it finds probative of actual malice, ante, at 682-685, deals exclusively with evidence extrinsic to the story itself. The absence of any discussion of Pape and Bresler also might be understood as a suggestion that the manner in which the contested statements are presented is irrelevant to the malice inquiry. Finally, the majority relies upon Butts in the course of its discussion of petitioner’s purposefully incomplete investigation of its story, ante, at 692-693, in a manner that suggests it might not have accorded significance to the difference between the forms of the respective stories in Butts and in this case. I am confident, however, that these aspects of the majority’s opinion are omissions in explanation rather than in analysis, and that the majority’s opinion cannot fairly be read to hold that the content of the article is irrelevant to the actual malice inquiry. Because I am convinced that the majority has considered the article’s content and form in the course of its painstaking “review of the entire record,” see ante, at 689, and because I conclude that the result the majority reaches is proper even when the contents of the story are given due weight, I concur. Justice Kennedy, concurring. I join the opinion of the Court, for in my view it is not inconsistent with the analysis set out in Justice Scalia’s separate concurrence. Justice Scalia, concurring in the judgment. I agree with the Court’s disposition of this case, and with its resolution of the second legal issue on which we granted certiorari, namely whether “highly unreasonable conduct constituting an extreme departure from ordinary standards of investigation and reporting” is alone enough to establish HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 697 657 Scalia, J., concurring in judgment (rather than merely evidence of) the malice necessary to assess liability in public figure libel cases. I disagree, however, with the Court’s approach to resolving the first and most significant question upon which certiorari was granted, which was the following: “Whether, in a defamation action instituted by a candidate for public office, the First and Fourteenth Amendments obligate an appellate court to conduct an independent review of the entire factual basis for a jury’s finding of actual malice—a review that examines both the subsidiary facts underlying the jury’s finding of actual malice and the jury’s ultimate finding of actual malice itself.” That question squarely raised the conflict that the Sixth Circuit perceived it had created with an earlier decision of the District of Columbia Circuit, en banc, concerning the requirement we set forth in Bose Corp. n. Consumers Union of United States, Inc., 466 U. S. 485 (1984), that judges “exercise independent judgment” on the question “whether the record establishes actual malice with convincing clarity,” id., at 514. The nub of the conflict, which is of overwhelming importance in libel actions by public figures, is whether this means, as the Sixth Circuit understood the District of Columbia Circuit to have held in Tavoulareas v. Piro, 260 U. S. App. D. C. 39, 817 F. 2d 762 (1987) (en banc), that the trial judge and reviewing courts must make their own “independent” assessment of the facts allegedly establishing malice; or rather, as the Sixth Circuit held here (explicitly rejecting Tavoulareas), that they must merely make their own “independent” assessment that, assuming all of the facts that could reasonably be found in favor of the plaintiff were found in favor of the plaintiff, clear and convincing proof of malice was established. Today’s opinion resolves this issue in what seems to me a peculiar manner. The Court finds it sufficient to decide the present case to accept, not all the favorable facts that the 698 OCTOBER TERM, 1988 Scalia, J., concurring in judgment 491 U. S. jury could reasonably have found, but rather only the adequately supported favorable facts that the jury did find. Exercising its independent judgment just on the basis of those facts (and the uncontroverted evidence), it concludes that malice was clearly and convincingly proved. The crucial passage of the Court’s opinion is the following: “Given the trial court’s instructions, the jury’s answers to the three special interrogatories, and an understanding of those facts not in dispute, it is evident that the jury must have rejected (1) the testimony of petitioner’s witnesses that Stephens was not contacted simply because Connaughton failed to place her in touch with the newspaper; (2) the testimony of Blount that he did not listen to the tapes simply because he thought they would provide him with no new information; and (3) the testimony of those Journal News employees who asserted that they believed Thompson’s allegations were substantially true. When these findings are considered alongside the undisputed evidence, the conclusion that the newspaper acted with actual malice inextricably follows.” Ante, at 690-691 (emphasis in original). This analysis adopts the most significant element of the Sixth Circuit’s approach, since it accepts the jury’s determination of at least the necessarily found controverted facts, rather than making an independent resolution of that conflicting testimony. Of course the Court examines the evidence pertinent to the jury determination—as a reviewing court always must—to determine that the jury could reasonably have reached that conclusion. But the Court does not purport to be exercising its own independent judgment as to whether Stephens was not contacted simply because Connaughton failed to place her in touch with the newspaper, whether Blount did not listen to the tapes because he thought they would provide no new information, or whether the Journal News employees believed Thompson’s allegations to be substantially true. HARTE-HANKS COMMUNICATIONS v. CONNAUGHTON 699 657 Scalia, J., concurring in judgment While I entirely agree with this central portion of the Court’s analysis, I do not understand the Court’s approach in conducting that analysis only on the basis of the three factual determinations the Court selects. To begin with, I am dubious of the Court’s conclusion that the jury must have made all three of those findings in order to bring in the verdict that it did under the judge’s instructions, and in order to answer as it did the only relevant “special interrogatory,” which was “Do you unanimously find by clear and convincing proof that the publication in question was published with actual malice?” It seems to me, for example, that even if one believed Blount’s explanation of why he did not listen to the tapes, it would still be reasonable to find (and I would find) clear and convincing proof of malice from the utterly inexplicable failure to interview Stephens plus the uncontroverted evidence. More important, however, even if each of these factual findings happened to be necessary to the verdict and interrogatory response, I see no reason to make them the exclusive focus of our analysis, instead of consulting (as the Sixth Circuit did, and as courts invariably do when reviewing jury verdicts) all the reasonably supported findings that the jury could have made. It may well be true that “we need only consider those factual findings that were essential to the jury verdict” in the sense that referring to those alone is enough to decide the case—i. e., those alone establish clear and convincing proof of malice. But one could pick out any number of categories of permissible jury findings that would meet that test. For example, it might be true that we could find the requisite proof of malice by considering, not all the evidence in its light most favorable to the plaintiff, but only that evidence produced by a particular witness. We could then say “we need only consider the findings the jury might have made based on the testimony of Mr. Smith to decide this case.” I see no more logic in limiting the inquiry the way the Court has done than in limiting it in this latter fashion. 700 OCTOBER TERM, 1988 Scalia, J., concurring in judgment 491 U. S. That can be made plain by applying the Court’s approach to a situation in which the facts essential to the jury verdict happen not to establish clear and convincing proof of malice. Assume a case in which there are innumerable controverted allegations, dozens of which, if the plaintiff’s version is credited, would suffice to establish malice; but in which only one controverted allegation—the defendant’s allegation that he knew firsthand the truth of the libelous charges—could not possibly have been found against the plaintiff if the jury was to come in with the verdict that it did. If we applied today’s analysis to that situation, we would then proceed to ask whether the fact that the defendant did not know firsthand the truth of the charges, and that he lied about that, is alone enough to establish clear and convincing proof of malice. It clearly would not be. Surely, however, we would not reverse the judgment for the plaintiff, when dozens of other disputed contentions which the jury might have resolved in the plaintiff’s favor would establish clear and convincing proof. We would, as the Sixth Circuit did, assume that all those disputes were resolved in the plaintiff’s favor—unless, of course, we again devised some nonfunctional category of the remaining disputes that we could look to, perhaps those pertaining to testimony by Mr. Smith. In sum, while the Court’s opinion is correct insofar as the critical point of deference to jury findings is concerned, I see no basis for consulting only a limited number of the permissible findings. I would have adopted the Sixth Circuit’s analysis in its entirety, making our independent assessment of whether malice was clearly and convincingly proved on the assumption that the jury made all the supportive findings it reasonably could have made. That is what common-law courts have always done, and there is ultimately no alternative to it. JETT v. DALLAS INDEPENDENT SCHOOL DIST. 701 Syllabus JETT v. DALLAS INDEPENDENT SCHOOL DISTRICT CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT No. 87-2084. Argued March 28, 1989—Decided June 22, 1989* Petitioner Jett, a white male, was employed by respondent Dallas Independent School District (DISD) as a teacher, athletic director, and head football coach at a predominantly black high school. After repeated clashes with the school’s Principal Todd, a black man, over school policies and Jett’s handling of the school’s football program, Todd recommended that Jett be relieved of his duties as athletic director and coach. The DISD’s Superintendent Wright affirmed Todd’s recommendation and reassigned Jett to a teaching position in another school, where he had no coaching duties. Alleging, inter alia, that Todd’s recommendation was racially motivated, and that the DISD, acting through Todd and Wright, had discriminated against him on the basis of race in violation of 42 U. S. C. §§ 1981 and 1983 and the Equal Protection Clause, Jett brought this action in the District Court, which upheld a jury verdict in his favor on all counts. The Court of Appeals reversed in part and remanded, finding, among other things, that the District Court’s jury instructions as to the DISD’s liability under § 1983 were deficient, since (1) they did not make clear that, under Monell v. 'New York City Dept, of Social Services, 436 U. S. 658, such liability could be predicated on the actions of Todd or Wright only if those officials had been delegated policymaking authority or acted pursuant to a well settled custom that represented official policy; and (2) even if Wright could be considered a policymaker for purposes of the transfer of personnel, the jury made no finding that his decision to transfer Jett was either improperly motivated or consciously indifferent to the improper motivations of Todd. The Court of Appeals also rejected the District Court’s conclusion that the DISD’s § 1981 liability for Todd’s actions could be predicated on a respondeat superior theory, noting that Monell had held that Congress did not intend that municipalities be subject to vicarious liability under § 1983 for the federal constitutional or statutory violations of their employees, and declaring that to impose such liability for only certain wrongs based on § 1981 apparently would contravene the congressional intent behind § 1983. *Together with No. 88-214, Dallas Independent School District v. Jett, also on certiorari to the same court. 702 OCTOBER TERM, 1988 Syllabus 491 U. S. Held: The judgment is affirmed in part, and the cases are remanded. 798 F. 2d 748 and 837 F. 2d 1244, affirmed in part and remanded. Justice O’Connor delivered the opinion of the Court with respect to Parts I, III, and IV, concluding that: 1. A municipality may not be held liable for its employees’ violations of § 1981 under a respondeat superior theory. The express “action at law” provided by § 1983 for the “deprivation of. . . rights secured by the Constitution and laws” provides the exclusive federal damages remedy for the violation of the rights guaranteed by § 1981 when the claim is pressed against a state actor. Cf., e. g., Brown v. GSA, 425 U. S. 820, where the Court, in holding that § 717 of Title VII of the Civil Rights Act of 1964 constitutes the exclusive remedy for racial discrimination in federal employment despite the possibility of an implied damages remedy under § 1981, invoked the general principle that a precisely drawn, detailed statute pre-empts more general remedies. Monell, supra, specifically held that a municipality cannot be liable under § 1983 on a respondeat superior theory, while the Courts of Appeals in post-Monell decisions have unanimously rejected the contention, analogous to petitioner’s argument here, that that theory is available against municipalities under a Bivens-type action implied directly from the Fourteenth Amendment. Given this Court’s repeated recognition that the Fourteenth Amendment was largely intended to embody and expand the protections of § 1981’s statutory predecessor as against state actors, this Court declines petitioner’s invitation to imply a damages remedy broader than § 1983 from § 1981’s declaration of rights. Creation of such a remedy would allow § 1983’s carefully crafted remedial scheme to be circumvented by artful pleading. Nor can a respondeat superior standard be implied from 42 U. S. C. § 1988, since, although that statute does authorize district courts in civil rights actions to look to the common law if federal remedies are deficient, the statute specifically withdraws that authority where, as here, the common law remedy is inconsistent with federal law; i. e., with § 1983. See Moor v. County of Alameda, 411 U. S. 693, 706, 710, n. 27. Thus, to prevail against the DISD, petitioner must show that the violation of his § 1981 “right to make contracts” was caused by a custom or policy within the meaning of Monell and subsequent cases. Pp. 731-736. 2. These cases are remanded to the Court of Appeals to determine whether, in light of the principles enunciated in Monell, supra, and clarified in Pembaur v. Cincinnati, 475 U. S. 469, and St. Louis v. Pra-protnik, 485 U. S. 112, Superintendent Wright possessed final policy-making authority under Texas law in the area of employee transfers, and if so whether a new trial is required to determine the DISD’s responsibility for the actions of Principal Todd in light of this determination. Although the Court of Appeals correctly ruled that the District Court’s JETT v. DALLAS INDEPENDENT SCHOOL DIST. 703 701 Syllabus jury instructions constituted manifest error, the case was tried before Pembaur and Praprotnik were decided, and the Court of Appeals issued its decision before Praprotnik. Pp. 736-738. Justice O’Connor, joined by The Chief Justice, Justice White, and Justice Kennedy, concluded in Part II that the text and legislative history of both the Civil Rights Act of 1866 and the Civil Rights Act of 1871, the precursors of §§ 1981 and 1983 respectively, demonstrate that § 1981 does not provide an independent federal damages remedy for racial discrimination by local governmental entities; rather, Congress intended that the explicit remedial provisions of § 1983 control in the context of § 1981 damages actions against state actors. Pp. 711-731. (a) The legislative history of the 1866 Act, which was originally enacted to implement the Thirteenth Amendment, demonstrates that that Act neither provided an express damages remedy for violation of its provisions nor created any original federal jurisdiction which could support such a remedy against state actors, and that the Act’s penal section—the only provision explicitly directed at state officials—was designed to punish only the official committing a violation and not the municipality itself. Two congressional actions subsequent to the passage of the 1866 Act— the submission of the Fourteenth Amendment to the States for ratification, which Amendment was based upon, and widely viewed as “constitutionalizing,” that Act’s protections, and the reenactment of that Act’s substance in the Enforcement Act of 1870, a Fourteenth Amendment statute—further evidence the relationship between §§ 1981 and 1983 and demonstrate that § 1981 is both a Thirteenth and a Fourteenth Amendment statute. Pp. 713-722. (b) The text and legislative history of the 1871 Act, which was expressly enacted to enforce the Fourteenth Amendment, establish that: (1) unlike any portion of the 1866 Act, that statute explicitly exposed state and local officials to liability for damages in a newly created “action at law” for deprivation of constitutional rights; (2) the Act expanded federal jurisdiction by explicitly providing original jurisdiction in the federal courts for prosecution of such actions; and (3) the provision of the Act which is now § 1983 was explicitly modeled on the penal provision of the 1866 Act and was intended to amend and enhance the protections of that provision by providing a parallel civil remedy for the same violations. Thus, Jett’s contention that the 1866 Act had already created a broader federal damages remedy against state actors is unpersuasive. Moreover, the fact that Congress rejected the Sherman amendment to the 1871 Act—which specifically proposed the imposition of vicarious liability on municipal governments for injuries caused by mob violence directed at the enjoyment or exercise of federal civil rights—demonstrates an awareness of, and a desire to comply with, the then-reigning constitu 704 OCTOBER TERM, 1988 Syllabus 491 U. S. tional doctrine of “dual sovereignty,” which indicated that Congress did not have the power to assign the duty to enforce federal law to state instrumentalities by making them liable for the constitutional violations of others. Given this constitutional background, Jett’s contention that the 1866 Act had already silently created a form of vicarious liability against municipal governments is historically untenable. Furthermore, the addition, in 1874, of the phrase “and laws” to the remedial provision of what is now § 1983 indicates an intent that the guarantees contained in what is now § 1981 were to be enforced against state actors through § 1983’s express damages remedy. Pp. 722-731. Justice Scalia concluded that the respondeat superior question is properly decided solely on the rudimentary principles of construction that the specific—here, § 1983, which precludes liability on that basis for the precise category of offense at issue—governs the general—here, § 1981—and that, where the text permits, statutes dealing with similar subjects should be interpreted harmoniously. Pp. 738-739. O’Connor, J., announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I and IV, in which Rehnquist, C. J., and White, Scalia, and Kennedy, JJ., joined; the opinion of the Court with respect to Part III, in which Rehnquist, C. J., and White and Kennedy, JJ., joined, and in which Scalia, J., joined, except insofar as that Part relies on legislative history; and an opinion with respect to Part II, in which Rehnquist, C. J., and White and Kennedy, JJ., joined. Scalia, J., filed an opinion concurring in part and concurring in the judgment, post, p. 738. Brennan, J., filed a dissenting opinion, in which Marshall, Blackmun, and Stevens, JJ., joined, post, p. 739. Stevens, J., filed a dissenting opinion, post, p. 753. Frank Gilstrap argued the cause for petitioner in No. ST-2084 and respondent in No. 87-214. With him on the briefs were Frank Hill and Shane Goetz. Leonard J. Schwartz argued the cause and filed a brief for respondent in No. 87-2084 and petitioner in No. 87-214. f tBriefs of amici curiae urging reversal were filed for the NAACP Legal Defense and Educational Fund, Inc., et al. by Julius LeVonne Chambers and Eric Schnapper; and for the National Education Association by Michael H. Gottesman and Jeremiah A. Collins. Benna Ruth Solomon, Joyce Holmes Benjamin, Beate Bloch, Donald B. Ayer, Glen D. Nager, and Robert D. Sweeney, Jr., filed a brief for the International City Management Association et al. as amici curiae urging affirmance. JETT v. DALLAS INDEPENDENT SCHOOL DIST. 705 701 Opinion of the Court Justice O’Connor announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I, III, and IV, and an opinion with respect to Part II, in which The Chief Justice, Justice White, and Justice Kennedy join. The questions before us in these cases are whether 42 U. S. C. § 1981 provides an independent federal cause of action for damages against local governmental entities, and whether that cause of action is broader than the damages remedy available under 42 U. S. C. § 1983, such that a municipality may be held liable for its employees’ violations of § 1981 under a theory of respondeat superior. I Petitioner Norman Jett, a white male, was employed by respondent Dallas Independent School District (DISD) as a teacher, athletic director, and head football coach at South Oak Cliff High School (South Oak) until his reassignment to another DISD school in 1983. Petitioner was hired by the DISD in 1957, was assigned to assistant coaching duties at South Oak in 1962, and was promoted to athletic director and head football coach of South Oak in 1970. During petitioner’s lengthy tenure at South Oak, the racial composition of the school changed from predominantly white to predominantly black. In 1975, the DISD assigned Dr. Fredrick Todd, a black, as principal of South Oak. Petitioner and Todd clashed repeatedly over school policies, and in particular over petitioner’s handling of the school’s football program. These conflicts came to a head following a November 19,1982, football game between South Oak and the predominately white Plano High School. Todd objected to petitioner’s comparison of the South Oak team with professional teams before the match, and to the fact that petitioner entered the officials’ locker room after South Oak lost the game and told two black officials that he would never allow black officials to work another South Oak game. Todd also objected to petitioner’s 706 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. statements, reported in a local newspaper, to the effect that the majority of South Oak players could not meet proposed National Collegiate Athletic Association academic requirements for collegiate athletes. On March 15, 1983, Todd informed petitioner that he intended to recommend that petitioner be relieved of his duties as athletic director and head football coach at South Oak. On March 17, 1983, Todd sent a letter to John Kincaide, the director of athletics for DISD, recommending that petitioner be removed based on poor leadership and planning skills and petitioner’s comportment before and after the Plano game. Petitioner subsequently met with John Santillo, director of personnel for DISD, who suggested that petitioner should transfer schools because any remaining professional relationship with Principal Todd had been shattered. Petitioner then met with Linus Wright, the superintendent of the DISD. At this meeting, petitioner informed Superintendent Wright that he believed that Todd’s criticisms of his performance as head coach were unfounded and that in fact Todd was motivated by racial animus and wished to replace petitioner with a black head coach. Superintendent Wright suggested that the difficulties between Todd and petitioner might preclude petitioner from remaining in his coaching position at South Oak, but assured petitioner that another position in the DISD would be secured for him. On March 25, 1983, Superintendent Wright met with Kincaide, Santillo, Todd, and two other DISD officials to determine whether petitioner should remain at South Oak. After the meeting, Superintendent Wright officially affirmed Todd’s recommendation to remove petitioner from his duties as coach and athletic director at South Oak. Wright indicated that he felt compelled to follow the recommendation of the school principal. Soon after this meeting, petitioner was informed by Santillo that effective August 4, 1983, he was reassigned as a teacher at the DISD Business Magnet School, a position that did not include any coaching duties. Petitioner’s at- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 707 701 Opinion of the Court tendance and performance at the Business Magnet School were poor, and on May 5, 1983, Santillo wrote petitioner indicating that he was being placed on “unassigned personnel budget” and being reassigned to a temporary position in the DISD security department. Upon receiving Santillo’s letter, petitioner filed this lawsuit in the District Court for the Northern District of Texas. The DISD subsequently offered petitioner a position as a teacher and freshman football and track coach at Jefferson High School. Petitioner did not accept this assignment, and on August 19, 1983, he sent his formal letter of resignation to the DISD. Petitioner brought this action against the DISD and Principal Todd in his personal and official capacities, under 42 U. S. C. §§ 1981 and 1983, alleging due process, First Amendment, and equal protection violations. Petitioner’s due process claim alleged that he had a constitutionally protected property interest in his coaching position at South Oak, of which he was deprived without due process of law. Petitioner’s First Amendment claim was based on the allegation that his removal and subsequent transfer were actions taken in retaliation for his statements to the press regarding the sports program at South Oak. His equal protection and § 1981 causes of action were based on the allegation that his removal from the athletic director and head coaching positions at South Oak was motivated by the fact that he was white, and that Principal Todd, and through him the DISD, were responsible for the racially discriminatory diminution in his employment status. Petitioner also claimed that his resignation was in fact the product of racial harassment and retaliation for the exercise of his First Amendment rights and thus amounted to a constructive discharge. These claims were tried to a jury, which found for petitioner on all counts. The jury awarded petitioner $650,000 against the DISD, $150,000 against Principal Todd and the DISD jointly and severally, and $50,000 in punitive damages against Todd in his personal capacity. 708 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. On motion for judgment notwithstanding the verdict, the defendants argued that liability against the DISD was improper because there was no showing that petitioner’s injuries were sustained pursuant to a policy or custom of the school district. App. to Pet. for Cert, in No. 87-2084, p. 46A. The District Court rejected this argument, finding that the DISD Board of Trustees had delegated final and unreviewable authority to Superintendent Wright to reassign personnel as he saw fit. Id., at 47A. In any event, the trial court found that petitioner’s claim of racial discrimination was cognizable under § 1981 as well as § 1983, and indicated that “liability is permitted on solely a basis of respondeat superior when the claim is one of racial discrimination under § 1981.” Ibid. The District Court set aside the punitive damages award against Principal Todd as unsupported by the evidence, found the damages award against the DISD excessive and ordered a remittitur of $200,000, but otherwise denied the defendants’ motions for judgment n.o.v. and a new trial and upheld the jury’s verdict in all respects. Id., at 62A-63A. Principal Todd has reached a settlement with petitioner and is no longer a party to this action. Id., at 82A-84A. On appeal, the Court of Appeals for the Fifth Circuit reversed in part and remanded. 798 F. 2d 748 (1986). Initially, the court found that petitioner had no constitutionally protected property interest “in the intangible, noneconomic benefits of his assignment as coach.” Id., at 754. Since petitioner had received both his teacher’s and coach’s salary after his reassignment, the change in duties did not deprive him of any state law entitlement protected by the Due Process Clause. The Court of Appeals also set aside the jury’s finding that petitioner was constructively discharged from his teaching position within the DISD. The court found the evidence insufficient to sustain the claim that petitioner’s loss of coaching duties and subsequent offer of reassignment to a lesser coaching position were so humiliating or unpleasant that a reasonable employee would have felt compelled to re- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 709 701 Opinion of the Court sign. Id., at 754-756. While finding the question “very close,” the Court of Appeals concluded that there was sufficient evidence from which a reasonable jury could conclude that Principal Todd’s recommendation that petitioner be transferred from his coaching duties at South Oak was motivated by impermissible racial animus. The court noted that Todd had replaced petitioner with a black coach, that there had been racial overtones in the tension between Todd and petitioner before the Plano game, and that Todd’s explanation of his unsatisfactory rating of petitioner was questionable and was not supported by the testimony of other DISD officials who spoke of petitioner’s performance in laudatory terms. Id., at 756-757. The court also affirmed the jury’s finding that Todd’s recommendation that petitioner be relieved of his coaching duties was motivated in substantial part by petitioner’s protected statements to the press concerning the academic standing of athletes at South Oak. These remarks addressed matters of public concern, and Todd admitted that they were a substantial consideration in his decision to recommend that petitioner be relieved of his coaching duties. The Court of Appeals then turned to the DISD’s claim that there was insufficient evidence to support a finding of municipal liability under 42 U. S. C. § 1983. The Court of Appeals found that the District Court’s instructions as to the school district’s liability were deficient in two respects. First, the District Court’s instructions did not make clear that the school district could be held liable for the actions of Principal Todd or Superintendent Wright only if those officials were delegated policymaking authority by the school district or acted pursuant to a well settled custom that represented official policy. Second, even if Superintendent Wright could be considered a policymaker for purposes of the transfer of school district personnel, the jury made no finding that Superintendent Wright’s decision to transfer petitioner was 710 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. either improperly motivated or consciously indifferent to the improper motivations of Principal Todd. Id., at 759-760. The Court of Appeals also rejected the District Court’s conclusion that the DISD’s liability for Principal Todd’s actions could be predicated on a theory of respondeat superior under § 1981. The court noted that in Monell v. New York City Dept, of Social Services, 436 U. S. 658 (1978), this Court held that Congress did not intend municipalities to be subject to vicarious liability for the federal constitutional or statutory violations of their employees. The Court of Appeals reasoned that “[t]o impose such vicarious liability for only certain wrongs based on section 1981 apparently would contravene the congressional intent behind section 1983.” 798 F. 2d, at 762. The Court of Appeals published a second opinion in rejecting petitioner’s suggestion for rehearing en banc in which the panel gave further explanation of its holding that respondeat superior liability against local governmental entities was unavailable under § 1981. 837 F. 2d 1244 (1988). The Court of Appeals noted that our decision in Monell rested in part on the conclusion that “‘creation of a federal law of respondeat superior would have raised all the constitutional problems’ ” associated with the Sherman amendment which was rejected by the framers of § 1983. 837 F. 2d, at 1247, quoting Monell, supra, at 693. Because the Court of Appeals’ conclusion that local governmental bodies cannot be held liable under a theory of respondeat superior for their employees’ violations of the rights guaranteed by § 1981 conflicts with the decisions of other Courts of Appeals, see, e. g., Springer v. Seamen, 821 F. 2d 871, 880-881 (CAI 1987); Leonard v. Frankfort Electric and Water Plant Bd., 752 F. 2d 189, 194, n. 9 (CA6 1985) (dictum), we granted Norman Jett’s petition for certiorari in No. 87-2084. 488 U. S. 940 (1988). We also granted the DISD’s cross-petition for certiorari in No. 88-214, ibid., to clarify the application of our decisions in St. Louis v. Pra- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 711 701 Opinion of O’Connor, J. protnik, 485 U. S. 112 (1988) (plurality opinion), and Pem-baur v. Cincinnati, 475 U. S. 469 (1986) (plurality opinion), to the school district’s potential liability for the discriminatory actions of Principal Todd. We note that at no stage in the proceedings has the school district raised the contention that the substantive scope of the “right ... to make . . . contracts” protected by §1981 does not reach the injury suffered by petitioner here. See Patterson n. McLean Credit Union, ante, at 176-177. Instead, the school district has argued that the limitations on municipal liability under §1983 are applicable to violations of the rights protected by § 1981. Because petitioner has obtained a jury verdict to the effect that Dr. Todd violated his rights under § 1981, and the school district has never contested the judgment below on the ground that § 1981 does not reach petitioner’s employment injury, we assume for purposes of these cases, without deciding, that petitioner’s rights under § 1981 have been violated by his removal and reassignment. See Canton v. Harris, 489 U. S. 378, 388-389, n. 8 (1989); United States v. Leon, 468 U. S. 897, 905 (1984). See also this Court’s Rule 21.1(a). II Title 42 U. S. C. § 1981, as amended, provides that: “All persons within the jurisdiction of the United States shall have the same right in every State and Territory to make and enforce contracts, to sue, be parties, give evidence, and to the full and equal benefit of all laws and proceedings for the security of persons and property as is enjoyed by white citizens, and shall be subject to like punishment, pains, penalties, taxes, licenses, and exactions of every kind, and no other.” In essence, petitioner argues that in 1866 the 39th Congress intended to create a cause of action for damages against municipal actors and others who violated the rights now enumerated in § 1981. While petitioner concedes that the text 712 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. of the 1866 Act itself is completely silent on this score, see Brief for Petitioner 26, petitioner contends that a civil remedy was nonetheless intended for the violation of the rights contained in § 1 of the 1866 Act. Petitioner argues that Congress wished to adopt the prevailing approach to municipal liability to effectuate this damages remedy, which was respondeat superior. Petitioner concludes that with this federal damages remedy in place in 1866, it was not the intent of the 42d Congress, which passed present day § 1983, to narrow the more sweeping remedy against local governments which Congress had created five years earlier. Since “repeals by implication are not favored,” id., at 15 (citations omitted), petitioner concludes that § 1981 must provide an independent cause of action for racial discrimination against local governmental entities, and that this broader remedy is unaffected by the constraints on municipal liability announced in Monell. In the alternative, petitioner argues that even if § 1981 does not create an express cause of action for damages against local governmental entities, 42 U. S. C. § 1988 invites this Court to craft a remedy by looking to common law principles, which again point to a rule of respondeat superior. Brief for Petitioner 27-29. To examine these contentions, we must consider the text and history of both the Civil Rights Act of 1866 and the Civil Rights Act of 1871, the precursors of §§ 1981 and 1983 respectively. Justice Brennan’s dissent errs in asserting that we have strayed from the question upon which we granted certiorari. See post, at 739-740. Jett’s petition for certiorari asks us to decide “[w]hether a public employee who claims job discrimination on the basis of race must show that the discrimination resulted from official ‘policy or custom’ in order to recover under 42 U. S. C. § 1981.” Pet. for Cert, in No. 87-2084, p. i. In answering this question, the lower court looked to the relationship between §§ 1981 and 1983, and refused to differentiate “between sections 1981 and 1983 with respect to municipal respondeat superior liability.” 837 F. 2d, at 1247. In both his petition for certiorari and his brief on the merits JETT v. DALLAS INDEPENDENT SCHOOL DIST. 713 701 Opinion of O’Connor, J. in this Court, petitioner Jett took issue with the Court of Appeals’ conclusion that the express damages remedy under §1983 militated against the creation or implication of a broader damages remedy under § 1981. See Pet. for Cert, in No. 87-2084, pp. 14-16; Brief for Petitioner 14-25. Moreover, petitioner concedes that “private causes of action under Sections 1981 and 1982 do not arise from the express language of those statutes,” Brief for Petitioner 27, and asks this Court to “look to state law or to fashion a single federal rule,” of municipal damages liability under §1981. Id., at 28-29 (footnote omitted). We think it obvious that the question whether a federal damages remedy broader than that provided by § 1983 should be implied from § 1981 is fairly included in the question upon which we granted certiorari. Equally implausible is Justice Brennan’s suggestion that we have somehow unwittingly answered this question in the past. See post, at 741. Most of the cases cited by the dissent involved private conduct, and thus quite obviously could not have considered the propriety of judicial implication of a federal damages remedy under § 1981 in the state action context we address here. The only two cases cited by Justice Brennan which involved state actors, Takahashi v. Fish and Game Comm’n, 334 U. S. 410 (1948), and Hurd n. Hodge, 334 U. S. 24 (1948), are completely inapposite. See post, at 745. Takahashi involved a mandamus action filed in state court, and thus understandably had nothing to say about federal damages remedies against state actors under § 1981. Hurd also involved only injunctive relief, and could not have considered the relationship of § 1981 to § 1983, since the latter statute did not apply to the District of Columbia at the time of our decision in that case. See District of Columbia n. Carter, 409 U. S. 418 (1973). A On December 18, 1865, the Secretary of State certified that the Thirteenth Amendment had been ratified and become part of the Constitution. Less than three weeks later, 714 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. Senator Lyman Trumbull, Chairman of the Senate Judiciary Committee, introduced S. 61, which was to become the Civil Rights Act of 1866. See Cong. Globe, 39th Cong., 1st Sess., 129 (1866). The bill had eight sections as introduced, the first three of which are relevant to our inquiry here. Section 1, as introduced to the Senate by Trumbull, provided: “That there shall be no discrimination in civil rights or immunities among the inhabitants of any State or Territory of the United States on account of race, color, or previous condition of slavery; but the inhabitants of every race and color, without regard to any previous condition of slavery or involuntary servitude, except as a punishment for a crime whereof the party shall have been duly convicted, shall have the same right to make and enforce contracts, to sue, be parties, and give evidence, to inherit, purchase, lease, sell, hold, and convey real and personal property, and to the full and equal benefit of all laws and proceedings for the security of person and property, and shall be subject to like punishment, pains, and penalties, and to none other, any law, statute, ordinance, regulation, or custom to the contrary notwithstanding.” Id., at 474. On January 29, 1866, Senator Trumbull took the floor to describe S. 61 to his colleagues. Trumbull indicated that “the first section will amount to nothing more than the declaration in the Constitution itself unless we have the machinery to carry it into effect.” Id., at 475. The Senator then alluded to the second section of the bill which provided: “That any person who under color of any law, statute, ordinance, regulation, or custom shall subject, or cause to be subjected, any inhabitant of any State or Territory to the deprivation of any right secured or protected by this act, or to different punishment, pains, or penalties on account of such person having at any time been held in a condition of slavery or involuntary servitude, . . . JETT v. DALLAS INDEPENDENT SCHOOL DIST. 715 701 Opinion of O’Connor, J. or by reason of his color or race, than is prescribed for the punishment of white persons, shall be deemed guilty of a misdemeanor, and, on conviction, shall be punished by fine not exceeding $1,000, or imprisonment not exceeding one year, or both, in the discretion of the court.” Ibid. Senator Trumbull told the Senate: “This is the valuable section of the bill so far as protecting the rights of freedmen is concerned.” Ibid. This section would allow for criminal prosecution of those who denied the freedman the rights protected by § 1, and Trumbull felt, in retrospect somewhat naively, that, “it will only be necessary to go into the late slaveholding States and subject to fine and imprisonment one or two in a State, and the most prominent ones I should hope at that, to break up this whole business.” Ibid. Trumbull then described the third section of the bill, which, as later enacted, provided in pertinent part: “That the district courts of the United States, within their respective districts, shall have, exclusively of the courts of the several States, cognizance of all crimes and offenses committed against the provisions of this act, and also, concurrently with the circuit courts of the United States, of all causes, civil and criminal, affecting persons who are denied or cannot enforce in the courts or judicial tribunals of the State or locality where they may be any of the rights secured to them by the first section of this act; and if any suit or prosecution, civil or criminal, has been or shall be commenced in any State court, against any such person, for any cause whatsoever . . . such defendant shall have the right to remove such cause for trial to the proper district or circuit court in the manner prescribed by the ‘Act relating to habeas corpus and regulating judicial proceedings in certain cases,’ approved March three, eighteen hundred and sixty three, and all acts amendatory thereof.” 14 Stat. 27. 716 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. Trumbull described this section as “giving to the courts of the United States jurisdiction over all persons committing offenses against the provisions of this act, and also over the cases of persons who are discriminated against by State laws or customs.” Cong. Globe, 39th Cong., 1st Sess., 475 (1866). Much of the debate in both the Senate and the House over the 1866 Act was taken up with the meaning of the terms “civil rights or immunities” contained in the first sentence of § 1 of the bill as introduced in the Senate. The phrase remained in the bill throughout the Senate’s consideration of S. 61, but was stricken by amendment in the House shortly before that body passed the bill. Discussion of §2 of the bill focused on both the propriety and constitutionality of subjecting state officers to criminal punishment for effectuating discriminatory state laws. Opponents of the bill consistently referred to criminal punishment and fines being levied against state judges and other state officers for the enforcement of state laws in conflict with § 1. See id., at 475, 499, 500 (Sen. Cowan); id., at 598 (Sen. Davis); id., at 1121 (Rep. Rogers); id., at 1154 (Rep. Eldridge). They never intimated that they understood any part of the bill to create a federal damages remedy against state officers or the political subdivisions of the States. Debate concerning §3 focused on the right of removal of civil and criminal proceedings commenced in state court. Senator Howard, an opponent, engaged in a section by section criticism of the bill after its introduction by Trumbull. As to § 3 he gave numerous examples of his perception of its operation. All of these involved removal of actions from state court, and none alluded to original federal jurisdiction except in the case of the exclusive criminal jurisdiction expressly provided for. Id., at 479 (“All such cases will be subject to be removed into the Federal courts”); see also id., at 598 (Sen. Davis) (“Section three provides that all suits brought in State courts that come within the purview of the previous sections may be removed into the Federal courts”). JETT v. DALLAS INDEPENDENT SCHOOL DIST. 717 701 Opinion of O’Connor, J. On February 2, 1866, the bill passed the Senate by a vote of 33 to 12 and was sent to the House. Id., at 606-607. Representative Wilson of Iowa, Chairman of the House Judiciary Committee, introduced S. 61 in the House on March 1, 1866. Of § 1 of the bill, he said: “Mr. Speaker, I think I may safely affirm that this bill, so far as it declares the equality of all citizens in the enjoyment of civil rights and immunities merely affirms existing law. We are following the Constitution. . . . It is not the object of this bill to establish new rights, but to protect and enforce those which already belong to every citizen.” Id., at 1117. As did Trumbull in the Senate, Wilson immediately alluded to § 2, the criminal provision, as the main enforcement mechanism of the bill. “In order to accomplish this end, it is necessary to fortify the declaratory portions of this bill with sanctions as will render it effective.” Id., at 1118. The only discussion of a civil remedy in the House debates surrounding the 1866 Act came in response to Representative Bingham’s proposal to send the bill back to the House Judiciary Committee with instructions “to strike out all parts of said bill which are penal and which authorize criminal proceedings, and in lieu thereof to give all citizens of the United States injured by denial or violation of any of the other rights secured or protected by said act, an action in the United States courts, with double costs in all cases of emergency, without regard to the amount of damages.” Id., at 1266, 1291. Bingham was opposed to the civil rights bill strictly on the grounds that it exceeded the constitutional power of the Federal Government. As to States “sustaining their full constitutional relation to the Government of the United States,” Bingham, along with several other Republicans, doubted the power of the Federal Government to interfere with the reserved powers of the States to define property and other rights. Id., at 1292. While Bingham realized that the same constitutional objections applied to his proposal 718 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. for modification of the bill, he felt that these would make the bill “less oppressive, and therefore less objectionable.” Id., at 1291. Representative Wilson responded to his Republican colleague’s proposal. Wilson pointed out that there was no difference in constitutional principle “between saying that the citizen shall be protected by the legislative power of the United States in his rights by civil remedy and declaring that he shall be protected by penal enactments against those who interfere with his rights.” Id., at 1295. Wilson did however see a difference in the effectiveness of the two remedies. He stated: “This bill proposes that the humblest citizen shall have full and ample protection at the cost of the Government, whose duty it is to protect him. The [Bingham] amendment . . . recognizes the principle involved, but it says that the citizen despoiled of his rights, instead of being properly protected by the Government, must press his own way through the courts and pay the bills attendant thereon. . . . The highest obligation which the Government owes to the citizen in return for the allegiance exacted of him is to secure him in the protection of his rights. Under the amendment of the gentleman the citizen can only receive that protection in the form of a few dollars in the way of damages, if he shall be so fortunate as to recover a verdict against a solvent wrongdoer. This is called protection. This is what we are asked to do in the way of enforcing the bill of rights. Dollars are weighed against the right of life, liberty and property.” Ibid. Bingham’s proposal was thereafter defeated by a vote of 113 to 37. Id., at 1296. The Senate bill was subsequently carried in the House, after the removal of the “civil rights and immunities” language in § 1, and an amendment adding a ninth section to the bill providing for a final appeal to the Supreme Court in cases arising under the Act. Id., at 1366- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 719 701 Opinion of O’Connor, J. 1367. On March 15, 1866, the Senate concurred in the House amendments without a record vote, see id., at 1413-1416, and the bill was sent to the President. After holding the bill for a full 10 days, President Johnson vetoed the bill and returned it to the Senate with his objections. The President’s criticisms of §§ 2 and 3 of the bill, and Senator Trumbull’s responses thereto, are particularly illuminating. As to § 2, the President declared that it was designed to counteract discriminatory state legislation, “by imposing fine and imprisonment upon the legislators who may pass such . . . laws.” Id., at 1680. As to the third section, the President indicated that it would vest exclusive federal jurisdiction over all civil and criminal cases where the rights guaranteed in § 1 were affected. Ibid. Trumbull took issue with both statements. As to the charge that §2 would result in the criminal prosecution of state legislators, Trumbull replied: “Who is to be punished? Is the law to be punished? Are the men who make the law to be punished? Is that the language of the bill? Not at all. If any person, ‘under color of any law,’ shall subject another to the deprivation of a right to which he is entitled, he is to be punished. Who? The person who, under the color of the law, does the act, not the men who made the law. In some communities in the South a custom prevails by which different punishment is inflicted upon the blacks from that meted out to whites for the same offense. Does this section propose to punish the community where the custom prevails? Or is it to punish the person who, under color of the custom, deprives the party of his right? It is a manifest perversion of the meaning of the section to assert anything else.” Id., at 1758. Trumbull also answered the President’s charge that the third section of the bill created original federal jurisdiction in all cases where a freedman was involved in a state court proceeding. He stated: 720 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. “So in reference to this third section, the jurisdiction is given to the Federal courts of a case affecting the person that is discriminated against. Now, he is not necessarily discriminated against, because there may be a custom in the community discriminating against him, nor because a Legislature may have passed a statute discriminating against him; that statute is of no validity if it comes in conflict with a statute of the United States; and it is not to be presumed that any judge of a State court would hold that a statute of a State discriminating against a person on account of color was valid when there was a statute of the United States with which it was in direct conflict, and the case would not therefore rise in which a party was discriminated against until it was tested, and then if the discrimination was held valid he would have a right to remove it to a Federal court.” Id., at 1759. Senator Trumbull then went on to indicate that “[i]f it be necessary in order to protect the freedman in his rights that he should have authority to go into the Federal courts in all cases where a custom [of discrimination] prevails in a State ... I think we have the authority to confer that jurisdiction under the second clause of the constitutional amendment.” Ibid. Two days later, on April 6, 1866, the Senate overrode the President’s veto by a vote of 33 to 15. Id., at 1809. On April 9, 1866, the House received both the bill and the President’s veto message which were read on the floor. Id., at 1857-1860. The House then promptly overrode the President’s veto by a vote of 122 to 41, id., at 1861, and the Civil Rights Act of 1866 became law. Several points relevant to our present inquiry emerge from the history surrounding the adoption of the Civil Rights Act of 1866. First, nowhere did the Act provide for an express damages remedy for violation of the provisions of § 1. See Jones n. Alfred H. Mayer Co., 392 U. S. 409, 414, n. 13 (1968) (noting “[t]hat 42 U. S. C. § 1982 is couched in declara- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 721 701 Opinion of O’Connor, J. tory terms and provides no explicit method of enforcement”); Sullivan v. Little Hunting Park, Inc., 396 U. S. 229, 238 (1969); Cannon n. University of Chicago, 441 U. S. 677, 690, n. 12 (1979); id., at 728 (White, J., dissenting). Second, no original federal jurisdiction was created by the 1866 Act which could support a federal damages remedy against state actors. See Allen n. McCurry, 449 U. S. 90, 99, n. 14 (1980) (§3 of the 1866 Act embodied remedy of “postjudgment removal for state-court defendants whose civil rights were threatened”); Georgia v. Rachel, 384 U. S. 780, 788-789 (1966); Strauder n. West Virginia, 100 U. S. 303, 311-312 (1880). Finally, the penal provision, the only provision explicitly directed at state officials, was, in Senator Trumbull’s words, designed to punish the “person who, under the color of the law, does the act,” not “the community where the custom prevails.” Cong. Globe, 39th Cong., 1st Sess., 1758 (1866). Two events subsequent to the passage of the 1866 Act bear on the relationship between §§ 1981 and 1983. First, on June 13, 1866, just over two months after the passage of the 1866 Act, a joint resolution was passed sending the Fourteenth Amendment to the States for ratification. As we have noted in the past, the first section of the 1866 Act “constituted an initial blueprint of the Fourteenth Amendment.” General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S. 375, 389 (1982). Many of the Members of the 39th Congress viewed § 1 of the Fourteenth Amendment as “constitutionalizing” and expanding the protections of the 1866 Act and viewed what became § 5 of the Amendment as laying to rest doubts shared by both sides of the aisle concerning the constitutionality of that measure. See, e. g., Cong. Globe, 39th Cong., 1st Sess., 2465 (1866) (Rep. Thayer) (“As I understand it, it is but incorporating in the Constitution of the United States the principle of the civil rights bill which has lately become a law”); id., at 2498 (Rep. Broomall); id., at 2459 (Rep. Stevens); id., at 2461 (Rep. Finck); id., at 2467 722 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. (Rep. Boyer). See also Hurd n. Hodge, 334 U. S., at 32 (“[A]s the legislative debates reveal, one of the primary purposes of many members of Congress in supporting the adoption of the Fourteenth Amendment was to incorporate the guaranties of the Civil Rights Act of 1866 in the organic law of the land”) (footnote omitted). Second, the 41st Congress reenacted the substance of the 1866 Act in a Fourteenth Amendment statute, the Enforcement Act of 1870. 16 Stat. 144. Section 16 of the 1870 Act was modeled after § 1 of the 1866 Act. Section 17 reenacted with some modification the criminal provisions of §2 of the earlier civil rights law, and § 18 of the 1870 Act provided that the entire 1866 Act was reenacted. See Civil Rights Cases, 109 U. S. 3, 16-17 (1883). We have thus recognized that present day 42 U. S. C. §1981 is both a Thirteenth and a Fourteenth Amendment statute. Runyon n. McCrary, 427 U. S. 160, 168-169, n. 8 (1976); id., at 190 (Stevens, J., concurring); General Building Contractors, supra, at 383-386. B What is now § 1983 was enacted as § 1 of “An Act to Enforce the Provisions of the Fourteenth Amendment to the Constitution of the United States and For other Purposes,” Act of April 20, 1871, ch. 22, 17 Stat. 13. The immediate impetus for the bill was evidence of widespread acts of violence perpetrated against the freedmen and loyal white citizens by groups such as the Ku Klux Klan. On March 23, 1871, President Grant sent a message to Congress indicating that the Klan’s reign of terror in the Southern States had “render[ed] life and property insecure,” and that “the power to correct these evils [was] beyond the control of State authorities.” Cong. Globe, 42d Cong., 1st Sess., 244 (1871). A special joint committee consisting of 10 distinguished Republicans, 5 from each House of Congress, was formed in response to President Grant’s call for legislation, and drafted the bill that became what is now known as the Ku Klux Act. As enacted, JETT v. DALLAS INDEPENDENT SCHOOL DIST. 723 701 Opinion of O’Connor, J. §§ 2 through 6 of the bill specifically addressed the problem of the private acts of violence perpetrated by groups like the Klan. Unlike the rest of the bill, §1 was not specifically addressed to the activities of the Klan. As passed by the 42d Congress, § 1 provided in full: “That any person who, under color of any law, statute, ordinance, regulation, custom, or usage of any State, shall subject, or cause to be subjected, any person within the jurisdiction of the United States to the deprivation of any rights, privileges, or immunities secured by the Constitution of the United States, shall, any such law, statute, ordinance, regulation, custom, or usage of the State to the contrary notwithstanding, be liable to the party injured in any action at law, suit in equity, or other proper proceeding for redress; such proceeding to be prosecuted in the several district or circuit courts of the United States, with and subject to the same rights of appeal, review upon error, and other remedies provided in like cases in such courts, under the provisions of the act of the ninth of April, eighteen hundred and sixty-six, entitled ‘An act to protect all persons in the United States in their civil rights, and to furnish the means of their vindication’; and the other remedial laws of the United States which are in their nature applicable in such cases.” 17 Stat. 13. Three points are immediately clear from the face of the Act itself. First, unlike any portion of the 1866 Act, this statute explicitly ordained that any “person” acting under color of state law or custom who was responsible for a deprivation of constitutional rights would “be liable to the party injured in any action at law.” Thus, “the 1871 Act was designed to expose state and local officials to a new form of liability.” Newport n. Fact Concerts, Inc., 453 U. S. 247, 259 (1981). Second, the 1871 Act explicitly provided original federal jurisdiction for prosecution of these civil actions against state 724 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. actors. See Will v. Michigan Dept, of State Police, ante, at 66 (“[A] principle purpose behind the enactment of § 1983 was to provide a federal forum for civil rights claims”); accord, Mitchum v. Foster, 407 U. S. 225, 239 (1972). Third, the first section of the 1871 Act was explicitly modeled on § 2 of the 1866 Act, and was seen by both opponents and proponents as amending and enhancing the protections of the 1866 Act by providing a new civil remedy for its enforcement against state actors. See Chapman n. Houston Welfare Rights Organization, 441 U. S. 600, 610-611, n. 25 (1979) (“Section 1 of the [1871] Act generated the least concern; it merely added civil remedies to the criminal penalties imposed by the 1866 Civil Rights Act”); Monroe v. Pape, 365 U. S. 167, 185 (1961); Mitchum, supra, at 238. Even a cursory glance at the House and Senate debates on the 1871 Act makes these three points clear. In introducing the bill to the House, Representative Shellabarger, who served on the joint committee which drafted the bill, stated: “The model for it will be found in the second section of the act of April 9, 1866, known as the ‘civil rights act.’ That section provides a criminal proceeding in identically the same case as this one provides a civil remedy for, except that the deprivation under color of State law must, under the civil rights act, have been on account of race, color or former slavery.” Cong. Globe, 42d Cong., 1st Sess., App. 68 (1871). Representative Shellabarger added that §1 provided a civil remedy “on the same state of facts” as §2 of the Civil Rights Act of 1866. Ibid. Obviously Representative Shellabarger’s introduction of § 1 of the bill to his colleagues would have been altogether different if he had been of the view that the 39th Congress, of which he had been a Member, had already created a broader federal damages remedy against state actors in 1866. The view that § 1 of the 1871 Act was an amendment of or supplement to the 1866 Act designed to create a new civil remedy against state actors was JETT v. DALLAS INDEPENDENT SCHOOL DIST. 725 701 Opinion of O’Connor, J. echoed throughout the debates in the House. See id., at 461 (Rep. Coburn); id., at App. 312-313 (Rep. Burchard). Opponents of § 1 operated on this same understanding. See id., at 429 (Rep. McHenry) (“The first section of the bill is intended as an amendment of the civil rights act”); id., at 365 (Rep. Arthur). Both proponents and opponents in the House viewed § 1 as working an expansion of federal jurisdiction. Supporters continually referred to the failure of the state courts to enforce federal law designed for the protection of the freedman, and saw § 1 as remedying this situation by interposing the federal courts between the State and citizens of the United States. See id., at 376 (Rep. Lowe) (“The case has arisen . . . when the Federal Government must resort to its own agencies to carry its own authority into execution. Hence this bill throws open the doors of the United States courts to those whose rights under the Constitution are denied or impaired”). Opponents recognized the expansion of original jurisdiction and railed against it on policy and constitutional grounds. See id., at 429 (Rep. McHenry) (“The first section of the bill . . . vests in the Federal courts jurisdiction to determine the individual rights of citizens of the same State; a jurisdiction which of right belongs only to the State tribunals”); id., at App. 50 (Rep. Kerr); id., at 365-366 (Rep. Authur); id., at 373 (Rep. Archer). The Senate debates on § 1 of the 1871 Act are of a similar tenor. Senator Edmunds, Chairman of the Senate Judiciary Committee, and one of the members of the joint committee which drafted the bill, introduced § 1 to the Senate in the following terms: “The first section is one that I believe nobody objects to, as defining the rights secured by the Constitution of the United States when they are assailed by any State law or under color of any State law, and it is merely carrying out the principles of the civil rights bill, which have 726 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. since become a part of the Constitution.” Id., at 568, quoted in Monroe v. Pape, supra, at 171. Again Senators addressed § 1 of the Act as creating a new civil remedy and expanding federal jurisdiction to accommodate it in terms incompatible with the supposition that the 1866 Act had already created such a cause of action against state actors. See Cong. Globe, 42d Cong., 1st Sess., 653 (1871) (Sen. Osborn) (“I believe the true remedy lies chiefly in the United States district and circuit courts. If the State courts had proven themselves competent... we should not have been called upon to legislate upon this subject at all. But they have not done so”); id., at App. 216 (Sen. Thurman) (“Its whole effect is to give to the Federal Judiciary that which does not belong to it—a jurisdiction that may be constitutionally conferred upon it, I grant, but that has never yet been conferred upon it”); see also id., at 501 (Sen. Frelinghuysen). The final aspect of the history behind the adoption of present day § 1983 relevant to the question before us is the rejection by the 42d Congress of the Sherman amendment, which specifically proposed the imposition of a form of vicarious liability on municipal governments. This history was thoroughly canvassed in the Court’s opinion in Monell, and only its broadest outlines need be traced here. Immediately prior to the vote on the bill in the Senate, Senator Sherman introduced an amendment which would have constituted a seventh section of the 1871 Act. Cong. Globe, 42d Cong., 1st Sess., 663 (1871). In its original form, the amendment did not place liability on municipal corporations per se, but instead rendered the inhabitants of a municipality liable in civil damages for injury inflicted to persons or property in violation of federal constitutional and statutory guarantees “by any persons riotously and tumultuously assembled together.” The initial Sherman amendment was passed by the Senate, but was rejected by the House and became the subject of a conference committee. The committee draft of the Sherman JETT v. DALLAS INDEPENDENT SCHOOL DIST. 727 701 Opinion of O’Connor, J. amendment explicitly provided that where injuries to person or property were caused by mob violence directed at the enjoyment or exercise of federal civil rights, “the county, city, or parish in which any of the said offenses shall be committed shall be liable to pay full compensation to the person or persons damnified by such offense.” Id., at 755. Judgments in such actions were to run directly against the municipal corporation, and were to be enforceable through a “lien. . . upon all moneys in the treasury of such county, city, or parish, as upon the other property thereof.” Ibid. Opposition to the amendment in this form was vehement, and ran across party lines, extending to many Republicans who had voted for § 1 of the 1871 Act, as well as earlier Reconstruction legislation, including the Civil Rights Act of 1866. See id., at 758 (Sen. Trumbull); id., at 798-799 (Rep. Farnsworth). The Sherman amendment was regarded as imposing a new and theretofore untested form of liability on municipal governments. As Representative Blair put it: “The proposition known as the Sherman amendment — and to that I shall confine myself in the remarks which I may address to the House—is entirely new. It is altogether without a precedent in this country. Congress has never asserted or attempted to assert, so far as I know, any such authority. That amendment claims the power in the General Government to go into the States of this Union and lay such obligations as it may please upon the municipalities, which are the creations of the States alone.” Id., at 795 (Rep. Blair), partially quoted in Monell, 436 U. S., at 673-674. See also Cong. Globe, 42d Cong., 1st Sess., 758 (1871) (Sen. Trumbull) (referring to the conference committee version of the Sherman amendment as “asserting principles never before exercised, on the part of the United States at any rate”). The strong adverse reaction to the Sherman amendment, and continued references to its complete novelty in the law of 728 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. the United States, make it difficult to entertain petitioner’s contention that the 1866 Act had already created a form of vicarious liability against municipal governments. Equally important is the basis for opposition. As we noted in Monell, a large number of those who objected to the principle of vicarious liability embodied in the Sherman amendment were of the view that Congress did not have the power to assign the duty to enforce federal law to state instrumentalities by making them liable for the constitutional violations of others. See Monell, supra, at 674-679. As Representative Farnsworth put it: “The Supreme Court of the United States has decided repeatedly that Congress can impose no duty on a State officer.” Cong. Globe, 42d Cong., 1st Sess., 799 (1871). Three decisions of this Court lent direct support to the constitutional arguments of the opponents, see Collector v. Day, 11 Wall. 113 (1871); Kentucky v. Dennison, 24 How. 66 (1861), and Prigg v. Pennsylvania, 16 Pet. 539 (1842). Day and Prigg were repeatedly cited in the House debates on the Sherman amendment. See Monell, supra, at 673-683, and n. 30. In Prigg, perhaps the most famous and most oft cited of this line of cases, Justice Story wrote for the Court that Congress could not constitutionally “insist that the states are bound to provide means to carry into effect the duties of the national government.” Prigg, supra, at 616. In Monell, we concluded that it was this constitutional objection which was the driving force behind the eventual rejection of the Sherman amendment. Monell, supra, at 676. Although the debate surrounding the constitutional principles established in Prigg, Dennison, and Day occurred in the context of the Sherman amendment and not § 1 of the 1871 Act, in Monell we found it quite inconceivable that the same legislators who opposed vicarious liability on constitutional grounds in the Sherman amendment debates would have silently adopted the same principle in §1. Because the “creation of a federal law of respondeat superior would have raised all the constitutional problems associated with JETT v. DALLAS INDEPENDENT SCHOOL DIST. 729 701 Opinion of O’Connor, J. the obligation to keep the peace” embodied in the Sherman amendment, we held that the existence of the constitutional background of Prigg, Dennison, and Day “compell[ed] the conclusion that Congress did not intend municipalities to be held liable [under § 1] unless action pursuant to official municipal policy of some nature caused a constitutional tort.” Monell, supra, at 691. Both Prigg and Dennison were on the books when the 39th Congress enacted § 1 of the 1866 Act. Supporters of the 1866 Act were clearly aware of Prigg, and cited the case for the proposition that the Federal Government could use its own instrumentalities to effectuate its laws. See, e. g., Cong. Globe, 39th Cong., 1st Sess., 1294 (1871) (Rep. Wilson). There was, however, no suggestion in the debates surrounding the 1866 Act that the statute violated Prigg’s complementary holding that federal duties could not be imposed on state instrumentalities by rendering them vicariously liability for the violations of others. Just as it affected our interpretation of § 1 of the 1871 Act in Monell, we think the complete silence on this score in the face of a constitutional background known to those who enacted the 1866 Act militates against imputing to Congress an intent to silently impose vicarious liability on municipalities under the earlier statute. Cf. Tenney v. Brandhove, 341 U. S. 367, 376 (1951). As originally enacted, the text of § 1983 referred only to the deprivation “of any rights, privileges, or immunities secured by the Constitution of the United States.” In 1874, Congress enacted the Revised Statutes of the United States. The words “and laws” were added to the remedial provision of § 1 of the 1871 Act which became Rev. Stat. § 1979. At the same time, the jurisdictional grant in § 1 of the 1871 Act was split into two different provisions, Rev. Stat. § 563(12), granting jurisdiction to the district courts of the United States to redress deprivations under color of state law of any right secured by the Constitution or “by any law of the United States,” and Rev. Stat. §629(16), granting jurisdic 730 OCTOBER TERM, 1988 Opinion of O’Connor, J. 491 U. S. tion to the old circuit courts for any action alleging deprivation under state authority of any right secured “by any law providing for equal rights.” In 1911, Congress abolished the circuit courts of the United States and the Code’s definition of the jurisdiction of the district courts was taken from Rev. Stat. § 629(16) with its narrower “providing for equal rights” language. This language is now contained in 28 U. S. C. §1343(3), the jurisdictional counterpart of §1983. Chapman, 441 U. S., at 608. There is no commentary or other information surrounding the addition of the phrase “and laws” to the remedial provisions of present day §1983. The revisers’ draft of their work, published in 1872, and the marginal notes to §§ 629(16) and 563(12), which appeared in the completed version of the Revised Statutes themselves, provide some clues as to Congress’ intent in adopting the change. The marginal note to § 629(16) states: “Suits to redress the deprivation of rights secured by the Constitution and laws to persons within jurisdiction of United States.” The note then cross cites to §1 of the 1871 Act, §§ 16 and 18 of the Enforcement Act of 1870, and § 3 of the 1866 Act. Both §§ 629(16) and 563(12) were followed by bracketed citations to Rev. Stat. § 1979, present day § 1983, and Rev. Stat. § 1977, present day § 1981. Rev. Stat. 95, 111 (1874). The revisers’ draft of 1872 contains the following notation concerning § 629(16): “It may have been the intention of Congress to provide, by this enactment [the Civil Rights Act of 1871], for all the cases of deprivations mentioned in the previous act of 1870, and thus actually to supersede the indefinite provision contained in that act. But as it might perhaps be held that only such rights as are specifically secured by the Constitution, and not every right secured by a law authorized by the Constitution, were here intended, it is deemed safer to add a reference to the civil rights act.” 1 Revision of the United States Statutes as JETT v. DALLAS INDEPENDENT SCHOOL DIST. 731 701 Opinion of the Court Drafted by the Commissioners Appointed for that Purpose 362 (1872). We have noted in the past that the addition of the phrase “and laws” to the text of what is now § 1983, although not without its ambiguities as to intended scope, was at least intended to make clear that that the guarantees contained in § 1 of the 1866 Act and § 16 of the Enforcement Act of 1870 were to be enforced against state actors through the express remedy for damages contained in § 1983. See Chapman, 441 U. S., at 617 (footnote omitted) (Section 1 of the 1871 Act “served only to ensure that an individual had a cause of action for violations of the Constitution, which in the Fourteenth Amendment embodied and extended to all individuals as against state action the substantive protections afforded by §1 of the 1866 Act”); id., at 668 (White, J., concurring in judgment). See also Maine n. Thiboutot, 448 U. S. 1, 7 (1980) (“There is no express explanation offered for the insertion of the phrase ‘and laws.’ On the one hand, a principal purpose of the added language was to ensure that federal legislation providing specifically for equality of rights would be brought within the ambit of the civil action authorized by that statute”) (some internal quotations omitted). Ill We think the history of the 1866 Act and the 1871 Act recounted above indicates that Congress intended that the explicit remedial provisions of § 1983 be controlling in the context of damages actions brought against state actors alleging violation of the rights declared in § 1981. That we have read § 1 of the 1866 Act to reach private action and have implied a damages remedy to effectuate the declaration of rights contained in that provision does not authorize us to do so in the context of the “state action” portion of § 1981, where Congress has established its own remedial scheme. In the context of the application of § 1981 and § 1982 to private actors, we “had little choice but to hold that aggrieved individuals 732 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. could enforce this prohibition, for there existed no other remedy to address such violations of the statute. ” Cannon, 441 U. S., at 728 (White, J., dissenting) (emphasis added; footnote omitted). That is manifestly not the case here, and whatever the limits of the judicial power to imply or create remedies, it has long been the law that such power should not be exercised in the face of an express decision by Congress concerning the scope of remedies available under a particular statute. See National Railroad Passenger Corporation v. National Assn, of Railroad Passengers, 414 U. S. 453, 458 (1974) (“A frequently stated principle of statutory construction is that when legislation expressly provides a particular remedy or remedies, courts should not expand the coverage of the statute to subsume other remedies”); accord, Fleischmann Corp. n. Maier Brewing Co., 386 U. S. 714, 720 (1967); Cannon, supra, at 718-724 (White, J., dissenting). Petitioner cites 42 U. S. C. §1988, and argues that that provision “compels adoption of a respondeat superior standard.” Brief for Petitioner 27. That section, as amended, provides in pertinent part: “The jurisdiction in civil. . . matters conferred on the district courts by the provisions of this [chapter and Title 18], for the protection of all persons in the United States in their civil rights, and for their vindication, shall be exercised and enforced in conformity with the laws of the United States, so far as such laws are suitable to carry the same into effect; but in all cases where they are not adapted to the object, or are deficient in the provisions necessary to furnish suitable remedies and punish offenses against law, the common law, as modified and changed by the constitution and the statutes of the State wherein the court having jurisdiction of such civil or criminal cause is held, so far as the same is not inconsistent with the Constitution and laws of the United States, shall be extended to and govern the said courts in the trial and disposition of the cause . . . .” JETT v. DALLAS INDEPENDENT SCHOOL DIST. 733 701 Opinion of the Court Far from supporting petitioner’s call for the creation or implication of a damages remedy broader than that provided by § 1983, we think the plain language of § 1988 supports the result we reach here. As we noted in Moor v. County of Alameda, 411 U. S. 693, 706 (1973), in rejecting an argument similar to petitioner’s contention here: “[Section 1988] expressly limits the authority granted federal courts to look to the common law, as modified by state law, to instances in which that law is not inconsistent with the Constitution and laws of the United States.’” Ibid. See also Johnson n. Railway Express Agency, Inc., 421 U. S. 454, 465 (1975). As we indicated in Moor, “Congress did not intend, as a matter of federal law, to impose vicarious liability on municipalities for violations of federal civil rights by their employees.” 411 U. S., at 710, n. 27. Section 1983 provides an explicit remedy in damages which, with its limitations on municipal liability, Congress thought “suitable to carry . . . into effect” the rights guaranteed by § 1981 as against state actors. Thus, if anything, § 1988 points us in the direction of the express federal damages remedy for enforcement of the rights contained in § 1981, not state common law principles. Our conclusion that the express cause of action for damages created by § 1983 constitutes the exclusive federal remedy for violation of the rights guaranteed in § 1981 by state governmental units finds support in our decision in Brown v. GSA, 425 U. S. 820 (1976). In Brown, we dealt with the interaction of § 1981 and the provisions of § 717 of Title VII of the Civil Rights Act of 1964, 42 U. S. C. §2000e-16, which proscribe discrimination in federal employment and establish an administrative and judicial enforcement scheme. The petitioner in Brown had been passed over for federal promotion on two occasions, and after the second occasion he filed a complaint with his agency alleging that he was denied promotion because of his race. The agency’s Director of Civil Rights concluded after investigation that race had not entered into the promotional process, and informed Brown by 734 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. letter of his right under § 717(c) to bring an action in federal district court within 30 days of the agency’s final decision. Forty-two days later Brown filed suit in federal court, alleging violations of both Title VII and § 1981. The lower courts dismissed Brown’s complaint as untimely under § 717(c), and this Court affirmed, holding that §717 of Title VII constituted the exclusive remedy for allegations of racial discrimination in federal employment. The Court began its analysis by noting that “Congress simply failed explicitly to describe §717’s position in the constellation of antidiscrimination law.” 425 U. S., at 825. We noted that in 1972, when Congress extended the strictures of Title VII to federal employment, the availability of an implied damages remedy under §1981 for employment discrimination was not yet clear. Id., at 828. The Court found that this perception on the part of Congress, “seems to indicate that the congressional intent in 1972 was to create an exclusive, pre-emptive administrative and judicial scheme for the redress of federal employment discrimination.” Id., at 828-829. The Court bolstered its holding by invoking the general principle that “a precisely drawn, detailed statute pre-empts more general remedies.” Id., at 834. In Brown, as here, while Congress has not definitively spoken as to the relationship of § 1981 and § 1983, there is very strong evidence that the 42d Congress which enacted the precurser of § 1983 thought that it was enacting the first, and at that time the only, federal damages remedy for the violation of federal constitutional and statutory rights by state governmental actors. The historical evidence surrounding the revision of 1874 further indicates that Congress thought that the declaration of rights in §1981 would be enforced against state actors through the remedial provisions of § 1983. That remedial scheme embodies certain limitations on the liability of local governmental entities based on federalism concerns which had very real constitutional underpinnings for the Reconstruction Congresses. As petitioner JETT v. DALLAS INDEPENDENT SCHOOL DIST. 735 701 Opinion of the Court here would have it, the careful balance drawn by the 42d Congress between local autonomy and fiscal integrity and the vindication of federal rights could be completely upset by an artifice of pleading. Since our decision in Monell, the Courts of Appeals have unanimously rejected the contention, analogous to petitioner’s argument here, that the doctrine of respondeat superior is available against a municipal entity under a Bivens-type action implied directly from the Fourteenth Amendment. See, e. g., Tarpley v. Greene, 221 U. S. App. D. C. 227, 237, n. 25, 684 F. 2d 1, 11, n. 25 (1982) (Edwards, J.) (“Because Congress has elected not to impose respondeat superior liability under § 1983, appellant invites this court to expand the remedial options under Bivens [v. Six Unknown Fed. Narcotics Agents, 403 U. S. 388 (1971)]. We can find no good logic nor sound legal basis for this view; we therefore decline the invitation”); accord, Owen v. Independence, 589 F. 2d 335, 337 (CA8 1978); Thomas v. Shipka, 818 F. 2d 496 (CA6 1987); Ellis v. Blum, 643 F. 2d 68, 85 (CA2 1981); Cale v. Covington, 586 F. 2d 311, 317 (CA4 1978); Molina n. Richardson, 578 F. 2d 846 (CA9), cert, denied, 439 U. S. 1048 (1978). Given our repeated recognition that the Fourteenth Amendment was intended in large part to embody and expand the protections of the 1866 Act as against state actors, we believe that the logic of these decisions applies with equal force to petitioner’s invitation to this Court to create a damages remedy broader than § 1983 from the declaration of rights now found in § 1981. We hold that the express “action at law” provided by § 1983 for the “deprivation of any rights, privileges, or immunities secured by the Constitution and laws,” provides the exclusive federal damages remedy for the violation of the rights guaranteed by § 1981 when the claim is pressed against a state actor. Thus to prevail on his claim for damages against the school district, petitioner must show that the violation of his “right to make contracts” protected 736 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. by § 1981 was caused by a custom or policy within the meaning of Mone ll and subsequent cases. IV The jury found that Principal Todd had violated petitioner’s rights under § 1981, the First Amendment, and the Equal Protection Clause in recommending petitioner’s removal from the athletic director and head coaching positions at South Oak. As to the liability of the DISD, the trial judge gave the jury the following instruction: “A public independent school district (such as and including the Dallas Independent School District), acts by and through its Board of Trustees and/or its delegated administrative officials (including the Superintendent and school principals), with regard to action taken against or concerning school district personnel. “A public independent school district (such as and including the Dallas Independent School District) is liable for the actions of its Board of Trustees and/or its delegated administrative officials (including the Superintendent and school principals), with regard to wrongful or unconstitutional action taken against or concerning school district personnel.” App. 31. We agree with the Court of Appeals that this instruction was manifest error. The instruction seems to rest either on the assumption that both Principal Todd and Superintendent Wright were policymakers for the school district, or that the school district is vicariously liable for any actions taken by these employees. Since we have rejected respondeat superior as a basis for holding a state actor liable under § 1983 for violation of the rights enumerated in § 1981, we refer to the principles to be applied in determining whether either Principal Todd or Superintendent Wright can be considered policymakers for the school district such that their decisions may rightly be said to represent the official policy of the DISD subjecting it to liability under § 1983. JETT v. DALLAS INDEPENDENT SCHOOL DIST. 737 701 Opinion of the Court Last Term in St. Louis v. Praprotnik, 485 U. S. 112 (1988), (plurality opinion), we attempted a clarification of tools a federal court should employ in determining where policymaking authority lies for purposes of § 1983. In Praprotnik, the plurality reaffirmed the teachings of our prior cases to the effect that “whether a particular official has ‘final policymaking authority’ is a question of state law.” Id., at 123 (emphasis in original), quoting Pembaur, 475 U. S., at 483 (plurality opinion). As with other questions of state law relevant to the application of federal law, the identification of those officials whose decisions represent the official policy of the local governmental unit is itself a legal question to be resolved by the trial judge before the case is submitted to the jury. Reviewing the relevant legal materials, including state and local positive law, as well as “‘custom or usage’ having the force of law,” Praprotnik, supra, at 124, n. 1, the trial judge must identify those officials or governmental bodies who speak with final policymaking authority for the local governmental actor concerning the action alleged to have caused the particular constitutional or statutory violation at issue. Once those officials who have the power to make official policy on a particular issue have been identified, it is for the jury to determine whether their decisions have caused the deprivation of rights at issue by policies which affirmatively command that it occur, see Monell, 436 U. S., at 661, n. 2, or by acquiescence in a longstanding practice or custom which constitutes the “standard operating procedure” of the local governmental entity. See Pembaur, supra, at 485-487 (White, J, concurring). We cannot fault the trial judge for not recognizing these principles in his instructions to the jury since this action was tried in October 1984, and the District Court did not have the benefit of our decisions in either Pembaur or Praprotnik to guide it. Similarly, the Court of Appeals issued its decision in this action before our decision in Praprotnik. Pursuant to its cross-petition in No. 88-214, the school district urges us 738 OCTOBER TERM, 1988 Opinion of Scalia, J. 491 U. S. to review Texas law and determine that neither Principal Todd nor Superintendent Wright possessed the authority to make final policy decisions concerning the transfer of school district personnel. See Brief for Respondent 6-8. Petitioner Jett seems to concede that Principal Todd did not have policymaking authority as to employee transfers, see Brief for Petitioner 30, but argues that Superintendent Wright had been delegated authority to make school district policy concerning employee transfers and that his decisions in this area were final and unreviewable. Id., at 30-32. We decline to resolve this issue on the record before us. We think the Court of Appeals, whose expertise in interpreting Texas law is greater than our own, is in a better position to determine whether Superintendent Wright possessed final policymaking authority in the area of employee transfers, and if so whether a new trial is required to determine the responsibility of the school district for the actions of Principal Todd in light of this determination. We thus affirm the judgment of the Court of Appeals to the extent it holds that the school district may not be held liable for its employees’ violation of the rights enumerated in § 1981 under a theory of respondeat superior. We remand these cases to the Court of Appeals for it to determine where final policymaking authority as to employee transfers lay in light of the principles enunciated by the plurality opinion in Praprotnik and outlined above. It is so ordered. Justice Scalia, concurring in part and concurring in the judgment. I join Parts I and IV of the Court’s opinion, and Part III except insofar as it relies upon legislative history. To hold that the more general provisions of 42 U. S. C. § 1981 establish a mode of liability for a particular category of offense by municipalities that is excluded from the closely related statute (42 U. S. C. § 1983) which deals more specifically with that precise category of offense would violate the rudimen- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 739 701 Brennan, J., dissenting tary principles of construction that the specific governs the general, and that, where text permits, statutes dealing with similar subjects should be interpreted harmoniously. Justice Brennan, with whom Justice Marshall, Justice Blackmun, and Justice Stevens join, dissenting. To anyone familiar with this and last Terms’ debate over whether Runyon v. McCrary, 427 U. S. 160 (1976), should be overruled, see Patterson v. McLean Credit Union, ante, p. 164, today’s decision can be nothing short of astonishing. After being led to believe that the hard question under 42 U. S. C. §1981—the question that prompted this Court, on its own initiative, to set Patterson for reargument, 485 U. S. 617 (1988)—was whether the statute created a cause of action relating to private conduct, today we are told that the hard question is, in fact, whether it creates such an action on the basis of governmental conduct. It is strange, indeed, simultaneously to question whether § 1981 creates a cause of action on the basis of private conduct (Patterson) and whether it creates one for governmental conduct (these cases)—and hence to raise the possibility that this landmark civil rights statute affords no civil redress at all. In granting certiorari in these cases we did not, as the plurality would have it, agree to review the question whether one may bring a suit for damages under § 1981 itself on the basis of governmental conduct. The plurality hints that petitioner Jett offered this issue for our consideration, ante, at 711 (“In essence, petitioner argues that in 1866 the 39th Congress intended to create a cause of action for damages against municipal actors and others who violated the rights now enumerated in §1981”), when in fact, it was respondent who raised this issue, and who did so for the first time in its brief on the merits in this Court.1 In six years of proceedings in 1 The plurality twice cites petitioner Jett’s opening brief, ante, at 712, as if it presents this question. Neither of the passages to which the plurality refers, however, even remotely suggests that Jett anticipated, let alone 740 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. the lower courts, including a jury trial and an appeal that produced two opinions, respondent never once suggested that Jett’s only remedy was furnished by § 1983. Petitioner was able to respond to this argument only in his reply brief in this Court. While it is true that we often affirm a judgment on a ground not relied upon by the court below, we ordinarily do so only when that ground at least was raised below. See, e. g., Heckler n. Campbell, 461 U. S. 458, 468, n. 12 (1983); Washington v. Yakima Indian Nation, 439 U. S. 463, 476, n. 20 (1979); Hankerson v. North Carolina, 432 U. S. 233, 240, n. 6 (1977); Massachusetts Mutual Life Ins. Co. v. Ludwig, 426 U. S. 479 (1976); Dandridge v. Williams, 397 U. S. 471, 475, n. 6 (1970). It is not only unfair to decide the action on this basis; it is unwise. The question is important; to resolve it on the basis of largely one-sided briefing, without the benefit of the views of the courts below, is rash. It is also unnecessary. The Court appears to decide today (though its precise holding is less than pellucid) that liability for violations by the government of § 1981 may not be predicated on a, theory of respondeat superior. The answer to that question would dispose of Jett’s contentions. In choosing to decide, as well, whether § 1983 furnishes the exclusive remedy for violations of § 1981 by the government, the Court makes many mistakes that might have been avoided by a less impetuous course. Because I would conclude that § 1981 itself affords a cause of action in damages on the basis of governmental conduct violating its terms, and because I would conclude that such an action may be predicated on a theory of respondeat superior, I dissent. I Title 42 U. S. C. § 1981, originally enacted as part of § 1 of the Civil Rights Act of 1866 (1866 Act), provides in full: raised, the argument that respondent advanced for the first time in its own brief on the merits. JETT v. DALLAS INDEPENDENT SCHOOL DIST. 741 701 Brennan, J., dissenting “All persons within the jurisdiction of the United States shall have the same right in every State and Territory to make and enforce contracts, to sue, be parties, give evidence, and to the full and equal benefit of all laws and proceedings for the security of persons and property as is enjoyed by white citizens, and shall be subject to like punishment, pains, penalties, taxes, licenses, and exactions of every kind, and to no other.” The question is whether this statute permits a cause of action in damages against those who violate its terms. The plurality approaches this issue as though it were new to us, recounting in lengthy and methodical detail the introduction, debate, passage, veto, and enactment of the 1866 Act. The story should by now be familiar to anyone with even a passing acquaintance with this statute. This is so because we have reviewed this history in the course of deciding—and reaffirming the answer to—the very question that the plurality deems so novel today. See Jones v. Alfred H. Mayer Co., 392 U. S. 409 (1968); Sullivan v. Little Hunting Park, Inc., 396 U. S. 229 (1969); Tillman n. Wheaton-Haven Recreation Assn., Inc., 410 U. S. 431 (1973); Johnson v. Railway Express Agency, Inc., 421 U. S. 454 (1975); Runyon v. McCrary, 427 U. S. 160 (1976); McDonald v. Santa Fe Trail Transportation Co., 427 U. S. 273 (1976); Delaware State College v. Ricks, 449 U. S. 250 (1980); General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S. 375 (1982); Saint Francis College v. Al-Khazraji, 481 U. S. 604 (1987); Shaare Tefila Congregation n. Cobb, 481 U. S. 615 (1987); Goodman v. Lukens Steel Co., 482 U. S. 656 (1987); Patterson v. McLean Credit Union, ante, p. 164. An essential aspect of the holding in each of these cases was the principle that a person injured by a violation of § 1 of the 1866 Act (now 42 U. S. C. §§ 1981 and 1982) may bring an action for damages under that statute against the person who violated it. 742 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. We have had good reason for concluding that § 1981 itself affords a cause of action against those who violate its terms. The statute does not explicitly furnish a cause of action for the conduct it prohibits, but this fact was of relatively little moment at the time the law was passed. During the period when § 1 of the 1866 Act was enacted, and for over 100 years thereafter, the federal courts routinely concluded that a statute setting forth substantive rights without specifying a remedy contained an implied cause of action for damages incurred in violation of the statute’s terms. See, e. g., Marbury v. Madison, 1 Cranch 137, 162-163 (1803); Kendall v. United States, 12 Pet. 524, 624 (1838); Pollard v. Bailey, 20 Wall. 520, 527 (1874); Hayes v. Michigan Central R. Co., Ill U. S. 228, 240 (1884); De Lima v. Bidwell, 182 U. S. 1, 176-177 (1901); Texas & N. O. R. Co. v. Railway Clerks, 281 U. S. 548, 569, 570 (1930); Bell v. Hood, 327 U. S. 678, 684, and n. 6 (1946); J. I. Case Co. v. Borak, 377 U. S. 426, 433 (1964). The classic statement of this principle comes from Texas & Pacific R. Co. v. Rigsby, 241 U. S. 33, 39-40 (1916), in which we observed: “A disregard of the command of the statute is a wrongful act, and where it results in damage to one of the class for whose especial benefit the statute was enacted, the right to recover the damages from the party in default is implied, according to a doctrine of the common law.” These cases fit comfortably within Rigsby’s framework. It is of small consequence, therefore, that the 39th Congress established no explicit damages remedy in § 1 of the 1866 Act.2 2 During the 1970’s, we modified our approach to determining whether a statute contains an implied cause of action, announcing the following four-part test: “First, is the plaintiff ‘one of the class for whose especial benefit the statute was enacted’—that is, does the statute create a federal right in favor of the plaintiff? Second, is there any indication of legislative intent, explicit or implicit, either to create such a remedy or to deny one? Third, is it consistent with the underlying purposes of the legislative scheme to imply such a remedy for the plaintiff? And finally, is the cause of action one traditionally relegated to state law, in an area JETT v. DALLAS INDEPENDENT SCHOOL DIST. 743 701 Brennan, J., dissenting Indeed, the debates on § 1 demonstrate that the legislators’ worry was not that their actions would do too much, but that they would do too little. In introducing the bill that became the 1866 Act, Senator Trumbull explained that the statute was necessary because “[t]here is very little importance in the general declaration of abstract truths and principles [contained in the Thirteenth Amendment] unless they can be carried into effect, unless the persons who are to be affected by them have some means of availing themselves of their benefits.” Cong. Globe, 39th Cong., 1st Sess., 474 (1866) (emphasis added). Representative Thayer of Pennsylvania echoed this theme: “When I voted for the amendment to abolish slavery ... I did not suppose that I was offering ... a mere paper guarantee.” “The bill which now engages the attention of the House has for its object to carry out and guaranty basically the concern of the States, so that it would be inappropriate to infer a cause of action based solely on federal law?” Cort v. Ash, 422 U. S. 66, 78 (1975) (citations omitted), quoting Texas & Pacific R. Co. v. Rigsby, 241 U. S., at 39. It would make no sense, however, to apply a test first enunciated in 1975 to a statute enacted in 1866. An inquiry into Congress’ actual intent must take account of the interpretive principles in place at the time. See Cannon v. University of Chicago, 441 U. S. 677, 698-699 (1979); Merrill Lynch, Pierce, Fenner & Smith, Inc. v. Curran, 456 U. S. 353, 375-378 (1982). See also Welch v. Texas Dept, of Highways and Public Transportation, 483 U. S. 468, 496 (1987) (Scalia, J., concurring) (advising against construing a statute on the basis of an interpretive principle announced after the statute was passed). Thus, I would interpret § 1981 in light of the principle described in Rigsby, rather than the one described in Cort. Application even of the test fashioned in Cort, however, would lead to the conclusion that Jett may bring a cause of action in damages against respondent under § 1981. Jett belongs to the special class of persons (those who have been discriminated against in the making of contracts) for whom the statute was created; all of the indicators of legislative intent point in the direction of an implied cause of action; such an action is completely consistent with the statute’s purposes; and, in view of the fact that this Reconstruction-era legislation was in part designed to curtail the authority of the States, it would be unreasonable to conclude that this cause of action is one relegated to state law. 744 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. the reality of that great measure. It is to give to it practical effect and force. It is to prevent that great measure from remaining a dead letter upon the constitutional page of this country.” Id., at 1151. In these circumstances, it would be unreasonable to conclude that inferring a private cause of action from § 1981 is incompatible with Congress’ intent. Yet in suggesting that § 2 of the 1866 Act demonstrates Congress’ intent that criminal penalties serve as the only remedy for violations of § 1, ante, at 715-721, this is exactly the conclusion that the plurality apparently would have us draw. Not only, however, is this argument contrary to legislative intent, but we have already squarely rejected it. In Jones n. Alfred H. Mayer Co., respondent argued that because §2 furnished criminal penalties for violations of § 1 occurring “under color of law,” § 1 could not be read to provide a civil remedy for violations of the statute by private persons. Dismissing this argument, we explained: “[Section] 1 was meant to prohibit all racially motivated deprivations of the rights enumerated in the statute, although only those deprivations perpetrated ‘under color of law’ were to be criminally punishable under §2.” 392 U. S., at 426.3 3 The Court’s heavy emphasis on §2 of the 1866 Act also ignores the fact that the modern-day descendant of § 1 of the Act, 42 U. S. C. § 1981, includes no remedy or penalty at all. Section 2 of the 1866 Act now appears at 18 U. S. C. § 242, see United States v. Classic, 313 U. S. 299, 327, n. 10 (1941), a part of the Code entirely separate from § 1981, and is applicable to provisions other than § 1981. These facts strongly argue against placing too much weight on the availability of criminal penalties in deciding whether § 1981 contains an implied cause of action. The plurality’s assertion that the 1866 Act created no original federal jurisdiction for civil actions based on the statute, see ante, at 721, is similarly unavailing. The language of § 3 easily includes original jurisdiction over such suits, and we have in fact concluded as much. See Moor v. County of Alameda, 411 U. S. 693, 704-705 (1973) (“The initial portion of §3 of the Act established federal jurisdiction to hear, among other things, civil actions brought to enforce § 1”). In addition, the plurality’s argument con JETT v. DALLAS INDEPENDENT SCHOOL DIST. 745 701 Brennan, J., dissenting The only way that the plurality can distinguish Jones, and the cases following it, from this action is to argue that our recognition of an implied cause of action against private persons did not include recognition of an action against local governments and government officials. But before today, no one had questioned that a person could sue a government official for damages due to a violation of § 1981. We have, in fact, reviewed two cases brought pursuant to § 1981 against government officials or entities, without giving the vaguest hint that the lawsuits were improperly brought. See Hurd v. Hodge, 334 U. S. 24 (1948); Takahashi v. Fish and Game Common, 334 U. S. 410 (1948). Indeed, in Jones v. Alfred H. Mayer Co., the dissenters relied on Hurd v. Hodge in arguing that § 1981 applied only to governmental conduct. 392 U. S., at 452. The lower courts have heeded well the message from our cases: they unanimously agree that suit may be brought directly under § 1981 against government officials who violate the statute’s terms. See, e. g., Metrocare v. Washington Metropolitan Area Transit Auth., 220 U. S. App. D. C. 104, 679 F. 2d 922 (1982); Springer v. Seamen, 821 F. 2d 871 (CAI 1987); Mahone v. Waddle, 564 F. 2d 1018 (CA3 1977), cert, denied, 438 U. S. 904 (1978); Jett v. Dallas Independent School Dist., 798 F. 2d 748 (CA5 1986), on motion for rehearing, 837 F. 2d 1244 (1988) (case below); Leonard v. Frankfort Electric and Water Plant Board, 752 F. 2d 189 (CA6 1985); Bell v. Milwaukee, 746 F. 2d 1205 (CA7 1984); Taylor v. Jones, 653 F. 2d 1193 (CA8 1981); Greenwood v. Ross, 778 F. 2d 448 (CA8 1985); Sethy v. Alameda County Water Dist., 545 F. 2d 1157 (CA9 1976) (en banc). Perhaps recognizing how odd it would be to argue that one may infer from § 1 of the 1866 Act a cause of action against private persons, but not one against government officials, the Court appears to claim that the 1871 Act erased whatever fuses the question of which courts (state or federal) will enforce a cause of action with whether a cause of action exists. 746 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. action against government officials previously existed under the 1866 Act. The Court explains: “That we have read § 1 of the 1866 Act to reach private action and have implied a damages remedy to effectuate the declaration of rights contained in that provision does not authorize us to do so in the context of the ‘state action’ portion of § 1981, where Congress has established its own remedial scheme. In the context of the application of § 1981 and § 1982 to private actors, we ‘had little choice but to hold that aggrieved individuals could enforce this prohibition, for there existed no other remedy to address such violations of the statute.’ . . . That is manifestly not the case here, and whatever the limits of the judicial power to imply or create remedies, it has long been the law that such power should not be exercised in the face of an express decision by Congress concerning the scope of remedies available under a particular statute.” Ante, at 731-732, quoting Cannon v. University of Chicago, 441 U. S. 677, 728 (1979) (White, J., dissenting) (emphasis in original; footnote omitted). This argument became available only after § 1983 was passed, and thus suggests that § 1983 changed the cause of action implicitly afforded by § 1981. However, not only do we generally disfavor repeals by implication, see, e. g., Morton v. Mancari, 417 U. S. 535, 549-550 (1974); Posadas v. National City Bank, 296 U. S. 497, 503 (1936); Henderson’s Tobacco, 11 Wall. 652, 656-658 (1871), but we should be particularly hostile to them when the allegedly repealing statute specifically rules them out. In this regard, § 7 of the 1871 Act is highly significant; it provided “[t]hat nothing herein contained shall be construed to supersede or repeal any former act or law except so far as the same may be repugnant thereto.” §7, 17 Stat. 15.4 4 Several amici argue that we need not conclude that § 1983 impliedly repealed the cause of action furnished by § 1981 in order to decide that JETT v. DALLAS INDEPENDENT SCHOOL DIST. 747 701 Brennan, J., dissenting The Court’s argument fails for other reasons as well. Its essential point appears to be that, in § 1983, “Congress has established its own remedial scheme” for the “‘state action’ portion of §1981.”5 Ante, at 731. For this argument, the Court may not rely, as it attempts to do, on the principle that “ ‘when legislation expressly provides a particular remedy or remedies, courts should not expand the coverage of the stat- § 1983 provides the sole remedy for violations of § 1981. See Brief for International City Management Association et al. as Amici Curiae 18-19. Their theory is that an implied cause of action did not exist when the 1871 Act was passed, and that therefore one may argue that the 1871 Act furnished the only remedy for the 1866 Act without arguing that the later statute in any way repealed the earlier one. To support their premise, they observe, first, that it was not until the 1960’s that courts recognized a private cause of action under § 1 of the 1866 Act. In doing so, they ignore our earlier cases approving actions brought directly under § 1981. See Hurd v. Hodge, 334 U. S. 24 (1948). In any event, the relevance of the date on which we expressly recognized that one could bring a suit for damages directly under § 1 escapes me; that we did so in the 1960’s does not suggest that we would not have done so had we faced the question in the 1860’s. Amici assert, in addition, that “[i]n recognizing an implied cause of action” under § 1981, we “rested in part on congressional actions that postdate the creation in 1871 of an explicit civil cause of action for violations of Section 1981.” Brief for International City Management Association et al. as Amici Curiae 19. It is true that Jones v. Alfred H. Mayer Co., 392 U. S. 409, 412, n. 1 (1968), and Sullivan v. Little Hunting Park, Inc., 396 U. S. 229, 238 (1969), cited 28 U. S. C. § 1343(4) in support of federal jurisdiction over those cases. I do not understand, however, how this shows that the 1866 Act as originally enacted did not confer federal jurisdiction over actions to recover damages for violations of the statute. Moreover, even if the 1866 Act did not confer such jurisdiction, the jurisdictional question is separate from the question whether a cause of action may be inferred from the statute. Indeed, amici appear to recognize as much when they argue that although § 1 did not establish federal jurisdiction to hear civil actions based on the statute, Congress “left the task of civil enforcement to the state courts.” Brief for International City Management Association et al. as Amici Curiae 17. I cannot imagine what “civil enforcement” amici have in mind, unless it is the civil remedy that Jett seeks. 6 The one bright spot in today’s decision is its reaffirmation of our holding in Maine v. Thiboutot, 448 U. S. 1 (1980). 748 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. ute to subsume other remedies.”’ Ante, at 732, quoting National Railroad Passenger Corporation v. National Assn, of Railroad Passengers, 414 U. S. 453, 458 (1974). That principle limits the inference of a remedy for the violation of a statute only when that same statute already sets forth specific remedies. It cannot be used to support the argument that the provision of particular remedies in § 1983 tells us whether we should infer a damages remedy for violations of § 1981. The suggestion, moreover, that today’s holding “finds support in” Brown v. GSA, 425 U. S. 820 (1976), is audacious. Ante, at 733. Section 1983—which, for example, specifies no exhaustion requirement, no damages limitation, no defenses, and no statute of limitations—can hardly be compared with §717 of the Civil Rights of 1964, at issue in Brown, with its many detailed requirements and remedies, see 425 U. S., at 829-832. Indeed, in Preiser v. Rodriguez, 411 U. S. 475, 489 (1973), we emphasized the “general” nature of § 1983 in refusing to allow former prisoners to challenge a prison’s withholding of good-time credits under § 1983 rather than under the federal habeas corpus statute, 28 U. S. C. §2254. We never before have suggested that § 1983’s remedial scheme is so thorough that it pre-empts the remedies that might otherwise be available under other statutes; indeed, all of our intimations have been to the contrary. See, e. g., Middlesex County Sewerage Authority n. National Sea Clammers Assn., 453 U. S. 1, 19-21 (1981). According to the Court, to allow an action complaining of government conduct to be brought directly under § 1981 would circumvent our holding in Monell v. New York City Dept, of Social Services, 436 U. S. 658 (1978), that liability under § 1983 may not be based on a theory of respondeat superior. Ante, at 735-736. Not only am I unconvinced that we should narrow a statute as important as § 1981 on the basis of something so vague and inconclusive as “federalism concerns which had very real constitutional underpinnings JETT v. DALLAS INDEPENDENT SCHOOL DIST. 749 701 Brennan, J., dissenting for the Reconstruction Congresses,” ante, at 734, but I am also unable to understand how MoneWs limitation on § 1983 liability begins to tell us whether the same restriction exists under § 1981, enacted five years earlier than § 1983 and covering a far narrower range of conduct. It is difficult to comprehend, in any case, why the Court is worried that construing § 1981 to create a cause of action based on governmental conduct would render local governments vicariously liable for the delicts of their employees, since it elsewhere goes to great lengths to suggest that liability under § 1981 may not be vicarious. See ante, at 718-720. The Court’s primary reason for distinguishing between private and governmental conduct under § 1981 appears to be its impression that, because private conduct is not actionable under § 1983, we “had little choice” but to hold that private individuals who violated § 1981 could be sued directly under §1981. See ante, at 731. This claim, however, suggests that whether a cause of action in damages exists under § 1981 depends on the scope of § 1983. In deciding whether a particular statute includes an implied cause of action, however, we have not in the past suggested that the answer will turn on the reach of a different statute. In National Sea Clam-mers, for example, we analyzed both the question whether the Federal Water Pollution Control Act included an implied cause of action for damages, 453 U. S., at 13-19, and the question whether an action could be brought under § 1983 for violations of that statute, id., at 19-21, thus indicating that the answer to the latter question does not tell us the answer to the former one. The Court’s approach not only departs from our prior analysis of implied causes of action, but also attributes an intent to the 39th Congress that fluctuates depending on the state of the law with regard to § 1983. On the Court’s theory, if this case had arisen during the period between our decisions in Monroe v. Pape, 365 U. S. 167 (1961), and Monell v. New York City Dept, of Social Services, supra, when we be 750 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. lieved that local governments were not “persons” within the meaning of § 1983, we would apparently have been required to decide that a cause of action could be brought against local governments and their officials directly under § 1981. The plurality, in fact, confirms this conclusion in distinguishing Hurd v. Hodge, 334 U. S. 24 (1948), solely on the ground that we decided it at a time when § 1983 did not apply to the District of Columbia. See ante, at 713. In other words, on the Court’s view, a change in the scope of § 1983 alters the reach of § 1981. I cannot endorse such a bizarre conception of congressional intent. II I thus would hold that Jett properly brought his suit against respondent directly under § 1981. It remains to consider whether that statute permits recovery against a local government body on a theory of respondeat superior. Because § 1981 does not explicitly create a cause of action in damages, we would look in vain for an express statement that the statute contemplates liability based on the doctrine of respondeat superior. In Monell n. New York City Dept, of Social Services, supra, however, our background assumption appears to have been that unless a statute subjecting institutions (such as municipalities) to liability evidences an intent not to impose liability on them based on respondeat superior, such liability will be assumed. Id., at 691. Thé absolute language of § 1981 therefore is significant: “All persons within the jurisdiction of the United States shall have the same right in every State and Territory to make and enforce contracts ... as is enjoyed by white citizens.” 42 U. S. C. § 1981. Certainly nothing in this wording refutes the argument that vicarious liability may be imposed under this law. Section 1983, in contrast, forbids a person to “subjec[t], or caus[e] to be subjected” another person to a deprivation of the rights protected by the statute. It is telling that § 1981 does not contain this explicit language of causation. In holding in Monell that liability under § 1983 may not be predi- JETT v. DALLAS INDEPENDENT SCHOOL DIST. 751 701 Brennan, J., dissenting cated on a theory of respondeat superior, we emphasized that § 1983 “plainly imposes liability on a government that, under color of some official policy, ‘causes’ an employee to violate another’s constitutional rights. . . . Indeed, the fact that Congress did specifically provide that A’s tort became B’s liability if B ‘caused’ A to subject another to a tort suggests that Congress did not intend § 1983 liability to attach where such causation was absent.” 436 U. S., at 692. The absence of this language in § 1 of the 1866 Act, now § 1981, argues against the claim that liability under this statute may not be vicarious. While it acknowledged that § 1 of the 1866 Act did not contain the “subjects, or causes to be subjected” language of § 1983, the Court of Appeals nevertheless emphasized that § 2 of the 1866 Act did contain this language. 837 F. 2d, at 1247. There is not the least inconsistency, however, in arguing that the criminal penalties under the 1866 Act may not be imposed on the basis of respondeat superior, but that the civil penalties may be. Indeed, it is no surprise that the history surrounding the enactment of §2, as the plurality stresses, ante, at 719-720, indicates that Congress envisioned criminal penalties only for those who by their own conduct violated the statute, since vicarious criminal liability would be extraordinary. The same cannot be said of vicarious civil liability. Nor does anything in the history of § 1981 cast doubt on the argument that liability under the statute may be vicarious. The Court of Appeals placed heavy reliance on Congress’ rejection of the Sherman amendment, which would have imposed a dramatic form of vicarious liability on municipalities, five years after passing the 1866 Act. 837 F. 2d, at 1246-1247. That the plurality appears to accept this argument, see ante, at 726-729, is curious, given our frequent reminder that “ ‘the views of a subsequent Congress form a hazardous basis for inferring the intent of an earlier one.’” Consumer Product Safety Common v. GTE Sylvania, Inc., 447 U. S. 752 OCTOBER TERM, 1988 Brennan, J., dissenting 491 U. S. 102, 117 (1980), quoting United States v. Price, 361 U. S. 304, 313 (1960). I do not understand how Congress’ rejection of an amendment imposing a very new kind of vicarious liability on municipalities can tell us what a different and earlier Congress intended with respect to conventional vicarious liability. According to the plurality, the history of the Sherman amendment is relevant to the interpretation of § 1981 because it reveals Congress’ impression that it had no authority to subject municipalities to the kind of liability encompassed by the amendment. See ante, at 727-729. The plurality fails to recognize, however, that the circumstances in which municipalities would be vicariously liable under the Sherman amendment are very different from those in which they would be liable under § 1981. As the plurality describes it, the Sherman amendment “provided that where injuries to person or property were caused by mob violence directed at the enjoyment or exercise of federal civil rights, ‘the county, city, or parish in which any of the said offenses shall be committed shall be liable to pay full compensation to the person or persons damnified by such offense.’” Ante, at 727, quoting Cong. Globe, 42d Cong., 1st Sess., 755 (1871). Because the threat of such liability would have forced municipalities to ensure that private citizens did not violate the rights of others, it would have run up against Justice Story’s conclusion in Prigg n. Pennsylvania, 16 Pet. 539, 616 (1842), that Congress could not “insist that the states are bound to provide means to carry into effect the duties of the national government.” To hold a local government body liable for the discriminatory cancellation of a contract entered into by that local body itself, however, is a very different matter. Even assuming that the 39th Congress had the same constitutional concerns as the 42d, therefore, those concerns cast no doubt on Congress’ authority to hold local government bodies vicariously liable under § 1 of the 1866 Act in circumstances such as those present here. JETT v. DALLAS INDEPENDENT SCHOOL DIST. 753 701 Stevens, J., dissenting I thus would conclude that liability under § 1981 may be predicated on a theory of respondeat superior. Ill No one doubts that §1983 was an unprecedented federal statute. See ante, at 723-726. The question is not whether § 1983 wrought a change in the law, but whether it did so in such a way as to withdraw a remedy that § 1 of the 1866 Act had implicitly afforded. Unlike the plurality, I would conclude that it did not. Justice Stevens, dissenting. My agreement with Justice Brennan’s dissent is buttressed by the views I expressed in Middlesex County Sewerage Authority n. National Sea Clammers Assn., 453 U. S. 1, 22 (1981) (opinion concurring in judgment in part and dissenting in part), and in Oklahoma City v. Tuttle, 471 U. S. 808, 834 (1985) (dissenting opinion). 754 OCTOBER TERM, 1988 Syllabus 491 U. S. INDEPENDENT FEDERATION OF FLIGHT ATTENDANTS V. ZIPES ET AL. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SEVENTH CIRCUIT No. 88-608. Argued April 25, 1989—Decided June 22, 1989 After protracted litigation, respondents, a class of female flight attendants alleging that Trans World Airlines’ policy of dismissing flight attendants who became mothers constituted sex discrimination in violation of Title VII of the Civil Rights Act of 1964, entered into a settlement agreement with TWA in which the airline agreed, inter alia, to credit class members with full company and union “competitive” seniority. At this point, petitioner, the collective-bargaining agent for TWA flight attendants, intervened in the lawsuit on behalf of incumbent flight attendants who would be adversely affected by the conferral of the seniority, challenging the settlement agreement on the grounds that (1) the court lacked jurisdiction to award equitable relief to one of the subclasses of respondents, and (2) the terms of the settlement would violate the existing collectivebargaining agreement. After this challenge was rejected, respondents petitioned the District Court for an award of attorney’s fees against petitioner under § 706(k) of the Act. The court awarded fees against petitioner, and the Court of Appeals affirmed. Held: District courts may award Title VII attorney’s fees against those who are not charged with Title VII violations but intervene to protect their own rights only where the intervention is frivolous, unreasonable, or without foundation. Assessing fees against blameless intervenors is not essential to § 706(k)’s central purpose of providing victims of wrongful discrimination an incentive to file suit. The prospect of uncompensated fees in litigation against such persons exists in any event, since they may choose to attack the decree collaterally instead of intervening—an undesirable result that the rule respondents urge would foster. While petitioner’s advocacy of its members’ bargained-for rights was not the specific type of conduct § 706(k) was intended to encourage, neither was it conduct that the statute aimed to deter. Pp. 758-766. 846 F. 2d 434, reversed and remanded. Scalia, J., delivered the opinion of the Court, in which Rehnquist, C. J., and White, O’Connor, and Kennedy, JJ., joined. Blackmun, J., filed an opinion concurring in the judgment, post, p. 766. Marshall, FLIGHT ATTENDANTS v. ZIPES 755 754 Opinion of the Court J., filed a dissenting opinion, in which Brennan, J., joined, post, p. 770. Stevens, J., took no part in the consideration or decision of the case. Steven A. Fehr argued the cause for petitioner. With him on the briefs were William A. Jolley and Janae L. Schaeffer. Aram A. Hartunian argued the cause for respondents. With him on the brief were Robert M. Weissbourd and Kevin M. Forde* Justice Scalia delivered the opinion of the Court. Section 706(k) of the Civil Rights Act of 1964, 42 U. S. C. §2000e-5(k), provides in relevant part that a “court, in its discretion, may allow the prevailing party, other than the [Equal Employment Opportunity] Commission or the United States, a reasonable attorney’s fee as part of the costs.” In this case we must determine under what circumstances § 706(k) permits a court to award attorney’s fees against intervenors who have not been found to have violated the Civil Rights Act or any other federal law. I This controversy began in 1970 when respondents, female flight attendants of Trans World Airlines, brought this class action against TWA claiming that its policy of terminating flight attendants who became mothers constituted sex discrimination that violated Title VII of the Civil Rights Act of 1964, 42 U. S. C. § 2000e et seq. Respondents were represented by petitioner’s predecessor union, the Air Line Stew *Briefs of amici curiae urging reversal were filed for the United States et al. by Acting Solicitor General Bryson, Acting Assistant Attorney General Turner, Deputy Solicitor General Merrill, Deputy Assistant Attorney General Clegg, Dennis J. Dimsey, and Charles A. Shanor; for the Americans United for Life Legal Defense Fund by Edward R. Grant and Clarke D. Forsythe; and for the International Association of Fire Fighters, AFL-CIO, by Thomas A. Woodley and Michael S. Wolly. Colleen K. Connell, Harvey Grossman, John A. Powell, and Steven R. Shapiro filed a brief for the American Civil Liberties Union et al. as amici curiae. 756 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ards and Stewardesses Association (ALSSA). Soon after the suit was filed, TWA abandoned the challenged policy and entered into a settlement agreement with ALSSA. This agreement was approved by the District Court, but class members dissatisfied with certain of its terms appealed. Discerning a potential conflict between ALSSA’s obligations to respondents and its obligations to incumbent flight attendants, the Court of Appeals reversed the District Court’s judgment and ordered that ALSSA be replaced as the representative of respondents’ class. Air Line Stewards and Stewardesses Assn., Local 550, TWU, AFL-CIO n. American Air Lines, Inc., 490 F. 2d 636, 643 (CA7 1973). On remand the District Court granted summary judgment to respondents on the merits. The Court of Appeals affirmed the District Court’s determination that TWA’s policy violated Title VIL In re Consolidated Pretrial Proceedings in Airline Cases, 582 F. 2d 1142, 1144 (CA7 1978). However, holding that the timely filing of charges with the Equal Employment Opportunity Commission (EEOC) is a jurisdictional prerequisite to suit in federal court, the court went on to find that over 90% of the respondents’ claims were on that ground jurisdictionally barred. Id., at 1149-1150. Both parties filed petitions for certiorari; at their request we deferred consideration of the petitions pending the outcome of ongoing settlement negotiations. Sub nom. Zipes v. Trans World Airlines, Inc., 442 U. S. 916 (1979). The parties again reached a settlement, in which TWA agreed to establish a $3 million fund to benefit all class members and to credit class members with full company and union “competitive” seniority from the date of termination.1 ’“Competitive status” seniority is used “to allocate entitlements to scarce benefits among competing employees,” Franks n. Bowman Transportation Co., 424 U. S. 747, 766 (1976), while “benefit” seniority is used “to compute noncompetitive benefits earned under the contract of employment,” ibid. FLIGHT ATTENDANTS v. ZIPES 757 754 Opinion of the Court At this point petitioner, which had replaced ALSSA as the collective-bargaining agent for TWA’s flight attendants, sought permission to intervene in the lawsuit on behalf of incumbent flight attendants not affected by the challenged TWA policy and flight attendants hired since TWA’s termination of respondents’ employment. Petitioner objected to the proposed settlement on two grounds: first, that the District Court lacked jurisdiction to approve equitable relief for the time-barred respondents (designated by the District Court as “Subclass B”); second, that reinstatement of respondents with full retroactive “competitive” seniority would violate the collective-bargaining agreement between petitioner’s members and TWA. The District Court permitted petitioner’s intervention but rejected its objections, approving the settlement in all respects. The Court of Appeals affirmed. Air Line Stewards and Stewardesses Assn., Local 550 v. Trans World Airlines, Inc., 630 F. 2d 1164 (CA7 1980). Petitioner then filed a petition for certiorari, raising essentially the same objections to the settlement agreement that it had pressed in the two lower courts. This Court granted the petition and consolidated it with the earlier petition filed by respondents, consideration of which had been deferred. In Zipes v. Trans World Airlines, Inc., 455 U. S. 385, 393 (1982), we agreed with respondents that the timeliness requirement of Title VII, 42 U. S. C. § 2000e-5(c), was not jurisdictional and thus that the District Court had jurisdiction to approve the settlement even as to members of Subclass B. We also rejected petitioner’s second challenge to the settlement agreement, concluding that reinstatement of all respondents with full competitive seniority was a remedy authorized by Title VII and appropriate in the circumstances of the case. 455 U. S., at 398-400. To come, finally, to the aspect of this lengthy litigation giving rise to the issues now before us: Respondents’ attorneys petitioned the District Court for an award of attorney’s fees against petitioner under § 706(k) of the Civil Rights Act 758 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. of 1964, 42 U. S. C. §2000e-5(k). The District Court held that “[u]nsuccessful Title VII union intervenors are, like unsuccessful Title VII defendants, consistently held responsible for attorneys’ fees,” Airline Stewards and Stewardesses Assn., Local 550, TWU, AFL-CIO n. Trans World Airlines, Inc., 640 F. Supp. 861, 867 (ND Ill. 1986), and thus awarded respondents a total of $180,915.84 in fees against petitioner-in addition to approximately $1.25 million it had earlier awarded against TWA from the settlement fund. A divided panel of the Court of Appeals affirmed. Zipes v. Trans World Airlines, Inc., 846 F. 2d 434 (1988). We granted the union’s petition for certiorari, 488 U. S. 1029 (1989). II In Alyeska Pipeline Service Co. v. Wilderness Society, 421 U. S. 240 (1975), this Court reaffirmed what has come to be known as the “American Rule.” Put simply, “[i]n the United States, the prevailing litigant is ordinarily not entitled to collect a reasonable attorneys’ fee from the loser.” Id., at 247. At issue in this case is one of the congressionally created exceptions to that rule. As part of the Civil Rights Act of 1964, Pub. L. 88-352, Tit. VII, 78 Stat. 253, Congress enacted §706(k), 42 U. S. C. §2000e-5(k), which provides that a federal district court “in its discretion, may allow the prevailing party, other than the [EEOC] or the United States, a reasonable attorney’s fee.” Although the text of the provision does not specify any limits upon the district courts’ discretion to allow or disallow fees, in a system of laws discretion is rarely without limits. In the case of §706(k) and other federal fee-shifting statutes,2 just as in the case of 2 The language of § 706(k) is substantially the same as § 204(b) of the Civil Rights Act of 1964, 42 U. S. C. § 2000a-3(b), which we interpreted in Newman n. Piggie Park Enterprises, Inc., 390 U. S. 400 (1968), and 42 U. S. C. § 1988, which we interpreted in Hensley v. Eckerhart, 461 U. S. 424 (1983). We have stated in the past that fee-shifting statutes’ similar language is “a strong indication” that they are to be interpreted alike. Northcross v. Memphis Bd. of Education, 412 U. S. 427, 428 (1973). See FLIGHT ATTENDANTS v. ZIPES 759 754 Opinion of the Court discretion regarding appropriate remedies, we have found limits in “the large objectives” of the relevant Act, Albemarle Paper Co. v. Moody, 422 U. S. 405, 416 (1975), which embrace certain “equitable considerations,” Christiansburg Garment Co. v. EEOC, 434 U. S. 412, 418 (1978). Thus, in Newman v. Piggie Park Enterprises, Inc., 390 U. S. 400, 402 (1968), we held that under § 204(b) of the Civil Rights Act of 1964, 42 U. S. C. § 2000a-3(b), a prevailing plaintiff should “ordinarily recover an attorney’s fee unless special circumstances would render such an award unjust.” We thought this constraint on district court discretion necessary to carry out Congress’ intention that individuals injured by racial discrimination act as “‘private attorney[s] general,’ vindicating a policy that Congress considered of the highest priority.” 390 U. S., at 402. See also Albemarle Paper Co., supra, at 415 (applying the Newman standard to § 706(k)); Northcross v. Memphis Bd. of Education, 412 U. S. 427, 428 (1973) (applying the Newman standard to §718 of the Emergency School Aid Act, 20 U. S. C. § 1617). Similarly, in Christiansburg Garment, supra, we held that even though the term “prevailing party” in § 706(k) does not distinguish between plaintiffs and defendants, the principle of Newman would not be applied to a prevailing defendant. Unlike the Title VII plaintiff, we reasoned, the Title VII defendant is not “‘the chosen instrument of Congress,’” 434 U. S., at 418, quoting Newman, supra, at 402; and unlike the losing defendant, the losing plaintiff is not “a violator of federal law,” 434 U. S., at 418. We also rejected, however, the losing plaintiff’s argument that sound exercise of § 706(k) discretion would remand the prevailing defendant to the American Rule, providing attorney’s fees only if the plaintiff’s suit was brought in bad faith. Such an unequal disposition, also Hanrahan v. Hampton, 446 U. S. 754, 758, n. 4 (1980) (noting that § 1988 was patterned on § 204(b) and § 706(k)); Hensley, supra, at 433, n. 7 (noting that the standards set forth in the opinion apply to all fee-shifting statutes with “prevailing party” language). 760 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. we thought, “giving the private plaintiff substantial incentives to sue, while foreclosing to the defendant the possibility of recovering his expenses in resisting even a groundless action unless he can show that it was brought in bad faith,” would so “distort” the “fair adversary process” that Congress could not lightly be assumed to have intended it. Id., at 419. We thus concluded that the prevailing defendant could be awarded fees under § 706(k) against the plaintiff whose suit was brought in good faith, but only “upon a finding that the plaintiff’s action was frivolous, unreasonable, or without foundation,” id., at 421. The dissent contends that construing § 706(k) in such fashion as to allow competing rights and equities to be taken into account “ignore[s] its express language,” post, at 771, in two ways: first, because “the only party mentioned in § 706(k) is ‘the prevailing party,’” and thus, “when a district court decides whether to award fees, it must be guided first and foremost by the interests of the prevailing party,” ibid. This seems to us something less than an “express language” argument—and also a non sequitur. To say that only the prevailing party gets fees is not to say that the prevailing party’s interests are always first and foremost in determining whether he gets them. In any case, as discussed above, we decided long ago that in some circumstances the interests of the losing party trump those of the prevailing party under §706(k), so that the latter cannot obtain fees. See Christiansburg Garment, supra. The second respect in which the dissent contends we ignore the “express language” of the statute is that we fail to give effect to its “hostility to categorical rules for the award of attorney’s fees,” post, at 771, supposedly enshrined in the language that the court “in its discretion, may allow” (emphasis added) a reasonable attorney’s fee. We have already described how the law in general, and the law applied to §706(k) in particular, does not interpret a grant of discretion to eliminate all “categorical FLIGHT ATTENDANTS v. ZIPES 761 754 Opinion of the Court rules.”3 In Newman, supra, at 402, we held that in absence of special circumstances a district court not merely “may” but must award fees to the prevailing plaintiff; and in Christiansburg Garment, supra, at 421, we held that unless the plaintiff’s action is frivolous a district court cannot award fees to the prevailing Title VII defendant. The prescriptions in those cases are no less “categorical” than the rule we set forth today. Proceeding, then, to interpret the statute in light of the competing equities that Congress normally takes into account, we conclude that district courts should similarly award Title VII attorney’s fees against losing intervenors only where the intervenors’ action was frivolous, unreasonable, or without foundation. It is of course true that the central purpose of §706(k) is to vindicate the national policy against wrongful discrimination by encouraging victims to make the wrongdoers pay at law—assuring that the incentive to such suits will not be reduced by the prospect of attorney’s fees that consume the recovery. See Newman, supra, at 401-402. Assessing fees against blameless intervenors, however, is not essential to that purpose. In every lawsuit in which there is a prevailing Title VII plaintiff there will also be a losing defendant who has committed a legal wrong. That defendant will, under Newman, be liable for all of the fees expended by the plaintiff in litigating the claim against him, and that liability alone creates a substantial added incentive for victims of Title VII violations to sue. In the present case, for example, TWA paid over $1.25 million in fees to respondents’ attorneys. Respondents argue that this incentive will be reduced by the potential presence of inter 3 The dissent, post, at 772, n. 1, distorts our holding in United States v. Monsanto, ante, at 613, by describing it as “conclud[ing] that statutory construction that transforms the word ‘may’ into the words ‘may not’ . . . impermissibly frustrates legislative intent.” What we plainly said there was that “may” cannot be transformed into “may not” in such fashion as to frustrate the legislative intent. 762 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. venors whose claims the plaintiff must litigate without prospect of fee compensation. It is not clear to us that that consequence will follow. Our decision in Martin v. Wilks, 490 U. S. 755, 762-763 (1989), establishes that a party affected by the decree in a Title VII case need not intervene but may attack the decree collaterally—in which suit the original Title VII plaintiff defending the decree would have no basis for claiming attorney’s fees. Thus, even if we held that fees could routinely be recovered against losing intervenors, Title VII plaintiffs would still face the prospect of litigation without compensation for attorney’s fees before the fruits of their victory can be secure. But even if the inability generally to recover fees against intervenors did create some marginal disincentive against Title VII suits, we would still have to weigh that against other considerations, as we did in Christiansburg Garment. Foremost among these is the fact that, in contrast to losing Title VII defendants who are held presumptively liable for attorney’s fees, losing intervenors like petitioner have not been found to have violated anyone’s civil rights. See Christiansburg Garment, 434 U. S., at 418. In this case, for example, petitioner became a party to the lawsuit not because it bore any responsibility for the practice alleged to have violated Title VII, but because it sought to protect the bargained-for seniority rights of its employees. Awarding attorney’s fees against such an intervenor would further neither the general policy that wrongdoers make whole those whom they have injured nor Title Vil’s aim of deterring employers from engaging in discriminatory practices. Our cases have emphasized the crucial connection between liability for violation of federal law and liability for attorney’s fees under federal fee-shifting statutes. In Kentucky v. Graham, 473 U. S. 159 (1985), the plaintiffs had brought suit under 42 U. S. C. § 1983 against police officers in their individual capacities, alleging that the officers had violated their constitutional rights. After settling with the officers, FLIGHT ATTENDANTS v. ZIPES 763 754 Opinion of the Court they sought attorney’s fees from the officers’ employer, the Commonwealth of Kentucky, under 42 U. S. C. § 1988. In rejecting that claim, we stated: “Section 1988 does not in so many words define the parties who must bear these costs. Nonetheless, it is clear that the logical place to look for recovery of fees is to the losing party—the party legally responsible for relief on the merits. That is the party who must pay the costs of litigation . . . and it is clearly the party who should also bear fee liability under § 1988.” 473 U. S., at 164. See also id., at 165 (“[Liability on the merits and responsibility for fees go hand in hand”); id., at 168 (“[F]ee liability runs with merits liability”); ibid. (“Section 1988 simply does not create fee liability where merits liability is nonexistent”); id., at 171 (“[F]ee and merits liability run together”). Cf. Supreme Court of Virginia n. Consumers Union of United States, Inc., 446 U. S. 719, 738 (1980) (holding that § 1988 fees were not recoverable against defendants immune from merits liability). We have also distinguished between wrongdoers and the blameless in the related area of constraints upon district courts’ discretion to fashion Title VII remedies. See, e. g., Ford Motor Co. v. EEOC, 458 U. S. 219, 239-240 (1982); General Building Contractors Assn., Inc. v. Pennsylvania, 458 U. S. 375, 399-400 (1982). While innocent intervenors raising non-Title VII claims are not, like Title VII plaintiffs, “the chosen instrument[s] of Congress,” Christiansburg Garment, supra, at 418, neither are they disfavored participants in Title VII proceedings.4 4 The dissent repeatedly implies that intervenors are no more than intermeddlers who get in the way of tidy settlement agreements between Title VII plaintiffs and wrongdoers. See post, at 770, 774, 775, 777, 778, 779. That characterization might be understandable if our opinion addressed intervenors who are not themselves affected by the outcome of the lawsuit; but it does not. See infra, at 765. What is at issue here is only the liability of intervenors who enter lawsuits to defend their own constitutional or statutory rights. It seems to us that the dissent dismisses out of 764 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. An intervenor of the sort before us here is particularly welcome, since we have stressed the necessity of protecting, in Title VII litigation, “the legitimate expectations of . . . employees innocent of any wrongdoing,” Teamsters v. United States, 431 U. S. 324, 372 (1977). Even less with regard to an innocent intervenor than with regard to an allegedly lawbreaking defendant would Congress have wished to “distort” the adversary process, see Christiansburg Garment, supra, at 419, by giving the plaintiff a disproportionate advantage with regard to fee entitlement. Moreover, establishing such one-way fee liability against intervenors would foster piecemeal litigation of complex civil rights controversies—a result that is strongly disfavored. See Martin v. Wilks, supra, at 768. Adopting the regime proposed by respondents — that those who intervene in a Title VII suit are presumptively liable for fees, while those who take the alternative course of becoming plaintiffs in independent lawsuits attacking provisions of the decree are presumptively shielded from liability—would encourage interested parties to await the entry of judgment and collaterally attack remedial schemes. This hand the legitimate claims of these people, not because they are intermeddlers, but rather because the dissenters have established a judge-made ranking of rights, authorizing Title VII claims to prevail over all others. That is the essential difference between us. Whereas we think that the fee-award provision is subject to “the competing equities that Congress normally takes into account,” supra, at 761, the dissent believes that we “must be guided first and foremost by the interests of the prevailing party” (so long as that is the Title VII plaintiff and not the defendant, see Christiansburg Garment Co. v. EEOC, 434 U. S. 412 (1978)), post, at 771, and that the only criterion of our decision is that it “respect the objectives of Title VII,” post, at 772. Those objectives must of course be respected. But nothing in the statute gives them hegemony over all the other rights and equities that exist in the world. Here as elsewhere, the judicial role is to reconcile competing rights that Congress has established and competing interests that it normally takes into account. See, e. g., Ford Motor Co. v. EEOC, 458 U. S. 219, 239-240 (1982). When Congress wishes Title VII rights to sweep away all others it will say so. FLIGHT ATTENDANTS v. ZIPES 765 754 Opinion of the Court would serve the interests of no one: not plaintiffs, not defendants, not intervenors. Intervention that is in good faith is by definition not a means of prolonging litigation, but rather of protecting legal rights—ranging from contract-based rights, see, e. g., Richardson v. Alaska Airlines, Inc., 750 F. 2d 763 (CA9 1984) (collective-bargaining agreement), to statutory rights, see, e. g., Prate v. Freedman, 583 F. 2d 42 (CA2 1978) (Title VII), to constitutional rights, see, e. g., Reeves n. Harrell, 791 F. 2d 1481 (CA11 1986) (Equal Protection Clause); Grano v. Barry, 251 U. S. App. D. C. 289, 783 F. 2d 1104 (1986) (Takings Clause)—which are entitled to no less respect than the rights asserted by plaintiffs in the subject suit. In this case petitioner intervened to assert the collectively bargained contract rights of its incumbent employees, rights that neither respondents nor TWA had any interest in protecting in their settlement agreement. Just this Term we recognized that competitive seniority rights—the specific interests asserted by petitioner—are among the most important ingredients in flight attendants’ collective-bargaining agreements. See Trans World Airlines, Inc. v. Flight Attendants, 489 U. S. 426, 428-430 (1989). While a labor union’s good-faith advocacy of its members’ vital interests was not the specific type of conduct § 706(k) was intended to encourage, it is certainly not conduct that the statute aimed to deter. Of course, an intervenor may sometimes raise an argument that brings into question not merely the appropriateness of the remedy but the plaintiff’s very entitlement to relief. Here, for example, petitioner advanced one argument that would have prevented the District Court’s approval of any relief for Subclass B respondents. But that an intervenor can advance the same argument as a defendant does not mean that the two must be treated alike for purposes of fee assessments. The central fact remains that petitioner litigated (and lost) not to avoid liability for violation of the law 766 OCTOBER TERM, 1988 Blackmun, J., concurring in judgment 491 U. S. but to prevent TWA’s bargaining away of its members’ seniority rights in order to settle with respondents. It was entitled, like any litigant, to pursue that legitimate end through arguments that go to the merits no less than through arguments that go only to the scope of the relief. It would hardly serve the congressional policy in favor of “vigorous” adversary proceedings, Christiansburg Garment, 434 U. S., at 419, to require intervenors to disguise or avoid their strongest arguments in order to escape liability for attorney’s fees. Moreover, it is often quite difficult to separate arguments directed to the appropriate remedy from arguments directed to the existence or extent of past violations, so that making fees turn upon that distinction would violate our admonition that “a request for attorney’s fees should not result in a second major litigation.” Hensley v. Eckerhart, 461 U. S. 424, 437 (1983). * * * Because the courts below incorrectly presumed that petitioner was liable for attorney’s fees to respondents, and accordingly made no inquiry as to whether petitioner’s intervention was frivolous, unreasonable, or without foundation, the judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Justice Stevens took no part in the consideration or decision of this case. Justice Blackmun, concurring in the judgment. For me, the Court’s approach to the difficult problem of an intervenor’s fee liability is not fully satisfying. The Court notes that an intervenor is not like a culpable Title VII defendant because it is not a wrongdoer, and holds that, as a result, the rule that a defendant is presumptively liable for fees if the Title VII plaintiff prevails cannot be applied to an intervenor. The Court also acknowledges that “innocent in- FLIGHT ATTENDANTS v. ZIPES 767 754 Blackmun, J., concurring in judgment tervenors raising non-Title VII claims” are not like Title VII plaintiffs, because they are not “‘the chosen instrument[s] of Congress’ ” for enforcing the antidiscrimination policies of Title VIL Ante, at 763, quoting Christiansburg Garment Co. v. EEOC, 434 U. S. 412, 418 (1978). I agree with each of these observations. Despite the fact, however, that, from Congress’ point of view, an intervenor is not like a Title VII plaintiff, the Court today fashions a fee-shifting rule that essentially ignores this difference. The result is presumptively to place the additional cost of litigating third-party rights on the prevailing Title VII plaintiff, whom Congress has assumed lacks the resources to bear them. This result is neither fair nor necessary. It seems to me that the first step toward solving the problem of intervenor fee liability is to recognize that it is the Title VII wrongdoer, and not the Title VII plaintiff, whose conduct has made it necessary to unsettle the expectations of a third party who itself is not responsible for the Title VII plaintiff’s injuries. The Court states that the “defendant will, under Newman, be liable for all of the fees expended by the plaintiff in litigating the claim against him,” ante, at 761 (emphasis added)— and thereby tacitly assumes that the defendant’s fee liability goes no further. I see no basis for that assumption. Addressing and adjusting the rights of a third party are parts of the social cost of remedying a Title VII violation. That cost, as well as the cost to the plaintiff of vindicating his or her own rights, would not have existed but for the conduct of the Title VII defendant. I see nothing in the language of the statute or in our precedents to foreclose a prevailing plaintiff from turning to the Title VII defendant for reimbursement of all the costs of obtaining a remedy, including the costs of assuring that third-party interests are dealt with fairly. Thus, where an intervenor enters the case to defend third-party interests and the plaintiff prevails, the costs of the intervention, in my view, should presumptively be borne by 768 OCTOBER TERM, 1988 Blackmun, J., concurring in judgment 491 U. S. the defendant. Such a rule would safeguard the plaintiff’s incentive to enforce Title VII by assuring that the costs of defending against an unsuccessful intervention will be recouped, and would give a plaintiff added incentive to invite intervention by interested third parties, whose concerns can be addressed most fairly and efficiently in the original Title VII proceeding. Cf. Martin v. Wilks, 490 U. S. 755 (1989). This is not to say that an intervenor may never be held liable for fees. The Court in Christiansburg held that § 706(k) of Title VII must be interpreted as a full-scale departure from the American Rule, in order to assure that no party to a Title VII case has an incentive to maintain a position that is taken in good faith but is nonetheless “groundless.” 434 U. S., at 419. That rule should apply to an intervenor, as well as to a plaintiff. But the adjustment that should take place is one between the Title VII defendant, whose conduct implicated third-party interests, and the intervenor who seeks to protect those interests. In my view, liability for fees should shift from the defendant to the intervenor if the intervenor’s position was “frivolous, unreasonable, or without foundation.” Id., at 421. There is no reason why the defendant should be made to pay the cost of frivolous assertions of third-party rights, or that an intervenor should be without incentive to exercise some self-restraint in the position it takes in a Title VII case. The only potential “disadvantage” to the rule I would adopt is that it would diminish, to some extent, the gains a Title VII defendant could reap from settlement: under my rule, the defendant’s fee liability would not cease with its decision to settle the case. The result will not be to deter all settlements, however: it will deter only those that unfairly impose disproportionate costs on third parties. An examination of the considerations that enter into a settlement decision explains why this is so. As a general rule, a defendant framing a settlement offer considers his remedial FLIGHT ATTENDANTS v. ZIPES 769 754 Blackmun, J., concurring in judgment exposure in the event the plaintiff prevails at trial, and discounts it by the likelihood that the plaintiff will not prevail. For those aspects of the settlement package that come at the employer’s expense—e. g., backpay—the employer’s settlement offer likely will reflect these considerations. But the Title VII defendant has little incentive to make a similar calculation for elements of the settlement package that burden only third parties—e. g., competitive seniority. Indeed, a defendant has every reason to impose a disproportionate share of the remedial costs of settling a case on third parties, whose interests are not represented in the settlement negotiations. For this reason, a settlement that reasonably serves the employer’s needs might well fall short of reasonableness from the point of view of a rational third party. Under the rule I would adopt, a district court would be permitted to consider the settlement agreement’s fairness to third parties as a factor in determining whether the intervenor’s opposition to the settlement was reasonable. The intervenor therefore would have the incentive to acquiesce in a settlement proposal that fairly assesses the likely result at trial, because intervention to oppose a settlement which is fair across the board will expose the intervenor to fee liability. And the defendant would have the incentive to consider third-party interests in its settlement proposal, lest it be assessed attorney’s fees when third parties reasonably intervene to object to a settlement that is unfair from their point of view. This would be a desirable result, not a reason to reject the fee-shifting rule I propose. Accordingly, I concur in the judgment of the Court to reverse the judgment of the Court of Appeals and to remand the case for further proceedings. But I do not join the Court’s opinion insofar as it requires a prevailing plaintiff to bear the cost of intervention-related attorney’s fees unless the intervenor’s position is found to be “frivolous, unreasonable, or without foundation.” That result needlessly burdens the Title VII plaintiff with litigation costs imposed on 770 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. the plaintiff by the unlawful conduct of the Title VII defendant, and compromises Congress’ interest in furthering private enforcement of Title VIL On remand of this case, the court, if it followed my view, first would determine whether the union’s position in opposition to the settlement was frivolous or unreasonable. If the court so concluded, the union would be liable for fees. But if the court concluded that the union’s position had sufficient merit to bar the assessment of fees against it, the court would go on to consider whether, in the posture of the case, the plaintiffs may recover their attorney’s fees from TWA. In particular, the court would determine whether the plaintiffs have preserved a claim for additional fees against TWA and, if so, whether the provisions of the settlement agreement that governs TWA’s fee liability foreclose any additional fee award. If the claim has been preserved and additional fees may be recovered from TWA consistent with the settlement agreement, the plaintiffs would be entitled to recover from TWA the attorney’s fees due to the intervention. Justice Marshall, with whom Justice Brennan joins, dissenting. Nearly two decades ago, female flight attendants of Trans World Airlines (TWA) brought a class action challenging the airline’s practice of terminating all female flight attendants who became mothers, while retaining their male counterparts who became fathers. After almost 10 years of litigation, the parties reached a comprehensive settlement. At this point, petitioner Independent Federation of Flight Attendants (IFFA) intervened to oppose the settlement on two grounds: first, that untimely filing of charges by certain plaintiffs deprived the District Court of jurisdiction to approve their claims for equitable relief; and second, that reinstatement of the plaintiffs with full retroactive “competitive” seniority would violate the collective-bargaining agreement between TWA and IFFA’s incumbent members. The plaintiffs spent nearly three years and $200,000 successfully de- FLIGHT ATTENDANTS v. ZIPES 771 754 Marshall, J., dissenting fending the settlement against the intervenor’s claims in the District Court, the Court of Appeals for the Seventh Circuit, and this Court. See Zipes v. Trans World Airlines, Inc., 455 U. S. 385, 398-399 (1982). Despite the fact that the plaintiffs prevailed against IFF A, and that IFF A was solely responsible for forcing them to invest additional time and money to defend the agreement and thereby vindicate their civil rights, the majority holds that the District Court had practically no discretion under § 706(k) of the Civil Rights Act of 1964 (Act), 42 U. S. C. §2000e-5(k), to award the plaintiffs attorney’s fees from IFFA. Because this result ignores both the language of § 706(k) and the objectives of Title VII of the Act, I dissent. The majority begins its opinion by quoting §706(k), but then proceeds to ignore its express language. Section 706(k) states that a “court, in its discretion, may allow the prevailing party, other than the Commission or the United States, a reasonable attorney’s fee.” While §706(k) provides no detailed rules as to when attorney’s fees should be awarded, its terms nonetheless make two things clear. First, the only party mentioned in § 706(k) is “the prevailing party.” Thus, when a district court decides whether to award fees, it must be guided first and foremost by the interests of the prevailing party. See Texas State Teachers Assn. v. Garland Independent School Dist., 489 U. S. 782, 790 (1989) (“Congress clearly contemplated that. . . fee awards would be available where a party has prevailed on an important matter in the course of the litigation . . .”) (internal quotations omitted); Charles v. Daley, 846 F. 2d 1057, 1064 (CA7 1988) (civil rights fee-shifting statutes “fashion the parameters of eligibility for fee awards, rather than ... fix with precision the bounds of liability for such awards”) (emphasis in original). Second, § 706(k) contains “permissive and discretionary language,” Christiansburg Garment Co. v. EEOC, 434 U. S. 412, 418 (1978), reflecting Congress’ hostility to categorical rules for the award of attorney’s fees. 772 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. The majority overlooks both of these textual directives. After Zipes v. Trans World Airlines, supra, there can be little doubt that the plaintiffs prevailed in the face of IFFA’s challenges to the settlement agreement. Disregarding § 706(k)’s focus on the success of the plaintiffs, however, the majority decrees that the propriety of a fee award turns instead on the motivations and claims of the losing party, in this case an intervenor. To make matters worse, the majority also ignores Congress’ explicit conferral of discretion on the district courts, and instead establishes an absolute rule that, in all circumstances, a court must treat an intervenor like a plaintiff for fee liability purposes.1 Section 706(k), of course, does not invest district courts with unfettered discretion to award attorney’s fees to prevailing parties. But this does not mean that this Court has a free hand to fashion limitations. Rather, the principles we articulate to guide a district court’s discretion in awarding attorney’s fees in civil rights cases should respect the objectives of Title VII. See Albemarle Paper Co. n. Moody, 422 U. S. 405, 416-417 (1975). Regrettably, the limitations formulated by the majority do nothing of the kind. The Civil Rights Act of 1964 embodies a national commitment to eradicate discrimination. Congress intended not only “to make the wrongdoers pay at law,” ante, at 761, but more broadly to make victims of discrimination whole. See Albemarle Paper Co., supra, at 418. Given the scarcity of public resources available for enforcement, individuals injured by discrimination serve as “the chosen instrument of Congress to vindicate ‘a policy that Congress considered of the highest priority.’” Christiansburg Garment, supra, at 418, quoting Newman v. Piggie Park Enterprises, Inc., 390 ‘Just today, in United States v. Monsanto, ante, at 613, the Court concluded that statutory construction that transforms the word “may” into the words “may not,” thereby substituting a command for congressionally mandated discretion, impermissibly frustrates legislative intent. I see no reason to depart from this commonsense rule in the civil rights context. FLIGHT ATTENDANTS v. ZIPES 773 754 Marshall, J., dissenting U. S. 400, 402 (1968). Congress recognized that victims of discrimination often lack the resources to retain paid counsel, and frequently are unable to attract lawyers on a contingency basis because many victims seek injunctive relief rather than pecuniary damages. See, e. g., S. Rep. No. 94-1011, pp. 1-4 (1976); H. R. Rep. No. 94-1558, pp. 1-3 (1976); Note, Promoting the Vindication of Civil Rights Through the Attorney’s Fees Awards Act, 80 Colum. L. Rev. 346, 350-351 (1980). It therefore enacted § 706(k) to ensure that victims of discrimination could obtain lawyers to bring suits necessary to vindicate their rights and to provide victorious plaintiffs with fully compensatory attorney’s fees. Newman, supra, at 402. Nothing in the legislative history indicates that Congress intended to limit the types of losing parties against whom attorney’s fees could be awarded. Indeed, given Congress’ broad remedial goals, the majority errs in casually presuming that such limits exist.2 2 Congress fully expected fee awards under civil rights fee-shifting statutes to turn on whether the party seeking civil rights relief prevailed and not on formal labels such as plaintiff, defendant, or intervenor. For example, when Congress adopted 42 U. S. C. § 1988 in response to Alyeska Pipeline Service Co. v. Wilderness Society, 421 U. S. 240 (1975), the Senate Committee on the Judiciary noted: “In the large majority of cases the party or parties seeking to enforce [civil] rights will be the plaintiffs and/or plaintiff-intervenors. However, in the procedural posture of some cases, the parties seeking to enforce such rights may be the defendants and/or defendant-intervenors.” S. Rep. No. 94-1011, p. 4, n. 4 (1976). The same Committee cited approvingly a pre-Alyeska decision in which the prevailing party was awarded attorney’s fees against intervenors. See id., at 4, n. 3, citing Sims v. Amos, 340 F. Supp. 691 (MD Ala.) (per curiam) (three-judge court), aff’d, 409 U. S. 942 (1972) (affirming fee award against state legislators who intervened to defend legislative apportionment scheme). Alyeska itself, which barred an attorney’s fees award against an intervenor in an action brought pursuant to the Mineral Lands Leasing Act of 1920, 30 U. S. C. § 185, did not reverse the fee award because the losing party was an intervenor, but only because there was no statute, such as § 706(k), authorizing an exception to the American Rule. See 421 U. S., at 263, 267-268. 774 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. The majority’s contention that its ruling will not discourage private plaintiffs from bringing civil rights suits, or that it will only “create some marginal disincentive,” ante, at 762, is hard to take seriously. The costs to plaintiffs are no less real when the person causing the financial expenditures is an intervenor than when he is a defendant. To vindicate their civil rights, many plaintiffs must respond to, and defeat, claims raised by intervenors in support of the challenged practice or in opposition to the proposed remedy. Such intervenors force victims of discrimination to spend additional scarce resources to obtain relief, often long after the named defendant has conceded a violation of the Act. See, e. g., Geier v. Richardson, 871 F. 2d 1310, 1313 (CA6 1989) (United States, as intervenor, challenged settlement reached after 15 years of litigation); Akron Center for Reproductive Health v. Akron, 604 F. Supp. 1268, 1272 (ND Ohio 1984) (individuals who intervened solely to defend an abortion ordinance that did not implicate their conduct filed 40 documents, at least 14 of which required independent responses from the plaintiffs); Vulcan Society of Westchester Co., Inc. v. Fire Dept, of White Plains, 533 F. Supp. 1054, 1062 (SDNY 1982) (union intervened and moved to dissolve a temporary restraining order granted to the plaintiffs). By denying plaintiffs the opportunity to be compensated for those expenditures simply because the losing party was an intervenor rather than a named defendant, the majority breaks the congressional promise that prevailing plaintiffs will be made whole for efforts to vindicate their civil rights. Cf. Sullivan v. Hudson, 490 U. S. 877 (1989) (right to fee awards for prevailing civil rights plaintiffs extends to work performed in administrative as well as judicial proceedings).3 3 While the majority pays lipservice to the objectives of Title VII, it is guilty of establishing its own “judge-made ranking of rights.” Ante, at 764, n. 4. By elevating intervenors to the same plane as Title VII plaintiffs, the majority undermines Congress’ determination that Title VII FLIGHT ATTENDANTS v. ZIPES 775 754 Marshall, J., dissenting The majority further states that a defendant’s liability for “all of the fees expended by the plaintiff in litigating the claim against him, . . . alone creates a substantial added incentive for victims of Title VII violations to sue.” Ante, at 761 (emphasis added). The majority apparently believes that the typical victim injured by discrimination will have available discretionary income, possibly running into the hundreds of thousands of dollars, to spend to counter intervenors’ claims. If the typical victim had access to such financial resources, however, there would have been less need in the first place for civil rights fee-shifting statutes. Or perhaps the Court is assuming that the initial fee award in this case of over $1.25 million is so large that it should cover whatever costs the plaintiffs have incurred, including those costs incurred in responding to the intervenor’s claims. But this ignores the fact that the District Court concluded that $1.25 million was a reasonable attorney’s fee only for the hours the plaintiffs’ attorneys spent reaching the settlement with the defendant. The notion that this award can also compensate the plaintiffs for the expenses of subsequent litigation against the intervenor presumes that the initial fee award was not reasonable, but rather far in excess of the amount warranted. To justify a result contrary to the language of § 706(k) and the objectives of Title VII, the Court offers two propositions: first, that liability on the merits is a prerequisite for liability for fees; and second, that the interests of intervenors are as important as the civil rights concerns of plaintiffs. Neither assertion withstands scrutiny. Nor does either explain why the majority has adopted a blanket rule that all intervenors must be treated like plaintiffs for purposes of fee liability. This Court has never held that one is immune from liability for attorney’s fees absent a finding of liability on the merits. On the contrary, we have expressly recognized that a district court’s authority to award fees in civil rights cases does not plaintiffs alone are “the chosen instruments” for vindicating the national policy against discrimination. 776 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. require a finding that any party caused a civil rights injury. See Maher v. Gagne, 448 U. S. 122, 129 (1980) (“Nothing in the language of § 1988 conditions the District Court’s power to award fees on ... a judicial determination that the plaintiff’s rights have been violated”).4 The majority’s alternative suggestion stems from a misreading of several of this Court’s precedents. In Christiansburg Garment, for example, we held that prevailing defendants could recover fees from civil rights plaintiffs only if the suit was “frivolous, unreasonable, or without foundation.” 434 U. S., at 421. We explained that the two “equitable considerations” that warrant an award of attorney’s fees when a plaintiff prevails—compensating the party who is the chosen instrument for enforcing civil rights laws, and assessing fees “against a violator of federal law”—are “wholly absent” when a defendant prevails against a plaintiff. Id., at 418. The majority reads Christiansburg Garment as mandating that both considerations be satisfied before attorney’s fees can be imposed. But our holding that a plaintiff could be assessed attorney’s fees in certain circumstances plainly demonstrates that liability on the merits is not always a precondition for liability for fees.5 4 By contrast, several fee-shifting statutes outside the civil rights field specify that attorney’s fees are available only upon a showing of injury in violation of the underlying statute. See, e. g., Bank Holding Company Act Amendments of 1970, 12 U. S. C. § 1975; Clayton Act, 15 U. S. C. §15. 5 Similarly, if liability for attorney’s fees is premised on liability on the merits, then it is hard to understand why a court could ever impose attorney’s fees against an intervenor. Yet, the majority applies the Christiansburg Garment rule to intervenors so that a district court may award attorney’s fees pursuant to § 706(k) “where the intervenors’ action was frivolous, unreasonable, or without foundation.” Ante, at 761. In permitting fee awards against intervenors under these limited circumstances, the majority thus implicitly recognizes that a district court should be able to impose a fee award solely on the ground that the intervenor’s claims did not warrant the added length and cost of the litigation. FLIGHT ATTENDANTS v. ZIPES 777 754 Marshall, J., dissenting Kentucky v. Graham, 473 U. S. 159 (1985), likewise provides no support for the majority’s assertion that civil rights wrongdoers are the only persons liable for fees. The plaintiffs in Graham sued employees of the Commonwealth of Kentucky in their personal capacity for civil rights violations, and named the Commonwealth for attorney’s fees that might result. Relying on the Eleventh Amendment, the District Court dismissed the Commonwealth as a party. The Commonwealth then refused to defend the individual defendants or to pay for their litigation expenses. Although the plaintiffs ultimately prevailed against the individual defendants, we concluded that § 1988 did not authorize a fee award against the Commonwealth because it “ha[d] not been prevailed against.” Id., at 165. We thus refused to impose vicarious liability for attorney’s fees on a nonparty who neither actively participated nor intervened in the litigation. That is hardly the situation in this case. The plaintiffs here prevailed against a party who voluntarily intervened in the litigation and who actively opposed the settlement for several years after the defendant had agreed to liability.6 Nor does this Court’s decision in Supreme Court of Virginia v. Consumers Union of United States, Inc., 446 U. S. 719 (1980), support the proposition that liability on the merits 6 The majority asserts that permitting fee assessments against intervenors will cause these parties to refrain from intervening in favor of attacking consent decrees through collateral actions in which the original Title VII plaintiffs will have no basis for claiming attorney’s fees. This argument is specious. First, Martin v. Wilks, 490 U. S. 755 (1989), on which the majority relies, is silent on whether a court may impose attorney’s fees against a party challenging a consent decree in a collateral action. The majority’s intimation to the contrary is conclusory dicta of the worst kind. Second, notwithstanding the possibility of fee liability, interested parties have strong incentives to intervene in a Title VII action rather than to wait and file a collateral attack. An intervenor may directly challenge the merits of a claim or defense in the underlying action, see 3B J. Moore & J. Kennedy, Moore’s Federal Practice II24.16[4] (2d ed. 1987), and may help craft an appropriate remedy. In so doing, an intervenor avoids the delay and increased costs of a collateral action. 778 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. is always a precondition to liability for fees. In Consumers Union, we absolved the Supreme Court of Virginia from fee liability because it had been acting in a legislative capacity when it promulgated the challenged regulations, and thus enjoyed common-law absolute legislative immunity. Id., at 738-739. Unlike the Commonwealth of Kentucky and the Supreme Court of Virginia, IFF A enjoys no special immunity warranting exemption from fee liability.7 Aside from its unpersuasive assertion that fee liability must be conditioned on a finding of wrongdoing, the majority never even attempts to explain why it adopts a categorical rule directing district courts to treat all intervenors like civil rights plaintiffs. Whatever validity such treatment might have where an intervenor raises a civil rights claim, there is absolutely no justification for it where, as in this case, an intervenor asserts non-civil-rights claims of third parties, or where an intervenor raises no third-party claims at all.8 The majority’s failure to differentiate among intervenors can- 7 Where Congress intends to exclude certain parties from fee entitlement or fee liability, it states so specifically. For example, § 706(k) itself expressly excludes the Federal Government from fee entitlement. See also Fair Labor Standards Act, 29 U. S. C. § 216(b) (authorizing fee liability only for “defendants” who are “employers”); Civil Rights Act of 1968, 42 U. S. C. § 3612(c) (authorizing fee entitlement only for “plaintiffs”). 8 Some parties intervene for the sole purpose of defending the challenged practice or opposing the relief sought by the civil rights plaintiffs. See, e. g., Diamond n. Charles, 476 U. S. 54 (1986) (pediatrician intervened to defend an abortion statute that neither implicated nor threatened his conduct); Akron Center for Reproductive Health v. Akron, 604 F. Supp. 1268, 1272 (ND Ohio 1984) (same). In most instances, intervenors not asserting the rights of third parties could adequately express their views by proceeding as amicus curiae. When they decline this option and instead voluntarily intervene, they benefit from “their ability to affect the course and substance of the litigation,” and thus should “fairly be charged with the consequences,” including the risk of attorney’s fees. Charles v. Daley, 846 F. 2d 1057, 1067 (CA7 1988); see also Akron Center, supra, at 1274. FLIGHT ATTENDANTS v. ZIPES 779 754 Marshall, J., dissenting not be squared with Congress’ conferral of discretion on the district courts. The majority also seeks to justify its interpretation of § 706(k) by asserting the importance of the claims asserted by intervenors. With respect to this case, the majority states that IFFA’s contract-based rights “are entitled to no less respect than the rights asserted by plaintiffs in the subject suit.” Ante, at 765. The issue, however, is not whether the claims are entitled to equal respect, but whether fees are beyond the discretion of the District Court.9 As the majority concedes, “intervenors raising non-Title VII claims are not, like Title VII plaintiffs, ‘the chosen instrument[s] of Congress.’” Ante, at 763, quoting Christiansburg Garment, 434 U. S., at 418. The central fact then is not, as the Court suggests, “that petitioner litigated ... to prevent TWA’s bargaining away of its members’ seniority rights in order to settle with respondents,” ante, at 765-766, or that IFFA did not violate Title VII, but rather that the plaintiffs who seek fees from IFFA are “the chosen instruments of Congress” to eradicate discrimination. In its rush to protect an intervenor who contributed almost $200,000 in costs and nearly three years to the plaintiffs’ struggle to achieve a settlement, the Court leaves behind the plaintiffs themselves, thereby reversing congressional priorities. The critical question in determining whether fees are awarded pursuant to § 706(k) should be whether the plaintiff prevailed, either against a named defendant or an intervenor. If the plaintiff has done so, fees ordinarily should—and certainly may—be awarded. Finally, the majority ignores the likely consequence of today’s decision. In the future, defendants can rely on in 9 The majority forgets, furthermore, that the Court has already recognized that vindicating the civil rights of victims of discrimination may require an award of retroactive seniority that may adversely affect innocent employees. See, e. g., Teamsters v. United States, 431 U. S. 324 (1977); Franks v. Bowman Transp. Co., 424 U. S. 747 (1976). 780 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. tervenors to raise many of their defenses, thereby minimizing the fee exposure of defendants and forcing prevailing plaintiffs to litigate many, if not most, of their claims against parties from whom they have no chance of recovering fees. Without the hope of obtaining compensation for the expenditures caused by intervenors, many victims of discrimination will be forced to forgo remedial litigation for lack of financial resources. As a result, injuries will go unredressed and the national policy against discrimination will go unredeemed. I dissent. WARD v. ROCK AGAINST RACISM 781 Syllabus WARD ET AL. v. ROCK AGAINST RACISM CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT No. 88-226. Argued February 27, 1989—Decided June 22, 1989 Respondent Rock Against Racism (RAR), furnishing its own sound equipment and technicians, has sponsored yearly programs of rock music at the Naumberg Acoustic Bandshell in New York City’s Central Park. The city received numerous complaints about excessive noise at RAR’s concerts from users of the nearby Sheep Meadow, an area designated by the city for passive recreation, from other users of the park, and from residents of areas adjacent to the park. Moreover, when the city shut off the power after RAR ignored repeated requests to lower the volume at one of its concerts, the audience became abusive and disruptive. The city also experienced problems at bandshell events put on by other sponsors, who, due to their use of inadequate sound equipment or sound technicians unskilled at mixing sound for the bandshell area, were unable to provide sufficient amplification levels, resulting in disappointed or unruly audiences. Rejecting various other solutions to the excessive noise and inadequate amplification problems, the city adopted a Use Guideline for the bandshell which specified that the city would furnish high quality sound equipment and retain an independent, experienced sound technician for all performances. After the city implemented this guideline, RAR amended a pre-existing District Court complaint against the city to seek damages and a declaratory judgment striking down the guideline as facially invalid under the First Amendment. The court upheld the guideline, finding, inter alia, that performers who had used the city’s sound system and technician had been uniformly pleased; that, although the city’s technician ultimately controlled both sound volume and mix, the city’s practice was to give the sponsor autonomy as to mix and to confer with him before turning the volume down; and that the city’s amplification system was sufficient for RAR’s needs. Applying this Court’s three-part test for judging the constitutionality of governmental regulation of the time, place, and manner of protected speech, the court found the guideline valid. The Court of Appeals reversed on the ground that such regulations’ method and extent must be the least intrusive upon the freedom of expression as is reasonably necessary to achieve the regulations’ purpose, finding that there were various less restrictive means by which the city could control excessive volume without also intruding on RAR’s ability to control sound mix. 782 OCTOBER TERM, 1988 Syllabus 491 U. S. Held: The city’s sound-amplification guideline is valid under the First Amendment as a reasonable regulation of the place and manner of protected speech. Pp. 790-803. (a) The guideline is content neutral, since it is justified without reference to the content of the regulated speech. The city’s principal justification—the desire to control noise in order to retain the sedate character of the Sheep Meadow and other areas of the park and to avoid intrusion into residential areas—has nothing to do with content. The city’s other justification, its interest in ensuring sound quality, does not render the guideline content based as an attempt to impose subjective standards of acceptable sound mix on performers, since the city has expressly disavowed any such intent and requires its technician to defer to the sponsor’s wishes as to mix. On the record below, the city’s sound quality concern extends only to the clearly content-neutral goals of ensuring adequate amplification and avoiding volume problems associated with inadequate mix. There is no merit to RAR’s argument that the guideline is nonetheless invalid on its face because it places unbridled discretion in the hands of city enforcement officials. Even granting the doubtful proposition that this claim falls within the narrow class of permissible facial challenges to allegedly unconstrained grants of regulatory authority, the claim nevertheless fails, since the guideline’s own terms in effect forbid officials purposely to select an inadequate system or to vary sound quality or volume based on the performer’s message. Moreover, the city has applied a narrowing construction to the guideline by requiring officials to defer to sponsors on sound quality and confer with them as to volume problems, and by mandating that amplification be sufficient for the sound to reach all concert ground listeners. Pp. 791-796. (b) The guideline is narrowly tailored to serve significant governmental interests. That the city has a substantial interest in protecting citizens from unwelcome and excessive noise, even in a traditional public forum such as the park, cannot be doubted. Moreover, it has a substantial interest in ensuring the sufficiency of sound amplification at bandshell events in order to allow citizens to enjoy the benefits of the park, in light of the evidence that inadequate amplification had resulted in the inability of some audiences to hear performances. The Court of Appeals erred in requiring the city to prove that the guideline was theleast intrusive means of furthering these legitimate interests, since a “less-restrictive-altemative analysis” has never been—and is here, again, specifically rejected as—a part of the inquiry into the validity of a time, place, or manner regulation. See Clark v. Community for Creative Non-Violence, 468 U. S. 288, 293; Regan v. Time, Inc., 468 U. S. 641. The requirement of narrow tailoring is satisfied so long as the regulation promotes a substantial governmental interest that would be WARD v. ROCK AGAINST RACISM 783 781 Syllabus achieved less effectively absent the regulation, and the means chosen are not substantially broader than necessary to achieve that interest. If these standards are met, courts should defer to the government’s reasonable determination. Here, the city’s substantial interest in limiting sound volume is served in a direct and effective way by the requirement that its technician control the mixing board. Absent this requirement, the city’s interest would have been served less well, as is evidenced by the excessive noise complaints generated by RAR’s past concerts. The city also could reasonably have determined that, overall, its interest in ensuring that sound amplification was sufficient to reach all concert ground listeners would be served less effectively without the guideline than with it, since, by providing competent technicians and adequate equipment, the city eliminated inadequate amplification problems that plagued some performers in the past. Furthermore, in the absence of evidence that the guideline had a substantial deleterious effect on the ability of performers to achieve the quality of sound they desired, there is no merit to RAR’s contention that the guideline is substantially broader than necessary to achieve the city’s legitimate ends. Pp. 796-802. (c) The guideline leaves open ample alternative channels of communication, since it does not attempt to ban any particular manner or type of expression at a given place and time. Rather, it continues to permit expressive activity in the bandshell and has no effect on the quantity or content of that expression beyond regulating the extent of amplification. That the city’s volume limitations may reduce to some degree the potential audience for RAR’s speech is of no consequence, since there has been no showing that the remaining avenues of communication are inadequate. Pp. 802-803. 848 F. 2d 367, reversed. Kennedy, J., delivered the opinion of the Court, in which Rehnquist, C. J., and White, O’Connor, and Scalia, JJ., joined. Blackmun, J., concurred in the result. Marshall, J., filed a dissenting opinion, in which Brennan and Stevens, JJ., joined, post, p. 803. Leonard J. Koerner argued the cause for petitioners. With him on the brief were Peter L. Zimroth, Larry A. Sonnenshein, and Julian L. Kalkstein. William M. Kunstler argued the cause for respondent. With him on the brief was Noah A. Kinigstein. * *Briefs of amici curiae urging reversal were filed for the United States by Solicitor General Fried, Assistant Attorney General Bolton, Deputy 784 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. Justice Kennedy delivered the opinion of the Court. In the southeast portion of New York City’s Central Park, about 10 blocks upward from the park’s beginning point at 59th Street, there is an amphitheater and stage structure known as the Naumberg Acoustic Bandshell. The bandshell faces west across the remaining width of the park. In close proximity to the bandshell, and lying within the directional path of its sound, is a grassy open area called the Sheep Meadow. The city has designated the Sheep Meadow as a quiet area for passive recreations like reclining, walking, and reading. Just beyond the park, and also within the potential sound range of the bandshell, are the apartments and residences of Central Park West. This case arises from the city’s attempt to regulate the volume of amplified music at the bandshell so the performances are satisfactory to the audience without intruding upon those who use the Sheep Meadow or live on Central Park West and in its vicinity. The city’s regulation requires bandshell performers to use sound-amplification equipment and a sound technician provided by the city. The challenge to this volume control technique comes from the sponsor of a rock concert. The trial court sustained the noise control measures, but the Court of Appeals for the Second Circuit reversed. We granted certiorari to resolve the important First Amendment issues presented by the case. I Rock Against Racism, respondent in this case, is an unincorporated association which, in its own words, is “dedicated to the espousal and promotion of antiracist views.” App. to Pet. for Cert. 3. Each year from 1979 through 1986, RAR has sponsored a program of speeches and rock music at the Solicitor General Ayer, Stephen L. Nightingale, and John F. Cordes; and for the National League of Cities by Benna Ruth Solomon, Joyce Holmes Benjamin, and Ogden N. Lewis. WARD v. ROCK AGAINST RACISM 785 781 Opinion of the Court bandshell. RAR has furnished the sound equipment and sound technician used by the various performing groups at these annual events. Over the years, the city received numerous complaints about excessive sound amplification at respondent’s concerts from park users and residents of areas adjacent to the park. On some occasions RAR was less than cooperative when city officials asked that the volume be reduced; at one concert, police felt compelled to cut off the power to the sound system, an action that caused the audience to become unruly and hostile. App. 127-131, 140-141, 212-214, 345-347. Before the 1984 concert, city officials met with RAR representatives to discuss the problem of excessive noise. It was decided that the city would monitor sound levels at the edge of the concert ground, and would revoke respondent’s event permit if specific volume limits were exceeded. Sound levels at the concert did exceed acceptable levels for sustained periods of time, despite repeated warnings and requests that the volume be lowered. Two citations for excessive volume were issued to respondent during the concert. When the power was eventually shut off, the audience became abusive and disruptive. The following year, when respondent sought permission to hold its upcoming concert at the bandshell, the city declined to grant an event permit, citing its problems with noise and crowd control at RAR’s previous concerts. The city suggested some other city-owned facilities as alternative sites for the concert. RAR declined the invitation and filed suit in United States District Court against the city, its mayor, and various police and parks department officials, seeking an injunction directing issuance of an event permit. After respondent agreed to abide by all applicable regulations, the parties reached agreement and a permit was issued. The city then undertook to develop comprehensive New York City Parks Department Use Guidelines for the Naum-berg Bandshell. A principal problem to be addressed by 786 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. the guidelines was controlling the volume of amplified sound at bandshell events. A major concern was that at some bandshell performances the event sponsors had been unable to “provide the amplification levels required and ‘crowds unhappy with the sound became disappointed or unruly.’” Brief for Petitioners 9. The city found that this problem had several causes, including inadequate sound equipment, sound technicians who were either unskilled at mixing sound outdoors or unfamiliar with the acoustics of the bandshell and its surroundings, and the like. Because some performers compensated for poor sound mix by raising volume, these factors tended to exacerbate the problem of excess noise.1 App. 30, 189, 218-219. The city considered various solutions to the soundamplification problem. The idea of a fixed decibel limit for all performers using the bandshell was rejected because the impact on listeners of a single decibel level is not constant, but varies in response to changes in air temperature, foliage, audience size, and like factors. Id., at 31,220,285-286. The city also rejected the possibility of employing a sound technician to operate the equipment provided by the various sponsors of bandshell events, because the city’s technician might have had difficulty satisfying the needs of sponsors while operating unfamiliar, and perhaps inadequate, sound equipment. Id., 'The amplified sound heard at a rock concert consists of two components, volume and mix. Sound produced by the various instruments and performers on stage is picked up by microphones and fed into a central mixing board, where it is combined into one signal and then amplified through speakers to the audience. A sound technician is at the mixing board to select the appropriate mix, or balance, of the various sounds produced on stage, and to add other effects as desired by the performers. In addition to controlling the sound mix, the sound technician also controls the overall volume of sound reaching the audience. During the course of a performance, the sound technician is continually manipulating various controls on the mixing board to provide the desired sound mix and volume. The sound technician thus plays an important role in determining the quality of the amplified sound that reaches the audience. WARD v. ROCK AGAINST RACISM 787 781 Opinion of the Court at 220. Instead, the city concluded that the most effective way to achieve adequate but not excessive sound amplification would be for the city to furnish high quality sound equipment and retain an independent, experienced sound technician for all performances at the bandshell. After an extensive search the city hired a private sound company capable of meeting the needs of all the varied users of the bandshell. The Use Guidelines were promulgated on March 21, 1986.2 After learning that it would be expected to comply with the guidelines at its upcoming annual concert in May 1986, respondent returned to the District Court and filed a motion for an injunction against the enforcement of certain aspects of the guidelines. The District Court preliminarily enjoined enforcement of the sound-amplification rule on May 1, 1986. See 636 F. Supp. 178 (SDNY 1986). Under the protection of the injunction, and alone among users of the bandshell in the 1986 season, RAR was permitted to use its own sound equip 2 In pertinent part, the Use Guidelines provide: “SOUND AMPLIFICATION “To provide the best sound for all events Department of Parks and Recreation has leased a sound amplification system designed for the specific demands of the Central Park Bandshell. To insure appropriate sound quality balanced with respect for nearby residential neighbors and the mayoralty decreed quiet zone of Sheep Meadow, all sponsors may use only the Department of Parks and Recreation sound system. DEPARTMENT OF PARKS AND RECREATION IS TO BE THE SOLE AND ONLY PROVIDER OF SOUND AMPLIFICATION, INCLUDING THOUGH NOT LIMITED TO AMPLIFIERS, SPEAKERS, MONITORS, MICROPHONES, AND PROCESSORS. “Clarity of sound results from a combination of amplification equipment and a sound technician’s familiarity and proficiency with that system. Department of Parks and Recreation will employ a professional sound technician [who] will be fully versed in sound bounce patterns, daily air currents, and sound skipping within the Park. The sound technician must also consider the Bandshell’s proximity to Sheep Meadow, activities at Bethesda Terrace, and the New York City Department of Environmental Protection recommendations.” App. 375-376. 788 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. ment and technician, just as it had done in prior years. RAR’s 1986 concert again generated complaints about excessive noise from park users and nearby residents. App. 127, 138. After the concert, respondent amended its complaint to seek damages and a declaratory judgment striking down the guidelines as facially invalid. After hearing five days of testimony about various aspects of the guidelines, the District Court issued its decision upholding the sound-amplification guideline.3 The court found that the city had been “motivated by a desire to obtain top-flight sound equipment and experienced operators” in selecting an independent contractor to provide the equipment and technician for bandshell events, and that the performers who did use the city’s sound system in the 1986 season, in performances “which ran the full cultural gamut from grand opera to salsa to reggae,” were uniformly pleased with the quality of the sound provided. 658 F. Supp. 1346, 1352 (SDNY 1987). Although the city’s sound technician controlled both sound volume and sound mix by virtue of his position at the mixing board, the court found that “[t]he City’s practice for events at the Bandshell is to give the sponsor autonomy with respect to the sound mix: balancing treble with bass, highlighting a particular instrument or voice, and the like,” and that the city’s sound technician “does all he can to accommodate the sponsor’s desires in those regards.” Ibid. Even with respect to volume control, the city’s practice was to confer with the sponsor before making any decision to turn the volume down. Ibid. In some instances, as with a New York Grand Opera performance, the sound technician accommodated the performers’ unique needs by integrating special microphones with the city’s equipment. The court specifically found that “[t]he City’s implementation of the Bandshell guidelines provides for a sound amplification system capable of meeting 3 The court invalidated certain other aspects of the Use Guidelines, but those provisions are not before us. WARD v. ROCK AGAINST RACISM 789 781 Opinion of the Court RAR’s technical needs and leaves control of the sound ‘mix’ in the hands of RAR.” Id., at 1353. Applying this Court’s three-part test for judging the constitutionality of government regulation of the time, place, or manner of protected speech, the court found the city’s regulation valid. The Court of Appeals reversed. 848 F. 2d 367 (CA2 1988). After recognizing that “[c]ontent neutral time, place and manner regulations are permissible so long as they are narrowly tailored to serve a substantial government interest and do not unreasonably limit alternative avenues of expression,” the court added the proviso that “the method and extent of such regulation must be reasonable, that is, it must be the least intrusive upon the freedom of expression as is reasonably necessary to achieve a legitimate purpose of the regulation.” Id., at 370 (citing United States v. O’Brien, 391 U. S. 367, 377 (1968)). Applying this test, the court determined that the city’s guideline was valid only to the extent necessary to achieve the city’s legitimate interest in controlling excessive volume, but found there were various alternative means of controlling volume without also intruding on respondent’s ability to control the sound mix. For example, the city could have directed respondent’s sound technician to keep the volume below specified levels. Alternatively, a volume-limiting device could have been installed; and as a “last resort,” the court suggested, “the plug can be pulled on the sound to enforce the volume limit.” 848 F. 2d, at 372, n. 6. In view of the potential availability of these seemingly less restrictive alternatives, the Court of Appeals concluded that the sound-amplification guideline was invalid because the city had failed to prove that its regulation “was the least intrusive means of regulating the volume.” Id., at 371. We granted certiorari, 488 U. S. 816 (1988), to clarify the legal standard applicable to governmental regulation of the time, place, or manner of protected speech. Because the Court of Appeals erred in requiring the city to prove that its regulation was the least intrusive means of furthering its le 790 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. gitimate governmental interests, and because the ordinance is valid on its face, we now reverse. II Music is one of the oldest forms of human expression. From Plato’s discourse in the Republic to the totalitarian state in our own times, rulers have known its capacity to appeal to the intellect and to the emotions, and have censored musical compositions to serve the needs of the state. See 2 Dialogues of Plato, Republic, bk. 3, pp. 231, 245-248 (B. Jowett transi., 4th ed. 1953) (“Our poets must sing in another and a nobler strain”); Musical Freedom and Why Dictators Fear It, N. Y. Times, Aug. 23, 1981, section 2, p. 1, col. 5; Soviet Schizophrenia toward Stravinsky, N. Y. Times, June 26, 1982, section 1, p. 25, col. 2; Symphonic Voice from China Is Heard Again, N. Y. Times, Oct. 11, 1987, section 2, p. 27, col. 1. The Constitution prohibits any like attempts in our own legal order. Music, as a form of expression and communication, is protected under the First Amendment. In the case before us the performances apparently consisted of remarks by speakers, as well as rock music, but the case has been presented as one in which the constitutional challenge is to the city’s regulation of the musical aspects of the concert; and, based on the principle we have stated, the city’s guideline must meet the demands of the First Amendment. The parties do not appear to dispute that proposition. We need not here discuss whether a municipality which owns a bandstand or stage facility may exercise, in some circumstances, a proprietary right to select performances and control their quality. See Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546, 570-574 (1975) (Rehnquist, J., dissenting). Though it did demonstrate its own interest in the effort to insure high quality performances by providing the equipment in question, the city justifies its guideline as a regulatory measure to limit and control noise. Here the bandshell was open, apparently, to all performers; and we de- WARD v. ROCK AGAINST RACISM 791 781 Opinion of the Court cide the case as one in which the bandshell is a public forum for performances in which the government’s right to regulate expression is subject to the protections of the First Amendment. United States v. Grace, 461 U. S. 171, 177 (1983); see Frisby v. Schultz, 487 U. S. 474, 481 (1988); Perry Education Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 45 (1983). Our cases make clear, however, that even in a public forum the government may impose reasonable restrictions on the time, place, or manner of protected speech, provided the restrictions “are justified without reference to the content of the regulated speech, that they are narrowly tailored to serve a significant governmental interest, and that they leave open ample alternative channels for communication of the information.” Clark v. Community for Creative Non-Violence, 468 U. S. 288, 293 (1984); see Heffron v. International Society for Krishna Consciousness, Inc., 452 U. S. 640, 648 (1981) (quoting Virginia Pharmacy Bd. v. Virginia Citizens Consumer Council, Inc., 425 U. S. 748, 771 (1976)). We consider these requirements in turn. A The principal inquiry in determining content neutrality, in speech cases generally and in time, place, or manner cases in particular, is whether the government has adopted a regulation of speech because of disagreement with the message it conveys. Community for Creative Non-Violence, supra, at 295. The government’s purpose is the controlling consideration. A regulation that serves purposes unrelated to the content of expression is deemed neutral, even if it has an incidental effect on some speakers or messages but not others. See Renton v. Playtime Theatres, Inc., 475 U. S. 41, 47-48 (1986). Government regulation of expressive activity is content neutral so long as it is “justified without reference to the content of the regulated speech.” Community for Creative Non-Violence, supra, at 293 (emphasis added); Heffron, supra, at 648 (quoting Virginia Pharmacy Bd., supra, at 792 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. 771); see Boos v. Barry, 485 U. S. 312, 320-321 (1988) (opinion of O’Connor, J.). The principal justification for the sound-amplification guideline is the city’s desire to control noise levels at bandshell events, in order to retain the character of the Sheep Meadow and its more sedate activities, and to avoid undue intrusion into residential areas and other areas of the park. This justification for the guideline “ha[s] nothing to do with content,” Boos v. Barry, supra, at 320, and it satisfies the requirement that time, place, or manner regulations be content neutral. The only other justification offered below was the city’s interest in “ensuring] the quality of sound at Bandshell events.” 658 F. Supp., at 1352; see 848 F. 2d, at 370, n. 3. Respondent urges that this justification is not content neutral because it is based upon the quality, and thus the content, of the speech being regulated. In respondent’s view, the city is seeking to assert artistic control over performers at the bandshell by enforcing a bureaucratically determined, valueladen conception of good sound. That all performers who have used the city’s sound equipment have been completely satisfied is of no moment, respondent argues, because “(t]he First Amendment does not permit and cannot tolerate state control of artistic expression merely because the State claims that [its] efforts will lead to ‘top-quality’ results.” Brief for Respondent 19. While respondent’s arguments that the government may not interfere with artistic judgment may have much force in other contexts, they are inapplicable to the facts of this case. The city has disclaimed in express terms any interest in imposing its own view of appropriate sound mix on performers. To the contrary, as the District Court found, the city requires its sound technician to defer to the wishes of event sponsors concerning sound mix. 658 F. Supp., at 1352-1353. On this record, the city’s concern with sound quality extends only to the clearly content-neutral goals of ensuring adequate WARD v. ROCK AGAINST RACISM 793 781 Opinion of the Court sound amplification and avoiding the volume problems associated with inadequate sound mix.4 Any governmental attempt to serve purely esthetic goals by imposing subjective standards of acceptable sound mix on performers would raise serious First Amendment concerns, but this case provides us with no opportunity to address those questions. As related above, the District Court found that the city’s equipment and its sound technician could meet all of the standards requested by the performers, including RAR. Respondent argues further that the guideline, even if not content based in explicit terms, is nonetheless invalid on its face because it places unbridled discretion in the hands of city officials charged with enforcing it. See Lakewood v. Plain Dealer Publishing Co., 486 U. S. 750, 769-772 (1988) (4-to-3 decision); Heffron v. International Society for Krishna Consciousness, Inc., supra, at 649; Freedman v. Maryland, 380 U. S. 51, 56 (1965); Thornhill v. Alabama, 310 U. S. 88, 97 (1940). According to respondent, there is nothing in the language of the guideline to prevent city officials from selecting wholly inadequate sound equipment or technicians, or even from varying the volume and quality of sound based on the message being conveyed by the performers. As a threshold matter, it is far from clear that respondent should be permitted to bring a facial challenge to this aspect of the regulation. Our cases permitting facial challenges to regulations that allegedly grant officials unconstrained authority to regulate speech have generally involved licensing schemes that “ves[t] unbridled discretion in a government official over whether to permit or deny expressive activity.” Plain Dealer, supra, at 755. The grant of discretion that re 4 As noted above, there is evidence to suggest that volume control and sound mix are interrelated to a degree, in that performers unfamiliar with the acoustics of the bandshell sometimes attempt to compensate for poor sound mix by increasing volume. App. 218, 290-291. By providing adequate sound equipment and professional sound mixing, the city avoids this problem. 794 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. spondent seeks to challenge here is of an entirely different, and lesser, order of magnitude, because respondent does not suggest that city officials enjoy unfettered discretion to deny bandshell permits altogether. Rather, respondent contends only that the city, by exercising what is concededly its right to regulate amplified sound, could choose to provide inadequate sound for performers based on the content of their speech. Since respondent does not claim that city officials enjoy unguided discretion to deny the right to speak altogether, it is open to question whether respondent’s claim falls within the narrow class of permissible facial challenges to allegedly unconstrained grants of regulatory authority. Cf. 486 U. S., at 787 (White, J., dissenting) (arguing that facial challenges of this type are permissible only where “the local law at issue require[s] licenses—not for a narrow category of expressive conduct that could be prohibited—but for a sweeping range of First Amendment protected activity”). We need not decide, however, whether the “extraordinary doctrine” that permits facial challenges to some regulations of expression, see id., at 772 (White, J., dissenting), should be extended to the circumstances of this case, for respondent’s facial challenge fails on its merits. The city’s guideline states that its goals are to “provide the best sound for all events” and to “insure appropriate sound quality balanced with respect for nearby residential neighbors and the may-orally decreed quiet zone of [the] Sheep Meadow.” App. 375. While these standards are undoubtedly flexible, and the officials implementing them will exercise considerable discretion, perfect clarity and precise guidance have never been required even of regulations that restrict expressive activity. See Grayned v. City of Rockford, 408 U. S. 104, 110 (1972) (“Condemned to the use of words, we can never expect mathematical certainty in our language”); see also Kovacs n. Cooper, 336 U. S. 77, 79 (1949) (rejecting vagueness challenge to city ordinance forbidding “loud and raucous” sound amplification) (opinion of Reed, J.). By its own terms the WARD v. ROCK AGAINST RACISM 795 781 Opinion of the Court city’s sound-amplification guideline must be interpreted to forbid city officials purposely to select inadequate sound systems or to vary the sound quality or volume based on the message being delivered by performers. The guideline is not vulnerable to respondent’s facial challenge.5 Even if the language of the guideline were not sufficient on its face to withstand challenge, our ultimate conclusion would be the same, for the city has interpreted the guideline in such a manner as to provide additional guidance to the officials charged with its enforcement. The District Court expressly found that the city’s policy is to defer to the sponsor’s desires concerning sound quality. 658 F. Supp., at 1352. With respect to sound volume, the city retains ultimate control, but city officials “mak[e] it a practice to confer with the sponsor if any questions of excessive sound arise, before taking any corrective action.” Ibid. The city’s goal of ensuring that “the sound amplification [is] sufficient to reach all listeners within the defined concertground,” ibid., serves to limit further the discretion of the officials on the scene. Administrative interpretation and implementation of a regulation are, of course, highly relevant to our analysis, for “[i]n evaluating a facial 3 The dissent’s suggestion that the guideline constitutes a prior restraint is not consistent with our cases. See post, at 808-809. As we said in Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546 (1975), the regulations we have found invalid as prior restraints have “had this in common: they gave public officials the power to deny use of a forum in advance of actual expression.” Id., at 553. The sound-amplification guideline, by contrast, grants no authority to forbid speech, but merely permits the city to regulate volume to the extent necessary to avoid excessive noise. It is true that the city’s sound technician theoretically possesses the power to shut off the volume for any particular performer, but that hardly distinguishes this regulatory scheme from any other; government will always possess the raw power to suppress speech through force, and indeed it was in part to avoid the necessity of exercising its power to “pull the plug” on the volume that the city adopted the sound-amplification guideline. The relevant question is whether the challenged regulation authorizes suppression of speech in advance of its expression, and the sound-amplification guideline does not. 796 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. challenge to a state law, a federal court must . . . consider any limiting construction that a state court or enforcement agency has proffered.” Hoffman Estates v. The Flipside, Hoffman Estates, Inc., 455 U. S. 489, 494, n. 5 (1982); see Plain Dealer, 486 U. S., at 769-770, and n. 11; United States v. Grace, 461 U. S., at 181, n. 10; Grayned v. City of Rockford, supra, at 110; Poulos v. New Hampshire, 345 U. S. 395 (1953). Any inadequacy on the face of the guideline would have been more than remedied by the city’s narrowing construction. B The city’s regulation is also “narrowly tailored to serve a significant governmental interest.” Community for Creative Non-Violence, 468 U. S., at 293. Despite respondent’s protestations to the contrary, it can no longer be doubted that government “ha[s] a substantial interest in protecting its citizens from unwelcome noise.” City Council of Los Angeles v. Taxpayers for Vincent, 466 U. S. 789, 806 (1984) (citing Kovacs n. Cooper, supra); see Grayned, supra, at 116. This interest is perhaps at its greatest when government seeks to protect “‘the well-being, tranquility, and privacy of the home,’” Frisby v. Schultz, 487 U. S., at 484 (quoting Carey v. Brown, 447 U. S. 455, 471 (1980)), but it is by no means limited to that context, for the government may act to protect even such traditional public forums as city streets and parks from excessive noise. Kovacs n. Cooper, 336 U. S., at 86-87 (opinion of Reed, J.); id., at 96-97 (Frankfurter, J., concurring); id., at 97 (Jackson, J., concurring); see Community for Creative Non-Violence, supra, at 296 (recognizing the government’s “substantial interest in maintaining the parks ... in an attractive and intact condition, readily available to the millions of people who wish to see and enjoy them”). We think it also apparent that the city’s interest in ensuring the sufficiency of sound amplification at bandshell events is a substantial one. The record indicates that inadequate WARD v. ROCK AGAINST RACISM 797 781 Opinion of the Court sound amplification has had an adverse affect on the ability of some audiences to hear and enjoy performances at the bandshell. The city enjoys a substantial interest in ensuring the ability of its citizens to enjoy whatever benefits the city parks have to offer, from amplified music to silent meditation. See Community for Creative Non-Violence, supra, at 296. The Court of Appeals recognized the city’s substantial interest in limiting the sound emanating from the bandshell. See 848 F. 2d, at 370. The court concluded, however, that the city’s sound-amplification guideline was not narrowly tailored to further this interest, because “it has not [been] shown . . . that the requirement of the use of the city’s sound system and technician was the least intrusive means of regulating the volume.” Id., at 371 (emphasis added). In the court’s judgment, there were several alternative methods of achieving the desired end that would have been less restrictive of respondent’s First Amendment rights. The Court of Appeals erred in sifting through all the available or imagined alternative means of regulating sound volume in order to determine whether the city’s solution was “the least intrusive means” of achieving the desired end. This “less-restrictive-alternative analysis . . . has never been a part of the inquiry into the validity of a time, place, and manner regulation.” Regan v. Time, Inc., 468 U. S. 641, 657 (1984) (opinion of White, J.). Instead, our cases quite clearly hold that restrictions on the time, place, or manner of protected speech are not invalid “simply because there is some imaginable alternative that might be less burdensome on speech.” United States v. Albertini, 472 U. S. 675, 689 (1985). The Court of Appeals apparently drew its least-intrusive-means requirement from United States v. O'Brien, 391 U. S., at 377, the case in which we established the standard for judging the validity of restrictions on expressive conduct. See 848 F. 2d, at 370. The court’s reliance was misplaced, 798 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. however, for we have held that the O'Brien test “in the last analysis is little, if any, different from the standard applied to time, place, or manner restrictions.” Community for Creative Non-Violence, supra, at 298. Indeed, in Community for Creative Non-Violence, we squarely rejected reasoning identical to that of the court below: “We are unmoved by the Court of Appeals’ view that the challenged regulation is unnecessary, and hence invalid, because there are less speech-restrictive alternatives that could have satisfied the Government interest in preserving park lands. . . . We do not believe . . . that either United States v. O'Brien or the time, place, or manner decisions assign to the judiciary the authority to replace the [parks department] as the manager of the [city’s] parks or endow the judiciary with the competence to judge how much protection of park lands is wise and how that level of conservation is to be attained.” 468 U. S., at 299. Lest any confusion on the point remain, we reaffirm today that a regulation of the time, place, or manner of protected speech must be narrowly tailored to serve the government’s legitimate, content-neutral interests but that it need not be the least restrictive or least intrusive means of doing so.6 6 Respondent contends that our decision last Term in Boos v. Barry, 485 U. S. 312 (1988), supports the conclusion that “a regulation is neither precisely drawn nor ‘narrowly tailored’ if less intrusive means than those employed are available.” Brief for Respondent 27. In Boos we concluded that the government regulation at issue was “not narrowly tailored; a less restrictive alternative is readily available.” 485 U. S., at 329 (citing Wygant v. Jackson Bd. of Ed., 476 U. S. 267, 280, n. 6 (1986) (plurality opinion)). In placing reliance on Boos, however, respondent ignores a crucial difference between that case and this. The regulation we invalidated in Boos was a content-based ban on displaying signs critical of foreign governments; such content-based restrictions on political speech “must be subjected to the most exacting scrutiny.” 485 U. S., at 321. While time, place, or manner regulations must also be “narrowly tailored” in order to survive First Amendment challenge, we have never applied strict scrutiny WARD v. ROCK AGAINST RACISM 799 781 Opinion of the Court Rather, the requirement of narrow tailoring is satisfied “so long as the . . . regulation promotes a substantial government interest that would be achieved less effectively absent the regulation.” United States v. Albertini, supra, at 689; see also Community for Creative Non-Violence, supra, at 297. To be sure, this standard does not mean that a time, place, or manner regulation may burden substantially more speech than is necessary to further the government’s legitimate interests. Government may not regulate expression in such a manner that a substantial portion of the burden on speech does not serve to advance its goals.7 See Frisby in this context. As a result, the same degree of tailoring is not required of these regulations, and least-restrictive-altemative analysis is wholly out of place. For the same reason, the dissent’s citation of Richmond v. J. A. Croson Co., 488 U. S. 469 (1989), is beside the point. See post, at 806, n. 4. Croson, like Boos, is a strict-scrutiny case; even the dissent does not argue that strict scrutiny is applicable to time, place, or manner regulations. Our summary affirmance of Watseka v. Illinois Public Action Council, 796 F. 2d 1547 (CA7 1986), aff’d, 479 U. S. 1048 (1987), is not to the contrary. Although the Seventh Circuit in that case did adopt the least-restrictive-alternative approach, see 796 F. 2d, at 1553-1554, its judgment was also supported by the alternative grounds that the regulation at issue did not serve to further the stated governmental interests and did not leave open alternative channels of communication. Id., at 1555-1558. As we have noted on more than one occasion: “A summary disposition affirms only the judgment of the court below, and no more may be read into our action than was essential to sustain that judgment.” Anderson v. Cele-brezze, 460 U. S. 780, 785, n. 5 (1983). 7 The dissent’s attempt to analogize the sound-amplification guideline to a total ban on distribution of handbills is imaginative but misguided. See post, at 806-807. The guideline does not ban all concerts, or even all rock concerts, but instead focuses on the source of the evils the city seeks to eliminate—excessive and inadequate sound amplification—and eliminates them without at the same time banning or significantly restricting a substantial quantity of speech that does not create the same evils. This is the essence of narrow tailoring. A ban on handbilling, of course, would suppress a great quantity of speech that does not cause the evils that it seeks to eliminate, whether they be fraud, crime, litter, traffic congestion, or noise. See Martin v. Struthers, 319 U. S. 141, 145-146 (1943). For that 800 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. v. Schultz, 487 U. S., at 485 (“A complete ban can be narrowly tailored but only if each activity within the proscription’s scope is an appropriately targeted evil”). So long as the means chosen are not substantially broader than necessary to achieve the government’s interest, however, the regulation will not be invalid simply because a court concludes that the government’s interest could be adequately served by some less-speech-restrictive alternative. “The validity of [time, place, or manner] regulations does not turn on a judge’s agreement with the responsible decisionmaker concerning the most appropriate method for promoting significant government interests” or the degree to which those interests should be promoted. United States v. Albertini, 472 U. S., at 689; see Community for Creative Non-Violence, supra, at 299. It is undeniable that the city’s substantial interest in limiting sound volume is served in a direct and effective way by the requirement that the city’s sound technician control the mixing board during performances. Absent this requirement, the city’s interest would have been served less well, as is evidenced by the complaints about excessive volume generated by respondent’s past concerts. The alternative regulatory methods hypothesized by the Court of Appeals reflect nothing more than a disagreement with the city over how much control of volume is appropriate or how that level of control is to be achieved. See Community for Creative Non-Violence, supra, at 299. The Court of Appeals erred in failing to defer to the city’s reasonable determination that its interest in controlling volume would be best served by requiring bandshell performers to utilize the city’s sound technician. The city’s second content-neutral justification for the guideline, that of ensuring “that the sound amplification [is] sufficient to reach all listeners within the defined concertreason, a complete ban on handbilling would be substantially broader than necessary to achieve the interests justifying it. WARD v. ROCK AGAINST RACISM 801 781 Opinion of the Court ground,” 658 F. Supp., at 1352, also supports the city’s choice of regulatory methods. By providing competent sound technicians and adequate amplification equipment, the city eliminated the problems of inexperienced technicians and insufficient sound volume that had plagued some bandshell performers in the past. No doubt this concern is not applicable to respondent’s concerts, which apparently were characterized by more-than-adequate sound amplification. But that fact is beside the point, for the validity of the regulation depends on the relation it bears to the overall problem the government seeks to correct, not on the extent to which it furthers the government’s interests in an individual case. Here, the regulation’s effectiveness must be judged by considering all the varied groups that use the bandshell, and it is valid so long as the city could reasonably have determined that its interests overall would be served less effectively without the sound-amplification guideline than with it. United States v. Albertini, supra, at 688-689; Community for Creative Non-Violence, 468 U. S., at 296-297. Considering these proffered justifications together, therefore, it is apparent that the guideline directly furthers the city’s legitimate governmental interests and that those interests would have been less well served in the absence of the sound-amplification guideline. Respondent nonetheless argues that the sound-amplification guideline is not narrowly tailored because, by placing control of sound mix in the hands of the city’s technician, the guideline sweeps far more broadly than is necessary to further the city’s legitimate concern with sound volume. According to respondent, the guideline “targets . . . more than the exact source of the ‘evil’ it seeks to remedy.” Frisby v. Schultz, supra, at 485. If the city’s regulatory scheme had a substantial deleterious effect on the ability of bandshell performers to achieve the quality of sound they desired, respondent’s concerns would have considerable force. The District Court found, 802 OCTOBER TERM, 1988 Opinion of the Court 491 U. S. however, that pursuant to city policy, the city’s sound technician “give[s] the sponsor autonomy with respect to the sound mix . . . [and] does all that he can to accommodate the sponsor’s desires in those regards.” 658 F. Supp., at 1352. The court squarely rejected respondent’s claim that the city’s “technician is not able properly to implement a sponsor’s instructions as to sound quality or mix,” finding that “[n]o evidence to that effect was offered at trial; as noted, the evidence is to the contrary.” App. to Pet. for Cert. 89. In view of these findings, which were not disturbed by the Court of Appeals, we must conclude that the city’s guideline has no material impact on any performer’s ability to exercise complete artistic control over sound quality. Since the guideline allows the city to control volume without interfering with the performer’s desired sound mix, it is not “substantially broader than necessary” to achieve the city’s legitimate ends, City Council of Los Angeles v. Taxpayers for Vincent, 466 U. S., at 808, and thus it satisfies the requirement of narrow tailoring. C The final requirement, that the guideline leave open ample alternative channels of communication, is easily met. Indeed, in this respect the guideline is far less restrictive than regulations we have upheld in other cases, for it does not attempt to ban any particular manner or type of expression at a given place or time. Cf. Frisby, supra, at 482-484; Community for Creative Non-Violence, supra, at 295; Renton n. Playtime Theatres, Inc., 475 U. S., at 53-54. Rather, the guideline continues to permit expressive activity in the bandshell, and has no effect on the quantity or content of that expression beyond regulating the extent of amplification. That the city’s limitations on volume may reduce to some degree the potential audience for respondent’s speech is of no consequence, for there has been no showing that the remaining avenues of communication are inadequate. See Taxpay- WARD v. ROCK AGAINST RACISM 803 781 Marshall, J., dissenting ers for Vincent, supra, at 803, and n. 23, 812, and n. 30; Kovacs, 336 U. S., at 88-89 (opinion of Reed, J.). Ill The city’s sound-amplification guideline is narrowly tailored to serve the substantial and content-neutral governmental interests of avoiding excessive sound volume and providing sufficient amplification within the bandshell concert ground, and the guideline leaves open ample channels of communication. Accordingly, it is valid under the First Amendment as a reasonable regulation of the place and manner of expression. The judgment of the Court of Appeals is Reversed. Justice Blackmun concurs in the result. Justice Marshall, with whom Justice Brennan and Justice Stevens join, dissenting. No one can doubt that government has a substantial interest in regulating the barrage of excessive sound that can plague urban life. Unfortunately, the majority plays to our shared impatience with loud noise to obscure the damage that it does to our First Amendment rights. Until today, a key safeguard of free speech has been government’s obligation to adopt the least intrusive restriction necessary to achieve its goals. By abandoning the requirement that time, place, and manner regulations must be narrowly tailored, the majority replaces constitutional scrutiny with mandatory deference. The majority’s willingness to give government officials a free hand in achieving their policy ends extends so far as to permit, in this case, government control of speech in advance of its dissemination. Because New York City’s Use Guidelines (Guidelines) are not narrowly tailored to serve its interest in regulating loud noise, and because they constitute an impermissible prior restraint, I dissent. 804 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. I The majority sets forth the appropriate standard for assessing the constitutionality of the Guidelines. A time, place, and manner regulation of expression must be content neutral, serve a significant government interest, be narrowly tailored to serve that interest, and leave open ample alternative channels of communication. See Frisby v. Schultz, 487 U. S. 474, 481-482 (1988); Perry Education Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 44 (1983). The Guidelines indisputably are content neutral as they apply to all bandshell users irrespective of the message of their music. App. 375; see Pacific Gas & Electric Co. v. Public Utilities Comm’n of Cal., 475 U. S. 1, 20 (1985).1 They also serve government’s significant interest in limiting loud noise in public places, see Grayned v. Rockford, 408 U. S. 104, 116 (1972), by giving the city exclusive control of all sound equipment. My complaint is with the majority’s serious distortion of the narrow tailoring requirement. Our cases have not, as the majority asserts, “clearly” rejected a less-restrictive-altemative test. Ante, at 797. On the contrary, just last Term, we held that a statute is narrowly tailored only “if it targets and eliminates no more than the exact source of the ‘evil’ it seeks to remedy.” Frisby n. Schultz, supra, at 485. While there is language in a few opinions which, taken out of ‘The majority’s reliance on Renton v. Playtime Theatres, Inc., 475 U. S. 41 (1986), is unnecessary and unwise. That decision dealt only with the unique circumstances of “businesses that purvey sexually explicit materials,” Id., at 49, and n. 2. Today, for the first time, a majority of the Court applies Renton analysis to a category of speech far afield from that decision’s original limited focus. Given the serious threat to free expression posed by Renton analysis, see Boos v. Barry, 485 U. S. 312, 335-337 (1988) (Brennan, J., concurring in part and concurring in judgment); Renton, supra, at 55 (Brennan, J., concurring in part and concurring in judgment), I fear that its broad application may encourage widespread official censorship. WARD v. ROCK AGAINST RACISM 805 781 Marshall, J., dissenting context, supports the majority’s position,2 in practice, the Court has interpreted the narrow tailoring requirement to mandate an examination of alternative methods of serving the asserted governmental interest and a determination whether the greater efficacy of the challenged regulation outweighs the increased burden it places on protected speech. See, e. g., Martin v. Struthers, 319 U. S. 141, 147-148 (1943); Schneider v. State, 308 U. S. 147, 162 (1939). In Schneider, for example, the Court invalidated a ban on handbill distribution on public streets, notwithstanding that it was the most effective means of serving government’s legitimate interest in minimizing litter, noise, and traffic congestion, and in preventing fraud. The Court concluded that punishing those who actually litter or perpetrate frauds was a much less intrusive, albeit not quite as effective, means to serve those significant interests. Id., at 162, 164; see also Martin, supra, at 148 (invalidating ban on door-to-door distribution of handbills because directly punishing fraudulent solicitation was a less intrusive, yet still effective, means of serving government’s interest in preventing fraud).3 * United States v. Albertini, 472 U. S. 675 (1985), for example, involved a person’s right to enter a military base, which, unlike a public park, is not a place traditionally dedicated to free expression. Id., at 687 (commanding officer’s power to exclude civilians from a military base cannot “be analyzed in the same manner as government regulation of a traditional public forum”). Nor can isolated language from Justice White’s opinion in Regan v. Time, Inc., 468 U. S. 641, 657 (1984), which commanded the votes of only three other Justices, be construed as this Court’s definitive explication of the narrow tailoring requirement. 3 The majority relies heavily on Clark v. Community for Creative Non-Violence, 468 U. S. 288 (1984), but in that case, the Court engaged in an inquiry similar to the one the majority now rejects; it considered whether the increased efficacy of the challenged regulation warranted the increased burden on speech. Id., at 299 (“[P]reventing overnight sleeping will avoid a measure of actual or threatened damage”; however, “minimiz[ing] the possible injury by reducing the size, duration, or frequency of demonstrations would still curtail the total allowable expression in which demonstrators could engage”). 806 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. The Court’s past concern for the extent to which a regulation burdens speech more than would a satisfactory alternative is noticeably absent from today’s decision. The majority requires only that government show that its interest cannot be served as effectively without the challenged restriction. Ante, at 799. It will be enough, therefore, that the challenged regulation advances the government’s interest only in the slightest, for any differential burden on speech that results does not enter the calculus. Despite its protestations to the contrary, the majority thus has abandoned the requirement that restrictions on speech be narrowly tailored in any ordinary use of the phrase.4 Indeed, after today’s decision, a city could claim that bans on handbill distribution or on door-to-door solicitation are the most effective means of avoiding littering and fraud, or that a ban on loudspeakers and radios in a public park is the most effective means of avoiding loud noise. Logically extended, the majority’s analysis would permit such far-reaching restrictions on speech. True, the majority states that “[g]overnment may not regulate expression in such a manner that a substantial portion of the burden on speech does not serve to advance its goals.” Ibid. But this means that only those regulations that “engage in the gratuitous inhibition of expression” will be invalidated. Ely, Flag Desecration: A Case Study in the Roles of Categorization and Balancing in First Amendment Analysis, 88 Harv. L. Rev. 1482, 1485 (1975). Moreover, the majority has robbed courts of the necessary analytic tools to make even this limited inquiry. The Court of Appeals examined “how much control of volume is appropriate [and] how that level of control is to be achieved,” ante, at 800, but the majority admonishes that court for doing so, stating that it should 4 In marked contrast, Members of the majority recently adopted a far more stringent narrow tailoring requirement in the affirmative-action context. See Richmond v. J. A. Croson Co., 488 U. S. 469, 507-508 (1989). WARD v. ROCK AGAINST RACISM 807 781 Marshall, J., dissenting have “deferred] to the city’s reasonable determination.” Ibid. The majority thus instructs courts to refrain from examining how much speech may be restricted to serve an asserted interest and how that level of restriction is to be achieved. If a court cannot engage in such inquiries, I am at a loss to understand how a court can ascertain whether the government has adopted a regulation that burdens substantially more speech than is necessary. Had the majority not abandoned the narrow tailoring requirement, the Guidelines could not possibly survive constitutional scrutiny. Government’s interest in avoiding loud sounds cannot justify giving government total control over sound equipment, any more than its interest in avoiding litter could justify a ban on handbill distribution. In both cases, government’s legitimate goals can be effectively and less intrusively served by directly punishing the evil—the persons responsible for excessive sounds and the persons who litter. Indeed, the city concedes that it has an ordinance generally limiting noise but has chosen not to enforce it. See Tr. of Oral. Arg. 5-6.5 By holding that the Guidelines are valid time, place, and manner restrictions, notwithstanding the availability of less intrusive but effective means of controlling volume, the majority deprives the narrow tailoring requirement of all meaning.6 Today, the majority enshrines efficacy but sacrifices free speech. 5 Significantly, the National Park Service relies on the very methods of volume control rejected by the city—monitoring sound levels on the perimeter of an event, communicating with event sponsors, and, if necessary, turning off the power. Brief for United States as Amicus Curiae 21. In light of the Park Service’s “experienc[e] with thousands of events over the years,” ibid., the city’s claims that these methods of monitoring excessive sound are ineffective and impracticable are hard to accept. 6 Because I conclude that the Guidelines are not nairowly tailored, there is no need to consider whether there are ample alternative channels for communication. I note only that the availability of alternative channels of communication outside a public park does not magically validate a 808 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. II The majority’s conclusion that the city’s exclusive control of sound equipment is constitutional is deeply troubling for another reason. It places the Court’s imprimatur on a quintessential prior restraint, incompatible with fundamental First Amendment values. See Near v. Minnesota ex rel. Olson, 283 U. S. 697 (1931). Indeed, just as “[m]usic is one of the oldest forms of human expression,” ante, at 790, the city’s regulation is one of the oldest forms of speech repression. In 16th- and 17th-century England, government controlled speech through its monopoly on printing presses. See L. Levy, Emergence of a Free Press 6 (1985). Here, the city controls the volume and mix of sound through its monopoly on sound equipment. In both situations, government’s exclusive control of the means of communication enables public officials to censor speech in advance of its expression. See Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546, 553 (1975). Under more familiar prior restraints, government officials censor speech “by a simple stroke of the pen,” Emerson, The Doctrine of Prior Restraint, 20 Law & Contemp. Prob. 648, 657 (1955). Here, it is done by a single turn of a knob. The majority’s implication that government control of sound equipment is not a prior restraint because city officials do not “enjoy unguided discretion to deny the right to speak altogether,” ante, at 794, is startling. In the majority’s view, this case involves a question of “different and lesser” magnitude—the discretion to provide inadequate sound for performers. But whether the city denies a performer a bandshell permit or grants the permit and then silences or government restriction on protected speech within it. See Southeastern Promotions, Ltd. v. Conrad, 420 U. S. 546, 556 (1975) (“‘[O]ne is not to have the exercise of his liberty of expression in appropriate places abridged on the plea that it may be exercised in some other place,’” quoting Schneider v. State, 308 U. S. 147, 163 (1939)). WARD v. ROCK AGAINST RACISM 809 781 Marshall, J., dissenting distorts the performer’s music, the result is the same—the city censors speech. In the words of Chief Justice Rehnquist, the First Amendment means little if it permits government to “allo[w] a speaker in a public hall to express his views while denying him the use of an amplifying system.” FEC v. National Conservative Political Action Committee, 470 U. S. 480, 493 (1985); see also Southeastern Promotions, supra, at 556, n. 8 (“A licensing system need not effect total suppression in order to create a prior restraint”). As a system of prior restraint, the Guidelines are presumptively invalid. See Southeastern Promotions, supra, at 558; Bantam, Books, Inc. v. Sullivan, 372 U. S. 58, 70 (1963). They may be constitutional only if accompanied by the procedural safeguards necessary “to obviate the dangers of a censorship system.” Freedman v. Maryland, 380 U. S. 51, 58 (1965). The city must establish neutral criteria embodied in “narrowly drawn, reasonable and definite standards,” in order to ensure that discretion is not exercised based on the content of speech. Niemotko v. Maryland, 340 U. S. 268, 271 (1951); see also Lakewood v. Plain Dealer Publishing Co., 486 U. S., 750, 758 (1988); Shuttlesworth v. Birmingham, 394 U. S. 147, 150-151 (1969). Moreover, there must be “an almost immediate judicial determination” that the restricted material was unprotected by the First Amendment. Bantam Books, supra, at 70; see also Southeastern Promotions, supra, at 560. The Guidelines contain neither of these procedural safeguards. First, there are no “narrowly drawn, reasonable and definite standards” guiding the hands of the city’s sound technician as he mixes the sound. The Guidelines state that the goals are “to provide the best sound for all events” and to “insure appropriate sound quality balanced with respect for nearby residential neighbors and the mayoralty decreed quiet zone.” App. 375; see also ante, at 794. But the city never defines “best sound” or “appropriate sound quality.” The bandshell program director-manager testified that quality of 810 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. sound refers to tone and to sound mix. App. 229, 230. Yet questions of tone and mix cannot be separated from musical expression as a whole. See The New Grove Dictionary of Music and Musicians 51-55 (S. Sadie ed. 1980) (tonality involves relationship between pitches and harmony); F. Everest, Successful Sound System Operation 173 (1985) (“The mixing console . . . must be considered as a creative tool”). Because judgments that sounds are too loud, noiselike, or discordant can mask disapproval of the music itself,7 government control of the sound-mixing equipment necessitates detailed and neutral standards. The majority concedes that the standards in the Guidelines are “undoubtedly flexible” and that “the officials implementing them will exercise considerable discretion.” Ante, at 794. Nevertheless, it concludes that “[b]y its own terms the city’s sound-amplification guideline must be interpreted to forbid city officials purposefully to select inadequate sound systems or to vary the sound quality or volume based on the message being delivered by performers.” Ante, at 794-795. Although the majority wishes it were so, the language of the Guidelines simply does not support such a limitation on the city’s discretion. Alternatively, the majority finds a limitation in the city’s practice of deferring to the sponsor with respect to sound mix, and of conferring “with the sponsor if any questions of excessive sound arise, before taking any corrective action.” 658 F. Supp. 1346, 1352 (SDNY 1987). A promise to consult, however, does not provide the detailed 7 “New music always sounds loud to old ears. Beethoven seemed to make more noise than Mozart; Liszt was noisier than Beethoven; Schoenberg and Stravinsky, noisier than any of their predecessors.” N. Slonim-sky, Lexicon of Musical Invective: Critical Assaults on Composers Since Beethoven’s Time 18 (1953). One music critic wrote of Prokofiev: “Those who do not believe that genius is evident in superabundance of noise, looked in vain for a new musical message in Mr. Prokofiev’s work. Nor in the Classical Symphony, which the composer conducted, was there any cessation from the orgy of discordant sounds.” Id., at 5 (internal quotations omitted). WARD v. ROCK AGAINST RACISM 811 781 Marshall, J., dissenting “neutral criteria” necessary to prevent future abuses of discretion any more than did the city’s promise in Lakewood to deny permit applications only for reasons related to the health, safety, or welfare of Lakewood citizens. Indeed, a presumption that city officials will act in good faith and adhere to standards absent from a regulation’s face is “the very presumption that the doctrine forbidding unbridled discretion disallows.” Lakewood, supra, at 770.8 Second, even if there were narrowly drawn guidelines limiting the city’s discretion, the Guidelines would be fundamentally flawed. For the requirement that there be detailed standards is of value only so far as there is a judicial mechanism to enforce them. Here, that necessary safeguard is absent. The city’s sound technician consults with the performers for several minutes before the performance and then decides how to present each song or piece of music. During the performance itself, the technician makes hundreds of decisions affecting the mix and volume of sound. Tr. of Oral Arg. 13. The music is played immediately after each decision. There is, of course, no time for appeal in the middle of a song. As a result, no court ever determines that a particular restraint on speech is necessary. The city’s admission that it does not impose sanctions on violations of its general sound ordinance because the necessary litigation is too costly and time consuming only underscores its contempt for the need for judicial review of restrictions on speech. Id., at 5. With neither prompt judicial review nor detailed and neutral standards fettering the city’s discretion to restrict protected 8 Of course, if the city always defers to a performer’s wishes in sound mixing, then it is difficult to understand the need for a city technician to operate the mixing console. See Tr. of Oral. Arg. 12 (city concedes that the possibilities for a confrontation over volume are the same whether the city technician directly controls the mixing console or sits next to a performer’s technician who operates the equipment). Conversely, if the city can control sound only by using its own equipment and technician, then it must not be heeding all the performer’s wishes on sound mixing. 812 OCTOBER TERM, 1988 Marshall, J., dissenting 491 U. S. speech, the Guidelines constitute a quintessential, and unconstitutional, prior restraint. Ill Today’s decision has significance far beyond the world of rock music. Government no longer need balance the effectiveness of regulation with the burdens on free speech. After today, government need only assert that it is most effective to control speech in advance of its expression. Because such a result eviscerates the First Amendment, I dissent. Reporter’s Note The next page is purposely numbered 901. The numbers between 812 and 901 were intentionally omitted, in order to make it possible to publish the orders with permanent page numbers, thus making the official citations available upon publication of the preliminary prints of the United States Reports. ORDERS FOR JUNE 15 THROUGH JUNE 21, 1989 June 15, 1989 Miscellaneous Order No. A-996 (88-7461). Woomer v. Evatt, Commissioner, South Carolina Department of Corrections, et al. Sup. Ct. S. C. Application for stay of execution of sentence of death, presented to The Chief Justice, and by him referred to the Court, granted pending the disposition by this Court of the petition for writ of certiorari. Should the petition for writ of certiorari be denied, this stay terminates automatically. In the event the petition for writ of certiorari is granted, this stay shall continue pending the sending down of the judgment of this Court. Certiorari Denied No. 88-7447 (A-995). Edwards v. Mississippi. Sup. Ct. Miss. Application for stay of execution of sentence of death, presented to Justice White, and by him referred to the Court, denied. Certiorari denied. Justice Brennan and Justice Marshall, dissenting. Adhering to our views that the death penalty is in all circumstances cruel and unusual punishment prohibited by the Eighth and Fourteenth Amendments, Gregg v. Georgia, 428 U. S. 153, 227, 231 (1976), we would grant the application for stay of execution and the petition for writ of certiorari and would vacate the death sentence in this case. June 19, 1989 Dismissal Under Rule 53 No. 88-1794. Northeast Illinois Regional Commuter Railroad Corp. v. Baker. C. A. 7th Cir. Certiorari dismissed under this Court’s Rule 53. Appeal Dismissed No. 88-1753. S. T., on Behalf of her Minor Child, N. T. v. Board of Education of the City of Millville, New Jer-901 902 OCTOBER TERM, 1988 June 19, 1989 491 U. S. SEY, et al. Appeal from Super. Ct. N. J., App. Div., dismissed for want of jurisdiction. Treating the papers whereon the appeal was taken as a petition for writ of certiorari, certiorari denied. Certiorari Granted—Vacated and Remanded No. 88-5795. Salazar et al. v. United States. C. A. 2d Cir. Motion of petitioner for leave to proceed in forma pauperis granted. Certiorari granted, judgment vacated, and case remanded for further consideration in light of Gomez v. United States, 490 U. S. 858 (1989). Reported below: 848 F. 2d 1324. No. 88-6122. Nittayanupap v. United States. C. A. 9th Cir. Motion of petitioner for leave to proceed in forma pauperis granted. Certiorari granted, judgment vacated, and case remanded for further consideration in light of Gomez v. United States, 490 U. S. 858 (1989). Reported below: 860 F. 2d 1090. No. 88-6725. Hayes v. Lockhart, Director, Arkansas Department of Correction. C. A. 8th Cir. Motion of petitioner for leave to proceed in forma pauperis granted. Certiorari granted, judgment vacated, and case remanded for further consideration in light of South Carolina v. Gathers, 490 U. S. 805 (1989). Reported below: 852 F. 2d 339. No. 88-6729. Daniels v. Indiana. Sup. Ct. Ind. Motion of petitioner for leave to proceed in forma pauperis granted. Certiorari granted, judgment vacated, and case remanded for further consideration in light of South Carolina v. Gathers, 490 U. S. 805 (1989). Reported below: 528 N. E. 2d 775. Miscellaneous Orders No.-----------. In re Dennis. Motion to direct the Clerk to file petition for writ of habeas corpus denied. No. A-944. Prevenslik, Individually and as Trustee of the Felix and Olga Davis Trust v. Rossko et al. C. A. 3d Cir. Application for stay, addressed to Justice Kennedy and referred to the Court, denied. No. D-784. In re Disbarment of Lillback. Disbarment entered. [For earlier order herein, see 490 U. S. 1044.] No. D-792. In re Disbarment of Schaefer. It is ordered that Brian D. Schaefer, of Longwood, Fla., be suspended from the practice of law in this Court and that a rule issue, returnable ORDERS 903 491 U. S. June 19, 1989 within 40 days, requiring him to show cause why he should not be disbarred from the practice of law in this Court. No. D-793. In re Disbarment of Cartwright. It is ordered that Prince Cartwright, Jr., of Houston, Tex., be suspended from the practice of law in this Court and that a rule issue, returnable within 40 days, requiring him to show cause why he should not be disbarred from the practice of law in this Court. No. D-794. In re Disbarment of Montemayor. It is ordered that Ruben R. Montemayor, of San Antonio, Tex., be suspended from the practice of law in this Court and that a rule issue, returnable within 40 days, requiring him to show cause why he should not be disbarred from the practice of law in this Court. No. 88-1353. United States v. Verdugo-Urquidez. C. A. 9th Cir. [Certiorari granted, 490 U. S. 1019.] Motion of Criminal Justice Legal Foundation for leave to file a brief as amicus curiae granted. No. 88-1640. Michigan Citizens for an Independent Press et al. v. Thornburgh, Attorney General of the United States, et al. C. A. D. C. Cir. [Certiorari granted, 490 U. S. 1045.] Motion of the parties to dispense with printing the joint appendix granted. Justice White took no part in the consideration or decision of this motion. No. 88-7281. In re Humphrey. Petition for writ of habeas corpus denied. Certiorari Granted No. 88-1719. Chauffeurs, Teamsters & Helpers, Local No. 391 v. Terry et al. C. A. 4th Cir. Certiorari granted. Reported below: 863 F. 2d 334. No. 88-1835. Florida v. Wells. Sup. Ct. Fla. Certiorari granted. Reported below: 539 So. 2d 464. No. 88-1595. Kaiser Aluminum & Chemical Corp, et al. v. Bonjorno et al.; and No. 88-1771. Bonjorno et al. v. Kaiser Aluminum & Chemical Corp, et al. C. A. 3d Cir. Certiorari granted, cases consolidated, and a total of one hour allotted for oral argument. Reported below: 865 F. 2d 566. 904 OCTOBER TERM, 1988 June 19, 1989 491 U. S. No. 88-6873. Clemons v. Mississippi. Sup. Ct. Miss. Motion of petitioner for leave to proceed in forma pauperis granted. Certiorari granted limited to Question 1 presented by the petition. Reported below: 535 So. 2d 1354. Certiorari Denied. (See also No. 88-1753, supra.) No. 87-1445. Hampton, Personal Representative of the Estate of Hampton v. Michigan. Ct. App. Mich. Certiorari denied. Reported below: 144 Mich. App. 794, 377 N. W. 2d 920. No. 87-1454. Skoblow v. Ameri-Manage, Inc. Sup. Ct. Fla. Certiorari denied. Reported below: 514 So. 2d 1077. No. 88-447. Vest v. Schafer et al. Sup. Ct. Alaska. Certiorari denied. Reported below: 757 P. 2d 588. No. 88-527. Bishop et ux., Administrators of the Estate of Bishop v. Gwaltney et al. Ct. App. Cal., 4th App. Dist. Certiorari denied. No. 88-542. Jackson v. Maine. Sup. Jud. Ct. Me. Certiorari denied. Reported below: 544 A. 2d 291. No. 88-660. Simkins Industries, Inc. v. Sierra Club. C. A. 4th Cir. Certiorari denied. Reported below: 847 F. 2d 1109. No. 88-709. Roe v. Superior Court of California, County of Marin (Doe, Real Party in Interest). Ct. App. Cal., 1st App. Dist. Certiorari denied. No. 88-1254. Schoo v. United States Postal Service et AL. C. A. 4th Cir. Certiorari denied. Reported below: 865 F. 2d 1259. No. 88-1382. Eagle-Picher Industries, Inc. v. United States; and No. 88-1418. Raymark Industries, Inc. v. United States. C. A. Fed. Cir. Certiorari denied. Reported below: 858 F. 2d 712. No. 88-1412. Randall et al. v. Thornburgh, Attorney General of the United States, et al. C. A. D. C. Cir. Certiorari denied. Reported below: 272 U. S. App. D. C. 63, 854 F. 2d 472. ORDERS 905 491 U. S. June 19, 1989 No. 88-1471. Walker v. Superior Court of Sacramento County. Sup. Ct. Cal. Certiorari denied. Reported below: 47 Cal. 3d 112, 763 P. 2d 852. No. 88-1571. DeLorenzo et al. v. New Jersey Department of Environmental Protection. Sup. Ct. N. J. Certiorari denied. Reported below: 113 N. J. 649, 552 A. 2d 172. No. 88-1604. Leach et al. v. Federal Deposit Insurance Corporation et al. C. A. 5th Cir. Certiorari denied. Reported below: 860 F. 2d 1266. No. 88-1616. Lombardo v. United States. C. A. 7th Cir. Certiorari denied. Reported below: 865 F. 2d 155. No. 88-1626. Sams v. United States. C. A. 6th Cir. Certiorari denied. Reported below: 865 F. 2d 713. No. 88-1642. Alaska v. Kenaitze Indian Tribe. C. A. 9th Cir. Certiorari denied. Reported below: 860 F. 2d 312. No. 88-1649. L & B Corp, et al. v. Commissioner of Internal Revenue. C. A. 8th Cir. Certiorari denied. Reported below: 862 F. 2d 667. No. 88-1657. S. H. et al. v. Edwards, Former Commissioner, Georgia Department of Human Resources, et al. C. A. 11th Cir. Certiorari denied. Reported below: 860 F. 2d 1045.' No. 88-1669. Hamilton et al. v. Stovall, Judge, Second Administrative Judicial District of Texas. C. A. 5th Cir. Certiorari denied. Reported below: 859 F. 2d 919. No. 88-1678. Walton v. Merriman. C. A. 9th Cir. Certiorari denied. Reported below: 856 F. 2d 1333. No. 88-1696. Sales Tax Collector, St. Charles Parish, Louisiana v. Westside Sand Co., Inc. Ct. App. La., 5th Cir. Certiorari denied. Reported below: 534 So. 2d 454. No. 88-1700. Central Louisiana Telephone Co. et al. v. McCarroll. Ct. App. La., 3d Cir. Certiorari denied. Reported below: 533 So. 2d 385. No. 88-1715. Woods et vir v. Bobrow, Greenapple & Skolnik. C. A. 2d Cir. Certiorari denied. Reported below: 865 F. 2d 43. 906 OCTOBER TERM, 1988 June 19, 1989 491 U. S. No. 88-1718. Schoemehl v. Edwards. Sup. Ct. Mo. Certiorari denied. Reported below: 765 S. W. 2d 607. No. 88-1720. Wilsmann v. Upjohn Co. et al. C. A. 6th Cir. Certiorari denied. Reported below: 865 F. 2d 1269. No. 88-1727. Video International Productions, Inc. v. Warner-Amex Cable Communications, Inc., et al. C. A. 5th Cir. Certiorari denied. Reported below: 858 F. 2d 1075. No. 88-1730. Attorneys’ Title Guaranty Fund, Inc. v. Richards. C. A. 10th Cir. Certiorari denied. Reported below: 866 F. 2d 1570. No. 88-1731. District of Columbia Nurses Assn, et al. v. District of Columbia. C. A. D. C. Cir. Certiorari denied. Reported below: 272 U. S. App. D. C. 231, 854 F. 2d 1448. No. 88-1736. Wemhoff v. Floria et al. Ct. App. D. C. Certiorari denied. No. 88-1740. Karzov v. Attorney Registration and Disciplinary Commission of Illinois. Sup. Ct. Ill. Certiorari denied. Reported below: 126 Ill. 2d 33, 533 N. E. 2d 856. No. 88-1746. Macurdy v. Pennsylvania. Super. Ct. Pa. Certiorari denied. Reported below: 381 Pa. Super. 657, 548 A. 2d 640. No. 88-1747. Grammer v. Patterson Services, Inc. C. A. 5th Cir. Certiorari denied. Reported below: 860 F. 2d 639. No. 88-1748. Tasher v. St. Tammany Parish Hospital et al. C. A. 5th Cir. Certiorari denied. Reported below: 862 F. 2d 873. No. 88-1749. Velez v. Connecticut. App. Ct. Conn. Certiorari denied. Reported below: 17 Conn. App. 186, 551 A. 2d 421. No. 88-1750. Daisley v. General Electric Co. C. A. 4th Cir. Certiorari denied. Reported below: 865 F. 2d 1257. No. 88-1752. South Carolina et al. v. Catawba Indian Tribe of South Carolina, Inc. C. A. 4th Cir. Certiorari denied. Reported below: 865 F. 2d 1444. No. 88-1759. United States v. Mandel et al. C. A. 4th Cir. Certiorari denied. Reported below: 862 F. 2d 1067. ORDERS 907 491 U. S. June 19, 1989 No. 88-1760. Worrell v. B. F. Goodrich Co. C. A. 9th Cir. Certiorari denied. Reported below: 845 F. 2d 840. No. 88-1780. Cooper v. Eisenman et al. C. A. 1st Cir. Certiorari denied. Reported below: 867 F. 2d 606. No. 88-1800. In re Powell. Sup. Ct. Ill. Certiorari denied. Reported below: 126 Ill. 2d 15, 533 N. E. 2d 831. No. 88-1810. Steffen v. Meridian Life Insurance Co. et al. C. A. 7th Cir. Certiorari denied. Reported below: 859 F. 2d 534. No. 88-1817. Barrow v. Wahl et al. C. A. 11th Cir. Certiorari denied. Reported below: 868 F. 2d 1274. No. 88-1823. Gladden v. Roach et al. C. A. 5th Cir. Certiorari denied. Reported below: 864 F. 2d 1196. No. 88-1853. Sterner v. Commissioner of Internal Revenue. C. A. 4th Cir. Certiorari denied. Reported below: 867 F. 2d 609. No. 88-1854. Davenport v. United States. C. A. 6th Cir. Certiorari denied. Reported below: 869 F. 2d 1489. No. 88-1869. Ayala v. United States. C. A. 9th Cir. Certiorari denied. Reported below: 872 F. 2d 430. No. 88-1876. Furnari v. United States. C. A. 2d Cir. Certiorari denied. Reported below: 868 F. 2d 524. No. 88-1888. Norton v. United States. C. A. 11th Cir. Certiorari denied. Reported below: 867 F. 2d 1354. No. 88-6582. Dudley v. Missouri. Ct. App. Mo., Western Dist. Certiorari denied. Reported below: 760 S. W. 2d 448. No. 88-6636. Embry v. Kentucky. Sup. Ct. Ky. Certiorari denied. No. 88-6913. Bruscino et al. v. Carlson et al. C. A. 7th Cir. Certiorari denied. Reported below: 854 F. 2d 162. No. 88-6917. Meade-Murphy v. City of Atlanta et al. C. A. 11th Cir. Certiorari denied. No. 88-6929. Wagstaff v. United States. C. A. 4th Cir. Certiorari denied. Reported below: 865 F. 2d 626. No. 88-6935. Leech v. United States. C. A. 3d Cir. Certiorari denied. Reported below: 869 F. 2d 592. 908 OCTOBER TERM, 1988 June 19, 1989 491 U. S. No. 88-6981. Rumney v. United States. C. A. 1st Cir. Certiorari denied. Reported below: 867 F. 2d 714. No. 88-7076. Halverson v. Barbour, Superintendent, Twin Rivers Correctional Center at Monroe, Washington. C. A. 9th Cir. Certiorari denied. Reported below: 868 F. 2d 1272. No. 88-7083. Williams v. Dugger, Secretary, Florida Department of Corrections, et al. C. A. 11th Cir. Certiorari denied. No. 88-7084. Patterson v. Transworld Drilling Co. C. A. 5th Cir. Certiorari denied. Reported below: 866 F. 2d 1418. No. 88-7092. Brooks v. City of Anniston, Alabama, et al. C. A. 11th Cir. Certiorari denied. No. 88-7094. Harper v. Nix, Warden, et al. C. A. 8th Cir. Certiorari denied. Reported below: 867 F. 2d 455. No. 88-7095. Hoffman v. Petsock, Superintendent, State Correctional Institution at Pittsburgh, et al. C. A. 3d Cir. Certiorari denied. No. 88-7102. Ross v. Zimmerman, Superintendent, State Correctional Institution (and Diagnostic and Classification Center) at Graterford, et al. C. A. 3d Cir. Certiorari denied. No. 88-7103. Martin v. Court of Appeals of Maryland et AL. C. A. 4th Cir. Certiorari denied. Reported below: 870 F. 2d 655. No. 88-7105. Dean v. West Virginia et al. C. A. 4th Cir. Certiorari denied. Reported below: 866 F. 2d 1415. No. 88-7115. Johnson v. District Court of Wyoming, First Judicial District. Sup. Ct. Wyo. Certiorari denied. No. 88-7117. Walker v. Ohio. Ct. App. Ohio, Stark County. Certiorari denied. No. 88-7141. Al-Hakim v. Tampa City Council. C. A. 11th Cir. Certiorari denied. Reported below: 871 F. 2d 123. No. 88-7150. Velilla v. Connecticut et al. C. A. D. C. Cir. Certiorari denied. ORDERS 909 491 U. S. June 19, 1989 No. 88-7195. F. G. v. Lawton et al. Sup. Ct. Mo. Certiorari denied. Reported below: 764 S. W. 2d 89. No. 88-7237. Moser v. United States. C. A. 6th Cir. Certiorari denied. Reported below: 870 F. 2d 658. No. 88-7238. Duffel v. United States. C. A. 6th Cir. Certiorari denied. Reported below: 869 F. 2d 1489. No. 88-7248. Naiyavaj v. United States. C. A. 9th Cir. Certiorari denied. Reported below: 869 F. 2d 1499. No. 88-7255. Thorpe v. United States. Ct. App. D. C. Certiorari denied. No. 88-7258. Kosko v. United States. C. A. 4th Cir. Certiorari denied. Reported below: 870 F. 2d 162. No. 88-7260. Gonzalez-Alvarez v. United States. C. A. 11th Cir. Certiorari denied. Reported below: 866 F. 2d 1422. No. 88-7264. Fahy v. United States. C. A. Fed. Cir. Certiorari denied. Reported below: 864 F. 2d 148. No. 88-7265. Cardenas, aka Ramirez v. United States. C. A. 10th Cir. Certiorari denied. Reported below: 864 F. 2d 1528. No, 88-7267. Klein v. United States. C. A. 6th Cir. Certiorari denied. Reported below: 861 F. 2d 720. No. 88-7268. Campos-Fuentes v. United States. C. A. 5th Cir. Certiorari denied. Reported below: 871 F. 2d 119. No. 88-7269. Rivera v. United States. C. A. 9th Cir. Certiorari denied. Reported below: 865 F. 2d 1271. No. 88-7270. Wright v. United States. C. A. 9th Cir. Certiorari denied. Reported below: 872 F. 2d 432. No. 88-7272. Stengel v. United States. C. A. 9th Cir. Certiorari denied. Reported below: 872 F. 2d 431. No. 88-7275. Trousdale v. United States. C. A. 11th Cir. Certiorari denied. Reported below: 872 F. 2d 434. No. 88-1466. People of Enewetak, Rongelap, and Other Marshall Islands Atolls v. United States. C. A. Fed. Cir. Motion of petitioners to defer consideration of the petition for writ 910 OCTOBER TERM, 1988 June 19, 1989 491 U. S. of certiorari denied. Certiorari denied. Reported below: 864 F. 2d 134. No. 88-1467. Credit Bureau Services-New Orleans, dba Chilton Corp. v. Hyde. C. A. 5th Cir. Motion of Associated Credit Bureaus, Inc., for leave to file a brief as amicus curiae granted. Certiorari denied. Reported below: 861 F. 2d 446. No. 88-1721. McCormick et al. v. Texas Commerce Bank, N. A. Ct. App. Tex., 14th Dist. Motion of respondent for costs denied. Certiorari denied. Reported below: 751 S. W. 2d 887. No. 88-1724. Arkansas v. Gibson. Sup. Ct. Ark. Certiorari denied. Justice Brennan and Justice White would grant certiorari. Reported below: 298 Ark. 43, 764 S. W. 2d 617. No. 88-1773. Ricoh Co., Ltd. v. Snellman, dba Norfin. C. A. Fed. Cir. Motion of International Electronics Manufacturers & Consumers of America, Inc., for leave to file a brief as amicus curiae granted. Certiorari denied. Reported below: 862 F. 2d 283. No. 88-1813. Raley v. Hughes et ux. Ct. Sp. App. Md. Motion of respondents for sanctions and attorney’s fees denied. Certiorari denied. Reported below: 76 Md. App. 796. No. 88-5024. No. 88-5685. No. 88-6006. No. 88-6154. No. 88-6468. No. 88-7073. No. 88-7120. No. 88-7132. No. 88-7135. No. 88-7136. Tenn.; and No. 88-7201. Travaglia v. Pennsylvania. Super. Ct. Pa.; Rice v. Washington. Sup. Ct. Wash.; Dawson v. Montana. Sup. Ct. Mont.; Beaty v. Arizona. Sup. Ct. Ariz.; Yarris v. Pennsylvania. Sup. Ct. Pa.; Thomas v. kLKRAMK. Sup. Ct. Ala.; Hernandez v. California. Sup. Ct. Cal.; Hoke v. Virginia. Sup. Ct. Va.; Webb v. Texas. Ct. Crim. App. Tex.; Caruthers v. Tennessee. Ct. Crim. App. Waye v. Townley, Warden. C. A. 4th Cir. Certiorari denied. Reported below: No. 88-5024, 359 Pa. Super. 630, 515 A. 2d 620; No. 88-5685, 110 Wash. 2d 577, 757 P. 2d 889; No. 88-6006, 233 Mont. 345, 761 P. 2d 352; No. 88-6154, 158 Ariz. 232, 762 P. 2d 519; No. 88-6468, 519 Pa. 571, 549 A. 2d 513; No. 88-7073, 539 Sc. 2d 399; No. 88-7120, 47 Cal. 3d ORDERS 911 491 U. S. June 19, 20, 1989 315, 763 P. 2d 1289; No. 88-7132, 237 Va. 303, 377 S. E. 2d 595; No. 88-7135, 760 S. W. 2d 263; No. 88-7201, 871 F. 2d 18. Justice Brennan and Justice Marshall, dissenting. Adhering to our views that the death penalty is in all circumstances cruel and unusual punishment prohibited by the Eighth and Fourteenth Amendments, Gregg n. Georgia, 428 U. S. 153, 227, 231 (1976), we would grant certiorari and vacate the death sentences in these cases. No. 88-7082. Solon v. United States. C. A. 5th Cir. Certiorari before judgment denied. Rehearing Denied No. 88-963. Culberson v. Veterans Administration, 490 U. S. 1034; No. 88-1011. Harman et al. v. Dole, Secretary of Labor, et al., 489 U. S. 1094; No. 88-1517. Dunning v. Commissioner of Internal Revenue, 490 U. S. 1047; No. 88-6515. Brathwaite v. United States, 490 U. S. 1048; No. 88-6726. Asberry v. United States Postal Service et AL., 490 U. S. 1037; No. 88-6734. Adamo v. Hotel, Motel, Bartenders, Cooks & Restaurant Workers Union Local 24 et al., 490 U. S. 1037; No. 88-6773. Barron v. Salt Lake City, Utah, 490 U. S. 1049; No. 88-6859. Gonzalez v. Department of the Navy, 490 U. S. 1050; and No. 88-6934. Reidt v. United States, 490 U. S. 1073. Petitions for rehearing denied. June 20, 1989 Miscellaneous Orders No. A-1008 (88-7235). Gilmore v. Armontrout, Warden, 490 U. S. 1114. Motion of respondent to vacate the order staying the execution of sentence of death entered by Justice Blackmun on June 16, 1989, granted. Justice Blackmun, dissenting. In an order issued June 12, 1989, 490 U. S. 1114, this Court denied George Gilmore’s petition for a writ of certiorari to the 912 OCTOBER TERM, 1988 Blackmun, J., dissenting 491 U. S. United States Court of Appeals for the Eighth Circuit, which had reversed the United States District Court’s grant of relief under 28 U. S. C. § 2254. On the following day the Chief Justice of the Missouri Supreme Court set Gilmore’s execution for 12:01 a.m., Wednesday, June 21, 1989, less than nine days after our order issued. On Friday, June 16, I issued a stay of execution to allow Gilmore to file a timely petition for rehearing with this Court. The State of Missouri moved later that day to vacate the stay, and today the Court has granted that motion. In doing so it has endorsed the State of Missouri’s unseemly rush to execution and has sanctioned the decision of the Chief Justice of Missouri to ignore the usual deadline of 25 days set by this Court for filing petitions for rehearing. See this Court’s Rule 51.1. Although I am far from certain that the claims presented by Gilmore would have moved the Court to grant rehearing, the claims are far from frivolous, as the District Court’s grant of habeas relief makes clear. The Court’s action today means that we shall never have the opportunity to consider those arguments as they might have been developed by counsel during the 25 days normally allowed for petitions for rehearing to be filed. I granted the stay which the Court vacates today because the initial application contained “only a synopsis of the arguments that counsel intend[ed] to make” in the petition for rehearing. Autry v. Estelle, 464 U. S. 1, 4 (1983) (Stevens, J., dissenting). Whether the relevant time period is that for filing a petition for a writ of certiorari, as was the case in Autry, or for a petition for rehearing, as in this case, our time limits are based upon the expectation that counsel needs the amount of time specified by our Rules in order to develop legal arguments that fit the specific requirements of this Court. Here, rather than just repeating the arguments contained in his petition for certiorari, Gilmore’s counsel was required to limit the grounds of the petition “to intervening circumstances of substantial or controlling effect or to other substantial grounds not previously presented.” This Court’s Rule 51.2. It is unrealistic to think that even the most resourceful counsel would be able to do this in less than nine days. Forcing Gilmore to present his claims in such a fashion “injects uncertainty and disparity into the review procedure, adds to the burden of counsel, distorts the deliberative process within this Court, and increases ORDERS 913 491 U. S. June 20, 1989 the risk of error.” Autry, supra, at 6. That was my view when I granted the stay in this case; it remains my view today. Justice Stevens, with whom Justice Brennan and Justice Marshall join, dissenting. For the reasons stated by Justice Blackmun, I would deny the motion to vacate the stay. No. A-1012. McKinney v. Idaho. Application for stay of execution of sentence of death, presented to Justice O’Connor, and by her referred to the Court, denied. Justice Brennan and Justice Marshall, dissenting. Adhering to our views that the death penalty is in all circumstances cruel and unusual punishment prohibited by the Eighth and Fourteenth Amendments, Gregg v. Georgia, 428 U. S. 153, 227, 231 (1976), we would grant the application for stay of execution in order to give the applicant time to file a petition for writ of certiorari and would grant the petition and vacate the death sentence in this case. No. A-1025. Laws, by and Through Laws, as his Next Friend v. Delo, Superintendent, Potosi Correctional Center. Application for stay of execution of sentence of death, presented to Justice Blackmun, and by him referred to the Court, granted pending the timely filing and disposition by this Court of a petition for writ of certiorari. Should the petition for writ of certiorari be denied, this stay terminates automatically. In the event the petition for writ of certiorari is granted, this stay shall continue pending the sending down of the judgment of this Court. No. A-1027. Delo, Superintendent, Potosi Correctional Center v. Gilmore. Application of the Attorney General of Missouri for an order to vacate the stay of execution of sentence of death entered by the United States District Court for the Eastern District of Missouri, presented to Justice Blackmun, and by him referred to the Court, denied. Certiorari Denied No. 88-7530 (A-1021). Edwards v. Black, Commissioner, Mississippi Department of Corrections. C. A. 5th Cir. Application for stay of execution of sentence of death, presented to 914 OCTOBER TERM, 1988 June 20, 21, 1989 491 U. S. Justice White, and by him referred to the Court, denied. Certiorari denied. Reported below: 876 F. 2d 377. Justice Brennan and Justice Marshall, dissenting. Adhering to our views that the death penalty is in all circumstances cruel and unusual punishment prohibited by the Eighth and Fourteenth Amendments, Gregg v. Georgia, 428 U. S. 153, 227, 231 (1976), we would grant the application for stay of execution and the petition for writ of certiorari and would vacate the death sentence in this case. June 21, 1989 Dismissal Under Rule 53 No. 88-1952. United States v. Trice et al. C. A. 8th Cir. Certiorari dismissed under this Court’s Rule 53. Reported below: 864 F. 2d 1421. INDEX ABSTENTION. Utility rates—State review process. —Where utility petitioned Federal District Court for declaratory and injunctive relief from state ratemaking authority’s order, court erred in abstaining from exercising jurisdiction in deference to state review process. New Orleans Public Service, Inc. v. New Orleans, p. 350. ACTUAL MALICE. See Libel. ADVISORY COMMITTEES. See Federal Advisory Committee Act; Jurisdiction, 1. AMERICAN BAR ASSOCIATION. See Federal Advisory Committee Act; Jurisdiction, 1. AMPLIFIED-MUSIC REGULATION. See Constitutional Law, VI, 2. ARBITRATION. See Railway Labor Act, 1. ASSETS FROM DRUG TRAFFICKING AS FORFEITABLE. See Comprehensive Forfeiture Act of 1984; Constitutional Law, III, 1; VIII. ATTORNEY-CLIENT PRIVILEGE. Internal Revenue Service summons—Crime-fraud exception—In camera review of communications. — Where respondents asserted privilege as a bar to production of tapes pursuant to an IRS summons, in camera review of allegedly privileged attorney-client communications could be used to determine whether communications fell within crime-fraud exception. United States v. Zolin, p. 554. ATTORNEY’S FEES. See Civil Rights Act of 1964; Civil Rights Attorney’s Fees Awards Act of 1976; Comprehensive Forfeiture Act of 1984; Constitutional Law, III, 1; VIII; IX, 1. BEER PRICES. See Constitutional Law, I, 2. BURDEN OF PROOF. See Civil Rights Act of 1866, 1. BURNING OF AMERICAN FLAG. See Constitutional Law, V. CALIFORNIA. See Constitutional Law, III, 3. CEDING COMMISSIONS. See Taxes. 915 916 INDEX CHILDREN’S PATERNITY DETERMINATIONS. See Constitutional Law, III, 3. CIVIL RIGHTS ACT OF 1866. 1. Racial discrimination actionable under 42 U. S. C. § 1981—Burden of proof in discriminatory promotion claim. — Where petitioner alleged that her employer had harassed her, failed to promote her, and then discharged her, all because of her race, stare decisis compels adherence to decision in Runyon v. McCrary, 427 U. S. 160, that § 1981 prohibits racial discrimination in making and enforcement of private contracts; racial harassment relating to conditions of employment is not actionable under § 1981; and District Court erred in instructing jury that petitioner had to prove that she was better qualified than white employee who allegedly received promotion. Patterson v. McLean Credit Union, p. 164. 2. Racial discrimination—Municipality’s liability—Respondeat superior theory—Damages.—A. municipality may not be held liable for its employees’ violations of 42 U. S. C. § 1981 under a respondeat superior theory; exclusive federal damages remedy for violation of rights guaranteed by § 1981 when claim is pressed against a state actor is provided by § 1983. Jett v. Dallas Independent School District, p. 701. CIVIL RIGHTS ACT OF 1871. See also Civil Rights Act of 1866, 2. “Person” under 42 U. S. C. § 1983—States and state officials.— Neither States nor state officials acting in their official capacities are “persons” within meaning of § 1983. Will v. Michigan Department of State Police, p. 58. CIVIL RIGHTS ACT OF 1964. Sex discrimination—Intervenors’ liability for attorney’s fees.—District courts may award Title VII attorney’s fees against those who are not charged with Title VII violations but intervene to protect their own rights—here, seniority rights adversely affected by a settlement agreement in a sex discrimination suit against an employer—only where intervention is frivolous, unreasonable, or without foundation. Flight Attendants v. Zipes, p. 754. CIVIL RIGHTS ATTORNEY’S FEES AWARDS ACT OF 1976. See also Constitutional Law, IX, 1. Compensation—Work of paralegals, law clerks, and recent law graduates.—In a major school desegregation case, District Court correctly compensated work of paralegals, law clerks, and recent law graduates at market rate for their services, rather than at their cost to attorneys. Missouri v. Jenkins, p. 274. CIVIL SANCTIONS. See Constitutional Law, VII. CLEAR AND CONVINCING EVIDENCE STANDARD. See Libel. INDEX 917 COLLECTIVE-BARGAINING AGREEMENTS. See Railway Labor Act. COMMERCE CLAUSE. See Constitutional Law, I. COMPENSATION FOR ATTORNEY’S FEES. See Civil Rights Attorney’s Fees Awards Act of 1976; Constitutional Law. COMPREHENSIVE ENVIRONMENTAL RESPONSE, COMPENSATION, AND LIABILITY ACT OF 1980. See also Constitutional Law, I, 1. Federal-state relations—States’ liability for Superfund site cleanup costs.— Act, as amended by Superfund Amendments and Reauthorization Act of 1986, clearly renders States liable in money damages in federal court for Superfund site cleanup costs. Pennsylvania v. Union Gas Co., p. 1. COMPREHENSIVE FORFEITURE ACT OF 1984. See also Constitutional Law, III, 1; VIII. 1. Attorney’s fees—Exemption from forfeiture of assets from drug-law violations. — Act, which requires forfeiture of assets acquired as a result of drug-law violations, contains no exemption from either its forfeiture or pretrial restraining order provisions for assets that a defendant wishes to use to retain an attorney; restraining order freezing respondent’s assets pending trial did not violate his right to counsel of choice as protected by Sixth Amendment or Due Process Clause of Fifth Amendment. United States v. Monsanto, p. 600. 2. Attorney’s fees—Pretrial restraining orders. —Whatever discretion district court judges have under Act, which requires forfeiture of assets acquired as a result of drug-law violations, to refuse pretrial restraining orders on potentially forfeitable assets, it does not grant them equitable discretion to allow a defendant to withhold assets to pay bona fide attorney’s fees. Caplin & Drysdale, Chartered v. United States, p. 617. CONDITIONAL ENFORCEMENT OF SUMMONSES. See District Courts. CONDITIONS OF EMPLOYMENT. See Civil Rights Act of 1866, 1; Railway Labor Act, 1. CONNECTICUT. See Constitutional Law, I, 2. CONSTITUTIONAL LAW. See also Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Comprehensive Forfeiture Act of 1984, 1. I. Commerce Clause. 1. Congressional power—Superfund site cleanup costs—State’s liability for monetary damages.— Court of Appeals’ decision that Congress, under 918 INDEX CONSTITUTIONAL LAW-Continued. Commerce Clause, had power to render States liable for monetary damages for Superfund site cleanup costs, is affirmed. Pennsylvania v. Union Gas Co., p. 1. 2. Discrimination against out-of-state shippers—Beer-price affirmation statute.—Connecticut statute requiring out-of-state shippers of beer to affirm that their posted prices for products sold to state wholesalers are no higher than prices at which these products are being sold in bordering States violates Commerce Clause. Healy v. Beer Institute, p. 324. IL Double Jeopardy. Multiple punishments—Satisfaction of one of alternative punishments. —Where prisoner had been convicted of both attempted robbery and first-degree murder arising out of same incident, two consecutive sentences had been imposed when state law permitted only one, and prisoner had completed shorter sentence, state-court order vacating shorter sentence and crediting time already served against remaining sentence fully vindicated prisoner’s double jeopardy rights. Jones v. Thomas, p. 376. III. Due Process. 1. Comprehensive Forfeiture Act of 198k—Attorney's fees. — Act requiring forfeiture of assets acquired as a result of drug-law violations, including assets used to pay attorney’s fees, does not upset balance of power between Government and accused in a manner contrary to Due Process Clause of Fifth Amendment. Caplin & Drysdale, Chartered v. United States, p. 617. 2. Jury instructions—Presumptions as to a crime’s core elements.— Where appellant was tried for grand theft for failure to return a rental car, jury instructions that a person “shall be presumed to have embezzled” a vehicle if it is not returned within 5 days of rental agreement’s expiration date and that “intent to commit theft by fraud is presumed” from failure to return property within 20 days of demand violated Due Process Clause of Fourteenth Amendment, since they could have been understood to require jury to find presumed facts if State proved certain predicate facts. Carella v. California, p. 263. 3. State-law presumption of paternity—Putative father’s visitation rights.— California Court of Appeal’s holding that state law providing that a child bom to a married woman living with her husband is presumed to be legitimate does not infringe upon either due process rights of a man wishing to establish his paternity of a child bom to another man’s wife or child’s due process or equal protection rights, and court’s rejection of child’s assertion of a right to continued visitation with putative father, are affirmed. Michael H. v. Gerald D., p. 110. INDEX 919 CONSTITUTIONAL LAW-Continued. IV. Equal Protection of the Laws. Board of freeholders—Land-ownership requirement.— Where Missouri Constitution allows voters to reorganize certain local governments based on a plan drafted by a board of freeholders, fact that board serves only to recommend a reorganization plan and does not enact any laws of its own cannot immunize it from equal protection scrutiny; land-ownership requirement for appointment to board violates Equal Protection Clause. Quinn v. Millsap, p. 95. V. Freedom of Expression. Flag desecration—Criminalization.—Respondent’s conviction for desecrating an American flag in violation of Texas law is inconsistent with First Amendment. Texas v. Johnson, p. 397. VI. Freedom of Speech. 1. Prohibition on posing or exhibiting nude minors. —Massachusetts Supreme Court’s decision reversing respondent’s conviction under state law prohibiting adults from posing or exhibiting nude minors on ground that statute was substantially overbroad under First Amendment, is vacated and remanded. Massachusetts v. Oakes, p. 576. 2. Sound-amplification guideline—Time, place, and manner restriction.— City’s sound-amplification guideline, which allows it to regulate volume of amplified music at public bandshell, is valid under First Amendment as a reasonable regulation of place and manner of protected speech. Ward v. Rock Against Racism, p. 781. VIL Freedom of the Press. Newspapers—Damages for publication of rape victim’s name.— Where newspaper lawfully obtained a rape victim’s name from a police report, imposing damages on paper for publishing victim’s name in violation of Florida law violates First Amendment. The Florida Star v. B. J. F., p. 524. VIII. Right to Counsel. Comprehensive Forfeiture Act of 198^—Attorney’s fees.— Failure of Act requiring forfeiture of assets acquired as a result of drug-law violations to exempt assets used to pay an attorney does not impermissibly burden a defendant’s Sixth Amendment right to retain counsel of his choice. Caplin & Drysdale, Chartered v. United States, p. 617. IX. States’ Immunity from Suit. 1. Attorney’s fees—Enhancement of award. —Eleventh Amendment does not prohibit enhancement of a fee award under Civil Rights Attorney’s Fees Awards Act of 1976 against a State to compensate for delay in payment. Missouri v. Jenkins, p. 274. 920 INDEX CONSTITUTIONAL LAW-Continued. 2. Education of the Handicapped Act—Tuition reimbursement.— Act does not abrogate States’ Eleventh Amendment immunity from suit, and, thus, Amendment bars handicapped student’s attempt to collect tuition reimbursement from State. Dellmuth v. Muth, p. 223. COURTS OF APPEALS. See Libel. CRIME-FRAUD EXCEPTION TO ATTORNEY-CLIENT PRIVILEGE. See Attorney-Client Privilege. CRIMINAL LAW. See Comprehensive Forfeiture Act of 1984; Constitutional Law, II; III, 1, 2; V; VI, 1; VIII. DAMAGES. See Civil Rights Act of 1866, 2; Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Constitutional Law, I, 1; VII. DEDUCTIBILITY OF CEDING COMMISSIONS ON FEDERAL INCOME TAX RETURNS. See Taxes. DEFAMATION. See Libel. DESECRATION OF AMERICAN FLAG. See Constitutional Law, V. DISCRIMINATION AGAINST OUT-OF-STATE BUSINESSES. See Constitutional Law, I, 2. DISCRIMINATION IN EMPLOYMENT. See Civil Rights Act of 1866; Civil Rights Act of 1871; Civil Rights Act of 1964. DISCRIMINATION ON BASIS OF RACE. See Civil Rights Act of 1866. DISCRIMINATION ON BASIS OF SEX. See Civil Rights Act of 1964. DISCRIMINATORY PROMOTIONS. See Civil Rights Act of 1866, 1; Civil Rights Act of 1871. DISTRICT COURTS. See also Abstention; Comprehensive Forfeiture Act of 1984, 2. Internal Revenue Service summons - Conditional-enforcement order.— Court of Appeals’ decision upholding District Court’s decision to condition enforcement of an IRS summons by placing restrictions upon dissemination of subpoenaed information by IRS, is affirmed. United States v. Zolin, p. 554. DOUBLE JEOPARDY. See Constitutional Law, II. DRUG TESTING. See Railway Labor Act, 1. DRUG TRAFFICKERS. See Comprehensive Forfeiture Act of 1984; Constitutional Law, III, 1; VIII. INDEX 921 DUE PROCESS. See Comprehensive Forfeiture Act of 1984, 1; Constitutional Law, III. DUTY TO ENGAGE IN COLLECTIVE BARGAINING. See Railway Labor Act, 2. EDUCATION OF THE HANDICAPPED ACT. See Constitutional Law, IX, 2. ELEVENTH AMENDMENT. See Constitutional Law, IX. EMPLOYER AND EMPLOYEES. See Civil Rights Act of 1866, 1; Civil Rights Act of 1871; Railway Labor Act. EMPLOYMENT CONDITIONS. See Civil Rights Act of 1866, 1; Railway Labor Act, 1. EMPLOYMENT DISCRIMINATION. See Civil Rights Act of 1866; Civil Rights Act of 1871; Civil Rights Act of 1964. ENHANCEMENT OF ATTORNEY’S FEE AWARDS. See Constitutional Law, IX, 1. ENVIRONMENTAL PROTECTION. See Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Constitutional Law, I, 1. EQUAL PROTECTION OF THE LAWS. See Constitutional Law, III, 3; IV. EVIDENCE. See Libel. EXPRESSIVE CONDUCT. See Constitutional Law, V. FEDERAL ADVISORY COMMITTEE ACT. See also Jurisdiction, 1. American Bar Association Standing Committee on Federal Judiciary.—FAC A—which authorizes establishment of administrative guidelines and management controls for any advisory committee established or utilized by President or an agency to give advice on public questions—does not apply to Justice Department’s solicitation of ABA Committee’s views on prospective judicial nominees. Public Citizen v. Department of Justice, p. 440. FEDERAL DISTRICT COURTS. See Abstention; Comprehensive Forfeiture Act of 1984, 2; District Courts. FEDERAL INCOME TAXES. See Taxes. FEDERAL JUDICIAL NOMINEES. See Federal Advisory Committee Act; Jurisdiction, 1. FEDERAL-STATE RELATIONS. See Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Constitutional Law, I, 1; IX, 2. 922 INDEX FEDERAL TAXES. See Attorney-Client Privilege. FEE AWARDS. See Civil Rights Attorney’s Fees Awards Act of 1976; Constitutional Law, IX, 1. FIFTH AMENDMENT. See Comprehensive Forfeiture Act of 1984, 1; Constitutional Law, III, 1. FIRST AMENDMENT. See Constitutional Law, V-VII. FLAG BURNING. See Constitutional Law, V. FLORIDA. See Constitutional Law, VII. FORFEITURE OF ASSETS RESULTING FROM DRUG-LAW VIOLATIONS. See Comprehensive Forfeiture Act of 1984; Constitutional Law, III, 1; VIII. FOURTEENTH AMENDMENT. See Constitutional Law, II; III, 2, 3; IV. FREEDOM OF EXPRESSION. See Constitutional Law, V. FREEDOM OF SPEECH. See Constitutional Law, VI. FREEDOM OF THE PRESS. See Constitutional Law, VII. FREEHOLDERS’ BOARD. See Constitutional Law, IV. GOVERNMENT REORGANIZATION. See Constitutional Law, IV. GRAND THEFT. See Constitutional Law, III, 2. HANDICAPPED STUDENTS. See Constitutional Law, IX, 2. HARASSMENT ON BASIS OF RACE. See Civil Rights Act of 1866,1. ILLEGITIMATE CHILDREN. See Constitutional Law, III, 3. IMMUNITY FROM SUIT. See Constitutional Law, IX. IN CAMERA REVIEW OF ALLEGEDLY PRIVILEGED MATERIAL. See Attorney-Client Privilege. INCOME TAXES. See Taxes. INDEMNITY REINSURANCE AGREEMENTS. See Taxes. INJUNCTIONS. See Railway Labor Act, 2. INSURANCE POLICIES. See Taxes. INTERNAL REVENUE SERVICE. See Attorney-Client Privilege; District Courts. INTERVENORS’ LIABILITY FOR TITLE VII ATTORNEY’S FEES. See Civil Rights Act of 1964. JUDICIAL DISCRETION. See Comprehensive Forfeiture Act of 1984, 2. INDEX 923 JUDICIAL NOMINEES. See Federal Advisory Committee Act; Jurisdiction, 1. JURISDICTION. See also Abstention. 1. Standing to sue—Federal Advisory Committee Act.—Refusal to permit appellants to scrutinize American Bar Association Standing Committee on Federal Judiciary investigations, reports, and votes on potential judicial nominees to extent allowed by FAC A—which authorizes establishment of guidelines for certain committees established or utilized by Executive Branch—constitutes a sufficiently distinct injury to provide standing, and fact that other groups or citizens might make same complaint as appellants does not lessen that injury. Public Citizen v. Department of Justice, p. 440. 2. Supreme Court.— Where State Supreme Court upheld validity of state constitutional provision on ground that Equal Protection Clause had no relevancy to case, appellees’ arguments that this Court has no power to hear case—because decision was based on an adequate and independent state ground, it presented a political question, appellants lacked Article III standing to bring appeal, or appeal’s adjudication would interfere with executive’s power to make discretionary appointments—are rejected. Quinn v. Millsap, p. 95. JURY INSTRUCTIONS. See Constitutional Law, III, 2. LABOR RELATIONS. See Railway Labor Act. LAND-OWNERSHIP REQUIREMENT FOR APPOINTMENT TO BOARD OF FREEHOLDERS. See Constitutional Law, IV. LAW CLERKS. See Civil Rights Attorney’s Fees Awards Act of 1976. LAWYERS. See Civil Rights Act of 1964; Civil Rights Attorney’s Fees Awards Act of 1976; Comprehensive Forfeiture Act of 1984; Constitutional Law, III, 1; VIII; IX, 1. LEGITIMACY OF CHILDREN. See Constitutional Law, III, 3. LIBEL. Public figure—Standard of proof—Standard of review.— In order to support a libel verdict in his favor, a public figure must prove by clear and convincing evidence that defendant published false and defamatory material with actual malice and cannot rely solely on a showing of highly unreasonable conduct constituting an extreme departure from standards of investigation and reporting ordinarily adhered to by responsible publishers; a reviewing court must exercise independent judgment and determine whether record establishes actual malice with convincing clarity to ensure that verdict is consistent with constitutional standard. Harte-Hanks Communications, Inc. v. Connaughton, p. 657. 924 INDEX LOCAL GOVERNMENT REORGANIZATION. See Constitutional Law, IV. MAJOR LABOR DISPUTES. See Railway Labor Act, 1. MARKET RATE FOR ATTORNEY’S FEES. See Civil Rights Attorney’s Fees Awards Act of 1976. MINOR LABOR DISPUTES. See Railway Labor Act, 1. MINORS POSED OR EXHIBITED IN NUDE. See Constitutional Law, VI, 1. MISSOURI. See Constitutional Law, IV. MONETARY DAMAGES. See Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Constitutional Law, I, 1. MUNICIPALITY’S LIABILITY UNDER 42 U. S. C. § 1981. See Civil Rights Act of 1866, 2. MURDER. See Constitutional Law, II. MUSIC-AMPLIFICATION REGULATION. See Constitutional Law, VI, 2. NARCOTICS TRAFFICKERS. See Comprehensive Forfeiture Act of 1984; Constitutional Law, III, 1; VIII. NEWSPAPERS. See Constitutional Law, VII. NUDITY. See Constitutional Law, VI, 1. OBSCENITY. See Constitutional Law, VI, 1. OVERBREADTH OF OBSCENITY STATUTE. See Constitutional Law, VI, 1. PARALEGALS. See Civil Rights Attorney’s Fees Awards Act of 1976. PARENTS AND CHILDREN. See Constitutional Law, III, 3. PATERNITY DETERMINATIONS. See Constitutional Law, III, 3. PERSONS LIABLE UNDER 42 U. S. C. § 1983. See Civil Rights Act of 1871. PHYSICAL EXAMINATIONS. See Railway Labor Act, 1. POLICE REPORTS. See Constitutional Law, VII. PRE-EMPTION OF STATE LAW BY FEDERAL LAW. See Abstention. PRESUMPTION OF A CRIME’S CORE ELEMENTS. See Constitutional Law, III, 2. INDEX 925 PRESUMPTION OF CHILDREN’S LEGITIMACY. See Constitutional Law, III, 3. PRETRIAL RESTRAINING ORDERS. See Comprehensive Forfeiture Act of 1984. PRIVACY RIGHTS OF RAPE VICTIMS. See Constitutional Law, VII. PRIVILEGE. See Attorney-Client Privilege. PROMOTIONS. See Civil Rights Act of 1866, 1. PROOF. See Civil Rights Act of 1866, 1; Constitutional Law, III, 2, 3; Libel. PROTECTED EXPRESSION. See Constitutional Law, V. PROTECTED SPEECH. See Constitutional Law, VI, 2. PUBLICATION OF RAPE VICTIM’S NAME. See Constitutional Law, VIL PUBLIC EMPLOYER AND EMPLOYEES. See Civil Rights Act of 1871. PUBLIC FIGURE LIBEL. See Libel. PUBLIC UTILITIES. See Abstention. PUNISHMENTS. See Constitutional Law, II. RACIAL DISCRIMINATION. See Civil Rights Act of 1866. RACIAL HARASSMENT. See Civil Rights Act of 1866, 1. RAILWAY LABOR ACT. 1. Major or minor disputes—Drug-testing program.— Where a railroad announced unilaterally that urinalysis drug screening would be included as part of all periodic and retum-from-leave physical examinations, dispute— which arises out of assertion of a contractual right to take contested action—is minor, since action is arguably justified by terms of collectivebargaining agreement, rather than major, since employer’s claims are not frivolous or obviously insubstantial; if terms of parties’ agreement arguably justify employer’s claim that agreement gives it discretion to make a particular change in working conditions without prior negotiations, employer may make change and courts must defer to arbitral jurisdiction of Adjustment Board. Consolidated Rail Corp. v. Railway Labor Executives’ Assn., p. 299. 2. Sale of assets—Duty to bargain with unions—Injunction against strike.— Where petitioner agreed to sell its assets to another railroad’s 926 INDEX RAILWAY LABOR ACT-Continued. subsidiary, which would not accept petitioner’s collective-bargaining agreements or hire all of petitioner’s employees, Act did not require or authorize an injunction against sale itself but did impose a limited duty on petitioner to bargain over unions’ notices proposing extensive changes in agreements to ameliorate proposed sale’s adverse impact on employees; record is insufficient to allow this Court to determine whether Court of Appeals correctly set aside injunction against unions’ strike, since lower courts did not consider whether Act creates a duty not to strike while its dispute resolution mechanisms are underway. Pittsburgh & L. E. R. Co. v. Railway Labor Executives’ Assn., p. 490. RAPE VICTIM’S PRIVACY RIGHTS. See Constitutional Law, VII. REINSURANCE AGREEMENTS. See Taxes. REMEDIES. See Civil Rights Act of 1866, 2. RESPONDEAT SUPERIOR THEORY OF LIABILITY UNDER 42 U. S. C. § 1981. See Civil Rights Act of 1866, 2. RIGHT TO COUNSEL. See Comprehensive Forfeiture Act of 1984, 1; Constitutional Law, VIII. SECTION 1981. See Civil Rights Act of 1866. SECTION 1983. See Civil Rights Act of 1866, 2; Civil Rights Act of 1871. SENTENCES. See Constitutional Law, II. SEX DISCRIMINATION. See Civil Rights Act of 1964. SEXUAL ASSAULT VICTIM’S PRIVACY RIGHTS. See Constitutional Law, VII. SIXTH AMENDMENT. See Comprehensive Forfeiture Act of 1984, 1; Constitutional Law, VIII. SOUND-AMPLIFICATION REGULATION. See Constitutional Law, VI, 2. SOVEREIGN IMMUNITY. See Constitutional Law, IX. STANDARD OF PROOF. See Libel. STANDARD OF REVIEW. See Libel. STANDING TO SUE. See Jurisdiction, 1. STARE DECISIS. See Civil Rights Act of 1866, 1. STATE ACTORS’ LIABILITY UNDER 42 U. S. C. § 1981. See Civil Rights Act of 1866, 2. INDEX 927 STATES AND STATE OFFICIALS AS “PERSONS” UNDER 42 U. S. C. § 1983. See Civil Rights Act of 1871. STATES’ IMMUNITY FROM SUIT. See Constitutional Law, IX. STATES’ LIABILITY FOR DAMAGES UNDER FEDERAL LAW. See Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Constitutional Law, I, 1. STATE UTILITY RATEMAKING AUTHORITIES. See Abstention. STRIKES. See Railway Labor Act, 2. SUBPOENAS. See Attorney-Client Privilege; District Courts. SUMMONSES. See Attorney-Client Privilege; District Courts. SUPERFUND SITE CLEANUP COSTS. See Comprehensive Environmental Response, Compensation, and Liability Act of 1980; Constitutional Law, I, 1. SUPREME COURT. See Jurisdiction, 2. TAXES. See also Attorney-Client Privilege. Federal income taxes—Insurance companies—Treatment of indemnity reinsurance ceding commissions.— Ceding commissions—up-front fees paid to a primary insurer when life insurance companies enter into reinsurance agreements whereby reinsurer agrees to assume primary insurer’s liabilities on reinsured policies in return for future income generated from policies and their associated reserve accounts — paid under an indemnity reinsurance agreement must be amortized over agreement’s anticipated life and thus are not fully deductible in year tendered. Colonial American Life Insurance Co. v. Commissioner, p. 244. TEXAS. See Constitutional Law, V. TIME, PLACE, AND MANNER RESTRICTION OF SPEECH. See Constitutional Law, VI, 2. TITLE VIL See Civil Rights Act of 1964. TUITION REIMBURSEMENT FOR HANDICAPPED STUDENTS. See Constitutional Law, IX, 2. UNIONS. See Railway Labor Act. URINALYSIS. See Railway Labor Act, 1. UTILITIES. See Abstention. VISITATION RIGHTS OF PUTATIVE FATHERS. See Constitutional Law, III, 3. 928 INDEX WORDS AND PHRASES. 1. “Advisory committee.” Federal Advisory Committee Act, 5 U. S. C. App. § 3(2). Public Citizen v. Department of Justice, p. 440. 2. “Person.” Civil Rights Act of 1871. 42 U. S. C. §1983. Will v. Michigan Department of State Police, p. 58. WORKING CONDITIONS. See Railway Labor Act, 1. U.S. GOVERNMENT PRINTING OFFICE : 1993 0 - 239-614