ARTICLE THE (NON)FINALITY OF SUPREME COURT OPINIONS Richard J. Lazarus CONTENTS I. THE PROCESS OF REVISING SUPREME COURT OPINIONS ......................................... 546 A. The Court’s Earliest Years and the Supreme Court Reporter’s Practice of Revising Opinions ........................................................................................................ 547 B. The Rise of Justices Providing Written Opinions and Increased Pressure for Early Release of Initial Opinions ............................................................................. 549 II. HISTORICAL AND CURRENT PATHWAYS FOR REVISING SUPREME COURT OPINIONS AFTER INITIAL PUBLICATION: THE QUESTION OF TRANSPARENCY ................................................................................ 555 A. The Institutional Reasons for Revision ......................................................................... 556 B. Types of Revisions ............................................................................................................. 562 1. Typographical, Spelling, Grammar, and Citation Errors....................................... 562 2. Word Additions, Deletions, and Substitutions ......................................................... 563 3. Erroneous Characterizations of Facts, the Record, the Positions of the Parties, the Positions of the Other Justices, Background Law, and the Court’s Opinion ........................................................................................................... 566 C. Pathways for Revising Opinions ..................................................................................... 573 1. Errata Lists .................................................................................................................. 573 2. Formal, Published Orders Amending a Prior Opinion .......................................... 575 3. “Printer Proofs” and “Change Pages” ....................................................................... 580 (a) The Meaning of “Printer Proofs” and “Change Pages”..................................... 581 (b) The Origin of Printer Proofs and Change Pages ............................................... 582 (c) The Internal Process for Preparing the Printer Proofs and Change Pages.... 584 (d) Source Materials for Printer Proofs and Change Pages ................................... 586 (i) National Archives Records 1808–1913........................................................ 586 (ii) Change Pages in Possession of GPO and Private Publishers .................. 586 (iii) The Official Papers of Former Justices ...................................................... 588 (iv) Electronic Scanning and Comparison of Digitized Versions of Supreme Court Opinions .............................................................................. 588 (e) Types of Corrections Made ................................................................................... 589 (i) Dred Scott v. Sandford (1857) ...................................................................... 589 (ii) Granny Goose Foods, Inc. v. Brotherhood of Teamsters (1974) ............... 593 (iii) Hodel v. Virginia Surface Mining & Reclamation Ass’n (1981) .............. 595 (iv) International Paper Co. v. Ouellette (1987) ................................................ 597 (v) Lawrence v. Texas (2003) ............................................................................... 599 (vi) Clapper v. Amnesty International USA (2013) (potentially forthcoming) ............................................................................... 601 (vii) EPA v. EME Homer City Generation, L.P. (2014) .................................... 603 (f) Frequency of Revisions ......................................................................................... 607 540 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 541 III. ENSURING THE COURT’S INSTITUTIONAL INTEGRITY BY IMPROVING THE PROCESS AND PRACTICE OF REVISION .................................................................. 608 A. Weaknesses of the Court’s Current Revision Practices ................................................ 610 1. Absence of Apparent Procedural Rigor .................................................................... 610 2. Absence of Appropriate Transparency and the Potential Undermining of Opinion Integrity ......................................................................................................... 611 B. Lessons from the Other Branches ................................................................................... 612 C. Possible Reforms for the Court’s Consideration ............................................................ 618 CONCLUSION ................................................................................................................................. 622 THE (NON)FINALITY OF SUPREME COURT OPINIONS Richard J. Lazarus∗ R emarking on the Supreme Court in his separate concurring opinion in Brown v. Allen,1 Justice Robert Jackson famously quipped, “We are not final because we are infallible, but we are infallible only because we are final.”2 As the Court itself makes clear, however, its published opinions are not immediately “final” at all. Just the opposite is true. The Court’s opinion in Brown itself was not final on the day it was announced, notwithstanding all the visible trappings and signature headings of the Court. It takes not just days or months but several years after the Court’s initial announcement and publication of its ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– * Howard J. & Katherine W. Aibel Professor of Law, Harvard University. © 2014 Richard J. Lazarus. I would like to thank my colleagues Naomi Campbell, Richard Fallon, Charles Fried, David Garrow, Vicki Jackson, Michael Klarman, Ron Levin, Peter Martin, Bill Popkin, Josh Schwartz, Joe Singer, Richard Taranto, Mark Tushnet, and Adrian Vermeule for their comments on an earlier draft that prompted significant improvements, and Karen Cordry, John MacKenzie, and Stephen Wasby, each of whom contacted me following a posting of an earlier version of this Article to offer additional examples of opinion revisions that they had come across in their work. I am also grateful to current and former Harvard Law School students Caitlin Halpern, Turner Smith, Leslie Griffith, Brendan Selby, Robert Niles, and Maya Brodziak for their editorial and research assistance in the preparation of this Article. Finally, special thanks are owed to LexisNexis for its kindness in providing me with the change pages for a volume of the United States Reports; former Supreme Court Reporter Frank Wagner; current members of the Office of the Supreme Court Reporter, the Office of the Supreme Court Curator, and the Office of the Supreme Court Library; the officials of the Library of Congress Manuscript Division; Robert Ellis of the National Archives; Stephen Wermiel and members of the family of Justice William J. Brennan, Jr.; and Meg Kribble of the Harvard Law School Library for their assistance in gaining access to the voluminous historical research base necessary for the preparation of this Article. None has reviewed, approved of, or is otherwise responsible for the Article’s findings or conclusions. This Article frequently cites to private correspondence between the Justices and between the Justices and the Supreme Court Reporter included in the collections of the official papers of the Justices maintained by the Library of Congress. See MANUSCRIPT DIV., LIBRARY OF CONG., BYRON R. WHITE PAPERS: A FINDING AID TO THE COLLECTION IN THE LIBRARY OF CONGRESS (rev. ed. 2012); MANUSCRIPT DIV., LIBRARY OF CONG., HARLAN FISKE STONE PAPERS: A FINDING AID TO THE COLLECTION IN THE LIBRARY OF CONGRESS (2010); MANUSCRIPT DIV., LIBRARY OF CONG., HARRY A. BLACKMUN PAPERS: A FINDING AID TO THE COLLECTION IN THE LIBRARY OF CONGRESS (rev. ed. 2010); MANUSCRIPT DIV., LIBRARY OF CONG., THURGOOD MARSHALL PAPERS: A FINDING AID TO THE COLLECTION IN THE LIBRARY OF CONGRESS (rev. ed. 2010); MANUSCRIPT DIV., LIBRARY OF CONG., WILLIAM J. BRENNAN, JR.: A REGISTER OF HIS PAPERS IN THE LIBRARY OF CONGRESS (2001); MANUSCRIPT DIV., LIBRARY OF CONG., WILLIAM O. DOUGLAS PAPERS: A FINDING AID TO THE COLLECTION IN THE LIBRARY OF CONGRESS (rev. ed. 2012). For citation purposes, any reference to a document from those collections will refer to the name of the Justice whose papers contain the document, followed by the box and folder number (if any) containing that document: for example, Stone Papers B82/F5. 1 344 U.S. 443 (1953). 2 Id. at 540 (Jackson, J., concurring in the result). 542 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 543 ruling before the Court releases what it is willing to describe as its “final” and “official” opinion. Hiding in plain sight at the top of a Supreme Court opinion when first issued is a formal notice that makes clear its nonfinal and nonofficial nature: NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D.C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.3 This notice appears on both “bench opinions,” distributed by the Court at the immediate conclusion of the opinion’s announcement from the bench, and “slip opinions,” which the Court releases several days later.4 A similarly worded notice appears on the “preliminary print” of the United States Reports, published several months after the original opinion announcement.5 The notice states that this formally published version of the Court’s opinion is likewise subject to revision before publication in the “bound volume[s]” of the United States Reports and invites members of the public to notify the Reporter of Decisions of errors so that corrections can be made.6 According to the Supreme Court, “[o]nly the bound volumes of the United States Reports contain the final, official text of [the Court’s] opinions.”7 Those volumes are published several years after the original opinion announcements. For instance, the Court handed down its final merits decisions of the October Term 2007 on June 26, 2008.8 The last volume of the corresponding set of United States Reports, including those final decisions, was not published until five years later.9 Five years is a long time to wait for the “final” and “official” version of a Supreme Court ruling. Since modern technology creates a public expectation of receiving information at lightning speed, a five-year delay ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 3 4 E.g., Burwell v. Hobby Lobby Stores, Inc., No. 13-354, slip op. at 1 (U.S. June 30, 2014). See, e.g., id. See generally Information About Opinions, SUPREME CT. U.S., http://www .supremecourt.gov/opinions/info_opinions.aspx (last visited Oct. 26, 2014) [http://perma.cc/TAW9 -8LBF]. Notwithstanding the Court’s stated distinction between “bench” and “slip” opinions, the opinion the Court releases on its website immediately after the announcement of the Court’s opinion from the bench is labeled “slip opinion.” The “bench opinion” version is distributed by the Court’s Office of Public Information that same morning to the news media and members of the public. 5 561 U.S. (2010) (prelim. print) (cover page). 6 Id. 7 Information About Opinions, supra note 4 (emphasis added). 8 See Davis v. FEC, 554 U.S. 724, 724 (2008); District of Columbia v. Heller, 554 U.S. 570, 570 (2008); Morgan Stanley Capital Grp. Inc. v. Pub. Util. Dist. No. 1, 554 U.S. 527, 527 (2008). 9 See 554 U.S. I (2013). 544 HARVARD LAW REVIEW [Vol. 128:540 might well be the psychological equivalent of a decades-long delay a half century ago.10 The origins and potential implications of these seemingly innocuous notices are fascinating and take us back through the Court’s history, long before the express acknowledgment was included in the Court’s opinions. In what is unquestionably the most extreme instance, Chief Justice Taney in Dred Scott v. Sandford11 added approximately eighteen pages to his majority opinion between the time of his original announcement of the Court’s ruling and the publication of the opinion several months later.12 And, as recently as Lawrence v. Texas13 in 2003, striking down as unconstitutional a Texas law criminalizing sodomy, and EPA v. EME Homer City Generation, L.P.14 in April 2014, upholding a major agency rule promulgated under the federal Clean Air Act,15 Justices have revised their opinions in significant, including highly substantive, ways prior to their final and official publication in the United States Reports.16 Examination of the Court’s practice naturally raises a series of questions. First, what does the Court mean by its suggestion that the initial slip opinions and preliminary prints are not “final, official” opinions? Unlike “proposed” rules that agencies publish in the Federal Register, Supreme Court opinions are legally effective as soon as they are first announced. Further, the Court prints slip opinions on inhouse equipment; no unofficial third party is involved.17 Second, what does the Court mean by “formal errors”?18 Are formal errors merely typographical errors, including technical errors in citation forms and grammar? Or does the term extend to factual mistakes? Further still, do the Justices make substantive changes in opinions under the rubric of a “formal error”? How frequent and how extensive are the changes that the Justices make? Third, what notices are provided when Justices make changes? Are other Justices notified of changes before or after they are made? What about the parties to the case? Or members of the public? Is there an opportunity to comment on proposed changes? ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 10 See JAMES GLEICK, FASTER 9 (1999) (“We are in a rush. We are making haste. A compression of time characterizes the life of the century now closing.”); Alex Kacelnik, The Evolution of Patience, in TIME AND DECISION 115 (George Loewenstein et al. eds., 2003). 11 60 U.S. (19 How.) 393 (1857). 12 See infra pp. 589–93. 13 539 U.S. 558 (2003). 14 134 S. Ct. 1584 (2014). 15 42 U.S.C. §§ 7401–7671q (2012). 16 See infra pp. 599–600, 603–07. 17 Information About Opinions, supra note 4. 18 E.g., Burwell v. Hobby Lobby Stores, Inc., No. 13-354, slip op. at 1 (U.S. June 30, 2014). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 545 Finally, what problems does the Court’s current process of revising opinions create? Do changes unwittingly introduce other errors? Does the lack of transparency invite Justices to make changes that they would be less likely to make if a more public acknowledgment were required? And, more particularly, do legal publications, lawyers, and judges sufficiently account for the Court’s revision process by ensuring that they are citing the “final, official” version? By what authority can Justices today correct “mistakes” in opinions published years or even decades before? Mistaken or not, those earlier opinions accurately state the words on which a majority of Justices, none of whom may currently serve on the Court, presumably relied in casting their votes at the time of decision. The purpose of this Article is to answer these questions by exploring the history and significance of the Court’s practice of revising opinions, which legal scholarship has largely ignored.19 The Article is di––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 19 Every sweeping statement that previous scholarship has “ignored” something has an exception. And when it comes to legal scholars ignoring something significant, that exception invariably is my colleague Professor Mark Tushnet. See Mark Tushnet, Sloppiness in the Supreme Court, O.T. 1935–O.T. 1944, 3 CONST. COMMENT. 73 (1986). This exception is no exception in that respect, although Tushnet’s focus is very different and far narrower: he focuses on a relatively brief moment between 1935 and 1945, when the Court utilized a wholly transparent practice of revising opinions, while this Article focuses on the strikingly different and nontransparent process that now dominates the Court’s practices. Tushnet alludes to this modern practice in a single sentence and does not consider the fuller implications of the scope of subsequent “corrections.” See id. at 81–82 (“Today the Reporter’s office undoubtedly corrects opinions, with the agreement of the Justices, as a routine matter.”). In addition, Tushnet expressly stops short of locating the original slip opinions that had been revised, which made it “impossible to determine the exact significance of a modification.” Id. at 75 n.12. This Article takes that additional step of comparing the final version to the original slip opinion in examined cases. Apart from Tushnet’s article, I have come across only three instances in which the practice has been acknowledged, but on each occasion only in application to a single, isolated case, and without any apparent awareness or discussion of the possibility or significance of a more widespread practice. The first is an incidental acknowledgment that the Court made a technical correction in a particular case. See Joshua Dressler, A Lesson in Incaution, Overwork, and Fatigue: The Judicial Miscraftsmanship of Segura v. United States, 26 WM. & MARY L. REV. 375, 407 (1985) (describing revision of a slip opinion to correct the misattribution of a quote of Justice Frankfurter to Justice Jackson). The second is a truly fascinating case comment that describes how the Court mistook the date of a Court ruling decided a century earlier and then relied heavily on that erroneous date to support its reasoning. See John P. MacKenzie, Comment, Hamm v. City of Rock Hill and the Federal Savings Statute, 54 GEO. L.J. 173 (1965). The case comment includes at the very end a brief “Editor’s Note” informing the reader, without analysis, of the author’s subsequent receipt of correspondence from the Supreme Court Reporter that the opinion will be revised prior to publication in the United States Reports. Id. at 182 (citing Letter from Reporter of Decisions, Supreme Court of the United States, to John P. MacKenzie, Supreme Court Reporter, Wash. Post (Oct. 27, 1965)). Finally, there is a short, recently published essay that describes changes made in a separate opinion by Chief Justice Burger, speaking just for himself, concurring in the denial of a petition for a writ of certiorari. See Douglas P. Woodlock, Chief Justice Burger Writes an Opinion on Palimpsest, 17 GREEN BAG 2D 37 (2013). The essay evidences no awareness that the practice is in fact far deeper and more widespread than the author discusses, extending throughout the Court’s history, to the entire Court, and to the Court’s actual opinions on the merits, and not just to an individual 546 HARVARD LAW REVIEW [Vol. 128:540 vided into three parts. Part I places the practice of revising Supreme Court opinions in its broader procedural and historical context. This includes a description of the formal stages in the Court’s opinionwriting process and a survey of the varied formal and informal processes for triggering and announcing revisions. Part I also catalogues the types of revisions that the Court makes, which from the Court’s earliest days to the present have included significant substantive changes. Part II examines the actual practice of revising opinions. Not surprisingly, discerning this practice proved challenging. Although the Court has long revised its opinions and disclosed the fact that it does so, the Court has done little to make clear what changes have been made in individual cases. Instead, the Court deliberately makes discovery difficult notwithstanding the public nature of the revisions. This Part of the Article illustrates the scope and potential depth of the revision process by describing examples of revisions, extending from the early nineteenth century to the present. These examples do not reflect an exhaustive effort to discover all possible revisions. They represent just a sampling of instances that I discovered based on my review of several sources. Part III considers options for improving the process and practice of revising Supreme Court opinions. This analysis requires identification of the advantages and disadvantages of the Court’s current process, and it discusses the contrasting processes that Congress and federal agencies use to revise statutes and regulations. To the extent, moreover, that such analysis and discussion of current practices implicitly or explicitly criticize the Court’s practices, any such criticism is not at all directed at the Office of the Reporter. At least since the early twentieth century, the Reporter’s Office has steadfastly performed its responsibilities, as instructed by the Court, with extraordinary skill, integrity, and professionalism. It is instead those Court instructions that this Article questions and that warrant rethinking and reform. The current Court did not itself create those instructions, which find their origins in practices that have persisted for decades if not centuries. The Court today, however, can easily fix them. I. THE PROCESS OF REVISING SUPREME COURT OPINIONS The process of revising Supreme Court opinions has changed considerably since its early days and has become both more and less transparent over time. Two fundamental reasons for the changing nature of the revision process are that the process within the Court of ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– opinion, of a single Justice, on an isolated occurrence, at the jurisdictional stage, and with no force of law. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 547 drafting opinions has changed, and the amount of public and news media interest in the Court’s opinions has substantially increased. In particular, the Supreme Court Reporter lost control over the decision of when opinions were to be published. And outside pressures from the public prompted earlier publication before the Reporter had a full opportunity to review initial drafts. In different but ultimately reinforcing ways, the changing Court procedures and increased public interest thereby combined to produce multiple published versions of the Court’s opinions and greater divergence between the Court’s original and final versions. A. The Court’s Earliest Years and the Supreme Court Reporter’s Practice of Revising Opinions During the Court’s earliest years, Supreme Court Reporters regularly revised opinions after they were announced orally by the Justices. Indeed, that was essentially their job.20 The Justices would publicly read their rulings in open Court, but only “in a few instances” would they provide the Reporter with the text.21 Instead, the Supreme Court Reporter would create a written opinion based on his notes of what a Justice said and any notes that the Justice provided. The Reporter would also secure notes from other attorneys who happened to be in the courtroom for the opinion announcement, especially if the Reporter had been absent.22 Somewhat counterintuitively, the first Reporters were not government employees and were essentially self-appointed, enhancing their ability to influence opinions.23 Nor was this a problem original to a then-new nation. The same practice had long persisted in England. During the early seventeenth century, Sir Edward Coke was notorious for being “too fond of making the law, instead of declaring the law, and of telling untruths to support his own opinions” in his reports.24 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 20 See MORRIS L. COHEN & SHARON HAMBY O’CONNOR, A GUIDE TO THE EARLY REPORTS OF THE SUPREME COURT OF THE UNITED STATES 4, 7 (1995). 21 5 THE DOCUMENTARY HISTORY OF THE SUPREME COURT OF THE UNITED STATES, 1789–1800, at xxiv (Maeva Marcus ed., 1994). 22 Id. at xxiv–xxv; COHEN & O’CONNOR, supra note 20, at 7. 23 Gerald T. Dunne, Early Court Reporters, in 1976 SUPREME COURT HISTORICAL SOC’Y, YEARBOOK 61, 62 (William F. Swindler ed., 1975). 24 JOHN WILLIAM WALLACE, THE REPORTERS 172 (4th ed. 1882) (emphasis omitted) (quoting NICHOLAS HARRIS NICOLAS, A TREATISE ON THE LAW OF ADULTERINE BASTARDY 80 (1836)) (internal quotation mark omitted). Blackstone’s Commentaries on the Laws of England, originally published between 1765 and 1769, were premised on the notion that judicial opinions were not law itself but “evidence of the law.” WILLIAM D. POPKIN, EVOLUTION OF THE JUDICIAL OPINION 13 (2007) (internal quotation marks omitted). The emergence of more formal announcements of judicial opinions in the mid-eighteenth century reportedly reflected an effort by the courts to compete with the rising role of legislatures by making their declarations of 548 HARVARD LAW REVIEW [Vol. 128:540 With no government salary, the Reporter’s sole source of income for this work derived from the sale of published volumes.25 The Reporter therefore had a strong interest in attracting buyers. Although in theory that could mean making the written rulings more interesting, one of the most significant impacts in practice was that Reporters exercised great discretion in deciding which rulings to report.26 Such discretion both reduced the publication costs and made the resulting volumes more substantively attractive.27 In other words, the Reporter’s ability to revise extended to eliminating opinions entirely from the public record based on his view that the rulings were too unimportant to merit the labor (and pages) required for their inclusion. The earliest Reporters — including in particular each of the first three, Alexander J. Dallas, William Cranch, and Henry Wheaton — had an established practice of not reporting all of the Court’s rulings.28 Dallas may have omitted as many as one-third of the Court’s rulings from his volumes.29 When Dallas took on the Reporter job, entirely on his own initiative, his reporting of Supreme Court rulings was very much secondary to his reporting of Pennsylvania state court rulings.30 His first volume included only Pennsylvania cases and no Supreme Court decisions; his second volume covered seventeen years of Pennsylvania decisions and was up to date, while that same volume covered only three years of Supreme Court rulings and was five years behind.31 The Supreme Court rulings were also at the back of the volume,32 underscoring their secondary status. In his sixteen Terms as Reporter, Dallas published fewer than seventy Supreme Court cases.33 Not until an act of Congress in 1817 did the Reporter (then Wheaton34) become a government employee and an officer of the Court.35 Even then the stipend ($1000) was substantially less than the ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– law more formal “as a parallel and legitimate form of law, side by side with statutes.” Id. at 7; see also id. at 15. 25 See COHEN & O’CONNOR, supra note 20, at 2. 26 See POPKIN, supra note 24, at 66. 27 Because Reporters were frequently advocates in cases they reported, there may have been a natural tendency to highlight cases in which the Reporters participated as counsel. See infra p. 549. 28 See COHEN & O’CONNOR, supra note 20, at 4, 7; see also Preface, 14 U.S. (1 Wheat.) iii, iii–iv (1816) (acknowledging the omission of cases that were factbound and unimportant). 29 COHEN & O’CONNOR, supra note 20, at 4. 30 See id. at 17–18. 31 See 2 U.S. (2 Dall.) (1798); 1 U.S. (1 Dall.) (1790); COHEN & O’CONNOR, supra note 20, at 17–21. 32 See 2 U.S. (2 Dall.) 399–480 (1798). 33 COHEN & O’CONNOR, supra note 20, at 21. 34 POPKIN, supra note 24, at 77. 35 An Act to Provide for Reports of the Decisions of the Supreme Court, ch. 63, 3 Stat. 376 (1817). Justice Story reportedly prompted the federal legislation. POPKIN, supra note 24, at 77. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 549 cost of preparing and publishing the volumes, so the Reporter remained dependent on sales.36 The Reporter’s exclusive right to publish opinions was essential to the position’s economic viability. The flip side, of course, was that neither the news media nor the public had access to written opinions until published by the Reporter. When the Court announced its opinion in Marbury v. Madison37 on February 24, 1803, for instance, the market demand was great, but the public had to wait more than a year for Cranch to publish his first volume, covering the Court’s rulings from 1801 to 1804.38 The potential for divergence between the Court’s orally announced ruling and the Reporter’s subsequent written opinion was great, especially when the Justice did not provide the Reporter with a draft written opinion. The Court’s first reporter, Dallas, “exercised his own editorial judgment” and “made changes in the language of the opinions.”39 Exercising a professional duality that would inspire envy in today’s Supreme Court Bar, Dallas reported cases in which he had been the advocate for one of the parties, including the first significant case ever published in the United States Reports, Chisholm v. Georgia.40 Nor did other early Reporters view their responsibilities as a full-time or even primary focus of employment. The second Reporter, Cranch, simultaneously served as “Chief Justice” of the U.S. Circuit Court of the District of Columbia, the forerunner to the D.C. Circuit.41 The third Reporter, Wheaton, argued more than twenty-five cases during his time in that position.42 B. The Rise of Justices Providing Written Opinions and Increased Pressure for Early Release of Initial Opinions The relationship among the Justices, the Reporter, and the Court’s published opinions, however, evolved during the first half of the nine––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 36 37 38 See COHEN & O’CONNOR, supra note 20, at 45 n.57, 46–47. 5 U.S. (1 Cranch) 137 (1803). See id. (decided February 24, 1803); 5 U.S. (1 Cranch) i (1804) (providing the volume’s date of publication); COHEN & O’CONNOR, supra note 20, at 30. 39 5 THE DOCUMENTARY HISTORY OF THE SUPREME COURT OF THE UNITED STATES, 1789–1800, supra note 21, at xxv. 40 2 U.S. (2 Dall.) 419 (1793); see id. at 419; see also COHEN & O’CONNOR, supra note 20, at 14–15. Dallas’s second volume, the first to include federal Supreme Court rulings, published nine Supreme Court opinions in eighty pages. See 2 U.S. (2 Dall.) 401–80 (1798). Chisholm constituted sixty-two of those eighty pages and included, as was the custom at the time, a description of the arguments of counsel. See Chisholm, 2 U.S. (2 Dall.) at 419–80. 41 COHEN & O’CONNOR, supra note 20, at 26. 42 Id. at 39. The fifth Reporter, Benjamin Chew Howard, advertised his advocacy services in the United States Reports, available when it was “inconvenient or impossible for the counsel who argued [a case] below to” do so. Id. at 82 (quoting A Card from the Reporter to His Professional Brethren and the Public Generally, 42 U.S. (1 How.) v, v (1843)). 550 HARVARD LAW REVIEW [Vol. 128:540 teenth century. The Justices increasingly began providing written drafts to the Reporter “in all cases of difficulty or importance”43 rather than relying on the Reporter’s doing so in the first instance.44 Draft written opinions by the Justices, however, eliminated neither the opportunity for friction between the Justices and the Reporter nor the opportunity for revision of the Court’s opinion between its initial announcement and final publication. In certain respects, the potential for both increased, especially as the pressure for ever-earlier release of the Justice’s draft written opinion for the Court likewise increased.45 The resulting rise in friction between individual Justices and various Reporters over the latter’s role in drafting and revising the Court’s opinions was reflected in the challenging tenures many of those Reporters faced in their jobs. On the one hand, the Justices were well aware that they needed the assistance of an able editor. On the other, the Justices became increasingly critical of the job that many Reporters were doing, and coalitions of Justices regularly lobbied, successfully, for the ouster of Reporters from their positions. For instance, with regard to “verbal and grammatical errors,” Justice Story strongly praised the role of the Reporter in correcting the opinions before their final publication in bound volumes.46 According to Justice Story, “it would be a disgrace to all concerned,” “mar the sense,” and “pain the author” not to correct such errors.47 And “[i]f a reporter do no more than acts of this sort, removing mere blemishes, he does all Judges a great favor.”48 The Justices, however, also regularly complained about Reporter inaccuracy. Justice Story complained about the (in)accuracy of Cranch’s reporting and helped push Cranch out in favor of Wheaton.49 Wheaton, in turn, may have resigned under pressure amidst concern on the Court about the extent to which Justice Story collaborated with Wheaton in drafting opinions.50 Wheaton’s successor, Richard Peters, ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 43 44 45 Preface, 5 U.S. (1 Cranch) iii, v (1804). Id. at iv–v. Publication delays remained a constant complaint. Dallas did not publish his last volume until 1807, three years after the second Reporter, Cranch, had published his first volume. See 5 U.S. (1 Cranch) at i; 4 U.S. (4 Dall.) i (1807); COHEN & O’CONNOR, supra note 20, at 30. When the third Reporter, Wheaton, took over in 1816, Cranch had not published volumes for cases reported between 1812 and 1815 because of financial difficulties. COHEN & O’CONNOR, supra note 20, at 47. 46 Alden I. Rosbrook, The Art of Judicial Reporting, 10 CORNELL L.Q. 103, 123 (1925) (quoting Letter from Joseph Story, Assoc. Justice, Supreme Court of the United States, to Richard Peters, Reporter of Decisions, Supreme Court of the United States (May 7, 1836), in 2 LIFE AND LETTERS OF JOSEPH STORY 231, 232 (William W. Story ed., 1851)). 47 Id. (quoting Letter from Joseph Story to Richard Peters, supra note 46, at 232). 48 Id. (quoting Letter from Joseph Story to Richard Peters, supra note 46, at 232). 49 POPKIN, supra note 24, at 76–80. 50 COHEN & O’CONNOR, supra note 20, at 52–53, 64. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 551 ultimately fared no better.51 Peters had the advantage of the Court’s new rule, established in 1834, that Justices provide their opinions in writing.52 But a draft written opinion is still just that: a “draft” and “written.” As a practical matter, it was frequently hard to read a Justice’s handwriting, and significant editorial assistance was needed to transition from draft to final published version.53 Peters, too, was ultimately forced out by a group of Justices, including Justice Catron,54 who published a lengthy letter listing all the errors that Peters had committed in publishing Catron’s opinions in five recent volumes.55 Once it became well known that the Justices were providing the Reporter with draft written opinions, the pressure naturally increased for earlier publication of those initial drafts. The reasons for delay until publication of the formal United States Reports seemed far less compelling once the Reporter was playing a less substantive role. And, even more important, with the rise of the national government during the mid-nineteenth century, the media and the public grew more impatient for a copy of the Court’s opinions,56 as did competing private publishers.57 The Court’s rulings were potentially big news and the nation’s newspapers, because of modern technology, were capable of publishing ever more quickly and distributing their product ever more widely.58 For the Reporter, however, early publication of opinions was plainly problematic. First, such early publication would be potentially catastrophic in terms of the basic economic viability of the job. What made the Reporter position profitable, at least on a marginal basis, ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 51 Controversy between Wheaton and Peters famously spilled over into litigation before the Court itself when Peters sought to republish Court opinions that had been handed down while Wheaton was the Reporter. See id. at 53–57. The Court rejected Wheaton’s claim of copyright infringement. See Wheaton v. Peters, 33 U.S. (8 Pet.) 591, 667–68 (1834). 52 Order of Mar. 14, 1834, 33 U.S. (8 Pet.) vii (1834). 53 COHEN & O’CONNOR, supra note 20, at 84–85, 85 n.38; see also 5 CARL B. SWISHER, THE OLIVER WENDELL HOLMES DEVISE HISTORY OF THE SUPREME COURT OF THE UNITED STATES 310 (1974). 54 COHEN & O’CONNOR, supra note 20, at 72–73. An example of a Peters transgression was his misspelling of Justice Daniel’s name. Daniel, as a result, favored his dismissal. Id. at 72 & n.54. 55 See List of Errata in the Opinions of the Court as Pronounced by Mr. Justice Catron, or in His Dissentient Opinions, Contained in the 12th, 13th, 14th, 15th and 16th Volumes of Peters’ Reports, 42 U.S. (1 How.) xv, xv–xix (1843). In what appears to have been intended as pointed sarcasm rather than a formal injunction, Justice Catron justified publication of his lengthy list of Peters’s errors by referring to the inevitable delays before Peters would be able to publish a corrected “second edition” of his reports. Id. at xv. Of course, no such “second edition” was ever forthcoming. 56 See, e.g., COHEN & O’CONNOR, supra note 20, at 85. 57 Id. at 68. 58 See generally, e.g., JAMES L. CROUTHAMEL, BENNETT’S NEW YORK HERALD AND THE RISE OF THE POPULAR PRESS (1989); SUSAN THOMPSON, THE PENNY PRESS (2004); Donald K. Brazeal, Precursor to Modern Media Hype: The 1830s Penny Press, 28 J. AM. CULTURE 405 (2005). 552 HARVARD LAW REVIEW [Vol. 128:540 was the Reporter’s exclusive right to publish opinions in the first instance.59 To be sure, the position provided significant professional stature, and Reporters routinely did more than just publish the Court’s opinions. They would write headnotes, summaries of the arguments of counsel, commentaries, and indices that both added substantive value to the publication and, no less important, increased the price that could be charged for each volume.60 But release of the Justices’ written opinions upon their initial announcement was perceived as a threat to the viability of the Reporter’s entire enterprise.61 Second, early publication of a Justice’s initial opinion for the Court, before the Reporter had fully reviewed the draft for errors or otherwise had an opportunity to make or at least propose editorial modifications, would dramatically increase the potential for divergence between the original published version and the final version appearing in the United States Reports. In the past, the revisions were all internal to the Court. The Reporter would review and edit the draft. The Justices would make changes upon further reflection, including based on statements made in concurring and dissenting opinions. The differences between the opinion announced by the Court orally and the subsequently published written opinion could be significant.62 But there was still only one written version: the opinion published by the Reporter in the United States Reports. There were no competing written versions and therefore no hard evidence of the changes made. That would no longer be true if the Court were to change its practices and release initial opinions immediately after the announcement of the Court’s ruling. Then, any subsequent revisions would result in contrasting written publications. The potential for confusion would be considerable. After a major battle between the Supreme Court Clerk and the Reporter,63 the Court ultimately sided with the Clerk and allowed the Clerk, in response to demand from the news media and the public, to release opinions to the public prior to their formal publication by the Reporter.64 To both allow for necessary revisions and reduce the potential for confusion, the Court appears to have taken three steps. The first was ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 59 60 61 62 63 See COHEN & O’CONNOR, supra note 20, at 55. See POPKIN, supra note 24, at 66–67. See COHEN & O’CONNOR, supra note 20, at 68–69. See, e.g., infra pp. 589–93. The Court initially sided with the Reporter: the Court would provide its opinions to the Reporter, who would deliver them to the Clerk “for preservation; as soon as the volume of the reports for the term at which they are delivered shall be published,” so the Reporter’s publication would take precedence. Order of Mar. 14, 1834, 33 U.S. (8 Pet.) vii, vii (1834). 64 COHEN & O’CONNOR, supra note 20, at 69 n.43; see also S. CT. R. 42 (1835), 42 U.S. (1 How.) xxxv (1843) (providing that the Court shall deliver all opinions to the Clerk, who will record them and then deliver them to the Reporter). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 553 to delay the announcement of the ruling until the written opinion had been more thoroughly drafted, edited, revised, and proofed within the Court. The best evidence of this change, though indirect and therefore not incontrovertible, is that the Court takes far more time to decide cases than it used to, even though it currently hears far fewer cases than it used to,65 and even though each Justice enjoys far more assistance from law clerks, whose presence was sporadic in the late nineteenth century, but standard at four per each Justice’s chambers today.66 For instance, the average time between oral argument and opinion announcement for cases argued in February Term 1825 was nine days.67 For October Term 1875, the corresponding time period was thirty days,68 and for October Term 1925, it was forty-nine days.69 In more recent decades, the time it has taken to announce opinions following oral argument has been even greater still: often more than one hundred days.70 The number of pages of opinions (majority, concurring, and dissenting) has significantly increased during this same time period,71 so that too could contribute to the length of time needed to review and revise opinions prior to announcement. Second, the Court developed formal procedures for releasing and publishing advance opinions prior to final publication in the United States Reports. Potentially as early as the late nineteenth century, and no later than the early twentieth century, the Court was regularly releasing copies of its opinions within a few days of the opinions’ announcements. Private companies published both those original opinions and, later on, the final opinions appearing in the bound United States Reports. The Lawyers Co-operative Publishing Company began publishing Lawyers’ Edition in 1882,72 and West Publishing Com––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 65 The Court’s plenary docket has dramatically shrunk since the late nineteenth century: from just shy of 300 in the late nineteenth century, to closer to 200 in the 1920s, to around 140 in the 1970s, and to around 75 in the past decade. See David R. Stras, The Supreme Court’s Declining Plenary Docket: A Membership-Based Explanation, 27 CONST. COMMENT. 151, 152–53 (2010); David R. Stras, The Supreme Court’s Gatekeepers: The Role of Law Clerks in the Certiorari Process, 85 TEX. L. REV. 947, 964–68 (2007) (book review). 66 ARTEMUS WARD & DAVID L. WEIDEN, SORCERERS’ APPRENTICES 30–53 (2006). 67 Author’s calculations based on 23 U.S. (10 Wheat.) (1825). 68 Author’s calculations based on 92 U.S. (1890) and 91 U.S. (1876). 69 Author’s calculations based on 271 U.S. (1927); 270 U.S. (1926); and 269 U.S. (1926). 70 Author’s calculations based on 428 U.S. (1978); 427 U.S. (1978); 426 U.S. (1978); 425 U.S. (1978); 424 U.S. (1977); and 423 U.S. (1977), as well as all slip opinions for October Term 2010. 71 Based on the author’s review of 428 U.S.; 427 U.S.; 426 U.S.; 425 U.S.; 424 U.S.; 423 U.S.; 272 U.S.; 271 U.S.; 270 U.S.; 93 U.S.; 92 U.S.; and 23 U.S. (10 Wheat.), as well as the slip opinions for October Term 2010, the average total number of opinion pages per case was 11.5, 6.6, 6.7, 26.9, and 30 during February Term 1825 and October Terms 1875, 1925, 1975, and 2010, respectively. 72 See 1 L. Ed. (1882); COHEN & O’CONNOR, supra note 20, at 5 (describing the commencement of the Lawyers’ Edition in 1882). 554 HARVARD LAW REVIEW [Vol. 128:540 pany commenced publishing its competing Supreme Court Reporter one year later in 1883.73 Congress confirmed the established nature of these “advance” publications in 1922 when it enacted legislation requiring publication of Supreme Court opinions by the Government Printing Office (GPO).74 This reform was part of a broader legislative effort, promoted by the new Chief Justice Taft, to improve the quality and efficiency of the Court’s decisionmaking.75 The 1922 legislation expressly provided for GPO’s publication of both the final bound volumes of the United States Reports, beginning with volume 257, and what the statute referred to as “advance copies” of opinions in “pamphlet installments.”76 Chief Justice Taft also made slip opinions routinely available to the public, converting what had commenced as an idiosyncratic process in which page proofs were “obtainable only from the Court-selected private printer or from a justice.”77 The Bureau of National Affairs’s United States Law Week began to publish slip opinions on September 5, 1933.78 And, by October Term 1946, anyone could subscribe to receive slip opinions directly from GPO.79 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 73 See 1 S. Ct. (1883); COHEN & O’CONNOR, supra note 20, at 5 (describing the commencement of the Supreme Court Reporter in 1883). 74 See Act of July 1, 1922, Pub. L. No. 67-272, § 1, 42 Stat. 816, 816. The immediate impetus for the legislation was that the Reporter was unable to find a publisher for the United States Reports because the statutorily established maximum price per volume deprived the publisher of the opportunity to make a reasonable return on its investment. See H.R. REP. NO. 67-963, at 2 (1922). The statute provided that GPO should print, bind, and issue the United States Reports “within eight months after said decisions have been rendered by the Supreme Court.” Act of July 1, 1922, § 1, 42 Stat. at 816. 75 See Jurisdiction of Circuit Courts of Appeals and United States Supreme Court, Pay of Supreme Court Reporter: Hearings Before the H. Comm. on the Judiciary, 67th Cong. (1922) (statement of William Howard Taft, C.J., Supreme Court of the United States), in 8 THE COLLECTED WORKS OF WILLIAM HOWARD TAFT 395, 395–422 (David H. Burton & Francis Graham Lee eds., 2004) [hereinafter Testimony of Chief Justice Taft Before the House Judiciary Committee]; ALPHEUS THOMAS MASON, WILLIAM HOWARD TAFT: CHIEF JUSTICE 88–120 (1965). 76 Act of July 1, 1922, § 1, 42 Stat. at 816; see also id. § 4, 42 Stat. at 818 (noting that the number of “advance pamphlet installments” to be printed by GPO is to be determined by the Reporter); FREDERICK C. HICKS, MATERIALS AND METHODS OF LEGAL RESEARCH 102 (2d ed. 1933) (describing GPO’s printing of “advance copies of the decisions in pamphlet installments” and slip opinions immediately after the opinions’ announcements). 77 THE OXFORD COMPANION TO THE SUPREME COURT OF THE UNITED STATES 933 (Kermit L. Hall et al. eds., 2d ed. 2005) (note by Professor Peter Fish). 78 See 1 U.S.L.W. 16 (Sept. 5, 1933). 79 Current Comment, Walter Wyatt Announces Availability of Supreme Court Slip Opinions, 40 LAW LIBR. J. 102, 102 (1947). The differences between the slip opinions and the preliminary prints are several. The former are the Court’s opinions in individual cases, issued within a day or two of the opinion announcement. Id. The latter contain, in a single paperback volume, all the opinions with “syllabi, names of counsel, indices, tables of cases, and other editorial additions,” in addition to all of the Court’s orders and other contemporaneously published rulings. Id.; see also THE OXFORD COMPANION TO THE SUPREME COURT OF THE UNITED STATES, supra note 77, at 20–21, 932–33 (defining “advance sheets,” id. at 20, and “slip opinion,” id. at 932). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 555 Third, and finally, the Court developed a series of pathways for revising opinions after initial publication. Some of these pathways are more transparent than others. It is to these varied procedures that this Article next turns.80 II. HISTORICAL AND CURRENT PATHWAYS FOR REVISING SUPREME COURT OPINIONS AFTER INITIAL PUBLICATION: THE QUESTION OF TRANSPARENCY A close examination of the Court’s opinions over the past two hundred–plus years reveals both that the Court makes a large number of revisions to its opinions following initial publication and that the process for revising Supreme Court opinions has changed considerably over time, especially since the early years. With regard to the former, the Court makes all sorts of revisions, ranging from the most mundane to the most intriguing, with the vast majority not surprisingly falling into the first category. With regard to the latter, several distinct pathways have emerged for revising opinions with varying degrees of transparency. Two are completely transparent. The first is formal publication of errata in the United States Reports. A list of errata typically singles out precise words or phrases as formal errors, along with corresponding deletions, additions, and substitutions to correct the errors. A second option, similarly transparent, is to publish an order in the United States Reports that formally revises a previously published opinion. The final pathway is the least transparent, apart from the formal announcement of its existence at the top of each slip opinion and preliminary print. Under this final pathway, the Reporter, in consultation with the Justices, corrects so-called “formal errors” in previously published opinions. The number of corrections made in this manner, through “change pages,” dwarfs those made under either of the other pathways, and the corrections range from the most precise technical changes to rewordings of plainly substantive import.81 But, unlike ei––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 80 The Reporter remains an important officer of the Court, but the role is far different from what it was in the early nineteenth century. See POPKIN, supra note 24, at 66, 79–82. Substantively, Justices draft their own written opinions, and the Court publishes early slip opinions. See id. at 81–82. Symbolically, the Reporter’s name no longer appears on the binding, a practice that ended when GPO assumed responsibility for publishing the United States Reports in 1922. Id. at 81. 81 The sheer number of changes made by the Court through “change pages” for each volume of the United States Reports is huge. Almost every page includes numerous changes. See infra note 111. But, as described in more detail below, the vast majority of those changes are technical updates — for instance, to reflect changes in pagination in the Court opinion cited, correction of citation form, or correction of truly incidental typographical errors. Precisely because the changes made by the “change page” method lack transparency, it is not possible to know with any level of 556 HARVARD LAW REVIEW [Vol. 128:540 ther the errata lists or the orders amending prior opinions, these changes are deliberately made hard to discover by the Court. Each of these three pathways, and their associated history, is described below. First, however, the varying institutional pressures for opinion revision are outlined and the differing types of changes made are more fully catalogued. A. The Institutional Reasons for Revision The Court’s practice of revising its opinions is surprising to most people, including those who follow the Court, and naturally raises the question: Why? Why does the Court make mistakes notwithstanding the talent of its personnel and the intensity of its internal review procedures? Why does the Court not do more to eliminate mistakes prior to publication of the slip opinions? And why does the Court insist on correcting all of its mistakes? The answers to some of these questions are quite obvious, but to others far less so. One obvious answer is that everyone needs a good editor, and Supreme Court Justices are no exception, even those who are especially talented writers. In internal Court correspondence, Chief Justice Stone described himself as “probably the most ineffective proof reader who ever sat on the Bench.”82 And the central role of the opinion, as the Court’s ultimate work product, makes it essential to write and edit carefully, which includes review and revision. Less obvious is why mistakes persist after publication of the slip opinion. After all, the structure of opinion writing within the Court provides ample opportunity for close scrutiny that invites revision. After the Justices vote at conference, the senior Justice in the majority assigns the responsibility of drafting the majority opinion. But that opinion becomes the opinion of the Court only if a majority of Justices subsequently join it. The Court’s internal procedures involve drafting opinions within chambers followed by formal circulation for careful review and harsh criticism by other chambers. Each Justice’s chambers relies on suggestions made by other chambers — including by Justices who have decided to join the opinion or are contemplating joining the opinion (and perhaps are conditioning joining on certain changes) — and, of course, responds to criticisms of the draft opinion in draft dissents.83 The internal revision process also extends to the ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– certainty to what extent substantive changes are made, but what is clear is that significant substantive changes can be and are made by this method. See infra pp. 593–95, 599–600. 82 Letter from Harlan F. Stone, Assoc. Justice, Supreme Court of the United States, to Ernest Knaebel, Reporter of Decisions, Supreme Court of the United States (Mar. 6, 1931), Stone Papers B82. 83 See generally SUSAN LOW BLOCH ET AL., INSIDE THE SUPREME COURT 574–635 (2d ed. 2008); WILLIAM H. REHNQUIST, THE SUPREME COURT 252–66 (2d ed. 2001). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 557 Reporter’s office, which scrutinizes opinions for errors in spelling, grammar, word usage, citation form, and cited authority before transmitting them for printing, publication, and distribution.84 The more elusive inquiry is therefore why, given all this exceedingly intense and skilled scrutiny, revision is still necessary after the Court’s opinion is first announced and published. As the Reporter himself acknowledged in private correspondence to the Chief Justice in 1984, by making a “considerable number of corrections and editorial changes in the Court’s opinions after their announcement and prior to their publication in the United States Reports[] . . . we actually operate a system that is completely at odds with general publishing practices.”85 The most fundamental reason is that mistakes are inevitable and will persist even after the rigorous reviewing process. The Justices and their chambers can, of course, reduce the number of mistakes by being more rather than less careful and by being more rather than less skilled. But no matter how much time and skill are applied, the possibility of mistakes cannot be eliminated. Of course, sometimes the case itself arises out of the kind of conflict that imposes a deadline on the Court, such as in Bush v. Gore,86 when the Court had to issue a ruling within days of granting review because federal constitutional procedures for deciding presidential elections would otherwise have begun to be triggered irreversibly.87 There are, however, far broader and more systemic reasons rooted in the Court’s structure, decisionmaking process, and governmental function for why these mistakes persist. In theory, the Court can take as long as it wants to decide a case and can delay publication until every possible layer of review has been applied. But there are costs to delay, which is why the Court instead has chosen to publish its initial opinions earlier, before they are reviewed to the full extent the Justices believe to be necessary for the “final” and “official” versions. That is also why the Court has voluntarily embraced a series of internal rules designed to put pressure on its members to issue rulings sooner rather than later. The Justices know the obvious. There is no substitute for a deadline of some consequence to promote closure in decisionmaking. And there are significant institutional and societal costs to cases that drag on, which would likely happen absent an internal rule that admits of ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 84 85 EUGENE GRESSMAN ET AL., SUPREME COURT PRACTICE 30–32 (9th ed. 2007). Letter from Henry C. Lind, Reporter of Decisions, Supreme Court of the United States, to Warren E. Burger, Chief Justice, Supreme Court of the United States 1 (May 11, 1984), White Papers BI:622/F5. 86 531 U.S. 98 (2000). 87 Notwithstanding these extraordinary time pressures, a review of Bush did not reveal any noteworthy changes in the opinions between the time of the publication of the original slip opinions and the final bound volumes of the United States Reports. 558 HARVARD LAW REVIEW [Vol. 128:540 no exceptions. With regard to the former, judicial chambers can quickly develop a backlog of work. And that backlog naturally becomes increasingly difficult to overcome over time as the law clerks most knowledgeable about a matter are replaced by new clerks who need time to get up to speed and who were not there either when the case was argued or when the Justices deliberated and recorded their preliminary votes. Broader societal interests also can be harmed by protracted judicial decisionmaking. The longer it takes the Court to decide, the longer legal uncertainty persists on an important legal issue, which necessarily undermines basic planning and investment, often in very personally profound and economically significant ways. For these reasons, once a sufficient consensus has been formally recorded to support a draft majority opinion, there is invariably significant institutional pressure within the Court to issue that ruling sooner rather than later, providing, of course, sufficient time for the preparation of dissenting and concurring opinions. Those Justices in the majority have no interest in delays that promote waffling, further negotiation, and possible shifts, let alone any unexpected changes in personnel that could occur the longer and longer it takes to issue the opinion. In January 1981, Justice Blackmun described these precise motivations in internal correspondence to the Reporter as part of an effort to explain why the Court operated “on a strange and ‘reverse’ basis, where the professional editing is done after initial public release. . . . I know of no other situation where delayed editing of this kind is the routine.”88 Justice Blackmun elaborated: There is a natural, and understandable, inclination to “rush to judgment” in the sense that every Justice, when all the votes are in, wants to get the decision down immediately and without delay. He wants to get “on the scoreboard.” I suppose, too, that he wants to guard against any last minute shifting of a vote, particularly in a close case. I suspect, therefore, that an element of impatience will be evident and will be rationalized with the thought that editing, after all, is “secondary” and “those details can be ironed out later.” I think that this attitude is wrong, but it is likely to be present.89 One of the most significant internal rules the Court imposes to promote closure relates to the summer recess. The Court’s practice is to reschedule for oral argument the following October Term any case that has not been decided prior to adjournment for the summer.90 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 88 Letter from Harry A. Blackmun, Assoc. Justice, Supreme Court of the United States, to Henry C. Lind, Reporter of Decisions, Supreme Court of the United States 1 (Jan. 5, 1981), Blackmun Papers B1425/F13. 89 Id. at 2. 90 See Stephen Wermiel, SCOTUS for Law Students: It Happens Every Spring, SCOTUSBLOG (June 20, 2012, 12:06 PM), http://www.scotusblog.com/2012/06/scotus-for-law-students-it-happens -every-spring-sponsored-by-bloomberg-law [http://perma.cc/83PB-WXVZ]. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 559 That tradition places a huge premium on ensuring that the Court reaches closure prior to adjournment. And the Court very rarely carries a case over to the next Term because of its inability to reach consensus. Reargument is generally reserved for instances in which the Justices want the parties to address an additional issue91 or when the Court lacks a full bench of nine Justices.92 The deliberate upshot is substantial pressure to produce opinions during the closing weeks before the summer recess. The Court’s practice is to require circulation of all draft opinions by the beginning of June.93 That timetable leaves relatively little time for revision. Because, moreover, the opinions being announced immediately before the summer recess are disproportionately the cases that have generated the most conflict within the Court (and therefore have taken more time), the limited time for review is particularly problematic. The Justices find themselves faced with simultaneously circulating majority, concurring, and dissenting opinions, each of which responds to language in the others and is subject to revision. Reaching closure on the wording of all opinions is not easy in a Term’s final days, hours, and minutes. Each Justice authoring a competing opinion seeks the last word, so closure can be elusive.94 Last-minute changes are the most risky because there is less time for review and revision, so the chance of a mistake is great. And as the final hours of the Term approach, the Supreme Court Reporter must struggle to keep up and carefully scrutinize each majority, concurring, and dissenting opinion. More than a half century ago, Justice Frankfurter was sufficiently concerned about the ability of the Justices to decide cases under this kind of time pressure that he recommended in a private memorandum to the other Justices that the Court change its internal procedures to eliminate the hard deadline of the summer recess: Changes in majority opinion and dissent, changes often fundamental to the theory of decision and certainly to its precise formulation, criss-crossed the Court in circulations that continued until almost the last minute before we went on the bench. It surely cannot be denied that such an atmosphere is hardly conducive to the full and mature consideration by every ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 91 See, e.g., Lyle Denniston, Kiobel to Be Expanded and Reargued, SCOTUSBLOG (Mar. 5, 2012, 2:01 PM), http://www.scotusblog.com/2012/03/kiobel-to-be-reargued [http://perma.cc/A9SA -M5PU]. 92 See Linda Greenhouse, Mysteries of Tie Votes and Calls for Reargument, N.Y. TIMES, Apr. 3, 1985, at A20. 93 See DAVID M. O’BRIEN, STORM CENTER: THE SUPREME COURT IN AMERICAN POLITICS 206 (8th ed. 2008). 94 See id. at 275–87. 560 HARVARD LAW REVIEW [Vol. 128:540 member of the Court of the implications of an opinion to which he has formally subscribed.95 Nor are the cases decided at the end of the Term a relatively small part of the Court’s docket. Not only do they typically include some of the most controversial and high-profile cases, they also reflect a disproportionately high number of the merits rulings. For instance, on June 12, 2013, two weeks before the Supreme Court was scheduled to recess for the summer, the Court had decided fifty-four of the cases that had been briefed and argued during the October 2012 Term. The Court had twenty-three cases yet to rule on, just under one-third of its total docket.96 Those twenty-three cases included the Texas affirmative action case, Fisher v. University of Texas at Austin,97 argued in October;98 a constitutional challenge to the federal Voting Rights Act,99 Shelby County v. Holder,100 argued in February;101 two cases involving the status of gay marriage under the U.S. Constitution, Hollingsworth v. Perry102 and United States v. Windsor,103 both argued in March;104 and the application of the Indian Child Welfare Act105 to adoption by a non-Indian parent under state law, Adoptive Couple v. Baby Girl,106 argued during the Court’s last argument session in April.107 In addition to the considerable challenge of reaching closure on five extremely difficult and divisive cases, the Court had to resolve eighteen other ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 95 Memorandum from Felix Frankfurter, Assoc. Justice, Supreme Court of the United States, to the Justices of the Supreme Court of the United States 9 (Sept. 25, 1961), microformed on Felix Frankfurter Papers, at Part 2, Reel 92 (Univ. Publ’ns of Am., Inc.) (on file with the Harvard Law School Library). 96 See Kedar S. Bhatia, Stat Pack for October Term 2012, SCOTUSBLOG 2 (June 12, 2013), h t t p : / / s b l o g . s 3. a m a z o n a w s . c o m / w p - c o n t e n t / u p l o a d s / 20 1 3/ 0 6/ S C O T U S b l o g _ s t a t _ p a c k _ 06 - 1 2- 20 1 3 .pdf [http://perma.cc/X25G-FTHP]; Kedar S. Bhatia, Merits Cases Remaining for October Term 2012, SCOTUSBLOG (June 10, 2013, 2:38 PM), http://www.scotusblog.com/2013/06/merits-cases -remaining-for-October-term-2012 [http://perma.cc/Z2MQ-GW88]. 97 133 S. Ct. 2411 (2013). 98 Id. at 2411. There was an obvious explanation for why Fisher had not been decided after eight months, related to the sheer difficulty of reaching closure should there be (as was then expected) many competing circulating opinions. That the Court’s ruling in Fisher ended up being shorter and marked by far more consensus than anticipated is not inconsistent with this explanation for the delay. It instead strongly suggests that the Court ultimately was unable (or ran out of time) to garner a majority for a more specific ruling and therefore opted to achieve consensus by deciding less rather than more. 99 42 U.S.C. §§ 1973 to 1973bb-1 (2012). 100 133 S. Ct. 2612 (2013). 101 Id. at 2612. 102 133 S. Ct. 2652 (2013). 103 133 S. Ct. 2675 (2013). 104 Id. at 2675; Hollingsworth, 133 S. Ct. at 2652. 105 25 U.S.C. §§ 1901–1963 (2012). 106 133 S. Ct. 2552 (2013). 107 See id. at 2552. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 561 cases in just two weeks. The Court’s accomplishments each June border on the Herculean. But the risk of introduction of unintended error, as Justice Frankfurter indicated decades ago, is also high. Nor are the few weeks immediately prior to summer recess the only time such pressures exist. The Court has internal expectations at other times, including in January, when the Justices are expected to have circulated draft opinions from cases argued earlier in the Term. In addition, a senior Justice responsible for assigning a majority opinion can fairly take into account the fact that a particular Justice is behind or slow in opinion production so far that Term.108 Justices who wish to receive more opinion assignments and more coveted cases have this additional incentive to demonstrate they are “moving” their assigned cases along. In addition to time pressures, the nature of the opinion can substantially reduce the opportunity for scrutiny prior to publication. The Court’s docket is dominated numerically by lower-profile, unanimous rulings.109 The Court no doubt reviews carefully the draft opinions in unanimous cases, too, but the incentives for review in such cases may be lower, so the resulting review might naturally be commensurately less demanding. The other chambers are less motivated to provide as careful a review, and the Justice’s chambers responsible for the opinion may anticipate less exacting scrutiny and do the same. Another structural reason why the pressures for post-publication revision of Supreme Court opinions are great is because the odds of discovering an error are very high. Most errors, like most trees falling in the woods, are never seen or heard. But that is not true when the Supreme Court commits an error. The number of people who read Supreme Court opinions carefully and the varied expertise of those readers are enormous. Every word, every fact, every characterization of the facts, and every discussion of background legal doctrine is subject to close scrutiny. Errors will be discovered and reported, if not immediately, then eventually, perhaps nearly a hundred years later.110 Finally, one might fairly ask why the Court insists on correcting all errors. As further elaborated below, the Court’s current and historic practices extend to correcting mistakes of all types, large and small, including those made decades (or even a century) earlier. Do the ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 108 See REHNQUIST, supra note 83, at 260 (“I try to be as evenhanded as possible as far as numbers of cases assigned to each justice, but as the term goes on I take into consideration the extent to which the various justices are current in writing and circulating opinions that have previously been assigned.”). 109 During the past five Terms, the Court has decided — in reverse chronological order — 48%, 49%, 44%, 48%, and 46% of its cases on the merits unanimously. See Kedar S. Bhatia, Stat Pack for October Term 2013, SCOTUSBLOG 5 (July 3, 2014), http://sblog.s3.amazonaws.com/wp -content/uploads/2014/07/SCOTUSblog_Stat_Pack_for_OT13.pdf [http://perma.cc/V87Q-CNE5]. 110 See infra p. 574. 562 HARVARD LAW REVIEW [Vol. 128:540 Court’s authority and stature truly require such an unrelenting drive to correct the record without temporal limitation? No other lawmaking institution seems similarly inclined to correct everything anytime. The short answer is that the Court clearly thinks that the stature and longevity of its opinions, and perhaps its own authority, ultimately require such an unqualified commitment to exactitude. In this regard, the true significance of the Court’s practice may be what it underscores about the Court’s aspirations for itself. B. Types of Revisions The types of errors that the Court corrects are many and varied. For analytical purposes, they can be grouped roughly into several categories, with the necessary caveat that specific revisions frequently fail to fall clearly within just one category. Specific examples of all these changes are provided in section II.C, which further describes the pathways for making revisions. 1. Typographical, Spelling, Grammar, and Citation Errors. — The Court routinely corrects errors of a strictly technical nature, such as typographical mistakes in spelling, duplicate words, punctuation, spacing, and citation form.111 In recent times, just as in the past, this category has extended to the misspelling of the name of a Justice.112 And, while these kinds of errors are not the only errors that the Court considers “typographical and other formal errors,” they have traditionally been the most prevalent.113 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 111 The striking breadth of these kinds of routine proofreader’s changes, and the tremendous skill displayed by the Office of the Supreme Court Reporter, are best illustrated by a review of the official “change pages,” which detail all the changes to be made to the original before printing the final version. See Letter from Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States, to Marianne Baylor, Dir. of Content Analysis, LexisNexis (May 17, 2012) (on file with the Harvard Law School Library) (attaching Reporter’s official “change pages” for volume 557, part II of the United States Reports); Letter from Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States, to Marianne Baylor, Dir. of Content Analysis, LexisNexis (Mar. 15, 2012) (on file with the Harvard Law School Library) (attaching Reporter’s official “change pages” for volume 557, part I of the United States Reports). I received from LexisNexis the change pages for one volume of the United States Reports, and then LexisNexis declined to send additional change pages after notifying me that it was the position of the Supreme Court Reporter that LexisNexis should not send further change pages. More than fourteen months later, on September 29, 2014, LexisNexis agreed to provide me with one part of one additional volume of the United States Reports in light of the publisher’s subsequent communication with the Reporter of Decisions. However, I received that part only a few days before this Article went to final page proofs, and therefore it was too late for the contents of the additional volume to be included. See infra note 268. 112 See Errata, 483 U.S. II, II (1990) (correcting Hobbie v. Unemp’t Appeals Comm’n of Fla., 480 U.S. 136, 148 (1986) (“SEVENS, J., concurring in judgment”)); COHEN & O’CONNOR, supra note 20, at 72 n.54 (noting misspelling of Justice Daniel’s name as “Daniels”). 113 See, e.g., Letter from Henry C. Lind, Reporter of Decisions, Supreme Court of the United States, to Warren E. Burger, Chief Justice, Supreme Court of the United States 1 (Oct. 11, 1979), Brennan Papers BII:141 (describing a review of the three preceding Terms, determining that “89% 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 563 Within this same general category, the Court further corrects errors in spelling and grammar that are unlikely to be merely typographical in the sense of inadvertent. These errors are intentional in that the author intended to use the words or punctuation but was mistaken about their correctness. Such errors in spelling, grammar, and usage are frequently and generously referred to as “typographical,” though their origins are more personal than mechanical. No less than Justice Holmes misspelled “capital” as “capitol” due to his “deliberate ignorance,” as he confessed in an apologetic letter to the Reporter, further admitting of a “double blush.”114 Citation errors extend to correcting quotations of authority by adding words mistakenly omitted or deleting words mistakenly included. The Court also has internal rules of style that it applies to promote a consistent institutional voice, including the spelling of certain words, such as “marijuana” with a “j,”115 and references to the Court itself as “we” rather than “the Court.”116 2. Word Additions, Deletions, and Substitutions. — Another category of corrections involves the addition, deletion, and substitution of words and phrases. This category is distinct from the first because these changes are not implemented to correct technical mistakes such as erroneous quotations. These changes can be merely stylistic, or they can be substantive modifications triggered because the prior phrasing did not convey the Court’s (or the Justice’s) intended meaning. Sometimes the error is blatant, such as when Justice White’s opinion for the Court in United States v. Fordice117 stated: “It is illogical to think that some percentage of black students who fail to score 15 do not seek admission to one of the historically white universities because ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– of the 2,876 editorial corrections and changes suggested and approved for the preliminary prints fell within” the category of “typographical errors, misquotes, and errors in citations” as well as “all items that do not conform to the [Court’s] Style Manual”). 114 Interview by Paul R. Baier, Assoc. Professor of Law, La. State Univ. Law Ctr., with Henry Putzel, jr., Reporter of Decisions, Supreme Court of the United States, in Baton Rouge, La. 7 (Oct. 1978) (transcript on file with the Harvard Law School Library). Putzel always used the lower case “j” in “jr.” as a personal expression of respect for his father. And the United States Reports reflect his personal preference. See, e.g., Note, 375 U.S. IV, IV (1964) (announcing the new Reporter of Decisions “Henry Putzel, jr.”). 115 See Memorandum from Henry C. Lind, Reporter of Decisions, Supreme Court of the United States, to the Conference (Apr. 22, 1986), in Tony Mauro, To the Bag, 12 GREEN BAG 2D 11, 14 (2008) (“I have now received all votes but one. There were four votes for ‘marijuana,’ one vote for ‘marihuana,’ and three Justices gave me their proxies, which I have cast for ‘marijuana.’ Therefore, the spelling from now on should be with a ‘j.’ I hope that this will settle the matter.”). 116 See Edited Opinion, Polar Tankers, Inc. v. City of Valdez, No. 08-310, slip op. at 4 (U.S. June 15, 2009) (rev. n.d.), in Reporter of Decisions Office, Supreme Court of the United States, Change Pages: Volume 557, Part I 1661, 1667 (2009) [hereinafter Change Pages: 557 U.S. Part I] (on file with the Harvard Law School Library) (changing “the Court has” in two instances to “we have”). 117 505 U.S. 717 (1992). 564 HARVARD LAW REVIEW [Vol. 128:540 of this automatic admission standard.”118 As pointed out by an academic in a letter to Justice White, “[t]he context suggests that the sentence needs a negative, which is omitted”119 and should instead say, “It is [not] illogical to think . . . .”120 Subsequent Court publications of the Court’s opinion in Fordice rewrote the sentence to state “It is logical to think . . . .”121 Quite often, however, these errors involve precision in word choice.122 For example, the Court has substituted “annexing” for “admitting,”123 “ensure” for “insure”124 (or “ensure” for “assure”125), “has authority to watch” for “watches,”126 and “hawkers or peddlers” for “itinerant vendors of merchandise.”127 Not surprisingly, the Justices can have sharply contrasting views on the proper definition of some words and even their existence. Justice Blackmun notified his colleagues he “would join no opinion this year which used the word pa––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 118 United States v. Fordice, 60 U.S.L.W. 4769, 4774 n.9 (U.S. June 26, 1992) (Nos. 90-1205, 906855). 119 Letter from Michael H. Hoffheimer, Assoc. Professor of Law, Univ. of Miss. Law Ctr., to Byron R. White, Assoc. Justice, Supreme Court of the United States (July 7, 1992), White Papers BII:174/F8. 120 Id. (internal quotation mark omitted). 121 Fordice, 505 U.S. at 735 n.9. 122 For instance, in Safford Unified School District No. 1 v. Redding, 129 S. Ct. 2633 (2009), Justice Ginsburg revised her separate concurring opinion at the Reporter’s suggestion. She added the words “on her” to avoid creating the misapprehension that the respondent might have hidden medicine “in her body.” See Edited Opinion, Safford Unified Sch. Dist. No. 1 v. Redding, No. 08479, slip op. at 1 (U.S. June 25, 2009) (rev. n.d.) (Ginsburg, J., concurring in part and dissenting in part), in Reporter of Decisions Office, Supreme Court of the United States, Change Pages: Volume 557, Part II 2043, 2062 (2009) [hereinafter Change Pages: 557 U.S. Part II] (on file with the Harvard Law School Library). To similar effect, Justice Ginsburg amended her 2009 dissenting opinion in Ricci v. DeStefano, 129 S. Ct. 2658 (2009), changing her wording from “One is left to wonder what cases would meet the standard and why the Court is so sure this case does not” to “One is left to wonder what cases would meet the standard and why the Court is so sure cases of this genre do not.” Edited Opinion, Ricci v. DeStefano, Nos. 07-1428, 08-328, slip op. at 20 (U.S. June 29, 2009) (rev. n.d.) (Ginsburg, J., dissenting), in Change Pages: 557 U.S. Part II, supra, at 2239, 2311 (emphasis added). 123 United States v. Texas, 340 U.S. 848, 848 (1950) (amending United States v. Texas, 339 U.S. 707 (1950)). 124 Compare Nat’l Ass’n of Home Builders v. Defenders of Wildlife, 551 U.S. 644, 673 (2007) (Stevens, J., dissenting) (“ensure”), with Nat’l Ass’n of Home Builders v. Defenders of Wildlife, Nos. 06-340, 06-549, slip op. at 1 (U.S. June 25, 2007) (Stevens, J., dissenting) (“insure”). 125 See Edited Opinion, Coeur Alaska, Inc. v. Se. Alaska Conservation Council, Nos. 07-984, 07-990, slip op. at 3 (U.S. June 22, 2009) (rev. n.d.) (Breyer, J., concurring), in Change Pages: 557 U.S. Part I, supra note 116, at 1935, 1969; Edited Opinion, Polar Tankers, Inc. v. City of Valdez, No. 08-310, slip op. at 9 (U.S. June 15, 2009) (rev. n.d.) (opinion of Breyer, J.), in Change Pages: 557 U.S. Part I, supra note 116, at 1661, 1672. 126 Cloverleaf Butter Co. v. Patterson, 315 U.S. 786, 786 (1942) (amending earlier version of Cloverleaf Butter Co. v. Patterson, 315 U.S. 148 (1942)) (substituting “is subject to” for “comes under,” “has authority to watch” for “watches,” and “subject” for “subjected”; and striking the word “continuous”). 127 Healy v. Ratta, 292 U.S. 607, 607 (1934) (amending earlier version of Healy v. Ratta, 292 U.S. 263 (1934)). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 565 rameter,”128 which he believed his colleagues and advocates were misusing.129 And Justice Harlan insisted on the use of the word “suppletive” in one opinion,130 which neither the Reporter nor future members of the Court considered a legitimate word at all.131 Other times, the changes could have significant substantive import. For instance, in Perry v. United States,132 Chief Justice Hughes revised his initial majority slip opinion to substitute the words “could have” for “has.”133 Without more, such a revision might appear merely technical in origin and import. Not so in this instance. Perry was one of four extremely significant and highly controversial cases that the Court decided on the same day in 1935 regarding the legality of restrictions that Congress had placed on the ownership of gold.134 The Chief’s ex post facto revision affected “an integral part of the opinion”135 wherein the Court justified its ruling against the plaintiff on the ground that although Congress lacked constitutional authority to abrogate “gold clauses” in its own obligations, such as legal tender, the plaintiff had failed to establish that the congressional action had caused the plaintiff an actual monetary loss.136 In what was a limited victory for the government,137 the Court’s reasoning that the plaintiff had suffered “no damage”138 was described as “confusing” even “at the hundredth” reading.139 The Court’s original slip opinion had described the “question of value” for assessing damages as “requir[ing] a consideration of the purchasing power of the dollars which the plaintiff has received,”140 while the subsequently published bound volume opinion shifted the inquiry ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 128 Letter from Harry A. Blackmun, Assoc. Justice, Supreme Court of the United States, to Henry Putzel, jr., Reporter of Decisions, Supreme Court of the United States (Nov. 17, 1975), Blackmun Papers B1425/F11. 129 Id.; cf. Letter from Harry A. Blackmun, Assoc. Justice, Supreme Court of the United States, to Henry Putzel, jr., Reporter of Decisions, Supreme Court of the United States (Mar. 29, 1976), Blackmun Papers B1425/F12 (discussing misuse of “parameter” by an Assistant Solicitor General in oral argument). 130 Labine v. Vincent, 401 U.S. 532, 540 (1971) (Harlan, J., concurring). 131 See Letter from Henry Putzel, jr., Reporter of Decisions, Supreme Court of the United States, to Byron R. White, Assoc. Justice, Supreme Court of the United States (Dec. 8, 1976), White Papers BI:366/F9. 132 294 U.S. 330 (1935). 133 Compare Perry v. United States, 2 U.S.L.W. 550, 552 (U.S. Feb. 18, 1935) (No. 532) (“has received”), with Perry, 294 U.S. at 357 (“could have received”). 134 Henry M. Hart, Jr., The Gold Clause in United States Bonds, 48 HARV. L. REV. 1057, 1057 n.2 (1935). 135 Id. at 1078. 136 Id. at 1077–78. 137 2 MELVIN I. UROFSKY & PAUL FINKELMAN, A MARCH OF LIBERTY 677–78 (2d ed. 2002). 138 Hart, supra note 134, at 1077. 139 Id. at 1057; see id. at 1077–81. 140 Perry v. United States, 2 U.S.L.W. 550, 552 (U.S. Feb. 19, 1935) (No. 532) (emphasis added). 566 HARVARD LAW REVIEW [Vol. 128:540 to “requir[ing] a consideration of the purchasing power of the dollars which the plaintiff could have received.”141 The difference in phrasing does not reflect a simple typographical error, and the correction is plainly substantive in nature. To make just such a substantive change is undoubtedly why the Chief amended the wording, albeit without formal notice. Finally, in Virginia State Board of Pharmacy v. Virginia Citizens Consumer Council, Inc.,142 decided in 1976, Chief Justice Burger deleted ten lines that appeared in the original slip opinion version of his concurring opinion, but which are not included in his concurring opinion as published in the United States Reports.143 Virginia State Board of Pharmacy was the Court’s seminal case holding that commercial speech is entitled to First Amendment protection,144 and on that ground the Court struck down a state law that had made unlawful a pharmacy’s advertisement of prescription drug prices.145 What the Chief Justice apparently considered a “formal error” warranting deletion was his extensive slip opinion discussion of the nature of the “tasks a professional pharmacist performs” in dispensing “dosage units already prepared by the manufacturer and sold to the pharmacy in that form.”146 According to the Chief’s concurring slip opinion, “it is clear that in this regard he no more renders a true professional service than does a clerk who sells lawbooks.”147 The Chief decided to eliminate much of this discussion from the final bound version, perhaps because some in the pharmaceutical industry took umbrage at his characterization of the profession. If so, Virginia State Board of Pharmacy offers an example of a very different kind of “error”: words that some members of the public found insulting. 3. Erroneous Characterizations of Facts, the Record, the Positions of the Parties, the Positions of the Other Justices, Background Law, and the Court’s Opinion. — Supreme Court opinions are replete with statements of fact. Those facts may be rooted in the record of the case, including assertions of the parties, or outside the record, including both commonly known facts subject to judicial notice and facts based on a Justice’s own research. The Justices commit errors of all varieties in ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 141 142 143 Perry v. United States, 294 U.S. 330, 357 (1935) (emphasis added). 425 U.S. 748 (1976). Compare Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, Inc., 44 U.S.L.W. 4686, 4693 (U.S. May 25, 1976) (No. 74-895) (Burger, C.J., concurring), with Va. State Bd. of Pharmacy, 425 U.S. at 773–74 (Burger, C.J., concurring). This example was brought to my attention by Professor Stephen Wasby. 144 See Cass R. Sunstein, The Supreme Court, 1995 Term — Foreword: Leaving Things Undecided, 110 HARV. L. REV. 4, 82 (1996). 145 See Va. State Bd. of Pharmacy, 425 U.S. at 759–62. 146 Va. State Bd. of Pharmacy, 44 U.S.L.W. at 4693. 147 Id. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 567 such factual assertions. Examples include misstating who was President of the United States in 1799,148 mischaracterizing the legal arguments of the Commissioner of Internal Revenue,149 erroneously labeling a prior “indictment” as a “conviction[],”150 miscalculating the number of copyrighted works that had entered the public domain,151 describing someone as a “librarian” rather than a “professor,”152 and misidentifying a legislator who made a particular comment.153 The Justices have also misquoted other opinions in the same case, presumably because they were working off outdated drafts. For instance, in Missouri ex rel. Missouri Insurance Co. v. Gehner,154 decided in 1930, then–Associate Justice Stone and Justices Holmes and Brandeis omitted language because “a statement in the majority opinion as originally circulated . . . was afterward taken out by the writer of the opinion.”155 Recent examples include Boumediene v. Bush156 and Hamdan v. Rumsfeld.157 Such errors can also reveal instances ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 148 Compare Am. Ins. Ass’n v. Garamendi, No. 02-722, slip op. at 16 (U.S. June 23, 2003) (“Making executive agreements to settle claims of American nationals against foreign governments is a particularly longstanding practice, the first example being as early as 1799, when the Washington administration settled demands against the Dutch Government . . . .”), with Am. Ins. Ass’n v. Garamendi, 539 U.S. 396, 415 (2003) (“Making executive agreements to settle claims of American nationals against foreign governments is a particularly longstanding practice, the first example being as early as 1799, when the Adams administration settled demands against the Dutch government . . . .”). 149 Helvering v. Stuart, 317 U.S. 602, 602–03 (1942) (amending earlier version of Helvering v. Stuart, 317 U.S. 154 (1942)). 150 Compare District of Columbia v. Heller, No. 07-290, slip op. at 49 (U.S. June 26, 2008) (“convictions”), with District of Columbia v. Heller, 554 U.S. 570, 621 (2008) (“indictment”). 151 Compare Eldred v. Ashcroft, No. 01-618, slip op. at 21 (U.S. Jan. 15, 2003) (Stevens, J., dissenting) (“[O]nly one year’s worth of creative work — that copyrighted in 1923 — has fallen into the public domain during the last 80 years.”), with Eldred v. Ashcroft, 537 U.S. 186, 241 (2003) (Stevens, J., dissenting) (“[W]ith the exception of works which required renewal and which were not renewed, no copyrighted work created in the past 80 years has entered the public domain or will do so until 2019.”). 152 Compare Ledbetter v. Goodyear Tire & Rubber Co., No. 05-1074, slip op. at 6 (U.S. May 29, 2007) (“librarian”), with Ledbetter v. Goodyear Tire & Rubber Co., 550 U.S. 618, 626 (2007) (“professor”). 153 Compare Boumediene v. Bush, No. 06-1195, slip op. at 49 (U.S. June 12, 2008) (attributing quotation to Senator Graham), with Boumediene v. Bush, 553 U.S. 723, 778 (2008) (attributing same quotation to Senator Kyl). 154 281 U.S. 313 (1930). 155 Letter from Harlan F. Stone, Assoc. Justice, Supreme Court of the United States, to Ernest Knaebel, Reporter of Decisions, Supreme Court of the United States (May 6, 1930), Stone Papers B82. 156 553 U.S. 723. Compare Boumediene, No. 06-1195, slip op. at 24, 25 (Roberts, C.J., dissenting), with Boumediene, 553 U.S. at 823, 824 (Roberts, C.J., dissenting) (revising two different quotations from majority opinion). 157 548 U.S. 557 (2006). In Hamdan, the majority opinion was revised to change “jurisdictionstripping” to “jurisdiction-ousting,” and “jurisdiction-conferring” to “jurisdiction-creating,” in apparent response to Justice Scalia’s change in his dissent, which the majority apparently did not 568 HARVARD LAW REVIEW [Vol. 128:540 when individual Justices have changed their opinions and votes during the internal deliberations. Working off an old draft, a slip opinion may mistakenly refer to another opinion as a majority, plurality, concurring, or dissenting opinion when, because of a vote change, it no longer has that same, precise status.158 Closely related are misstatements about background law and about the Court’s opinion. Here, too, there is no question that the change is substantive in nature and may have substantive import. Quite often, the misstatement in question may be “mere dictum,” but even nondicta misstatements can be revised without a corresponding formal modification of the Court’s judgment or holding. Examples include mischaracterization of the operation of federal and state statutes,159 and of the Court’s own prior precedent.160 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– spot until after the original opinion was issued. Compare Hamdan v. Rumsfeld, No. 05-184, slip op. at 17 (U.S. June 29, 2006), with Hamdan, 548 U.S. at 582. 158 My colleague, Professor Larry Tribe, identified for me just such a discrepancy in Justice Blackmun’s separate opinion in Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992). The slip opinion version of Justice Blackmun’s opinion takes issue with “the plurality’s views of homosexuality as sexual deviance,” Planned Parenthood of Se. Pa. v. Casey, 60 U.S.L.W. 4795, 4825 n.11 (U.S. June 29, 1992) (No. 91-744) (Blackmun, J., concurring in part, concurring in the judgment in part, and dissenting in part) (emphasis added), while the final bound volume of the United States Reports refers instead to “The Chief Justice’s views of homosexuality as sexual deviance,” Casey, 505 U.S. at 941 n.11 (Blackmun, J., concurring in part, concurring in the judgment in part, and dissenting in part) (emphasis added). The difference is consistent with assumptions that vote changes in Casey changed the outcome in that case. See Linda Greenhouse, Documents Reveal the Evolution of a Justice, N.Y. TIMES, Mar. 4, 2004, at A1. 159 In Morton Salt Co. v. G.S. Suppiger Co., 314 U.S. 488 (1942), the Supreme Court modified its description of the legal effect of the Clayton Act. See Morton Salt Co. v. G.S. Suppiger Co., 315 U.S. 788, 788 (1942) (amending earlier version of Morton Salt, 314 U.S. 488). Compare Morton Salt Co. v. G.S. Suppiger Co., 10 U.S.L.W. 4126, 4127 (U.S. Jan. 5, 1942) (No. 49), with Morton Salt, 314 U.S. at 490 (omitting statement that the Clayton Act “imposes criminal penalties for engaging in specified unlawful trade practices tending to monopoly and restraint of trade”). In Cudahy Packing Co. v. Holland, 315 U.S. 357 (1942), the Supreme Court eliminated footnotes and significant text to delete an apparent mischaracterization of the Bituminous Coal Act of 1937. See Cudahy Packing Co. v. Holland, 315 U.S. 788, 788 (1942) (amending earlier version of Cudahy Packing, 315 U.S. 357). Compare Cudahy Packing Co. of La. v. Holland, 10 U.S.L.W. 4252, 4253 (U.S. Mar. 2, 1942) (No. 245), with Cudahy Packing, 315 U.S. at 364–65. In Hysler v. Florida, 315 U.S. 411 (1942), the Court struck five sentences that apparently mischaracterized the Florida Supreme Court’s responsibilities under Florida state law. See Hysler v. Florida, 316 U.S. 642, 642 (1942) (amending earlier version of Hysler, 315 U.S. 411). Compare Hysler v. Florida, 10 U.S.L.W. 4236, 4238 (U.S. Mar. 2, 1942) (No. 64), with Hysler, 315 U.S. at 422. In Texas & Pacific Railway Co. v. Pottorff, 291 U.S. 245 (1934), the Supreme Court deleted several sentences that had described the nature of government oversight of national banks and what activity constituted a federal crime. See Tex. & Pac. Ry. Co. v. Pottorff, 291 U.S. 649, 649 (1934) (amending earlier version of Tex. & Pac. Ry., 291 U.S. 245); see also Tex. & Pac. Ry., 291 U.S. at 259 n.15. 160 For instance, in Lapides v. Board of Regents of University System of Georgia, 535 U.S. 613 (2002), the Court revised its slip opinion to correct an error in its characterization of a case, Ford Motor Co. v. Department of Treasury, 323 U.S. 459 (1945), that it was overruling. The question presented in both Lapides and Ford was whether a state that litigated a case in the lower courts on the merits, without raising an Eleventh Amendment defense, could invoke that defense in the Supreme Court. Lapides, 535 U.S. at 616, 621–23; see also Lapides v. Bd. of Regents of the Univ. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 569 The most substantive and potentially significant category of revisions pertains more directly to the reasoning, ruling, and formal judgment of the Court. These kinds of changes are the least frequent but do not appear to be especially unusual. The Court has not categorically shied away from adding and deleting words, phrases, and sentences of substantial import to correct erroneous characterizations of what the Justices intended to convey. For instance, in Virginian Railway Co. v. System Federation No. 40,161 then–Associate Justice Stone subsequently added the critical word “not” to the first sentence of the final paragraph of the Court’s opinion, which summarized the Court’s holding: “It suffices to say that the Norris-LaGuardia Act can affect the present decree only so far as its provisions are found not to conflict with those of § 2, Ninth, of the Railway Labor Act . . . .”162 Likewise of clear substantive import, in Connick v. Myers,163 Justice White eliminated a sentence from a Court opinion he had authored that had asserted that because a particular inquiry into the protected status of speech is one of law, not fact, the Court is “not bound to the views of the District Court unless clearly erroneous.”164 Five years later, to similar substantive effect, Justice White changed his opinion for the Court in Monessen Southwestern Railway Co. v. Morgan165 to eliminate entirely the unambiguous statement of law in the originally published slip opinion that “there is no right to a jury trial in [Longshoremen’s and Harbor Workers’ Compensation Act] actions.”166 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Sys. of Ga., No. 01-298, slip op. at 1, 7–8 (U.S. May 13, 2002). Overruling Ford, the Court concluded that the State had waived its defense in those circumstances and could not invoke it later. Lapides, 535 U.S. at 623–24; Lapides, No. 01-298, slip op. at 8–10. However, in doing so the Court mistakenly asserted that the State in Ford had litigated and “lost” in the lower courts, a mistake the Court corrected by the time of publication of the preliminary print. Compare Lapides, No. 01-298, slip op. at 7 (“Georgia adds that . . . in Ford a State regained immunity . . . even after the State litigated and lost a case brought against it in federal court.”), with Lapides v. Bd. of Regents of the Univ. Sys. of Ga., 535 U.S. 613, 622 (2002) (prelim. print) (“Georgia adds that . . . in Ford a State regained immunity . . . even after the State litigated a case brought against it in federal court”). This example was brought to my attention by Karen Cordry. 161 300 U.S. 515 (1937). 162 Id. at 563 (emphasis added); see Letter from Harlan F. Stone, Assoc. Justice, Supreme Court of the United States, to Ernest Knaebel, Reporter of Decisions, Supreme Court of the United States (Oct. 14, 1937), Stone Papers B82 (requesting the addition). 163 461 U.S. 138 (1983). 164 Connick v. Myers, 51 U.S.L.W. 4436, 4438 n.7 (U.S. Apr. 20, 1983) (No. 81-1251); see Letter from Henry C. Lind, Reporter of Decisions, Supreme Court of the United States, to Byron R. White, Assoc. Justice, Supreme Court of the United States (Mar. 13, 1984), White Papers BI:622/F5; Letter from Byron R. White, Assoc. Justice, Supreme Court of the United States, to Henry C. Lind, Reporter of Decisions, Supreme Court of the United States (Mar. 13, 1984), White Papers BI:622/F5. Compare Connick, 51 U.S.L.W. at 4438 n.7, with Connick, 461 U.S. at 148 n.7. 165 486 U.S. 330 (1988). 166 Monessen Sw. Ry. Co. v. Morgan, 56 U.S.L.W. 4494, 4497 (U.S. June 6, 1988) (No. 86-1743); see Letter from Byron R. White, Assoc. Justice, Supreme Court of the United States, to Frank D. 570 HARVARD LAW REVIEW [Vol. 128:540 In Hamm v. City of Rock Hill,167 decided in 1964, the Court made and subsequently corrected a seemingly trivial mistake that in fact had substantive implications for the Court’s reasoning.168 Hamm was a major civil rights case involving the validity of the state trespass convictions of several civil rights protestors who, after being denied service at store lunch counters, remained seated and refused to leave.169 One of the legal issues raised in Hamm was whether the federal saving statute,170 originally enacted in 1871,171 which presumptively preserved prosecutions under laws amended after the commencement of prosecution,172 limited the reach of the Civil Rights Act of 1964173 to invalidate the trespass convictions.174 The Court held that the federal saving statute did not limit the application of the Civil Rights Act because the 1871 law “was meant to obviate mere technical abatement such as that illustrated by application of the rule in [United States v. Tynen175] decided in 1870.”176 The Court’s reasoning rested on its assumption that Tynen was decided a year before Congress passed the saving statute and therefore that Congress had enacted the law in response to Tynen.177 Tynen, however, was in fact decided in 1871,178 af––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Wagner, Reporter of Decisions, Supreme Court of the United States (July 1, 1988), White Papers BII:83/F10. Compare Monessen Sw. Ry., 56 U.S.L.W. at 4497 (“Of course, because there is no right to a jury trial in LHWCA actions, the ‘trier of fact’ in Pfeifer was a judge rather than a jury.”), with Monessen Sw. Ry., 486 U.S. at 341 (“Of course, because Pfeifer was tried to the bench, the ‘trier of fact’ in that case was a judge rather than a jury.”). 167 379 U.S. 306 (1964). 168 This example was brought to my attention by John MacKenzie, who served as the Supreme Court reporter for the Washington Post at the time of the Court’s ruling and who published a case comment on Hamm, including the Court’s error. See MacKenzie, supra note 19. The Court’s decision to change the opinion in the United States Reports occurred after completion of the case comment and just before publication and is accordingly noted only in an “Editor’s Note” at the close of the comment. Id. at 182. 169 See Hamm, 379 U.S. at 307–08. 170 Act of July 30, 1947, Pub. L. No. 80-278, § 109, 61 Stat. 633, 635 (codified at 1 U.S.C. § 109 (2012)). 171 Act of Feb. 25, 1871, ch. 71, § 4, 16 Stat. 431, 432. 172 See 1 U.S.C. § 109. 173 Pub. L. No. 88-351, 78 Stat. 241 (codified as amended at 42 U.S.C. §§ 2000a to 2000h-6 (2012)). 174 See Hamm, 379 U.S. at 314. 175 78 U.S. (11 Wall.) 88 (1871). 176 Hamm v. City of Rock Hill, 33 U.S.L.W. 4079, 4081 (U.S. Dec. 14, 1964) (Nos. 2, 5). 177 See id. 178 MacKenzie, supra note 19, at 174 n.15. The Court’s mistake was apparently caused by its mixing up the Term that the case was decided (December Term 1870) with the year the case was decided (1871). Although not the kind of mistake one would expect of a Supreme Court Justice, this particular mistake was a bit more understandable because the Supreme Court Reporter during the December Term 1870, John William Wallace, did not include the date of the decision in any of his volumes of the United States Reports. He provided only the year of the Term of the Court at the outset of the volume and as a running header throughout the volume. See 78 U.S. (11 Wall.) (1871) (volume for December Term 1870); MacKenzie, supra note 19, at 174 n.15. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 571 ter Congress had already passed the federal savings statute.179 The Court subsequently corrected its error, changing the date to 1871 in the United States Reports.180 But the Court did so without any acknowledgment that the sentence, as corrected, no longer supported its apparent thesis that Congress had acted in response to Tynen. Other examples include modifying opinions to clarify what the Court “did not rule,”181 to explain what legal issues were remanded to the court below,182 and to “vacate” rather than “reverse” a lower court judgment.183 Additionally, the Court has deleted assertions that “[t]here was evidence to support this finding”184 and that the facts of a prior precedent of the Court were “practically on all fours with those of the present case.”185 And the Court has changed a characterization of “gross negligence” to one of just “negligence.”186 The Court has even retreated from a clear conclusion of law that it claimed “[w]e do not doubt,” which concerned the result of a conflict between a federal statutory provision and a rule of the Federal Rules of Civil Procedure.187 Sufficient doubt plainly arose, however, and the claim was eliminated prior to publication in the United States Re––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 179 180 See MacKenzie, supra note 19, at 174–75. See Hamm v. City of Rock Hill, 379 U.S. 306, 314 (1964). The Washington Post’s Supreme Court reporter notified the Supreme Court Reporter of the error. The Reporter announced the changes would be made in the United States Reports in correspondence back to the Post reporter, but there was otherwise no public acknowledgment by the Court of the change. See MacKenzie, supra note 19, at 182 (Editor’s Note). 181 Elgin, Joliet & E. Ry. Co. v. Burley, 327 U.S. 661, 666 (1946). In Elgin, Joliet & Eastern Railway Co. v. Burley, 327 U.S. 661, the Court had granted rehearing for the sole purpose of clarifying its prior ruling to make clear what it had not ruled. See id. at 666–67 (“[W]e did not rule, and there is no basis for assuming we did, that an employee can stand by with knowledge or notice of what is going on with reference to his claim. . . . No such ruling was necessary for their preservation and none was intended.”); Elgin, Joliet & E. Ry. Co. v. Burley, 326 U.S. 801, 801–02 (1945) (order granting rehearing); Elgin, Joliet & E. Ry. Co. v. Burley, 325 U.S. 711 (1945). 182 See Seminole Nation v. United States, 316 U.S. 651, 651 (1942) (amending earlier version of Seminole Nation v. United States, 316 U.S. 286 (1942)) (making clear that certain legal issues were available for consideration on remand); see also Edited Opinion, Forest Grove Sch. Dist. v. T. A., No. 08-305, slip op. at 16 (U.S. June 22, 2009) (rev. n.d.), in Change Pages: 557 U.S. Part I, supra note 116, at 1902, 1920 (adding words “and the case is remanded for further proceedings consistent with this opinion”). 183 INS v. Lavoie, 389 U.S. 908, 908 (1967) (amending earlier version of INS v. Lavoie, 387 U.S. 572 (1967) (per curiam)) (providing that “the judgment of the United States Court of Appeals for the Ninth Circuit be vacated rather than reversed”). 184 Libby, McNeil & Libby v. United States, 340 U.S. 916, 916 (1951) (amending earlier version of Libby, McNeil & Libby v. United States, 340 U.S. 71 (1950)). 185 Clearfield Trust Co. v. United States, 318 U.S. 744, 744 (1943) (amending earlier version of Clearfield Trust Co. v. United States, 318 U.S. 363 (1943)). 186 See Edited Opinion, The Iroquois, No. 200, slip op. at 4 (U.S. May 2, 1904) (rev. n.d.) (on file with the National Archives, Record Group 267, Entry 22, Box 81, Folders 18752, 18803) (decision filed and announced on May 2, 1904, then revised and corrected on May 4, 1904). 187 Granny Goose Foods, Inc. v. Bhd. of Teamsters, 42 U.S.L.W. 4331, 4335 (U.S. Mar. 4, 1974) (No. 72-1566). 572 HARVARD LAW REVIEW [Vol. 128:540 ports,188 forcing the concurring opinion of four Justices to change as well.189 The Court has also added an entirely new paragraph of reasoning and citations weeks after the original opinion announcement.190 Concurring and dissenting opinions have backed off on characterizations of the legal import of other Justices’ opinions in some of the Court’s most significant cases,191 added new characterizations,192 and significantly amended their own substantive reasoning.193 On none of those occasions has the Court felt compelled to grant rehearing or provide public notice prior to modifying its opinion. And while the Court has sometimes made clear that a particular amendment had been made, such as through the publication of an order, the Court has not done so in many of these instances. The same is true for the separate opinions of individual Justices. In recent years and in the past, individual Justices, like the Court, have regularly made ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 188 Compare id. (“We do not doubt that were there an actual conflict between § 1450 and Rule 65(b), the statute would control. See 28 U.S.C. § 2071 (1970). But no such conflict exists in this case.”), with Granny Goose Foods, Inc. v. Bhd. of Teamsters, 415 U.S. 423, 435 (1974) (sentences deleted). 189 Compare Granny Goose Foods, 42 U.S.L.W. at 4339 & n.2 (Rehnquist, J., concurring in the judgment), with Granny Goose Foods, 415 U.S. at 447 (Rehnquist, J., concurring in the judgment) (eliminating footnote 2 and accompanying text that appeared in the slip opinion). 190 In Late Corp. of the Church of Jesus Christ of Latter-Day Saints v. United States, 136 U.S. 1 (1890), the Court announced the opinion on May 19, 1890, and Justice Bradley made significant revisions several days later, as confirmed by a copy of those changes dated May 28, 1890. Among those changes were insertions, including an entirely new paragraph. See Draft Opinion at 23, Late Corp. of the Church of Jesus Christ of Latter-Day Saints v. United States, Nos. 1031, 1054 (U.S. May 19, 1890) (drft. May 28, 1890) (on file with the National Archives, Record Group 267, Entry 22, Box 65, Folders 13400–13497) (revising decision filed and announced on May 19, 1890) (adding a paragraph explaining a detail of Spanish property law). 191 Compare Lawrence v. Texas, No. 02-102, slip op. at 3 (U.S. June 26, 2003) (O’Connor, J., concurring in the judgment) (“The dissent apparently agrees that if these cases have stare decisis effect, Texas’ sodomy law would not pass scrutiny under the Equal Protection Clause . . . .”) (citing id., slip op. at 17–18 (Scalia, J., dissenting)), with Lawrence v. Texas, 539 U.S. 558, 580 (2003) (O’Connor, J., concurring in the judgment) (statement about dissenting opinion omitted); see also infra pp. 599–600. 192 Compare Williams v. Georgia, 23 U.S.L.W. 4297, 4304 (U.S. June 6, 1955) (No. 412) (Clark, J., dissenting), with Williams v. Georgia, 349 U.S. 375, 396 (1955) (Clark, J., dissenting) (adding two sentences relating to the majority opinion: “In this case, unlike Patterson, the Court determines the state law itself” and “Furthermore, I agree with Mr. Justice Minton that the majority has misconstrued Georgia’s law”). 193 Compare District of Columbia v. Heller, No. 07-290, slip op. at 39 (U.S. June 26, 2008) (Breyer, J., dissenting) (“It is similarly ‘treacherous’ to reason from the fact that colonial legislatures did not enact certain kinds of legislation an unalterable constitutional limitation on the power of a modern legislature cannot do so.”), with District of Columbia v. Heller, 554 U.S. 570, 718 (2008) (“It is similarly ‘treacherous’ to reason from the fact that colonial legislatures did not enact certain kinds of legislation to a conclusion that an unalterable constitutional limitation on the power of a modern legislature cannot do so.”) (text with strikethrough supplied to highlight omission of language previously present in Heller, No. 07-290, slip op. at 39 (Breyer, J., dissenting)). For a further example, see infra pp. 595–96 (discussing Hodel v. Va. Surface Mining & Reclamation Ass’n, 452 U.S. 264 (1981)). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 573 nonincidental, substantive changes in their opinions under the guise of correcting “typographical” and “other formal errors” and without any notice. C. Pathways for Revising Opinions As described above, the Court revises its opinions following their initial publication in three different ways, only the first two of which are transparent: (a) a formal “erratum sheet” published in the bound volumes of the United States Reports; (b) a formal order, also published in the United States Reports, modifying its opinions; and (c) changes correcting “typographical and other formal errors” pursuant to a generic notice that sits atop all slip opinions and preliminary prints that appear in the bound volume of the United States Reports. There is no discernible pattern or justification for when the Court has used one pathway rather than another. One can find roughly similar examples in each category. Each pathway, and its historical use, is described below. 1. Errata Lists. — The use of errata lists has the most enduring tradition at the Court. Recent volumes of the United States Reports contain a list of errata,194 as do some of the earliest volumes. The first Reporter, Dallas, included a list of “Errata, et Addenda” in his third volume, published in 1799, which was the second volume to include Supreme Court opinions.195 The corrections were, according to Dallas, based on both his own review and “the liberal communications of the Bar.”196 Although not every subsequent volume has included an errata sheet, they do appear regularly. Other early Reporters, including Cranch, Wheaton, and Peters, all included them.197 In the early years, errata sheets tended to modify text in the same volume. But, presumably because the changing nature and cost of printing technology allowed for last-second revisions in printed volumes, errata sheets increasingly corrected prior volumes.198 Errata ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 194 195 196 197 See, e.g., Errata, 555 U.S. II (2014); Erratum, 554 U.S. II (2013). See Errata, et Addenda, 3 U.S. (3 Dall.) v, v (1799) (unpaginated in original). Id. Errata, 27 U.S. (2 Pet.) 735 (1829); Errata, 20 U.S. (7 Wheat.) (1822) (errata sheet appended, unpaginated, to end of volume); Errata, 17 U.S. (4 Wheat.) Index 76 (1819) (errata sheet appended, separately paginated, to end of volume); Errata, 13 U.S. (9 Cranch) iv (1817); Errata, 11 U.S. (7 Cranch) iv (1816); Errata, 10 U.S. (6 Cranch) vii (1812); Errata, 9 U.S. (5 Cranch) viii (1812). 198 The earliest errata sheet I found that corrected a prior volume was in 13 U.S. (9 Cranch) at iv (listing errata from 12 U.S. (8 Cranch) (1816), as they “escaped notice until it was too late to correct them in the table of errata of that volume”), and the latest one I found correcting the same volume that the errata sheet itself appears in was in 1886, see Errata, 118 U.S. iv (1886). 574 HARVARD LAW REVIEW [Vol. 128:540 sheets have been used exclusively to correct prior volumes for more than a century.199 The errata sheets make all kinds of corrections, including to citations, spelling, grammar, word choice, names, and facts.200 The corrections address errors from decades earlier, or more. A recent volume of the United States Reports, published in 2013, corrected an error in a volume published in 1991.201 A volume published in 2010 corrected an error in a 1933 opinion, changing “enabling” to “employing,”202 and a 1980 volume corrected an error in an 1888 opinion,203 further underscoring the Court’s commitment to error correction no matter how distant in time. The errata sheets have acknowledged the mistaken omission of the fact that Justices dissented204 and of an entire separate opinion of a Justice.205 The Court has even used the errata sheet to delete part of its rationale. In Cass County v. Leech Lake Band of Chippewa Indians,206 decided in 1998, the Court explained that it had declined to consider a particular legal issue concerning the scope of the Indian Nonintercourse Act207 because “the parcels at issue here are not alienable — and therefore not taxable — under the terms of the Indian Nonintercourse Act.”208 In a subsequent volume of the United States Reports, the Court described that reason as erratum and ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 199 See, e.g., Erratum, 551 U.S. II (2011); Errata, 528 U.S. II (2001); Errata, 500 U.S. II (1995); Errata, 440 U.S. II (1981); Corrections, 146 U.S. iv (1893). 200 See, e.g., Erratum, 551 U.S. at II; Erratum, 550 U.S. II (2010); Errata, 549 U.S. II (2010); Errata, 528 U.S. at II; Errata, 500 U.S. at II; Errata, 440 U.S. at II. In 1980, Justice White acknowledged that a correspondent “has me cold” in suggesting Justice White had erroneously used the word “union” instead of “employer” in his opinion for the Court in NLRB v. Local Union No 103 (Iron Workers), 434 U.S. 335 (1978), and agreed that an erratum notice should be published in 437 U.S., Letter from Byron R. White, Assoc. Justice, Supreme Court of the United States, to Henry C. Lind, Reporter of Decisions, Supreme Court of the United States (Feb. 28, 1980), White Papers BI:463/F9. See Letter from Henry C. Lind, Reporter of Decisions, Supreme Court of the United States, to Byron R. White, Assoc. Justice, Supreme Court of the United States (Feb. 28, 1980), White Papers BI:463/F9; Letter from Henry C. Lind, Reporter of Decisions, Supreme Court of the United States, to Byron R. White, Assoc. Justice, Supreme Court of the United States (Feb. 27, 1980), White Papers BI:463/F9 [hereinafter Feb. 27, 1980 Letter from Henry C. Lind to Byron R. White]; Letter from Robert Scolnick, Attorney, to Michael Rodak, Jr., Clerk, Supreme Court of the United States (Feb. 22, 1980), White Papers BI:463/F9; see also Iron Workers, 434 U.S. at 346. 201 See Erratum, 554 U.S. II (2013) (correcting Basic Inc. v. Levinson, 485 U.S. 224, 226 (1988)). 202 Errata, 549 U.S. at II (correcting Conrad, Rubin & Lesser v. Pender, 289 U.S. 472, 478 (1933)). 203 Errata, 436 U.S. II (1980) (correcting Freedman’s Sav. & Trust Co. v. Shepherd, 127 U.S. 494, 494 (1888)). 204 Memorandum by the Reporter, 48 U.S. (7 How.) vi (1849) (acknowledging erroneous omissions in preceding volume of dissents in four cases). 205 See Errata, 9 U.S. (5 Cranch) viii (1812). 206 524 U.S. 103 (1998). 207 25 U.S.C. § 177 (2012). 208 Cass Cnty., 524 U.S. at 115 n.5. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 575 added that “this issue is outside the question presented in the petition for certiorari.”209 2. Formal, Published Orders Amending a Prior Opinion. — The second and equally transparent way that the Court revises a previously released opinion is by issuing a formal, published order amending the opinion, which includes the precise language to be added and deleted. The Court has sometimes done so in response to formal motions by parties. For instance, the Court sometimes grants a petition for rehearing and decides to modify its opinion in light of that rehearing.210 The Court rarely grants rehearing, though it has done so because of a “single clause in the prior opinion.”211 The Court has even more frequently modified an original opinion while denying rehearing. For example, in Kennedy v. Louisiana,212 which addressed the constitutionality of the death penalty applied to crimes against individuals in which no person was killed, the Court formally denied a petition for rehearing arguing that the Justices had relied on erroneous statements in the Solicitor General’s brief about the availability of the death penalty under federal statutory law.213 While denying that petition, the Court simultaneously modified its prior opinion to make clear that the Solicitor General’s mistake had not affected its reasoning or ruling.214 Such a denial of review on rehearing, coupled with a modification of the Court’s prior opinion, is not historically unusual, although this practice has been narrowly concentrated within one discrete period of time. There are thirty-nine additional instances in which the Court simultaneously denied a petition for rehearing and published an order amending its opinion, often in significant ways. The practice dates back to 1928,215 with the vast majority (thirty instances) occurring between 1934 and 1944.216 The Court’s reports reveal only four instanc––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 209 210 533 U.S. II, II (2002) (internal quotation mark omitted). See, e.g., Elgin, Joliet & E. Ry. Co. v. Burley, 327 U.S. 661, 666–67 (1946); see also id. at 668 (Frankfurter, J., dissenting). 211 Flora v. United States, 362 U.S. 145, 167 (1960). 212 554 U.S. 407 (2008), reh’g denied, 554 U.S. 945 (2008). 213 See Kennedy, 554 U.S. at 946. 214 See id.; Kennedy, 554 U.S. at 426 n.*. 215 See New Mexico v. Texas, 276 U.S. 557, 557–58 (1928) (mem.). 216 See City of Va. Beach v. Howell, 411 U.S. 922, 922 (1973); Parks v. Simpson Timber Co., 389 U.S. 909, 909 (1967); Slochower v. Bd. of Higher Educ., 351 U.S. 944, 944 (1956) (per curiam); Union Trust Co. v. E. Air Lines, Inc., 350 U.S. 962, 962 (1956); SEC v. Drexel & Co., 349 U.S. 910, 910 (1955); Libby, McNeill & Libby v. United States, 340 U.S. 916, 916 (1951); United States v. Texas, 340 U.S. 848, 848 (1950); Standard Oil Co. of Cal. v. United States, 338 U.S. 808, 808 (1949); Flournoy v. Wiener, 321 U.S. 754, 754–55 (1944); L.T. Barringer & Co. v. United States, 319 U.S. 729, 729–30 (1943); Helvering v. Stuart, 317 U.S. 602, 602–03 (1942); Seminole Nation v. United States, 316 U.S. 651, 651 (1942); Swift & Co. v. United States, 316 U.S. 649, 649 (1942); Hysler v. Florida, 316 U.S. 642, 642 (1942); Cloverleaf Butter Co. v. Patterson, 315 U.S. 786, 786 (1942); Toucey v. N.Y. Life Ins. Co., 314 U.S. 585, 585 (1941); Detrola Radio & Television Corp. v. Hazeltine Corp., 314 U.S. 576, 576 (1941); Just v. Chambers, 312 U.S. 668, 668 (1941); Equitable 576 HARVARD LAW REVIEW [Vol. 128:540 es in the Court’s history in which the Court simply granted a petition to modify its opinion.217 And in one of those instances, it did so only by formally treating the rehearing petition as a petition to modify.218 Finally, on twenty-seven additional occasions, the Court amended its prior opinion by published order without any reference to a pending petition or motion.219 All but three of those orders were issued between 1934 and 1943.220 It is not clear whether the amendments in those instances were prompted by correspondence from outside the Court identifying a problem or resulted from the Court’s internal review process. In all of these instances, both the fact of an opinion amendment and the substance of those changes were published by the Court. Accordingly, one need not compare the final published opinion in the ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Life Ins. Co. v. Halsey, Stuart & Co., 312 U.S. 668, 668 (1941); Mo.-Kan. Pipe Line Co. v. United States, 312 U.S. 665, 665 (1941); Sibbach v. Wilson & Co., 312 U.S. 655, 655 (1941); SchriberSchroth Co. v. Cleveland Trust Co., 312 U.S. 654, 654 (1941); FCC v. Sanders Bros. Radio Station, 309 U.S. 642, 642 (1940); Morgan v. Comm’r, 309 U.S. 626, 626 (1940); Boteler v. Ingels, 308 U.S. 521, 521 (1939); Bonet v. Yabucoa Sugar Co., 307 U.S. 613, 613–14 (1939); Elec. Storage Battery Co. v. Shimadzu, 307 U.S. 613, 613 (1939); Eichholz v. Pub. Serv. Comm’n, 306 U.S. 622, 622 (1939); Mackay Radio & Tel. Co. v. Radio Corp. of Am., 306 U.S. 618, 618 (1939); SchriberSchroth Co. v. Cleveland Trust Co., 305 U.S. 573, 573–74 (1938); S.C. State Highway Dep’t v. Barnwell Bros., 303 U.S. 625, 625 (1938); United Gas Pub. Serv. Co. v. Texas, 303 U.S. 625, 625 (1938); Groman v. Comm’r, 302 U.S. 654, 654 (1937); Stone v. White, 302 U.S. 639, 639–40 (1937); Ickes v. Parks, 300 U.S. 640, 640 (1937); Schwartz v. Irving Trust Co., 300 U.S. 636, 636 (1937); Helvering v. Ill. Life Ins. Co., 299 U.S. 516, 516 (1936); Radio Corp. of Am. v. Radio Eng’g Labs., Inc., 293 U.S. 522, 522 (1934); New Mexico, 276 U.S. at 557–58. 217 See Washington v. United States, 444 U.S. 816, 816–17 (1979); Mahan v. Howell, 411 U.S. 922, 922 (1973); Swenson v. Stidham, 410 U.S. 904, 904 (1973); Elec. Storage Battery Co. v. Shimadzu, 307 U.S. 616, 616–17 (1939). 218 See Mahan, 411 U.S. at 922. 219 See Emich Motors Corp. v. Gen. Motors Corp., 340 U.S. 945, 945 (1951); Ala. Great S. R.R. Co. v. United States, 340 U.S. 926, 926 (1951); Frazier v. United States, 336 U.S. 912, 912 (1949); Clearfield Trust Co. v. United States, 318 U.S. 744, 744 (1943); Fisher v. Whiton, 317 U.S. 602, 602 (1942); Miller v. United States, 317 U.S. 601, 601–02 (1942); Seminole Nation v. United States, 316 U.S. 647, 647 (1942); United States v. New York, 316 U.S. 643, 643 (1942); Cudahy Packing Co. v. Holland, 315 U.S. 788, 788 (1942); Morton Salt Co. v. G.S. Suppiger Co., 315 U.S. 788, 788 (1942); Holiday v. Johnston, 313 U.S. 550, 550 (1941); United States v. Chi., Milwaukee, St. Paul & Pac. R.R. Co., 313 U.S. 543, 543 (1941); Fashion Originators’ Guild of Am., Inc. v. FTC, 312 U.S. 668, 668 (1941); Opp Cotton Mills, Inc. v. Adm’r of the Wage & Hour Div. of the Dep’t of Labor, 312 U.S. 657, 657 (1941); United States v. Darby, 312 U.S. 657, 657 (1941); Phila. Co. v. Dipple, 312 U.S. 656, 656 (1941); Newark Fire Ins. Co. v. State Bd. of Tax Appeals, 307 U.S. 616, 616 (1939); NLRB v. Fainblatt, 307 U.S. 609, 609 (1939); Taylor v. Standard Gas & Elec. Co., 306 U.S. 618, 618 (1939); Lone Star Gas Co. v. Texas, 304 U.S. 551, 551–52 (1938); Gen. Talking Pictures Corp. v. W. Elec. Co., 304 U.S. 546, 546–47 (1938); New Negro Alliance v. Sanitary Grocery Co., 304 U.S. 542, 542 (1938); Helvering v. Davis, 301 U.S. 672, 672–73 (1937); Legg v. St. John, 297 U.S. 695, 695 (1936); Rowley v. Chi. & Nw. Ry. Co., 293 U.S. 532, 532–33 (1934); Healy v. Ratta, 292 U.S. 607, 607 (1934); Tex. & Pac. Ry. Co. v. Pottorff, 291 U.S. 649, 649 (1934). 220 See cases cited supra note 219. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 577 United States Reports to prior slip opinions or preliminary prints to determine what changes were made.221 The types of modifications that the Court has made in amending prior opinions pursuant to published orders are varied in kind and significance. They can, however, be roughly grouped into the same categories described in section II.B, above. They run the full gamut from the most insignificant and highly technical to the exceedingly significant and substantively important. Many orders include corrections of technical citations and words,222 word choice substitutions or deletions,223 and revisions of factual statements or descriptions of background law.224 Closely related are corrections of how the Court char––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 221 On six occasions between 1945 and 1948, the Court inexplicably departed from this otherwise routine practice by publishing orders that a prior opinion was being amended without any indication of the nature of those amendments. See Woods v. Hills, 334 U.S. 856, 856 (1948) (amending earlier version of Woods v. Hills, 334 U.S. 210 (1948)); Schwabacher v. United States, 334 U.S. 814, 814 (1948) (amending earlier version of Schwabacher v. United States, 334 U.S. 182 (1948)); Aircraft & Diesel Equip. Corp. v. Hirsch, 332 U.S. 752, 752 (1947) (amending earlier version of Aircraft & Diesel Equip. Corp. v. Hirsch, 331 U.S. 752 (1947)); United States v. Nat’l Lead Co., 332 U.S. 751, 751 (1947) (amending earlier version of United States v. Nat’l Lead Co., 332 U.S. 319 (1947)); United States v. Beuttas, 325 U.S. 836, 836 (1945) (amending earlier version of United States v. Beuttas, 324 U.S. 768 (1945)); Price v. Gurney, 324 U.S. 825, 825 (1945) (amending earlier version of Price v. Gurney, 324 U.S. 100 (1945)). The reason, if any, for the Court’s distinct treatment in these few instances is hard to discern because of lack of explicit explanation from the Court. But, a rough review of one instance, Schwabacher v. United States, 334 U.S. 182, suggests that the changes were extremely minor and technical in nature, including substituting “were” for “was” and “,” for “.” and the addition of an “&” to a citation of authority. Compare Schwabacher v. United States, 16 U.S.L.W. 4378, 4380, 4381, 4382 (U.S. May 3, 1948) (No. 258), with Schwabacher, 334 U.S. at 182, 192, 194, 199. None of these appears to be the kind of change that could plausibly require formal notice. 222 See, e.g., United States v. Texas, 340 U.S. 848, 848 (1950) (amending earlier version of United States v. Texas, 339 U.S. 707 (1950)) (correcting a quotation and several citations). 223 See, e.g., id. (amending earlier version of Texas, 339 U.S. 707) (substituting “annexing” for “admitting”); Swift & Co. v. United States, 316 U.S. 649, 649 (1942) (amending earlier version of Swift & Co. v. United States, 316 U.S. 216 (1942)) (substituting “establish or become interested in” for “use”); Cloverleaf Butter Co. v. Patterson, 315 U.S. 786, 786 (1942) (amending earlier version of Cloverleaf Butter Co. v. Patterson, 315 U.S. 148 (1942)) (substituting “is subject to” for “comes under,” “has authority to watch” for “watches,” and “subject” for “subjected,” and striking the word “continuous”); Detrola Radio & Television Corp. v. Hazeltine Corp., 314 U.S. 576, 576 (1941) (amending earlier version of Detrola Radio & Television Corp. v. Hazeltine Corp., 313 U.S. 259 (1941)) (substituting the number “8” for “9”); Equitable Life Ins. Co. v. Halsey, Stuart & Co., 312 U.S. 668, 668 (1941) (amending earlier version of Equitable Life Ins. Co. v. Halsey, Stuart & Co., 312 U.S. 410 (1941)) (deleting the word “unsuccessful”); Healy, 292 U.S. at 607 (amending earlier version of Healy v. Ratta, 292 U.S. 263 (1934)) (substituting “hawkers or peddlers” for “itinerant vendors of merchandise”). 224 See, e.g., Mahan v. Howell, 411 U.S. 922, 922 (1973) (amending earlier version of Mahan v. Howell, 410 U.S. 315 (1973)) (correcting a factual error); Morton Salt, 315 U.S. at 788 (amending earlier version of Morton Salt Co. v. G.S. Suppiger Co., 314 U.S. 488 (1942)) (rewording the description of the legal effect of the Clayton Act); Cudahy Packing, 315 U.S. at 788 (amending earlier version of Cudahy Packing Co. v. Holland, 315 U.S. 357 (1942)) (eliminating footnotes and significant text to delete apparent mischaracterization of the Bituminous Coal Act of 1937); Hysler v. 578 HARVARD LAW REVIEW [Vol. 128:540 acterizes the record, including the positions of the parties.225 Substantive changes include plainly significant matters such as whether the judgment below was affirmed, reversed, or vacated.226 They also include clarification of the issues the Court had and had not ruled upon, and what legal issues therefore were still available for consideration on remand.227 The most substantive changes significantly alter the Court’s reasoning in support of its judgment. The Court has done so by deleting language228 and by adding language, including words and phrases of precedential significance.229 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Florida, 316 U.S. 642, 642 (1942) (amending earlier version of Hysler v. Florida, 315 U.S. 411 (1942)) (striking five sentences that apparently mischaracterized the Florida Supreme Court’s responsibilities under Florida state law); Lone Star Gas, 304 U.S. at 551–52 (amending earlier version of Lone Star Gas v. Texas, 304 U.S. 224 (1938)) (substituting “intrastate” for “interstate” in describing nature of property with particular claimed value); Tex. & Pac. Ry., 291 U.S. at 649 (amending earlier version of Tex. & Pac. Ry. Co. v. Pottorff, 304 U.S. 224, 245 (1934)) (deleting several sentences that had described the nature of government oversight of national banks and what activity constitutes a federal crime). 225 See, e.g., Helvering v. Stuart, 317 U.S. 602, 602–03 (1942) (amending earlier version of Helvering v. Stuart, 317 U.S. 154 (1942)) (rewording Court’s characterizations of arguments made by the Commissioner of Internal Revenue); United States v. New York, 316 U.S. 643, 643 (1942) (amending earlier version of United States v. New York, 315 U.S. 510 (1942)) (substituting “the total of such claims” for “the sum of the assets available” in describing New York’s theory). 226 See, e.g., INS v. Lavoie, 389 U.S. 908, 908 (1967) (amending earlier version of INS v. Lavoie, 387 U.S. 572 (1967) (per curiam)) (providing that “the judgment of the United States Court of Appeals for the Ninth Circuit be vacated rather than reversed”); Elec. Storage Battery Co. v. Shimadzu, 307 U.S. 616, 616–17 (1939) (amending earlier version of Elec. Storage Battery Co. v. Shimadzu, 307 U.S. 5 (1939)) (modifying opinion and judgment); Newark Fire Ins. Co. v. State Bd. of Tax Appeals, 307 U.S. 616, 616 (1939) (amending earlier version of Newark Fire Ins. Co. v. State Bd. of Tax Appeals, 307 U.S. 313 (1939)) (striking “The judgments in both cases are affirmed.”). 227 See, e.g., SEC v. Drexel & Co., 349 U.S. 910, 910 (1955) (amending earlier version of SEC v. Drexel & Co., 348 U.S. 341 (1955)) (adding a lengthy footnote to clarify that “[w]e do not prejudice” a particular argument available to a party in further proceedings); Seminole Nation v. United States, 316 U.S. 651, 651 (1942) (amending earlier version of Seminole Nation v. United States, 316 U.S. 286 (1942)) (clarifying that certain legal issues were available for consideration on remand). 228 See, e.g., Libby, McNeill & Libby v. United States, 340 U.S. 916, 916 (1951) (amending earlier version of Libby, McNeill & Libby v. United States, 340 U.S. 71 (1950)) (striking the words “There was evidence to support this finding”); Clearfield Trust Co. v. United States, 318 U.S. 744, 744 (1943) (amending earlier version of Clearfield Trust Co. v. United States, 318 U.S. 363 (1943)) (deleting assertion that another precedent’s “facts are practically on all fours with those of the present case”); Miller v. United States, 317 U.S. 601, 601–02 (1942) (amending earlier version of Miller v. United States, 317 U.S. 192 (1942)) (deleting more than nine lines of text in three places and adding further statement of the Court’s holding). 229 See, e.g., Frazier v. United States, 336 U.S. 912, 912 (1949) (amending earlier version of Frazier v. United States, 335 U.S. 497 (1948)) (adding language); Flournoy v. Wiener, 321 U.S. 754, 754–55 (1944) (amending earlier version of Flournoy v. Wiener, 321 U.S. 253 (1944)) (adding language explaining why a particular legal issue was not before the Court); L.T. Barringer & Co. v. United States, 319 U.S. 729, 729–30 (1943) (amending earlier version of L.T. Barringer & Co. v. United States, 319 U.S. 1 (1943)) (adding language to formally dispose of “appellant’s attack on the sufficiency of the evidence”); Mo.-Kan. Pipe Line Co. v. United States, 312 U.S. 665, 665 (1941) (amending earlier version of Mo.-Kan. Pipe Line Co. v. United States, 312 U.S. 502 (1941)) 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 579 Cases in which the Court subsequently modified its decision, without granting rehearing, extend to some of its most significant rulings. For instance, the Court amended its opinion in FCC v. Sanders Brothers Radio Station,230 which has been fairly characterized as “the most famous of standing cases.”231 In revising Sanders, the Court added a key word and deleted an entire sentence in the portion of its ruling that defined the power of Congress to confer standing. The excerpt below highlights the added word and the deleted sentence: Congress had some purpose in enacting § 402(b)(2). It may have been of the opinion that one likely to be financially injured by the issue of a license would be the only person having a sufficient interest to bring to the attention of the appellate court errors of law in the action of the Commission in granting the license. In this view, while the injury to such person would not be the subject of redress, that person might be the instrument, upon an appeal, of redressing an injury to the public service which would otherwise remain without remedy. It is within the power of Congress to confer such standing to prosecute an appeal.232 The substantive import of the changes is considerable. The language narrows the identified purpose of Congress to aid those “financially” injured rather than those injured in any sense at a time when a ripening issue, addressed years later, was whether noneconomic injuries could support standing.233 And by deleting that single sentence, the Court significantly undercut Congress’s potentially sweeping authority to eliminate the requirement that a plaintiff demonstrate a personal injury subject to redress to possess Article III standing. The omitted language provided that, even without such personal injury, a plaintiff could serve as “the instrument . . . of redressing an injury to ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– (adding new paragraph to respond to “[a] final contention in support of the order”); Fashion Originators’ Guild of Am., Inc. v. FTC, 312 U.S. 668, 668 (1941) (amending earlier version of Fashion Originators’ Guild of Am., Inc. v. FTC, 312 U.S. 457 (1941)) (adding two sentences of support for the Court’s ruling); Schriber-Schroth Co. v. Cleveland Trust Co., 312 U.S. 654, 654 (1941) (amending earlier version of Schriber-Schroth Co. v. Cleveland Trust Co., 311 U.S. 211 (1940)) (adding language in support of Court’s reasoning). 230 309 U.S. 470 (1940). 231 Louis L. Jaffe, Standing to Secure Judicial Review: Private Actions, 75 HARV. L. REV. 255, 257 (1961); see also Kenneth Culp Davis, Standing to Challenge Governmental Action, 39 MINN. L. REV. 353, 363 (1955) (describing Sanders as “the most prominent Supreme Court case recognizing standing in absence of violation of a ‘legal right’ of the plaintiff”). 232 Compare FCC v. Sanders Bros. Radio Station, 8 U.S.L.W. 542, 544 (U.S. Mar. 25, 1940) (No. 499), with Sanders, 309 U.S. at 477. Text with strikethrough is supplied to highlight omission of language previously present in Sanders, 8 U.S.L.W. at 544, while bolded text is supplied to highlight additions made in Sanders, 309 U.S. at 477. 233 See Ass’n of Data Processing Serv. Orgs., Inc. v. Camp, 397 U.S. 150, 154 (1970) (“That interest, at times, may reflect ‘aesthetic, conservational, and recreational’ as well as economic values. . . . We mention these noneconomic values to emphasize that standing may stem from them as well . . . .” (citations omitted) (quoting Scenic Hudson Pres. Conference v. Fed. Power Comm’n, 354 F.2d 608, 616 (2d Cir. 1965))). 580 HARVARD LAW REVIEW [Vol. 128:540 the public service which would otherwise remain without remedy.”234 And because of the immediately following sentence, especially its reference to “such standing,” the Court would have made clear that “[i]t is within the power of Congress to confer such standing to prosecute an appeal.”235 Such a broad notion of congressional authority to confer standing is far different than the Court’s current standing jurisprudence, which describes the redressability of the plaintiff’s personal injury as one of the “three elements” that constitutes the “irreducible constitutional minimum of standing” that Congress lacks authority to displace,236 although the legislature can “loosen” that “stricture[].”237 What is even more remarkable than the large number of orders modifying the Court’s opinions in both very small and very large ways during the 1930s and the 1940s is that there are hardly any such orders beforehand or afterward. The possibility that the Court was particularly sloppy in its production of initial opinions during this timeframe seems clearly wrong. The need to correct errors is a constant. The anomaly is far more likely to reflect a shift in procedure for how the Court corrected errors. In other words, during the 1930s and 1940s, the Supreme Court Reporter chose to embrace the most transparent approach. But before that period and ever since, the Court has made changes, only in a far less transparent fashion.238 3. “Printer Proofs” and “Change Pages.” — The third and final way that the Court has revised its opinions subsequent to their initial ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 234 235 Sanders, 8 U.S.L.W. at 544 (emphasis added). Id. Notwithstanding the Court’s deletion of this substantively significant language, both commentators and the Court itself have nonetheless sometimes characterized the Court’s holding in Sanders in ways that would have more force had the language not been deleted. See, e.g., United States v. Richardson, 418 U.S. 166, 193 (1974) (Powell, J., concurring) (“The Court has confirmed the power of Congress to open the federal courts to representatives of the public interest through specific statutory grants of standing.” (citing Sanders, 309 U.S. 470)); Scripps-Howard Radio, Inc. v. FCC, 316 U.S. 4, 14 (1942) (“[T]hese private litigants have standing only as representatives of the public interest.” (citing Sanders, 309 U.S. at 477)); Elizabeth Magill, Standing for the Public: A Lost History, 95 VA. L. REV. 1131, 1139 (2009) (“As these challengers [to administrative action] had no cognizable rights of their own, they had standing — and the courts were explicit about this — to raise the rights of the public. The case in which this approach was born [was] FCC v. Sanders Brothers Radio Station . . . .”); Richard B. Stewart, The Reformation of American Administrative Law, 88 HARV. L. REV. 1667, 1731 (1975) (“Sanders construes statutory review provisions as affording standing to plaintiffs who have no legally protected interest so that they may act as surrogates for those who do.” (footnote omitted)). 236 Lujan v. Defenders of Wildlife, 504 U.S. 555, 560 (1992). 237 Summers v. Earth Island Inst., 555 U.S. 488, 497 (2009). 238 In commenting on an earlier draft of this Article, Professor Bill Popkin raised the intriguing possibility that the Court’s short-lived practice might somehow relate to the Court’s condemnation during that same time period of the executive branch for failing to have a regularized system, accessible to the public, of publishing executive orders, which led to the creation of the Federal Register. See United States v. Pub. Utils. Comm’n of Cal., 345 U.S. 295, 320 (1953) (Jackson, J., concurring) (describing Court’s actions and oral argument in Panama Refining Co. v. Ryan, 293 U.S. 388 (1935)). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 581 announcement and publication is the least transparent. It is accomplished through what are sometimes referred to euphemistically as “printer proofs” or “change pages.” This process of revising the Court’s opinion prior to final publication in the United States Reports has become the most common method for making changes in recent decades and has included extensive changes of all types. (a) The Meaning of “Printer Proofs” and “Change Pages.” — The terms “printer proofs” and “change pages” are interchangeable. They both refer to the marked up pages that result from editing and review within the Court following the initial opinion announcement (bench and slip opinions) and again after the publication of the preliminary print. They are literally the pages that the Court sends to the publisher of the United States Reports to indicate what changes to make for the “final” and “official” version.239 Unlike the errata sheets or the orders amending opinions, these pages are not released to the public.240 The absence of any contemporaneous public record makes these changes hidden as a practical matter. Without the change pages, the only way to discover changes is a tedious process of comparing prior versions of the opinion to the final version. As illustrated by specific examples below, the types of changes that the Court makes in this fashion are not categorically different from changes made through other pathways. Every kind of example — large and small, technical and substantive — that is evidenced in one category can be found in the others. Corrections made through change ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 239 See Email from Frank D. Wagner, Former Reporter of Decisions, Supreme Court of the United States, to author (June 22, 2012, 8:51 AM) (on file with the Harvard Law School Library) (“The Reporter’s Office compiles all changes suggested (whether by chambers, Reporter’s Office staff, or the public) on copies of slip opinion pages for the preliminary print and copies of preliminary print pages for the bound volume and sends these change pages to the authoring Justice for approval. Chambers either accepts or exes out the suggestions and returns the change pages to the Reporter, who sends them to the Court’s Publications Unit to input the approved suggestions. Copies of the change pages are then sent to any print or online republisher of opinions (who is willing and able to pay the annual $600 copying fee) to allow the republisher to update its version of the opinions.”). 240 A former Supreme Court Reporter suggested that the change pages might be available following publication of the bound volumes but that no one had previously requested them. See id. (“The original change pages that were returned by chambers are kept in the Reporter’s Office for several years, then sent to the National Archives for permanent storage. The change pages are not separately available to the public prior to their implementation in the preliminary print or bound volume.”); Email from Frank D. Wagner, Former Reporter of Decisions, Supreme Court of the United States, to author (June 29, 2012, 11:22 AM) (on file with the Harvard Law School Library) (“No one ever asked to see them while I was Reporter. The pages remain in the Reporter’s Office for several years, then are sent to the Curator, who forwards them to the Archives on her own schedule. Neither my staff nor I ever dealt with the Archives.”). According to the Archives, the Court has not transmitted to the Archives any additional records of change pages since 1913 and according to Office of the Curator at the Court, any records they might have are commingled with records that are not publicly available. See infra note 266 and accompanying text. 582 HARVARD LAW REVIEW [Vol. 128:540 pages can also be prompted by errors discovered by the Court itself or by problems identified by outside commentators, which the Court subsequently decides warrant amendment of a previously released opinion.241 (b) The Origin of Printer Proofs and Change Pages. — As described above, although the Court has long revised opinions following their initial oral announcement, the prospect of differing written opinions in the same case did not arise until the Court began releasing earlier written versions. The historic record is not entirely clear regarding the precise timing, but by the mid-nineteenth century, the Court appears to have been releasing copies of its written opinions prior to publication of the United States Reports, at least when there was sufficient public and media demand for advance release.242 Before then, the only published version was the final version in the United States Reports.243 Differences between what was orally announced in Court and the subsequently published written versions were sometimes great but were much harder to pin down. The need for formal procedures for revising previously printed and published versions became acute in the late nineteenth century once private publishers began routinely publishing both the original opinions and the final opinions appearing in the United States Reports. The need became even greater in the 1920s when, as also described above, Congress assigned GPO the responsibility of publishing both the “advance copies” of opinions in “pamphlet installments” and the bound United States Reports.244 Indeed, in justifying the legislation, Chief Justice Taft’s testimony before Congress underscored the extent to which the Reporter’s role included making “corrections” to the original opinion. The Chief stressed that “the number of corrections he has to make, even in a careful judge’s opinion, you would hardly credit.”245 In what can be fairly assumed to be a not-so-veiled reference to some of his colleagues ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 241 For instance, according to the Reporter’s official “change pages” for volume 557 of the United States Reports, Paul Wolfson — an attorney affiliated with the law firm representing the respondent in Cuomo v. Clearing House Ass’n, 129 S. Ct. 2710 (2009), but not himself listed as counsel — contacted the Reporter to suggest that the Court’s opinion contained a grammatical error. See Edited Opinion, Cuomo v. Clearing House Ass’n, No. 08-453, slip op. at 5 (U.S. June 29, 2009) (rev. n.d.), in Change Pages: 557 U.S. Part II, supra note 122, at 2202, 2209 (referring to Paul Wolfson correspondence dated June 30, 2009). The Court subsequently corrected the error. Id. 242 See supra pp. 551–52. 243 This includes the Reports compiled and published in the late eighteenth and early nineteenth centuries by private individuals. 244 Act of July 1, 1922, Pub. L. No. 67-272, § 1, 42 Stat. 816, 816. 245 Testimony of Chief Justice Taft Before the House Judiciary Committee, supra note 75, at 411; see also 62 CONG. REC. 7756 (1922) (remarks of Rep. Walsh) (“He [the Reporter] has to verify every opinion which is written by the justices of that court.”). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 583 on the bench, the Chief further acknowledged that not all judges are “careful,” by adding: “And when you come to the opinion of a judge who is not so careful, who only tries to reach a conclusion, and is bothered with references, the number of corrections which the reporter has to make are [sic] very numerous and that adds greatly to the burden of his work.”246 The Reporter would himself need to “prepare and furnish the references.”247 As contemplated by the legislation, GPO began to publish the advance copies in paperback volumes entitled “preliminary prints,” which consisted in effect of advance sheets of what would later appear in the United States Reports.248 At least by their formal title — preliminary print — and later by express notice, the Court made clear for the first time that the initially published versions were subject to change.249 Only several decades later, in January 1970, did the Court begin its current practice of including an express notice in both its “bench opinions” (which the Court releases immediately upon the opinion’s announcement) and its “slip opinions” (which GPO prints and makes available in greater numbers a few days later)250 that the opinions are subject to formal revision.251 The change in practice was ap––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 246 Testimony of Chief Justice Taft Before the House Judiciary Committee, supra note 75, at 411. 247 Id. The Chief was referring to the need for the Reporter “to go through every opinion when written and run down every reference in the original opinions, and correct them.” Id. The term “reference” is susceptible to a narrow interpretation, nothing more than citation format, but it is also susceptible to a far broader application, consistent with the Court’s use of the term “formal error.” 248 The earliest citation I located to a “preliminary print” was in 1926, only four years after GPO took over the printing responsibilities. See Judicial Decisions Involving Questions of International Law, 20 AM. J. INT’L L. 782, 815 (1926) (reproducing L. Littlejohn & Co. v. United States, No. 94 (U.S. Mar. 1, 1926)) (citing as “Official Reports of the Supreme Court, Preliminary Print, Vol. 270 U.S., No. 2, p. 215”). 249 The original House version of the 1922 legislation used the word “preliminary” to refer to versions of Court opinions to be released prior to the bound versions, and the Senate, without explanation, amended the bill to replace “preliminary” with “advance.” See 62 CONG. REC. 8226 (1922) (Senate amendment to H.R. 11450, 67th Cong. (1922)); id. at 9531 (House agreeing to Senate amendments). 250 The traditional difference between the “bench opinion” and the “slip opinion,” which is still reflected on the Court’s website, is that the former is produced within the Court and is made available in limited numbers to the public and media immediately upon the opinion’s announcement, while the latter is printed in far greater numbers a few days later and is distributed to the public, judges, and government offices. See Information About Opinions, supra note 4. The slip opinions also include the first set of “corrections.” KENNETH JOST, THE SUPREME COURT A TO Z 344 (5th ed. 2012); 2 DAVID G. SAVAGE, GUIDE TO THE U.S. SUPREME COURT 999 (5th ed. 2010). With the advent of electronic release of the Court’s opinion within seconds of the opinion announcement, however, any difference between the bench opinion and the slip opinion seems moot. See supra note 4. 251 The first bench opinion to include the notice was released in Wade v. Wilson, 396 U.S. 282 (1970). See Wade v. Wilson, No. 55, bench op. at 1 (U.S. Jan. 13, 1970), Brennan Papers 584 HARVARD LAW REVIEW [Vol. 128:540 parently ordered by Chief Justice Burger on his own initiative.252 (c) The Internal Process for Preparing the Printer Proofs and Change Pages. — The Reporter has long been responsible for supervising the process of reviewing and modifying the initial, published versions of the Court’s opinions prior to publication of the final, official versions in the United States Reports. The Reporter, moreover, works closely with the Justice who authored the opinion of the Court (or a dissenting or concurring opinion potentially joined by others) in seeking that Justice’s approval prior to sending the final version to GPO for printing. The Reporter regularly sends the proofs to that Justice, and only to that Justice, for review and approval in formal, internal correspondence, which ultimately characterizes the opinion of the Court as “your opinion.”253 The revision process, therefore, contrasts sharply with the procedure followed prior to a slip opinion’s initial publication when all chambers are part of the formal review process and a majority vote is achieved only after the most exacting review of a draft opinion’s precise wording and detail. The Reporter follows the same procedure of forwarding only to the opinion’s author correspondence from persons outside the Court who write the Reporter to point out an error in an initial, published version of the Court opinion warranting correction in the United States Re––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– BI:208/F7. A memorandum from Justice Brennan’s clerk to the Justice describes “[t]he notice at the top of the opinion” as “new.” Memorandum from Judicial Clerk, Supreme Court of the United States, to William J. Brennan, Assoc. Justice, Supreme Court of the United States (undated), Brennan Papers BI:208/F7. The United States Law Week’s publication of the Court’s slip opinions confirms this same timing. Compare Wade v. Wilson, 38 U.S.L.W. 4071, 4071 (U.S. Jan. 13, 1970) (No. 55) (containing the notice, like all Supreme Court opinions published on this date), with Sullivan v. Little Hunting Park, Inc., 38 U.S.L.W. 4059, 4059 (U.S. Dec. 15, 1969) (No. 33) (containing no notice). 252 Memorandum from Judicial Clerk to William J. Brennan, supra note 251 (“The notice at the top of the opinion is new; apparently it was ordered by the CJ.”). According to handwritten notes on another copy of a draft opinion in Wade, Justice Brennan’s law clerk for that case was Richard Cooper. See Annotated Draft Opinion at 1, Wade v. Wilson, No. 55 (U.S. Jan. 13, 1970) (drft. Dec. 23, 1969) (Black, J., dissenting), Brennan Papers BI:208/F7 (containing a note to Justice Brennan followed by the initials “RMC”); see also Email from Richard M. Cooper to author (Dec. 2, 2013, 9:33 AM) (on file with the Harvard Law School Library); Email from author to Richard M. Cooper (Nov. 30, 2013, 4:09 PM) (on file with the Harvard Law School Library). 253 E.g., Letter from Henry Putzel, jr., Reporter of Decisions, Supreme Court of the United States, to Byron R. White, Assoc. Justice, Supreme Court of the United States (Nov. 2, 1964), White Papers BI:52 (“On the attached copies of your opinions for that volume, . . . we have suggested certain changes of a formal nature.”); Memorandum from Walter Wyatt, Reporter of Decisions, Supreme Court of the United States, to William J. Brennan, Assoc. Justice, Supreme Court of the United States (May 22, 1959), Brennan Papers BII:141 (“Please advise whether you desire us to make these changes in your opinions before publishing the bound volume of 358 U.S.”); Letter from Walter Wyatt, Reporter of Decisions, Supreme Court of the United States, to William O. Douglas, Assoc. Justice, Supreme Court of the United States (May 24, 1949), Douglas Papers B1133/F2 (“Before going to press with the bound volume of 335 U.S., this office has gone through the preliminary prints with a fine-toothed comb . . . . [W]e have noticed a few minor matters of form in some of your opinions which you may or may not want to change.”). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 585 ports (or, if too late for that, an erratum notice for a final published bound volume).254 This is true even if the Justice has since retired from the Court. It is to the retired Justice that the Reporter sends the proofs for review and approval of “your opinion,”255 and the retired Justice can unilaterally make further suggested changes.256 The “ground rules”257 for making editorial changes in initial, published opinions prior to the final United States Reports publication are set forth in a letter from the Reporter to a Justice “[i]n connection with a Justice’s first opinion for the Court.”258 According to those ground rules as described in 1971, “[n]o changes are made in an opinion except those that are specifically approved or authorized by the author or the Court,”259 with the exception of incomplete citations and “obviously misspelled words.”260 It was up to the “author of each opinion to determine whether the changes suggested require clearance with the Court and to obtain such clearance when needed.”261 That practice apparently changed sometime in 1994 when Chief Justice Rehnquist reportedly informed the Reporter and the other Justices that any substantive changes could in the future be made only after review and approval by the Court.262 It is not clear, however, how that then-new procedure has been administered, including how strictly it has been applied, if at all.263 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 254 See, e.g., Feb. 27, 1980 Letter from Henry C. Lind to Byron R. White, supra note 200 (referring to a letter describing a factual error in Justice White’s opinion for the Court in Iron Workers, 434 U.S. 335 (1978)). 255 See, e.g., Letter from Frank D. Wagner, Reporter of Decisions, Supreme Court of the United States, to Harry A. Blackmun, Assoc. Justice (Retired), Supreme Court of the United States (Nov. 14, 1997), Blackmun Papers B1426/F6 (“On the enclosed copies of your opinions for that volume, . . . we have suggested certain changes of a formal nature.”). 256 See, e.g., Letter from Harry A. Blackmun, Assoc. Justice (Retired), Supreme Court of the United States, to Frank D. Wagner, Reporter of Decisions, Supreme Court of the United States (May 6, 1996), Blackmun Papers B1426/F6 (providing changes, described as a “few generally insignificant suggestions,” to his opinions). 257 Letter from Henry Putzel, jr., Reporter of Decisions, Supreme Court of the United States, to Harry A. Blackmun, Assoc. Justice, Supreme Court of the United States 1 (Jan. 21, 1971), Blackmun Papers B1425/F11 (internal quotation marks omitted). 258 Id. 259 Id. (emphasis added). 260 Id. 261 Id. at 2. 262 Telephone Interview with Frank D. Wagner, Former Reporter of Decisions, Supreme Court of the United States (Feb. 27, 2012); Email from Frank D. Wagner, Former Reporter of Decisions, Supreme Court of the United States, to author (Nov. 30, 2013, 5:55 PM) (on file with the Harvard Law School Library); Email from author to Frank D. Wagner, Former Reporter of Decisions, Supreme Court of the United States (Nov. 30, 2013, 2:02 PM) (on file with the Harvard Law School Library). 263 In October 1979, the Reporter proposed streamlining the process to allow for the vast majority of changes, such as typographical errors, misquotes, and citation errors, to be changed by the Reporter without the need for a Justice’s specific approval and having “[o]nly [certain] mat- 586 HARVARD LAW REVIEW [Vol. 128:540 Even more fundamentally, the term “substantive” is a notoriously opaque and flexible term in discrete application. And, for that same reason, it is no more obvious that Justices would agree on the affixing of that label to specific modifications of an opinion than on other matters about which they routinely disagree. Given, moreover, the tight logic and extreme precision necessary in many Court opinions to achieve a majority, it is not at all speculative to posit that a change of a single word could prove significant. Yet, as far as can be discerned, even under the procedures as modified by Chief Justice Rehnquist, it is up to the Justice authoring the opinion to decide what is sufficiently “substantive” to warrant consultation with other chambers. No other notice is provided. (d) Source Materials for Printer Proofs and Change Pages. — There is no single repository containing all the printer proofs and change pages. Instead, the examples described below were derived from the identification and review of several different sources: (i) National Archives Records 1808–1913. — While always retaining ownership and control over its records, the Court has historically sent some of its records to the National Archives for storage.264 One set of boxes in those records includes the printer proofs that the Court used internally between 1808 and 1913 to revise opinions subsequent to their initial announcement.265 Sometime after 1913, the Court stopped sending those records to the National Archives.266 (ii) Change Pages in Possession of GPO and Private Publishers. — The Court provides GPO and private publishers, including Westlaw and LexisNexis, with change pages so they can make necessary corrections and publish final versions of the Court’s opinions. According to the Reporter’s Office, publishers pay an annual fee of $600 for this service.267 In response to a request, LexisNexis provided the change pages for one past volume but declined to send additional volumes after apparently learning that the Reporter objected to their release.268 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– ters, primarily changes of a substantive nature” require submission to a Justice for approval. Letter from Henry C. Lind to Warren E. Burger, supra note 113, at 1. 264 See Email from Robert Ellis, Archivist, Fed. Judicial Records, Nat’l Archives, to author (July 16, 2012) (on file with the Harvard Law School Library); Email from Frank D. Wagner to author, supra note 240. 265 See Email from Robert Ellis to author, supra note 264. 266 See id. The Archives records are located in “Record Group 267, Records of the United States Supreme Court, Entry 22, Appellate Opinions, 1808–1913, with gaps.” Id. I made several contacts with Court officials to try to locate the documents, including speaking to officials at the Supreme Court Library and the Curator’s Office, but was unable to discover where the documents are currently located or to gain access to them. 267 Email from Frank D. Wagner to author, supra note 239. 268 LexisNexis initially indicated that it would send more volumes, extending as far back as Volume 542. See Email from Meg Kribble, Harvard Law Sch. Library, to author (Sept. 18, 2012, 5:00 PM) (on file with the Harvard Law School Library) (“Are you looking for a specific starting 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 587 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– point? We have corrections starting with Vol 542 US and sometimes 2nd and 3rd corrections as the court sends them.”) (quoting email correspondence from LexisNexis to Meg Kribble). The publisher subsequently changed its position a few weeks later, suggesting that I would need to “contact the ROD [Reporter of Decisions] directly for additional pages,” Email from Meg Kribble, Harvard Law Sch. Library, to author (Oct. 15, 2012, 6:30 PM) (on file with the Harvard Law School Library) (appending Email from Karen Gray, LexisNexis, to Meg Kribble, Harvard Law Sch. Library (Oct. 11, 2012, 8:55 PM)), and no additional volumes were forthcoming. Several months later, LexisNexis formally wrote me to describe more fully the reasons for its change of position: “Unfortunately,” it lacked the authority to release the volumes to me; to do so required the permission of “the Reporter of Decisions for the Supreme Court” and it was “her position that [the author] would need to establish [his] own subscription to receive these.” Email from Marianne Baylor, Dir. of Content Analysis, LexisNexis, to author (June 10, 2013, 10:45 PM) (on file with the Harvard Law School Library). LexisNexis, accordingly, recommended that I “contact her [the Reporter] directly to request a reduced cost or complimentary subscription.” Id. Conversations with GPO, which also receives official change pages from the Court, similarly failed to yield any documents, presumably as a result of the Reporter’s position. Consistent with the advice received from LexisNexis, I wrote directly to the Reporter in April 2013, explained the difficulty I was having in securing the corrections from the publisher, asked if the Reporter had any objection to my receipt of the change pages directly from the publishers, and, in the alternative, expressly offered to pay the Court any necessary subscription fee to receive the change pages. See Email from author to Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States (Apr. 25, 2013, 1:19 PM) (on file with the Harvard Law School Library). The April 2013 correspondence also attached prior correspondence with the Reporter, in which I described my plans, consistent with a then-recent telephone conversation with the Reporter, and the Reporter’s agreement “to inquire whether the Court would have any formal objection to [the author’s] gaining access directly from the publishers of the Court’s opinions of the ‘change pages’ that the Court routinely sends to those publishers.” Letter from author to Christine Fallon, Reporter of Decisions, Supreme Court of the United States (July 25, 2012) (on file with the Harvard Law School Library). The Reporter did not respond to either the July 2012 or the April 2013 correspondence. More than a year later, in response to the May 2014 online posting of a draft of this Article, the Reporter wrote an email to me stating that “[o]nce the publisher takes possession of the ‘change’ pages, they belong to the publisher, which has the right to distribute the pages as it sees fit,” including providing the “pages to any interested parties, including [the author].” Email from Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States, to author (Aug. 16, 2014, 7:09 PM) (on file with the Harvard Law School Library). I cannot square the Reporter’s most recent statement of her position about the publisher’s authority to release the change pages with the contradictory, express written statements made by the publishers to me about the Reporter’s position against their release of the change pages. In May 2014, the New York Times, in an effort to confirm that the Court was denying access to the change pages, made its own formal request of the Court to receive the pages, and was denied. See Adam Liptak, Final Word on U.S. Law Isn’t: Supreme Court Keeps Editing, N.Y. TIMES, May 25, 2014, at A1. In all events, I wrote back to LexisNexis in September, appended a copy of the Reporter’s recent email statement to me, and inquired whether the publisher might reconsider its position in light of the Reporter’s stated position. See Email from author to Marianne Baylor, Dir. of Content Analysis, LexisNexis (Sept. 16, 2014, 8:20 AM) (on file with the Harvard Law School Library). The publisher responded that there had “apparently” been “a misunderstanding regarding the Supreme Court’s policy” and that, after checking directly with the Reporter, the publisher would be able to “furnish [the author] with one additional set of change pages.” Email from Marianne Baylor, Dir. of Content Analysis, LexisNexis, to author (Sept. 29, 2014, 11:58 AM) (on file with the Harvard Law School Library). Further email correspondence with the publisher provided that only one set from one volume (volume 561, part 2) was available. Email from Marianne Baylor, Dir. of Content Analysis, LexisNexis, to author (Sept. 30, 2014, 2:33 PM) (on file with the Harvard Law School Library). The new information contained in that volume is not reflected in this Arti- 588 HARVARD LAW REVIEW [Vol. 128:540 (iii) The Official Papers of Former Justices. — There are references to revisions of opinions in the official papers of former members of the Court, including Chief Justice Stone,269 and Justices Blackmun,270 Brennan,271 Douglas,272 Marshall,273 and White.274 The Library of Congress or academic institutions maintain these collections, depending on the instructions provided by the Justice. Discovering these references, however, depends largely on happenstance because they frequently appear only in files for particular cases and not in any central location. (iv) Electronic Scanning and Comparison of Digitized Versions of Supreme Court Opinions. — A final way to identify changes is to compare multiple versions of opinions electronically. The susceptibility of opinions to such electronic comparison varies: the earlier slip opinions and United States Reports were not created electronically in the first instance, so any electronic versions derived from those print versions (for example, by scanning) are necessarily of lower quality and are more difficult to compare. By contrast, the more recent versions, those produced since the 1990s, of slip opinions, preliminary prints, and final, bound reports, are initially published in a similar electronic format and are available on the Court’s website and other websites. They are accordingly far more susceptible to electronic comparison for the identification of changes. But the quality of even those more recent electronic versions has varied over time, and the comparison software proved in actual operation (in contrast to hypothetical explanation) not yet sufficiently sophisticated to identify the changes clearly and efficiently, even for recent opinions. The results were often a puzzle to decipher and time-consuming to compare.275 For this Article, comparisons were made using electronic scanning and comparison software ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– cle because it was received on October 21, 2014, which was too late for inclusion prior to this Article’s publication. 269 See supra p. 567. 270 See supra pp. 558, 585. 271 See supra notes 251–53. 272 See supra note 253. 273 See infra p. 596. 274 See supra pp. 564–65, 569. 275 Following significant media attention to a posting of a draft of this Article in May 2014, see Liptak, supra note 268, a service appeared on Twitter promising to “make[] it easy to identify when changes have been made” using computer technology no doubt far more sophisticated than within my grasp, SCOTUS Servo, GITHUB, https://github.com/vzvenyach/scotus-servo (last visited Oct. 26, 2014) [http://perma.cc/PGZ9-NTQM]. Yet even that new service acknowledged a few weeks later unanticipated difficulties: “Gosh, everyone, I’m so sorry. The Supreme Court’s website is flooding me with false alarms. None of the last tweets were valid.” SCOTUS Servo, TWITTER, (Aug. 21, 2014, 10:52 AM), https://twitter.com/SCOTUS_servo/status/489188709095337984 [http://perma.cc/ET7X -9FEE]. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 589 programs for a sampling of cases decided during fourteen Terms between 1933 and 2008.276 (e) Types of Corrections Made. — The best way to illustrate the nature and degree of revisions that the Court has made through change pages and printer proofs is through examples in individual cases. The examples described below were, like many of those discussed above, identified using each of the kinds of source materials just described.277 (i) Dred Scott v. Sandford (1857).278 — No doubt the most extreme and most notorious instance of the Court amending its opinion after its original announcement was in Dred Scott v. Sandford.279 But both because of the passage of more than 150 years and because the original reading of that opinion was not accompanied by simultaneous publication, even this example has largely faded from memory for all but a very few. When Chief Justice Taney announced the opinion of the Court on March 6, 1857, he followed the Court’s then-typical practice and orally read the opinion.280 Justices Wayne, Nelson, Grier, Daniel, Campbell, and Catron each filed his own concurring opinion.281 And the next day, the last day of the Term, Justice McLean filed a dissenting opinion,282 and Justice Curtis filed a separate dissenting opinion.283 That same day, Justice Curtis unilaterally provided a copy of his dissent to the Clerk of the Court and to a Boston editor for publication, even before the Chief had released the majority opinion to the public.284 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 276 I am most grateful to Professor Peter Martin, who co-founded Cornell’s Legal Information Institute and is an expert on uses of digital technology in law, for his comments on an initial version of this Article posted in May 2014. Martin described the history of the Court’s use of digital technology in the production of its opinions in ways that should make it easier for future researchers to undertake the kind of research I performed for this Article. Martin also published a thoughtful blog essay on the topic of revising court opinions roughly contemporaneous to my posting of an initial draft of this Article and, by chance, on the very day of Justice Scalia’s error in EME Homer. See Peter Martin, Judges Revising Opinions After Their Release, CITING LEGALLY (Apr. 29, 2014), http://citeblog.access-to-law.com/?p=157 [http://perma.cc/7TJV-P27G]. 277 These examples are not exhaustive of all changes made during those Terms or even in the individual cases. Some other changes have been briefly cited above in the general description of types of revisions. See supra section II.B, pp. 562–73. 278 Ironically, the name “Sandford” is itself an error. The party’s actual name was “Sanford.” See infra note 413. 279 60 U.S. (19 How.) 393 (1857). 280 GEORGE TICKNOR CURTIS, 1 A MEMOIR OF BENJAMIN ROBBINS CURTIS, LL. D. 211 (Benjamin R. Curtis ed., 1879). 281 Dred Scott, 60 U.S. (19 How.) at 454–56 (Wayne, J., concurring); id. at 457–69 (Nelson, J., concurring); id. at 469 (Grier, J., concurring); id. at 469–93 (Daniel, J., concurring); id. at 493–518 (Campbell, J., concurring); id. at 518–29 (Catron, J., concurring). 282 Id. at 529–64 (McLean, J., dissenting). 283 Id. at 564–633 (Curtis, J., dissenting). 284 VINCENT C. HOPKINS, DRED SCOTT’S CASE 156 (1967). 590 HARVARD LAW REVIEW [Vol. 128:540 When Justice Curtis subsequently learned that the Chief was making extensive and material revisions, he sought a copy from the Clerk.285 The Clerk referred the request to Chief Justice Taney, who responded by issuing an order imposing an embargo on the release of any more written opinions in the case.286 He also declined Justice Curtis’s request for a copy of his opinion, which he explained was undergoing “accustomed”287 and “usual”288 revisions before publication. Justice Curtis complained that he needed to learn what revisions Chief Justice Taney was making in order to respond with revisions in his dissent “before my own opinion should be published by the reporter in a permanent form.”289 Chief Justice Taney asserted that the problem was Justice Curtis’s effort to disparage the Court by prematurely releasing his written dissent without first allowing the Chief’s opinion to undergo the normal process of revision. According to the Chief, Justice Curtis was the one who had “rendered it impossible that the opinions could come out together” and had done so to “encourage attacks upon the court and upon the judges who gave the opinion, by political partisans.”290 The Chief further denied that he was materially altering his opinion from what he had read aloud in open court. Chief Justice Taney’s denial was categorical and sweeping. He maintained that Justice Curtis’s accusation of material alterations in the opinion “had no foundation in truth”291 and that Justice Curtis would find “nothing altered, nothing in addition but proofs to maintain the truth of what was announced and affirmed in the opinion delivered”292: [N]ot one historical fact, nor one principle of constitutional law, or common law, or chancery law, or statute law, in the printed opinion, which was not distinctly announced and maintained from the bench; nor is there any one historical fact, or principle, or point of law, which was affirmed in the opinion from the bench, omitted or modified, or in any degree altered, in the printed opinion.293 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 285 286 287 Id. at 157; CURTIS, supra note 280, at 212. HOPKINS, supra note 284, at 157. Letter from Roger B. Taney, Chief Justice, Supreme Court of the United States, to Benjamin R. Curtis, Assoc. Justice, Supreme Court of the United States (Apr. 28, 1857), in CURTIS, supra note 280, at 213, 214. 288 Letter from Roger B. Taney, Chief Justice, Supreme Court of the United States, to Benjamin R. Curtis, Assoc. Justice, Supreme Court of the United States (June 11, 1857), in CURTIS, supra note 280, at 221, 222. 289 Letter from Benjamin R. Curtis, Assoc. Justice, Supreme Court of the United States, to Roger B. Taney, Chief Justice, Supreme Court of the United States (May 13, 1857), in CURTIS, supra note 280, at 217, 219. 290 Letter from Roger B. Taney to Benjamin R. Curtis, supra note 288, at 224. 291 Id. at 221. 292 Id. at 222. 293 Id. at 221. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 591 The Chief characterized his changes as no more than the addition of “proofs and authorities to maintain the truth of the historical facts and principles of law” that, until they were “denied in the dissenting opinions,” he would not have “thought it necessary” to include.294 In his memoirs, by contrast, Justice Curtis claimed that the Chief had added “upwards of eighteen pages,” and that “they are in reply to my opinion.”295 A close review of the “printer proofs” for Dred Scott in the National Archives, which include Chief Justice Taney’s handwritten changes to two versions of a printed opinion,296 generally supports Justice Curtis’s characterization of the extent and nature of the Chief’s changes. Although it is uncertain what kinds of pages — typescript or printed — Justice Curtis was referring to in his claim of “eighteen pages,” the archival records certainly reveal pages and pages of substantive changes made in the Chief Justice’s own handwriting or by his inser––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 294 295 Id. at 222. Benjamin Robbins Curtis, Some Observations on the Above Correspondence (undated), in CURTIS, supra note 280, at 229, 229. 296 The National Archives Records for Dred Scott, available at Record Group 267, Entry 22, Box 28, Case 3230, consist of three different versions of the Chief Justice’s opinion in the case, which show the changes from the original opinion read on March 6, to the final version published in the United States Reports. The records include two sequential typed drafts that each have Chief Justice Taney’s handwritten changes in the margins and on attachments. The identity of Chief Justice Taney’s handwriting was confirmed by an employee in the Office of the Curator of the U.S. Supreme Court. See Email from Matthew Hofstedt, Assoc. Curator, Supreme Court of the United States, to author (May 7, 2013, 10:50 AM) (on file with the Harvard Law School Library) (“I don’t think there is any doubt that the major edits and insertions are in his hand.”). The first of these two drafts includes far more changes than the second. The third version is a handwritten version of the entire opinion, which reflects changes made to the first two typed drafts. The Court did have a practice of having a “scrivener” write up a draft provided by the Justice in order to promote legibility for the printer and reduce errors. This version is somewhat paradoxical, however, given that it clearly was produced after the two typed drafts. The Court records are incomplete in another significant respect. Missing from the National Archives is any clear record establishing that the first typed draft upon which Chief Justice Taney made his extensive handwritten changes was the original opinion he read in Court. It is possible that Chief Justice Taney read an earlier version. I have not discovered a different original opinion in any records of the Court, the Archives, or any of the Justices, but the records may have been destroyed in a fire in 1898. See WALTER EHRLICH, THEY HAVE NO RIGHTS 229 n.8 (1979) (“The original opinions were all filed with the clerk but were destroyed in a file room fire in November 1898.”). Indeed, there was a fire in November 1898 in the part of the Capitol then used by the Court. See GLENN BROWN, GLENN BROWN’S HISTORY OF THE UNITED STATES CAPITOL 494–95 (William B. Bushong ed., 2d ed. 2008). Further corroboration is provided by the National Archives’s copy of the Dred Scott judgment, which evidences fire damage. See Dred Scott v. Sandford, No. 7 (U.S. Mar. 6, 1857), available at Dred Scott v. Sanford (1857), OUR DOCUMENTS, http://www.ourdocuments.gov/doc.php?flash=true&doc=29 (last visited Oct. 26, 2014); see also Judgment in the U.S. Supreme Court Case Dred Scott v. John F.A. Sanford, March 6, 1857, PRESERVATION AT THE NAT’L ARCHIVES (Mar. 6 2012, 1:44 PM), http:// preservearchives.tumblr.com/post/18854832815/the-dred-scott-document-has-water-and-soot-damage [http://perma.cc/JSP4-4G2Z]. 592 HARVARD LAW REVIEW [Vol. 128:540 tion of additional typewritten text.297 Justice Curtis also seems correct that some significant substantive changes discuss issues raised in his dissent, although Chief Justice Taney’s changes also go far beyond that dissent. Chief Justice Taney’s changes begin with ordinary spelling corrections (for example, correcting “arrisen”)298 and word modifications (for example, changing “prove” to “disprove”)299 routinely seen in printer proofs. The Chief Justice’s amendments also include numerous additions and deletions of single words, phrases, sentences, and paragraphs that shifted the opinion’s substantive emphasis. For example, in his first set of recorded revisions, Chief Justice Taney extended his original contention that African Americans were not intended to be part of the political community created by the Constitution to assert the further claim that they were “not intended to be embraced in [it].”300 And the Chief deleted an entire paragraph that had embellished his argument that African Americans should not receive the protection of the Privileges and Immunities Clause on the ground that the Framers could not have possibly intended such a result.301 But Chief Justice Taney’s changes were of a different order of magnitude still, not remotely as narrow and modest as he suggested in his contemporaneous correspondence to Justice Curtis. In the final reported version, Chief Justice Taney added two pages to the opinion’s justification for reaching the question of whether Scott was still a slave even after concluding that the Court lacked jurisdiction regardless of whether he was still a slave.302 Chief Justice Taney added more than four pages of discussion to contest dissenting Justice McLean’s reliance on a prior Court ruling regarding the constitutionality of the Missouri Compromise.303 And he added more than a page to contend that Scott had acted improperly, as a procedural matter, by not appealing his ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 297 The first typed version in the Archives records consists of thirty-eight pages, and the final version in the United States Reports is approximately fifty-five pages long. Compare Edited Draft Opinion at 1–38, Dred Scott v. Sandford, No. 7 (U.S. Mar. 6, 1857) (1st rev. n.d.) [hereinafter First Printer Proofs, Dred Scott v. Sandford] (on file with the National Archives, Record Group 267, Entry 22, Box 28, Case 3230), with Dred Scott v. Sandford, 60 U.S. (19 How.) 393, 399–454 (1857). Two other commentators have analyzed the changes: Austin Allen more closely and Don Fehrenbacher more briefly. See AUSTIN ALLEN, ORIGINS OF THE DRED SCOTT CASE (2006); DON E. FEHRENBACHER, THE DRED SCOTT CASE 320–21 (1978). 298 Edited Draft Opinion at 1, Dred Scott v. Sandford, No. 7 (U.S. Mar. 6, 1857) (2d rev. n.d.) [hereinafter Second Printer Proofs, Dred Scott v. Sandford] (on file with the National Archives, Record Group 267, Entry 22, Box 28, Case 3230). 299 Id. at 3. 300 First Printer Proofs, Dred Scott v. Sandford, supra note 297, at 6. 301 See id. at 21. 302 Compare First Printer Proofs, Dred Scott v. Sandford, supra note 297, at A1–A4, with Dred Scott, 60 U.S. (19 How.) at 428–30. 303 Compare First Printer Proofs, Dred Scott v. Sandford, supra note 297, at B1–B12, with Dred Scott, 60 U.S. (19 How.) at 442–46. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 593 original loss in the Missouri Supreme Court to the U.S. Supreme Court. According to Chief Justice Taney’s opinion, as revised, Scott knew that following the correct procedures would have made plain the Court’s lack of jurisdiction to second-guess the Missouri court’s ruling on whether Scott was a slave under Missouri law.304 In light of the strikingly hostile tones evident in the correspondence between Chief Justice Taney and Justice Curtis, the exchange’s most important legacy may be the institutional necessity of settled understandings regarding the timing, revising, and final publication of opinions of individual Justices and of the Court. The contrast between then and now certainly gives credit to the Court’s current procedures and reportedly amiable relations between the Justices, notwithstanding the inevitable friction created in producing the final opinions each year. Despite Justice Curtis’s denial,305 it would be easy to speculate that the tense relations that developed between Chief Justice Taney and Justice Curtis over the timing of opinions in Dred Scott played some role in Justice Curtis’s decision to resign from the Court soon thereafter. (ii) Granny Goose Foods, Inc. v. Brotherhood of Teamsters (1974). — At issue in Granny Goose Foods, Inc. v. Brotherhood of Teamsters306 was the meaning of 28 U.S.C. § 1450, a federal statute providing that when a lawsuit is removed from state court to federal court, any injunction or order issued by the state court prior to removal “shall remain in full force and effect until dissolved or modified by the district court.”307 The district court had held the respondent labor union in contempt for violating the state court’s temporary restraining order.308 That order would have lapsed as a matter of state law, but the district court argued that § 1450 provided that only the federal court could lift it.309 The court of appeals disagreed,310 and the Supreme Court affirmed in an opinion by Justice Marshall.311 Justice Rehnquist filed a separate opinion concurring in the judgment, which three other Justices joined.312 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 304 Compare First Printer Proofs, Dred Scott v. Sandford, supra note 297, at C1–C4, with Dred Scott, 60 U.S. (19 How.) at 453–54. 305 See CURTIS, supra note 280, at 243 (“The correspondence with the Chief Justice . . . had no influence upon the determination to which he finally came.”). 306 415 U.S. 423 (1974). 307 28 U.S.C. § 1450 (2012); see Granny Goose Foods, 415 U.S. at 425. 308 See Granny Goose Foods, Inc. v. Bhd. of Teamsters, No. C-70-1057, 1970 WL 720, at *1–3 (N.D. Cal. Dec. 2, 1970). 309 Id. at *1–2. 310 Granny Goose Foods, Inc. v. Bhd. of Teamsters, 472 F.2d 764, 765 (9th Cir. 1973). 311 See Granny Goose Foods, 415 U.S. at 425. 312 See id. at 445 (Rehnquist, J., concurring in the judgment). 594 HARVARD LAW REVIEW [Vol. 128:540 The majority held that Congress did not intend that § 1450 give state court injunctions greater duration than they otherwise would have had: “The ‘full force and effect’ provided state court orders after removal of the case to federal court was not intended to be more than the force and effect the orders would have had in state court.”313 The Court further held that once the case was removed, “federal rather than state law governs,”314 and the state’s order had lapsed under the applicable time limitation for temporary restraining orders established by Rule 65(b) of the Federal Rules of Civil Procedure.315 The Court’s original slip opinion, published in the United States Law Week, included substantive language that was omitted in the United States Reports. The slip opinion purported to reach the issue concerning what the result would be if § 1450 and Rule 65(b) were in conflict. The Court was unequivocal: “We do not doubt that were there an actual conflict between § 1450 and Rule 65(b), the statute would control.”316 Justice Rehnquist’s concurring opinion sought to exploit the majority’s statement. Justice Rehnquist stressed in the text of his opinion the majority’s concession that the statute would be controlling.317 And the concurring opinion went on to include a lengthy footnote that used the majority’s concession as the primary basis for distinguishing a Court precedent upon which the majority had heavily relied.318 In the final version published in the United States Reports, however, the majority’s strong statement about what it “[did] not doubt,” is gone.319 Some “doubt,” apparently, subsequently arose. The majority’s deletion of a significant assertion, moreover, necessarily undermined Justice Rehnquist’s concurrence. The final version of his concurrence correspondingly eliminated both the textual discussion and the accompanying footnote.320 Nor was this deletion substantively inconsequential. Legal scholars have long debated the validity of the so-called “suppression clause” of the Federal Rules Enabling Act,321 which purports to provide the Su––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 313 314 315 316 Id. at 436 (majority opinion) (quoting 28 U.S.C. § 1450 (1970)). Id. at 437. Id. at 437–40. Granny Goose Foods, Inc. v. Bhd. of Teamsters, 42 U.S.L.W. 4331, 4335 (U.S. Mar. 4, 1974) (No. 72-1566). 317 Id. at 4339 (Rehnquist, J., concurring in the judgment). 318 See id. at 4339 n.2 (citing Ex Parte Fisk, 113 U.S. 713 (1885)). 319 Compare id. at 4335 (majority opinion), with Granny Goose Foods, 415 U.S. at 435. Also missing, more incidentally, was the majority’s claim that the temporary restraining order expired “by its [own] terms.” Compare Granny Goose Foods, 42 U.S.L.W. at 4337, with Granny Goose Foods, 415 U.S. at 440. 320 Compare Granny Goose Foods, 42 U.S.L.W. at 4339 & n.2 (Rehnquist, J., concurring in the judgment), with Granny Goose Foods, 415 U.S. at 447 (Rehnquist, J., concurring in the judgment). 321 28 U.S.C. § 2072 (2012). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 595 preme Court with the authority to “prescribe general rules of practice and procedure”322 for federal courts that deprive inconsistent statutory provisions of “force or effect.”323 The Court’s own opinions oscillate between showing support for that provision and indicating that the Court could not promulgate rules in conflict with a statute.324 Had the Court retained the original language, it might have (even if unintentionally) resolved a conflict in its own precedent. (iii) Hodel v. Virginia Surface Mining & Reclamation Ass’n (1981). — In Hodel v. Virginia Surface Mining & Reclamation Ass’n,325 the Court rejected a series of challenges to the constitutionality of the federal Surface Mining Control and Reclamation Act of 1977.326 The challenges included allegations of violations of the Fifth Amendment Due Process and Just Compensation Clauses, the Tenth Amendment, and the Commerce Clause.327 Several Justices wrote separately,328 but none dissented. Justice Rehnquist concurred in the judgment only.329 The thrust of Justice Rehnquist’s opinion was that “there are constitutional limits on the power of Congress to regulate pursuant to the Commerce Clause.”330 Justice Rehnquist stressed that the regulated activity must have a “substantial effect”331 on interstate commerce and that congressional findings of such an effect are subject to judicial review and must be supported by a “rational basis.”332 Of course, the full import of then-Justice Rehnquist’s concurrence in Hodel was not realized until 1995, when Chief Justice Rehnquist authored the majority opinion in United States v. Lopez.333 In Lopez, his earlier concurring opinion effectively became the constitutional rule. The Court applied the “substantial effects” and “rational basis” tests to strike down, for the first time in decades, a federal statute for exceeding Congress’s Commerce Clause authority.334 Lopez remains one of the Rehnquist Court’s most significant and well-known rulings. What is not widely known is that Justice Rehnquist changed the wording of his opinion in Hodel after releasing the slip opinion, and ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 322 323 Id. § 2072(a). Id. § 2072(b); see Leslie M. Kelleher, Separation of Powers and Delegations of Authority to Cancel Statutes in the Line Item Veto Act and the Rules Enabling Act, 68 GEO. WASH. L. REV. 395, 398 (2000). 324 Kelleher, supra note 323, at 439 & nn.307–08. 325 452 U.S. 264 (1981). 326 Pub. L. No. 95-87, 91 Stat. 445 (codified as amended in scattered sections of 30 U.S.C.). 327 Hodel, 452 U.S. at 273. 328 See id. at 305 (Burger, C.J., concurring); id. (Powell, J., concurring). 329 See id. at 307 (Rehnquist, J., concurring in the judgment). 330 Id. at 309. 331 Id. at 310. 332 Id. at 311 (internal quotation marks omitted). 333 514 U.S. 549 (1995). 334 See id. at 567–68. 596 HARVARD LAW REVIEW [Vol. 128:540 the change became relevant to the Court’s ruling in Lopez. In his original slip opinion in Hodel, the future Chief Justice provided that “Congress must show that its regulatory activity has a substantial effect on interstate commerce.”335 In the final United States Reports version, however, Justice Rehnquist’s proposed test became “Congress must show that the activity it seeks to regulate has a substantial effect on interstate commerce,”336 language consistent with that used in his later opinion for the Court in Lopez. The modification makes substantive sense. As first drafted, the inartful use of “its” suggests that the activity that must have the substantial effect on commerce is Congress’s own. As reworded, it is clearly the activity that Congress regulates that must have that substantial effect. The official papers of Justice Marshall, moreover, include a document in which Justice Rehnquist formally seeks permission to make the change to his original slip opinion. Justice Rehnquist was clearly amused that he was seeking to make a change to his concurring opinion in a case in which Justice Marshall wrote the majority opinion when, just a few years earlier in Granny Goose Foods, Justice Marshall had amended his slip opinion for the Court in a way that had compelled Justice Rehnquist to change his concurring opinion. In a “Memorandum to the Conference,” Justice Rehnquist wrote to his colleagues: At the suggestion of a law professor who shall remain unnamed, I would like to change the penultimate sentence in my opinion concurring in the Court’s judgment last Term to read: “Congress must show that the activity it seeks to regulate has a substantial effect on interstate commerce.”337 In a special plea directed to Justice Marshall, Justice Rehnquist wrote: Thurgood, I hope you will indulge me in allowing me to make this change; if you don’t, I will write for my memoirs the infamous account of your deletion of an entire paragraph in the “Granny Goose” opinion in one of the earlier Terms of the 70’s after the case had come down, after I had written a dissent directed largely at that paragraph. Seriously, please let me know if you have any objection to my advising the Reporter of Decisions to go ahead with this change.338 Justice Rehnquist’s memorandum may relate to a then-existing procedure, which assigned the author of the majority opinion some responsibility for overseeing proposed revisions in any of the opinions ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 335 Hodel v. Va. Surface Mining & Reclamation Ass’n, 49 U.S.L.W. 4654, 4666 (U.S. June 15, 1981) (Nos. 79-1538, 79-1596) (Rehnquist, J., concurring in the judgment). 336 Hodel, 452 U.S. at 313 (Rehnquist, J., concurring in the judgment). 337 Memorandum from William H. Rehnquist, Assoc. Justice, Supreme Court of the United States, to the Conference (Sept. 3, 1981), Marshall Papers B273 (concerning Hodel v. Va. Surface Mining & Reclamation Ass’n, No. 79-1538, slip op. (U.S. June 15, 1981)). 338 Id. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 597 filed in the case. If so, Justice Marshall apparently had no objection because Justice Rehnquist’s proposed change was made. (iv) International Paper Co. v. Ouellette (1987). — At issue in International Paper Co. v. Ouellette339 was the extent to which, if any, the federal Clean Water Act340 preempted the application of a state’s common law of nuisance.341 The lawsuit had been brought by parties injured in a downstream state (affected state) by discharges of pollutants originating in an upstream state (source state).342 The Court held that the affected state’s common law of nuisance was entirely preempted, that the source state’s nuisance law was not preempted, and that the Clean Water Act superseded otherwise applicable conflict of law rules, thus requiring that courts apply the source state’s nuisance doctrines in all lawsuits.343 Justice Powell wrote the majority opinion.344 In describing the operation of the Clean Water Act, his original slip opinion supported his view that the Act preempted application of the affected state’s nuisance law to interstate water pollution by stressing the extensive remedies provided under the federal statute.345 That description included the statement that “[t]he CWA also provides its own remedies, including civil and criminal fines for permit violations, and ‘citizen suits’ that allow individuals (including those from affected States) to compel the EPA to enforce a permit.”346 The majority opinion published in the United States Reports amends that sentence in significant respects. As modified, the final version reads: “The CWA also provides its own remedies, including civil and criminal fines for permit violations, and ‘citizen suits’ that allow individuals (including those from affected States) to sue for injunctions to enforce the statute.”347 The Court’s opinion no longer provides that citizen plaintiffs can sue to “compel the EPA” to take an action, and it no longer provides that citizen plaintiffs can enforce the “permit” rather than just the “statute.” Here again, the official papers of one of the former Justices reveal the reasons for the substantive changes. According to a memorandum ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 339 340 341 342 343 344 345 479 U.S. 481 (1987). 33 U.S.C. §§ 1251–1387 (2012); see also 33 U.S.C. §§ 1251–1376 (1982). Ouellette, 479 U.S. at 483. Id. at 483–84. Id. at 497–99, 499 n.20. See id. at 483. See Int’l Paper Co. v. Ouellette, No. 85-1233, slip op. at 10 (U.S. Jan. 21, 1987) (attached with proposed edits to Memorandum from Lewis F. Powell, Assoc. Justice, Supreme Court of the United States, to the Conference (Mar. 5, 1987), Blackmun Papers B470). 346 Id. 347 Ouellette, 479 U.S. at 492 (emphasis added); see also Memorandum from Lewis F. Powell to the Conference, supra note 345. 598 HARVARD LAW REVIEW [Vol. 128:540 to the conference prepared by Justice Powell, the Solicitor General wrote to the Reporter to suggest the change in the opinion.348 The gist of the correction was that the Court had addressed an issue that was not yet settled in the lower courts: “The Solicitor General is correct that it has not yet been decided whether the Clean Water Act authorizes a ‘citizens’ suit’ to compel the EPA to enforce a permit.”349 The “formal error” that justified the revision of the opinion, therefore, was the presence of dictum in the Court’s opinion that unwittingly resolved a circuit split. Justice Powell recommended that, “[a]bsent objection,”350 the changes be made, and they were made.351 This example, known only because of historical happenstance,352 may be particularly telling. It cannot be gainsaid that the changes were intended to be substantively significant. If not for their potential significance, the Solicitor General would not have asked the Court to make the revisions, and the Court would not have agreed to the request. If the Court widely applied the concept that uninformed dictum constitutes a “formal error” within the meaning of the notice that an opinion is subject to revision, then opinions could undergo substantial substantive revision without notice following the publication of slip opinions. The Ouellette example is also procedurally important because it highlights a problem, at least in appearance, with the potentially ex parte character of the procedures by which “readers” contact the Reporter to notify the Reporter of an error in a Court slip or preliminary opinion. Under that established procedure, only the Reporter receives notice that the Court may be considering a possible “correction” to the opinion. None of the parties receive such notice, even when, as in Ouellette, the reader notifying the Reporter is himself counsel for one of the parties, including amicus, appearing before the Court in the case. Wholly apart from any possible question of whether parties are always entitled to such notice under broad notions of due process — which seems highly remote — the absence of any notice to other par––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 348 349 350 351 352 Memorandum from Lewis F. Powell to the Conference, supra note 345. Id. Id. See Ouellette, 479 U.S. at 492. While serving as Assistant to the Solicitor General, I represented the United States in this litigation. I authored the amicus brief on the merits filed, at the Court’s invitation, on behalf of the United States and EPA. See Int’l Paper Co. v. Ouellette, 475 U.S. 1081, 1081 (1986); Brief for the United States as Amicus Curiae Supporting Affirmance, Ouellette, 479 U.S. 481 (No. 85-1233) (on file with the Harvard Law School Library). After EPA expressed concern about the substantive import of certain language in the Court’s slip opinion, I drafted a letter on behalf of the Solicitor General to the Court identifying the language of concern and recommending a modification that the Court ultimately adopted. Because the records were, as required, left at the Department of Justice, I have no record of the letter drafted besides my own recollection. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 599 ties from the Court when the “reader” is itself a party is at least troubling enough to suggest the need for rethinking.353 (v) Lawrence v. Texas (2003). — The Supreme Court’s decision in Lawrence v. Texas354 is one of the most significant rulings in recent decades. The Court dramatically overturned its prior decision in Bowers v. Hardwick355 and struck down on federal constitutional grounds a state statute that made it a crime for two persons of the same sex to engage in certain intimate sexual conduct.356 According to the Court, the state law violated the Due Process Clause of the Fourteenth Amendment by interfering with the liberty interests of homosexual persons to engage in “the most private human conduct . . . and in the most private of places, the home.”357 Justice Kennedy wrote the opinion of the Court for a five-Justice majority.358 Justice O’Connor authored a separate opinion, concurring in the judgment.359 Justice Scalia filed a dissenting opinion that Chief Justice Rehnquist and Justice Thomas joined,360 and Justice Thomas filed his own dissenting opinion.361 Justice O’Connor’s concurrence was significant because she declined to agree that Bowers should be overturned and instead argued that the state sodomy law was unconstitutional on equal protection grounds.362 Justice O’Connor’s concurrence reviewed the Court’s equal protection precedent and described how the Court had “appl[ied] rational basis review . . . where, as here, the challenged legislation inhibits personal relationships.”363 Justice O’Connor then included the following claim about Justice Scalia’s dissent: “The dissent apparently agrees that if these cases have stare decisis effect, Texas’ sodomy law would not pass scrutiny under the Equal Protection Clause, regardless of the type of rational basis review that we apply.”364 Justice O’Connor’s point about the dissent was sufficiently telling that judges and commentators singled it out. In dissenting from a denial of rehearing en banc a year later, a federal appellate judge agreed with Justice O’Connor’s characterization of Justice Scalia’s dissent, noting that “[a]n examination of these cases bears out Justice ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 353 354 355 356 357 358 359 360 361 362 363 364 See infra p. 611. 539 U.S. 558 (2003). 478 U.S. 186 (1986). See Lawrence, 539 U.S. at 578. Id. at 567. Id. at 561. Id. at 579 (O’Connor, J., concurring in the judgment). Id. at 586 (Scalia, J., dissenting). Id. at 605 (Thomas, J., dissenting). Id. at 579 (O’Connor, J., concurring in the judgment). Id. at 580. Lawrence v. Texas, No. 02-102, slip op. at 3 (U.S. June 26, 2004) (O’Connor, J., concurring in the judgment) (citing id., slip op. at 17–18 (Scalia, J., dissenting)). 600 HARVARD LAW REVIEW [Vol. 128:540 O’Connor’s point.”365 Recent editions of constitutional law casebooks excerpt this language from Justice O’Connor’s concurrence.366 And anyone who “Googles” the case will quickly find the same language, including in a highly regarded website administered by a leading law school.367 When, moreover, websites list the various opinions in a single case separately, the Justice O’Connor concurrence does not include the notice from the first page of the majority opinion that slip opinions are preliminary and subject to revision.368 The problem is that the “final,” “official” version of Justice O’Connor’s opinion no longer includes this sentence. It has been deleted in its entirety.369 Without “Papers of the Justices” like those that exist for substantially older cases decided by Justices no longer on the Court, one can do little more than speculate about the reasons for the subsequent deletion. Perhaps Justice Scalia complained, and Justice O’Connor agreed to make the change as a matter of accommodation. Or perhaps Justice O’Connor’s characterization of the import of the dissent was based on language contained in an earlier version of that dissent that was changed prior to the opinion announcement, but escaped the attention of Justice O’Connor’s chambers when she first published her concurrence. Lawrence was decided on the final day of the October 2002 Term before the summer recess.370 The very end of the Term is when multiple majority, concurring, and dissenting opinions are most likely to be crisscrossing, and therefore the possibility of such missteps is greatest. Whatever the precise nature of the behind-the-scenes cause for Justice O’Connor’s decision to change her Lawrence concurrence, this example highlights one of the more significant consequences of the Court’s current practice of making such changes without meaningful notice. Supreme Court opinions are cited and quoted frequently, and immediately: by lower courts, by other lawmaking branches, and by legal scholars and teachers. And they are cited in this way and to such an extent for understandable institutional reasons, given the Court’s prestige and the tremendous weight of its precedential authority. ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 365 Lofton v. Sec’y for the Dep’t of Children & Family Servs., 377 F.3d 1275, 1292 (11th Cir. 2004) (Barkett, J., dissenting from denial of rehearing en banc). 366 See, e.g., JEROME A. BARRON ET AL., CONSTITUTIONAL LAW 640 (8th ed. 2012); 2 DONALD P. KOMMERS ET AL., AMERICAN CONSTITUTIONAL LAW 363 (3d ed. 2010). 367 See Lawrence, No. 02-102, slip op. at 3 (O’Connor, J., concurring in the judgment), http://www.law.cornell.edu/supct/pdf/02-102P.ZC [http://perma.cc/N5C8-RGPP]. 368 See Lawrence v. Texas, LEGAL INFO. INST., http://www.law.cornell.edu/supct/html/02-102.ZC .html (last visited Oct. 26, 2014) [http://perma.cc/62B4-QM74]. 369 See Lawrence v. Texas, 539 U.S. 558, 580 (2003) (O’Connor, J., concurring in the judgment). 370 See Opinions, J. SUP. CT. U.S., Oct. Term 2002, at 1103–04, http://www.supremecourt.gov /orders/journal/jnl02.pdf [http://perma.cc/5VUQ-6SWL]. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 601 But serious practical problems arise when the version of the Court’s opinion upon which lower courts, other branches of government, and scholars and teachers rely can change, without notice, as many as five years after initial publication. Not only do those relying on the Court’s opinions not know of the need to correct their own work, let alone have any practical way to discover the changes made, but their own writings — whether a judicial opinion, casebook, or treatise — can unwittingly perpetuate the error, long after the Court itself has changed its opinion. (vi) Clapper v. Amnesty International USA (2013) (potentially forthcoming). — In Clapper v. Amnesty International USA,371 decided in February 2013, respondents had claimed that section 702 of the Foreign Intelligence Surveillance Act372 is unconstitutional.373 The Supreme Court ruled that respondents lacked Article III standing because they failed to establish the necessary “concrete injury” to themselves from the Act’s administration.374 Because the case was so recently decided, it is too soon to know whether the Court will identify any “formal errors” in the opinion warranting revision prior to publication in the final United States Reports. Review of Justice Alito’s opinion for the Court, however, reveals a possible error and therefore a contemporaneous occasion to witness the Court’s current practices regarding the revision of its opinions. One footnote in the Court’s opinion is especially important because it is where the Court seeks to distinguish its ruling in Clapper from a precedent upon which the respondents had heavily relied. In making that distinction, the Court reasoned: “But to the extent that the ‘substantial risk’ standard is relevant and is distinct from the ‘clearly impending’ requirement, respondents fall short of even that standard, in light of the attenuated chain of inferences necessary to find harm here.”375 There seems to be an error in that crucial footnote. The use of quotation marks on either side of “clearly impending” suggests that the phrase appears earlier in the opinion. That exact phrasing, however, does not appear anywhere else in the Clapper opinion. The phrasing that the Court uses elsewhere is “certainly” impending,376 not “clearly” impending, which are words of potentially very different substantive import. Although the two terms have obvious overlap, their respective references to certainty and clarity can also be differently directed, de––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 371 372 373 374 375 376 133 S. Ct. 1138 (2013). 50 U.S.C. § 1881a (2012). Clapper v. Amnesty Int’l USA, No. 11-1025, slip op. at 1 (U.S. Feb. 26, 2013). See id., slip op. at 2, 10. Id., slip op. at 15 n.5 (citing id., slip op. at 11–15). Id., slip op. at 2, 10, 11, 15, 17, 18, 24. 602 HARVARD LAW REVIEW [Vol. 128:540 pending entirely on the precise context for their application.377 In Supreme Court opinions, perhaps more than any other, the precise meaning of the words used by the Court matter. If this is in fact an error, there are two ways a correction could be made. “Clearly” in footnote five could be changed to “certainly,” which would be consistent with the earlier phrasing. Or the quotation marks around “clearly” could be deleted, which would eliminate the suggestion that the words appeared earlier in the opinion but would leave some ambiguity between the footnote’s phrasing of the applicable test and the rest of the opinion. Until the preliminary print is published in a couple of years, or the United States Reports a couple of years after that, it will not be known for sure whether the Court has decided that this was an error and, if so, whether it warrants revision.378 Until then, moreover, lower courts will continue to rely on language in a Supreme Court opinion that, because it appears likely to result from an initial drafting error, may be changed. To date, nine courts have cited to Clapper’s “clearly impending” language,379 and one of those courts never once uses “certainly impending.”380 Unlike courts citing Lawrence, where the removed statement was a concurring opin––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 377 Compare WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY 367 (1981) (referring to “certainty” in the primary definition of “certainly”), with id. at 420 (defining “clearly” as “in a clear manner”). The application of Clapper to probabilistic injuries is one example where the difference in wording at least arguably makes a difference. Cf. Jonathan Remy Nash, Standing’s Expected Value, 111 MICH. L. REV. 1283, 1297 (2013) (“In spite of the seeming hostility of the Clapper holding to less-than-certain harm constituting injury for standing purposes, there is much in the Clapper opinion that suggests that standing based on probabilities is far from foreclosed, and indeed that the holding in Clapper may be tied to the unique nature of the circumstances raised by the case.”). 378 Consistent with the notice on the slip opinion to notify the Reporter about any “formal errors,” I sent an email to the Reporter regarding this possible discrepancy in the Court’s opinion. See Email from author to Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States (May 14, 2013, 5:28 PM) (on file with the Harvard Law School Library). The Reporter replied that she “will query Justice Alito about your suggestion when we prepare the opinion for publication in the U.S. Reports.” Email from Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States, to author (May 20, 2013, 3:16 PM) (on file with the Harvard Law School Library). 379 See City of Rock Island v. United States, No. 4:13-CV-04047, 2014 WL 4748326, at *5 (C.D. Cal. Sept. 24, 2014); N.Y. Bankers Ass’n, Inc. v. City of New York, No. 13 Civ. 7212, 2014 WL 4435427, at *9 n.7 (S.D.N.Y. Sept. 9, 2014); In re Adobe Sys., Inc. Privacy Litig., No. 13-CV -05226, 2014 WL 4379916, at *7 n.7 (N.D. Cal. Sept. 4, 2014); Wallace v. New York, No. 12-CV -5866, 2014 WL 4243564, at *12 (E.D.N.Y. Aug. 28, 2014); U.S. Hotel & Resort Mgmt., Inc. v. Onity, Inc., No. 13-1499, 2014 WL 3748639, at *4 & n.2 (D. Minn. July 30, 2014); Chevron Corp. v. Donziger, 974 F. Supp. 2d 362, 554 n.1257 (S.D.N.Y. 2014); Galaria v. Nationwide Mut. Ins. Co., Nos. 2:13-CV-118, 2:13-CV-257, 2014 WL 689703, at *5 n.8 (S.D. Ohio Feb. 10, 2014); Taylor v. Bernanke, No. 13-CV-1013, 2013 WL 4811222, at *6 (E.D.N.Y. Sept. 9, 2013); State Nat’l Bank of Big Spring v. Lew, 958 F. Supp. 2d 127, 162 (D.D.C. 2013). 380 See Chevron, 974 F. Supp. 2d at 376–644. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 603 ion’s characterization of a dissenting opinion,381 these courts relied on language from a majority opinion on the relevant standard of law. Moreover, the very subtlety of this potential mistake makes it possibly far more problematic than the full sentence excision in Lawrence. Even if the Court chooses to correct its apparent mistake in Clapper, there will be no ready occasion to correct lower court precedent, also published in official versions, that cites to and quotes from the Court’s original language. (vii) EPA v. EME Homer City Generation, L.P. (2014). — In EPA v. EME Homer City Generation, L.P.,382 decided in April 2014, the Court upheld the validity of a major federal agency rule promulgated by EPA under the Clean Air Act,383 which was designed to curb air pollution from upwind states that interfered with the ability of downwind states to meet and maintain compliance with national ambient air quality standards.384 The Court agreed with EPA that the relevant statutory language was sufficiently ambiguous to allow EPA to consider compliance costs in allocating the relative amount of reduction that was required of the multiple upwind states that were all contributing to a downwind state’s pollution problems.385 The Agency could, accordingly, decide to require greater emissions reductions from those upwind states that could achieve such reductions less expensively than could other upwind states.386 The majority rejected the contrary view, promoted by regulated industry subject to the rule and some states, that the plain meaning of the Clean Air Act mandated a strictly proportionalreduction approach without any consideration of which states could reduce their emissions more cost-effectively.387 Justice Scalia, joined by Justice Thomas, dissented on the ground that the Clean Air Act unambiguously requires EPA to pursue a strict proportional-reduction approach.388 The depth of Justice Scalia’s disagreement with the majority in the case was underscored both by the rhetorical harshness of his slip opinion’s criticism and by his decision to announce the reasons for his dissent orally from the bench immediately after Justice Ginsburg announced the opinion for the Court.389 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 381 382 383 384 See supra pp. 599–600. 134 S. Ct. 1584 (2014). 42 U.S.C. §§ 7401–7671q (2012). EPA v. EME Homer City Generation, L.P., Nos. 12-1182, 12-1183, slip op. at 1–2 (U.S. Apr. 29, 2014). 385 See id., slip op. at 2, 25–28. 386 See id., slip op. at 25–28. 387 See id., slip op. at 2, 26. 388 Id., slip op. at 1, 3 (Scalia, J., dissenting) (joined by Thomas, J.). 389 See Robert Barnes, Scalia v. the Blogosphere: Gaffe Sparks Partisan Outrage, WASH. POST (May 4, 2014), http://www.washingtonpost.com/politics/2014/05/03/8629d9d8-d231-11e3-937f-d3026234b51c _story.html [http://perma.cc/L9GU-B63V]. 604 HARVARD LAW REVIEW [Vol. 128:540 Justice Scalia’s written dissent described EPA personnel as “unelected agency officials” and characterized EPA’s actions as seeking to exercise “broad lawmaking authority” not delegated to the Agency by Congress.390 According to the dissent, because EPA was “unsatisfied” with the way Congress had “specified quite precisely” how much upwind states should reduce emissions that contributed to downwind problems, the Agency had decided to replace, in effect, the congressionally mandated approach with the Agency’s view of a better and “more efficient”391 approach, which of course is exactly what Justice Scalia made clear an agency cannot do.392 As part of Justice Scalia’s narrative on EPA’s departure from its proper function, the dissent highlighted that “[t]his is not the first time EPA has sought to convert the Clean Air Act into a mandate for costeffective regulation.”393 The dissent placed this part of its discussion under the heading “Plus Ça Change: EPA’s Continuing Quest for CostBenefit Authority.”394 And, in support of its narrative, the dissent described how in a prior case, Whitman v. American Trucking Ass’ns,395 the Court had “confronted EPA’s contention that it could consider costs in setting NAAQS” and rejected it.396 Justice Scalia’s dissent correctly characterized the Court’s ruling in American Trucking, but erred in its characterization of EPA’s legal argument in that case. In American Trucking, EPA had argued the exact opposite of what Justice Scalia’s dissent claimed and upon which its narrative about EPA depended. EPA had argued that it could not consider costs in setting national ambient air quality standards.397 EPA’s opponents were the parties in that case that had argued costs could be considered.398 And the Court had therefore agreed with EPA’s view, not rejected it.399 EPA’s actual legal position in American Trucking was entirely inconsistent with the dissent’s narrative about the Agency’s history of overreaching. Justice Scalia was not the first Justice to make this mistake about what EPA had argued in American Trucking,400 but what made the ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 390 391 392 393 394 395 396 EME Homer, Nos. 12-1182, 12-1183, slip op. at 1 (Scalia, J., dissenting). Id. Id., slip op. at 1–2. Id., slip op. at 12. Id. 531 U.S. 457 (2001). EME Homer, Nos. 12-1182, 12-1183, slip op. at 12 (Scalia, J., dissenting); see id., slip op. at 12–13. 397 See Am. Trucking, 531 U.S. at 468–71. 398 See id. 399 See id. at 464–71. 400 Justice Stevens similarly mischaracterized EPA’s argument in American Trucking in his dissent in Entergy Corp. v. Riverkeeper, Inc., 556 U.S. 208 (2009) (prelim. print). See id. at 239 (Stevens, J., dissenting) (“In that case, the Court reviewed the EPA’s claim that § 109 of the Clean Air 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 605 dissent’s error especially remarkable was that Justice Scalia had himself authored the opinion for the Court in American Trucking that had upheld EPA’s reading of the Clean Air Act.401 The Justice’s error in EME Homer was therefore the opinion’s mischaracterizing at length aspects of Justice Scalia’s own prior Court opinion. And he did so with much gusto, which made the mistake all the more problematic. Unlike on other occasions, however, this error was identified not months or years later, but only a few hours after the EME Homer slip opinions were released. The Court was immediately notified through the formal procedures that the Court has established for notifying the Supreme Court Reporter of such “formal errors,” and the Court released a revised slip opinion early the very next morning.402 The original slip opinion heading that began with “Plus Ça Change” was changed to the more benign title “Our Precedent” in order to eliminate any characterization of EPA’s prior practices.403 And the prior references to what “EPA” had argued before were changed to make clear ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Act (CAA), 42 U.S.C. § 7409(a) (2000 ed.), authorized the Agency to consider implementation costs in setting ambient air quality standards.”). The source of the common error, though not remotely an excuse, might be that the Court in American Trucking had granted review of EPA’s petition for a writ of certiorari to consider a nondelegation doctrine issue and a cross petition by industry to consider the cost issue, see Am. Trucking, 531 U.S. at 462–64, and therefore the most casual reader might wrongly assume, based on the case name, that EPA was the petitioner who had lost on both issues in the court below. The relevant language in Justice Stevens’s Entergy dissent has since been corrected in the bound United States Reports. Compare Entergy, 556 U.S. at 239 (Stevens, J., dissenting) (prelim. print) (“In that case, the Court reviewed the EPA’s claim that § 109 of the Clean Air Act . . . authorized the Agency to consider implementation costs in setting ambient air quality standards.” (citation omitted)), with Entergy Corp. v. Riverkeeper, Inc., 556 U.S. 208, 239 (2009) (Stevens, J., dissenting) (“In that case, the Court reviewed a claim that § 109 of the Clean Air Act . . . authorized the EPA to consider implementation costs in setting ambient air quality standards.” (citation omitted)). (I was counsel of record for respondent Riverkeeper in Entergy.) 401 See Am. Trucking, 531 U.S. at 462. 402 See EPA v. EME Homer City Generation, L.P., Nos. 12-1182, 12-1183, slip op. (U.S. Apr. 29, 2014) (rev. Apr. 30, 2014) (Scalia, J., dissenting). As it happens, I was the first to identify the error, within two hours after the opinion was first released. I immediately reported the error to a listserv of academic colleagues (who broadcast the news more broadly through blogs), to a few members of the news media who cover the Court, and within two hours emailed the Supreme Court Reporter a formal letter. Pursuant to the Court’s established procedures, the letter identified the “possible” error, suggested alternative ways to correct the error, and recommended the letter’s referral more immediately to Justice Scalia’s chambers because of the potential to correct the error in an expedited fashion. See Letter from author to Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States 1 (Apr. 29, 2014) (on file with the Harvard Law School Library). 403 Compare EME Homer, Nos. 12-1182, 12-1183, slip op. at 12 (Apr. 29, 2014) (Scalia, J., dissenting) (“Plus Ça Change: EPA’s Continuing Quest for Cost-Benefit Authority”), with EME Homer, Nos. 12-1182, 12-1183, slip op. at 12 (rev. Apr. 30, 2014) (Scalia, J., dissenting) (“Our Precedent”). 606 HARVARD LAW REVIEW [Vol. 128:540 that other parties and not EPA had made the argument about costs that Justice Scalia’s opinion for the Court had rejected.404 The absence of any significant delay before the correction was made contrasts greatly with past experiences, no doubt largely because of the speed of Internet-based communications. The instantaneous availability of the slip opinions on the Internet, within seconds of the opinion’s announcement from the bench, allowed for immediate scrutiny by anyone with Internet access; and, once the error was spotted, the Internet further allowed for broad public dissemination of the error as well as for specific notification to the Court itself.405 Both the error and the correction were consequently quickly publicized throughout the nation and within the Court’s own judicial chambers, even if not at the instigation of the Court itself. The speedy correction is plainly an improvement over corrections that take years to be made. But the risk of significant confusion still remains. After the revisions were made to the slip opinions in EME Homer, there were two different versions of those slip opinions publicly available: the first released on April 29 and the second on April 30. Both have been widely disseminated through electronic media, and there is nothing apparent on the face of either one to indicate which one is the “correct” version. In particular, the second, more recent version nowhere indicates on its face that it constitutes a “revision” of the prior opinion. To that extent, the change is far less transparent, and for that reason even more problematic, than when changes are made between the original slip opinions, the preliminary print, and the United States Reports. At least then, one can discern on the face of the document whether the version was produced earlier or later and therefore is authoritative. The Court has notified the public ahead of time that changes may be made between the slip opinion and the preliminary print, and between the preliminary print and the United States Reports. No notice is provided either that the Court may silently ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 404 Compare EME Homer, Nos. 12-1182, 12-1183, slip op. at 12 (Apr. 29, 2014) (Scalia, J., dissenting) (“This is not the first time EPA has sought to convert the Clean Air Act into a mandate for cost-effective regulation. Whitman . . . confronted EPA’s contention that it could consider costs in setting NAAQS.”), with EME Homer, Nos. 12-1182, 12-1183, slip op. at 12 (rev. Apr. 30, 2014) (Scalia, J., dissenting) (“This is not the first time parties have sought to convert the Clean Air Act into a mandate for cost-effective regulation. Whitman . . . confronted the contention that EPA should consider costs in setting NAAQS.”). 405 Both the initial error and the correction received substantial publicity. See, e.g., Jonathan H. Adler, Justice Scalia Is Not the Only One Making Errors, WASH. POST: VOLOKH CONSPIRACY (May 1, 2014), http://www.washingtonpost.com/news/volokh-conspiracy/wp/2014/05/01 /justice-scalia-is-not-the-only-one-making-errors [http://perma.c c/K8C7-RBX6]; Barnes, supra note 389; Sahil Kapur, Antonin Scalia’s Blunder Is Unprecedented, Legal Experts Say, TALKING POINTS MEMO (May 1, 2014, 6:00 AM), http://talkingpointsmemo.com/dc/antonin-scalia-blunder -unprecedented-epa [http://perma.cc/U4HT-4DBW]. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 607 change its online electronic versions of its slip opinions or, as in EME Homer, that specific slip opinions have in fact been changed.406 (f) Frequency of Revisions. — The full extent and frequency of opinion revisions through printer proofs and change pages are, as a practical matter, unknowable. They are nominally in plain sight because the Court does not shy away from acknowledging that it makes changes. Indeed, the Court expressly invites readers to notify the Reporter of errors. The changes themselves are in a strictly technical sense visible because they can theoretically be discovered by comparing the original bench opinion and the final version published in the United States Reports. Only relatively recently have technological advances both in the electronic digitization of legal information, including court opinions, and in comparison-software technology combined to promise the potential for ready identification of revisions made in Supreme Court opinions. Before then, the revisions were wholly invisible as a practical matter. That is perhaps why, until this Article, the public and academics have apparently assumed that only revisions of a technical, nonsubstantive nature were made under the banner of correcting “typographical or other formal errors.” Based on the examples identified above, it is clear that any such assumption would be mistaken. The revisions that the Court makes through this pathway are categorically no different from the changes it has made through other methods. And the decrease in errata lists and opinion-modification orders in recent decades is most likely explained by an increase in the use of change pages instead. The revisions made through change pages have plainly included changes of substance, the examples of which reveal Justices changing opinions for the Court and for themselves individually for wide-reaching and potentially openended reasons. The Justices reconsider the strengths of legal arguments made and revise language accordingly.407 They reword core legal conclusions in light of further thinking.408 They make revisions in response to concerns about the possible import of unintended dictum.409 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 406 Nor is there any reason to presume that such unannounced slip opinion revisions are limited to only certain Justices. A few weeks after Justice Scalia changed his opinion in EME Homer, Justice Kagan corrected a factual error in her slip opinion dissenting in Town of Greece v. Galloway, 134 S. Ct. 1811 (2014). Compare Town of Greece v. Galloway, No. 12-696, slip op. at 23 (U.S. May 5, 2014) (Kagan, J., dissenting) (describing Newport, Rhode Island as “the home of the first community of American Jews”), with Town of Greece v. Galloway, No. 12-696, slip op. at 23 (U.S. May 5, 2014) (rev. n.d.) (Kagan, J., dissenting) (describing Newport as “the home of one of the first communities of American Jews”). 407 See supra pp. 593–95. 408 See supra p. 596. 409 See supra pp. 597–99. 608 HARVARD LAW REVIEW [Vol. 128:540 And they make changes after reconsidering the accuracy of their initial characterizations of the published opinions of other Justices.410 It is no safer to assume that the Court is less likely to make significant substantive changes in this manner because it lacks the transparency provided by errata lists and opinion-modification orders. Indeed, precisely the opposite could be true. The lack of transparency could make this pathway a more attractive option for making changes, including some with substantive import, because it allows the Court to correct mistakes without any direct, public acknowledgment. Lending further support to the possibility that the Court is making more revisions than ever through change pages is the remarkably increasing amount of time required to prepare the United States Reports. Notwithstanding huge advances in technology, the diminished role of the Reporter, and the marked decrease in the number of cases decided, publication of the United States Reports currently requires four to five times longer than it used to take. In 1825, the Reporter published the United States Reports the same year the decisions were announced. In 1875, they were published one year later; in 1925 within a year; and in 1975, in less than two years.411 By contrast, in recent years, it has typically taken at least four or five years after the original opinion announcement for the corresponding volume of the United States Reports to be published.412 The Justices seem to be trying to have it both ways. On the one hand, the Justices make no effort to hide the fact that they engage in the practice of changing opinions — they publicly announce it. Yet, at the same time, they deliberately make it hard for anyone to determine when changes are made, although they could easily make that information public. In short, the Court is simultaneously transparent and opaque. Part III of this Article next discusses whether such duality is sensible and appropriate. III. ENSURING THE COURT’S INSTITUTIONAL INTEGRITY BY IMPROVING THE PROCESS AND PRACTICE OF REVISION The Supreme Court’s practice of revising its opinions is praiseworthy in its objective. As the nation’s highest court, an unbending com––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 410 411 See supra pp. 599–600. Author’s calculations based on the dates of the publication for the United States Reports in 428 U.S. (1978); 427 U.S. (1978); 426 U.S. (1978); 425 U.S. (1978); 424 U.S. (1977); 423 U.S (1977); 271 U.S. (1927); 270 U.S. (1926); 269 U.S. (1926); 91 U.S. (1876); and 23 U.S. (10 Wheat.) (1825). Omitted from this analysis is Volume 92 for October Term 1875, only because of an apparent anomaly relating to the timing of its publication. That volume says it was “entered” with the Library of Congress in 1876 but not “published” until 1890, suggesting some kind of oddity not representative of that time period. See 92 U.S. i–ii (1890). 412 See, e.g., supra p. 543. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 609 mitment to precision and accuracy is plainly a good thing.413 A court that decides some of the nation’s most pressing legal issues should not have typographical, punctuation, or grammatical errors in its rulings. Nor should its rulings rest on or otherwise perpetuate significant factual errors. The Court’s opinions are intended to have a long life and provide an unquestioned source of controlling authority even when ruling on controversial matters, including striking down actions taken by the executive and legislative branches on constitutional grounds. Opinion errors, even of the most seemingly innocuous nature, would seriously erode the Court’s prestige and authority. To its credit, the Court freely acknowledges the possibility of opinion revisions to correct mistakes by providing a formal notice on every slip opinion and preliminary print, although the Court has never acknowledged that it makes changes in electronic versions of its slip opinions made available online. The Court’s practice of revising its slip opinions and preliminary prints is also exceedingly impressive in application. The kind of fundamental error made by the third Supreme Court Reporter in no less than McCulloch v. Maryland,414 misquoting the Necessary and Proper Clause,415 would most certainly not happen under the rigorous review procedures that the Court currently follows. The Justices, their law clerks, and the Reporter today are painstakingly precise and rigorous at each stage. A review of the change pages for one recent volume of the United States Reports makes that clear.416 It reveals a depth and breadth of review fitting for the High Court. That is most likely why, notwithstanding the far ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 413 In some instances, however, the Court has allowed fairly fundamental factual errors about the names of parties in famous cases to persist, perhaps because the error has no substantive significance, but the case name has become so well known that it would be too disruptive to correct. Examples include “Sanford” in Dred Scott v. Sandford, 60 U.S. (19 How.) 393 (1857), “Tysen” in Swift v. Tyson, 41 U.S. (16 Pet.) 1 (1842), and “McCulloh” in McCulloch v. Maryland, 17 U.S. (4 Wheat.) 316 (1819). See CHARLES ALAN WRIGHT & MARY KAY KANE, LAW OF FEDERAL COURTS § 54 n.5 (7th ed. 2011); see also id. (discussing Ford Motor Credit Co. v. Milhollin, 444 U.S. 555 (1980)) (“[T]he Court, in an introductory footnote *, says that the respondents spell their name ‘Millhollin’, but since it has been misspelled throughout the litigation and since ‘legal research catalogs and computers are governed by the principle of consistency, not correctness, we feel constrained to adhere to the erroneous spelling.’” (quoting Ford Motor Credit, 444 U.S. at 555 n.*)). 414 17 U.S. (4 Wheat.) 316. 415 Compare U.S. CONST. art. I, § 8, cl. 18 (“To make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers and all other Powers vested by this Constitution in the Government of the United States, or in any Department or Officer thereof.”), with McCulloch, 17 U.S. (4 Wheat.) at 412 (mistakenly omitting the words “or Officer”). This error does not appear to have been previously corrected, and an academic colleague recently notified the Reporter of the “formal error” and suggested its possible correction, presumably in a future errata listing. See Letter from John F. Manning, Professor, Harvard Law Sch., to Christine L. Fallon, Reporter of Decisions, Supreme Court of the United States (Oct. 15, 2014) (identifying error in McCulloch for possible correction) (on file with the Harvard Law School Library). 416 See supra notes 81, 111. 610 HARVARD LAW REVIEW [Vol. 128:540 greater resources now available in each Justice’s chambers and the enormous advances in technology, it takes far longer to produce the final bound volumes of the United States Reports than it did decades ago.417 The many examples described above, however, make a strong case in favor of the Court’s revising its current practices. In particular, they demonstrate the considerable and potentially increasing costs of the Court’s lack of transparency. Described below are some of those costs, some possible lessons from how the other branches of the federal government address similar mistakes, and some possible reforms for the Court’s consideration. A. Weaknesses of the Court’s Current Revision Practices The Court’s current practices for revising its opinions fall short in at least two significant respects. First, the Court lacks a coherent, sensible set of procedures for determining when revisions should be made that is commensurate to the degree of procedural rigor applied prior to the initial opinion publication. And, second, the Court makes it far too difficult for anyone to discover what corrections have in fact been made. Both problems warrant changes in the Court’s current practices. The first has led to seemingly haphazard revision of “formal errors” and to too few safeguards for determining when a revision is warranted. The second simply makes no sense and disserves both the Court and the public. Every piece of the puzzle is made public, from the fact that changes are made to the various versions of an opinion. There is nothing remotely confidential about the information contained in those change pages and no good reason not to show the changes themselves. 1. Absence of Apparent Procedural Rigor. — One of the most striking aspects of the Court’s current and historic practices for revising an opinion is the apparent lack of procedural rigor within an institution known for such rigor. There is no discernible pattern for how or why the Court has made different kinds of changes in different ways over time. It seems almost haphazard. The most incidental technical changes have been made in the most public way (for example, erratum ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 417 Of course, as all authors (and law review editors) know, it is frustratingly impossible to eliminate all errors. The Supreme Court is no exception in this respect. One amusing example relates to the Reporter’s apparently mistaken elimination of language from a concurring opinion. In Lee v. Weisman, 505 U.S. 577 (1992), Justice Souter’s concurring slip opinion included a footnote paragraph that ended with the acknowledgment that “Homer nodded,” see MARK TUSHNET, A COURT DIVIDED 60, 353 (2005), but that statement is missing from the bound volume, see Lee, 505 U.S. at 623 n.5 (Souter, J., concurring). The Reporter corrected the error in an errata page a decade later. See Erratum, 535 U.S. II, II (2003) (“505 U.S. 624, n. 5, line 10: add ‘Homer nodded.’ after ‘. . . n. 3, supra.’”). 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 611 notices and formal orders modifying opinions),418 and changes in wordings with obvious substantive import have been made in the least transparent way.419 The answers to basic questions — including what constitutes a “formal error,” the full extent to which it embraces substantive changes, and how the Justices are consulted about possible changes — are unclear.420 What is clear, however, is that substantive changes are made under the rubric of “formal error,” and that there are few, if any, procedural assurances that modifications are made with awareness of all chambers, let alone the approval of those whose votes were necessary in the first instance to render an opinion of a single Justice anything more than that. The role of the public and the parties in prompting these changes likewise lacks any clear structure. Before a case is decided, there are strict rules that everyone must follow in any communications with the Court. All parties must receive copies of any document filed by any other party or amicus. Ex parte communication is strictly and appropriately forbidden. But under the Court’s remarkably informal process for revising its opinions, any member of the public — including any party, amicus, or counsel — may and apparently routinely does unilaterally write the Court about a “formal error” and seek its revision.421 No service or notice is required. And the Court may revise its opinion in substantive ways without any opportunity for input from interested parties. 2. Absence of Appropriate Transparency and the Potential Undermining of Opinion Integrity. — The Court’s dominant method of making changes to its opinions is the least transparent, and deliberately so. The costs, moreover, are considerable. For instance, as described above, the lack of notice invariably perpetuates the Court’s initial error long after the correction has been made and risks considerable confusion about what the law actually is. During the five years between the time of the Court’s release of its original slip opinion and the Court’s publication of the final, official version in the United States Reports, lower federal and state courts, lawmakers from other branches of government, legal scholars, and law teachers rely on, cite to, and quote from that initial language. And those citations and quotations naturally persist long after the Court has changed its opinions. Those relying on the slip opinions’ precise wording do not truly appreciate its tentative nature, notwithstanding the formal notice at––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 418 419 420 See supra pp. 573–80. See, e.g., supra pp. 597–99. As discussed above, in 1994, Chief Justice Rehnquist apparently notified the Court Reporter that, in the future, the Court would need to approve substantive changes in the Court’s opinions. See supra p. 585. 421 See, e.g., supra pp. 581–82, 598–99. 612 HARVARD LAW REVIEW [Vol. 128:540 tached to it, and the Court’s deliberately nontransparent procedures make it exceedingly difficult to discover the changes made even if those lower courts, lawmakers, scholars, and teachers tried to discover them. Certainly, those textbook authors who, many years after Justice O’Connor changed her concurring opinion in Lawrence, continue to reproduce in their texts the prior language would immediately update their publications were notice of the Court’s changes more readily available.422 Finally, the Court’s current policy also undermines the integrity of the Court’s central product: its written opinion. It makes little sense to suggest, as the Court currently does, that its opinions formally announced in open Court are not “final” and “official.” Perhaps that policy made sense a century ago, but it no longer does. We are far from the early nineteenth century when the initial opinions announced from the bench were truly preliminary and the Supreme Court Reporter played a heavy hand in converting those oral opinions into written opinions. The Court’s current process for drafting opinions is appropriately demanding and produces a written text that becomes law only when thoroughly reviewed by all chambers and when a majority of the Justices formally and expressly “join” in its endorsement. That slip opinions and preliminary prints may be subject to change does not make them any less “final” or “official” than are the bound volumes of the United States Reports, which are always subject to possible subsequent correction through errata lists. B. Lessons from the Other Branches The other two branches of the federal government provide contrasting examples of how a lawmaking entity can correct its mistakes in a more systematic, careful, and transparent manner. To be sure, the judicial branch has different lawmaking responsibilities than the legislative and executive branches. Congress can enact a law only upon adherence to certain strict procedural requirements set forth in the Constitution, which circumscribes the legislature’s ability to correct mistakes made in its initial lawmaking efforts. So too, executive branch agencies must make and administer law pursuant to congressionally mandated procedures, and those procedures invariably govern those agencies’ abilities to correct their mistakes. Judicial opinionwriting is not similarly circumscribed, and the courts, including the Supreme Court, are far freer to determine the best processes for drafting and revising opinions.423 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 422 423 See supra p. 600. The Supreme Court has long embraced the inherent power of courts “to correct mistakes of the clerk or other officer of the court, inadvertencies of counsel, or to supply defects or omissions 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 613 A judicial opinion is also plainly different from both statutory and regulatory text in an even more fundamental respect. The opinion accompanies and explains the judgment, but only the judgment has discrete legal effect. Judges possess tremendous discretion in using an opinion to explain the reasoning underlying the court’s judgment. The practice of routinely including a written opinion to explain a judgment is one that emerged over time here in the United States in fits and starts during the nineteenth century, with some states even enacting statutes to require the issuance of written opinions.424 While the distinction may be strictly true as a historical matter, the role of the judicial opinion generally today, and certainly the function of a Supreme Court opinion in particular, cannot be so easily cabined. The judgment, alone, has no precedential significance beyond the facts of the particular case. It is the precedential effect of the Court’s opinion that renders the High Court such an important lawmaking institution.425 Of course, not all words in a Supreme Court opinion are of equal stature. The Court’s opinions include extensive explanatory text, some of which is central to the ruling but much of which is dicta without binding precedential effect on the lower courts or on the other governmental branches. And, therefore, the safeguards appropriate for the revision of such explanatory text are not strictly analogous to those necessary in revising statutory or regulatory language, all of which states binding law. The inquiry here, however, is not what the Supreme Court must do, but what practices it should follow in revising its opinions to be consistent with both the rigorous procedures that the Court imposes on the advocacy that frames its decisionmaking and the exalted role the Court’s opinions serve in our nation’s laws. And in ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– in the record.” Gagnon v. United States, 193 U.S. 451, 456 (1904) (citing Gonzales v. Cunningham, 164 U.S. 612, 623 (1896)); see In re Wight, 134 U.S. 136, 143–44 (1890); United States v. Vigil, 77 U.S. (10 Wall.) 423, 426–27 (1870); see also Am. Trucking Ass’ns v. Frisco Transp. Co., 358 U.S. 133, 145 (1958) (“It is axiomatic that courts have the power and the duty to correct judgments which contain clerical errors or judgments which have issued due to inadvertence or mistake.”). The Federal Rules of Civil Procedure expressly authorize courts to correct “a clerical mistake or a mistake arising from oversight or omission” in a “judgment, order, or other part of the record,” FED. R. CIV. P. 60(a), and the Federal Rules of Criminal Procedure include a comparable provision, see FED. R. CRIM. P. 36. The former provides that the court “may do so on motion or on its own, with or without notice” (until an appeal has been docketed), FED. R. CIV. P. 60(a), and the latter “[a]fter giving any notice [the court] considers appropriate,” FED. R. CRIM. P. 36. 424 POPKIN, supra note 24, at 93–95. 425 For this same reason, the text of judicial opinions cannot be fairly analogized to the legislative history accompanying a statutory enactment. In all events, Congress does not (to my knowledge) revise the language of legislative reports after a bill has become law, and when individual legislators subsequently revise or add to remarks made on a chamber floor, Congress includes (since 1978) a bullet or other explicit designation to make clear to readers that this language was added later. See Richard J. McKinney, An Overview of the Congressional Record and Predecessor Publications, L. LIBR. LIGHTS, Winter 2002, at 16, 20–21. 614 HARVARD LAW REVIEW [Vol. 128:540 addressing that latter question, the Court could learn from the experiences and practices of the other two branches. For instance, Congress has developed a host of procedural pathways for correcting errors, depending on when the mistakes are discovered and corrected.426 If Congress identifies an error in an enrolled bill after its passage by both chambers, but before the President has acted, both chambers can correct that error by passing a concurrent resolution directing changes in enrollment.427 Such resolutions direct the House Clerk or the Secretary of the Senate to make specified corrections needed to ensure that the bill’s language accurately reflects congressional intent. If the legislative officers have already signed the bill, the resolution rescinds their signatures, and if the bill has already been presented to the President, the resolution requests that the President return the bill to Congress.428 Corrections can include typographical mistakes, substantive errors, and clerical errors that resulted in each chamber agreeing to a different text.429 If the mistake is not discovered until after the President has signed the enrolled bill — which, therefore, is now enacted law — Congress follows a different procedure, which requires Congress to pass and the President to sign a new statute that revises the prior law.430 To this end, Congress has an entire office committed to correcting mistakes in its statutory enactments and the United States Code: the Office of the Law Revision Counsel. One of the Office’s primary responsibilities is to “remove ambiguities, contradictions, and other imperfections both of substance and of form” in statutory enactments and their initial codification, with the aim of having Congress pass the corrected version as positive law.431 Routine changes include improving wording and revising errors in spelling, punctuation, and grammar.432 In addition, Congress routinely enacts narrowly directed “technical ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 426 WM. HOLMES BROWN ET AL., HOUSE PRACTICE: A GUIDE TO THE RULES, PRECEDENTS, AND PROCEDURES OF THE HOUSE 772–76 (2011); VALERIE HEITSHUSEN, CONG. RESEARCH SERV., RL34480, ENROLLMENT OF LEGISLATION: RELEVANT CONGRESSIONAL PROCEDURES 3–5 (2013). 427 See HEITSHUSEN, supra note 426, at 3. An enrolled bill is one that has been agreed to by both chambers, and enrollment occurs in the chamber where the bill originated. Then the head of each chamber must sign the bill to authenticate that the text reflects what the chamber passed. Only then is the bill presented to the President. Id. at 1–2. 428 Id. at 3–4. For instance, Senate Concurrent Resolution 32 in the 112th Congress instructed the House Clerk to correct the enrollment of H.R. 470, the Hoover Power Allocation Act. See Hoover Power Allocation Act of 2011, Pub. L. No. 112-72, 125 Stat. 777; S. Con. Res. 32, 112th Cong. (2011) (enacted). 429 HEITSHUSEN, supra note 426, at 3. 430 Id. at 4–5. 431 2 U.S.C. § 285b(1) (2012). 432 See generally OFFICE OF THE LAW REVISION COUNSEL, U.S. HOUSE OF REPRESENTATIVES, POSITIVE LAW CODIFICATION IN THE UNITED STATES CODE (2014), http://uscode .house.gov/codification/positive_law_codification.pdf [http://perma.cc/H4MY-HBSX]. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 615 amendments”433 to correct unintended formatting and substantive errors in specific legislation.434 However loosely Congress might apply the “technical” label, its process for identifying and correcting errors that invariably creep into its work product offers lessons for the Court: Congress does not engage in the strange fiction of labeling its initial enactments as neither “final” nor “official.” Congress publishes its procedures for making revisions. And Congress openly describes the changes that it has made, without mystery or embarrassment. The ultimate safeguard is that Congress itself must approve, by formal passage, any proposed changes in the relevant statutory language. Federal executive branch agency procedures for making corrections are similarly far more coherent and transparent than the Court’s. The Administrative Procedure Act435 (APA) generally governs agency exercises of legislatively delegated lawmaking authority, supplemented by additional requirements and modifications set forth in specific laws under which agencies act.436 More specifically, the APA sets forth procedural requirements applicable to agency rulemakings and adjudications, for which prior notice and public participation are central to the legitimacy of the agency’s exercise of authority.437 Those same APA requirements generally apply to an agency’s efforts to correct errors. As the U.S. Court of Appeals for the District of Columbia Circuit explained, agencies possess no “‘inherent power’ to correct ‘technical errors[]’ . . . without complying with the APA’s procedural requirements.”438 Moreover, where the APA provides various ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 433 434 Envtl. Def. v. Duke Energy Corp., 549 U.S. 561, 576 (2007). See, e.g., id. (“Nothing in the text or legislative history of the technical amendments that added the cross-reference to NSPS [(new source performance standards)] suggests that Congress had details of regulatory implementation in mind when it imposed PSD [(prevention of significant deterioration)] requirements on modified sources . . . .”); Lockhart v. United States, 546 U.S. 142, 145–46 (2005) (holding it is irrelevant that Congress may not have foreseen all the legal consequences of the language it enacted in the Higher Education Technical Amendments); see also, e.g., 159 CONG. REC. H2169 (daily ed. Apr. 23, 2013) (regarding “Technical Corrections and Improvements in Title 36, United States Code”). The Court considers the “technical” nature of these changes in construing their legal import. See, e.g., Muniz v. Hoffman, 422 U.S. 454, 472 (1975) (“[I]t would seem difficult at best to argue that a change in the substantive law could nevertheless be effected by a change in the language of a statute without any indication in the Reviser’s Note of that change.”). 435 Pub. L. No. 79-404, 60 Stat. 237 (1946) (codified as amended in scattered sections of 5 U.S.C.). 436 See 5 U.S.C. §§ 551–554 (2012). 437 See id. §§ 553–554. A recent law review article ambitiously proposes that courts more fully embrace the administrative agency model by engaging in “notice-and-comment judicial decisionmaking” to reduce the likelihood of judicial error. See Michael Abramowicz & Thomas B. Colby, Notice-and-Comment Judicial Decisionmaking, 76 U. CHI. L. REV. 965 (2009). 438 Util. Solid Waste Activities Grp. v. EPA, 236 F.3d 749, 752 (D.C. Cir. 2001). In Utility Solid Waste Activities Group v. EPA, 236 F.3d 749, EPA sought “to correct language resulting from an 616 HARVARD LAW REVIEW [Vol. 128:540 avenues for correcting mistakes that do not require compliance with all the requirements that apply to original decisions, safeguards remain in place to ensure that the transparency inherent in the usual procedures is not sacrificed. For instance, there is a “good cause” exception applicable to agency rulemaking that allows an agency to forego prior notice and public procedure when it is “impracticable, unnecessary, or contrary to the public interest.”439 But, even under that narrow exception “limited to emergency situations,”440 the agency must publish an after-the-fact notice of specific changes made in the Federal Register.441 Agencies may forego prior notice and comment rulemaking using either “interim final rulemaking” or “direct final rulemaking.”442 With the former, the agency adopts a rule without prior notice and comment, makes the rule effective immediately, and invites comment on whether the rule should be changed in the future. With the latter, there is similarly no prior public notice and opportunity for comment, and the agency states that if it receives a “significant adverse”443 comment, the final rule will be rescinded and normal notice and comment rulemaking requirements triggered.444 Interim final rulemaking is used for important, urgent rulemaking, which may well therefore be controversial, while direct final rulemaking is designed for presumptively noncontroversial rulemakings, extending to corrections to prior rulemakings, which is why the receipt of a significant adverse comment automatically triggers rule rescission and the need for full notice and comment rulemaking.445 Finally, some agencies have occasionally published “Errata Notices” and “Corrections” to revise errors in prior Federal Register publications, including in the content of final rules. Examples of corrections ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– erroneous use of the Word Perfect find/replace command in the drafting of the regulation,” id. at 752. 439 5 U.S.C. § 553(b)(3)(B); see also Ellen R. Jordan, The Administrative Procedure Act’s “Good Cause” Exemption, 36 ADMIN. L. REV. 113 (1984). 440 Util. Solid Waste Activities Grp., 236 F.3d at 754 (quoting Am. Fed’n of Gov’t Emps. v. Block, 655 F.2d 1153, 1156 (D.C. Cir. 1981)) (internal quotation mark omitted). 441 See 5 U.S.C. § 552(a)(1)(E). 442 Ronald M. Levin, Direct Final Rulemaking, 64 GEO. WASH. L. REV. 1, 2 (1995); see id. at 1–4. 443 Id. at 32 (internal quotation marks omitted). 444 See Adoption of Recommendations, 60 Fed. Reg. 43,108, 43,110–12 (Aug. 18, 1995); Michael Kolber, Rulemaking Without Rules: An Empirical Study of Direct Final Rulemaking, 72 ALB. L. REV. 79, 88–91 (2009); Levin, supra note 442, at 2–3, 4–7 (describing differences between “interim final rulemaking” and “direct final rulemaking,” id. at 2, and how EPA “invented direct final rulemaking in the early 1980s,” id. at 4); Ronald M. Levin, More on Direct Final Rulemaking: Streamlining, Not Corner-Cutting, 51 ADMIN. L. REV. 757, 759–61 (1999); Lars Noah, Doubts About Direct Final Rulemaking, 51 ADMIN. L. REV. 401, 419–23 (1999). 445 Levin, supra note 442, at 2–3, 6. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 617 include typographical errors, misspellings,446 missing words,447 incorrect numbers,448 and even corrections to prior corrections.449 In all these circumstances, the federal agencies published the precise corrections and changes being made. Notwithstanding the obvious differences between judicial opinions, statutory provisions, and agency regulations, there are potentially important lessons to be gleaned by the Court from the contrasting examples of how the legislative and executive branches correct their own errors. Most important, their examples strongly suggest that the institutional costs to the Court of adopting a more rigorous and transparent process for error correction might be less than assumed by the justifications presumably underlying the Court’s current practices. They further underscore the shared judgment of two other branches that the public can fairly expect a high level of transparency from its lawmaking institutions. First, Congress and federal agencies have demonstrated that publicly acknowledging revisions need not prove embarrassing or delegitimizing to a government branch. Such error correction is embarrassing only when one attempts to hide it, and it is discovered nonetheless. As soon as it becomes routine and appreciated as an ordinary and entirely understandable part of an institution’s necessary practices, as has happened with the two other branches, the risk of any damaging stigma quickly disappears. Second, the legislative and executive branch examples demonstrate that there can be flexibility in the precise procedures adopted for correcting errors. Errors can be properly classified based on their nature. And different kinds of errors can warrant different kinds of correction procedures and levels of transparency, including potentially no transparency at all for some types of errors given the different nature of a judicial opinion and the types of revisions frequently needed. Finally, the contrasting examples of the legislative and executive branches suggest the possibility that the absence of rigor and transparency in the Court’s own practices may be interrelated. In particular, ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 446 Correction, 65 Fed. Reg. 81,961, 81,961 (Dec. 27, 2000) (correcting eight misspellings and typographical errors in prior NOAA guidance). 447 Errata Notice, 67 Fed. Reg. 43,309, 43,309 (June 27, 2002) (adding words to make clear FERC’s obligation to consider and incorporate the views of “affected stakeholders” and not only state public utility commissions). 448 Errata Notice, 74 Fed. Reg. 22,773, 22,773 (Apr. 17, 2013) (changing time frame for filing modifications with FERC from forty-five to sixty days). 449 In 1979, the Department of the Interior published seventy corrections to an errata sheet that had corrected agency regulations and the accompanying preamble. See Surface Coal Mining and Reclamation Operations Permanent Regulatory Program; Corrections, 44 Fed. Reg. 53,507, 53,507–09 (Sept. 14, 1979) (correcting errata notice at Surface Coal Mining and Reclamation Operations, Permanent Regulatory Program; Corrections, 44 Fed. Reg. 49,673, 49,673–87 (Aug. 24, 1979)). 618 HARVARD LAW REVIEW [Vol. 128:540 the existing absence of transparency may partly explain why the Court lacks any systematic approach for determining how and what changes are made. Accountability can be a strict, even if indirect, disciplinarian. A modicum of exposure would likely produce a more sensible process of correcting Court opinions. Once the Court routinely makes known the corrections, the Justices would invariably develop and apply more consistent and disciplined procedures in determining what types of corrections are warranted in specific circumstances. C. Possible Reforms for the Court’s Consideration There are myriad ways that the Court could improve its existing procedures for revising its opinions following their initial publication. At a minimum, however, the Court should consider the following. First, the Court should consider clarifying what it means by claiming that its slip opinions (and preliminary prints) are not “final” or “official.” The Court’s opinion is plainly “final” with reference to how the Court itself uses that term in describing the culmination of lawmaking processes. It has “all of the hallmarks”450 of finality: it has immediate legal effect, has legal consequences for the parties, and ends the Court’s decisionmaking process.451 The Court can reserve the right to correct unintended errors without that inapt disclaimer. Nor is the notion that the Court’s opinion is not “official” until the publication of the bound United States Reports entirely correct. The opinion that the Court initially releases is not some informal writing that the Reporter then takes and crafts into a finished work. It is the result of extraordinarily careful and rigorous drafting, negotiating, and revising conducted within the chambers of the Justices. The Reporter’s job, by contrast, is ministerial. The slip opinions are physically produced within the Court and contain all the official imprimatur of the Court. In short, it is about as “official” as one could possibly get. Second, the Court should consider providing more guidance on what constitutes a “formal error” and adopting different procedures for addressing different kinds of errors. The inherent vagueness of the Court’s current notice accompanying slip opinions and preliminary prints invites arbitrariness. Certain kinds of errors are obviously within the Reporter’s purview, such as typographical and formatting errors. But, as demonstrated by the examples above, the Court’s application of the term “formal error” extends to plainly substantive changes that are far from obvious and are not of a comparably technical nature. The Court can fairly distinguish among different types of corrections and the kinds of procedural safeguards and degrees of transpar––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 450 451 Sackett v. EPA, 132 S. Ct. 1367, 1371 (2012). Cf. id. at 1371–72. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 619 ency they require. Misspellings, typographical errors, and citation errors or updates trigger, numerically, the vast majority of revisions and presumptively warrant no special safeguards. The same would be true for almost all (but not all) grammatical errors — even something as mundane and incidental as a shift in punctuation could have significant substantive import.452 The addition, deletion, or substitution of words or phrases, by contrast, should presumptively trigger more procedural safeguards and transparency because of the greater likelihood of substantive import. Proposed revisions aimed at correcting erroneous characterizations of facts, the record, the positions of parties, the positions of other Justices, or background law likewise warrant greater safeguards because of their potential substantive import. Corrections about background law, illustrated by the Court’s revisions of its description of the Clean Water Act in Ouellette,453 are especially susceptible to unintended consequences, making prior notice to and input from the parties more valuable. At the very least, the Court should reconsider its current practice of delegating the decision whether to make these kinds of word changes to the Justice who authored the opinion, including even when that Justice has since retired.454 Wholly apart from the legitimate threshold question whether such retroactive changes should be made at all, especially when the membership of the Court might well have since significantly changed, it is far from obvious that the decision should be left to the Justice who authored the opinion. As the Court’s own guidance to oral advocates underscores, “an opinion of the Court” is not the opinion of the single Justice who writes it for the Court;455 it is an opinion of the Court. For this same reason, it can be forcefully maintained that any such change in the wording of the Court’s opinion is presumptively substantive enough to warrant at least the opportunity for other chambers (at least those in the majority) to review. Third, the Court should also consider distinguishing among possible revisions based on the identity of the person proposing the change. The procedures need not be identical for suggestions proposed by those outside the Court, whether a party, amicus, or third person, and those proposed within the Court. For the former, there should be a thresh––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 452 453 454 455 Cf. District of Columbia v. Heller, 554 U.S. 570 (2008). See supra pp. 597–98. See supra pp. 584–85. See CLERK OF THE COURT, SUPREME COURT OF THE UNITED STATES, GUIDE FOR COUNSEL IN CASES TO BE ARGUED BEFORE THE SUPREME COURT OF THE UNITED STATES 11 (2013), h t t p : / / w w w . s u p r e m e c o u r t . g o v / o r a l _ a r g u m e n t s / g u id e f o r c o u n s e l . p d f [h t t p: // perma.cc/D968-LMES] (“Do not refer to an opinion of the Court by saying: ‘In Justice Ginsburg’s opinion.’ You should say: ‘In the Court’s opinion, written by Justice Ginsburg.’”). 620 HARVARD LAW REVIEW [Vol. 128:540 old presumption that such communications are public or, at the very least, made known to the parties in the case (and typically to the Solicitor General in light of his extensive involvement with the Court’s merits docket).456 The Court could implement formal service requirements applicable to parties and amici, but the Clerk of the Court would ultimately be responsible for ensuring that the parties received prior notice when the impetus for the revision originated outside the Court from a nonparty or amicus. The procedures could fairly be different if the initial impetus for the correction came from one of the Chambers or an officer of the Court, including the Reporter. In that circumstance, a presumption against prior notice to the parties could fairly apply, consistent with the strictly confidential nature of the Court’s deliberative process once a case is submitted. Even then, however, the Court might decide to provide prior notice of its consideration of such changes to the parties (and to the Solicitor General as an amicus). The Justices are well aware of the value that briefs and oral arguments provide. The Court is highly dependent on effective advocacy in its initial decisionmaking and opinion writing, and there is no obvious reason why such input would not be similarly valuable when the Court (or an individual Justice) is contemplating a revision. The parties and their counsel are frequently aware of aspects of the case and the possible consequences of rewordings not readily anticipated by the chambers. That is especially true when changes are being made long after the opinion was originally announced, so that the clerks who worked on the opinion are no longer there. Without the assistance of the parties, the Court risks substituting one kind of error for another. Fourth, the Court should provide after-the-fact public notice of any revisions made, just as Congress does in revising its legislation and federal agencies do in correcting errors in regulations. This could be easily accomplished. All the Court would need to do is post to its website the same change pages that it already provides to publishers.457 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 456 See Margaret Meriwether Cordray & Richard Cordray, The Solicitor General’s Changing Role in Supreme Court Litigation, 51 B.C. L. REV. 1323, 1338–60 (2010). 457 As described in note 462, infra, just as this Article was going to press, the Court publicly announced corrections being made by Justice Ginsburg to a recent slip opinion that she authored in dissent. The announcement suggests the possibility that the Court as a whole has now decided to provide such post-correction notification, most likely in response to significant media attention received by an earlier draft of this Article. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 621 Further, because the Court’s revisions occur in the context of litigation where there are individual parties with concrete stakes, the Court should provide specific notice to those parties. As with “direct final rulemaking,” a significant adverse comment might prompt the Court to reconsider its revision, albeit in rare instances. Of course, any notice requirement (whether before or after the revision) may chill postpublication changes of a substantive nature simply because the Court would prefer not to contact the parties and their counsel. But that possibility counsels in favor of notice because any such chilling might well be warranted. The Court should likewise make clear when it is has revised a slip opinion available on its website before the issuance of a preliminary print. At the very least, the slip opinion itself should include an incidental notation of the fact of a revision and the date of the revision, making clear that the version currently up on the website differs from a prior version that the Court had made available. Otherwise, as has occurred with EME Homer described above,458 there will invariably be competing electronic versions available on various websites and readers will be unable readily to discern from the slip opinion’s face whether the version they have is an earlier version that has since been corrected or the updated, corrected version. To be sure, a posting of change pages or the formal acknowledgment of the fact of a revised slip opinion will publicly acknowledge that the Court, too, makes mistakes. That should hardly come as a surprise given that the Court already says just that in the notice on slip opinions and preliminary prints. Nevertheless, public acknowledgment that the Court makes some mistakes not just in theory, but in reality will no doubt generate some initial headlines and uneasiness with the Court. So too, of course, may this Article by describing the Court’s past practices so fully. But those headlines will quickly dissipate and the advantages of such a posting moving forward are considerable. The posting should substantially reduce the likelihood of reliance on language that has been superseded. It may also prompt websites that reproduce slip opinions to change that practice or at the very least include revision postings for their readers. ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 458 See supra pp. 603–07. 622 HARVARD LAW REVIEW [Vol. 128:540 Even more important, such public postings will invariably prompt the Court to be more careful and systematic in deciding when and how to make revisions. Transparency might also encourage the Justices to think twice before venturing in their opinions into areas that are outside the record and beyond their own areas of expertise. Justices and their law clerks are exceedingly smart, but even they have their limits, and a few public corrections can be an effective reminder of the pitfalls of making unilateral assumptions about facts not grounded in the actual record.459 Publicity is not always a good thing, but in this context, it should and can be. Finally, in considering all of these reforms, the Court can fairly conclude that the same procedural rules and degree of transparency that apply to revising majority opinions need not apply to revising concurring and dissenting opinions, at least where the votes in support of the concurring opinion in no manner limit the precedential effect of the majority opinion. The “opinion of the Court” alone has force of law and precedential effect, so revisions to that opinion are of an entirely different character and import than revisions to opinions of individual Justices. The case for greater procedural safeguards and transparency is accordingly far more compelling. That is not to suggest that there should be no safeguards and no transparency applicable to revisions of concurring and dissenting opinions. The public interest in such opinions and their precise wording is very high. They play a major role in interpreting the Court’s opinion and in future discussions and litigation. At the very least, there should be public, after-the-fact notice of such changes. And, when a Justice is considering a correction to a concurring or dissenting opinion suggested by someone outside the Court, the Court may well change its current practices by providing prior notice to the parties (and the Solicitor General). As before, confidentiality concerns would not be implicated, and the rules against ex parte communication would strongly favor such notice and opportunity for input. CONCLUSION The Supreme Court’s traditions are a source of the institution’s great strength and durability. Such traditions provide a useful and visible reminder of the Court’s commitment to the institutional framework for lawmaking established by the Framers more than two centuries ago. They enable the Court to issue enormously controversial rulings while maintaining its essential stature and respect, not––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 459 See, e.g., Jacob Gershman, Alito’s Statistics Lesson Misses the Mark in Death-Penalty Dissent, Experts Say, WALL ST. J. L. BLOG (May 28, 2014, 2:43 PM), http://blogs.wsj.com/law/2014 /05/28/justice-alitos-statistics-lesson-misses-the-mark-in-recent-dissent-experts-say. 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 623 withstanding commentators’ frequent predictions that the Court has irreversibly squandered its authority. Unlike in the early days of Chief Justice Marshall,460 contemporary Justices dissent regularly and loudly, but without ever suggesting a lack of shared commitment to the Court’s institutional integrity. The persistent inability of the oftenstrident acrimony surrounding individual cases to erode the Court’s essential authority is a stunning institutional achievement. Yet, as important a role as tradition has played in the Court’s successes, the Court has also quietly and frequently changed its procedures, including in drafting and revising opinions. The Justices long ago abandoned the English tradition of issuing their opinions seriatim in favor of an “opinion of the Court.” The Justices no longer delegate to a Reporter, as they did to a potentially unsettling extent in the Court’s earliest years, the drafting of the written opinion or the decision whether a ruling is sufficiently significant to warrant publication. Since the mid-nineteenth century, the Justices themselves have been heavily involved in drafting written opinions, and the Reporter’s substantive role has diminished.461 The Court’s procedures for revising its opinions have likewise undergone changes. The Justices no longer routinely make extensive substantive changes after the opinion is first announced in Court, as they did during much of the nineteenth century. Nor, on the other hand, do the Justices insist on publishing every modification as a formal order in the United States Reports, as they often did during the 1930s. The Justices also consistently release the majority, concurring, and dissenting opinions at the same time. Happily, long gone are the days, underscored by the destructive institutional discordance in Dred Scott, in which Justices failed to provide each of their colleagues with draft opinions and fair opportunity to comment prior to formal release and publication. The Justices all adhere strictly to basic norms of institutional process regarding the drafting and release of their opinions. They do so no matter how deep and intense their disagreement about the results in a particular case. The Court nonetheless has retained practices regarding the revision of opinions that are long overdue for improvement. Unlike any other domestic court, the nation’s Highest Court formally issues its opinions before they have been fully reviewed and checked for errors within the ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 460 See Charles F. Hobson, Defining the Office: John Marshall as Chief Justice, 154 U. PA. L. REV. 1421, 1431–35 (2006). 461 The practice of having the Reporter draft summaries of the arguments of counsel seems to have dissipated during the 1940s. See Terminal R.R. Ass’n of St. Louis v. Bhd. of R.R. Trainmen, 318 U.S. 1, 1–2 (1943); United States v. Local 807 of Int’l Bhd. of Teamsters, 315 U.S. 521, 522–24 (1942); United States v. Ragen, 314 U.S. 513, 514–16 (1942). 624 HARVARD LAW REVIEW [Vol. 128:540 Court itself. The Justices do so apparently because of institutional pressures they feel to release opinions as quickly as possible, perhaps related to the desire to formalize an achieved majority before it might potentially change. But, whatever the reason, the Justices have long embraced the odd, but by-now settled practice of routinely revising their initial opinions and justifying their doing so by asserting, in a largely ignored notice, that only the United States Reports volume published many years later constitutes the “final,” “official” version of the Court’s opinions. What reforms are warranted? At the very least, the Court should end the fiction of labeling its initial, published opinions as neither “final” nor “official,” and any specific changes should be subject to afterthe-fact public notice. The latter reform in particular should mitigate electronic media’s tendency to perpetuate inaccurate versions of opinions. The mere fact of publication will also likely have a positive effect on the Court’s procedures for determining when and what revisions are necessary and appropriate. As described above, the Court should also consider a series of changes in its current practices that would provide greater structure and coherence to the revision process. Such reforms would extend to the possibility of notifying parties about certain categories of proposed changes prior to their adoption. However, precisely how those reforms should, in distinguishing between types of error, strike the balance between providing greater transparency and maintaining the essential confidentiality of the Court’s deliberative processes is less clear — at least to an academic on the outside looking in. It is plainly not costfree to invite overeager advocates to have another round of input. And those within the Court are far better equipped than anyone on the outside to determine how best to restrike that balance. Justice Jackson’s famous quip about the source of the Court’s “infallibility” is no doubt true. But the Court need not pretend to be infallible to maintain its stature and authority. The strength of the Court’s opinions is firmly rooted in the persuasiveness of its reasoning and, no less importantly, in the fairness and openness of its decisionmaking process. The Court can both make mistakes and admit mistakes without placing its institutional integrity at risk. And the Court need not compromise the confidentiality of the internal deliberative processes that is essential to candid debate, careful reasoning, and the exercise of sound judgment. The Court need only ensure that its process for acknowledging the need to make changes and for correcting mistakes is no less fair and open than that used in producing its opinions in the first instance. The current Court did not create the problems with its existing procedures, which find their origins in prac- 2014] (NON) FINALITY OF SUPREME COURT OPINIONS 625 tices that can be traced back to the days of the Court’s first great Chief Justice, John Marshall. But the Court today can easily fix them, and now should.462 ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 462 Indeed, just as this Article was going to final page proofs, there was a sign that the Court might be moving to a significantly more transparent practice of error correction, most likely in response to the significant attention that an earlier draft of this Article received in the national news media. See supra note 275. On October 22, 2014, the Supreme Court’s Public Information Office notified by email a large number of reporters who cover the Court for the national print and broadcast media that Justice Ginsburg had “amended her dissent” in a slip opinion released on October 18, 2014, because “[p]age 4 of the dissent contained an error” and that the Court had, accordingly, “updated” the opinion on the Court’s website. See Email from Kathleen L. Arberg, Pub. Info. Officer, Supreme Court of the United States, to Adam Liptak, Reporter, N.Y. Times, et al. (Oct. 22, 2014, 12:51 PM) (describing Veasey v. Perry, Nos. 14A393, 14A402, 14A404 (U.S. Oct. 18, 2014) (Ginsburg, J., dissenting)). The email notice identified the specific error that had been made and the precise language that was being deleted, and also added that the new version would include “small stylistic changes on pages 2 and 4.” Id. The Court’s announcement represents a welcome change from the Court’s historical practice and underscores this Article’s view that such candor about the need for corrections and that corrections are in fact made ultimately bolsters rather than undermines the Court’s stature and authority. It is, however, too soon to know whether Justice Ginsburg’s action represents a decision by the Court as a whole to disclose when such corrections are made or just Justice Ginsburg’s own new practice. Also unclear is whether further changes in the Court’s practices related to revising its opinions will be forthcoming, including others similarly recommended by this Article. In all events, the Court’s announcement well illustrates how the Court can so easily — in this instance, a simple email — increase its transparency.