WASHINGTON — The Supreme Court has been quietly revising its decisions years after they were issued, altering the law of the land without public notice. The revisions include “truly substantive changes in factual statements and legal reasoning,” said Richard J. Lazarus, a law professor at Harvard and the author of a new studyexamining the phenomenon.
The court can act quickly, as when Justice Antonin Scalia last month corrected an embarrassing error in a dissent in a case involving the Environmental Protection Agency.
But most changes are neither prompt nor publicized, and the court’s secretive editing process has led judges and law professors astray, causing them to rely on passages that were later scrubbed from the official record. The widening public access to online versions of the court’s decisions, some of which do not reflect the final wording, has made the longstanding problem more pronounced.
Unannounced changes have not reversed decisions outright, but they have withdrawn conclusions on significant points of law. They have also retreated from descriptions of common ground with other justices, as Justice Sandra Day O’Connor did in a major gay rights case.
The larger point, said Jeffrey L. Fisher, a law professor at Stanford, is that Supreme Court decisions are parsed by judges and scholars with exceptional care. “In Supreme Court opinions, every word matters,” he said. “When they’re changing the wording of opinions, they’re basically rewriting the law.”
Supreme Court opinions are often produced under intense time pressure because of the court’s self-imposed deadline, which generally calls for the announcement of decisions in all cases argued during the term before the justices leave for their summer break. In this term, 29 of the 70 cases argued since October remain to be decided in the next five weeks or so.
The court does warn readers that early versions of its decisions, available at the courthouse and on the court’s website, are works in progress. A small-print notice says that “this opinion is subject to formal revision before publication,” and it asks readers to notify the court of “any typographical or other formal errors.”
But aside from announcing the abstract proposition that revisions are possible, the court almost never notes when a change has been made, much less specifies what it was. And many changes do not seem merely typographical or formal.
Four legal publishers are granted access to “change pages” that show all revisions. Those documents are not made public, and the court refused to provide copies to The New York Times.
The final and authoritative versions of decisions, some published five years after they were announced, do not, moreover, always fully supplant the original ones. Otherwise reliable Internet resources and even the court’s own website at times still post older versions.
The only way the public can identify most changes is by painstaking comparison of early versions of decisions to ones published years later.
But there have been recent exceptions. Last month, Justice Scalia made a misstep in a dissent in a case involving the E.P.A. Under the heading “Plus Ça Change: E.P.A.’s Continuing Quest for Cost-Benefit Authority,” he criticized the agency for seeking such authority in a 2001 case. But he got its position backward. Worse, he was the author of the majority opinion in the 2001 decision.
Law professors pointed out the mistake, and Justice Scalia quickly altered his opinion, revising the text and substituting a bland heading: “Our Precedent.”
The court did not draw attention to the changes, but they did not go unnoticed. Other revisions have. A sentence in a 2003 concurrencefrom Justice O’Connor in a gay rights decision, Lawrence v. Texas, has been deleted from the official record. She had said Justice Scalia “apparently agrees” that a Texas law making gay sex a crime could not be reconciled with the court’s equal protection principles.
Lower court judges debated the statement, and law professors used it in teaching the case. The statement continues to appear in Internet archives like Findlaw and Cornell Law School’s Legal Information Institute.
But it has vanished from the official version published in 2006 and from the one available on Lexis, a legal database.
“They deliberately make it hard for anyone to determine when changes are made, although they could easily make that information public,” Professor Lazarus wrote in the study, which will be published in The Harvard Law Review.
In revisions to two 2009 opinions, onschool searches and race-conscious hiring, Justice Ruth Bader Ginsburg added phrases to clarify and broaden the points she had made. The changes appear in Lexis, but the court’s website still features the original versions.
The court also corrects factual errors, including, in recent years, ones about who was president in 1799, which senator made a particular statement and whether a defendant was convicted or merely indicted.
After-the-fact editing is not a new phenomenon. “The current court did not begin this practice, which finds its origins in the court’s earliest days and has extended to all justices over the years, liberal and conservative, but the court today can take the steps to correct it,” Professor Lazarus said. “Easy to do, and long overdue.”
The court seems to have been even more freewheeling in the past. Chief Justice Roger B. Taney added approximately 18 pages to his 1857 majority opinion in the Dred Scott decision after it was announced.
There are indications in former justices’ papers that the court knows that its editing practices are open to question.
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,” a court official wrote to Chief Justice Warren E. Burger in 1984, “we actually operate a system that is completely at odds with general publishing practices.”
In an internal memorandum in 1981, Justice Harry A. Blackmun offered reasons that the court operated “on a strange and ‘reverse’ basis, where the professional editing is done after initial public release.” Once an opinion has garnered the five votes needed to have it speak for the court, he said, the author wants to issue it immediately to guard against defections and “get ‘on the scoreboard.’ ”
There are four generations of opinions, and only the last is said to be final. So-called bench opinions, in booklet form, are available at the court when decisions are announced. Slip opinions are posted on the court’s website soon after. They are followed by preliminary softcover prints and then by the only official versions, which are published in hardcover volumes called United States Reports. The official versions of opinions from 2008 were published in 2013.
There are two exceptions to the general practice of quietly slipping changes into opinions. One happens only after the decisions are published in final form. The hardcover books sometimes contain a page of “errata.”
The court also issues an occasional order formally revising an opinion. The most recent notable example was in 2008, when the court learned that it had banned capital punishment for child rapists partly based on the faulty premise that no federal law allowed such executions. In denying a motion for rehearing, the court issued an order revising parts of the original decision to reflect the correct information.
But most changes can be found only by careful comparison or in the “change pages” that the court does not make public. Professor Lazarus obtained a year’s worth of the pages but was denied access to more. He said the court should consider posting them on its website.
“Of course the justices make mistakes,” he said. “And, of course, they can correct them. They just need to use a process that is more in keeping with the integrity and rigor of the process that produces the opinions in the first instance.”