Here's a small supplement to Perry Dane's very interesting discussion of the jurisprudential significance of the fact that today in the United States law clerks have substantial input, including the development of specific arguments and doctrinal formulations, into the opinions that are ultimately published under "their" judge's name (and the follow-up comments here and here). A short and therefore distorted version of Dane's argument is that the appearance of the judge's name on an opinoin acts as an "icon" or signal about something -- on my reading, about the fact that the opinion should be understood to be part of a larger project of developing "the judge's" vision of the law that is consistent across large-ish domains and over large-ish periods of time (the "ish's" inserted to show that a person's views and understanding of the relevant domains can change without casting doubt on the proposition that the person is pursuing a project). But, Dane worries, how much confidence can we have that the words published under the judge's name are part of such a project when they are produced by an ever-changing poplation of law clerks?

I draw on my experience as a law clerk to Justice Thurgood Marshall in 1972-73 for this supplemental comment. The law clerks did substantially all of the opinion-drafting in the chambers, with the judge providing relatively modest pre-drafting guidance and equally modest post-drafting review. Two practices within the chambers seem to me relevant to Dane's argument. First, and less important, because the clerks knew that post-drafting review would be light, we were extremely careful not to send a draft to the judge before it had been carefully vetted by all three of us.

Second, and more important here, we operated with a strong norm of what (as I recall) we called "personal stare decisis." That is, we thought that we should draft opinions that were as consistent as we could make them with the judge's prior decisions. (Note that this was Marshall's fifth year on the Court, and he had a small-ish relevant body of decisions from his time on the Second Circuit.) Clerks following the norm of personal stare decisis would make the opinions published under the judge's name part of the Dane-like project.

One anecdote that we built into our understanding (recounted as accurately as I can, which might not be all that accurate): As I recall, until 1971 the judge had joined and "written" opinons (which reflected his views as transmitted to his law clerks) that all new decisions in matters of constitutional criminal procedure should be given complete retroactive effect. In the 1970-71 Term, the judge had a law clerk who thought that Justice Harlan had a better view of retroactivity -- full retroactivity for all cases pending on direct review, and no retroactivity for all other cases, especially those on habeas review. A concurring opinion published under Marshall's name in Williams v. United States endorsed Harlan's view.

In 1972 we were given to understand that this had been an abuse of the law clerk's role. Although the opinion as published began, "After studying afresh the pattern of the Court's retroactivity decisions," and therefore formally satisfied the "consistent project" requirement, the fact that we law clerks knew that the statement was inaccurate as an account of the judge's mental processes meant that the drafting law clerk had misbehaved.

The example is one of departure from the norm of personal stare decisis. But the more important point is that, as far as I know, such departures were extremely rare. (I don't recall any that occurred during the remainder of the judge's tenure with respect to opinions we drafted in 1972-73.) The praatice of personal stare decisions, I think, fits well with Dane's overall account (and he might have mentioned it in a part of this article that I overlooked). [rushing off, so no time to proofread this -- aplogies for errors.]