May I live so long without my mistakes being noticed

New York Times reporter Cornelia Dean writes about a unique retraction by 84-year-old chemist Homer Jacobson:

Nobody paid much attention to the paper at the time, he said in a telephone interview from his home in Tarrytown, N.Y. But today it is winning Dr. Jacobson acclaim that he does not want - from creationists who cite it as proof that life could not have emerged on earth without divine intervention.
So after 52 years, he has retracted it.

Wow, can you imagine how much time it would save us if people would retract old papers more often? I mean, there are a lot of scientists who are open and honest when new results overturn their old papers, and that is always refreshing. But even then, it can be hard to figure out which earlier results were shown to be definitely wrong, and which were just forgotten. A good retraction, linked to the original, would make everything much clearer.

Still, most papers don't deserve retracting. Most papers are full of good, logical ideas, even if some of the empirical results were later shown to be wrong.

Tracing the history of these old ideas is a very fruitful source of new ideas. I was telling a student just today that reading the literature supporting old, discredited ideas is much more intellectually demanding than reading about the new, trendy ideas, and often leads to more new thoughts.

I'm not sure that's universally true, but it certainly seems to be the case in paleoanthropology. The modern human origins problem emerged, more or less in its current form, more than 60 years ago. Old papers in this area sometimes seem quaint, and of course we have many more empirical points now in support of various arguments, but the basic themes have hardly changed.

These themes were established 20 years before any substantial molecular information about human evolution, purely on the basis of skeletal comparisons. Better, if you strip away much of the recent empirical evidence, the basic claims are often clearer and more succinct in old publications.

So, I think retractions should be limited to cases where genuine errors were made -- the point and power of a retraction is to prevent a real mistake from propagating itself. A retraction may help stop some high-profile misinterpretations, but it is almost never worth it.

In Jacobson's case, he identified what he considers to be genuine errors in the paper; errors that tend to support the misinterpretations he now rues. So, I suppose a retraction is appropriate. But it will hardly stop misinterpretation.

In paleoanthropology, misinterpretation usually rules the day!