Perfectionism

"Academics who delivered their promised manuscripts twenty years late used to cause us amusement; but it was a respectful amusement, because we knew the delay to be due, not to idleness, but to perfectionism. Perfectionism can be obsessive, like that which prevented Wittgenstein from publishing another book in his lifetime, and probably would have done so however long he had lived; but, as the phrase goes, it is a fault on the right side. Every learned book, every learned article, adds to the weight of things for others to read, and thereby reduces the chance of their reading other books or articles. Its publication is therefore not automatically justified by its having some merit: the merit must be great enough to outweigh the disservice done by its being published at all. Naturally, no individual writer can be expected to be able accurately to weigh the one against the other; but he should be conscious of the existence of such a pair of scales. We used to be trained to believe that no one should put anything into print until he no longer sees how to make it any better. That, I still believe, is the criterion we should apply; it is the only means that exists of keeping the quality of published work as high as possible, and its quantity manageably low. The ideologues who in their arrogance force their misconceived ideals upon us attempt to make us apply virtually the opposite criterion: publish the moment you can get editor or publisher to accept it. We are compelled outwardly to comply with their demands; let us inwardly continue to maintain our own values." (Michael A. E. Dummett, Frege: Philosophy of Mathematics)

On the flip side, it is hard to read or cite a book not published. There are a number of cases where books waiting for perfection have got cited as "unpublished work", which is pretty limiting to those not in the inner circle that get to see such work. Or a graduate student gets told that the hypothesis she's been working on disproving has already been proved by Smith; maybe it will get published in 20 years, errors still intact, or someone will point out the errors to Smith in 10 years and in another decade Smith will published and maybe then the rumors that Smith proved that hypothesis will start to go away. Or maybe Smith was right, and could have saved that grad student a couple years of work.

If one spends decades publishing, that's decades nobody can respond to that research. In practice, instead the deeply inferior citation to unpublished work has occurred and will occur. Rapidly progressing subjects will be unlearnable without access to a friendly scholar in the subject.

I would think this would be especially bad for mathematics, as so much circles around the proof, and until we do computer-checked proofs, the best way to get a good proof is write lots of proofs and feed them to peer-review until you can write a proof that passes peer-review, and then write proofs to pass peer-review. Without that, it's hard to write bulletproof proofs; it's too easy to skip over the fine details and miss something critical.

It would interesting were it possible to disengage the tenure-track focus on publishing, from the utility of publishing as outlined in >2 prosfilaes:.

I understand the use of publishing as an indicator of professional competence, but believe (at least in many disciplines) it's too indirect and loose to be persuasive. Instead, there is a push to publish not for reasons of the greater academic conversation, but for the narrower incentive to further one's career. And this pressure on publishing only messes up the enterprise of publication, further diluting its effectiveness.

It's this context, I think, that is the appropriate one for understanding the Dummett quote. That perfectionism is a real concern, only more concerning in the context in which scientific progress is diluted by publishing for reasons other than that progress.