With Great Power Comes Great Responsibility

My post on trackbackgate below drew a plaintive comment from Matti Pitkanen who is, he says, blacklisted from the ArXiv. I won't pretend to have any opinion on the merits of his work or of his particular case, but it is clear that the existence of the ArXiv has radically changed the landscape of scientific publication. If, as I expect, it gradually supercedes all paper journals, as it largely already has in some fields of physics, those who control the ArXiv will have more or less exclusive control of the dissemination of new results.

This is a radical change from the system of fifteen years ago, when a fringe scientist rejected by mainstream journals could usually find some journal that would publish his results. The history of science shows that although new ideas often arise from the scientific consensus, it is far from rare for them to originate on the fringes, e.g., Wegner, Boltzmann, Margulies, Alvarez.

Trackbackgate illustrates how a few individuals with authority arising mainly from accidental circumstances can make decisions in an essentially unaccountable fashion, without any need to explain or even define their criteria, and that these decisions potentially affect the whole future of science.

I would prefer a system that was truly open, with perhaps three levels of papers: submitted papers, which would be completely open, endorsed papers (similar to the present system), and refereed papers. It would still be necessary to have some method for rejecting commercial and other spam, but I would envision the submitted papers system open to anyone willing to go to the trouble of writing one, no matter how crazy. In any case, it seems imperative that all decisions be made based on objective and published criteria.

It's not clear how this type of system can introduce the kind of market mechanisms and competition which worked after a fashion for paper journals.

Comments

Popular posts from this blog

Anti-Libertarian: re-post

Uneasy Lies The Head

We Call it Soccer