What scientific journals do

Michael Clarke over at Scholarly Kitchen has an interesting post today Why Hasn't Scientific Publishing Been Disrupted Already?. As he points out, the World Wide Web was invented to help scientists communicate. Shelly and I recognized this the instant we heard about WWW in late 1993 - pretty early in web history - and submitted a proposal to Physical Review Letters at the time to use web technology to put the journals online. So, all that has in fact happened, but without, for the most part, much disruption of the organizations and companies that do scientific publishing. Is it because those organizations are very smart and nimble and have been able to keep up with the times? I'd like to think so, but Michael has some good points about specific features in journal publishing that may have made it harder than first recognized to displace the existing system. Perhaps not particularly new points, but nicely organized.

I added a comment (still in moderation at this writing) that there was one more issue he seemed to have left out:

Michael – wonderful article. On your list of things scientific journals do, there’s one more that perhaps seems obvious, but that really is slightly peculiar in the context of other types of publication: uniqueness. This is complementary to, but distinct from, your criterion of Registration. No reputable scientific publication re-prints articles that have been published elsewhere: each article is supposed to be original and unique. The condensed output of a particular scientific episode, experiment, theoretical effort, story or whatever you want to call it, appears in only one, unique place, and is expected to be always cited by that fixed representation. There are some variations on this theme by type of article (a letter, a regular article, and a review paper may cover the same research) and certainly some authors publish multiple articles that are not terribly far apart in their content, but the extraordinary sanctions journals place on authors who try to violate it (submitting work whose full substance they’ve already published elsewhere) suggests an importance beyond copyright or the foibles of editors.

In contrast a journalistic article, review or opinion essay printed in one newspaper or magazine may with little fear of reprisal be sold to countless others as well; syndicators and press agencies exist to spread good writing around without bothering to re-write. Short stories find themselves reprinted in many disparate collections. Books often come out in many editions with large portions intact from one version to the next.

So why must the scientific article be unique in its published journal-based citable representation? I think it’s a good thing, but I’d like to better understand why.