Re: Self-Archiving Refereed Research vs. Self-Publishing Unrefereed Research

From: Arthur Smith <apsmith_at_APS.ORG>
Date: Fri, 17 Aug 2001 11:33:44 -0400

David Goodman wrote:
> The publication of material in an free archival system will permit much
> more open and effective review and comment than the present system does.

Permit perhaps - but will it actually happen? So much is published these
days that the vast majority is unlikely to ever receive a single
thoughtful comment unless specifically requested by a reviewing
authority of some sort. The Math Society set up a preprint system that
allowed commentary - and received at most a handful of comments on a
collection of several hundred papers. That experience has been very
typical - the journal "Electronic Transactions of Artificial
Intelligence" specifically set up such an open review system, with
commentary published along with articles, but still receives only a tiny
number of unsolicited comments from researchers.

> In the case you postulate, how long do you think it will take until the
> discrepancy is noticed and publicized? I'd say less than one day.

If the original article received little publicity, the discrepancy may
never be noticed, or will itself be unlikely to receive much publicity
unless a tragedy happens. The natural state of scientific research
papers is to receive essentially no publicity at all; they are
distributed and archived and read at random by people from a wide
spectrum of backgrounds, some very knowledgeable about the field and the
journals in question, some practitioners who may try to apply the
knowledge they gain, some simply members of the lay public.


> [...]
> We all could add analogous situations. How does the beginner learn
> these things?

Through coming to know a hierarchy of authoritative sources - for what
it's worth, the lay public relies on mass media at the top level, and
then more science-oriented popular books or publications (Scientific
American, space.com) or even the specific journals cited in the media
(Science and Nature, JAMA, NEJM, etc.) for greater understanding. Aside
from general reputation, the criteria for a source to be seen in this
way is (1) responsibility for the content it publishes, and (2)
presentation in a useful context (e.g. great articles have to be much
easier to find than the needle in the haystack).

A raw database for which nobody but the author is responsible supplies
little of this potential for authority.

A raw database with a thin layer of criticism for which only the critics
are responsible for their comments is hardly better (is there any
working example of such an open system? There are the "slashdot" or
"kuro5hin" models, but that means thinking about peer review in a
completely new light too - and those news/discussion sources have some
pretty substantial defects).

> They learn from the criticism in the literature (and from education). Open
> archives would facilitate this, not hinder it.

Criticism (and education) is but another form of communication, for
which the same issues of authority and responsibility arise. Who do you
read first? Who do you trust?

The point is we have now a system, likely more expensive than it needs
to be, but on the other hand pretty cheap in the whole scheme of things,
that supplies these layers of authority to the literature, through
hierarchies of journals, media, and other sources. We could work on
fixing this existing system (and most conscientious publishers are
working hard to be more timely, more fully electronic, and provide more
services such as interlinking), or we could focus attention on a great
idea (such as author self-archiving) that may not solve any of the real
problems in communication of scientific and scholarly information.

                Arthur (apsmith_at_aps.org)
Received on Wed Jan 03 2001 - 19:17:43 GMT