Re: Self-Selected Vetting vs. Peer Review: Supplement or Substitute?

From: Stevan Harnad <>
Date: Sun, 13 Jul 2003 21:39:30 +0100

[Identity deleted] wrote:

> I cite a couple of your papers in an overview section about the broader
> and deeper crises in academic publishing and I would like to hear more
> of your thoughts on the possibilities (or failure) to automate
> moderation and some of the refereeing process.

That 1997 cardhouse paper is rather dated, given the pace of
developments in the subsequent half-decade. See the latest ones in:

> the separation of controls in the dissemination of esoteric literature.

I don't know what you mean by "esoteric literature." (I meant the
full fleet of 20K peer-reviewed journals when I used the term, since
dropped), nor what the separation of controls means. Peer review just
means qualified experts testing, improving and certifying the quality
of reported research. (Everything gets published. The only question is:
at what quality level in the journal hierarchy? Then the user can adjust
his usage accordingly. And referees are scarce, they don't self-select,
and they need to be answerable, as the author is, to a qualified,
experienced, neutral, and answerable mediator: the editor.)

> knowledge communities have implemented the growing communications and
> database management technologies in order to develop instant, global and
> free-of-charge repositories of research results with extended citation
> services, data harvesting and search functions, of unprecedented
> capacity.

But how are researchers to know whether these results are worth the time
of reading and the risk of using? That's what the journal's peer-review
level and track-record tags.

> This has caused a crisis in the traditional approach to the
> dissemination of scientific and other scholarly literature.

There is no crisis, just a powerful new possibility (open access) that
has not yet been exploited (except in a few areas).

> I will introduce and examine some of the key arguments set
> forth by pioneers and advocates of unrefereed electronic and networked
> resources.

I am certainly not one of these, as I advocate self-archiving of both
pre- and post-refereeing research, with an emphasis on the latter. As it
happens, if you look at what they *do* (as opposed to what they *say*
they do), this is also exactly also what all the users of the Physics
ArXiv are doing, and have been doing since the beginning. There is not,
in physics, any change whatsoever insofar as peer reviewed publication is
concerned. The change is purely in the *access,* pre- and post-peer

> I will focus mainly on a dissemination system introduced first to
> physicists and later to mathematicians and other related fields. I will
> show the separation of controls in the development of this system.

There has been no separation of controls whatsoever. Everything is exactly
as before. It is just that both preprints and postprints are now openly
accessible online.

> One of the supposed disadvantages of an on-line preprint exchange
> as compared to formally refereed publications is its open-to-all access,
> risking to flood such repositories with erroneous results or irrelevant
> content.

There is not (and never was) such a preprint repository. There is merely
a much wider way to disseminate both pre- and post-refereeing papers.
Everything else is exactly the same as it was before. (Everyone knows
-- and alway knew -- that with unrefereed preprints the rule is "caveat
emptor", whether they are online or on-paper, toll-access or free.)

> Some object that the archives will become hosts for poor science and
> readers will be misled,

Users are perfectly capable of judging for themselves whether and when
they want to risk their time and efforts on pre-refereeing preprints
(names and reputations are a cue), but peer review remains in place
everywhere, including in Physics, and within about 8-12 months the
refereed postprint is available, as always.

> Some object that a flood of irrelevant articles will slow scientists
> down in finding what is relevant to their current work.

There has been no evidence of such a flood: just the usual preprints
(now openly accessible), submitted to journals for peer review, then
the usual postprints, 8-12 months later.

> I will argue that such objections are of no importance
> because expert readers are self-policing the quality
> of a scientific claim and the problem is really just one
> of classification rather than detailed refereeing.

Readers can "self-police" now, with the "invisible hand" of peer review
keeping everyone honest, and the postprint only 8-12 months away. But that
would not be the case if what some people are *supposing* is the case now
(but is not) *were* indeed the case, namely, that the journals vanished,
papers were no longer submitted to peer review, and no postprint were
around the corner. If/when that ever happened, human nature being what it
is, quality standards would decline and decline, and the literature would
become less and less usable -- until we simply re-invented peer review.
Peer-review is merely pre-vetting by appointed, qualified, answerable
experts. It is not an anarchic, ad-lib self-policing system.

But nothing of the sort is happening now. Peer review proceeds apace and
preprints and postprints are merely being made openly accessible.

> Finally, if it is the case that an archive of unrefereed scientific
> preprints can operate without presenting an problem, why do scientists
> still publish nearly all their output through the conventional refereed
> channels?

Because this is not "unrefereed" material but *pre-refereeing* material,
all written with the expectation of being answerable to peer review. You
are talking about a nonexistent hypothetical literature, not what is in
the archives today. Authors are self-archiving to get around
the tolls blocking the access to and hence the impact of their work;
they are not self-archiving to get around peer review.

> Journals continue to exist only because formal publication is still a
> necessary gatekeeper, to assure outsiders, who have to make decisions
> about appointments, promotions and funding, that a piece of work has
> significance.

This is simply a repetition of the conjectures (in my opinion, incorrect
ones) of those self-archiving physicists and mathematicians who *say*
that this is why everyone (everyone!) is still doing exactly the same
thing they have always done.

But it is more prudent to disregard what Simon *says* he is doing and to
look instead at what Simon actually *does*. This turns out to be:
preprint, journal submission, postprint -- exactly as before, but both now
openly accessible.

And the peer review is not primarily for the "outsiders" but for the
would-be *users*, to ensure the quality of the work, and minimize the
risk of wasting time and resources on untested, unreliable work.

> Conventional journal publications have a classificatory rather
> than an informational role for these outsiders. Whether this analysis
> would apply to the less esoteric sciences is a task for further
> exploration.

I don't know what "esoteric sciences" are yet, but peer review
serves peer-users primarily and research-evaluators and -funders only

Stevan Harnad

NOTE: A complete archive of the ongoing discussion of providing open
access to the peer-reviewed research literature online is available at
the American Scientist September Forum (98 & 99 & 00 & 01 & 02 & 03):

Discussion can be posted to:
Received on Sun Jul 13 2003 - 21:39:30 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:59 GMT