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

From: Stevan Harnad <>
Date: Fri, 10 Aug 2001 17:14:12 +0100

> On Fri, 10 Aug 2001 Arthur P. Smith <> wrote:
> sh> We are talking about authors self-archiving their final, refereed
> sh> draft, the one that has been peer-reviewed, revised, finalized and
> sh> accepted.
> Sez who? Take for example:
> (our friend Podkletnov of the anti-gravity device)
> What is to prevent this fellow from adding a journal-ref to some
> obscure supposedly peer-reviewed journal subscribed to by only a
> handful of institutions (such as Physica C: Superconductivity...)?
> Podkletnov could even insert a journal-ref to Physical Review, or
> Reviews of Modern Physics, or Applied Physics Letters, or some such; I
> am not aware of any substantial verification mechanism that is
> currently in place for that data; perhaps something will come along for
> the major journals, but that still leaves the obscure ones...

So what? Let us check all the cases to see whether there is any one of
them for which the "solution" is not patently obvious (as long as one
has no special interest vested in there being some sort of substantive
problem or obstacle inherent in the strategy of freeing the refereed
literature by self-archiving it publicly on-line):

    (1) Podkletnov actually has an article published in a low-level
    journal. It reports a cure for cancer (in reality bogus),
    involving drugs that are in reality toxic and do not cure cancer.

This is a serious problem in every instance -- whether the article
appears only on-paper, "filtered" by the access-costs, or it is
self-archived publicly on-line, free for all. This is the case of a
bogus result that has slipped through peer review in a low-level

But it has nothing to do with self-archiving! And the "solution" is the
same with or without self-archiving: The journal's name and quality
matters. Beware of low-level journals.

    (2) Podkletnov actually has an unrefereed, unpublished article but
    falsely self-archives it as having been published in a low-level
    journal. It reports a cure for cancer (in reality bogus), involving
    drugs that are in reality toxic and do not cure cancer.

Caveat emptor applies doubly: (a) be careful about self-archived articles:
they may not have appeared in the journal indicated; (b) beware of
low-level journals.

An obvious cross-check for (a) is to check the reference in another
index (e.g. a free one like PubMed) to confirm the authorship.

    (3) Podkletnov actually has an article published in a high-level
    journal. It reports a cure for cancer (in reality bogus), involving
    drugs that are in reality toxic and do not cure cancer.

This is a still more serious problem. Peer review has failed here at a
high level. But this again has nothing whatsoever to do with whether or
not the article is publicly self-archived.

    (4) Podkletnov actually has an unrefereed, unpublished article but
    falsely self-archives it as having been published in a high-level
    journal. It reports a cure for cancer (in reality bogus), involving
    drugs that are in reality toxic and do not cure cancer.

Same as (2) above.

I don't think these obvious "solutions" require much intelligence or
imagination. It is an empirical question what the proportion of
Podkletnov-like papers will be as more and more of the refereed
literature is self-archived. (Can anyone give us an estimate of their
current proportion in ArXiv?) It is also an empirical question how many
users will NOT have the sense to do the obvious cross-checks and will
hence fall afoul of the caveats mentioned. I expect the same sorts of
worries could be adduced against printing itself, or perhaps the
written word (in favor of the oral tradition): All systems can be
abused or misused (including peer-reviewed publication!).

But there is nothing whatsoever here to hold back the bona fide
researcher, even for one millisecond, from going ahead and
self-archiving anyway. (You can be sure the bogus ones will all do it
in any case!)

One advantage of institution-based self-archiving is their greater
answerability, relative to public central archives: The Eprints
software creates archives that can be moderated for content, and even
restricted to institutional logins, if need be.

In short, I think this is all a red herring.

> And even
> if he actually DOES have a published paper on this subject (as he does
> in Physica C) it is up to the author to ensure that there is no
> substantive difference between the two versions.

Same as above, but less shrilly.

> Hence "responsibility"
> and "context" are issues - whose responsibility is the integrity of the
> content and metadata (the author's or the journal's), and what context
> is the article presented in (a big database with little tags to
> indicate publication status, or a journal ToC)?

It can be made as clear as we like that one is accessing only an
institutional or central archive for author self-archived eprints
rather than an official journal's website (if for no other reason than
that the latter requires institutional S/L/P for access!).

Self-archived eprints can be designed to carry "health warnings" that
are as shrill as we like a priori (or, more sensibly, a posteriori,
once we get an idea of the size of the bogus paper problem -- if there
is any).

This is also so obvious that it is hard to see why anyone would want to
raise it as any kind of objection to or worry about going ahead and
self-archiving all refereed research, immediately!

> Now that's not to say we wouldn't publish papers like this - we do on
> occasion. But very rarely do they make it through the highest-level
> filters in our journals (Phys Rev Letters, or the Rapid Communications
> sections of the others) and then only through a deliberate editorial
> decision to allow a controversial group of author to present their
> case.

Fine, but utterly irrelevant to the question at hand, because it
applies equally to self-archived and non-self-archived papers in these
high-level journals.

> Lundberg is clearly worried about even worse cases in the medical
> fields - recent news about the scurrilous things drug companies will do
> with reports of scientific research are enough to tread very carefully
> in this area.

Could you elaborate on this?

But remember we are talking about the refereed journal literature, and
your burden is to show how the refereed-journal's name FAILS to perform
its usual filtration and sign-posting function for self-archived
(refereed) papers. What scurrilous things can drug companies do with
the for-free versions of refereed papers that they could not already do
with the for-fee versions? (Ditto for unrefereed papers!)

> Of course it really just means that engineers can't rely on the ArXiv
> for finding what's going on in physics; they will more likely go first
> to secondary services that filter on a group of relevant and
> highly-regarded journals (such as INSPEC in physics, or Medline in
> bio-medicine). And so that group of readers bypasses the author
> self-archive altogether. Which is probably fine.

Then they can go back to the Archives and set their journal-name tag
for those journals alone, and they will retrieve the full text for free
(whereas from the secondary services they will get at most only the

Or, to put it even more transparently, they can access the "official"
full-texts for those journals that their institutions can afford to
subscribe to, and then, as a SUPPLEMENT (not a SUBSTITUTE), access the
slightly less reliable "unofficial" versions self-archived by their
authors, for those journals that their institutions cannot afford to
subscribe to.

Sorry Arthur, I was unable to discern anything but red herrings in all

Stevan Harnad
Professor of Cognitive Science
Department of Electronics and phone: +44 23-80 592-582
             Computer Science fax: +44 23-80 592-865
University of Southampton
Highfield, Southampton

NOTE: A complete archive of the ongoing discussion of providing free
access to the refereed journal literature online is available at the
American Scientist September Forum (98 & 99 & 00 & 01):

You may join the list at the site above.

Discussion can be posted to:
Received on Wed Jan 03 2001 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:12 GMT