Self-Archiving the Refereed Journal Literature

From: Stevan Harnad <harnad_at_COGSCI.SOTON.AC.UK>
Date: Sun, 18 Apr 1999 10:42:40 -0400

On Mon, 12 Apr 1999, Donald Forsdyke wrote:

> Letter posted on Bionet.journals.note
> Dr. Harnad,
> Thank you for this great proposal for bringing electronic
> publication to the biological sciences. Could you briefly tell us what
> steps are being taken to ensure that material, once deposited, is not
> interfered with. I would love, for example, to publish an article
> discovering the structure of DNA which I would date 1952, the year
> before Watson and Cricks' famous paper.
> The only way to cover this point would seem to be to have
> simultaneous deposition in multiple sites (>2), and have a search engine
> of some kind constantly monitoring the sites to ensure that all copies
> remain identical. This is something like GenBank,where depositions of
> DNA sequences are made simultaneously in the USA, Europe and Japan.

Dr. Forsdyke,

Your concern is legitimate, but the situation is in some ways more
complicated and in some ways simpler than you seem to be envisioning.

I have described the overall pattern in my writings
<>, so I will only give the
pertinent gist here:

Don't confuse eprint archives with journals at this point. Eventually
they will coalesce, but right now it is critical to distinguish them.

Self-archiving by authors in free, public Eprint Archives (like Los
Alamos, and CogPrints <>) is NOT meant as a
substitute for journals: Journals provide peer review and
certification. Archives merely provide access, and in this case, access
to texts that their depositors "certify" are their own; and, where
appropriate, authors also "certify" that they have appeared, verbatim,
in the journal they indicate.

That is all that Eprint Archives need to provide at this time. It is
journals themselves that continue to provide the peer review, and that
certify the final, accepted draft as published. Hence, for now, the only
"authentic" draft is the paper (or online) one provided by the publisher
(the online version can be encrypted, finger-printed,
password-protected, you name it).

Do not mix up these two functions at this time. The user who draws upon
the Eprint literature does it with a knowing (but not very risky)
caveat: The text retrieved from the free archive might be corrupted (if
the depositor has been careless) or it might even be false (if the depositor
has been dishonest, say, falsely claiming that it has been published in
journal X, volume Y). To be sure of authorship and authenticity, one
would still have to check the "authenticated" version, which is not
free, and may not be online -- but for most purposes this will simply be

The effect of the existence of this self-archived literature, however,
will be to draw most of the useage to it, as it has already done in
Physics, where this is PRECISELY the state of affairs today
<>. The world physics community uses the Archive,
not the journals, although the journals are still there, providing the
"silent hand" of peer review that protects and certifies the quality of
the literature <>.

What will happen next is that the de facto use of the free archive will
force publishers to scale down to providing only the essential service
they provide -- quality control and certification thereof -- and the
cost of this (less than 1/3 of the cost per page in the present system
of subscription/site-license/pay-per-view -- S/L/P) will be paid up-front, by
authors, not out of their pockets, of course, but funded by 1/3 of the
3/3 savings from the cancellation of all their institution's S/L/P
costs). This will make the CERTIFIED literature free too, and then it
will be official journal overlays on the public archives (with all the
protections you want: distributedness, "finger-printing," encryption,
redundancy, etc.) that provide the authentication. (This scenario was
the gist of my "subversive proposal"

You will neither see nor arrive at this if you think directly in terms
of having Archives do this a priori on their own at this point in

The question of how to cite an unrefereed, unpublished preprint is
trivial and already answered, and in practise in LANL: See

Youngen, G. (1998) Citation patterns of the physics preprint literature
with special emphasis on the preprints available electronically. UIUC
Physics and Astronomy library [online] (c. 5 Nov. 1998)

The question of how to cite a published reprint is also trivial: Same
way you do now.

How to cite both, when there is both a paid paper incarnation and a
free publicly archived incarnation? Obviously by combining both
formats: journal citation and archive access number.

What will happen when Archives coalesce with journal overlays, as
above? The hybrid citation will be simplified (no more need for page
spans or page spans, of course, numbered paragraphs do much better;
probably no need for volume numbers either, year is enough, and issues
will vanish, replaced by single items, appearing at the moment the
final version is accepted), and hence it will look like exactly the
logical combination of the old and new citation parameters that any
sensible child would have predicted. Nothing to ruminate about here at all.


Stevan Harnad
Professor of Cognitive Science
Department of Electronics and phone: +44 1703 592-582
Computer Science fax: +44 1703 592-865
University of Southampton
Highfield, Southampton

> There is also the question of how to cite this
> information. Rather than some abstract number, why not follow the
> pattern of joural citations. e.g. Harnad, S. Cogprints 1999, 4:10-1610.
> In this case the "volume" number (4) would be the month and the "page"
> numbers would be the day and hour-min (preferably corrected to GMT).
> Thus the citation would contain an implicit time-stamp.

> Sincerely, Donald Forsdyke. Discussion Leader. Bionet.journals.note
Received on Wed Feb 10 1999 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:29 GMT