From: Stevan Harnad <>
Date: Mon, 11 Sep 2000 12:04:45 +0100

On Mon, 11 Sep 2000, Chris Armstrong wrote:

> Steve Hitchcock may have answered this to some extent:
> "a good eprint archive is an authority because it
> provides a service that is clearly labelled."

The Eprints software -- -- provides
this clear labelling:

(1) It is fully complaint with the Santa Fe convention for meta-data
tagging and interoperability. Hence every institution that mounts the
(free) Eprints software can immediately register as an Open Archive
-- --
interoperable with all the other registered Open Archives. The contents
of all these distributed Open Archives can and will all be harvested
into one seamless, global "virtual" Archive.

(2) Anyone searching the literature can restrict their search to these
Eprints Archives, thereby ensuring that they only access
Santa-Fe-compliant data.

(3) Among the (actual and potential) metadata fields for the Eprints
Archives are the following:

   (b) (actual) JOURNAL NAME (etc.)
   (c) (potential) UNOFFICIAL/OFFICIAL (author-supplied/publisher-supplied)

> My worries
> have to do with uninformed or naive use - something which
> may be impossible or unlikely if items ARE labelled and
> if the label is sufficiently informative.

If you can think of further pertinent labels, can be
configured to include them too. It is upgradable to keep pace
with evolving metadata-tagging practices.

>sh>trivial semantics of "publishing" ... The typescript Q
>sh>sent to me above, is a "publication," by this general
>sh>(and irrelevant) legal definition.
> NO it is not - publication means making available or
> disseminating to a mass not to one individual (Publish:
> to make generally known; publisher: one who distributes
> to the public). You are confusing the issues with these
> statements.

I am not a lawyer, and nothing of any substance rides on this, but I
believe you are factually wrong. ONE COPY is sufficient to qualify as a
publication, by the legal definition -- a text I scrawl on a piece of
paper, for example, and give to you, is legally a "publication" (and
can, for example, be used to assert copyright and intellectual

I repeat: This technical definition is of no interest here. What is of
interest is not what the lawyers consider publication, but what the the
research community, funding agencies, and academic review committees
consider publication, and that is publication in a refereed journal,
with a publisher, whose imprimatur is the official "certification" of
the quality-level of that paper.

There is also Vanity Press (unrefereed) publication, of course,
including self-publication on the Web, but as this form
of publication is not officially quality-controlled, the question of
whether or not a draft is "official" becomes rather less consequential.

However, even for Vanity-Press/self-publication, it is possible to
date-stamp and authenticate versions (as Adrian Pickering has indicated
in this Forum). So even here, there is no principled problem.

Perhaps the root of the misunderstanding here is that you are confusing
self-archiving with self-publication. It is certainly possible to
self-archive an unrefereed preprint and leave it at that; and that would
indeed constitute self-publication. But that is not the primary purpose
of the Eprint Archives under discussion here.

Self-Publication is a SUBSTITUTE for conventional, refereed-journal
publication, and that is not a model being advocated or particularly
discussed here. What is being discussed here is self-archiving (not
self-publication) as a SUPPLEMENT to refereed-journal publication,
for the purpose of freeing the published refereed (and not
yet-published, pre-refereeing) research literature.

Being merely a supplement to and not a substitute for conventional
refereed-journal publication, self-archiving should not be regarded
as an alternative form of publication; it is and should be regarded
as an alternative means of access to refereed publication (including
to unofficial copies as well as pre-refereeing copies). Those versions
can be tagged and authenticated as minutely as we wish, but existing
practise already suggests that such heroics are not necessary -- and
they are certainly nothing to defer self-archiving for, not even for a

>sh>For if there IS an official, published version, and it
>sh>resides in the pages of journal Q, then it seems that
>sh>the web versions are very much like author-supplied
> Not at all. The difference lies in their authority in
> being published by the "official" publisher on the web
> site of that publisher.

I think we have now repeated our respective positions sufficiently
often. You set more store by the "official" stamp than the research
community does (for the huge portion of the refereed literature to
which their institution does not give them access): Having an unofficial
version of everything in hand for free now is preferable to waiting for the
off-chance that the official versions can be made free someday -- or
that the unofficial versions can be made official.

>sh>The sign-posts can be made as clear as we like.
>sh>(Self-archived drafts could even carry a health warning,
>sh>if people thought it necessary!)
> So now we come to the crux of the matter: I DO,
> EMPHATICALLY think it is necessary that they carry a
> health warning (your phrase): a label stating clearly
> what the paper is. The label should be a part of _each_
> paper (much like a copyright statement or a title page)
> and should have standard form and content.

Fine. As I said, the format and labelling of the Eprints Archives is
configurable. If the rest of the research community feels that they
would like such a health-warining on every item that is not from the
publishers proprietary archives, it can be provided.

(Now can we move on from these a-priori worries to more substantive

> Only this will give true LIBERATION of access; without
> labelling the liberation may entrap the unwary! I know
> physicists have been doing this for some years (archiving
> not entrapping, sorry!). My concern is that should your
> approach be adopted wholesale, some archives may not be
> so well administered and may not make their function and
> status clear to naive users.

As free official publishers' versions will initially be the exceptions
rather than the rule in Eprints Archives, I would make the more
parsimonious proposal that when a browser is set to search Eprints
Archives (only), it is the (rare) official publishers' versions that
should carry the tag OFFICIAL; and the default assumption should be
unofficial. That will save a lot of screen-space. I think the
research community is capable of keeping that in mind, while profiting
from the windfall of a newly freed research literature.

The important thing to remember is that these are not issues of
substance, and should not detain us from mounting Eprint Archives
and self-archiving, so as to free the refereed literature, now.

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 this ongoing discussion of providing free
access to the refereed journal literature is available at the American
Scientist September Forum (98 & 99 & 00):

You may join the list at the site above.

Discussion can be posted to:
Received on Mon Jan 24 2000 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:51 GMT