Re: Leading academics back UK Research Councils on self-archiving

From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
Date: Mon, 5 Sep 2005 16:43:09 +0100

On Mon, 5 Sep 2005, Sally Morris (ALPSP) wrote:

> As I rather suspected, most people (with the exception of Stevan) think
> that 'the problem' is access - that's why most presentations about OA (both
> kinds) are usually prefaced by the standard ARL slide about the rate of
> increase in journal prices and library budgets respectively.

Access and impact are two sides of exactly the same coin: There can be
no impact without access (pace Don Sannella, who points out that some
"scholars" cite what they have not read).
ttp://openaccess.eprints.org/index.php?/archives/20-Journal-Publishing-and-Author-Self-Archiving-Peaceful-Co-Existence-and-Fruitful-Collaboration.html#c6

The reason to focus on impact, though, is that the only ones who can
*provide* open access are authors: first, by being willing to give away
their texts in the first place, rather than seeking royalties or fees
from their sales, as most sane authors do; second, by either publishing
them in an OA journal (if a suitable one exists) or by self-archiving
their final author's drafts in an OA IR.

So whereas the research community will be eternally in the debt of the
library community for having helped to raise the hue and cry about the
access problem, its solution can only come from the access-provider
end, hence must address the needs and interests of the access-provider;
it cannot come from the access-user end, even though the access-user
is of course a co-beneficiary of OA, and invariably the author wearing
another hat: the user of the access provided by other authors. There is
a Golden Rule inherent in all that, but it is not sufficient.

    "Self-Archive Unto Others as Ye Would Have Them Self-Archive Unto You"
    http://cogprints.org/3022/

What is missing is the direct appeal to impact, not access. Access
is implicit in impact: The way I enhance the impact of my research is
by making it accessible to as many of its would-be users as possible:
optimally, *all* of them. Access is a necessary (though not a sufficient)
condition for impact, but impact is a sufficient condition for access: a
*cited* article is necessarily (pace Don Sannella's fine point about
not-so-fine scholarship) an accessed *article.*

This is why it is the 50-250% impact gap that is the relevant measure
(for the researcher, the researcher's institution, and the research's
funder, and the tax-paying public that has invested in the research by
funding the funder) of the access gap -- not libraries' serials crises,
nor publishers' journal prices.

    http://opcit.eprints.org/oacitation-biblio.html

> But if the problem is access, and we reach a situation where a large
> proportion of the content of some journals (even, perhaps, if it's in a
> less-than-final version) is freely and easily available, what responsible
> cash-strapped librarian would not choose to cancel those journals in order
> to retain the rest? I just don't see the logic in doing otherwise...

The point you are missing, Sally, is that the problem is impact for
authors and access for users, but the user's access problem is on an
article by article basis, varying from user to user, as a function of
the user's field and whatever journals his institution can afford. (This
is the only link with journal affordability, and it is a very
distributed one.) *If* an article happens to be published in a journal
that that particular user's institution cannot afford, then that user
will have to use the author's self-archived OA draft; otherwise, the user
can and will use the publisher's value-added version.

There is no basis for whole-journal cancellation in what I have just
described. It is an anarchic author/article/user-based question of who
uses which version of which article in which journal. There are no
implications there for whole-journals and their potential cancelation.
The role of the OA version is, as has been noted over and over, a
supplement, not a substitute -- and a supplement anarchically
distributed worldwide, depending on the article, journal, field, and
institution of the author and the user.

And that is without even reminding Sally that she also needs to add to
her logical equation the fact that there is (a) still a robust demand
for the print edition of the journal and (b) that, if, as she rightly
stated, the publisher's online edition has many values added, those who
can afford those added values will want to keep paying for them.

So in second-guessing the logic of the "cash-strapped librarian," one
should, in the service of seemliness, not portray the problem as that
of continuing to bleed stones dry, but the more upbeat one that journals
will continue to compete for librarians' limited available cash, and they
should compete for it on the value-added end, not on the access-denial
end. Journal budgets are and always have been finite, and well below the
bounds of being able to afford most or even many out of the total number of
journals on offer; it's always been about which among the small affordable
fraction to keep or to swap, year by year, seriatim.

So it continues now. Self-archiving does not advantage or disadvantage
any particular journal in this regard. Even less so does a mere UK
self-archiving mandate.

But if and when the demand for the print edition should dry up, and if
and when the online added values should no longer prove to be enough to
sustain demand -- and all this may well be a long way off yet -- then
there will definitely be ways to cover the costs of the irreducible and
indispensable value that journals (with the free help of peers) always
have and always will add, namely, peer review.

http://www.publications.parliament.uk/pa/cm200304/cmselect/cmsctech/399/399we152.htm

That contingency, however, has shown no sign of being imminent or even
likely, even in the (few) fields where OA self-archiving has effectively
been at 100% for a number of years now.

What is certain is that 100% self-archiving is highly beneficial --
indeed optimal -- for research *now*, and that it is well within reach
and already well overdue. To put it very baldly, the publishing community
has absolutely no justification for asking the research community to
defer maximising its research impact -- closing the 50-250% access gap --
for one moment longer. The interests of research are not to be denied in
favor of the anxieties of publishers about the future course of research
journal publishing in the online age. Evolution will prevail here,
researchers will adapt, and publishers will adapt. There is a logic to
that, and it is futile (and perhaps also less than admirable) to try
to delay or divert it. Research publishing is done in the service of
research, not vice versa.

> The 'straws in the wind' that we reported in our letter to RCUK (which,
> like STM's, was supposed to be private - we mistakenly made it free on our
> site for about a day)

Why mistakenly? Are arguments not to be answered? The research community's
recommendations and critiques have all been made in the open, ready
to meet criticism head-on: why not the publishing community's?

And here is the publishing community asking the RCUK (privately, but for
the grace of the godhead) to *delay* things still further, for still
more debate (now going on for at least a half-dozen years, with every
substantive and trivial point already aired many, many times over, on
both sides) -- whilst deeming it a *mistake* to have aired their views
and grounds openly!

> are strong and worrying indications that this might happen;

Repeating and raising the volume on the adjectives and gerundives does
not make them truer the Nth time round. All the *indications* to date are
of peaceful co-existence between journal publishing and self-archiving:
no cancellations, no revenue loss. What is strong and getting stronger
is merely the *worrying*, on the part of publishers. It is not that
there are zero *grounds* for worry (there are always grounds for worry):
there is merely zero *evidence* for worry, and some not insubstantial
evidence for the contrary.

> loss of usage for those physics journals most of whose content can
> also be found in ArXiv would be worrying not because of citations, but
> because clued-up librarians would rapidly work out that they don't need to
> subscribe, since the journal site is not where the content is actually
> being read.

*Rapidly work out*? How rapidly? since some areas of physics have been
at 100% self-archiving for a number of years now! Do you really think
the "cash-strapped" librarians are that slow-witted, if they've a mind
to cancel on these grounds?

But it's not libraians, in any case, who have the last word on
cancellations. It is institutional researchers. And physicists have
no wish to cancel their journals, or journals in general. They use
the prepublication preprint first, and because the preprints tend to
be in the same central archive (Arxiv) as the refereed final drafts
(postprints), they tend to use that same source even after publication
too, even when their own institution has licensed online access to the
journal version. (I have no actually stats on this, but it seems evident
from what is known.)

Nevertheless, if you ask these physicists, as either authors or users,
whether the journals should be canceled, they will say no -- and they are
right! As authors, they want the peer-review and certification for their
work; and as users, they want to be sure there is an authoritative
version-of-record to cite -- and to check in any cases of doubt about
the self-archived version. They all quite naturally see the self-archived
corpus, even if it is the one accessed most, as a supplement to the
official journal version, not a substitute for it.

Not to mention that it still does many of their hearts good to know that
there's a print version of the journal sitting on their institution's
book-shelves somewhere...

(I personally believe that the days of central archives, like Arxiv, are
numbered. OAI-interoperability and distributed OAI-compliant IRs make far
more sense, are far more likely to generate the missing self-archived
content (85%) and can easily be integrated with search and harvesting
services that direct a user toward his own institutionally licensed
official version rather than the poor man's supplement in those cases
where it exists [even google scholar could do it]: This is just a simple
and natural extension of the thinking behind reflink, OpenURL etc. Takes
no imagination, but requires a much bigger self-archived database in
order to make it worthwhile implementing. Download stats can be pooled
and credited to the official journal version this way too, and citations
consolidated and credited. It's all pretty obvious, once you fill that
85% content gap...)

> I have yet to hear any confirmation from physics publishers
> that the anecdotal 'evidence' reported by Alma Swan (the only shred of such
> 'evidence', as far as I can tell), suggesting that ArXiv is not hurting
> journals, is actually correct or the full story. I hope they will tell us?

Even better, if the "evidence" -- of zero correlation between cancelations
and self-archiving (controlling for the other factors affecting baseline
cancellations across time) is merely "anecdotal," it would be very
useful to see whether there actually exists qunatitative evidence of a
nonzero correlation between cancelations and self-archiving (controlling
for the other factors affecting baseline cancellations across time) --
for that is what Sally needs to sustain her (otherwise already refuted)
hypothesis about what the "cash-strapped librarians" will/would do --
"rapidly"....

> National licences would help to solve the problem - consider, for example,
> the National Research Council in Canada which has made all its journals
> freely available throughout Canada. Someone has to pay, of course (I think
> most people have now grasped that journal publishing, in whatever form,
> costs money) - in that case, the Canadian Government.

An NRC National license is a splendid solution -- for the NRC journals,
and their Canadian usership.

What about all the non-NRC journals? And what about the non-Canadian
usership of the NRC journals? What about all the rest of that lost
impact and access?

No, Sally, global licensing does not generalise to a solution of the
problem of the 50-250% impact gap -- and (n.b.) it would not do so even
if all journals were sold at-cost (zero profit). The world is simply too
"cash-strapped" to be able to afford to pay, in advance, everywhere, for
any product a vendor might offer, not even in the case of commodities for
which there is a basic inelastic need, such as food, water, medicine,
shelter.

    "Journal Publisher Click-Through Monopoly: A Trojan Horse"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0436.html
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0210.html
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0238.html
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci01/0447.html
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1507.html
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1525.html

The solution is far more realistic and direct: The research community
must plug up the 50-250% gap, now, instead of continuing to bleed
needless weekly, monthly, yearly research impact. The publishing
community will then adapt, as dictated by the actual course of events
and evidence (not pre-emptive worries). No global oligopolies, no
delays, no filibustering. To repeat: research publishing is done in the
service of research, not vice versa.

Stevan Harnad
Received on Mon Sep 05 2005 - 17:06:01 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:59 GMT