Re: Self-Archiving vs. Self-Publishing

From: Stevan Harnad <>
Date: Tue, 9 Dec 2003 14:35:16 +0000

On Mon, 8 Dec 2003, John Unsworth wrote:

>sh> (1) The open-access movement is about journal-articles, not books.
> Why isn't that an arbitrary distinction?

Because journal-articles are author give-aways and books are not.
Book authors seek royalties, article authors seek only research
impact. Hence toll-barriers are acceptable to (most) book
authors but are unacceptable and counterproductive to (all)
journal-article authors.

    "On not conflating the give-away and non-give-away literature"

    "1.1. Distinguish the non-give-away literature from the give-away literature"

    Harnad, S., Varian, H. & Parks, R. (2000) Academic
    publishing in the online era: What Will Be For-Fee And
    What Will Be For-Free? Culture Machine 2 (Online Journal)

>sh> Peer-reviewed journals that happen to be published by
>sh> university presses (e.g., Behavioral and Brain Sciences
>sh> [published by Cambridge University
>sh> Press], which I edited for 25 years) are simply journals (and mostly
>sh> toll-access journals!). Nothing special to do with open access. And
>sh> especially not with open-access provision for that university's
>sh> own journal-article output, with which it is orthogonal.
> I thought the topic was peer-review, and whether university presses
> should or could supervise it. They do, in some cases, and they clearly
> can, and it's not self-review. If the topic is, instead, tolls for
> access, then sure, journals published by university presses (or books,
> for that matter) are just like other journals (or books).

Several topics were being conflated:

(1) Saving money for the university by lowering journal access-toll
expenditures (journal cancellations) or by helping to lower journal

(2) Making money for the university through electronic publishing ventures
(journal and/or book)

(3) University publishing of new low-toll-access journals

(4) University publishing of new open-access journals (or taking over
and converting existing toll-access journal-titles to open access)

(5) Maximizing the visibility and impact of the university's own journal
article output (by either (5a) publishing articles in open-access journals
or (5b) self-archiving articles published in toll-access journals, before and
after refereeing)

I think universities need to carefully separate -- and keep separate
-- these five topics, separating especially those that concern saving
money or generating direct revenue for the university (1), (2), (3),
from those that concern enhancing the university's research impact
(which is a special, indirect source of revenue) (4), (5a), (5b). The
open-access movement is dedicated to the latter (4, 5a and 5b).

All universities want to reduce their journal expenditures (1). That is
the serials crisis. It was the serials crisis that first alerted us to the
access/impact problem (5) but its relation to that problem and its solution
is complicated and indirect. There are both high-toll and low-toll journals
published by University Presses and Learned Societies. So it is not at all
clear that having University Presses take over journal publication from
the existing journal publishers would result in lower journal prices. Nor
is it realistic to imagine any substantial number of such take-overs or
title-migrations (2,3). The same is true for new journal start-ups.

And in any case, lower-toll journals, even if published by the
universities, are not a solution to universities' research access/impact
problems (for their own research output). Not even a university publishing
an open-access journal or two (4) is a solution for that university's
research access/impact problem (for its own research output).

Only (5) addresses this problem, but here the university is in the
position to have its greatest influence, both on the impact of its own
research output (5b) and, through reciprocity, on its access to the
research output of other universities (if they do likewise). Here lies
the more complicated causal link between 5 and 1: For once open access
prevails through 5, savings are likely for 1.

> I'm not arguing for new journals in the CIC summit paper -- maybe new
> publishers, but not new journals.

The prospects for title-migration and journal take-overs are
minimal right now. After open-access provision has been mandated
by universities, on the other hand, the landscape may well change:

> Societies who own their journals could move them from publishers who
> are less likely to behave in ways friendly to the system of scholarly
> communication, to publishers who are more likely to do so. Societies
> that don't own their journals can abandon them and move to journals that
> are more friendly.

Society-owned and society-based journals are, first, a minority
among journals, hence even in principle only amenable to a minority
solution. Migration of society-owned journals to lower-toll publishers (3)
is certainly a useful strategy, and might eventually help (1). This was
the old SPARC strategy (consortial support for lower-toll journals). So
far, though, this has nothing to do with open access.

Migrating a society-owned journal to an open-access publisher is also
an option, though a somewhat more risky one at this time, when the
open-access cost-recovery model is still new, untested, and, most
important, trying to survive in what is still a 97.5% toll-access
environment, with vast amounts of university money still tied up in the
toll-access budgets, forcing universities to cover open-access journal
costs for their own research output from other sources. The problem is
minor while there are negligibly few open-access journals, but it may
not remain minor if more such costs need to be paid -- as long as the
resources on which it would be most natural to draw to pay them are
still locked up in paying existing annual journal-tolls.

This is why it is so important to prepare the ground for open-access
journals through institutional open-access provision (self-archiving).
That immediately maximizes the access to and the impact of a university's
own research output, while paving the way for their eventual windfall
savings from toll-cancellations that will free up the funds to pay for
open-access journal costs for their own research output.

And if voluntary migration at this time to the open-access cost-recovery
model by those learned societies that own their journal-titles is risky,
a more drastic step by those learned societies that do not own their
titles is even more risky.

Open-access provision for each university's own research output seems
a much more sensible and promising step.

> The Association for Computers and the Humanities,
> of which I'm currently president, falls (unfortunately) into the latter
> category, and we are in the process of doing exactly what I'm suggesting:
> we've severed ties with Computers and the Humanities, our society journal
> since the '80s, but owned and published by Kluwer, and we're consolidating
> our publishing activity with the Association for Literary and Linguistic
> Computing, which owns their own journal, which is published by Oxford UP
> (which, as you know, is engaged in an open-access experiment with the
> Oxford library services).

I applaud the step, but I don't think it scales up to all, most, or even many of
the 24,000 peer-reviewed journals that exist today. And they, and today, are the

> No new journals in that picture, but a shift from commercial to
> university press publishing.

A laudable minority solution (if OUP's open-access experiment succeeds)
but not a solution that scales to the majority at this time. And it is
this time that we must deal with. And open access provision is already
feasible today, indeed long overdue. We need only do it, instead of
continuing to grasp only at straws and long-shots!

>sh> And again, it has nothing to do with providing open access to their *own*
>sh> university research output (hence should not be conflated with it).
> I'm not conflating it--but you keep doing so, which makes me wonder why.

Vide supra: 1-5.

>sh> But what is the point, then? Is it about creating university-based
>sh> lower-toll-access journals, as competitors for the existing ones? (Nothing
>sh> to do with open access then, and both a risky and an a circuitous route
>sh> to trying to cut university costs or increase university revenue.)
> It's puzzling to me that you are so determined to keep university
> presses (and libraries) out of the picture.

I'm not determined to keep them out of the picture. I am determined
to promote the already proven, tried-and-true shortest/fastest/surest
path to open access at a time when -- although we are at last beginning
to grasp that open access is indeed possible and desirable -- we still seem
to be going for the lower-probability solutions only: Let us pursue
those lower-probability solutions *also,* by all means, but not *only*!
Our efforts should be invested in proportion to the odds,
and the odds (empirically, practically, and logically) are at
this time all tilted c. 95/5 in favor of the "green" strategy
of open-access provision through institutional self-archiving
of all research articles, relative to the "golden" strategy
of creating/converting and publishing in open-access journals:

>sh> I still don't understand. What does it mean to peer-review a paper and
>sh> not publish it (if it passes peer review)? Who/what peer-reviewed it,
>sh> and what for?
> The phrase you're worrying here ('conduct peer review independent of
> a decision to publish') is quoted, in the CIC essay, from another talk
> given at ACLS. In that talk, I was suggesting that the problem university
> presses have--of very small print runs--might be mitigated by allowing
> freely available scholarship to find its audience first, and then be
> selected for print on the basis of the size of the audience. At the
> risk of incurring a blizzard of citations, I'd argue that my argument on
> that point is essentially the same as yours on the subject of "esoteric
> publishing."

Although I continue to hold that this entire problem is so simple and
low-dimensional as to be trivial (and hence that it continues to be a
real head-shaker as to why so many, many people still don't "get it"),
I have to confess that I myself did not have the picture nearly well
enough in focus when I invoked the "esoteric" criterion (in the
early '90's)!

What you say above, though, is absolutely correct for certain books
(esoteric monographs, actually), and I don't disagree at all. I am sure
that such books will fall under the same rubric as that under which every
single one of the 2,500,000 annual articles in the 24,000 peer-reviewed
journals falls.

But the relevant distinction -- which was not yet in focus in my "esoteric"
days but is crystal-clear now -- is between non-giveaway writings written
for royalty revenue and give-away writings written only for research impact.

The esoteric/exoteric distinction captured this more fundamental
distinction only approximately and fuzzily. The logical refutation is that
even the highest citation-impact journal article (with a readership far
greater than a good deal of the royalty-based literature) is a give-away,
written only for research impact and not royalty revenue, even though
it is hardly esoteric!

    Harnad, S. (1995) Universal FTP Archives for Esoteric Science and
    Scholarship: A Subversive Proposal. In: Ann Okerson & James O'Donnell
    (Eds.) Scholarly Journals at the Crossroads; A Subversive Proposal
    for Electronic Publishing. Washington, DC., Association of Research
    Libraries, June 1995.


    Harnad, S., Varian, H. & Parks, R. (2000) Academic
    publishing in the online era: What Will Be For-Fee And
    What Will Be For-Free? Culture Machine 2 (Online Journal)

What you say above, however, about separating peer-review from
"publication" for esoteric monographs definitely does not apply or scale
to the annual 2,500,000 articles appearing in the 24,000 peer-reviewed
journals for which generic pre-vetting is not a viable option: The moment
a journal has certified them as having met its peer-review quality
standards, they are *published* (insofar as posterity and promotion-panels
are concerned)! The rest is just about providing access to them.

    Peer Review Reform Hypothesis-Testing

    A Note of Caution About "Reforming the System"

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

    The Invisible Hand of Peer Review.

>sh> What needs to be separated is (1) peer-review provision and
>sh> certification from (2) access-provision and archiving, *not*
>sh> peer-review from publication.
>sh> (1) What is meant by academic "publish or perish" policy is "publish in
>sh> a peer reviewed journal" (let us leave aside books in this discussion).
> Well, in the humanities we don't leave aside books in that context, but
> we can bracket them, if you wish.

The book/journal distinction is critical here, because it so closely
matches the non-giveaway/giveaway distinction, which is absolutely
fundamental to open access. But I agree that "publish or perish" in
many disciplines includes -- or even primarily means -- publishing
books. To the extent that those books are *not* esoteric monographs,
likewise published only for research impact and not for royalty revenue,
open access may not be the solution for them.

>sh> (5) It follows (with considerable force, both logically and empirically)
>sh> that in the online era, the (PostGutenberg) journal reduces to the sole
>sh> remaining essential component of academic publish-or-perish publication,
>sh> which is peer-review service-provision and certification.
> >
>sh> (6) Access-provision will eventually be off-loaded completely onto
>sh> the network of institutional OAI archives in which the authors have
>sh> self-archived the eprints of their peer-reviewed publications (sic).
>sh> (But NB: Today, the OAI archives are merely an open-access supplement,
>sh> and cover only a small fraction of annual refereed research output.)
> >
>sh> Hence it is peer-review/certification service-provision (=
>sh> publication) that is unbundled from archiving/access-provision in the
>sh> online/open-access age, and *not* peer-review that is unbundled from
>sh> publication! Peer review is the invariant essential component of
>sh> academic publish-or-perish publication.
> Well, I understand the distinction you're making, but I think it is
> perverse to use the term "publication" for a concept that explicitly
> excludes distribution. My guess is that doing so has led to much
> confusion.

Yes, the PostGutenberg era requires revising our notion of
(peer-reviewed research) "publication":

     Used to mean: "peer-reviewed, accordingly imprimatur-certified,
     printed, and disseminated by the publisher."

     Now means: "peer-reviewed and accordingly imprimatur-certified
     by the publisher"

but with access provided in a variety of optional ways, among them
on-paper printing/dissemination by the publisher, online dissemination
by the publisher, and online archiving and dissemination by the author's

        Garfield: "Acknowledged Self-Archiving is Not Prior Publication"

A short way to put it is "Publisher" does not equal "Printer."

>sh> Who provides peer review for papers (we are not speaking of grant
>sh> proposals), other than the journals that publish them?
> The societies or editorial boards who run them. And those are not
> generally the same business entity as the publisher that's producing
> the journal.

The peer-review service-provider and certifier (in the online era)
will be seen as the publisher, and the rest will be seen as merely
access-provision. And in this case the publication has the Society's
imprimatur, and not the printer's. (Sounds like these "publishers"
are just printers! But surely printing is not publishing!)

> I am more inclined than you are to think that an
> open-access system might include university-based publishers.

It might, and will, but for now, the proportions among
university/society/commercial publishers of the existing 24,000
peer-reviewed journals are not likely to change significantly. *After*
open access prevails, perhaps, but not before. Hence it is open-access
provision (by universities, to their own journal-article output), as
soon as possible, that is needed, not dreams of university take-overs
of existing journals or start-ups of new ones.

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 Open Access Forum (98 & 99 & 00 & 01 & 02 & 03):
    Post discussion to:

Dual Open-Access-Provision Policy:
    BOAI-2 ("gold"): Publish your article in a suitable open-access
            journal whenever one exists.
    BOAI-1 ("green"): Otherwise, publish your article in a suitable
            toll-access journal and also self-archive it.
Received on Tue Dec 09 2003 - 14:35:16 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:11 GMT