Re: Self-Archiving vs. Self-Publishing

From: Stevan Harnad <>
Date: Mon, 8 Dec 2003 14:32:20 +0000

Hi John,

I am answering in some detail because it looks as if there are some points
on which our differences are substantive, though perhaps exchanging a
little more information may resolve them. I'd like permission to post
the discussion (having taken the time to reply in detail); otherwise I'd
like to post just my replies, compressing and anonymizing your portions,
if you prefer. These issues are too important -- and the questions arise
too often -- to just discuss offline.

>U> "the things that university presses [can] contribute... [include]
>U> administering an online authoring and peer-review environment that
>U> encourages authors to produce content in forms that lower library
>U> costs for collection and preservation"
>h> I know you just mean software for reducing the cost of peer review,
>h> hence publication. But I know from experience that this will be
>h> misconstrued as universities peer-reviewing their own output.
>U> University presses already routinely run peer review processes for
>U> books--but since those books are rarely written by faculty at the same
>U> university that houses the press, I don't think it counts as
>U> self-review. In any case, presses have been doing peer review for a
>U> long, long time.


(1) The open-access movement is about journal-articles, not about books.

(2) Peer-reviewed journals that happen to be published by university
presses (e.g., Behavioral and Brain Sciences
[published by Cambridge University Press], which I edited for 25 years)
are simply journals (and mostly toll-access journals!). Nothing special
to do with open access. And especially not with open-access provision
for that particular university's journal-article output, with which it
is orthogonal.

(3) New journals -- whether toll-access or open-access, and whether
published by university presses, learned societies, nonprofit publishers,
or for-profit publishers -- are merely new journals: competitors to
existing journals. The online era and its economies have opened up a
few new journal niches, but not many. University Presses (and others)
can try to fill these niches, but again, this has nothing to do with
open access.

(4) The open-access journal cost-recovery model is still under trial. It
is not yet clear whether it will prove viable and stable. (I for,
one, believe it is the optimal and inevitable cost-recovery model --
but only *after* open access has prevailed through self-archiving,
which will at the same time force journal cost-cutting and downsizing *and*
generate the university windfall toll-savings out of which to pay the
open-access journal publication costs.)

So it's just as risky a business for university presses to launch new
open-access journals as it is for all other publishers. And again, it
has nothing to do with providing open access to their *own* university
research output (hence should not be conflated with it).

>h> That interpretation needs to be blocked. There is also slight crossing
>h> here of another pair of wires: *All* publishers need to lower (and
>h> most of them are lowering) peer-review costs with online processing
>h> software. Singling out university presses here again sounds like
>h> take-over plans for what are currently independent journals.
> No, I mean that university presses should compete for some of the
> business currently owned by commercial publishers, and universities
> should look favorably on them, if they compete by lowering costs for
> the university. Of course, if that encourages commercial publishers to
> compete on the same terms, all to the good.

I couldn't follow: How does *my* university launching a new open access
journal (in these still-uncertain times for the open-access journal
cost-recovery model) lower costs for my university (or provide open
access for my university's research output -- or even reduce library
serial costs)?

>h> The number of new-journal niches is small (this has not changed), so
>h> the prospect of achieving open access solely by creating new, competing
>h> open-access journals (PLOS- and BMC-style) and trying to win away the
>h> authors of the established toll-access journals (23,400 of them!) is
>h> extremely uncertain and certainly extremely slow.
> I'm more optimistic, actually--but the point I'm making doesn't rest on
> creating new open-access journals.

But what is the point, then? Is it about creating university-based
lower-toll-access journals, as competitors for the existing ones?
(Nothing to do with open access then, and both a risky and a circuitous
route to trying to cut university costs or increase university revenue!)

>h> The self-archiving
>h> road needs to be clearly differentiated from the open-access journal
>h> road. Subsuming them both under a university function only encourages
>h> their conflation (while again evoking the spectre of vanity press
>h> publication).
> Fine--I understand.

It's a bit subtler than that, because I am afraid that there continue to
be lots of indications in this exchange that that conflation continues to
be made, implicitly!

>U> "we should conduct peer review independent of a decision to
>U> publish."
>h> Who are "we" (the university? the author?).
>U> journals, societies, presses.

I still don't understand. What does it mean to peer-review a paper
and not publish it (if it passes peer review)? Who/what peer-reviewed
it, and what for?

>h> And here is an important one: What needs to be separated is (1)
>h> peer-review provision and certification from (2) access-provision and
>h> archiving, *not* peer-review from publication.
> That's splitting a hair, seems to me. Isn't peer-review provision =
> peer review, and access-provision = publication?

No, and this is actually a central, substantive point that couldn't be
further from hair-splitting:

(1) What is meant by academic "publish or perish" policy is "publish in
a peer reviewed journal" (let us leave aside books in this discussion).

(2) Hence neither writing a research paper and putting it in a desk-drawer
nor publishing it with a vanity press counts as "publication" on an
academic CV, and before an academic promotion committee.

(3) In the on-paper (Gutenberg) era, authors provided (paper) offprints
of their peer-reviewed publications [articles] (by mail) to would-be
users who asked.

(4) In today's online (PostGutenberg) era, authors can provide limitless
"eprints" of their peer-reviewed publications to all would-be users by
self-archiving them online (in their own institutional OAI archives).

(5) It follows (with considerable force, both logically and empirically)
that in the online era, the (PostGutenberg) journal reduces to the sole
remaining essential component of academic publish-or-perish publication,
which is: peer-review service-provision and certification.

(6) Access-provision will eventually be off-loaded completely onto
the network of institutional OAI archives in which the authors have
self-archived the eprints of their peer-reviewed publications (sic).
(But NB: Today, the OAI archives are merely an open-access supplement,
and cover only a small fraction of annual refereed research output.)

Hence it is peer-review/certification service-provision (equals:
publication) that is unbundled from archiving/access-provision in the
online/open-access age, and *not* peer-review that is unbundled from
publication! Peer review is the invariant essential component of
academic publish-or-perish publication.

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

>h> Hence journals are and will remain (autonomous) peer-review
>h> providers/certifiers, and to be certified as having successfully met
>h> a particular established set of peer-review quality standards *is*
>h> to be published.
> Well, that's silly, isn't it? If I peer review something and then
> stick it in a drawer, have I published it?

Not only would it be silly to stick a peer-reviewed publication in a
desk-drawer without making it openly accessible by self-archiving the
eprint, but it would be silly for any "entity" to provide a peer-review
service for desk-drawer papers! So it is not a coincidence that no such
silly service or practice exists: Journals peer review articles; they
are also still providing access to them, on paper and online, for the
time being. But now that the author can provide open access, the
access-provision by the journal is no longer an *essential* function of
publish-or-perish publishing. But the peer-review/certification still is.
Indeed, that is exactly what publish-or-perish (journal) publishing in
the open-access era amounts to!

>h> Hence there is no dissociating publishing from peer review!
> Well, sure there is.

Could you please map that out for me? Who provides peer review for
papers (we are not speaking of grant proposals), other than the
journals that publish them? And why on earth would they provide peer
review for any other purpose, or in any other sense?

What *can* be dissociated form peer-review/publication is

See these AmSci threads:

    "Distinguishing the Essentials from the Optional Add-Ons"

    "The True Cost of the Essentials (Implementing Peer Review)"

    "Re: The True Cost of the Essentials (Implementing Peer Review -

    "Re: Scientific publishing is not just about administering

>h> There is only dissociation of *access-provision* from peer
>h> review.
> Sophistry, or else muddle.

With all due respect, it is you who appear not to have thought this one

>h> And even that cannot be done now, when 23,400 journals are
>h> toll-based access-providers. All an author can do is *supplement* the
>h> toll access to his own work by self-archiving it, thereby making it
>h> open-access. This too is not a case of making peer review independent
>h> of a decision to publish, but merely (merely!) making access-provision
>h> independent of peer-reviewed publication!
>U> "peer-reviewed information would be freely available *soon after*
>U> its creation"
> >
>h> This "soon after" also occurs repeatedly in your proposal, and I
>h> would suggest removing it, because it is quite easily misconstrued
>h> as not being about open-access at all, but only the journal's
>h> willingness to allow free access after an interval (from 6 months to
>h> 2 years or more) has elapsed after which toll-revenues become
>h> negligible anyway.
>U> Nope, I just mean that there shouldn't be a great lag time in between
>U> creation and publication. I'll stand by that as a desideratum for an
>U> ideal system of scholarly communication.

We're all for minimizing publication lag (both the time taken to
perform the peer review and the time taken for the journal to
go to press). But that has nothing to do with open access, which is
about when *access* should be provided.

Access to the unrefereed preprint is the author's business (though many
have found self-archiving preprints to be useful, both as authors and
as users).

But all authors want maximal access to the refereed (hence published)
postprint, immediately. If only the publisher is relied on to provide
this access, not only is there the temporary publication lag to
worry about, but permanent access-denial (to all would-be users whose
institutions cannot afford the access-tolls). And the latter is what
open-access-provision is intended to remedy.

>h> This is not how research progresses, and not what open access is for
>h> or about! Open access means *immediate* open access. It in fact starts
>h> *before* refereeing, with the preprint (optional), and continues,
>h> unbroken, to the moment when the accepted, peer-reviewed final draft
>h> (the "postprint") exists, which is immediately self-archived too.
>U> Sure--we don't disagree on that.
>h> Surely with this "soon after" you don't want to reduce open access to
>h> the Dave Shulenburger old and inadequate "NEAR" proposal of free online
>h> access only after an agreed embargo period!
> It's much simpler than that: not a prescription for a particular
> publishing model, just a desirable characteristic for a system of
> communication.

If I understand correctly then, all you were talking about here was
the importance of trying to minimize publication lag, irrespective of
open-access matters?

>U> "self-archiving and open-access journals, by themselves, do not
>U> guarantee 'permanent open access'."
> >
>h> This raises the red herring of preservation in a place where it
>h> harms rather than helps: Of course digital contents need to be
>h> preserved permanantly. But in the special case of the peer-reviewed
>h> journals, the preservation problem is squarely on the shoulders of
>h> the 23,400 journals and the libraries that subscribe to them at this
>h> time.
>U> Well, journals don't do preservation. Libraries do. So, at least on
>U> half of that point we agree.

We fully agree on half the point: There is a definite preservation
problem for the toll-access journal literature: Who will ensure that
the online versions of the contents of the 24,000 journals with their
annual 2,500,000 articles remain accessible and usable permanently?
This is a problem between the toll-access publishers and the libraries
that pay the tolls to purchase these contents.

But open-access-provision concerns something else: the self-archived
open-access *duplicates* of (or approximations to) those toll-access
articles. These duplicates are merely supplements to (not substitutes for)
the literature in question; they are provided for the sake of immediate
ongoing access to those who cannot afford the canonical version.

Until and unless publishing does downsize to peer-review/certification
service-provision alone, and offloads all access provision onto the
institutional network of self-archived versions (which are currently
merely access-supplements to the canonical versions), it is not the
self-archived supplementary system that has the preservation burden but
the primary toll-based system.

>h> That is all toll access.
> Libraries don't charge for access. They do limit access when the
> publishers they rent from require them to do so.

We are talking here only about restricted toll-access articles:
articles that are licensed by the library to be made accessible only
to their own institutional users and no one else.

> A library-based open-access model is much more powerful than one based
> on individual self-archiving, because preservation is an important
> characteristic of a system of scholarly communication, and individuals
> can't guarantee it (nor can publishers).

What is "a library-based open-access model"? Where do the open-access
articles come from? Whose articles are they?

(1) We are presumably not talking here about the small number of
open-access journals that exist today (600, versus 23,400 toll-access

(2) Nor, as we have just confirmed, are we talking about the restricted
local access journals the library pays license-tolls to access.

So exactly what form of open-access do you have in mind here, if not
institutional self-archiving?

The university-press new-journals we spoke about earlier? Are
they open-access journals? Then that is just (1) again. Are they
toll-access? Then that is just (2) again.

So where's the open access?

>s> the physics corpus, doing this since 1991, is all with us today
> all of 12 years later.

Yes, all of 12 years -- and 250,000 open-access articles being actively
accessed and used across those years, and all the research impact and
progress that that has generated. Compared to those of us who during
those 12 years, and still today, are not self-archiving but instead
fretting about preservation of the contents of our empty cupboards,
worrying that, if we self-archive, those open-access versions may not
be accessible in 12 years! Meanwhile, the years keep going by, and the
cumulating research impact loss keeps growing...

Stevan Harnad
Received on Mon Dec 08 2003 - 14:32:20 GMT

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