Serials Review Interview

From: Stevan Harnad <harnad_at_COGSCI.SOTON.AC.UK>
Date: Tue, 13 Oct 1998 10:59:18 -0400

Date: Wed, 7 Oct 1998 21:47:13 +0100 (BST)
From: Stevan Harnad <>
To: Ellen Finnie Duranceau <efinnie_at_MIT.EDU>
Subject: Re: Questions about scholarly skywriting for Serials Review

> 1. You've written extensively about how scholarly publication could be
> redesigned for the digital world. Can you characterize your model briefly
> for readers who may not be familiar with it?

Very simple, though, based on experience, remarkably easily misunderstood:

(1) The model applies only to the refereed journal literature,
not to books, textbooks, magazine articles, best-sellers, films, etc.
The simple test is: Does the author get paid for the text (whether
by fee or royalties)? If the answer is yes, my model does NOT apply.

(2) The authors of the refereed journal literature, not writing for
fee, wish only to maximise the visibility and accessibility of their

(3) Because of the genuine expenses of the paper era, the only way
refereed journal authors could get published at all was by allowing
publishers to recover their expenses and a fair profit through
Subscription, Site-License or Pay-Per-View (S/SL/PPV). Copyright was
assigned to the publisher to prevent theft of the product.

(4) In the online-ONLY era (note the "only") the costs per page shrink
to at most 1/3 of what they are for paper publication; those remaining
costs are for peer review and editing (i.e., quality control for
content and form). Printing, Distribution, Fulfillment and Marketing
costs vanish.

(5) Rather than recovering that remaining 1/3 the old way, through
S/SL/PPV, with the access blockage it entails, it should be recovered
as page charges paid by the author (out of publication funds provided
by the author's institution, out of -- and for the sake of -- the 2/3
savings from S/SL/PPV).

(6) The result is free online availability to everyone, which is the
optimal outcome, and also inevitable. However, arriving at this will
take too long, and researchers and researchers will be denied its
benefits, if they wait for publishers to adopt it spontaneously.
(Instead, publishers will try to continue selling both paper and online
editions via S/SL/PPV, in the hope that the online-only era will
continue to be funded that way.)

(7) To hasten the optimal and inevitable, authors should publicly
archive their unrefereed preprints and their refereed reprints on their
home servers as well as in a global archive such as the Los Alamos
Eprint Archive [ ].

(8) XXX already has at least 35,000 users daily and archives at least
14,000 papers annually. Once this is generalised to the other
disciplines, library subscription cancelations will place pressure on
finding an alternative funding model, publishers will switch to
online-only and page charges, and the windfall savings from S/SL/PPV
will become available to fund them.

There will, however, be an unstable interim period during which some
tide-over subsidy will be needed, because online-only and its attendant
savings cannot be switched to overnight, and the user population needs
preparation, both as readers, to become addicted to free online access,
and as authors, to overcome the current stigma of page-charges
(associated with paper publication, where there is no justification for
page charges, and vanity-press publication, which is irrelevant, but
linked to page charges in authors' minds at the moment).

> 2. Has your/How has your vision for the future of scholarly publication
> changed since you originally conceived it? What factors have led to the
> change? Has the influx of web versions of commercial journals played a role?
> Have things moved more rapidly or less rapidly than you'd expected?

Originally I had been an advocate of online-only because of the unique
power of the new medium, for speed, access, and interactivity (via
commentary). I disagreed at the time with those who thought it could be
provided for free, because they tended to be against quality control,
and in favour of either no peer review at all, or an absurd form of
peer commentary which I knew (from 20 years of experience editing a
commentary journal) was no substitute for peer review, only a
supplement to it.

But then I realised that although quality control could not be
cost-free in the online-only medium, its cost would be so much lower
that it would make it possible to do away with the toll-gate and
fire-wall barriers to access posed by S/SL/PPV out of only a small
part of the savings from the cancelation of S/SL/PPV.

So I too became an advocate of free access. I think I formulated my
Subversive Proposal (
at a time when the whole thing was not yet entirely clear in my head: I
advocated authors archiving all their papers on their home servers
mainly just to hasten the online-only era, but it soon became evident
that the free access that that would provide, once available, would never
again be given up, by either users or authors. Then all it took was
arithmetic to realise that if costs would be 1/3 for online-only, they
could be paid out of the 3/3 savings from S/SL/PPV cancelations, still
leaving 2/3 saving -- plus a free refereed journal literature for
everyone, everywhere.

Meanwhile, of course, the publishers quite naturally turned instead to
the offer I've called the "Trojan Horse": Hybrid publication, both
paper and online, offering the paper edition for the usual price, the
online edition for a bit lower, and both editions for a bit higher, and
then letting demand shift to online-only whenever its time comes, but
always supported by S/SL/PPV (and its attendant toll-booths and
fire-walls blocking free access).

If that Trojan Horse were to succeed, it would suffice to delay the
optimal and the inevitable for a while (I gave up predicting when it
would all come to pass a long time ago: all I say is that we COULD do
it very quickly if we wanted to), but meanwhile, I and others will
continue along the evangelical path of subversion, by promoting author
eprint archiving.

Have things moved more rapidly or less rapidly than I expected? I
haven't really tried to second guess the nature of the horse, just to
lead it to water! XXX has certainly prevailed faster than anyone
expected, and I am certain it will eventually prevail in all

> 3. You have described the physics pre-print archive run by Paul Ginsparg at
> the Los Alamos National Laboratory [I think I have his affiliation correct?]
> as a model for how all disciplines could make their research known to other
> scholars. This archive, if I understand correctly, is funded in part by the
> National Science Foundation and seems to rely to a great extent on the
> dedication and vision of one individual. Is the Ginsparg model applicable
> outside physics (or the sciences), and without such a visionary leader? How
> is this model sustainable and scalable, so that it can really offer a
> longterm solution to publication in other (especially non-scientific)
> disciplines? To what extent is the Ginsparg model an exact replica of what
> you would extend to other disciplines?

XXX is funded by both NSF and DOE. The funds are mostly for development
of new features, because just the upkeep of the archive, along with its
steady linear growth rate, is not expensive at all. It will generalise
to all disciplines. It is already doing so. It lately subsumed computer
science too. My own cognitive science archive, CogPrints, is being
groomed for eventual subsumption by XXX. I am keeping it separate now
just to develop a submission front-end that is more congenial to other
disciplines, and especially users of text-processing software over and
above TeX, such as MS Word.

As I have said many times before, I am just playing John the Baptist to
Paul Ginsparg's Messiah. All the historical credit shall and should be
given him. It was his vision, his software, his implementation that set
the literature on its inexorable course to the optimal and the
inevitable (though I doubt that he saw all the implications explicitly
at the beginning either).

> [if of interest or helpful to extend the discussion, could you comment on
> the example Thomas Walker uses of the Florida Entomological Society?]

I've done that in my paper for the AmSci Forum. Walker's heart is in the
right place, but his proposal to tide over the transition (from
S/SL/PPV-supported paper publication to page-charge-supported
online-only, free for all) out of author reprint charges is incoherent:
Why should I pay reprint charges to a journal when I can archive all
my\papers on my home-server and on XXX for free?

> 3a. Would you envision that the quality-controlled, peer-reviewed results of
> a preprint archive's article would be officially sanctioned at the preprint
> archive itself, or only through traditional publication? (That is, how are
> preprints and complete articles differentiated in your model?)

Peer review is medium-independent. Refereed journals are simply
implementers of peer review. They should continue to do that; there is
no alternative I know of. And there should continue to be a
hierarchical spectrum of peer-reviewed journals, varying in their
subject matter as well as their quality and rigour. That should all be
financed out of the page charges. The archive is just the means of
access. Papers in the Archive should be explicitly tagged as UNREFEREED
or REFEREED (and if the latter, tagged with the brand-name of the

During the subversive phase, authors could simply tag their own papers
in the archive as unrefereed or refereed, specifying the journal in the
latter case. (I am not worried about whether they can be trusted.)

Eventually, though, arrangements will be made between journal
publishers, such as the American Physical Society (APS), and XXX: APS have
already agreed that authors can submit to APS journals via XXX; later
the refereed, accepted APS version's tag could be officially
authenticated by the APS itself.

But frankly, that's just technical detail. In broad strokes, the
picture is clear.

The Archive can also include peer commentary (both unrefereed and
refereed) and authors' responses. But this peer commentary should not be
confused with peer review, for which it is a supplement, not a

> 4. Could you comment in particular on the archiving arrangements that you
> would expect to provide satisfactory redundancy and reliability in the
> scholarly skywriting model? In particular, how would the mirror sites you
> have mentioned be motivated to provide a backrun of an electronic journal or
> a group of preprints indefinitely and for all interested parties? (Or would
> libraries play a role here?) What about the costs of storing, moving and
> refreshing files and porting them to new platforms as needed? (Here I would
> like to address head-on the concerns that research libraries have, that we
> don't have a model for the digital world that offers the same redundancy and
> neutral, longterm access that academic and research libraries have offered,
> as a group, in the US in modern history. Do publishers and authors care
> about preservation and longterm access the way libraries do? Can they be
> expected to take this on?)

I have just today come back from an eLib preservation meeting in London,
and, as usual, librarians' hearts are in the right place, but their
heads are full of needless and misplaced worries, motivated, I now
believe, by a very simple, paper-based "intuition pump": They think of
the "preservation" problem as requiring some analogue of paper, some
undying object, multiplied many times all over the world, to fend off a
Library of Alexandria calamity.

The truth is that once text is digitized, bits are bits, and all that's
needed is a means of turning the bits into a form (be it paper or
screen) that is accessible to the human senses. THAT hybrid bytes-to-eyes
interfacing capability is what needs to be preserved, not objects
strewn over the planet.

And, yes, multiple mirror and backup sites are part of preserving that;
so is centralisation, because the more authors' and disciplines'
intellectual goods you have in one collective (but distributed and
redundant) basket, the more collective interest is vested in
continually keeping the bits accessible to the eyes; that is the
continuous upgrading that is these days called "migration."

There are details and technicalities, but the essence of it is that
technology will keep evolving to keep making those bits viewable and
navigable by the senses, and the more eggs are in the same basket,
the more eggs will benefit from the same fate. There was no such
collective dimension to paper, nor was there the part about the
bytes-to-eye link to be continually upgraded, but those are all soluble

But if you keep thinking of it in terms of the dreaded "orphaned
CD-Rom" that either gets blown off the face of the earth or that no one
can make head or tail of any more, you are just pumping the wrong
intuition pump!

The motivation comes from the community that vested its interests in
the continuity and upgrading. Moreover, the expenses go down, per
item, with scale. Imagining any government ready to pull the plug on
the a virtual basket containing scholarship's intellectual goods is as
plausible as imagining them pulling the plug on the national electrical
grid, or slaughtering, Moses-like, every last volume on the planet.

> 5. Much has been said about the reduced costs of publishing electronically.
> In your model, if I understand it, authors absorb the largest portion of the
> cost of the editorial side of publication (refereeing and editing) through
> page charges. Would the additional costs of running a live archive, mirror
> sites,and preservation archiving be funded through grants? Is this a sound
> foundation?

Page charges are to cover online-only journal publication expenses (at
most 1/3 of what they are now), to be paid by authors instead of
libraries (not out of their pockets, of course, but out of the library
saving) so that the learned serial literature can be free for all.

As I said, more of the current XXX subsidy is for development than for
maintenance, and a lot of the developmental cost is one-time only. With
economies of scale (I am improvising now), XXX could probably scale up
to include all the annual papers in all fields in the 14,000 refereed
journals covered by Ulrich's for only a few times what it costs for
just covering Physics now. And that expense is tiny and sustainable.
Take it out of part of the remaining 2/3 savings from going from paper
to online-only, in your reckoning, but it'll be a lot less than another

The point is that the learned serial literature is already being
subsidised through S/SL/PPV anyway; switching to online and redirecting
the remaining costs will not only make the subsidy much lower, but it
will provide a much more powerful and efficient mode of access, and
provide it for free for one and all.

> 6. Is there any role for commercial publishers in your model? Are commercial
> publishers a negative force in terms of achieving your goals?

Anyone prepared to scale down to this new niche -- providing quality
control to authors -- can be a player.

But commercial publishers have other products too, and those do not fit
the refereed-journal model (wide-spectrum monographs, textbooks,
popular magazines, etc.), and I'm not sure how such big enterprises
will mix with cottage-scale ones like what online-only refereed journal
publication will be once it has achieved the optimal and the

It's not just commercial publishers who will want to cling to the
S/SL/PPV status quo for as long as possible; most big publishers will,
including Learned Society and University presses. It's only natural.
They will fail, of course, because they will be fighting against the
optimal and the inevitable for scholars and scholarship, research and
researchers, but it is, I suppose, natural in the Darwinian marketplace
to try to prevail along the old lines as long as possible. But the
conflict of interest is a great vulnerability: their constituency,
after all, is us, the authors and readers. We acquiesced in the
Faustian Bargain of bartering copyright for publication while there was
no option, but now that there is an option, we will realise it sooner
or later (helped by subversion).

Learned Society publishers like the APS, whose Editor, Martin Blume,
was a co-signator of the excellent manifesto in Science last month (for
mandating that funded authors should retain the right to archive their
papers publicly --,
are more likely allies than commercial publishers, as are University
Provosts, like Steve Koonin, whose similar proposal was debated in the
Chronicle of Higher Education

> 6a. Right now, commercial publisher own the copyright of most authors' work.
> In your model, do individual scholars retain copyright?

For refereed journal articles they need only retain the right to
archive their papers for free for all, in perpetuo. We are not talking
about movie rights here! Publishers can have all the rights to sell the
paper version (or their own online version, for that matter) for
S/SL/PPV, as long as the author retains the right to archive it for
free, forever.

> Does this create
> barriers to innovation and marketing of research tools and services? (e.g.
> the indexing and abstracting or fulltext services that now negotiate with
> large publishers to provide access to their material; would they now have to
> track down individuals?)

Utter nonsense. The innovation will all be in the expanded XXX
full-text archive itself, and the navigation resources that users
will design on that rich, fire-wall-free, toll-gate-free corpus:

If primary publishers, especially commercial ones, need to do some
rethinking about their future course, I'd like even less to be a
secondary or tertiary publisher right now: Here's the proof. Just
imagine the subversion being successful and complete. XXX now contains
the 14K refereed corpus, suitably tagged. Now, do you want to pay to
use ISI or Medline or some other service, or would you rather just
send off a good old Alta Vista searcher armed with those all important

Who is there to track down? And what for? Authors give it away, as they
always have; XXX houses it; generic engines find and retrieve it.

That goes for citation searches too. (See what the Open Journal Project
showed could be done with a few journals plus the ISI database; now
imagine doing that on the full-text corpus directly: all articles have
their own reference lists!)

> 7. Is it useful to offer a parallel to another marketplace that operates in
> some way along the lines of the Ginsparg model? Is the historical model of
> scholarship shared through learned societies a valid precedent?

I don't understand this question.

> 8. Recently you devised and coordinated an interactive discussion forum at
> the American Scientist web site, offering readers a chance to debate about
> the public eprint archive model and other ideas about scholarly publishing
> in the digital age. How was this forum successful? In what ways was it
> disappointing? Did it advance any line of argument?

It was fine; opposing views were aired. I think readers will be able to
draw their own conclusions. Numerically, people who knew less or had
an interest in nay-saying prevailed, but rationally, I don't think they
did. But you'll need to get other people's verdict on that.

There was also a lot of naive nonsense, which ought to be filtered out.
But I think the AmSci debate could be distilled down into a useful
Quote/Comment Virtual Symposium, mainly involving the contributions of
Arthur Smith, Mark Doyle, and myself, plus some individual
interventions by a few others.

> 9. What is the greatest barrier to achieving your vision?

Human nature. (See above, about horses, water, and drinking.)

> 10. What, if anything, would you like to say to librarians?

Stop worrying about preservation; support subversion; don't take in
any Trojan Horses.

Stevan Harnad
Professor of Cognitive Science
Department of Electronics and phone: +44 1703 592582
Computer Science fax: +44 1703 593281
University of Southampton
Highfield, Southampton
Received on Tue Aug 25 1998 - 19:17:43 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:28 GMT