Re: The Green and Gold Roads to Open Access

From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
Date: Mon, 13 Dec 2004 01:04:00 +0000

On Sun, 12 Dec 2004, Jan Velterop wrote:

> Economic viability, sustainability and scalability don't need to be
> shown. The only thing that needs to be shown is 'cultural' acceptance
> in the research community. Or even just in the funder community, which
> will do fine. Economic viability and sustainability will follow.

I of course agree that OA journals do not need to show their economic
viability, sustainability and scalability in advance in order to be
created by OA publishers and used by authors and readers. I said economic
viability, sustainability and scalability need to be shown before non-OA
publishers (95%) will consider converting to OA.

> But self-archiving carries risks for those publishers, too. Even though your
> stock-in-trade answer is that such risk is 'counterfactual', given what
> happened so far in the high energy physics field. However, just as in
> investment banking, past performance is a poor indicator of future
> results.

I never say self-archiving carries no risk for publishers. I always say
it carries far less risk than converting to OA publishing (until its
economic viability, sustainability and scalability have been shown).

    "The Green Road to Open Access: A Leveraged Transition"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/3378.html
    http://www.ecs.soton.ac.uk/~harnad/Temp/greenroad.html

>sh> Now, because only about 5% of the total 24,000 peer-reviewed
>sh> journals have taken the risk of trying the OA cost-recovery model
>sh> today, it follows that only about 5% of articles can be published
>sh> in an OA journal today, even if the author, undeterred by the
>sh> author-institution publication cost (as, I agree, he should not
>sh> be, if the journal is otherwise suitable) wishes to publish in
>sh> an OA journal.
>
> This is a logical flaw that presumes that paperflow is always static,
> from journal to journal, and that there can be no shift in submissions
> from one journal to another. It is plausible that not all articles at
> the moment can find an appropriate OA journal to be published in, but
> the implied proportionality to the number of journals in your argument
> is wrong.

I agree that there can and should (and will) be a reduction in the
total number of journals (with or without OA). And I agree that the 5%
OA journals that exist now could probably publish 10% perhaps even 20% of
current articles (if authors prefer to submit there). But the fundamental
point is that there are no suitable OA journals for at least 80% of the
literature today, and no inclination on the part of the non-OA journals
in which those articles appear to convert to the OA cost-recovery model.

(Nor, let's face it, is there as yet such a pressing author demand for
OA journals! Authors sign petitions for OA, but they do little else,
even though OA is for their benefit, and they ought to! Three times as
many authors self-archive for OA (15%) as publish in OA journals (5%),
but that total of 20% still doesn't amount to a hill of beans; and if I
were a non-OA publisher trying to find evidence that OA is not all it's
cracked up to be, and that perhaps researchers don't really want it after
all, *that* is the evidence I would use! Fortunately, there is a good
deal of counter-evidence, and the main missing element is researcher
awareness of the strong causal connection between OA and impact. Once
that evidence is wider known and better understood, researchers will be
more ready to provide OA and their institutions and funders more ready
to mandate that they do so, for their mutual good! [Using data on
correlations between research impact and research funding, as well
as between research impact and researchers' salary, I shall soon be
translating the OA impact-advantage data into dollars -- the only language
that seems to talk!])

>sh> Distributed institutional self-archiving is simple, easy
>
> If only. Plenty of institutions do not have a repository yet,
> unfortunately. Only a concerted central archiving in
> discipline-oriented archives (such as PubMed Central in medicine and
> biology), which you seem to abhor, could conceivably deliver the
> immediacy you're looking for.

Jan, I am afraid you are quite wrong on both counts!

(1) The fact that plenty of institutions don't have OA archives yet
is certainly not evidence that they are not simple and easy to create!

(If BMC really intends to get into the business of helping institutions
create archives, they had better get a better idea of what is involved.
I suggest you start with reading the Handbook:
http://software.eprints.org/handbook/
and then look more closely at the small subset of existing OA archives that
not only exist, but have successfully overcome the much bigger hurdle,
which is getting themselves filled. CalTech's is a good example:
http://www.arl.org/sparc/pubs/enews/aug01.html#6
Southampton's ECS archive is another (moreover, they wrote the book
on the subject!):
http://software.eprints.org/handbook/departments.php

(2) I don't abhor central OA archives (I founded one of the first of them
myself: http://cogprints.org/ ). My preference for distributed institutional
OA archives is purely practical, in the interests of propagating the
practice of self-archiving and accelerating the growth of OA. And a recently
published JISC study seems to have come to the same conclusion:

    "For technical and cultural reasons, this study recommends that
    the centralised model should not be adopted... [The] central archiving
    approach is the 'wrong way round' with respect to e-print provision.
    [For] reasons of academic and institutional culture and so long
    as effective measures are implemented, individual institution-based
    e-print archives are far more likely to fill (and fill quickly)
    than centralised archives, because institutions and researchers
    share a vested interest in the impact of their research output,
    and because institutions are in a position to mandate and monitor
    compliance..."

    Swan, A., Needham, P., Probets, S., Muir, A., O'Brien, A., Oppenheim,
    C., Hardy, R. & Rowland, F. (2004) Delivery, Management and Access
    Model for E-prints and Open Access Journals within Further and
    Higher Education.
    http://www.keyperspectives.co.uk/OpenAccessArchive/E-prints_delivery_model.pdf

    "Central vs. Distributed Archives" (1999)
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0293.html

    "PubMed and self-archiving" (2003)
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/2973.html

    "Central versus institutional self-archiving" (2003)
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/3205.html

> I have compared self-archiving with a painkiller, as you know: it works
> to relieve the symptoms, but doesn't cure the underlying problem.

Jan, as you know, I have answered this "symptoms-vs-cure" argument once already:

> Harnad (from the floor): What is the purpose of the OA movement?
>
> Velterop: Can only answer for BMC: BMC believes OA is good, and that
> it can be done on a commercial basis.
>
> And, one can only add, for other than BMC, that OA is not only good,
> but it is the purpose of the OA movement! And OA can also be done on
> a non-commercial basis, by author/institution self-archiving!
>
> Harnad: Non-Open Access is the pain.
>
> Velterop: BMC is also killing the cause of the pain.
>
> The cause of the pain is needless impact-loss from Non-Open Access and
> 100% OA self-archiving cures that pain, completely. The rest is purely
> speculation about the *consequences* of 100% OA self-archiving, not
> causes. (Hypotheses non Fingo.)

    Re: Cologne Summit on Open Access Publishing
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4222.html

But let me venture to give another piece of strategic advice, specifically
to counter that (hypothetical) non-OA publisher I mentioned earlier, who
might be interested in adducing evidence that researchers don't really
want/need OA as they are purported to do: Your suggestion (Jan) that 100%
OA from self-archiving would merely be treating the symptoms and not
the underlying cause (which only a change in cost-recovery model would
cure) comes dangerously close to undermining the case for OA. Because
the case for OA is not that journal prices are too high and journals
make too much profit!

The essential case for OA is that research usage and impact are being
needlessly lost. 100% OA (by whatever means) would solve that problem --
*if* it is indeed the problem. But if you were really right (Jan) that the
true problem is not research access/impact but journal cost-recovery
models and excess profits, then I am not surprised that most authors
do not bother either to publish in OA journals or to self-archive, nor
would I expect even our forthcoming evidence about impact enhancement
to persuade them to do it either. For if the only ones who are in the
position to *provide* OA to their own work (whether via gold or green)
won't do it for the sake of their own research impact (because that is
not the real problem), then they *certainly* won't do it for the sake of
journal cost-recovery models and excess profits! And that would certainly
make the non-OA publisher who is trying to show that researchers don't
really want or need OA, as alleged, very happy indeed!

I would accordingly urge that you drop this "only a pain-killer" argument
and recognise the fact that either OA itself is the cure for needless,
chronic research impact loss or the whole case for OA is bogus!

>sh> The same is true of archive creation and maintenance. It costs a
>sh> little (very little) to an institution, to provide OA archives
>sh> for all its authors article output, but it costs incomparably
>sh> less that creating and maintaining a new journal, and has none
>sh> of the attendant risks.
>
> It has the risk that the journals whose articles are being
> self-archived disappear for lack of subscription income. Not a bad
> thing, perhaps, but surely a risk.

Before they "disappear" (on this doomsday hypothesis) their editorial
boards, refereeships, authorships, titles and track-records can migrate
to one of the brave new OA publishers like BMC: Next hypothesis? (One
can keep playing this empty game of speculation and counter-speculation
for another decade, easily, without any concrete progress...)

> If they don't disappear, then the institution pays two bills:
> first for the subscriptions, and then for the maintenance of the
> repository. So where does the idea come from that it's cheaper?

Jan, your one-sided view of OA has tied you into knots that you don't
even notice. What I said was that creating an institutional OA archive is
far cheaper and easier than creating an OA journal! (And journals are
created by the publisher, not the author's institution!) The "two-bills"
argument has *nothing* to do with this. The institution continues to pay
for incoming journals, as before, but its own outgoing research articles
are self-archived so as to maximize their research impact, at a cost of
<$10 per article. My impact/income figures will demonstrate that this is
a bargain, generating far more impact-income than the pittance it costs.

The idea that something should be cheaper for the institution overall
comes from *you* (and all the others who have conflated the research
output access/impact problem with the research input affordability/pricing
problem)! The purpose of OA, the reason researchers sign petitions for it,
publish in OA journals for it and self-archive for it -- and the reason
(let's hope!) that their institutions and funders will require those
who do not now do that to do it -- is research impact (and income):
The purpose is *not* to solve the journal affordability/pricing problem
(by making journals cheaper)! It *might* eventually have that aftereffect
too, but OA certainly will not be provided merely or primarily for the
sake of that hypothetical aftereffect.

What is cheaper is creating OA archives, compared to creating OA journals.
What is *not* cheaper (and not intended to be) is institutions' existing
journal costs plus their (small) new archiving costs, compared to institutions'
existing journal costs alone.

> For a handful of very well endowed and research-intensive institutions
> subscriptions may be (not *are*; just *maybe*) cheaper than paying for
> dissemination, as in the OA journals that charge article processing
> fees. For the majority OA is cheaper.

Jan, you are still tied in a knot, unable to see the calculation and
comparison I was actually making, because you seem to be able to think
only in terms of OA publishing and its costs, not in terms of just OA
archiving and its costs -- alongside current journal costs!

To repeat, what I said was that it is cheap and easy to create an OA archive,
and much cheaper and easier than to create an OA journal.

This did not imply that institutional journal costs plus institutional
self-archiving costs would be the same as or less than institutional journal
costs alone (though they will not be much more).

But least of all did it imply anything about whether institutions would be paying
less or more overall if there were 100% OA publishing, compared to what they
are paying now. No doubt they pay more now, but the problem is with getting
there from here. And inasmuch as reaching 100% OA depends on either
persuading all or most existing journals to convert voluntarily to the
OA cost-recovery model (forget it!) or persuading all or most authors
and their institutions and funders to provide OA to their research article
output (either by publishing it in an OA journal or by self-archiving it),
it all depends critically on there being a far more persuasive reason
to induce authors (and their institutions and funders) to provide OA
than: "it will reduce journal expenditures."

> A simple calculation tells us that the average price of an article paid
> by Academia is in the order of $3000-4000 at least. No OA publisher
> charges that, and certainly for the bulk of scientific articles the
> fees necessary for OA publishing to be sustainable, viable and
> scalable, are nowhere near such amounts. But even if they were; or even
> if they were higher; OA publishing would still be the better system, as
> for that amount all scientific research results could be freed up. Why
> would that be the case? Because "no achievement in science is
> exclusively the product of one brain" (John Waller, in 'Leaps in the
> Dark'). With OA, there can be more and faster 'interconnectivity' with
> more brains, so to speak.

You seem to see it as the pricing/affordability problem in the first
instance, and the access/impact problem only as an afterthought: I
suggest that a change in priorities and strategies might be in order,
if researchers are to be persuaded to go along for the ride: It is their
research, after all, and their research impact and income is not merely
a secondary afterthought for them: it is their primary concern. Tell
them that's merely a headache, and 100% OA via self-archiving (which,
unlike 100% OA via OA journal publication, is *entirely* within their
immediate reach) is merely hiding symptoms, not curing them, and you
lose the best and only ally you have, in the quest for OA (and you also
lose the real rationale for OA itself)

>sh> But OA self-archiving can be mandated it. And that will bring 100%
>sh> OA (wireless) with 100% probability.

> Of course. As it is beginning to happen. But to think that that means
> that traditional publishing is *not* moribund if it were take up even
> to 50%, is not understanding the business of publishing. If OA
> archiving is mandated immediately upon publication, traditional
> publishing is dead. If it is with a 6-months delay, subscriptions pay
> for no more than a brief time advantage, and for that, current price
> levels may just prove far too high, so cancellations are pretty sure to
> follow. I know you call this speculation, but I prefer calling it
> 'looking ahead' or 'anticipation' and would like to plan for that
> future which I regard as inevitable (though it's anybody's guess how
> long it will actually take).

I call it speculation and plead nolo contendere. What we need is an
immediate self-archiving mandate, and a mandate for immediate (not
delayed) self-archiving. That will generate 100% OA. How things may or
may not evolve after we have 100% OA is a matter of far less interest
and urgency than putting an end to needless access/impact less as soon
as possible. It is in fact already well overdue.

>sh> It may or may not also eventually usher in the era of OA publishing.
>
> OA publishing is already here, albeit still small.

We have an OA era when we are at or near 100% OA. We have an OA publishing
era when we are at or near 100% OA publishing. We in fact have 5%
OA publishing, 15% OA self-archiving, and hence 20% OA today. Hardly an era
for either, yet...

>sh> We need to create/convert roughly 23,000 more OA journals to
>sh> scale up to 100% OA [via OA publishing] whereas we could already
>sh> do so virtually overnight via OA self-archiving.
>
> This presumes the absurd notion that we actually *need* 23,000
> journals. We may have them, and OA publishing may result in 50,000 or
> 100,000 journals, but we only *need* enough journals to have an
> appropriate outlet for each paper.

I don't disagree with this, but it still leaves us at 5% now, which is next
to nowhere -- and nowhere near enough.

>sh> But alas it is easier to squeeze a camel through the eye of a
>sh> needle than to get most publishers to adopt the OA cost-recovery
>sh> model today. It is hence incomparably easier and more sensible
>sh> to squeeze researchers just a little, so they self-archive:
>sh> After all, the OA is all for their benefit anyway!
>
> Some authors do realize the benefits, but some don't.

Perhaps those who seek to maximize their research impact (and already know
OA will do that) realize OA's benefits, whereas those who think OA's just
meant to solve the journal affordability/pricing problem don't...

Stevan Harnad
Received on Mon Dec 13 2004 - 01:04:00 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:43 GMT