Re: On the Need to Take Both Roads to Open Access

From: Stevan Harnad <>
Date: Sat, 11 Oct 2003 14:05:50 +0100

These are comments on two October 9 articles on open access in Nature by
Declan Butler (plus an accompanying letter by John Ewing).

> Who Will Pay for Open Access/DECLAN BUTLER
> Will scientists, their host institutions and those who fund their research
> embrace the author-pays model? And if they do, is $1,500 per article enough
> to cover the costs of producing a journal of the highest quality?

The quality of a journal depends on the quality of its submissions
and the rigor and selectivity of its peer review. Authors give
their papers for free; referees referee for free. The only cost is
administering the peer-review service. The highest-end estimate for
the cost of implementing peer review alone has been $500 per paper:

> For most researchers in the physical sciences PLoS's campaign is a side
> issue. They routinely make their papers freely available before formal
> publication using online preprint archives such as arXiv org. is a central archive, mainly for physics, but also for
mathematics, computer science, and (as noted in Butler's other article,
below), now for quantitative biology too. But neither Arxiv nor any of
the growing number of institutional open-access eprint archives is or has
ever been for unrefereed preprints alone, or even primarily. Open-access
archives are for both the pre-refereeing preprint and the post-refereeing
postprint. The preprint comes, logically and chronologically, before
the postprint in the embryology of an article, but it is the refereed
postprint that is the most important to self-archive and thereby make
open-access. It is incorrect and misleading to equate open-access
self-archiving with preprint-archiving.

I think the reason opponents of self-archiving keep misrepresenting
self-archiving as being only or mainly preprint self-archiving may be
that they wish to sound a note of warning bout self-archiving that
would simply make no sense if it were frankly admitted that both the
preprint and postprint stage of research are being self-archived. Here
is an example:

> But for biologists who are not generally comfortable with prepublication
> the answers to the questions thrown up by the launch of PLoS Biology
> may define the future of scientific communication.

Here is that usual wishful discouraging note again!

The future of scientific communication will indeed be permanently altered
by open access, but open access is not just open-access publication (and
open-access self-archiving is not just preprint-archiving)!

> PLoS's ... letter attracted more than 30,000 signatures although
> few signatories seem to have followed through on their pledge to stop
> submitting to and reviewing for journals that have not acceded to PLoS's
> call for open access. These journals remain in the majority hence PLoS's
> decision to launch its own publishing enterprise.

It is certainly true that the many researchers who signed the toll-access
journal boycott petition had no place to go when their petition failed to
convert the 23,500 toll-access journals into open-access journals (of
which there are still only around 500). That's why PLoS created its two
new open-access journals.

But waiting passively for the one-by-one conversion or replacement of
the 23,500 toll-access journals is not the only road to open access,
nor the fastest: There is also self-archiving, and each researcher can
do that on his own, right now, with no need to wait for anything. In
fact, 55% of journals sampled already officially support author
self-archiving, and many others will agree if asked. Becoming a Romeo
"green" (self-archiving-friendly) publisher is a way that publishers
can provide their support to open access and its benefits to research and
researchers without necessarily having to take the radical and risky step
of converting to open-access publishing at this time:

> Some journals such as the American Physiological Society's Physiological
> Genomics are allowing authors to pay for open online access for individual
> papers while retaining a subscription model for the journal as a whole.

Authors who can afford it are welcome to pay toll-access journals to do
their self-archiving for them, but for those who cannot afford that,
self-archiving for themselves, in their own institutional eprint
archives, is surely the preferable option.

> "I feel that PLoS's estimate is low by four to sixfold says cell biologist"
> Ira Mellman of Yale University editor of the The Journal of Cell Biology.

The true cost of the essentials has been a matter of much debate and
speculation since at least 1998:
and today's open-access journals have had to make a pre-emptive guess,
ranging from $500 per paper to $1500 per paper.

[My own guess is that it is impossible to determine the cost
of the essentials a priori, because we don't yet know what the
essentials are -- and we will not know until/unless competition
from the self-archived vanilla postprints causes cancellation
pressure on toll-access journals, forcing them to cut costs and
downsize to the essentials, which (I again guess) may turn out to be
just peer-review administration, with text-production offloaded onto
authors' XML word-processors and access and storage offloaded onto the
interoperable network of OAI-compliant institutional eprint archives
in which the postprints are self-archived: .]

> in the longer term PLoS and other open access groups must persuade the
> organizations and institutions that fund and host biology research to
> pay their fees. Grants from the US National Institutes of Health already
> allow for the charging of publication fees. Other bodies are moving in
> the same direction.

Yes, funding to cover open-access publication costs is beneficial
and welcome, but funding 500 open-access journals does not solve the
problem of providing open access for the contents of the remaining
23,500 toll-access journals. That depends on formulating systematic
institutional and national self-archiving policies for all research
output, which should be explicitly coupled with these new open-access
publication-funding policies, to maximize the overall return on the

> Peter Suber... notes that many institutions will be reluctant to cover
> dissemination fees while still paying subscriptions to traditional
> journals...

Yes, this double-payment burden is a problem. But if coupled with a
concerted institutional and national self-archiving policy, it is a
rational investment into the future of universal open access.

> PLoS Biology will have to overcome the hurdle that faces any new journal
> irrespective of its business model, convincing scientists -- particularly
> young researchers who need to publish in high profile journals to further
> their careers -- that it is worth taking the risk on a new and unknown
> quantity.

All new journal startups have to first establish their track-records,
but in the online age, and if the editorial boards are strong and the
refereeing quality standards and selectivity are high, they can do this
quite quickly. The Journal of Higher Energy Physics started only a few
years ago (as an open-access journal, incidentally!) and very quickly
attained one of the highest impact factors in its field.

The fact that JHEP later converted from open-access to toll-access is
interesting, but not as negative as it sounds: Its contents all continue to
be openly accessible online because all of its authors self-archive.
JHEP came before the recent momentum for open-access journals. If the
new funding sources for covering open-access publishing costs grow and
extend from biology to physics, JHEP may again be able to revert to the
open-access publishing cost-recovery model. The important thing is that
all of its contents remain open access online, thanks to

> Most publishers remain sceptical about the viability of PLoS's eventual
> goal of converting the entire scientific literature to the open access
> model. But many now accept that the author pays approach may have its
> place. In August the Association of Learned and Professional Society
> Publishers declared itself wholly in favour of maximizing access to
> research literature.

All this support and enthusiasm for funding open access publishing is
desirable and welcome, but I hope it is clear from the above example
that self-archiving must be systematically promoted and practised
at least as vigorously, if the benefits of open access are to extend
to the contents of the remaining 23,500 journals, and not just to the
existing 500 open-access journals, or the ones we succeed in creating
or converting across the coming decade.

> the various proposals for achieving [open access] raise complex economic
> logistical and sociological questions which differ from field to field
> as well as between different sizes and types of publishers. Much more
> information needs to be gathered through experimentation and analysis.

There are two essential facts that do not vary from field to field:

    (1) There is no field that does not benefit from maximizing its
    research impact by maximizing its research access through open access.

    (2) All fields can have immediate open-access through self-archiving,
    now regardless of the current availability of suitable open-access

> Biologists join physics preprint club/ DECLAN BUTLER
> [Arxiv has] created q-bio, an archive for quantitative biology. But
> papers on ArXiv are not peer-reviewed, and there is concern this could
> create problems if medical papers are accessed by physicians or patients.

It is incorrect that self-archived papers -- in arXiv or elsewhere,
including in their authors' institutional open-access archives -- are not
peer-reviewed. The first version that authors may decide to self-archive
might be the unrefereed "preprint," prior to journal submission, but the
peer-reviewed "postprint" also gets self-archived once the refereeing is
complete and the final version is available (except when the differences
are trivial).

Researchers have decades of experience distinguishing unrefereed papers
(tagged clearly as "preprints" online) from refereed papers (which are
tagged clearly by the journal name). The minority of biomedical papers
that have implications for human health could even be assigned a "health
warning" tag, if the biomedical community judges it important enough.
All this ground was already covered in the public discussion of the
original 1999 Ebiomed proposal out of which PubMed Central and
eventually PLoS arose:

> 'Open access' will not be open to everyone: John Ewing Amer Math Soc
> Public Library of Science (PLoS)... will support their [open access]
> journals by charging authors, not subscribers... [about] US$1,500
> per paper. [But] not all researchers are funded by research grants...
> PLoS [says] authors... unable to pay won't have to. But this assumes
> that few authors are unable to pay -- a false premise in many disciplines.
> [H]ow will universities and departments decide which faculty and which
> areas of research are supported? What happens to faculty in small colleges
> with limited resources?

First, there are only 500 open-access journals so far (and only 2 PLoS
journals) out of a total of 24,000 refereed journals. PLoS has a $9 million subsidy. If they say
impecunious authors won't have to pay, you can believe them.

Second, for the the authors in the 500 open-access journals whose
institutions can't afford the publication charge, and for the authors in
the 23,500 toll-access journals that do not yet have suitable open-access
counterparts, there is always the option of open-access self-archiving:
Making their toll-access articles openly accessible by depositing them
in their institution's open-access eprint archives.

> When a scientist doesn't have a subscription, he or she can nonetheless
> get information about the article (the abstract and perhaps a list of
> references); requesting a copy of the article can be as easy as sending
> an e-mail.

But each individual request-for-a-copy costs money (if it is through
publisher's pay-per-view or interlibrary loan) and takes time; it takes
even more time (and is altogether uncertain) if the request is sent to
the author. Those toll-barriers and turnaround times have been made
obsolete by today's online media and instant click-through capabilities!
And it is so that research and researchers can benefit from these
powerful new capabilities of the online medium -- unconstrained by
needless toll-barriers for research that its own authors want all
users to have, toll-free, so as to maximize its research impact -- that
the open access movement has evolved!

Requesting a paper reprint was the old, restricted, papyrocentric
way researchers shared their findings. The new, far more powerful and
beneficial PostGutenberg way is to make an unlimited supply of reprints
available to all users, as online, openly-accessible eprints. No need to
take the time or trouble to ask each time you see or want a paper. And
no access-toll-fees.

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 September Forum (98 & 99 & 00 & 01 & 02 & 03):

Discussion can be posted to:
Received on Sat Oct 11 2003 - 14:05:50 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:05 GMT