[OAI-eprints] Re: Public Access to Science Act (Sabo Bill, H.R. 2613)

[Stevan Harnad]

On Fri, 5 Sep 2003, Sally Morris made a very good point:

>         Stevan Harnad wrote:
>sh>     "Most of the existing 24,000 journals would not
>sh>      accept to publish public-domain texts"
> 
> I think this is probably inaccurate. I would guess that practically all of
> those journals do publish works which are currently governed by the Public
> Domain status of US Government works.

Sally is quite right to point out that I had overlooked the fact that many
publishers are already at home with the fact that a certain percentage
of their authors cannot sign copyright transfer agreements because they
are government employees. Effectively, the Sabo Bill, if it passed, would
simply increase the percentage of such authors. So it was incorrect on
my part to say that they would not accept to publish them: Given the
percentage of journal content that is based on US funded research, they
would be forced to.

But the Bill has not passed yet, and the publishers (and authors) will
still have their say. The percentage of authors who did not sign copyright
transfer in the past (for this reason, or even for other reasons)
was small enough so that publishers could discount it as statistical
variation. But publishers are likely, I think, to try to contest it if it
risks becoming the majority case. Do they have a valid argument?

I think they do, for the simple reason that if the public-domain
constraint is being introduced in order to create open access, then it is
a far stronger constraint than it needs to be. Merely forcing publishers
to allow authors to self-archive accomplishes the very same goal in a
far less radical and risky way -- for both publishers and authors.

For authors, putting their texts into the public domain leaves them
less protected from plagiarism and text-alteration. For publishers,
a large increase of public-domain content could easily threaten
their viability. In this day and age, all we have to imagine is that
another copycat company could systematically (and legally) harvest and
aggregate open-access public-domain contents as soon as they appear, and
immediately offer them, at cut-rate prices, both online and on-paper. Why
subscribe to journal X, which published the contents, if you can subscribe
to journal or aggregator Y for the same contents, at a far lower
price? (US funded research is a huge chunk of many journals'
contents. That's why this Bill is so important. But that's also why it's
so important that it should avoid overkill.)

Wouldn't exactly the same risk be there if instead of mandating that the
contents be public-domain, the Bill mandated only that they be
open-access? Definitely not. With just open access, copyright continues to
be asserted, whether the author transfers it to the publisher (retaining
only the open-access self-archiving right) or merely licenses the content,
retaining the copyright. Self-archived contents cannot be harvested and
re-sold, online or on-paper. The publisher (or author) could immediately
take legal action against that, as before. Self-archiving is the
prerogative of the author, not of third parties, and it does not even
include the *author's* right to sell or re-sell his own texts (that has
to be negotiated separately, as always, if copyright is transfered).

The self-archiving right is merely the author's right to make his own
full-text freely accessible online to any would-be user on the web,
worldwide. That means any individual user, webwide, can read it
on-screen, navigate it computationally, download it, save it, and print
it off. It also means that harvesters like google can link to, invert
and index the full text but they cannot -- if it is copyrighted --
re-sell the text, online or on-paper (otherwise they would effectively
become the rival publisher mentioned above).

Now we come to a delicate point that it is best if all parties --
open-access advocates and adversaries alike -- face up to squarely
and frankly : Although it is true that mandating only open-access rather
than public-domain entails far less risk for publishers (and authors), it
is not true that it entails *zero* risk for publishers. Whereas putting the
contents into the public domain would allow -- even invite -- attempts to
undercut the publisher's business, making the contents openly accessible
online could eventually have a similar effect, indirectly: For whereas
with public-domain contents, rival cut-rate publishers could legally
capture the publisher's market, it is conceivable that with open-access
contents the demand for the publisher's version would simply shrink,
because users could could get everything they needed using the open-access
version.

There is no point denying that this could be the eventual effect
of the Sabo Bill even if it mandated only open-access rather than
public-domain. The glue of OAI-interoperability
http://www.openarchives.org/ for all self-archived research means that
eventually all users would have the option of accessing all research
journal content for free from one global virtual OAI archive in the
sky. But the speed and certainty of abrupt demand-loss is far greater
with the public-domain option -- and with no advantage at all to
anyone. Publishers would be put at a needless increased risk of a
forced, hasty transition to open-access publishing -- a choice that
should be left to them, and to supply and demand, and to time: *as
long as open-access itself is provided right now.* Nor would authors
be better-served by abruptly being forced to place their texts into an
unprotected, untested limbo.

The transition to open-access publishing, if and when it actually took
place, would be a far slower, smoother, and more natural one if it
occurred as the gradual result of author/institution self-archiving
rather than a draconian public-domain mandate. The open-access
cost-recovery model is not yet a tested one. Nor are the sources
from which to cover the costs yet available or assured. Moreover,
the model's probability of succeeding is far greater once journal
publishers have had a chance to accommodate to it gradually -- as they
would if authors were gradually beginning to self-archive their own
texts in their own institutional archives. 
http://www.nature.com/nature/debates/e-access/Articles/harnad.html#B1

With the help of the glue of OAI-interoperability and cross-archive
search engines like OAIster http://oaister.umdl.umich.edu/o/oaister/
users too would gradually learn the benefits and possibilities of a
growing online open-access literature. No specific journal would be put
abruptly at risk, because self-archiving is a distributed, anarchic
process, even with the glue of OAI-interoperability. It would take
a long time for libraries or users to ascertain at what point enough of
the contents of a given journal were openly accessible to make it safe
to cancel the subscription or license to it.

And if and when cancellation pressure did begin to build up, publishers
would have a chance to experiment gradually with cost-cutting and
down-sizing, to keep making ends meet. There is much talk today about
immediate transitions to open-access publishing, with wide variation in
how much it should cost per paper (from <$500 to >$1500) but nobody really
knows what the the true costs of open-access publishing should or will
be. Some think the cost per paper will be the same as now, but instead
of being paid by a set of subscribing institutions, it will be paid by
the author's institution. But if the author's institution is already
self-archiving the paper, it is not at all clear what service the
publisher will need to perform, over and above peer review and editing.

All these variables would have a chance to adjust themselves gradually
if the only thing mandated by the Sabo Bill were open access itself.
We would have immediate open-access, but none of the other risky and
unpredictable consquences of abruptly mandating that all US
funded-research papers must be put in the public domain.

> To my mind, the question really is whether either the authors or their
> employer actually do anything to avail themselves of the works' Public
> Domain status.  No one seems to have been able to answer this question.
> If they don't, why should the Sabo Bill's extension of identical status to
> Federally funded works, in itself, be expected to achieve anything for the
> Open Access agenda?

You are quite right that mandating public domain alone does not even
ensure that the research will be made open access! It only provides the
*possibility* of making it open access (which again boils down to
self-archiving, for publishers always had the possibility of becoming
open-access publishers!). Mandating open access instead -- by making
it a condition of research funding that all reulting refereed journal
articles must be open-access -- thereby mandates an *action* on the part
of the grant-recipient, namely, that he either publishes the research
in an open-access journal (if a suitable one exists -- 5% of journals
currently) or he publishes it in a conventional journal (95%), and also
self-archives it in his institutional open-access archive (or a central
one, such as PubMedCentral, where one exists).

Mandating public domain forces publishers to immediately accept
public-domain texts (and forces authors to make their texts public
domain) without any assurance that the texts will be made openly
accessible online by anyone.

Mandating open access forces the grant recipient and institution to
ensure that their own research output is openly accessible. For 5%
of research, this mandate can be fulfilled by publishing it in an
open-access journal. For the rest of the 95%, it will force publishers to
allow self-archiving by exactly the same token as it would have forced
them to accept public-domain content -- but at far less risk to both
publisher and author.

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):

    http://amsci-forum.amsci.org/archives/september98-forum.html
                            or
    http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Amsci/index.html

Discussion can be posted to: september98-forum at amsci-forum.amsci.org