[Redirected from "No Free Lunches" Thread.]
On Tue, 9 Oct 2001, Andrew Wray wrote:
> But this still doesn't address the point of stable financing of the
> essential peer review process. This problem of stability is the main
> message I took from John Ewing's arcticle.
>> Subscriptions are a financial firewall, author charges per page or
> per article might work for some authors but not all, and funding by
> governments, science funding bodies or universities seems likely to be
> unsustainable in the long term (>5-10 years).
I was too cryptic, but as it has been written out in longhand elsewhere,
just a brief preamble and then some quoted excerpts:
(1) The CHE Essay on which I was commenting
was about the PLoS boycott/petition enjoining publishers to give away
their full-text contents, and the likely economic consequences should
that petition succeed (the failure of the smaller, less expensive,
learned-society journals and their possible absorption by the larger,
more expensive, commercial publishers, who were better equipped
to survive a boycott or a temporary full-text give-away period).
(2) The essay argued that the "frills" (online enhancements for
indexing and search, such as reference-linking) might become
"essentials" and would justify the continuation of the financial
firewalls blocking free access.
(3) My comment was that this reasoning was based on "conflating the
essentials and the add-on options." Peer review is the only essential:
the refereed text, whether on-paper or on-line, plus all other
enhancements, such as reference-linking, are all frills, because authors
can give away their own refereed texts (in their own institutional
Eprint Archives) without the need for publishers to give away their
(4) Once it is clear that the only problem is the "stable financing of
the essential peer review process" (as you correctly state), it is also
clear that it is NOT a matter of page-charges: Publishers' pages,
whether on-line or on-paper are NOT essentials, they are frills (with
the advent of author/institution self-archiving of all refereed
research), and they can and should be sold and paid for as such, i.e.,
separately, not force-wrapped into the essentials.
(5) This de-conflation also makes it clear what funds will finance the
essential peer review process, if/when it is ever needed: The annual
institutional windfall savings from the cancelled financing of the
frills (if/when that ever occurs). No author page-charges or
government subsidies needed.
(6) But until then, author/institution self-archiving can proceed apace,
freeing access to the entire refereed research literature without
waiting for petitions, boycotts, or publisher give-aways.
4. The Subversive Proposal
4.1 Enough to free entire refereed corpus,
Eight steps will be described here. The first four are not
hypothetical in any way; they are guaranteed to free the entire
refereed research literature from its access/impact-barriers right
away. The only thing that researchers and their institutions need
to do is to take these first four steps. The second four steps are
hypothetical predictions, but nothing hinges on them: The refereed
literature will already be free for all as a result of steps i-iv,
irrespective of the outcome of predictions v-viii.
i. Universities install and register OAI-compliant Eprint Archives
The Eprints software is free and will be open-sourced. It in
turn uses only free software; it is quick and easy to
install and maintain; it is OAI-compliant and will be kept
compliant with every OAI upgrade:
http://www.openarchives.org/. Eprints Archives are all
interoperable with one another and can be (e.g.,
http://arc.cs.odu.edu/) of the entire research literature,
both pre- and post-refereeing.
ii. Authors self-archive their pre-refereeing preprints and
post-refereeing postprints in their own university's Eprint
This is the most important step; it is insufficient to
create the Eprint Archives. All researchers must
self-archive their papers therein if the literature is to be
freed of its access- and impact-barriers. Self-archiving is
quick and easy; it need only be done once per paper, and the
result is permanent, and permanently and automatically
uploadable to upgrades of the Eprint Archives and the
iii. Universities subsidize a first start-up wave of
self-archiving by proxy where needed.
Self-archiving is quick and easy, but there is no need for
it to be held back if any researcher feels too busy, tired,
old or otherwise unable to do it for himself: Library staff
or students can be paid to "self-archive" the first wave of
papers by proxy on their behalf. The cost will be negligibly
low per paper, and the benefits will be huge; moreover,
there will be no need for a second wave of help once the
palpable benefits (access and impact) of freeing the
literature begin to be felt by the research community.
Self-archiving will become second-nature to all researchers
once its benefits have become palpable.
iv. The Give-Away corpus is freed from all access/impact barriers
Once a critical mass of researchers has self-archived, the
refereed research literature is at last free of all access-
and impact-barriers, as it was always destined to be.
4.2 Hypothetical Sequel:
Steps i-iv are sufficient to free the refereed research literature.
We can also guess at what may happen after that, but these are
really just guesses. Nor does anything depend on their being
correct. For even if there is no change whatsoever -- even if
Universities continue to spend exactly the same amounts on their
S/L/P budgets as they do now -- the refereed literature will have
been freed of all access/impact barriers forever.
However, it is likely that there will be some changes as a
consequence of the freeing of the literature by author/institution
self-archiving. This is what those changes might be:
v. Users will prefer the free version?
It is likely that once a free, online version of the
refereed research literature is available, not only those
researchers who could not access it at all before, because
of S/L/P-barriers at their institution, but virtually all
researchers will prefer to use the free online versions.
Note that it is quite possible that there will always
continue to be a market for the S/L/P options (on-paper
version, publisher's on-line PDF, deluxe enhancements) even
though most users use the free versions. Nothing hangs on
vi. Publisher S/L/P revenues shrink, Library S/L/P savings grow?
But if researchers do prefer to use the free online
literature, it is possible that libraries may begin to
cancel journals, and as their S/L/P savings grow, journal
publisher S/L/P revenues will shrink. The extent of the
cancellation will depend on the extent to which there
remains a market for the S/L/P-based add-ons, and for how
If the S/L/P market stays large enough, nothing else need
vii. Publishers downsize to providers of QC/C service+ optional
It will depend entirely on the size of the remaining market
for the S/L/P options whether and to what extent journal
publishers will have to down-size to providing only the
essentials: The only essential, indispensable service is
viii. QC/C service costs funded by author-institution out of
reader-institution S/L/P savings?
If publishers can continue to cover costs and make a decent
profit from the S/L/P-based optional add-ons market, without
needing to down-size to QC/C provision alone, nothing much
But if publishers do need to abandon providing the S/L/P
products and to scale down instead to providing only the
QC/C service, then universities, having saved 100% of their
annual S/L/P budgets, will have plenty of annual windfall
savings from which to pay for their own researchers'
continuing (and essential) annual journal-submission QC/C
costs (10%); the rest of their savings (90%) they can spend
as they like (e.g., on books -- plus a bit for Eprint
So petitions, boycotts and publisher give-aways are unnecessary (apart
from being unlikely to succeed). And the only ostensible obstacle to
the author/institution self-archiving of all ~2M annual refereed papers
in all ~20K refereed journals (apart from author sluggishness!) is
author misperceptions about copyright restrictions, which can likewise
be completely circumvented, legally:
5. PostGutenberg Copyright Concerns
There is a great deal of concern about copyright in the digital
age, and some of it may not be easily resolvable (e.g., what to do
about the pirating of software and music). But none of that need
detain us here, because digital piracy is only a problem for
non-give-away work, whereas we are concerned here only with
give-away work. (Again, failing to make the give-away/non-give-away
distinction leads only to confusion, and the misapplication of the
much bigger and more representative non-give-away model to the
anomalous give-away corpus, which it does not fit.)
The following digital copyright concerns are relevant to the
non-give-away literature only:
5.1. Protecting Intellectual Property (royalties)
This is as much of a concern to authors of books as to
authors of screenplays, music, and computer programs. It is
also a concern to performers who have made digital audio or
video disks of their work. They do not wish to see that work
stolen; they want their fair share of the gate-receipts in
return for their talent and efforts in producing the work.
But the producers of refereed research reports do not wish
to have protection from "theft" of this kind; on the
contrary, they wish to encourage it. They have no royalties
to gain from preventing it; they have only research impact
to lose from access-blockage of any kind.
5.2. Allowing Fair Use (user issue)
"Fair Use" is another worthy concern. It has to do with
certain sanctioned uses of non-give-away material, such as
all or parts of books, magazine articles, etc., often for
teaching purposes; the producers of these works do not wish
to lose their potential royalty/fee-income from these
The producers of refereed research reports, in contrast,
wish to give their work away; hence fair-use issues are moot
for this special give-away literature.
5.3. Preventing Theft of Text (piracy)
The producers of refereed research reports do not wish to
prevent the theft of their texts; they wish to facilitate it
as much as possible. (In the on-paper era they used to
purchase and mail reprints to requesters at their own
The following digital copyright concern is relevant to all
literature, both give-away and non-give-away:
5.4. Preventing Theft of Authorship (plagiarism)
No author wants any other author to claim to have been the
author of his work. This concern is shared by all authors,
give-away and non-give-away. But it has nothing whatsoever
to do with concerns about theft-of-text, and should not be
conflated with such concerns in any way: Give-away work
need not be held hostage to non-give-away concerns about
theft-of-text under the pretext of "protecting" it from
theft-of-authorship. (Unfortunately, many journal publishers
try to write and use their copyright transfer agreements for
precisely this purpose, and authors need to become aware of
The following digital copyright concern is relevant to the
give-away literature only:
5.5. Guaranteeing Author Give-Away Rights
Apart from the protection from plagiarism and the assurance
of priority that all authors seek, the only other
"protection" the give-away author of refereed research
reports seeks is protection of his give-away rights!
(The intuitive model for this is advertisements: what
advertiser wants to lose his right to give away his ads for
free, diminishing their potential impact by charging for
access to them!)
Well, there is no need for the authors of refereed research
to worry about exercising their give-away rights, for they
can do it, legally, even under the most restrictive
copyright agreement, by using the following strategy.
6. How to get around restrictive copyright legally so
as to self-archive your own give-away refereed research
6.1. Self-archive the pre-refereeing preprint
Self-archiving the preprint is the critical first step. Before it
has even been submitted to a journal, your intellectual property is
your own, and not bound by any future copyright transfer agreement.
So archive the preprints (as physicists have done for 10 years now,
with over 130,000 papers, and cognitive scientists have done for 3
years now, with over 1000 papers). This is a good way to establish
priority, elicit informal feedback, and keep a public record of the
embryology of knowledge.
[Note that some journals have, apart from copyright policies, which
are a legal matter,"embargo policies," which are merely policy
matters (nonlegal). Invoking the " Ingelfinger (Embargo) Rule,"
some journals state that they will not referee (let alone publish)
papers that have previously been "made public" in any way, whether
through conferences, press releases, or on-line self-archiving. The
Ingelfinger Rule, apart from being directly at odds with the
interests of research and researchers and having no intrinsic
justification whatsoever -- other than as a way of protecting
journals' current revenue streams -- is not a legal matter, and
unenforceable. So researchers are best advised to ignore it
completely (Harnad 2000a, 2000b), exactly as the authors of the
130,000 papers in the Physics Archive have been doing for 10 years
now. The "Ingelfinger Rule" is under review by journals in any
case; Nature has already dropped it, and there are indications that
Science may soon follow suit too.]
6.2. Submit the preprint for refereeing (revise etc.)
Nothing changes in author publication practises; nothing needs to
be given up. Submit your preprint to the refereed journal of your
choice, and revise it as usual in accordance with the directive of
the Editor and the advice of the referees.
6.3. At acceptance, try to fix the copyright transfer agreement to
Copyright transfer agreements take many forms. Whatever the wording
is, if it does not explicitly permit online self-archiving, modify
it so that it does. Here is a sample way to word it
I hereby transfer to [publisher or journal] all rights to sell
or lease the text (on-paper and on-line) of my paper
[paper-title]. I retain only the right to distribute it for
free for scholarly/scientific purposes, in particular, the
right to self-archive it publicly online on the Web.
Some publishers (about 10%) already explicitly allow self-archiving
of the refereed postprint (e.g., the American Physical Society:
ftp://aps.org/pub/jrnls/copy_trnsfr.asc). Most other publishers
(perhaps 70%) will also accept this clause, but only if you
explicitly propose it yourself (they will not formulate it on
their own initiative).
6.4. If 6.3 is successful, self-archive the refereed postprint
Hence, for about 80% of journals, once you have done the above, you
can go ahead and self-archive your paper.
Some journals (about 20%), however, will respond that they decline
to publish your paper unless you sign their copyright transfer
agreement verbatim. In such cases, sign their agreement and proceed
to the next step:
6.5. If 6.3 is unsuccessful, archive the"corrigenda"
Your pre-refereeing preprint has already been self-archived since
prior to submission, and is not covered by the copyright agreement,
which pertains to the revised final ("value-added") draft. Hence
all you need to do is to self-archive a further file, linked to the
archived preprint, which simply lists the corrections that the
reader may wish to make in order to conform the preprint to the
refereed, accepted version.
Everyone chuckles at this point, but the reason it is so easy is
that this is the author give-away literature. No non-give-away
author would ever dream of doing such a thing (archiving the
prepublication draft for free, along with the corrigenda). And
copyright agreements (and copyright law) are designed and conceived
to meet the much more representative interests of non-give-away
authors and their much larger body of royalty/fee-based work. Hence
this simple and legal expedient for the special, tiny, anomalous,
give-away literature has no constituency anywhere else.
Yet this simple, risible strategy is also feasible, and legal
(Oppenheim 2001) -- and sufficient to free the entire current
refereed corpus of all access/impact barriers immediately!
So, to repeat, the CHE Essay was based on conflating the problem of
stably financing the essentials with the nonproblem of how to
continue funding the optional frills:
3.1. S/L/P [Subscription/Site-License/Pay-Per-View]:
S/L/P tolls are the access-barriers, hence the impact-barriers, for
researchers and their give-away research. S/L/P is the journal
publisher's means of recovering costs and making a fair profit.
High costs were inescapable in the expensive and inefficient
on-paper Gutenberg era; but today, in the on-line PostGutenberg
era, continuing to do it all the old Gutenberg way, with its high
costs, must be clearly seen as the optional add-on (for this
give-away literature only: not for the royalty/fee-based
literature!) that it has become, rather than as the obligatory
feature it used to be.
Beware of the language of obligatory "value-added," with which the
peer-reviewed literature must, by implication, continue to be
inextricably wrapped. The only essential service still provided by
journal publishers (for this anomalous, author-give-away literature
in the PostGutenberg era) is peer review itself.
The rest -- on-paper versions, PDF on-line page images, deluxe
online enhancements -- are all potentially valuable features, to be
sure, but only as take-it-or-leave-it options. In the on-line era
there is no longer any necessity, hence no longer any justification
whatsoever, for continuing to hold the refereed research itself
hostage to S/L/P tolls and whatever add-ons they happen to pay
Beware also of any attempt to trade off S for L or L for P: Pick
your poison, all three are access-barriers, hence impact-barriers,
and hence all three must go -- or rather, they must all now become
only the price-tags for the add-on, deluxe options that they buy
for the researcher and his institution, but no longer also for the
peer-reviewed essentials, which can now be self-archived for free
3.2. QC/C [Quality-Control & Certification]: peer review
Peer review itself is not a deluxe add-on for research and
researchers: This quality-control service and its certification
(QC/C) is an essential (Harnad 1998/2000). Without QC/C, the
research literature would be neither reliable nor navigable, its
quality uncontrolled, unfiltered, un-sign-posted, unknown,
But the peers who review it for the journals are the researchers
themselves, and they review it for free, just as the researchers
report it for free. So it must be made quite clear that the only
real QC/C cost is that of implementing the peer review, not
actually performing it.
Estimates (e.g., Odlyzko 1998) as well as the real experience of
online-only journals (e.g., Journal of High Energy Physics
http://www.cogsci.soton.ac.uk/psycoloquy/) have shown that the QC/C
implementation cost is quite low -- about 10% of the total amount
that the world's institutional libraries (or rather, the small
subset of them that can afford any given journal at all!) are
currently paying annually per article in S/L/P tolls .
Once the 90% S/L/P add-ons become optional, the essential 10% QC/C
cost could easily be paid out of the 100% S/L/P savings -- if ever
the world's libraries decide they no longer need the add-ons. (The
other 90% savings can be used to buy other things, e.g., books,
which are not, and never will be, author give-aways.)
3.3. Separating (i) QC/C service-provision from (ii) eprint
access-provision (and from (iii) optional add-ons)
Researchers need not and should not wait until journal publishers
voluntarily decide to separate the provision of the essential QC/C
service from all the other optional add-on products (on-paper
version, publisher's PDF version, deluxe enhancements) before their
give-away refereed research can at last be freed of all access- and
All researchers can free their own refereed research now, virtually
overnight, by taking the matter into their own hands; they can
self-archive it in their institutional Eprint Archives:
http://www.eprints.org. Access to the eprints of their refereed
research is then immediately freed of all S/L/P barriers, forever.
3.4. Interoperability: The Open Archive initiative (Oai)
Papers self-archived by their authors in their institutional Eprint
Archives can be accessed by anyone, anywhere, with no need to know
their actual location, because all Eprints Archives are compliant
with the Open Archives Initiative (OAI) meta-data tagging protocol
for interoperability: http://www.openarchives.org
Because of their OAI-compliance, the papers in all Eprints Archives
can be harvested and searched by Open Archive Services such as the
Cross Archive Searching Service http://arc.cs.odu.edu/, providing
seamless access to all the eprints, across all the Eprint Archives,
as if they were all in one global, virtual archive.
[Excerpt from: Harnad, S. (2001) For Whom the Gate Tolls? How and
Why to Free the Refereed Research Literature Online Through
Author/Institution Self-Archiving, Now.
Stevan Harnad harnad at cogsci.soton.ac.uk
Professor of Cognitive Science harnad at princeton.edu
Department of Electronics and phone: +44 23-80 592-582
Computer Science fax: +44 23-80 592-865
University of Southampton http://www.cogsci.soton.ac.uk/~harnad/
Highfield, Southampton http://www.princeton.edu/~harnad/
SO17 1BJ UNITED KINGDOM
NOTE: A complete archive of the ongoing discussion of providing free
access to the refereed journal literature online is available at the
American Scientist September Forum (98 & 99 & 00 & 01):
You may join the list at the amsci site.
Discussion can be posted to:
september98-forum at amsci-forum.amsci.org