Re: Napster: stealing another's vs. giving away one's own

From: Stevan Harnad <harnad_at_cogprints.soton.ac.uk>
Date: Sun, 3 Mar 2002 03:36:44 +0000

I don't know if it was the author, Stephen Strauss, or newspapers'
quaint habit of having uninformed copy-writers coin titles in complete
ignorance or disregard of content, with an eye only to catching eyes,
but the Globe & Mail has an article with an even more unfortunate title
than the original title of Sam Vaknin's 2-part UPI Story, which had
been "Copyright-free online scholarship" but was later sensibly changed
to "Copyright and Scholarship" in response to our request. No second
chances here:

[Note to Stephen Strauss: Stephen, it wouldn't have taken nearly as
many words as this to fix your story if you had let me see it in
advance! A few changes and additions here and there and it would
have been just fine...]

> Napster for scientists?
>
> http://makeashorterlink.com/?S22820B7
>
> The Internet is disrupting the voodoo economics of scholarly journals
> with an information revolution, STEPHEN STRAUSS reports. The
> question is if free knowledge is always the best knowledge
>
> By STEPHEN STRAUSS
>
> Saturday, March 2, 2002 Print Edition, Page F6
> http://www.theglobeandmail.com/
> Copyright 2002 Bell Globemedia Interactive Inc. All Rights Reserved.
>
> The ever-subversive economics of the Internet have recently
> spawned an unlikely question: What do a million teenagers
> listening to an illicitly downloaded version of Britney Spears
> trilling I'm Not a Girl, Not Yet a Woman have to do with a few
> biologists engrossed in an article entitled "A Cytosolic ADP
> Glucose Pyrophosphorylase is a Feature of Graminaceous
> Endosperms, but Not of Other Starch-Storing Organs?"

No problem with the opening paragraph, which is a good rhetorical way
of putting it -- in fact, one that I myself have found useful to bring
out the paradox -- but without having to compete with a title that
implies the exact opposite:

    Harnad, S. (2001) The (Refereed) Literature-Liberation Movement. The
    New Scientist. http://www.ecs.soton.ac.uk/~harnad/Tp/newsci2.htm
    http:/ /www.ecs.soton.ac.uk/~harnad/Tp/newsci2.htm

    "Napster: stealing another's vs. giving away one's own"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0671.html

> The answer is that they represent two very different faces of
> one of the modern world's greatest paradoxes: If you can get
> something for nothing, how much is it worth?

That is of course NOT the paradox at all! The paradox is that the kind
of creative work you steal with napster, its creators DO NOT want
you to steal, yet you can; whereas with peer-reviewed research,
its creators DO want you to "steal" it, yet you cannot! Napster is
consumer rip-off, whereas self-archiving is creator give-away. Napster
is against the law, whereas access-denial for peer-reviewed research is
merely against the interests of researchers, research progress, and the
best interests of society.

Does Stephen Strauss manage to resolve the parallel paradox created
by equating the two instead of contrasting them?

> Stopping the "free" appropriation of songs by every kid armed
> with a CD burner has become the holy grail of the world's
> music industry. At the same moment, the foundation set up by
> billionaire Hungarian-American philanthropist George Soros
> announced last week that it is contributing $3-million (U.S.)
> aimed at "completely free and unrestricted access to [scientific
> information] by all scientists, scholars, teachers, students, and
> other curious minds."

Soros is doing this by supporting two perfectly legal open-access
strategies, both author give-aways, rather than consumer rip-offs like
Napster: (1) public self-archiving of peer-reviewed papers by their
own authors and/or (2) the foundation of open-access journals.

Stephen Strauss has likened Soros's philanthropy to napster-style
theft, in that they both provide free access to a product that would
otherwise be behind a financial firewall, without ever troubling to
count the ways in which they are unlike, viz:

    one is taking away another's work, the other is giving away one's
    own work;

    one is illegal the other is legal;

    one is done against the wishes of the work's creators, one is
    done according to the wishes of the work's creators;

    one is against the interests of productivity and creativity, the
    other is in the interests of productivity and creativity;

    one is a piracy, the other a service to society.

> To do this, the foundation admits, it will have to undermine a
> system of scientific publishing whose parameters have remained
> roughly the same since the Royal Society of London created the
> first scientific journal in 1665.

This is a bit apocalyptic, and the language of "undermining" is
far too shrill, but it could still be used in the service of
dispelling the confusion created by the napster koan: by going on
to enlighten the reader -- if only Stephen Strauss actually went on to
do so, but read on...:

> For nearly 350 years, scientists have done their research, written
> up their results and then sent a paper off to a scholarly journal.
> The journal, in turn, sent the article off to other scholars, who
> freely judged its validity and suggested changes. When this peer
> review was satisfactorily completed, the paper was published in
> a journal -- which the researcher's library, and all other libraries,
> then had to buy.

This is correct. But now what has happened to change this?

> There's a paradox here quite different than the music-industry
> situation, says psychologist Stevan Harnad: "I give away what I
> have done for free, and then I have to pay to buy it back when
> it is published." In the past decade, Harnad, who works at the
> University of Quebec at Montreal, has become one of the
> intellectual bulldozers pushing for a scientific-publishing
> revolution,

I wish I were a bull-dozer! If I had been, the access-barriers to the
20,000 peer-reviewed research journals would have been long gone by
now! (And the revolution is not mine, but the Net's.)

But I am fairly careful about what I say on this score. And although
the institutional libraries, reeling from their serials crises, are
fond of putting it in the above way, I believe I pointed out to Stephen
Strauss that the problem is not that I or my institution have to buy
back OUR OWN give-away research (why would we want to do that? We have
it already!). It is that my institution has to buy in all the OTHER
researchers' give-away research, and they ours, whereas we would all
dearly love to access one another's give-away work, and have our own
give-away work accessed, for free. That's the reciprocity guaranteed
by opening access to one's own give-away work + the Golden Rule.

Notice that I keep inserting "give-away." I need to, in order to
immunize against the still unresolved "napster" analogy.

What should have been mentioned at this point, before segueing to
Elsevier below, is that what has brought this 350-year era to an end is
the specific new capability that the PostGutenberg technology (networked
digital storage and retrieval) has ushered in for this special, tiny,
give-away literature that never sat comfortably with the
access-restrictions dictated by the Gutenberg technology and economy
in the first place: Throughout all those 350 years, the toll-barriers
blocking access to their give-away work were always Faustian Bargain
for these anomalous authors, tolerated because there was no way to make
their findings PUBLIC at all unless the high Gutenberg PUBLICation
costs were allowed to be recovered through access-tolls.

> Publishing what they didn't pay for is often an extremely
> profitable exercise for companies. Elsevier, the large Dutch
> publishing house that publishes roughly 20 per cent of the
> peer-reviewed scientific publications in the world, reported a
> 33-per-cent profit last year on its scientific journal operation.
> Part of the reason is that subscription prices can be enormous.

This is correct, and I rather like the subtle implication that
peer-reviewed journal publishers, unlike book publishers (and
magazine/newspaper publishers -- or, for that matter, the music
industry) are indeed selling work for which they have not paid a
penny to its authors, in either royalty, salary or fees. (Stephen
Strauss should have kept in mind the ironic contrast with the
economics of his own articles, which are not author give-aways!)

Yet peer-reviewed journal publishers are not guilty of napster-like
piracy in so doing either, for they have added value to the work, in
the form of paying for the peer review, the editing, the mark-up, and
the production and distribution of the on-paper and on-line PDF product.
So the fact that Elsevier's profit margins in so doing happen to be
quite large is not, in and of itself, a sin, and certainly no crime.
So the paradox is still unresolved.

> Each monthly copy of Elsevier's journal Adverse Reactions, for
> example, costs a subscriber $1,000 (U.S.). For libraries,
> individual journal prices since 1986 have increased on average
> by 226 per cent. As a consequence, research libraries have cut
> back their journal purchases by an average of 7 per cent.

True, and familiar. But if their prices were lower, would that change
the fact that the works they are selling are author give-aways that
their anomalous authors (rather more like advertisers) would much
rather see openly accessible, toll-free, to any and every would-be
user, rather than sitting behind financial firewalls that many (in
fact most) would-be users' institutions would not be able to scale no
matter how low they were?

There are 20,000 peer-reviewed journals published annually
http://www.ulrichsweb.com/ulrichsweb/, and even the richest research
library, such as Harvard, cannot afford more than a minor proportion of
them http://fisher.lib.virginia.edu/cgi-local/arlbin/arl.cgi?task=setupstats
How low would the price have to be so that all research institutions, all over the
planet, could afford all of it?

The author-researcher-advertisers have known very well how low it would
have to be, in making their works give-aways: Zero. Open access is the
only way to resolve the paradox. But then how to recover costs?

> Harnad and others agree that this is the way things had to be
> when scientists were forced to rely on third parties with printing
> presses to communicate with other scientists. "In the Gutenberg
> era, there was no way around the real costs of printing and
> circulating on paper," he says.
>
> Many electric journals came into being when irritated scientists
> resigned en masse from the editorial boards of for-profit
> publications and started up non-profit competitors.

This unfortunately does not answer the question, which is about open
access, i.e., NO charge to the user-institution. Electronic-only costs
are lower, but not zero.

Lower prices do not resolve any paradoxes. Everyone wants lower prices,
for every kind of commodity. Producers are also happy to lower prices
in exchange for more sales. But the issue here was open access,
give-aways, NO price. How to resolve it? Otherwise we are left
believing that napster-like consumer-theft would be the only answer!

> Another approach, and one that has already been extensively used in
> physics and astronomy, is to publish "pre-prints." In this case,
> papers that have been peer-reviewed but not yet published
> elsewhere are posted on a server site.

Alas, this is just plain wrong! Preprints are the opposite of this. They
are papers that have NOT yet been peer-reviewed or published. What
physicists did was to self-archive (not self-publish!) these unrefereed,
unpublished preprints on a server site.

But it is not too far wrong, because in addition to self-archiving
their unrefereed, unpublished preprints, those physicists, 8-12 months
later, after peer review and revision was completed, also self-archived
the peer-reviewed, published postprints of those same papers
http://opcit.eprints.org/tdb198/opcit/embryology/
(when the difference was big enough to make it worth bothering --
otherwise they merely updated the preprint's reference, to include the
publication information for citation purposes:
http://opcit.eprints.org/tdb198/opcit/q2/).

But in getting this seemingly minor point wrong, Stephen Strauss will
later go on to compound the error, as we shall see, at the cost of
compounding the paradox instead of resolving it.

(Even more mysterious to me than the ceding of journalists' prerogative
to decide the title of their own articles to mindless copy-writers is their
extreme reticence about checking their texts with their sources, even
when time permits: This can't be because they are afraid their own
sources may scoop them. Nor is merely checking their drafts with their
sources tantamount to granting them vetting or veto power. This
journalistic casualness about accuracy [except where there is a risk of
libel or litigation] is an utter mystery to me, but it invariably leads
to needless, and sometimes serious and misleading errors. It's also one
of the basis differences between popular press journals and peer
reviewed journals.)

> University of Minnesota
> mathematician Andrew Odlyzko has argued that conventional
> publishing of the 20,000 papers one such site stores a year
> would require an investment of $40-million to $80-million. The
> electronic pre-prints do the same thing for $100,000.

Those are merely archiving costs, and to all intents and purposes --
especially if distributed across the archives of the universities that
produced the research, instead of being concentrated in one central
archive like the Physics archive -- would be essentially negligible per
paper, a small piece of the online infrastructure that universities all
standardly offer their faculty already.

But the trouble is that those archived eprints also include postprints
-- the "value-added" contribution of peer review. Not the refereeing
itself, for the peers review for free, just as the peer-reviewed
authors give away their papers for free. But the peer-review process
has to be administered, implemented, even when it is all accomplished
electronically. And that can cost $200-$500 per paper, and has nothing
to do with the negligible archiving costs per paper mentioned above!

Stephen Strauss has here unknowingly stumbled into a very controversial
area, but one that is in reality independent of the question of free
access, hence a red herring: Should we free peer-reviewed research not
only from access-tolls, but also from peer review itself? (If Stephen
had contacted me with his draft, I would have pointed out to him the
controversy and confusion he was courting if he did not steer clear of
this side-issue.)

For there is nothing that those who would like to defend preserving
toll-based access to this give-away literature would like to hear more
than that a toll-free literature = a peer-review-free literature! If
that were really the choice, then Elsevier would say (and I would side
with them!): "Take your pick between the quality-controlled literature
you purchase from us now, paying for the added value, and the
anything-goes literature you would have if there were no peer review."

In fact, look below, and you will see that that is precisely what
Jeffrey Drazen, editor-in-chief of the New England Journal of
Medicine, is put in the convenient position of saying, because of
the gratuitous napster analogy, and its implication of somehow not
"playing by the rules"!

(Note that "preprints" are not a sample of what the literature would be
like if there were no longer peer review, for, as noted, all those
preprints are destined for peer review, and re-appear as postprints:
http://www.nature.com/nature/webmatters/invisible/invisible.html)

This is very controversial stuff, however, and not the way to resolve
paradoxes, but rather the way to compound confusion. The Budapest Open
Access Initiative (BOAI) http://www.soros.org/openaccess is about
freeing access to the peer-reviewed literature, such as it is, now, not
about access to some hypothetical, future, unfiltered literature. And
the fact is that implementing peer review can cost up to $500 per
paper. Who's to pay the piper if access to this literature is freed?

> Finally, there is self-archiving: Scientists simply put versions of
> their work on their own Web sites and notify colleagues who
> want to see them that they exist.

This is not ANOTHER approach. Self-archiving is exactly the same as
what was already mentioned as the previous approach. It just has two
stages: the self-archiving of the unrefereed preprint, and then (8-12
months later) the self-archiving of the peer-reviewed postprint.

A minor detail is that "self-archiving" as I have been advocating it
happens to favor distributed, institutional eprint Archives
(for various specific strategic reasons:
http://www.ecs.soton.ac.uk/~harnad/selfaq.htm#institutional-vs-central).
But in central eprint Archives like the one in Physics, it is likewise
the authors who are self-archiving their own preprints and postprints!

> "For the most part, scholars aren't paid for what they do and
> don't expect to be," Odlyzko says. However, it has become
> apparent in recent years that Internet publishing isn't actually so
> much free as free-ish. When computer time, administrative work
> and other costs are factored in, it could cost anywhere between
> $200 and $500 per academic paper. Who pays for that? The
> free electronic publisher Pubmed Central, for one, has tried
> asking the scientists to pay up front as well as asking libraries to
> pay an annual fee.

Here is a golden opportunity, lost, to decisively resolve the paradox of
how the peer-reviewed literature is to be made accessible to all users
toll-free (without anything faintly napsterlike): The last sentence
above contains the contradiction that has to be resolved: Remember what
was wrong with the librarians' "buyback" formula earlier? It is that
the incoming articles that they pay for as the reader-institution are
not the same ones as their outgoing articles, as the author-institution!

Ask yourself what, exactly, a university library would be committing
itself to paying for if it agreed to "subscribe" to a publisher's
output by agreeing to pay an annual fee? (Something was garbled above,
by the way, as Pubmed Central is not a publisher.)

Again, there are 20,000 journals, most of which do not contain any of
that particular university's authors' outgoing papers, although many of
them may contain papers that those authors would like to read. We are
back to the toll-based access model again! And that model will not
scale up to free access to the full 20,000 no matter how low the tolls,
for one simple reason (also known as the prisoner's dilemma): If the
journals are really open-access journals, it's much more sensible to
opt out of the annual fee and get a free ride than to contribute!

So library access tolls for free access make no sense: If they are
mandatory, the access is not free (and we are just back to lower
prices: "free-ish" is a weasel-word); if they are optional, it makes
much more sense not to pay.

[The logic of the notion of something's being "free-ish" calls to mind
two old jokes, one about being "a little bit pregant(-ish)," the other,
especially apt, Shaw's: "Madame, we have already established your
profession. We are merely haggling over the price."
http://www.elise.com/quotes/shawquotes.htm ]

The other option, which is that the authors themselves pay the costs
per paper looks, on the face of it, as if it is adding insult to injury:
These anomalous authors already give away their papers for free; they
already do peer-review for free. Are they now expected to PAY for that
privilege?

Yet authors, rather than readers, do look to be the right "client" for
the peer-review service, which, after all, is a cost per paper
submitted, not per paper subscribed to. The answer is that it is
authors' INSTITUTIONS that are the critical link, and the natural
source for the requisite peer-review service charges:
http://www.ecs.soton.ac.uk/~harnad/selfaq.htm#institutional-vs-central

It remains only to complete Andrew Odlyzko's arithmetic: Peer-review
costs $200-$500 per paper, but the planet (or rather, that small portion
of its research libraries that can afford to subscribe to any given
one of the 20,000 journals) collectively pays $2000-$5000 for access
per article. This makes it obvious to anyone who can count that if
all the reader-institutions got back their respective portions of
the collective $2000-$5000 subscription tolls they are currently paying
annually for the subset of the 20,000 that they can afford, then they
would have abundant annual windfall savings from which to redirect the
10%-30% that would be needed to pay for the peer-review service on all
of their own outgoing papers. The only thing that needs to be noted to
see this is that the reader-institution = the author-institution.

This is the fair, sensible, PostGutenberg way to pay for the essential
service of peer review, per paper published, thereby guaranteeing free
access to it for all users, everywhere. And there is not a trace of
napster-like theft in it!

The other expenses (print-on-paper, publisher's PDF page-images,
any further publisher enhancements) could continue to be paid for as
options, as long as there was still a market for them. But what must
end now is forcibly bundling them in with the essential peer-review
costs as an excuse for continuing to hold the peer-reviewed draft
hostage to Gutenberg access-tolls. The only way to test whether those
add-ons are still valued in the PostGutenberg era is to unbundle them,
try to sell them as separates, and see whether they still have any
market value even when there is open access to the vanilla
peer-reviewed version.

> Scientific publishers, unlike the music industry, have started to
> embrace the free-ish revolution. Several announced in January
> that they would provide free access to biomedical literature in
> their journals to people living in the world's 70 poorest
> countries. And many allow the general public on-line access to
> all papers once they are a year or so old.

The BOAI is not a "free-ish" revolution; it is an open access
initiative (open access to give-away peer-reviewed research). Subsidies
for the biomedical literature to the poorest countries are welcome, but
something else. They do not solve the basic problem, which is freeing
access to all 20,000 for all researchers, everywhere. (The
http://fisher.lib.virginia.edu/cgi-local/arlbin/arl.cgi?task=setupstats
site shows that there are far more Have-Nots than Harvards in the
developed world too!)

And, thank you very much, it is not a solution for the researcher to
have his give-away findings embargoed for 12 months before becoming
freely accessible to all their would-be users! That is not how research
progresses, nor is it why it is published, which is in order to make
the peer-reviewed findings PUBLIC, not in order to keep subsidizing a
an access-blocking cost-recovery model long after it has ceased to be
necessary, simply because it had been the only possibility in the
Gutenberg era.

> Yet there are those who say while it may be less free, the old
> model of publishing may always be better.
>
> Jeffrey Drazen, editor-in-chief of the New England Journal of
> Medicine, says that what doctors get when they subscribe to his
> journal is judgment, which is vital because what they read
> quickly becomes medical treatment. The NEJM is very picky.
> Of the 3,500 hundred articles the journal receives a year, only
> 400 or 500 are published -- often with qualifiers that the authors
> would rather leave out or play down.

This is all just a fancy way of referring to peer review! Peer review
is peer review, whether it is practised in biomedicine, physics, or
philosophy, whether it is protecting public health or the other
important quality standards of technology, science and scholarship.

And it costs $200-$500 per paper, whether that paper is in the New
England Journal of Medicine (NEJM) or Plant Physiology (which is where
the esoteric "Graminaceous Endosperms" paper Stephen Strauss cited
appeared: I found it with Biosis. If it had been self-archived, I
could have given you the URL for the full text.)

With its stated 1% acceptance rate, NEJM is up there with Science
and Nature among the most selective, highest-quality, highest-impact
journals. But that does not make them different in any other respect.

NEJM's referees, like all referees, do NEJM's peer reviewing for
them for free. NEJM's authors give them their papers for free. (Though
NEJM is so prestigious, many authors would no doubt love to buy their
way in! But that's not the way it works: You can't buy off peers who
are reviewing for free. And if you ever could, that would simply amount
to giving up the very quality and selectivity that makes NEJM so
prestigious, influential, and desirable for authors!)

All NEJM will have to do in the open-access era is to continue to
maintain its high peer-review standards. This is purely a gate-keeping
matter; it has no connection whatsoever with toll-gating the access to
the outcome:

    "Conflating Gate-Keeping with Toll-Gating"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1118.html

> "What a journal such as ours has to offer is the imprimatur of a
> dispassionate evaluation of the veracity and importance of the
> information. That is what people are paying for," Drazen says.

Aka, peer review. So what else is new?

> "If they want their paper published in the New England Journal
> there are certain rules they have to follow. If they don't, they
> don't have to publish in us."

The rules are those dictated by the standards of peer review; they
have nothing whatsoever to do with the obsolete and indefensible
Gutenberg cost-recovery model.

But casting the Budapest Open Access Initiative (for which Stephen
Strauss omitted to provide the URL:
http://www.openaccess.org/openaccess) as if it were Napster-like piracy
and crime certainly provides a needless opening for this familiar,
specious equation of access-tolls with quality by the defenders of
the no longer defensible Gutenberg-era revenue-streams and modera
operandi.

(For more on playing by the "Ingelfinger rule" -- which has nothing to
do with copyright, law, or napster, by the way -- see below.)

Stevan Harnad

    "Copyright, Embargo, and the Ingelfinger Rule"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1118.html

Harnad, S. (2000) Ingelfinger Over-Ruled: The Role of the Web in the
Future of Refereed Medical Journal Publishing. Lancet Perspectives 256
(December Supplement): s16.
http://www.ecs.soton.ac.uk/~harnad/Papers/Harnad/harnad00.lancet.htm

Harnad, S. (2000) E-Knowledge: Freeing the Refereed Journal Corpus
Online. Computer Law & Security Report 16(2) 78-87. [Rebuttal to Bloom
Editorial in Science and Relman Editorial in New England Journal of
Medicine]
http://www.ecs.soton.ac.uk/~harnad/Papers/Harnad/harnad00.scinejm.htm
Received on Sun Mar 03 2002 - 03:40:19 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:27 GMT