Re: Elsevier Science Policy on Public Web Archiving Needs Re-Thinking

From: Stevan Harnad <>
Date: Fri, 25 Sep 1998 21:16:58 +0100

Arthur Smith's apologia for Elsevier copyright policy can be completely
understood (and rejected) as being based on a false opposition:
Researchers have a choice between (1) a toll-based refereed literature
they can trust, or (2) a toll-free unrefereed literature they cannot
trust. The truth is that researchers can now have BOTH literatures
toll-free (though not entirely cost-free), and they can be trusted to
know the difference between the two, as they always have.

Arthur Smith writes:

> It is clearly in Elsevier's best interest in retaining the journal
> itself as the authoritative "final" version of papers - it is possibly
> also in the best interest of general support for peer review. It is not
> in the best interest of the immediate researchers who use these
> articles, I agree.

I think this speaks volumes: In the reporting of research, whose
interests are paramount?

"Peer review," by the way, is not an interested party; it is a service
performed by referees and editors. Rigorous, reliable, authoritative
peer review is in the interests of research and researchers but that
has nothing whatsoever to do with the specific access and toll issues
under discussion here.

This red herring keeps getting coughed up because of a completely
unnecessary (though habitual) linkage that is being implicitly made
between peer review and S/SL/PPV (Subscription/Site-License/Pay-Per-View).

The only relevant point that is true is that peer review cannot be
implemented for free, but see the 30/70 discussion thread on for a
discussion of how much peer review really costs, and of sensible ways
to recover that cost without toll-gates or fire-walls to block access.

> (1) The pre-print. This could exist in a universe where the publisher
> (and peer review) does not exist, and the value seen by all researchers
> is 4.072. Is this "optimal"?

When readers consult the unrefereed preprint literature, they know very
well that caution is in order. There is nothing new about this. Paper
preprints and tech reports have been in circulation for decades. It is
hence nonsensical to imply that special risks of some kind (what kind?) are
entailed by electronic preprints.

A well-tagged public electronic serial corpus will have a +REFEREED and
-REFEREED sector (with journal brand-names specified). The peer
community can continue to be trusted to exercise their judgment in
distinguishing between the two.

Here is a substantive question, though: Can the AUTHOR be trusted to
tag his own papers as +REFEREED and -REFEREED, specifiying journal-name
(or can we only trust the journal publisher to do so)?

There can be debate about this (no one knows for sure), but I'm
inclined to say yes to the former, for the time being, and that there
is no need at all for this matter to detain us along the road to the
optimal and inevitable (the entire serial literature, pre- and
post-refereeing, online for free for all).

> (2) The "final article". This would almost certainly not exist without
> the publisher.

Correct. And the publisher is entitled to recover necessary costs and a
fair return. But how much are those necessary costs for a refereed
online-only journal? And what is the optimal way to recover them: With
or without access tolls?

Now note that I said online-only. The rest is all about how to get
there from here. The course of action favoured by 24,000 authors per
year and 65,000 readers per weekday at
appears to be that researchers should archive their unrefereed
preprints and refereed reprints now, rather than waiting for the
publishers to sort out a way to make the transition, somehow, someday.
(It might even hasten them along the way: Necessity is the Mother of
Invention, and there's nothing like the wake-up call from serials
cancellations as the user community shifts its preferences to xxx
to stimulate resourcefulness among forward-looking publishers.)

Should research and researchers wait?

> And don't argue that the "referee" would see all this on
> the author's web site and fix it - first of all, Elsevier is ALREADY
> allowing the author to post it on his web site - why has not the
> referee or somebody else come and told the author about the Starrken
> correction already?

Straw man. No one here has proposed doing away with refereeing or
replacing it with a free-for-all. Yes, informal feedback on preprints
is welcome and helpful, but it is no substitute for peer review.

Harnad, S. (1996) Implementing Peer Review on the Net: Scientific
Quality Control in Scholarly Electronic Journals. In: Peek, R. &
Newby, G. (Eds.) Scholarly Publishing: The Electronic Frontier.
Cambridge MA: MIT Press. Pp. 103-118.

Harnad, S. (1997) Learned Inquiry and the Net: The Role of Peer Review,
Peer Commentary and Copyright. Antiquity 71: 1042-1048.

> Most people seeing a paper on the web "missing the
> Starrken correction" would either say "stupid author, won't trust any
> of his papers any more" or would send the author a note which the
> author would have no reason to trust or obligation to believe.

Irrelevant hypothetical sociometry. "Caveat emptor" will continue to
prevail when researchers consult the unrefereed literature, as it always
did; and informal peer feedback will continue to be weighted
differently from formal referee reports, as it always was. There is
nothing new here, or relevant to the issues at hand.

> The publisher sets up a trust relationship where the author is obligated to
> listen to and satisfy the selected expert referee, or at least the
> editor overseeing the relationship. This is what peer review is all
> about, and is what the majority of irreducible publishing expenses boil
> down to.

You mean the fortune currently being spent on S/SL/PPV is being spent
for "trust"? Well, forgive me if that revelation does not inspire too
much of the commodity in question!

You are simply ignoring what has already been conceded and repeatedly
confirmed in the 70/30 thread: Recalculated for online-only, the
"irreducible publishing expenses" boil down to at most 1/3 of what they
are now per page.

> It is what the publisher is investing its resources in. Why
> should the publisher give away the product of these investments, when
> they are accessible to most researchers through their libraries, or for
> a small fee directly?

Again, a false premise and a false opposition: The publisher of course
has the right (and the responsibility) to recover this 30% for the
service (peer review) provided, but that need not be recovered through
access-blocking S/SL/PPV. Authors, moreover, need not be held hostage
to the paper-based 100% today. They have done their bit by contributing
their papers to their publishers for free, to be sold through S/SL/PPV
for as long as that lasts. To assure full access to them now, they can
and should also archive all their papers publicly on their Home-Servers
and on

Why should the communication of research wait? And how long?

> sh> Is it coherent to declare that "you may publicly archive your work, but
> sh> not the correct, final version?" (Will you shoot me if I just fix the
> sh> "it"?)
> They could sue the institution for infringement - easy to get people to
> back down when their department chair is breathing down their neck.
> The policy is certainly enforceable, whether they will actually try
> I do not know.

I think this is complete nonsense, and that is borne out by the fact
that the seven years of existence of xxx have not inspired a single
attempted lawsuit. The reasons are as simple as they are numerous:

(1) It is AUTHORS themselves who are publicly archiving their own texts
there, not bootleg-xerox USERS.

(2) The authors have not been paid a penny for their texts.

(3) They all want to maximise their texts' visibility and accessibility.

(4) Provost initiatives like that of Steve Koonin, under discussion
here, don't bode well for any prospects of support from Departmental

(5) The arbitrariness of any attempt to specify how different a
preprint and reprint would have to be does not make the case more

(6) The global furor caused by the few days when xxx went off-line a few
years ago suggests little sympathy for any publisher who attempted to
achieve the same result, and a possible class-action response.

    Taubes, (G) E-mail withdrawal prompts spasm. Science 262 (Oct. 8,
    1993), 173-174.

> sh> Is there an apter word than "tortuous" to describe the attempt to
> sh> justify this constraint as being in the service of "preserving
> sh> textual integrity"?
> What if the author comes in again with another "correction" and changes
> the number to 3.995? Are we relying on authors for versioning and
> keeping their readers aware of changes? Publishers have (an admittedly
> inadequate) method of dealing with such corrections through
> errata/corrigenda in later publication. Textual integrity is a real
> issue, not to be lightly dismissed.

Please say you're joking! What's the best and broadest-band way to get
out corrigenda and link them to the original? Let the final, refereed
1st edition be tagged as such. And let all updates likewise be so
tagged, including whether they are +REFEREEED or -REFEREED (with the
journal's brand-name, to boot). The idea that there is some definitive
entity that must be etched in stone over and above all of this is just
paleolithic nonsense; that it requires paper and S/SL/PPV to preserve
this "integrity" is positively pre-Cambrian!

> sh> Does "ludicrous" not quite capture portraying such a self-serving
> sh> restriction as "adding value"?
> They didn't say the restriction added value. You seem to be confusing
> the economics of the issue here. The restriction is a protection of
> what they perceive as the value added through the editorial process,
> which you do not dispute (3.972 is better than 4.072, so value has been
> added).

Arthur, I knew perfectly well that I was punning on "added value"
there, and that is precisely what is ludicrous about it: Is there a
better descriptor for mentioning the "adding of value" in the same
breath with a measure intended to force authors either to stick to an
uncorrected public statement, or not to make it public at all,
entrusting that instead exclusively to a restricted pay-channel channel
called "PUBLICation"?

> sh> Is the conflict of interest with which this is all but exploding blatant
> sh> only to my ears?
> Sure there's a conflict - everybody involved has different interests
> (authors, referees, editors, readers, and librarians and publishers) -
> but I don't quite see what you're driving at with this accusation -
> Elsevier as publisher wants to publish its own way, not through free
> pages on the web - if authors are willing to agree to this, why not let
> them?

24,000 authors a year on xxx alone seem to feel otherwise, and I'm
betting that, once they realise and understand the possibility (and I
HOPE you don't disagree that it is inevitable and optimal that they
should realise and understand it!), ALL authors will feel otherwise.
Why on earth wouldn't they?

By the way (and this is important): It is not a coincidence that the
conflict of interest that you think is so natural and universal exists
only for this very special literature: the refereed journal literature,
where authors give publishers their texts for free and ask not a penny
in return. No such conflict of interest among the very same parties
(authors, referees, editors, readers, and librarians and publishers)
exists for the rest of the literature, which includes most of it:
books, magazines, etc.

There is clearly something different going on in this very small corner
of publication space; the peculiar conflict of interest has always been
there, dormantly, because unresolvable; but the means of resolving it
have now become available, in the PostGutenberg Galaxy.

> sh> And here is the last step of the reductio ad absurdum: Both the author
> sh> and the referees have "added their value" to the product for absolutely
> sh> nothing, not a penny!
> You place too much economic value on money itself, and not enough on
> the intangible (sometimes even tangible) benefits authors and others
> (yes, even referees) gain from the process. My argument is that without
> the publisher, this "added value" would, at least in 99% of cases,
> never be created, no matter how much authors and referees thought of
> the value of their contributions.

The added value you are referring to is peer review itself, and what
you are saying is tautologous: Of course without peer review there
would be no peer review. The benefits you speak of are those vouchsafed
by a certified, peer-reviewed literature. But no one is talking about
doing away with peer review, or with the publishers who implement it.
At issue is only how to provide and fund it. My own candidate is to
support Skywriting in the interests of free Skyreading out of the
savings from the S/SL/PPV that had subsidised paper-based reading in
the Gutenberg Era.

Stevan Harnad
Received on Tue Aug 25 1998 - 19:17:43 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:28 GMT