Zeno's Paradox and the Road to the Optimal/Inevitable

From: Stevan Harnad <harnad_at_COGPRINTS.SOTON.AC.UK>
Date: Sat, 2 Sep 2000 22:40:56 +0100


Zeno's Paradox was the one about the philosopher who thought: "How can
I possibly get across this room? For before I can do that, I have to
get half-way across, and that takes time. And before I can get half-way
across, I have to get half-way-half-way across, and that takes time
too. And so on. So how can I possibly even begin?"

I don't know what the theoretical solution to Zeno's Paradox is, but
the practical solution is to walk across the room. Otherwise we have
Zeno's Paralysis.

The equivalents of Zeno's Paradox have been repeatedly aired in this
Forum: "How can we possibly free the refereed literature? For before
we can do that..."

The faithful participant in this Forum will by now be familiar with
many of the prima facie candidates for "...." What they all have in
common is that they leave us where we are, paralyzed, within a
financially fire-walled refereed literature.

By my count, all the prima facie rationales for Zeno's Paralysis so far
aired have also been rebutted (energetically, repeatedly, and
decisively) in this Forum, and the steps are clearly within sight and
reach for freeing the refereed literature overnight through author

Joseph Ransdell's rationale for Zeno's Paralysis has been that there is
a bigger problem than freeing the refereed journal literature, and that
this bigger problem is the real one to solve.

The simple (and correct) answer is that solving the smaller problem is
not contingent upon solving the bigger problem -- except if we insist
on seeing it as such, in which case it is that erroneous conflation
that IS the problem.

I will not do a full point-by-point rebuttal of Joseph's rather lengthy
posting, because much of it is simply irrelevant, and would amount to
my joining in the endless and needless splitting of the half-distances,
Zeno-style. I will merely correct the substantive errors and

> Date: Fri, 1 Sep 2000 17:57:13 -0700 From: "ransdell, joseph m."
> <ransdell_at_DOOR.NET>
> Your plan for achieving your goal ["to free the refereed literature online"]
> ... is [1] to encourage self-archiving of
> professional papers, including functional equivalents of papers already
> published or being published in the refereed journals, thus [2] initiating
> an economically determined process which you believe will inevitably
> result in the elimination of such fees in due time.

There is a profound misapprehension here which it would probably be
useful to clear up in the minds of anyone who might have fallen prey to

The refereed literature is already freed by [1]! The rest [2] is merely
a prediction. There remains the possibility that even when the refereed
literature is freed by [1], there will still be add-ons that
individuals and institutions are willing to pay fees for. That is not
pertinent to the goal of this Forum, which is to free the refereed
literature, which will have been realized. The refereed literature will
no longer be held hostage to those add-ons.

> Since the idea is to eliminate costs to the reader this plan shifts
the > costs back either to the author or to somebody -- a third party
-- who > benefits sufficiently from this kind of publication to be
willing to > bear the expense involved in mediating the author's work
to the > prospective readership -- the costs of publication, in other
words -- > or at least is willing to share the burden of expense with
the author.
> You think the best candidate for this third party is the university,
> who will subsidize the mediation, perhaps supervise it, perhaps even be
> fully in charge of it. This is on the assumption, though, that this
> third party would not impose access restrictions of its own.

You are conflating [1, freeing the refereed literature] and [2, the
probable economic restructuring that may follow from 1]. The University
is an ally in establishing Open Archives in which its researchers can
self-archive their preprints and postprints (http://www.eprints.org).
That is all that is needed for attaining [1].

(Universities do not currently think of levying fees on JOURNALS for
access to their researchers' give-away research; it is equally absurd
to imagine that they would levy access fees on the very researchers on
whom the universities seek to make a research IMPACT in publishing that
same give-away research in the first place! It is IMPACT that generates
research and revenue for researchers and universities. To instead seek
direct income by blocking access to research reports is tantamount to
blocking their impact, and would be equivalent to advertisers seeking
direct income by levying fees for access to their ads. In other words,
Joseph's Zeno-like worries on this score are not only unfounded but

As regards [2] (on which [1] is not contingent, and which it need not
and should not, indeed cannot await in advance, Zeno-style), it is an
arithmetic fact that IF the economic consequence of [1] is that
universities' readers prefer the free version of the literature (rather
than just that portion of it that their university can afford via S/L/P
fees), and hence that their annual S/L/P fees need no longer be paid,
THEN there will be many times more than enough to pay for the tiny
remaining peer-review costs out of the annual savings. The universities
need to have their research peer-certified for the same reasons their
researchers' do -- for the "credibility" of their "ads" -- so it is in
both their interests to use a bit of the annual wind-fall savings to
pay for the quality-control.

But ignore prediction [2] if you like, because [1] is not contingent on
it! Self-archiving, eo ipso, frees the literature for everyone,
everywhere, forever, even if the universities continue to pay for the
fire-walled version the old way!

> the problems implicit in the assumption that whoever acts in behalf of
> the author in mediating (or financing the mediation) of the research
> paper from the author to the reader will not impose access restrictions
> of their own. The common sense dictum that whoever pays the piper is
> likely to want to dictate the tune suggests that anyone who assumes the
> former role of the commercial publisher might find reasons of their own
> for restricting access in some way, and your idea that there is no
> problem in this worth discussing if the agency is the university or a
> multi-national consortium of universities seems to me extraordinarily
> naive, but perhaps explicable in terms of your own background as a
> British academician which may mislead you in trying to understand the
> rather more volatile and sometimes rapacious character of American
> universities.

Vide supra. (I am, by the way, Hungarian born; a Canadian citizen, and
lived and worked in the US until 1994...)

> The weakness... is in the description of the motive, which
> everybody agrees with but which says much less than it seems to say
> because it does not take into account the realities of communication,
> one of which is that in order to say something to somebody it is not
> enough to utter the words and make them public: the person addressed
> must be in position to hear them and be willing to listen to them
> attentively, especially when the words are in the form of lengthy and
> difficult scientific reports.

This question might have been raised in the era before Gutenberg and
before peer review. The answer is, one does not say ones words, one
publishes them. And one does not publish them in a Vanity Press, but in
the highest-standard peer-reviewed journal that will accept and certify
them as having met its standards. And a "peer" is merely a qualified
fellow-expert in the domain of expertise in question.

One can only repeat: The objective is to free the REFEREED LITERATURE
(such as it is), not to save the world first. Whatever constrained the
"realities of communication" in the Gutenberg Era (i.e., peer-reviewed
publication) will continue to do so in the Online Era. All that's
needed is to free that already constrained literature from
access-blocking tolls.

> Now we come to the point: what the research author wants is not
> merely to write his or her results in the sky but to press them upon a
> targeted readership -- his or her research peers -- in the special form
> of research claims, the communicational vehicle of which is the primary
> research publication. Why? Because the researcher aims at making a
> contribution to his or her field, and the way you make a contribution
> is to make a research claim and have it accepted by your research
> peers.

Correct. And that is precisely why one self-archives both the
pre-refereeing preprint and the post-refereeing postprint. The
journal's peer-review tag is then the certification of having met the
quality standards of the peers at that level.

> Now acceptance of a research claim is something that occurs
> when and only when what the researcher has concluded and presents in
> the form of a claim is in fact taken up and actually used by others as
> something taken for granted by them in their own work.

Correct. And this second phase, after peer-review certification, is
called IMPACT. And it is precisely to maximize the potential impact of
research that we are trying to free the peer-reviewed papers from
obsolete access-barriers (= impact-barriers) online.

> It seems to me, Stevan, that in your zeal for the protection of peer
> review practices you have lost sight of why people want to publish.
> According to you, in one of many statements to the same effect:
> SH > Authors work hard for recognition and certification
> SH > by their PEERS, and it is the service of peer-review
> SH > (refereeing) that a refereed journal implements (the
> SH > peers review for free too!).
> The first half of the sentence is clearly intended to convey that it is
> the approval of peer reviewers -- as distinct from acceptance by one's
> research peers generally -- that motivates normal publication

I'm afraid I couldn't follow this: The peer-reviewers selected by the
editor are indeed intended to be a valid and representative sample of
the relevant peer expertise in general.

Apart from that, Joseph seems to be confusing the first, filtering
stage of IMPRIMATUR (Quality-Control & Certification [QC/C] through
peer review), which sign-posts the published literature, with the
second stage, IMPACT, something that only comes after publication, as
the peers navigate this sign-posted literature and try to build on it.

> since the
> fictitious office of certification could only be imputed to peer
> reviewers or editors; and the second part of the sentence characterizes
> the research function of refereed journals as being that of
> "implementing" this bogus act of certification.

Strong words. But note that they could just as well have been spoken in
the paper era. Joseph is dissatisfied with peer review: Fine. Let him
find and test something better, something that gives us a literature
that is of at least the same level of quality as the one we have, and
then we can talk about implementing his system.

Till then his animus against refereeing is completely irrelevant to the
objective of freeing the current refereed literature from needless
impact barriers. It is merely a red herring.

> ... certification is a convenient fiction...
> authoritarian conception of editors
> nothing more than... a very questionable means -- of effective
> access to research peers.

Until Joseph provides an at least equally effective means of access, I
suggest that these fulminations serve no constructive purpose.

> My guess is that had you not come up with the "invisible hand" argument
> to reconcile the apparent discrepancy between your devotion to the
> refereed journal, construed as a policing system based on peer review,
> and your devotion to the Ginsparg Archive, which dispenses with
> editorial and peer review policing altogether, you would long since
> have recognized that there is no need to resort to such a questionable
> hypothesis since the explanation of how critical control is maintained
> in the research fields served by the Ginsparg Archive is readily
> apparent when you look closely at how Ginsparg constructed the Archive
> and ask yourself what would keep YOU in line if you were publishing
> there. Fear of a possible future peer review by a person unknown whose
> opinion will count only so far as some editor decides to take account
> of it? I don't think so.

The "invisible hand argument" follows from a very simple fact (not a
hypothesis) about the Ginsparg Archive: From the very beginning (and
still earlier), all those papers were and still are destined for peer
reviewed journals. They are written with the expectation of being
answerable to peer review. There has been no change in that respect.
The Archive is not, and never was, a publisher, or an alternative to
publishing. And publishing in Physics (and in all other academic
disciplines) does not refer to vanity-press self-publication, but
refereed publication. Self-archiving, in other words, is not
self-publication and never has been. The journal hierarchy continues to
exist, as it did before the Archive existed; the only difference is
that now the peer community has access to both the pre-refereeing
preprints and the post-refereeing postprints for free (and earlier).

Those are the facts. The speculative part is about what would happen if
the invisible hand of peer review were actually removed -- if the
journals all vanished, for some reason. Joseph thinks everything would
stay pretty much the same, whereas I think they would dip toward Usenet
-- until we re-invented peer review.

Nothing much hinges on that speculation for me, for I am merely
advocating freeing the peer-reviewed literature from impact-barriers
through self-archiving -- which is precisely what the Ginsparg Archive
is doing for Physics; hence this is a proven principle. Joseph, in
contrast, seems to be advocating freeing the literature from peer
review, through self-publication. That is rather a different matter,
and raises a wealth of empirical questions that the facts and history
of the Ginsparg Archive certainly do not address, let alone

    Harnad, S. (1998g/2000) The invisible hand of peer review. Nature
    [online] (5 Nov. 1998) and in Exploit Interactive 5 (2000)
    Longer version:

> The fact that quality control is maintained in the fields served by the
> LANL system can be accounted for easily enough if we bear in mind what
> the critical control of research is based on: uninhibited peer
> criticism by the generality of one's peers

Yes, peer review does not stop with formal journal refereeing; that is
part of the continuous, collective, self-corrective nature of science
and of learned inquiry in general. But post-publication "open peer
commentary" is not to be confused with pre-publication peer review. For
the peer commentary is transpiring in a filtered, quality-controlled,
sign-posted, and hence navigable published literature; the peer review,
in contrast, is operating on the raw, unfiltered substrate which, even
with the help of the Invisible Hand, is something the
community-at-large needs to be spared having to navigate unaided, given
the unspeakable (let alone readable) volume that is being written.

    Harnad, S. (1998/1997) Learned Inquiry and the Net: The Role of
    Peer Review, Peer Commentary and Copyright. Learned Publishing
    4(11): 283-292
    html http://citd.scar.utoronto.ca/EPub/talks/Harnad_Snider.html

Only Editors, paid to do the job, and referees, stealing from their
precious research time at the behest of those Editors [and the Golden
Rule], should ever have to contend with those raw first drafts:
Joseph's imagination is failing him if he thinks that science would be
well-served if those drafts all went straight on-line, without
mediation, and the filtration began only afterwards, with "open peer
commentary." What peer would have the time or the courage to tackle
generic papers willy-nilly so as to let the rest of the peerage know
whether it had all been worthwhile? How is the labour to be divided?
Indeed, is it not likely that classical peer-review would simply have
to be re-invented so as to divide the labour rationally?

In any case; this is all speculation. If Joseph wants to test an
alternative, he has to do so empirically, with a representative sample
of raw manuscripts, NOT written for submission to refereed journals,
but written only for this sort of free-for-all; and he needs a
representative sample of peers, who must navigate and sign-post this
literature with comments. And the experiment must be conducted with a
large enough sample size and for long enough so that we can have
confidence that its results generalize beyond a Hawthorn Effect. When
he comes back to report his findings, then they will warrant our
attention; but not now, on the basis of the Ginsparg Archive, which is
certainly NOT an implementation of the experiment in question.

> A rejected paper on the basis of negative peer
> review can perhaps be published elsewhere, but a damaged reputation
> arising from such direct communication is all but impossible to
> rehabilitate because people often will not even bother reading later
> work that might redeem an author's reputation because they already have
> too much to read, anyway.

And with so much to read, how does Joseph imagine that anyone has the time to
test and monitor reputations?

> (I omit here a discussion of the evidence
> for believing that the Ginsparg Archive actually does work as a system
> of direct primary publication for those fields for which it was
> created.)

It would be interesting to see that evidence; on the face of it, until
and unless those authors stop submitting all those papers to refereed
journals, they constitute no kind of publication, but merely what they
look like on their face: public self-archiving, before and after

> Your assumption has
> been that getting people to self-archive is just a matter of providing
> software that enables people to upload, download, and so forth, given
> their interest in making their work public. But individual researchers
> have no interest in using a communicational medium that doesn't connect
> them effectively with people in their research field, and merely
> self-archiving their papers doesn't do that.

Correct. It is the Quality-Control & Certification (QC/C) provided by peer
reviewed journals that does that.

As to "connecting effectively": The old way was with impact-barriers,
the new (optimal/inevitable) way will be without impact-barriers. Which
is more effective?

And what (if anything) would be still more effective?

(Note that self-archiving both preprints and postprints on-line
SUBSUMES Joseph's own preferred option -- which I suppose is one
without peer review: Each author has the option of performing the
empirical experiment I urged on Joseph, namely, to self-archive only
preprints, and not bother with journal peer review. It is surely of
some empirical relevance that virtually no author in the Physics
Archive has chosen this option, in its decade of existence. Discipline
differences? I'll wait to see; apart from a few crackpot papers,
exactly the same pattern occurs with the multidisciplinary CogPrints
Archive http://cogprints.soton.ac.uk.)

    Harnad, S. & Carr, L. (2000) Integrating, Navigating and Analyzing
    Eprint Archives Through Open Citation Linking (the OpCit Project).
    Current Science (special issue honour of Eugene Garfield) (in

> Paul Ginsparg ported a
> research community, or at least its nucleus, from one medium of
> communication (the email distributed preprint) to another (the preprint
> archive and server system), and any authentic emulation of his
> accomplishment has to do something analogous to that.

I don't understand this "porting" terminology. Speaking literally and
factually, the Archive gave Physicists a means of publicly
self-archiving their preprints and postprints online. That's all. No
other practises changed (apart from the user preference for the free
online archive).

> My impression is that because of your conviction that the refereed
> journal system of communication is the be-all and end-all of scientific
> communication you have thought of what Ginsparg did as being primarily
> a feat of computer applications, whereas the real feat was in creating
> an on-line system of primary research publication, which was a very
> different sort of accomplishment.

I am afraid that there is no evidence whatsoever for this -- except if
you are doing a play on words with the term "primary": certainly both
the unrefereed preprint and the refereed postprint precede the
appearance of the print version, but that is not a new form of "primary
research publication" but a welcome primacy in real time. (By the same
token, paper preprints, which were always disseminated earlier than
final drafts and journal offprints, were never a new form of "primary
research publication" but merely a faster way of disseminating papers.)

> Let me quote a couple of passages from an address of [Ginsparg's] of 1996:
> PG > It is ordinarily claimed that journals play two
> PG > intellectual roles: a) to communicate research
> PG > information, and b) to validate this information
> PG > for the purpose of job and grant allocation.
> PG > As I've explained, the role of journals as
> PG > communicators of information has long since been
> PG > supplanted in certain fields of physics, so let's
> PG > consider their other role. Having queried a number
> PG > of colleagues concerning the criteria they use in
> PG > evaluating job applicants and grant proposals, it
> PG > turns out that the otherwise unqualified number of
> PG > published papers is too coarse a criterion and plays
> PG > essentially no role. Researchers are typically familiar
> PG > with the research in their own field, and must in any
> PG > event independently evaluate it together with letters
> PG > of recommendation from trusted sources. Recent
> PG > activity levels of candidates were mentioned as a
> PG > criterion, but that too is independent of publication
> PG > per se: "hot preprints" on a CV can be as important as
> PG > any publication.
> http://xxx.lanl.gov/blurb/pg96unesco.html
> That seems straightforward enough as a statement of research autonomy,
> in contrast with your view of the dependence of the LANL Archive on the
> "invisible hand" of the refereed journal system

I am afraid I do not see that at all! We have here some anecdotes about
how colleagues say they go about doing evaluations, but that has no
bearing whatsoever on the "invisible hand" factor (which I am beginning
to fear you have not understood: to repeat, that factor would only
vanish [to mix metaphors] if physicists stopped preparing their papers
for, and submitting their papers to, peer review; nothing of the sort
has happened).

> he indicates
> clearly in a later passage in the same paper that his intention was
> specifically to create at Los Alamos a system for primary research
> publication:
> PG > It is important to distinguish the form of communication
> PG > facilitated by these systems from that of usenet newsgroups
> PG > or garden variety "bulletin board" systems. In "e-print
> PG > archives," researchers communicate exclusively via research
> PG > abstracts that describe material otherwise suitable for
> PG > conventional publication. This is a very formal mode of
> PG > communication in which each entry is archived and indexed
> PG > for retrieval at arbitrarily later times; Usenet newsgroups and
> PG > bulletin boards, on the other hand, represent an informal mode
> PG > of communication, more akin to ordinary conversation, with
> PG > unindexed entries that typically disappear after a short time.

Ahem. Even taken out of context, it is hard to imagine that anyone
would interpret this as implying that mere format constraints such as
providing an abstract are what distinguish the Usenet literature from
the Los Alamos literature. Might I point out that Usenet postings are
not, and never will be, answerable to peer review, whereas Los Alamos
papers all are?

> How did he go about establishing the Archive as a place of primary
> publication? First, he did this by so constructing it that the basic
> necessary conditions of primary publication were met and its users were
> made aware of this so that they would understand what they were doing
> in depositing their papers in the archive. This included such things
> as providing the software to be used in putting the contributions into
> a form suitable for primary publication, requiring an abstract, setting
> up a system of notification of availability to interested parties,
> dating the submission, insuring against tampering, keeping successive
> versions, and so forth. (Paul could perhaps be prevailed upon to fill
> in more detail on this, but much can be derived merely from studying
> his informal papers on the topic, which make it clear that he takes
> account of all of the factors which Lederberg identified as necessary
> for literature of this very special function.) His diligence and good
> judgment in attending to these seemingly trivial details was the basis
> of his accomplishment. His achievement in constructing that system was
> not the achievement of a computer programmer but of a professional
> physicist.

What has been provided is a self-archiving system. Try to remove the
layers of interpretation and see it factually for what it is. It will
not be a place of primary publication until the REAL place of primary
publication (the journal, and its all-refereeing QC/C) are out of the
loop. Or is this just a play-of-words on "place": for I certainly agree
that the "locus classicus" of the papers has become the virtual locus,
rather than the real locus: yet the citations are invariably to the
journal-published version, as soon as it is available...)

> Second, he also accomplished it by defending it against
> misrepresentations of its functions, as he is doing in the passage just
> quoted. Attempts to trivialize its significance -- which are attempts
> to show that it is not a place of primary publication -- typically take
> the form of mischaracterizing it by describing it in such a way that
> one ignorant of how it actually works gets the impression that it is
> like a chatroom or a usenet news bulletin board or a listserver based
> forum or general discussion group or a MUD, etc., which are on-line
> communicational forms rarely capable of functioning as places of
> primary publication, whatever their value may be otherwise, because
> some one or more of the necessary conditions for that are normally
> absent.

Yes, it is false to characterize it as a MUD; but it is equally false
to characterize it as a "place of primary publication": It is a place
to publicly self-archive primary publications (and prepublications)

> The bad news, from the point of view of your use of it as a model for
> self-archiving, is that it is such a formally austere environment as to
> be suitable only for a highly developed research community, accustomed
> to independence from editors and peer reviewers because their regular
> practices as scientists embodies a higher level of critical control
> than the refereed journal system can offer. For all your admiration
> for Ginsparg's accomplishment you have not been able to perceive the
> basis for it in the exceptional character of the research communities
> themselves, whose members have not required the kind of policing you
> take for granted as the paradigm of critical control because their
> normal practices already exemplify the paradigm of scientific
> self-control in a more pure and developed form. Thus it apparently
> hasn't occurred to you that it cannot function as the universal model
> for stimulating self-archiving for that very reason.

Nothing of the sort, as I hope I've pointed out in enough ways by now.
If there is reputed to be something different about physics (and it is
just a rumour, I don't really know the hard evidence), it is that (1)
physics papers undergo less revision under refereeing than papers in
other disciplines, (2) physics journal rejection rates are lower, at
all levels of the hierarchy, and (3) physics papers tend to be aimed
more directly at their proper level of the hierarchy (rather than
trickling down via multiple submissions, until they find their level,
as they do in some other disciplines).

It would follow from (1) that unrefereed Physics preprints would be
safer for immediate peer consumption than in fields where QC/C
typically requires more substantive modifications. If these things are
true, they suggest that physicists are perhaps more disciplined, more
conscientious, or smarter. But be that as it may, it does not give a
hint of what they would be like if the Invisible Hand were removed,
i.e., if they stopped being answerable to journal peer review. My guess
is that they would prove to be like all human beings when they are not
answerable to quality-control standards; and the quality-control would
quickly have to be brought back again.

> I have no idea how many research communities like the ones that make
> good use of the Ginsparg Archives there may be, but in reflecting on
> the many reasons why a given research community might not be matured
> enough to handle a system as austere and demanding as the Ginsparg
> Archive is -- sometimes because of factors beyond anyone's control,
> sometimes from situations in which privilege and power has corrupted
> leadership and demoralized the research community, sometimes from
> continuing inability to find agreement on basic research aims -- it
> quickly became clear to me that it is not likely that we will find
> nearly enough fields mature enough to set in motion the economic
> process you envision as leading to the "inevitable and the optimal",
> even supposing the problem which I touched upon in my previous message
> of recruiting research communities rather than individuals has been
> solved. You may want to think in terms of making common cause with the
> Digital Journal people and others who take a more realistic view of
> what can be done.

I can't follow all this. It seems to me any "research community" that
currently has a fleet of refereed journals, currently behind S/L/P
impact-barriers, is "mature" enough to free it through self-archiving
-- if it can free itself from Zeno's Paralysis.

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.ecs.soton.ac.uk/~harnad/
Highfield, Southampton http://www.princeton.edu/~harnad/

NOTE: A complete archive of this ongoing discussion of providing free
access to the refereed journal literature is available at the American
Scientist September Forum (98 & 99 & 00):


You may join the list at the site above.

Discussion can be posted to:

Received on Mon Jan 24 2000 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:50 GMT