Charles Oppenheim (C.Oppenheim AT lboro.ac.uk) (CO) has given me permission to
post this exchange:
CO:
"Stevan, whilst I have a lot of sympathy with your position, some of
your statements are not quite correct.
"On (1), the approach IS pedantic, but there have been cases of fixing
RAE returns. So, RAE panels are concerned that academics might
amend postprints to improve them after the event. Unfortunately,
the 0.01% who cheat have caused RAE panels not to trust the other
99.99% of submissions.
"On (2), if authors are so stupid as to assign copyright to publishers
then indeed they (or their employers) do not have the freedom to
pass copies to the RAE panels for evaluation. The exceptions to
copyright in UK law are few, and do not apply when an RAE panel
member wishes to read an item. The way round this problem is to
ensure that academics never assign copyright to publishers!
"On (3), allowing any number of publications to be considered would
make the panel members' burden massive, as they would be obliged to
read as many or all of the items listed by an academic - they have
a hard enough job evaluating the four at the moment.
"However, thank heavens for (4) because, as you rightly point out,
a metrics based system will sweep all this nonsense aside.
"Overall, then, unfortunately there are valid reasons for (1),
(2) and (3), but this is the last time we need be crippled by such
niceties....
Professor Charles Oppenheim
Head
Department of Information Science
Loughborough University
Loughborough
Leics LE11 3TU
Here are my replies to CO:
> CO:
> Stevan, whilst I have a lot of sympathy with your position, some of your
> statements are not quite correct.
>
> On (1), the approach IS pedantic, but there have been cases of fixing RAE
> returns. So, RAE panels are concerned that academics might amend postprints
> to improve them after the event. Unfortunately, the 0.01% who cheat have
> caused RAE panels not to trust the other 99.99% of submissions.
Dear Charles,
All these points came up in our last exchange on AmSci
"Question for publishers - Research Assessment Exercise 2008"
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/subject.html#5122
We don't disagree about the objective facts; I think we simply disagree
about both the substance and the consequences of the objective facts.
(a) The premise is that this panel re-review is absurd and needs to be
scrapped in favour of metrics: the journals already did the peer review,
once; no need to repeat it yet again, locally.
(b) So what if authors correct or improve their publications after
publication? That's *good*, not bad!
(c) If the panel *must* do a re-review of published output, then surely
it is the *content* that should be reviewed; and that content should be
in its best form, not riven with errors that happen to have have slipped
into the published version, undetected by the first peer review!
(d) And, last, given all the above, surely the prospect of being
named-and-shamed (and otherwise punished-and-penalised) by RAE
for fraudulent submissions is a more cost-effective deterrent, for
both researchers and their institutions, than submitting/licensing
publishers' PDFs! Hence it is not *RAE* that should be worrying about
and taking precautions for the 0.01% who actually do cheat (if there
really is even 0.01%, in view of (1)-(3) above), but researchers and
their own institutions: It is *institutions* who should vet their own
RAE submissions to ensure there is no fraud, not the RAE!
> CO:
> On (2), if authors are so stupid as to assign copyright to publishers then
> indeed they (or their employers) do not have the freedom to pass copies to
> the RAE panels for evaluation. The exceptions to copyright in UK law are
> few, and do not apply when an RAE panel member wishes to read an item. The
> way round this problem is to ensure that academics never assign copyright to
> publishers!
I would like to ask you a set of very direct questions about this, because
I find these interpretations of UK copyright law rather hard to construe:
(i) If I publish an article and personally mail a reprint or photocopy
of it, along with my CV, to a university that I am applying to for
a job, so that their hiring committee can evaluate my candidacy, is
there a country or law foolish enough to (a) declare that illegal and
(if, per impossibile, (a) were indeed the case) to (b) imagine that
that "illegality" can or could or would or should ever be detected or
enforced, by requiring job candidates to seek permission to submit their
own published work to be evaluated for their own career and credit by
their own prospective employers? (In other words, if this does not fall
under "Fair Use" by the author, then surely nothing on earth does!)
(ii) Precisely the same reasoning applies to submitting my own work to
the RAE for evaluation. It is simply a piece of arbitrary nonsense to
view this submission "formally" as a transaction between a university
and a national organisation! It is the individual submission, by an
individual researcher, of his own work, for evaluation: Use that is
every bit as individual and Fair as (ii).
(iii) With the deposit of the author's final, peer-reviewed postprint in
the author's own Institutional IR already at least 94% Open Access, with
the publisher's blessing, not only to the RAE, but to the entire WWW, and
the remaining 6% deposited in Closed Access, with access provided only to
individuals who request a single reprint from the author by email, again,
the RAE panelists, if they ask nicely, fall under individual authorial
Fair Use. This is almost certainly (a) incontestable in a legal sense
but absolutely certainly (b) unpreventable in a practical sense and (c)
with even contemplated/attempted prevention unjustifiable in any sense
other than taking complete leave of one's senses (as the RAE has done,
in self-declaring that it must license the content -- and then going
ahead and self-imposing that absurd licensing arrangement upon itself).
(If the RAE did this on legal advice, I would have to say it was done on
very bad, unrealistic, far-fetched, and out-of-touch advice. And Charles,
I am sure you will agree that copyright lawyers quite often freely
dispense, with supreme confidence, very bad, unrealistic, far-fetched, and
out-of-touch advice on online-age matters that they do not even remotely
understand -- and laymen are sometimes foolish enough to heed it!)
> CO:
> On (3), allowing any number of publications to be considered would make the
> panel members' burden massive, as they would be obliged to read as many or
> all of the items listed by an academic - they have a hard enough job
> evaluating the four at the moment.
Are we forgetting here that the purpose of the new metric RAE is to put an
end to this unnecessary, redundant and time-money-wasting panel re-review
of *already-peer-reviewed* publications? Are we forgetting that even now,
it is almost certain that the panelists don't read all or most of the
submissions, but merely spot-check? And that the RAE rankings that result
are nevertheless highly correlated with citation metrics, making all
this solemn submission and re-reading all the more absurd and wasteful?
I agree that RAE 2008 is already too far-gone now to scuttle the panel
review for metrics alone; that is why it is being billed as a "parallel
exercise" (perhaps not altogether a dead loss, if it generates more
explicit discipline-based correlations, to lay to rest any lingering
doubts in some disciplines about their being exceptions to the predictive
power of metrics). But it does no harm to remind all concerned of the
needlessness and profligacy of the current panel re-review process,
lest the slightest temptation to resurrect it be tempted to re-surge.
(I would also like to make a side-bet with you, Charles, which is that if
you took the very useful and important RAE/citation correlational studies
that you have already done, and broke them down into the correlation
between the citation metric and the RAE outcome using only the citation
counts for the four submitted articles, versus the citation counts for all
the published works (in the reviewed time-frame) for all the submitted
researchers, you would find that the correlation was significantly
higher and more stable when you use all their works than when you use only
four of them!)
> CO:
> However, thank heavens for (4) because, as you rightly point out, a metrics
> based system will sweep all this nonsense aside.
Yes, we are now talking about the dying gasps of a moribund system:
The RAE seems determined to keep it as absurd and wasteful as possible
right up to the last gasp!
> CO:
> Overall, then, unfortunately there are valid reasons for (1), (2) and (3),
> but this is the last time we need be crippled by such niceties.
I would say there are no *valid* reasons for (1), (2) and (3)! The
only thing that can be said is that the RAE continues to insist,
needlessly, and for no good reason whatsoever, by arbitrary and
counterfunctional fiat, on (1) receiving the "originals" rather than
the author's postprints, on (2) construing this as an institution/HEFCE
transaction, instead of the Fair Use by the individual author, for
individual evaluation, that it really is, and on (3) cleaving to the
limit of 4 submissions as a purely self-imposed consequence of the
(soon to be abandoned) cleaving to panel review itself.
Charles Oppenheim subsequently wrote this:
> CO:
> Happy for you to forward our exchange in any way you like, but add the
> following comment from me:
>
> I agree that no publisher would be so foolish as to sue for infringement
> if an individual author passed a copy of his/her papers to an appointments
> panel, etc., but when an entire University does so for all its staff,
> then a publisher might well complain about lost sales and would sue.
But Dear Charles: that's my point! It is simply a "facon de parler" to
say "an entire university does so for all its staff"! The RAE is a
research assessment exercise. Individual participation on the part of
staff is *voluntary*, and each staff member does so, individually, in
his own interests, for the sake of being assessed and thereby
contributing to his departmental RAE ranking and funding.
Submitting to RAE need no more be treated as something "an entire
university does so for all its staff" than submitting grant applications
for RCUK or EC funding, or submitting for internal performance review.
In every case, it is the individual author providing his own individual
papers for his own particular needs: funding and evaluation.
Call it what it really is, and it is obviously individual Fair Use. Call
it something else, needlessly -- some sort of transaction between
universities and national bodies -- and you are simply asking for
needless, gratuitous complications, which the pedants, bureaucrats,
copyright lawyers (and of course the publishers and licensing bodies)
are only too happy to play along with!
Individual Fair Use by the author: that's what RAE submissions amount
to: no more, no less. And there is no need to submit to RAE anyway,
just to deposit the author's final draft in his own Institutional IR,
set access as OA in 94% of the cases, and give RAE special one-off access
for assessment purposes if/when they need it for the remaining 6%. End
of story.
Stevan Harnad
Received on Mon Sep 18 2006 - 16:49:05 BST