Re: Open Letter about OA to the Royal Society by Fellows of the Royal Society

[Stevan Harnad <harnad@ecs.soton.ac.uk>]

On Thu, 8 Dec 2005, Marc Brodsky wrote:

> I do not speak as one who is against OA. I am neutral and open to 
> experiment and results.

I will now make the charitable assumption that Mark Brodsky does mean 
exactly what he says, and is simply unaware of the contradictions inherent 
in it. I will try to hold up a mirror, reflecting exactly what is latent 
in his own words and their meanings, and where they actually lead, if made 
fully explicit:

     "I am not against OA. I am open to experiment and results."

In physics, a 14-year experiment has already been conducted. Here are the 
results:

(1) OA means free online access for all would-be users.
(2) Self-archiving provides OA.
(3) Physicists have self-archived spontaneously.
(4) Spontaneous self-archiving (OA) has reached 100% in some areas of
     physics.
(5) Repeated experimental comparisons of research impact (between OA and
     non-OA articles in the same journal issue) have revealed an OA citation
     advantage of 25%-250% or more, and a download advantage of c. 300%.
(6) APS and IOP report that no detectable decline in subscriptions has been
     associated with self-archiving.
(7) RCUK now proposes to apply these positive experimental results to
     articles that have *not* yet been self-archived spontaneously, in order
     to raise the percentage of OA to 100%

Mark says he is not against OA, and open to experiment and results: Then 
why object to the RCUK experiment? Why delay it? And delay it for what?

> AIP has the same liberal pre-print and post-print posting policies as 
> the APS ones positively referred to below.

It is commendable and welcome for publishers to offer an official green 
light to go ahead with self-archiving ("authors may self-archive"): It is 
less commendable if they object to being taken up on it, when RCUK 
proposes to require the research it funds to "go" on green (i.e., "RCUK 
fundees must self-archive"), exactly as the AIP liberally agrees they may.

(I also note in passing that spontaneous self-archiving in physics did not 
wait for publishers to go green: Physicists went ahead and did the 
experiment first; then, to their credit, APS and AIP and IOP -- and 
eventually also Elsevier and Springer -- went green.)

> AIP has offered "author Select" OA options in three of its journals this 
> year, plans to do so for more in the near future, and will offer fully 
> OA journals soon as well. We are not biased one way or the other.

This too (optional "gold" OA publishing) is nothing but commendable, and 
an excellent experiment; but what I cannot determine is whether Mark (and 
many other publishers) are being deliberately obtuse or (on the charitable 
assumption) have simply failed to understand that RCUK is not proposing to 
require gold (OA publishing)! It is proposing to require green (OA 
self-archiving). Hence the RCUK policy is not an experiment in requiring 
an alternative publishing model that has not yet been tested and 
demonstrated to work experimentally. (The RCUK policy has nothing to do 
with gold or alternative publishing in any respect except that RCUK offers 
some *optional* help in paying for the "author Select" option that AIP is 
offering!)

What the RCUK is proposing to *require* (as opposed to merely offering to 
fund as an option) is that its fundees self-archive: i.e. that they all do 
it, in order to all enjoy the experimentally tested and demonstrated 
benefits of self-archiving, not just those that happen to do it 
spontaneously.

> AIP aims to disseminate info about physics to the widest possible
> audiences within the most affordable economic models.

This is all commendable, but not at issue at all, because the RCUK's 
application of the experimental findings to date does not require AIP (or 
any other publisher) to disseminate their contents any more widely than 
they already do; nor to abandon their affordable economic models. RCUK 
does not require publishers to do anything at all (how could it?). RCUK 
requires its own *fundees* to self-archive (as AIP has so liberally given 
them the green light to do), in order to derive the benefits of the 
spontaneous self-archiving that have already been experimentally 
demonstrated.

If there is an element of further experimentation in the RCUK policy, 
rather than merely the application of existing experimental results, it is 
the following:

(1) The RCUK is testing whether required self-archiving provides the same 
usage and impact benefits that spontaneous self-archiving does. (I think 
it is a very safe bet that it does, but let us test and see.)

(2) The RCUK is also testing whether the peaceful co-existence between 
spontaneous self-archiving and journal subscription revenue that has been 
experimentally demonstrated to date (even in fields that reached 100% 
spontaneous self-archiving years ago) will continue to hold with required 
self-archiving.

The outcome of (2) is not yet known, but let us consider the options:

(a) Peaceful co-existence between self-archiving and subscription revenues 
continues after the RCUK policy is adopted, but only because the RCUK 
percentage of journal output in every field is too small to make a 
significant change in the spontaneous self-archiving effects for any given 
journal. (This would be a very positive outcome of the experiment for UK 
research, but not a decisive test of the effect of overall 100% OA 
self-archiving.)

OR

(b) After a period of time, detectable declines in subscription revenue 
occur that can be statistically associated with the RCUK self-archiving 
requirement.

Pause for a question about experimental methodology: Is the possibility of 
outcome (b) grounds for not performing the RCUK experiment in the first 
place? (I think not.) Would outcome (b) be experimental grounds for 
abandoning the RCUK policy (I again think not: publishing would have to 
adapt to the outcome, rather than expecting research to renounce the 
benefits.)

What is virtually certain, however, is that the outcome of the RCUK 
experiment alone will not be catastrophic decline in subscription revenue, 
necessitating a change in publishing cost-recovery model. The UK share of 
journal content is almost certainly not big enough to have that effect.

That is the end of the plausible immediate outcomes of the RCUK 
experiment.

But let us consider the possibilities even further: What if the RCUK 
policy is adopted in other countries, because of its manifest benefits to 
research. The following outcomes are then possible:

(c) Peaceful co-existence between self-archiving and subscription revenues 
continues after the RCUK policy is adopted worldwide, and OA 
self-archiving reaches 100%. (This too would be a happy outcome for all 
concerned.)

OR

(d) After a period of time, as worldwide OA self-archiving approaches 
100%, detectable declines in subscription revenue occur worldwide that can 
be statistically associated with the worldwide growth of self-archiving 
requirement.

Pause again for another question about experimental methodology: Is the 
possibility of outcome (d) grounds for not performing the RCUK experiment 
in the first place? (I think not.) Would outcome (d) be experimental 
grounds for abandoning the RCUK policy or its worldwide emulation (I again 
think not: wordlwide publishing would have to adapt to the outcome, rather 
than expecting research to renounce the benefits.)

But now what about:

(e) After a period of time, as worldwide OA self-archiving approaches 
100%, catastrophic declines in subscription revenue begin to occur 
worldwide, necessitating either a change to the OA publishing model or an 
abandonment of the self-archiving requirement.

Questions: Is the possibility of (e) grounds for not performing the RCUK 
experiment in the first place? (I think not.) Would outcome (e) be 
experimental grounds for abandoning the RCUK policy or its worldwide 
emulation (I again think not: worldwide publishing would have to adapt to 
the outcome, rather than expecting research to renounce the benefits.)

> If our authors, readers or subscribers indicate by real actions what 
> they want, we will try our best to respond appropriately.

Is it clearer now why the above is a non-sequitur? The proposed RCUK 
experiment, which its opponents are trying to defer conducting, pending 
the outcome of other prior "experiments" (what experiments? experiments on 
what?) has nothing whatsoever to do with what authors, readers or 
subscribers currently want by way of journal economic models or OA 
publishing, because what the RCUK experiment is applying is the results of 
existing experiments demonstrating the positive effects of self-archiving 
on research impact (i.e., what *researchers* want) and their absence of 
any effect on subscription revenue. What RCUK is requiring is OA 
self-archiving, not OA publishing.

> I was talking about the thoughtfulness of the various letters referred 
> to. One seems to open to options and experiments, the other to a more 
> pre-judged one.

Those who are arguing for delaying or deferring the RCUK experiment 
pending the outcome of other "options and experiments" have not even 
specified what "options and experiments" they have in mind, nor what they 
would be testing, let alone specifying why we should not be applying and 
extending the positive experimental results we already have, to the 
benefit of research, researchers, their institutions, their funders, and 
the tax-paying public that funds the funders and is meant to derive the 
benefits of the experimental findings.

The Open Letter of the Royal Society Fellows is for opening the options 
and experiments, and the Royal Society statement (undoubtedly drafted by 
its publishing wing) is the one pre-judging the outcomes (and trying to 
filibuster the conduct of the experiment, despite the uniformly positive 
results to date).

Stevan Harnad