Re: Mandating OA: Don't Let the Best Be the Enemy of the Good

From: Stevan Harnad <amsciforum_at_GMAIL.COM>
Date: Tue, 2 Dec 2008 11:44:03 -0500

On 2-Dec-08, at 7:08 AM, leo waaijers wrote (on SPARC-OAForum):

      Dear Stevan,

      I will focus on your comments on my Recommendation 1 and
      leave the judgment on the Recommendations 2 and 3 and
      your criticism thereon to the readers.

      So this is about a comparison between a 'mandate to
      self-archive' and the usage of a 'licence to publish'.

Dear Leo,

For comparability, it needs to be a comparison between a 'mandate to
self-archive' and a 'mandate to successfully adopt a license to
publish'. Neither self-archiving nor licensing is being done
spontaneously by authors, hence we are talking about mandates in both
cases, and the question is (1) for which mandate is it more likely
that consensus on adoption will be achieved at all, and (2) what is
the likelihood of compliance, if mandated. 

Moreover, both questions have to be considered separately for funder
mandates and for institutional mandates, as funders and institutions
have different prerogatives. 

(Institutional consortia on mandates are yet another category, though
at a time when consensus on adopting even individual institutional
mandates is still hard to achieve, consortial consensus on mandates
seems even more difficult; the analogy with consortial consensus on
subscription licensing is, I think, very misleading. Subscription
licensing consortia are based on strong shared interests on the part
of institutional libraries, and no countervailing interests on the
part of institutional authors; author licensing mandates, in
contrast, involve the problem of authors' free choice of journals and
author risk of journal nonacceptance.) 

The reason I think an author licensing mandate has a much higher
hurdle to mount is that it raises the problem of authors' free choice
of journals and author risk of journal nonacceptance, whereas author
deposit mandates face only the inertia about doing the few extra
author keystrokes required -- and both surveys and outcome studies
on actual practice show that most authors will comply, willingly.

      Both tools only apply to the domain of toll-gated
      publishing where they try to improve the accessibility of
      publications. It is the copyright owner who decides about
      the conditions of access and reuse and the toll-gated
      domain is characterized by many access limitations and
      conditions that only may be lifted after payment.
      However, there is an important legal exception to that
      model; the fair use clause states that these access
      limitations do not apply for a personal copy.

Digital documents that are made freely accessible on the web can be
accessed, read on-screen, downloaded, stored, printed-off, and
data-crunched by any individual user. (They can also be harvested by
harvesters like google.) That is all the use that researchers need,
and that is all the use that ("Gratis") OA need provide.

      In the self-archiving approach the author assigns the
      full copyrights to a publisher and subsequently utilises
      the fair use clause to facilitate access to the
      publication. The licence to publish leaves the copyrights
      with the author, gives the publisher the right of first
      publishing and adopts an embargo period for other
      publishing modes.

This is incorrect, I am afraid. In the self-archiving approach, the
author makes the article freely accessible online, and the rest comes
with the territory. As I said, this is done with the official
blessing of the journal for 63% of journals, providing full OA for
those articles. For the remaining 37%, the author has the option of
depositing in Closed Access and letting users rely on the "email
eprint request" button during any publisher embargo. This is "Almost
OA" -- and once immediate-deposit mandates are universally adopted,
over and above immediately providing 63% OA + 37% almost-OA, they
will soon usher in 100% OA as a matter of natural course.

In my view, it makes no sense at all to delay still further the
certainty of providing 63% OA + 37% almost-OA (by mandating
immediate-deposit), to wait instead and try to adopt a much stronger
mandate (mandatory author licensing) for which consensus on adoption
is much harder to achieve -- because of the problem of authors' free
choice of journals and author risk of journal nonacceptance -- simply
because of the possibility of 37% almost-OA owing to publisher

63% OA + 37% almost-OA is already fully within reach and long
overdue. We should not delay grasping it for one minute longer, in
quest of something stronger yet not now within rich and far less

      The Immediate-Deposit/Optional Access (ID/OA) Mandate:
      Rationale and Model

(Note that there can be -- and are -- stronger self-archiving
mandates than ID/OA, but that ID/OA is the default option, the
mandate for which consensus on adoption is most easily achieved and
all legal concerns are completely mooted. If success can be achieved
on adopting a stronger mandate -- such as Immediate
Deposit/Immediate-OA, or a 6-month cap on embargoes, or Immediate
Deposit plus Author Licensing with Opt-Out -- then by all means adopt
the stronger mandate. But on no account delay the adoption of the
weaker, certain mandate that is already within immediate reach, in
order to hold out for a stronger, uncertain mandate that is not!

      For a fair comparison of the two tools, let's assume that
      in both cases an institutional mandate applies.

I am afraid that that is not at all a fair comparison, because it
presupposes (a) that consensus on adopting the mandate is equally
probable in both cases, weaker and stronger, and that (b) compliance
is equally probable in both cases, weaker and stronger. Neither of
these assumptions is correct because of the problem of authors' free
choice of journals and author risk of journal nonacceptance. Neither
handicap is shared by the weakest, easiest and surest IDOA mandate,
which is only about keystrokes, and already fully within all
institutions' immediate rich.

      When it comes to mandating self-archiving, the only party
      involved is the author. That makes such a mandate
      relatively easy of course. But it also has a high price.
      Open Access remains to the publisher's discretion.

The weakest mandate is always relatively the easiest, but it is
infinitely preferable to no mandate at all, and to continued delay
and debate instead about stronger mandates. Let the weaker mandates
be adopted universally, now, and then let's debate about
strengthening them!

What the ID/OA guarantees, immediately, is 63% OA + 37% almost-OA.
Every day we keep delaying and speculating about stronger mandates,
less certain of consensus and compliance, we are simply losing yet
another day of 63% OA + 37% almost-OA, needlessly, after already
having allowed years and years of needless delay, and needlessly lost
research access, progress and impact.

      Currently that's a complete mess. Publishers' policies
      vary widely when it comes to permitting access to
      different versions (pre-print, post-print, pdf) for
      different uses (author's web site, institutional window,
      educational usage, commercial usage) after different
      embargo periods. In the meantime for personal copies an
      end user may use the request button in the same way as
      she uses the SFX button of her library. (Why not combine
      the two buttons?). Under the circumstances the request
      button is a smart invention. Kudos for you!

(1) The variation in publisher policy is completely described and
covered by 63% OA + 37% almost-OA for OA's target content: the
peer-reviewed postprint. (For the preprint, the figure is even
higher; and the publisher's proprietary PDF is completely

(2) The SFX button is only for licensed (i.e., toll-gated) content
for institutional-internal users. The "email eprint request"
("Almost-OA") Button is for all IR-deposited content, worldwide, for
any user. There is a world of difference there (and again, the
institutional-library subscription-licensing perspective is
profoundly misleading, just as it is with the analogy of
multi-institution subscription-licensing consortia.)

      When an institution considers mandating the usage of the
      licence to publish they should involve the publishers as
      well. It would be unfair just to issue such a mandate and
      leave the authors to the mercy of the publishers.

(a) Institutions' employees, and funders' fundees can be mandated,
but publishers cannot: They are not within employers' and funders

(b) A big funding agency like NIH might be in a position to invite
publishers to the table, and to apply their clout, but individual
universities certainly cannot.

(c) Institutional library consortia have clout on subscription
licensing agreements, but author licensing is another matter, and
involves another party: authors.

(d) Meanwhile, authors would indeed be left "to the mercy of their
publishers"  (because of the problem of authors' free choice of
journals and author risk of journal nonacceptance) if individual
institutions adopted the stronger author licensing mandates without
prior successful consortial negotiations.

(e) So let's go ahead and adopt the weaker IDOA mandate that is
already within universal reach, at no risk to the author; and let's
pursue stronger mandates thereafter, rather than delaying OA still
longer, this late in the day, to hold out for a possible future

      It's my guess that negotiations with publishers may not
      be prospectless. A common interest, not only for authors
      and their institutions but also for (some) publishers is
      to raise their social and academic profile and clear the
      operational situation. In order to have a stronger
      position institutions should combine their efforts in
      (national) consortiums. By the way, I allready know of
      several occasions where a publisher (including Elsevier
      and even Wiley) has published articles without the
      copyrights being transferred to them.   

(i) It is not at all evident that the interests in and benefits from
OA to authors, institutions and funders are matched by corresponding
interests and benefits to publishers!

(ii) But by all means, if such consortial negotiations are "not
prospectless," pursue them: But not at the price failing to grasp
what is already immediately within reach, which is individual
institutional (and funder) Green OA self-archiving mandates.

(iii) (The occasional individual exception to copyright transfer that
has been accommodated for decades by publishers is not the same as a
blanket acceptance of universal copyright retention.)

      To conclude. Indeed, in the toll gated domain I prefer
      mandating the usage of the licence to publish over
      mandating of self-archiving. The first option involves a
      higher commitment of the institutions which makes it
      tougher of course. But the operational result is much
      clearer and better sustainable.

It is fine to prefer to have a stronger benefit rather than a weaker
one if both are within reach and you have a choice; but it is
certainly not fine to fail to grasp a weaker benefit that is already
fully within reach in order to hold out for a stronger but much less
certain benefit that is not yet within reach. 

Le mieux est l'ennemi du bien.
Stevan Harnad

      Stevan Harnad wrote:

            On 1-Dec-08, at 5:55 AM, leo waaijers wrote
            (in SPARC-OAForum:

      Dear Stevan,

      Most authors do not self-archive their
      publications spontaneously. So they must be
      mandated. But, apart from a few, the
      mandators do not mandate the authors. In a
      world according to you they themselves must
      be supermandated. And so on. This approach
      only works if somewhere in the mandating
      hierarchy there is an enlightened echelon
      that is able and willing to start the
      mandating cascade.

Leo, you are quite right that in order to induce authors
to provide Green OA, their institutions and funders must
be induced to mandate that they provide Green OA
(keystrokes). Authors can be mandated by their
institutions and funders, but institutions and funders
cannot be mandated (except possibly by their governments
and tax-payers), so how to persuade them to mandate the

The means that I (and others) have been using to persuade
institutions and funders to mandate that authors provide
OA have been these:

(1) Benefits of Providing OA: Gather empirical
evidence to demonstrate the benefits of OA to the author,
institution, and funder, as well as to research progress
and to tax-paying society (increased accessibility,
downloads, uptake, citations, hence increased research
impact, productivity, and progress, increased visibility
and showcasing for institutions, richer and more valid
research performance evaluation for research assessors,
enhanced and more visible metrics of research impact --
and its rewards -- for authors, etc.).

(2) Means of Providing OA: Provide free software for
making deposit quick, easy, reliable, functional, and
cheap, for authors as well as their institutions.
Provide OA metrics to monitor, measure and reward OA and
OA-generated research impact.

(3) Evidence that Mandating (and Only Mandating) Works:
Gather empirical data to demonstrate that (a) most
authors (> 80%) will deposit willingly if it is mandated
by their institutions and/or funders, but they will not
deposit if it is not mandated (< 15%)  (Alma Swan's
studies); and that (b) most authors (> 80%) actually do
what they say they would do (deposit if it is mandated [>
80%] and don't deposit if it is not mandated [< 15%] even
if they are given incentives and assistance [< 30%]
(Arthur Sale's Studies).

(4) Information about OA: Information and evidence about
the means and the benefits of providing OA has to be
widely and relentlessly provided, in conferences,
publications, emails, discussion lists, and blogs. At the
same time, misunderstanding and misinformation have to be
unflaggingly corrected (over and over and over!) 

There are already 58 institutional and funder Green OA
mandates adopted and at least 11 proposed and under
consideration. So these efforts are not entirely falling
on deaf ears (although I agree that 58 out of
perhaps 10,000 research institutions [plus funders]
worldwide -- or even the top 4000 --  is still a sign of
some hearing impairment! But the signs are that audition
is improving...)

      To create such a cascade one needs water
      (i.e. arguments) and a steep rocky slope
      (i.e. good conditions). The pro OA arguments
      do not seem to be the problem. In all my
      discussions over the last decade authors,
      managers and librarians alike agreed that the
      future should be OA also thanks to you, our
      driving OA archevangelist.

But alas it is not agreement that we need, but mandates
(and keystrokes)! And now, not in some indeterminate

      So, it must be the conditions that are
      lacking. This awareness brought me to the
      writing of an article about these failing
      conditions. Only if we are able to create
      better conditions mandates will emerge and be
      successful on a broad scale. A fortiori, this
      will make mandates superfluous.

I am one of the many admirers of your splendid efforts
and success in the Netherlands, with SURF/Dare, "Cream of
Science," and much else.

But I am afraid I don't see how the three recommendations
made in the Ariadne article will make mandates emerge
(nor how they make mandates superfluous). On the
contrary, I see the challenge of making the three
recommendations prevail to be far, far greater than the
challenge of getting mandates to be adopted. Let me

            Recommendation 1: Transferring
            the copyright in a publication
            has become a relic of the past;
            nowadays a "licence to publish"
            is sufficient. The author retains
            the copyrights. Institutions
            should make the use of such a
            licence part of their
            institutional policy.

Persuading authors to retain copyright is a far bigger
task than just persuading them to deposit (keystrokes):
It makes them worry about what happens if their publisher
does not agree to copyright retention, and then their
article fails to be published in their journal of

Doing the c.  6-minutes-worth of keystrokes that it takes
to deposit an article -- even if authors can't be
bothered to do those keystrokes until/unless it is
mandated -- is at least a sure thing, and that's the end
of it. 

In contrast, it is not at all clear how long copyright
retention negotiations will take in each case, nor
whether they will succeed in each case.

Moreover, just as authors are not doing the deposit
keystrokes except if mandated, they are not doing the
copyright retention negotiations either: Do you really
think it would be easier to mandate doing copyright
retention than to mandate deposit? 

(Harvard has adopted a kind of a copyright-retention
mandate, though it has an opt-out, so it is not clear
whether it is quite a mandate -- nor is it clear how well
it will succeed, either in terms of compliance or in
terms of negotiation [nor whether it is even thinkable
for universities with authors that have less clout with
their publishers than Harvard's]. But there is a simple
way to have the best of both worlds by upgrading the
Harvard copyright-retention mandate with opt-out into a
deposit mandate without opt-out that is certain to
succeed, and generalizable to all universities -- the
Harvards as well as the Have-Nots. To require successful
copyright renegotiation as a precondition for providing
OA and for mandating OA, however, would be needlessly and
arbitrarily to raise the goal-post far higher than it
need be -- and already is for persuading institutions and
funders to mandate deposit at all.)

      Upgrade Harvard's Opt-Out Copyright Retention
      Mandate: Add a No-Opt-Out Deposit Clause

            Recommendation 2: The classic
            impact factor for a journal is
            not a good yardstick for the
            prestige of an author. Modern
            digital technology makes it
            possible to tailor the
            measurement system to the author.
            Institutions should, when
            assessing scientists and
            scholars, switch to this type of
            measurement and should also
            promote its further development.

This is certainly true, but how does using these
potential new impact metrics generate OA or OA mandates,
or make OA mandates superfluous? On the contrary, it is
OA (and whatever successfully generates OA) that will
generate these new metrics (which will, among other
things, in turn serve to increase research impact, as
well as making it more readily measurable and

      Brody, T., Carr, L., Gingras, Y., Hajjem, C.,
      Harnad, S. and Swan, A. (2007) Incentivizing
      the Open Access Research Web:
      Publication-Archiving, Data-Archiving and
      Scientometrics. CTWatch Quarterly 3(3).

      Harnad, S. (2007) Open Access Scientometrics
      and the UK Research Assessment Exercise. In
      Proceedings of 11th Annual Meeting of the
      International Society for Scientometrics and
      Informetrics 11(1), pp. 27-33, Madrid, Spain.
      Torres-Salinas, D. and Moed, H. F., Eds.

      Harnad, S. (2008) Validating Research
      Performance Metrics Against Peer Rankings.
      Ethics in Science and Environmental Politics
      8 (11) doi:10.3354/esep00088  The Use And
      Misuse Of Bibliometric Indices In Evaluating
      Scholarly Performance

            Recommendation 3: The traditional
            subscription model for
            circulating publications is
            needlessly complex and expensive.
            Switching to Open Access,
            however, requires co-ordination
            that goes beyond the level of
            individual institutions.
            organisations, for example the
            European University Association,
            should take the necessary

The European University Association has already taken the
initiative to recommend that its 791 member universities
in 46 countries should all mandate Green OA
self-archiving! Now the individual universities need to
be persuaded to follow that recommendation. The European
Heads of Research Councils have made the same
recommendation to their member research councils. (I am
optimistic, because, for example, 6 of the 7 RCUK
research funding councils have so far already followed
the first of these recommendations -- from the UK
Parliamentary Select Committee on Science and Technology.
And the 28 universities that have already mandates show
that institutional mandates are at last gathering
momentum too.

But if it is already considerably harder to mandate
author copyright-retention than it is to mandate author
self-archiving in their institutional repositories (Green
OA), it is surely yet another order of magnitude harder
to mandate "Switching to Open Access" from the
"traditional subscription model." 

If author's are likely to resist having to renegotiate
copyright with their journal of choice at the risk of not
getting published in their journal of choice, just in
order to provide OA, they are even more likely to resist
having to publish in a Gold OA journal instead of in
their journal of choice, just in order to provide OA. 

And journal publishers are likely to resist anyone trying
to dictate their economic model to them. (Moreover, this
goes beyond the bounds of what is within the university
community's mandate to mandate!) 

So mandating Green OA is still the fastest, surest, and
simplest way to reach universal OA. Let us hope that the
"enlightened echelon" of the institutional hierarchy will
now set in motion the long overdue "mandating cascade."

Best wishes,

Stevan Harnad
Received on Wed Dec 03 2008 - 01:53:15 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:36 GMT