Sunday, March 04, 2007

Mandate debate

Peter Suber is weighing in on the mandate debate. In one of the comments on my previous post on his blog (March 3, 2007) he says the following about his own position on mandates:
"One objection is that a mandate paternalistically coerces [authors] for their own good. If true, this would be a serious problem for me, though perhaps not for everyone who defends mandates. I cannot support paternalism over competent adults....Fortunately, the paternalism objection misses the target and is easily answered....First, I only support mandates that are conditions on voluntary contracts. They might be funding contracts: if you take our money, you'll have to provide OA to your research; if this bothers you, then don't take our money. They might be employment contracts: if you work here, you'll have to provide OA to your research; if this bothers you, then don't work here....Second, I only support mandates with reasonable exceptions....Third, an OA mandate [advances other interests beyond the author's]. The [author] interest is greater visibility and impact. The university [or funder] interest is that an OA mandate will better fulfill the university [or funder] mission to share the knowledge it produces, and better assist researchers elsewhere who could benefit from this knowledge...."
Peter is a philosopher, and thus can be expected to be more careful with choosing his words than a mere mortal like me. Yet I cannot square the idea of a mandate, given its usual definition of 'an official or authoritative command; an order', with the idea of a condition, a stipulation, in a voluntary contract. If you mean starter pistol, don't say machine gun. You might confuse some people. If you mean contract stipulation, don't say mandate. Such a heavy word is, well, too 'loaded' (no pun intended).

And how voluntary is a funding contract actually? Only in the sense that if you don't sign, you have the option of leaving science altogether. In comparison, the condition in a voluntary contract that asks authors to transfer their copyright to a publisher seems a very mild and decidedly benign one, especially if the publisher is 'green'.

Jan Velterop

Saturday, March 03, 2007

Challenge for open access

(This is a long post. If you don't want to read it all, go straight to the last two paragraphs.)

Stevan Harnad has posted his “Challenge to OA Publishers” in some form or other on a number of email lists and after I responded on two lists (I chopped my response up for clarity, and to make it possible to discuss each issue he raised separately), I became aware that he has posted a similar, maybe the same, piece on other lists as well. Perhaps a response on The Parachute is more efficient than posting to all these lists. I will still separate the issues out, and my responses here will differ in some detail from the ones I have posted on the AMSCI and SOAF lists, as I now have the benefit of having received responses to my responses, as many off-line as on the lists themselves (the latter can be found in the archives of the respective lists).

I identified at least seven issues in Stevan’s piece that I think are misconceptions and misunderstandings.

Misconception 1: The idea that publishers and the research establishment are each other’s natural adversaries.

Stevan pits the interests of science publishers against the interests of "research, researchers, universities, research institutions, research funders, the vast research and development (R&D) industry, and the tax-paying public that funds the research." This seems to assume that the researchers establishment lives in a parallel universe to the one in which science publishers live – a universe which is not 'tainted' by anything that might appear to have anything to do with economics or business.

That doesn’t appear to be particularly perspicacious or observant. The interests of the global scientific enterprise and publishing enterprises are necessarily in line with one another. Stevan himself makes the point that "...research publishing [...] is a service [...]. It will have to adapt to what is best for research, and not vice versa." Quite right. Precisely because publishing is a service, the interests of the global research enterprise are in line with the interests of publishers. No service industry can survive by rendering services that are against the interests of its clientèle. In fact, publishing is so intertwined with academia that it is part of the global research enterprise. Access to – and sustainability of – formal publication channels (a.k.a. journals) are two lattices of the same clear crystal.

Somewhat cryptically, Stevan dismisses this as ideology, and adds his ceterum censeo* that an OA publisher, by definition pro-OA, cannot at the same time withhold support for a mandate to self-archive non-OA-published material. This brings us to:

Misconception 2: OA publishers opposing OA.

Stevan Harnad calls it "disappointing, if not deplorable" if OA publishers take a stance "against Open Access itself." I couldn't agree more, if that were indeed the case. But it isn't. It's an absurd notion that they are.

'Gold' OA publishers are definitely for open access. Strongly so. And they are not against 'green' (open self-archiving of authors’ manuscript versions). After all, they endorse 'green'. They are just not necessarily so fanatically for it to support a self-archiving mandate (which is not the same as an OA mandate) for non OA-published materials. Stevan seems to adhere to the idea that says: "if you're not entirely, unquestioningly, and unequivocally for an open self-archiving mandate, you're against open access." To illustrate why this is rather absurd, imagine being strongly in favour of promoting health through physical exercise. Does it follow that if you do not support a mandate for everyone to run the half-marathon every week, your health-promoting credentials are questionable?

Needless to say, Stevan’s response to the above consists of his ceterum censeo.

Misconception 3: Publishers think protecting their risks outweighs the benefits of OA.

Stevan mentions two risks that publishers face. The risk of OA self-archiving mandates undermining subscription income and the risk of authors (or their institutions and funders) not willing to pay enough for OA publishing. Perhaps unlike some tenured scientists, publishers are used to living with risk. And there are more than the ones Stevan mentions. For instance the risk of not engaging in OA at all.

When Stevan talks about the 'benefits of OA' he means the benefits of having open access to the formally published, peer-reviewed and certified literature. OA to research results themselves is easy enough. Authors can just post their work on n'importe quel web server.

Outfits that are asked to arrange this formal publication process are known as 'publishers'. The benefits of OA are the benefits of access to the formal literature. Without 'publishers' (who are not necessarily the ones currently in existence, of course), there is no formal literature. The risk to publishers (or rather, the journals that they publish) is the risk to the benefits of OA.

Stevan’s response to this point? You guessed it: ceterum censeo.

Misconception 4: Articles are a 'product', presented as a 'gift' to publishers.

Though the difference between 'product' and 'service' is somewhat artificial (some speak of a 'service product'), what publishers have provided has always been a 'service'. The service consisted - and still consists - of arranging all that's necessary to make a scientifically non-recognised piece of work (pretty much 'worthless' for the scientific establishment), into a scientifically recognised addition to the knowledge pool (a valuable piece of work, identifiable as such by the fact that it is formally published in a peer-reviewed journal).

For the purpose of communicating information it may be good enough, but for the purpose of constituting the scientific record what the author delivers is only raw material, at best a semi-product, an intermediate good.

I was criticised by Andrew Adams (of the University of Reading in the UK) for the use of the word ‘worthless’ here. He has a point and I haven’t been clear enough why I used that word. Andrew thought it was an indication of my "contempt for the scientist as author and communicator." Let me categorically say that I do not harbour the least contempt for scientists as writers and communicators. Far from it. I used the word 'worthless' in inverted commas. Informal research papers are far from worthless in my opinion. But scientific culture insists on formally published research papers for things like priority, tenure, funding, recognition of researchers and recognition of the scientific record (at least in many disciplines, and there may well be exceptions, where formal journals are indeed not necessary). If they are not formally published, they simply don't count. So informal publications are not at all worthless per se; but they are seen as pretty 'worthless' in the context of career advancement in science. Most scientists are not fortunate enough not to need to have a list of formal publications to their CV in order to earn the approbation of their fellow-scientists.

The prevailing scientific culture, world-wide, is extremely conscious of, and sensitive to, 'brand identities' of journals. Isn't that at the heart of the matter?

The author doesn't 'give' anything to a publisher, but instead, asks for a service. Stevan thinks that such a service should be delivered at “vastly reduced costs” (whatever that means). He is most welcome to set up as a publisher and do just that (in fact I think he has done so a long time ago already). There are virtually no barriers to entry for would-be publishers. Even less so for the minimalist 'administrators' of the publishing process if that is what he thinks publishing entails (the word ‘administrators’ was actually Andrew’s). Why is it, then, that such an approach hasn't taken over the position of the existing publishers like a storm?

But Stevan doesn't seem to like the risk that's associated with setting up such a service to replace existing journals, so he tries to off-load any risk to the existing publishers by getting politicians to mandate subversion.

OA publishers already offer the service he seeks. Authors have by now a wide range of journals with OA to choose to submit to. What is he waiting for? Well, authors' uptake. We all do.

Stevan responds to this point with:
It is interesting how Jan's financial analysis fits, indifferently, the writings author sell to their publishers for a fee, or against royalties, on the one hand, and, on the other hand, the writings in question here, where the author gives them to their publishers, the peer review is likewise done for free, and all publishers do is administer it, paying no fees, no royalties.

I note that the authors of fee/royalty-based writings are not interested in making their writings OA. Researchers, the authors of the give-away writings in question, are.
He misses the point. Sure enough, the content of their articles is ‘given’ to the world by researchers in the same way that this piece I’m penning here on this blog is my ‘gift’ to the world (magnanimous of me, isn’t it?). But the ‘gift’ of an author is only accepted by the scientific establishment if it comes with a ‘certificate’. If it comes with proof that it has undergone peer-review and that it has been formally accepted for publication in a journal.

Oh, and of course he adds his ceterum censeo.

Misconception 5: Expecting non-OA journals to suffer from self-archiving mandates is hypothetical, but expecting subscriptions to continue to be paid for by institutions when the content is openly and freely available is evidence-based.

In fact, both are hypotheses, the former just more logical than the latter.

The 'evidence' that subscriptions will continue is based on the situation that subscriptions to physics journals, on the whole, seem to be co-existing with their free availability in Arxiv. As evidence goes, it doesn't deserve that moniker. It's the equivalent of saying that driving under the influence is safe, just because you've done it for years without having an accident. Or giving a number of unsupervised toddlers a packet of matches and when none of their houses have burned down by the end of the week, infer that matches are safe in the hands of toddlers.

The hypothesis that subscriptions will suffer is based on the mainstream economic observation that if goods or services are easily available for free elsewhere, it will be very difficult to sell them.

Stevan response is this:
Fact (not Hypothesis): Research today is losing access, usage and impact daily, weekly, monthly, because not all researchers can afford access to all the research they can use.
I agree. I never questioned that. He also poses this:
Fact (not Hypothesis): Journals today are not losing subscription revenue because of OA self-archiving, not even in the fields where OA has been at or near 100% for years.
I agree, too. But Stevan makes it sound a rather more generally found fact than it is. He really speaks about physics and physics alone. OK, some maths as well. Long before the web, perhaps even before Tim Berners-Lee was born, physics already developed a culture of communicating via preprints, not journals. For decades, journals have been seen as the formal record only and if they communicated anything at all, it was primarily the fact that a certain ‘label of acceptance’ (the journal reference) could be added to a given article. Arxiv, now seen as a self-archiving repository, is really the electronic manifestation of the preprint circuit that was part of the physics culture.

This ‘fact’, however (which I accept as a fact), has no predictive value. Just like the fact that not having had an accident while driving under the influence cannot be taken as evidence that you never will. The fact that physics developed a preprint culture didn't mean that most other scientific disciplines developed it, too. So why would one now believe that something that might work in physics would necessarily work elsewhere as well?

Misconception 6: If an author 'pays' for the services of a publisher by handing over rights, that payment is in addition to subscription charges.

Stevan Harnad must not have understood what I said, and it's entirely possible that I wasn't clear enough. Mea culpa. (He subsequently assured me he did understand.)

No, Stevan, you can't just add these. When an author 'pays' by transferring rights, these rights only represent 'potential' money. This 'potential' money has to be converted into 'actual' money for the publisher to be able to pay his bills. That's what subscriptions do, they convert rights into money.

Why exclusive rights? 'Exclusive' here means that the same article may not be published in more than one journal. Virtually everybody in the scientific establishment agrees with that principle. Well, not absolutely everybody. Some articles appear in more than one journal. When this happens, it is frowned upon, even regarded as scientific fraud.

The notion of an author paying seems to be anathema for Stevan. He justifies this by saying that authors ‘give’ their articles away; they are not given royalties, and not even expect to receive them.

As for royalties to the author, of course they are given, and they make sense if the publisher really wants to publish the work because in his judgement he can sell it well. For instance text-books or good review-articles. For research articles this doesn't apply, because the judgement of sales potential isn’t there. In fact, it's not up to the publisher at all to decide which article to publish and which not. Just as well. Editors and editorial boards - scientists - decide, on the basis of scientific merit, not financial potential. This is as true for subscription-based journals as it is for OA journals and hybrid ones.

Stevan responds to that with this question:
“And because referees referee (for free) and editors decide, it follows that the author should not self-archive his article?”
Did I say that it did? He then continues:
“Or that the author's funder or employer should not mandate that the fundee/employee self-archive his article?”
Did I say that it did? Funders and employers can mandate what they like. And if they are aware of the potential consequences of what they’re doing, it’s entirely up to them. If they realise the value of formal, peer-reviewed journals, as an increasing number of funders do, following the lead of the Wellcome Trust, we are finding that they are prepared to create other ways to keep the journals going than via the traditional subscription system, as long as these journals offer open access. That's the way to go.

What Stevan asks – demands – is that the publishers of those journals lobby for a mandate that articles that do not contribute to the support of these journals are nonetheless self-archived in open repositories. I refer to what I said above about healthy exercise and the compulsory half-marathon.

Stevan also says
“The logic of "excluding" the right to self-archive, or to mandate self-archiving, continues to escape me. (Could it be because I keep thinking of access and impact, and you keep thinking of funding and revenues? But then why do you portray yourself as being for OA?)”
This is a version of his ceterum censeo, of course, but who is actually excluding the right to self-archiving? The publishers, who are virtually all ‘green’? There is a difference, though, between the right to self-archive and the compulsion to do so. Stevan equates ‘OA’ to ‘a mandate to self-archive’. ‘Healthy exercise’ to ‘a compulsion to run the half marathon every week’. He’s taken his eyes off the ball of the ‘end’ and fixed them firmly onto the ‘means’.

Misconception 7: The notion that OA publishing takes away from scarce research funds.

I'm tempted to start believing in one of the religions of the physics domain, parallel universes. Stevan seems to live in the universe where OA publishing - 'gold' - costs money and subscriptions don't.

In the universe where I live, formal publishing in peer-reviewed journals costs money. Luckily, Stevan agrees. In that universe, research budget allocations and research grants typically include earmarked overhead charges. These overhead charges are taken by the research institution to pay for all manner of infrastructural costs, including the library budget. From which subscriptions are paid.

Formal publication is part and parcel of research, and thus the cost of publication is part and parcel of the cost of research. Any kind of formal publishing 'eats away' a portion of scarce research funds. But unpublished research is pretty much regarded as research not done, so money on publication is generally well-spent.

-OA publishing, with an aggregate cost to the scientific establishment of X per article published (total per article: X);
-OA via self-archiving of non-OA articles, with an aggregate cost to the scientific establishment of all the subscriptions taken (necessary in a self-archiving model), amounting to X per article published, plus the aggregate cost of thousands of institutional repositories and the staffing to keep them going, amounting to Y per article (total per article: X+Y).
Which is the greatest drain on scarce research funds?

Stevan doesn’t really respond to this, but he says:
“… until and unless subscription money is no longer paying for non-OA publishing (as it is now), and can be redirected to paying for OA publishing (Gold OA), there is no payment issue in connection with OA self-archiving mandates (Green OA): The publications that are being self-archived today have been paid for. This remains true until and unless OA self-archiving ever actually does cause cancellations and makes subscriptions unsustainable. Till then, it's Green OA and nothing more to pay.”
Sounds a bit like if your parachute fails, don’t worry about it until you hit the ground. Till then, you’re alive and well. Exquisitely logical. Yet some of us would rather like to try and pull a couple of cords here and there to see if we can manage to make a soft landing after all.

Now, a challenge to Stevan Harnad cum suis. Would he be campaigning for a mandate imposed by funders, that institutions, when paying for published research literature out of any budgets that benefit from overheads taken from research grants, pay only for article charges for OA and not for subscriptions anymore?

Mandates are of course last-resort measures and my liberal inclinations would prefer persuasion over mandates any time. But should mandates really be the only possibility, the advantages of this mandate would be clear, and these are just some of them: structural open access, no 'double' payment, only e few tens of thousands of institutions to deal with instead of millions of researchers, no need for self-archiving mandates, no multiple-version publishing.

Jan Velterop

* ‘Furthermore I am of the opinion’, from Cato the Elder, who famously ended every speech with “Ceterum censeo Carthaginem esse delendam” – “Furthermore I am of the opinion that Carthage must be destroyed.”