One Small Step for NIH, One Giant Leap for Mankind

Peter Suber wrote in Open Access News:

PS: "It's one thing to argue that the NIH policy should mandate deposit in the author's institutional repository (when they have one)..."

Most universities have an Institutional Repository (IR).

Even more would, if NIH mandated IR deposit as the preferred default option.

And those universities who don't yet have an IR are only a piece of free software and a few days' sysad start-up time from having one -- and not just for their NIH output, but for all their research output, funded and unfunded, in all disciplines.

The goal of the OA movement is to make all research output OA. But it is not just the OA content itself that needs to be "interoperable": OA mandates from funders need to be interoperable with OA mandates from institutions.

Institutions are -- without exception -- the source, the providers, of all research output, worldwide.

Hence funder OA mandates should not be competing with institutional OA mandates, needlessly and counterproductively, but adapting to, facilitating and reinforcing them.

It is not at all too late to correct this small -- but crucial and easily-fixed -- bug in the recent, welcome, timely flowering of funder OA mandates, to create a synergy with the potentially far bigger blooming of institutional OA mandates that is also on the horizon (as heralded by Harvard's recent OA mandate).

NIH need merely specify that the preferred means of fulfilling the NIH OA mandate is for NIH fundees to deposit their articles in their own institution's IR, and just send NIH each deposit's URL, so that PubMed Central can harvest it therefrom.

One small step for NIH, one giant leap for mankind.

PS: "But as long as the NIH is mandating deposit in PMC, and as long as a journal meets the NIH's criteria for depositing articles on behalf of authors, then I don't see any reason why authors shouldn't take advantage of the option."

The reason is simple:

The NIH mandate, as it stands, does not scale up to providing a systematic means of covering all of institutional research output, NIH and non-NIH, funded and unfunded, across all disciplines worldwide.

NIH research output is just a small -- but extremely important -- subset of US and worldwide research output:

NIH, the world's biggest (nonmilitary) research funder, is providing a model for research funder mandates worldwide, a model that will be influential, closely watched, and widely emulated.

It is all the more critical, therefore, that the NIH mandate should be systematically scalable -- that it should interoperate coherently (rather than compete or conflict) with OA mandates from the research providers themselves -- the universities and research institutions worldwide -- as well as with other funder mandates, in other fields and other countries worldwide.

If, instead, authors and their institutions were now to begin ceding responsibility for compliance with the NIH OA mandate to their publishers as their proxies, relying on them to deposit their work in PubMed Central in their place, this would deprive the NIH mandate of any possibility of growing to cover all of research output, in all fields, worldwide. (It would also add to the compliance-monitoring and fulfillment problem that the Wellcome Trust, which has a similar funder mandate, is just now discovering -- and NIH will soon discover it too.)

Publisher proxy deposit would at the same time tighten the control of publishers over a process that should be entirely in the hands of authors themselves: the provision of supplementary free access to their give-away work for those who cannot afford paid access to the publisher's proprietary version. (Proxy deposit would also encourage publishers to charge for compliance with the NIH mandate.)

Publisher proxy deposit would also lose the three special, scalable strengths of the NIH mandate, which are (1) that the NIH mandate applies specifically to the researcher's peer-reviewed final draft (the postprint, on which restrictions are the fewest), not necessarily to the publisher's proprietary PDF; (2) that the NIH mandate is a researcher self-archiving mandate, binding on researchers (not their publishers), and based on each researcher's right (and responsibility) to maximize access to his own give-away findings; and (3) that the NIH mandate is a coherent component of a universal mandate to provide OA to all research output, not just to NIH-funded research output, in PubMed Central.

It is crucially important to remind ourselves very explicitly that what we are talking about here is just keystrokes -- i.e., about who should do the few keystrokes that make a piece of peer-reviewed research OA. We are talking, very specifically, about a few minutes' worth of keystrokes per paper (over and above the many keystrokes that already went into writing it in the first place). The natural ones to do those keystrokes are the authors themselves (or their assistants, students or assigns); and the natural place for them to do it is in their own IRs. It makes as little sense to consider offloading the task of performing those few keystrokes onto publishers (or even onto institutional librarians) as it would be to offload onto any other party the task of keying in the paper itself, in these days of personal word-processing.

So although most authors today are still not doing those few extra keystrokes of their own accord (and that is precisely the problem that the OA mandates are meant to remedy) it would be exceedingly short-sighted to propose that the remedy is to invite authors' publishers to do those keystrokes for them (possibly even at additional cost). That dysfunctional remedy is remarkably reminiscent of the grotesque degree of control over the dissemination of our own giveaway research findings we have unwittingly been ceding to our publishers throughout the paper era (the "Faustian Bargain"): the very disease that OA is meant to cure, in the online era, at long last.

And needlessly insisting on direct deposit in PubMed Central is the very heart of the problem. Yet the cure is ever so simple: NIH need merely stipulate that the preferred means of fulfilling the NIH OA mandate is for each researcher to deposit the postprint in his own university's IR and send NIH the URL.

PS: "I did object to journal deposit under the older, voluntary policy, because it gave publishers the decision on the length of the embargo. Under the new policy, however, the length of the embargo is already set by the time the author signs the copyright transfer agreement. Hence, journal deposit cannot change the terms of the deal."

That leaves only the six other serious reasons militating against publisher deposit: (1) Publisher proxy deposit in an external repository needlessly competes with institutional IR self-archiving mandates instead of facilitating them; (2) it defeats the benefits of an immediate-deposit mandate, where the IR's "email eprint request" button could have tided over worldwide research usage needs during any publisher embargo by providing almost-immediate, almost-OA; (3) it loses the benefits of having specified that the OA deposit target is the author's postprint, not necessarily the publisher's PDF; (4) it leaves publishers in control of providing OA (and even facilitates their charging for it); (5) it leaves IRs empty, and non-NIH content non-OA; (6) it leaves researchers' fingers paralyzed.

PS: [update] "My response above was limited to publishers who do not charge fees, and I share Stevan's objections to those who would charge fees."

My objections are not just limited to publishers who charge fees: They concern any publisher proxy deposit, and indeed any funder mandate that does not stipulate that the author's own institutional IR is the preferred default locus for deposit wherever possible.

PS: "Or if there's some subtle way in which it can, then I'll join Stevan's call on authors to make the deposits themselves. I already agree with him that, if the policy were to mandate deposit in the author's IR, then author deposits would make much more sense than journal deposits."

Peter, the ways in which both publisher proxy deposit and direct institution-external deposit are counterproductive for the growth of OA and OA mandates are far from subtle. I fervently hope that you will support my call on authors (or their collaborators) to make the deposits themselves, preferably in their own IRs, providing NIH with the URL. And of course also the call on NIH to allow -- indeed welcome -- IR deposit and PubMed Central harvesting rather than just direct PubMed Central deposit. (And, while we're at it, the call on universities, like Harvard, to mandate deposit, without opt-out, rather than just mandating copyright-retention, with opt-out!)

This slight change in the implementational details of the NIH policy would be a small step for NIH, but a huge step for the growth of OA worldwide.

Stevan Harnad
American Scientist Open Access Forum