Saturday, June 28, 2014

The Subversive Proposal at 20

Twenty years ago yesterday, cognitive scientist Stevan Harnad posted a message on a mailing list, a message he headed A Subversive Proposal. This called on all researchers to make copies of the papers they published in scholarly journals freely available on the Internet.

The message sparked a protracted discussion, and eventually led to the publication of a book called Scholarly Journals at the Crossroads: A Subversive Proposal for Electronic Publishing.

Today the Subversive Proposal is viewed as one of the seminal texts of the open access movement.

To celebrate the 20th anniversary of the Subversive Proposal, I emailed Harnad nine questions yesterday. These questions are published below, with Harnad’s answers attached. 
Stevan Harnad

Q&A

RP: Today is the 20th anniversary of the Subversive Proposal, a 496-word online message you posted to a mailing list on June 27th 1994 in which you called on researchers to make copies of all the papers they published in scholarly journals freely available on the Internet. The message sparked a heated online debate that later formed the basis of a book. What stimulated you to make that posting, and why do you think it attracted as much attention and disagreement as it did?

SH: Two things impelled me to do it:

(1)   I had been editing a journal of open peer commentaryBehavioral and Brain Sciences — for 16 years at the time, and had always had the feeling that the print-on-paper medium was not the optimal medium for scholarly communication.

(2)   I also had a strong belief in the creative power of interactive written dialogue, which became even stronger with the advent of the online medium. (I had dubbed this “scholarly skywriting.”)

For scholarly skywriting to work, it has to be accessible online. But although I knew about the price of subscriptions and the serials crisis at the time, that was not my primary motivation: open online access and interaction was (and still is). (I explained this more fully in your 2007 interview.)

As to attention: I’d have much been much happier if it had attracted action rather than just attention! The disagreement (which is always welcome, and can even be creative) was about the things we will go on to discuss further below: Green vs. Gold OA and, to a lesser extent, Gratis vs. Libre OA.

RP: Looking back, what contribution would you say the Subversive Proposal has made to the development of the OA movement, which in fact really only became a movement 7 years later (in 2001), when the term open access was adopted at the meeting where the Budapest Open Access Initiative was planned and articulated?

SH: I’m not sure. What I tried to urge all scholars to do in 1994 (self-archive their journal articles) some had already been doing for years (notably computer scientists in anonymous FTP archives since the 1980s and physicists in arXiv since 1991), but I’m not aware that the self-archiving rate increased appreciably after my proposal. The proposal may have created a bit of a flurry, but it was a notional flurry: it was not heeded when it came to actual action (self-archiving).

At the 2001 BOAI meeting, self-archiving got a name — it became “BOAI OA Strategy I” (later dubbed “Green OA”).

“BOAI OA Strategy II” was OA journal publishing (“Gold OA”) and that option (though it too was mentioned in the Subversive Proposal as the likely end-game, after universal Green OA had prevailed) seems to have captured people’s imaginations more than Green OA did. In fact, across the years since 1990 authors were providing little OA at all, though of the minority who were providing OA, 2-3 times as many provided Green than Gold (and this is still true).

So, again, I don’t see much practical effect of the Subversive Proposal, either in 1994 or in the subsequent half-decade. Nor did Green OA begin to come into its own when I commissioned (and Rob Tansley created) the first free software for creating Green OA institutional repositories in 2000. BOAI helped; but the first real sign of progress came with the outcome of the 2004 UK Parliamentary Committee (which you phoned me in Barcelona to report, Richard!). The committee recommended following the proposal — by me and others — that UK research funders and universities should mandate (require) Green OA. (The Committee only recommended some experimental support for Gold OA.) After that, mandates began to grow (though still very slowly).

Gold vs. Green

RP: As you note, the Subversive Proposal invited researchers to adopt what later became known as Green OA.  Shortly before the BOAI meeting, Vitek Tracz founded the first open access publisher BioMedCentral, pioneering what became known as Gold OA. In the intervening years there has been a frequently bitter debate about the respective merits of Green and Gold OA. I realise you are an advocate for Green OA, but how would you characterise the pros and cons of these two types of OA?

SH: Pros of Gold OA: (1) Gold OA is immediate. (2) Gold OA can be made not just Gratis OA (freely accessible online) but also Libre OA (freely accessible online plus further re-use rights such as data-mining, re-mixing and re-publishing). (3) Gold OA could solve the journal affordability problem.

Cons of Gold OA: (1) Gold OA costs extra money (author publication fees), over and above what institutions already pay for subscriptions as long as subscription journals prevail (and they still do). (2) Gold OA payment to publish risks a decline in journals’ quality standards for acceptance (because journals are paid by authors to publish their work, not by users to access their work) as long as subscription journals prevail. (3) Gold OA payment cannot be mandated (required) as long as subscription journals prevail. (4) Pre-Green Gold OA is vastly overpriced (which is why I call it “Fool’s Gold”) as long as subscription journals prevail.

Pros of Green OA: (1) Green OA costs no extra money. (2) Green OA has no effect on journal quality standards. (3) Green OA can be mandated (required). (4) Green OA, once it is universally mandated by funders and institutions, can allow journal subscriptions to be cancelled, inducing all journals to cut obsolete costs (print edition, online edition, archiving, access-provision), downsize to just the provision of peer review, and convert to “Fair Gold” post-Green OA, paid for out of just a fraction of each institution’s subscription cancellation savings.

Cons of Green OA: (1) Authors do not self-archive spontaneously: like “publish or perish,” Green OA has to be mandated by their institutions and funders. (2) Publishers can (and 40% do) embargo Green OA self-archiving for 6-12 months or longer.  (3) Not all Green OA mandates are effective: it is important to adopt the most effective mandate model (which is the Liège/HEFCE mandate now also recommended after 10 years by BOAI-10).

You asked about the pro’s and con’s of Green and Gold OA. I’ve tried to list all of them. Although the numbers look balanced, I think anyone who gives it some thought will see that Green OA needs to be mandated first and Fair Gold will be scaleable and sustainable and fair only after Green OA has prevailed globally.

Mixed blessing?


RP: I think it fair to say that publishers were initially highly resistant to open access. Today, by contrast, I suspect no scholarly publisher would say that they did not support it. However, publishers clearly prefer Gold OA.  As an advocate for Green OA, would you say that publisher support for OA has been a mixed blessing? If so, why?

SH: Maximizing the access, uptake, usage, progress, productivity, applications and impact of their publicly funded research output is a research community (and tax-payer) matter, not a publishing industry matter. Publishers provide a service to the research community (the management of peer review); the web has made publishers’ other traditional service — access-provision — (along with its costs) obsolete. Journal publishers already realize this, but the research community has not yet realized it.

Journal publishers earn a great deal of revenue from subscriptions — disproportionally great; they know this too. But they would like to hold onto it for as long as possible.

Consequently, (some) journal publishers have embargoed Green OA to try to slow the growth of OA and to try to redirect it to (Fool’s) Gold OA, priced on their own terms, so as to sustain their current income levels, with the research community double-paying (subscriptions plus Fool’s Gold OA fees) until there is a full transition to Fool’s Gold on publishers’ terms. This route is not only pricey but it is extremely slow, with precious research access, uptake, usage, progress, productivity, applications and impact being continuously lost during the wait.

Green OA (universally mandated by research institutions and funders) could make the transition to 100% Green OA happen almost overnight (just as I had hoped in 1994 that spontaneous self-archiving would do), and the transition to Fair Gold OA (for peer review alone, all access-provision and archiving having been offloaded onto the global network of institutional Green OA repositories) won’t be far behind.

This is what (some) publishers are desperately trying to forestall, by embargoing Green OA and offering Fool’s Gold OA instead — under the pretext of supporting OA. These days, publishers no longer have any choice but to profess support for OA. But the research community has a choice about whether to provide OA in the publishers’ slow and expensive Fool’s-Gold way, or to provide it themselves, the Green way.  Researchers could have done it all spontaneously in 1994; let’s hope that their institutions and funders will now see to it that providing Green OA is effectively mandated before we lose yet another two decades of research access, uptake, usage, progress, productivity, applications and impact needlessly.

The HEFCE Policy

RP: As I understand it, you believe that the recently announced UK HEFCE open access policy has finally tipped the balance in favour of Green OA. There have also been a number of other green OA policies announced in the past few months, suggesting you may be right. However, as publishers are now actively promoting pure or Hybrid Gold OA (which most if not all now offer I think) and funders and research institutions have been busy creating gold OA funds, I have been wondering if we might not see many of these green policies fulfilled by means of pay-to-publish Gold OA.  Would that in your view be a good or a bad thing? Clearly it would increase the costs to the research community, but it would have the merit of providing immediate OA, and in many cases I assume it would also allow for reuse, so researchers would be free to text and data mine the papers.

SH: The UK HEFCE policy model for funders — immediate repository deposit required for eligibility for research evaluation, irrespective of whether access to the deposit is made immediately OA or OA is embargoed — together with its counterpart Liège policy model for institutions, once they are adopted globally, will guarantee 100% OA in short order (and will induce a transition to Fair Gold OA publishing not long after, with all the re-use rights users need). (Meanwhile, embargoed Green OA deposits are made almost-OA during any publisher embargo via the institutional repository’s automated request-copy Button, with which any user can request and the author can provide a copy for research purposes with one extra click each.)

But what (some) journal publishers would prefer is to continue to hold research hostage to the tolls they dictate, whether via subscriptions or via bloated Fool’s Gold OA publishing fees. Offering Hybrid Gold OA together with Green OA embargoes — a Trojan Horse to penetrate the research community — would be the optimal way for publishers to accomplish this: A subscription journal adopts a Green OA embargo to prevent its authors from providing immediate Green OA, but, by way of compensation, offers to sell authors immediate Fool’s Gold OA, for their article only, for a fee. So institutions must keep on paying subscriptions, at their current rates, until and unless all authors find extra money from some other source (usually already-scarce research funds) to pay the bloated Fool’s Gold fee: their research is held hostage to subscriptions until the research community coughs up the same inflated sums that they are paying now for subscriptions, but in the form of a Fool’s Gold fee; until then, either double payment or no OA. (And publishers offer as a sop that they will continuously adjust their prices so that they do not earn more than they did from subscriptions). 

What do I think? (not that it matters what I think): I couldn’t care less whether 100% OA is reached via Green or Gold. What I care about is that it is reached, and reached as soon as possible: it is already vastly overdue, at great cost in lost research access, uptake, usage, progress, productivity, applications and impact (at least 20 years worth). I think the research community was extremely foolish not to provide spontaneous Green OA in 1994. If they now prefer to double-pay for overpriced Fool’s Gold OA in order to prop up the revenue streams to which publishers have become accustomed, instead of just providing cost-free Green OA, that’s their own lookout. I am pushing for the universal adoption of effective mandates that ensure that all researchers do provide OA, one way or the other, now, and not another 20 years from now.

A word about re-use rights (“Libre OA”: data-mining, re-mixing and re-publishing):  I’m all for them, just as I’m all for Fair-Gold OA. I just don’t want them to get in the way of OA itself — already long overdue — as Fool’s Gold OA has done. Hybrid Gold subscription publishers are now dangling the prospect of Libre OA as an extra perk for paying for Fool’s Gold (as if that further justified embargoing Green Gratis OA). If researchers (or their funders or institutions) are able and willing to pay for this extra perk now, fine. But the fact is that funds are short and most fields don’t need Libre OA anywhere near as urgently as they need Gratis OA, whereas all fields need Gratis OA — and we are still nowhere near 100% (or 75%, or even 50%) Gratis OA yet.

So let the institutional and funder immediate-deposit mandates be complied with in any way that authors choose — cost-free Green or costly Gold — but let the immediate-deposit mandates be immediately adopted and immediately complied with, one way or the other.

Accessibility vs. Affordability


RP: OA advocacy was initially driven by two different, but related, concerns. As the people who have to pay the constantly-rising costs of journal subscriptions, librarians viewed OA as an answer to what you refer to as the affordability problem. That is, they supported OA because they assumed it would lower the costs of scholarly communication. By contrast, researchers (initially at least) viewed OA as a solution to the accessibility problem — i.e. they wanted their work to be accessible to as many other researchers as possible. But as librarians began to threaten journal cancelations the threat was that researchers would have access to fewer and fewer scholarly journals, thereby reducing accessibility. While we should not doubt that OA can solve the accessibility problem, it is far from clear today that it will also resolve the affordability problem — not least because publishers (and, it seems, many funders and governments) believe that OA should have no negative impact on the revenues of scholarly publishers. What are your current thoughts on this matter? Do you have any concern about the affordability problem, or is it only a desire for greater accessibility that motivates you?

SH: Many funders and governments? Not by my count. Only the UKs RCUK, with its Finch Folly (now detoxified by HEFCE) and the well-intentioned but rather wilful Welcome Trust, as far as I know (with some echoed intentions — sans action — from Netherlands MP Sander Dekker).

But my sole interest in the journal affordability problem is inasmuch as it impinges on the research accessibility problem. Let me try to speak very directly, because there is a lot of double-talk on both sides, concerning OA:

(1)   Some publishers profess to be all for OA, but claim that unless subscriptions are protected during any transition, they will be ruined, and both research publication and peer review will die. This is self-interested disinformation, used to lobby against Green OA and Green OA mandates. It is these publishers’ current inflated revenue streams that they are trying to protect, not research publication, peer review or even the subscription model, let alone OA.

(2)   Some librarians profess to be for OA, but in practice, their pre-emptive threat to cancel subscriptions as a journal’s Green OA percentage increases is an extremely foolish policy, and against the interests of OA. Equally foolish and false, however, is any assurance from librarians that once all journals’ articles have reached 100% immediate-OA (anarchically, via Green OA mandates) they will not cancel their journals. (Of course they will! And should! It’s the only way to force journals to cut obsolete costs, downsize and convert to Fair-Gold OA at a scaleable, sustainable price — paid for out of institutions’ windfall subscription cancellation savings.)

(3)   If governments ignored publisher lobbying and did the arithmetic properly, they would immediately see that the interests of publicly funded research vastly eclipse those of the research publishing industry, and that the benefits of immediate Green OA vastly outweigh the arguments of (some) publishers for a slow double-paid transition to Fool’s Gold OA at their current asking prices, while holding Green OA at bay with embargoes.

So my reply to your question is that I care only about solving the research accessibility problem, as soon as possible. (It is already grotesquely overdue.) I believe the fastest and surest way to OA is to mandate immediate-deposit Green OA. And I am confident that once that has generated 100% Green OA, thereby solving the accessibility problem, Fair-Gold OA will follow, solving the affordability problem.

I also believe it would be foolish to go for a direct transition to Fool’s Gold OA instead. That would solve the accessibility problem, but certainly not the affordability problem. Publishers would be getting paid as much as before. But I don’t care. At this point, the only thing that matters is precious time, so much of it already having been needlessly wasted for two decades. If 100% immediate-OA can be attained faster by throwing scarce funds at Fool’s Gold OA during a double-paid transition period held in place by embargoing Green OA, so be it  But I don’t believe for a moment that this would indeed be the fastest way to reach 100% OA.

And I hope that researchers, institutions and funders are not foolish enough to buy it.

RP: You often express surprise at the time it is taking for the world to wake up to OA, which you have long argued is inevitable and optimal. Looking back, why do you think it is taking so long for something inevitable and optimal to be realised? And what in your view could the OA movement have done/still do to speed the transition up? Or do the kind of changes required inevitably take a very long time to achieve?

SH: I do not believe there is any natural or social or psychological or neurological law that says the transition to 100% OA had to take this long. I think the reasons it has been so slow are multiple (and I’ve been enumerating them in the BOAI self-archiving FAQ since 2002 as various forms of “Zeno’s Paralysis”). If I had to pick the two most prevalent ones they would be (1) groundless researcher fears of legal consequences if they self-archive and (2) equally groundless researcher fears that self-archiving is a lot of work. Of course, (3) publisher embargoes and (4) lobbying are designed specifically to stoke such groundless fears.

The cure for all this is for funders and institutions to adopt the HEFCE/Liège immediate-deposit Green OA mandate model.

OA and the developing world

RP: OA advocates argue that open access has a great deal to offer the developing world. In saying this, they point out that the high costs of subscribing to scholarly journals means that the accessibility problem is far greater in the global South, and so OA will be all the more beneficial. But if, as increasingly seems likely, the dominant model for providing open access becomes pay-to-publish Gold OA will researchers in the global South not find that while they can read as much third-party research as they wish, they will not be able to afford to publish their own work? I realise that OA advocates dismiss this fear by pointing out that OA publishers offer APC waivers (or reduced APCs) for researchers in the developing world. Researchers in the developing world, however, respond that they do not want charity, not least because charity can at any time be removed. Perhaps they have a point. Initially, the non-profit OA publisher PLOS provided waivers on a “no-questions-asked” basis. This was withdrawn in 2010. Then earlier this year PLOS further tightened up its waiver rules. Meanwhile, Elsevier’s APC waiver policy is even more daunting, reading “If an author would like their article to be published open access, but cannot afford these fees, then individual waiver requests are considered on a case-by-case basis and may be granted in cases of genuine need”. I am wondering who would relish going cap-in-hand to Elsevier in order to make a case for “genuine need”? What are your views on this?

SH: Pre-Green Fool’s Gold is Fool’s Gold, any way you cut the cake; North, South, East or West. Right now, any institution, in any country, that wants access to the refereed research literature must pay subscriptions for as much of it as it can afford, and must make do without access to the rest. Asking them to pay Gold OA fees now, on top of what they are already paying for subscriptions that they can already ill afford would be ludicrous, if it weren’t for all the publisher disinformation, lobbying, and embargoes  coupled with a goodly dose of “gold fever” in the research community that is even more foolish than I was in imagining that the research community would immediately do what was in its own best interests in response to the Subversive Proposal in 1994.

So my reply is, yes, Pre-Green Fool’s Gold is unscaleable, unsustainable, and unfair, not just for the Developing World but for everyone who is not rolling in excess cash. Let the Developing World (and everyone else) keep publishing in the best journals they can, without paying an extra penny beyond what they are already spending on subscriptions, and let the Developing World (and everyone else) mandate immediate-deposit Green OA. Do that, and the optimal and inevitable (and long overdue) will at long last be upon us. (Is that yet another ill-fated subversive proposal, doomed to fail over the next 20 years?)

A Subversive Proposal for 2014

RP: If you were composing the Subversive Proposal today how different would it be? Would it be different? If so, would you care to rephrase it to fit today’s environment? In other words, how would the Subversive Proposal look if written for a 2014 audience (in less than 500 words)?

SH: Knowing now, in 2014, that researchers won’t do it of their own accord, I would have addressed the proposal instead to their institutions and funders, and in less than 200 words:

To maximize the access, uptake, usage, progress, productivity, applications and impact of your publicly funded research output, mandate (require) that the refereed, revised, accepted final draft of all articles must be deposited in the author’s institutional repository immediately upon acceptance for publication as a condition for research evaluation and funding. If you allow an embargo on making the deposit OA (freely accessible to all online), implement the automated almost-OA Button (and don’t let the embargo exceed 6-12 months at most). This is called “Gratis Green OA.” Do not pay for Gold OA journal publication fees (“Fool’s Gold”) until global Green OA has made subscriptions unsustainable; then you can pay for Fair-Gold out of your subscription cancellation savings. Fair-Gold will also be Libre OA (with re-use rights such as data-mining, re-mixing and re-publishing). Ignore publishers’ lobbying to the effect that Green OA will destroy peer-reviewed journal publishing: it will re-vitalize it and save the research community a lot of money while maximizing the access, uptake, usage, progress, productivity, applications and impact of their research.

RP: Thanks for answering my questions. Readers may also be interested in a 10-year review of the Subversive Proposal I wrote in 2004 — Ten Years After; and the 15-year review you wrote — The 1994 “Subversive Proposal” at 15: A Critique.

1 comment:

Stevan Harnad said...

For Paleontologists:
The 1994 Subversive Proposal at 15, 10 & 5


http://j.mp/subprop15
http://j.mp/subprop10
http://j.mp/subprop5