Stevan Harnad

Unrestricted access to the work of researchers will benefit all.

MOST OF the 2.5 million articles published yearly in our planet's 24,000 research journals are inaccessible to a large portion of their potential users worldwide, but especially to those in the developing world. One might think that the reason for this is that no research institution can afford to subscribe to all 24,000 journals, and most can only afford a fraction of them. And this is true, but it is not the whole story, nor the main part of it. For, even if all those journals were sold at cost — not a penny of profit — they would still remain unaffordable for many of the research institutions worldwide. The only way to make all those articles accessible to all their potential users is to provide open access to them on the Web, so anyone can access them, anywhere in the world, at any time, for free.

One could have said the same of food, medicine, and all other human essentials, of course, but one cannot eat digital food or cure diseases with strings of 0s and 1s. Nor, alas, are all the producers of digital products — let alone of physical food or medicine — interested in giving away their products for free. So what makes research different (if it is different) and why is it urgent for all of its potential users to have access to it?

Research is the source from which future improvements in the quality, quantity, and availability of food, medicine, technology, and all other potential benefits to mankind will come, if they are to come at all. And researchers — unlike the producers of commercial products — give their findings away. Unlike writers or journalists, researchers do not seek or get fees or royalties for their articles. They give them to their journals for free, and they even mail (and these days email) free copies to any potential user who writes to ask for one.

Why do they give their research away for free? Partly for the same reason they are researchers rather than businessmen. They want to make a contribution to knowledge, to research progress. Partly also because that is the nature of the reward structure of science and scholarship. Research is funded, and researchers are employed and paid, on the strength of their "research impact." This used to mean how much they publish, but these days it also means how much their publications are read, used, and built upon, to generate further research and applications, to the benefit of the tax-paying society that funds their research and their institutions.

And now we can see both why researchers give away their articles and why it is so important that all their potential users should be able to access and use them. Because all access-barriers are barriers to research progress and its benefits (as well as to the advancement of researchers' careers and productivity).

Researchers are not businessmen, but they are not always very practical either. The reason publications need to be counted and rewarded by their employers and those providing funds — "publish or perish" — is that otherwise many researchers would just put their findings in a desk drawer and move on to the next piece of research. (That is part of the price that humanity must pay for nurturing a sector that is curiosity-driven rather than profit-driven.) So, since researchers do need to fund their research and to feed themselves and family, their publications are counted and then rewarded proportionately. But counting publications is not enough. It has to be determined whether the research was important enough to have been worth doing and publishing in the first place; its "research impact" has to be measured. What was its uptake, usage, influence? How much further research and applications did it generate? Although the measure is crude, and far richer measures are under development, citation counts — the number of times an article is cited by other articles — are an indicator of research impact.

So, along with publications, citations are counted, in paying researchers and funding their research. And recent studies have shown that the citation counts of articles that are freely available on the web (Open Access) are far higher than the citation counts of articles that are only available to those researchers whose institutions can afford a subscription to the journal in which they were published.

One would think, in view of these findings, and the fact that researchers give away their articles anyway, that researchers would all be making their published articles open access by now — by self-archiving their articles in their own institution's online repositories, free for all. Ninety-four per cent of journals already endorse self-archiving by their authors. Yet in fact only about 15 per cent of researchers are self-archiving their publications spontaneously today. Perhaps that is about the same percentage that would be publishing at all, if it were not for the "publish or perish" mandate. So it is obvious what the solution is, for research and researchers worldwide, in the online era: the existing publish-or-perish mandate has to be extended to become a "publish and self-archive" mandate.

International surveys have shown that 95 per cent of researchers would comply with a self-archiving mandate. Seven research institutions worldwide (two in Australia; two in Switzerland; one of them CERN; one in Portugal, one in the U.K.; and one in India, the National Institute of Technology, Rourkela) have already mandated self-archiving, and their self-archiving rates are indeed rapidly climbing from the 15 per cent baseline towards 100 per cent.

But those are spontaneous institutional mandates, and there are only seven of them so far. There are also a few systematic national mandates: four of the eight U.K. research-funding councils and the Wellcome Trust have now mandated self-archiving. And there are several other national proposals to mandate self-archiving, by the European Commission, a Canadian research council (CIHR), and all of the major U.S. funding agencies (FRPAA).

There is no need, however, for developing countries to wait for the developed countries to mandate self-archiving. Developing countries have even more to gain — both in the impact of their own research on the research of others and in their own access to the research of others — because currently both their access and their impact is disproportionately low, relative to their actual and potential research productivity.

In the past few years there have been many abstract avowals of support for the principle of open access but these have all merely declared that providing open access is a "good thing" and "should be done" — without saying exactly what should be done, and without committing themselves to doing it!

What the whole world needs now is concrete commitments to open access. Under the guidance of India's tireless open access advocate, Subbiah Arunachalam, there will be a two-day workshop on research publication and open access in Bangalore on November 2 and 3, at which the three most research-active developing countries — India, China, and Brazil — will gather in order to frame the "Bangalore Commitment." A commitment to mandate open access self-archiving in their own countries and thereby set an example for emulation by the rest of the world.

(The writer, born in Hungary, is Canada Research Chair in Cognitive Sciences, Université du Québec à Montréal; adjunct Professor, Electronics and Computer Science, Southampton University, U.K.; Past President, Society for Philosophy and Psychology; Corresponding Member, Hungarian Academy of Science; and author of over 200 publications on categorisation, communication and cognition.)