It is time to stop making apocalyptic predictions about the coming of the electropublication era and to start providing concrete strategies for hastening the day. Before proposing anything, I have to describe an important parting of ways that will be taking place as the literature is launched into cyberspace. The "trade" literature (for want of a better word, though Shakespeare was hardly a tradesman) will go one way, whereas the specialised scholarly and scientific research literature will go another.

The trade literature includes all texts that are written to be sold. Trade authors wish to sell their words, and readers wish to buy them. The trade literature includes everything from poetry and fiction to journalism and entertainment. Little or nothing I say here will be relevant to its publication in the electronic era, for my proposal applies only to the scholarly literature.

Scientific and scholarly research has become increasingly specialised. There are few individuals with the expertise or interest to follow and understand what is being published in any given subspecialty. Yet it is the pursuit of this specialised expertise that has brought us all the benefits of science and scholarship. These specialised pursuits are what are revealing the mysteries of the atom, the gene, the cancer cell, our language, our past, and human nature itself.

Although the readership of any particular scholarly article is small by the standards of the trade press, the size of the scholarly serial literature as a whole is substantial. The Institute for Scientific Information indexes 6,500 journals. These include the core learned journals in science and engineering and a good portion of the ones in the arts and humanities as well. However, there are many more serial publications, and some counts of them reach 130,000.

ÖйúAƬ

Most of these publications fit the criteria for non-trade literature, namely, that (1) the authors are not paid for their texts and (2) there are few readers per article. Non-trade authors not only do not expect to be paid for their words, but they are so eager to reach the eyes and minds of their fellow-specialist readership that (3) they are often willing to pay to do so, by purchasing and mailing reprints of their articles to those who request copies (and some who do not). In some fields authors also pay page charges to accelerate the publication of their work.

Why are non-trade authors prepared to go to such lengths? For them, publication is a means to a much more indirect end than being paid for their words. The scholarly/scientific reward structure looks more like this: (a) Publish a lot.

ÖйúAƬ

(b)Publish work that makes an impact.

Steps (a) and (b) help advance the scholar's career. And scholars and scientists, it must be recalled, are not looking only to advance their careers. They wish to contribute to human knowledge, and this requires that their work be noticed and built upon by their fellow scholars. This explains why scholars publish their words for free, and even pay to have them distributed all the more widely and quickly.

The first step in getting the word to one's peers, however, is to publish it at all. In the print world created by the Gutenberg invention of movable type, the only way to do this was through the mediation of the slow and expensive medium of printing and paper distribution. It was because of the high cost of this process that non-trade authors have stood ready to make the "Faustian" bargain of trading the copyright for their words in exchange for having them published. From the publishers' standpoint, the bargain was eminently fair. They asked for nothing more than they asked from trade authors, which was the right to protect the product from theft, so costs could be recovered and both author and publisher could make a fair profit. For the trade author, this bargain was not Faustian, because both he and his publisher stood to gain from it - and to lose from theft. For non-trade authors, however, there was always a conflict of interest built into the act of publishing. They want to get their words out to everyone who might be interested, but reluctantly agree to erect a price-tag as a barrier, to cover the costs (not their own, but those of their publisher) and a fair return (again not to themselves, but to their publisher).

The news is that with the advent of electronic publication, the Faustian era for non-trade authors can be brought to an end. The reason is that the publication cost - if one reckons it properly - can be much lower for purely electronic publications than for traditional print ones. It makes much more sense - and matches much better the indirect reward structure I have just described - to recover those costs (and a fair return) from those who gain the most from the much broader scope of electronic scholarly publication, namely, authors, through some version of page charges. This would enable scholarly publications to be available for free for all on the Net. This would benefit: (1) authors' universities, who gain in many ways from the the publications of their staff (in the United Kingdom not least in the form of their rankings in the Research Assessment Exercises that determine departments' level of funding); (2) authors' research funding bodies, who subsidise the research not only so that it should be performed, but so that it should be publicly reported; (3) learned societies, who collectively benefit, both as authors and readers (as does society as a whole), from a freely available learned literature; (4) university libraries, whose budgets will be perhaps the greatest immediate financial beneficiaries.

What will be the true cost in the post-Gutenberg era of purely electronic publication? Paper publishers have been estimating that it will only be 10 to 30 per cent lower than paper costs, but their figures are based on reckoning only what electronic processing can save if one continues to do things as one did them in paper. Most categories of expenses (for example, not just paper, printing and distribution, but subscription, marketing and fulfilment) vanish with freely accessible purely electronic publication. The only inherent expenses of purely electronic publication are those of (1) peer review (which requires only editorial administration, because the reviewers have always done their work for free) and (2) editing (including formatting, mark-up and archiving). My own estimate (based on experience from editing both a paper journal, Behavioral and Brain Sciences, published by Cambridge University Press, and a purely electronic one, Psycoloquy, sponsored by the American Psychological Association, is that the savings would be more like 70 to 90 per cent. Translated into annual page charges for even the most prolific author, it makes much more sense to recover these minimal costs in advance from some strategic combination of the four sources mentioned earlier, particularly in view of the enormous added value of the electronic medium compared to the paper one.

It remains only to tally up these post-Gutenberg values: If the world's non-trade scholarly/scientific literature were available to everyone for free in electronic form, the first benefit to authors would be the great increase in the visibility, accessibility and hence the potential impact of their work. Some fear that such a literature would be overwhelming and un-navigable, but stop and think: How do we currently manage it in paper? If the entire corpus were transferred to the Net, instead of our eyes and fingers and feet doing the walking to get to the papers or to get the papers to us, electronic directories containing everything could be searched using the kinds of keyword search already used in searching electronic databases that contain the titles and abstracts (but not the articles) in the paper literature. Then, one more click, and you have the paper itself! In the future, clever "knowbots" (automatic search programs) will be designed to go out instead of us and look for papers fitting our profile of interests, leaving scholars even more time to read what they want and to do their research.

ÖйúAƬ

Apart from free availability to all, there would be an advantage in greater speed, because although peer review and editing probably cannot be speeded up much beyond present rates, the time it takes for an accepted, edited paper to go to press and be distributed would be eliminated. Electronic searchability of the entire scholarly literature, hypertext links allowing readers to jump to other relevant papers, and other electronic enhancements will add still further value.

The savings in library budgets, plus other sources of support, can be used to increase the already growing global access to the Net even in poorer universities and countries. The general public, which is likewise gaining greater access to the Net, also stands to benefit from the free availability of the scholarly literature, especially in the biomedical area.

Perhaps the greatest added value of the electronic medium has not yet been mentioned. It is interactive publication. Behavioral and Brain Sciences has an extremely high impact factor (that is, articles in it are cited frequently) because, besides publishing articles, it also publishes commentaries - sometimes as many as 20 to 30 per article - from specialists across disciplines and around the world, analysing, amplifying, criticising and supplementing the target article. It is this "open peer commentary" feature that has not only given BBS its impact with its readership, but has made it such a sought-after place to publish for authors. This interactive dimension is missing from conventional publication, even though it is a natural and important aspect of learned inquiry.

ÖйúAƬ

Peer commentary is expensive to provide in paper, and not every article merits that much attention. That is why there are few journals like BBS. But is it only the 12 to 15 articles per year that BBS publishes that would benefit from peer commentary? BBS's electronic counterpart, Psycoloquy, offers peer commentary for all articles. Once peer-reviewed, accepted, and published, every paper is open to commentary, and the commentaries (and responses) are published rapidly, to keep the momentum up. (Psycoloquy, thanks to an annual subsidy from the APA, is free for all.) With electronic publication, commentaries are easy to provide.

So what is the strategy for ushering in this era? It is a simple subversive proposal that I make to all scholars and scientists right now: if from this day forward, everyone were to make available on the Net, in publicly accessible archives on the World-Wide Web, the texts of all their current papers, then the transition to the post-Gutenberg era would happen virtually overnight. Readers would quickly form the habit of accessing the free, globally available electronic versions of articles, rather than the late, remote, expensive paper ones. Publishers would be encouraged to restructure themselves for the transfer of cost-recovery to the much lower advance page-charge model rather than the subscription model.

Currently, publishers tend to experiment with what I think is a doomed hybrid model, offering paper subscribers a paper plus an electronic subscription for a bit more than the paper-alone subscription price, or an electronic-only subscription for somewhat less than the paper price; the hope is that this will provide a gradual transition to electronic-only publication, if and when demand dictates it, but always retaining the subscription model. I hope I have already given some reasons why this scenario is not in the best interests of scholars, the pursuit of knowledge, or the public that supports it all. If my subversive proposal is followed, the inevitable will arrive soon, and it will force publishers to keep pace. For if they do not adapt, scholars are likely to take matters into their own hands and create brand new electronic-only journals unencumbered by the old trade model.

Something like this is already happening in the physics community. In four short years one person, Paul Ginsparg at the Los Alamos National Laboratory, has managed to bring much of this to pass. He started out in 1991 with a proposal to exchange preprints electronically among 100 fellow high-energy theoretical physicists. Since then, the remarkable electronic preprint archive he created has grown to encompass more than half of the entire physics literature, and virtually all the papers in many fields of physics: 25,000 physicists worldwide are accessing the archive 45,000 times a day, with 350 new papers deposited per week. This is truly revolutionary, and when the post-Gutenberg history is written, Paul Ginsparg will be duly credited with having set the inevitable firmly in motion. The physics literature still faces some potential crises, however, because all those papers in Ginsparg's archive eventually appear in paper journals. It is the paper publishers who pay for the peer review and the editing. Something clearly has to be done to keep the Invisible Hand of peer review intact, to preserve the quality of the literature. Perhaps when the library revenue begins to show signs of dwindling, the publishers will begin to recognise the virtues of the electronic-only, page-charge model over the hybrid one . . .

ÖйúAƬ

Stevan Harnad is professor of psychology and director, Cognitive Sciences Centre at the University of Southampton. harnad@ecs.soton.ac.uk ftp://ftp.princeton.edu/pub/harnad/ ftp://cogsci.ecs.soton.ac.uk/pub/harnad/ gopher://gopher.princeton.edu/11/.libraries/.pujournals