Humpty Dumpty

Imagine an egg that has just rolled off the edge of a table. Imagine interviewing the egg halfway through its descent. How would it feel? Anxious about an uncertain future perhaps, but so far fine and not aware of any evidence of danger. Now interview the egg after it’s hit the floor. You can’t. And it’s too late to do anything about it. You can’t put the egg back together again. It’s potentially the same with scholarly journals and institutional repositories (IRs). Those who choose to can say, ‘Look – there’s no evidence of any damage’, ‘Don’t worry – it’ll all be fine.’ But what if they are wrong? What if filling IRs with journal content does mean that the libraries that used to buy the journals stop doing so, and the journals go out of business? What will we do then? And by ‘we’ I don’t mean just publishers. At the launch of its report on the RIN/DTI/RCUK EvidenceBased Analysis of Scholarly Journal Publishing, Michael Jubb, Director of the Research Information Network in the UK, reported that ‘a marginal increase in efficiency of communication could have dramatic effects on the impact of research on the innovation system (of the UK)’. He suggested benefits for the UK of $171m based on the lowest assumptions of improvement in efficiency, and $3.6bn on mid-range assumptions.1 What he was saying was clear: the scholarly communication system is a sensitive one and small improvements can yield significant benefits. But sensitive systems usually work both ways – the wrong small changes can be disastrous. The scholarly journals system (by which I mean the discipline-specific management of peer-review, collection, and dissemination of research articles, regardless of business model) is complex, delicate, and highly interdependent. Publishers, librarians, editors, societies, and authors all want it to be as effective as possible in disseminating the outputs of research. We may well disagree from time to time about how to achieve that increased effectiveness, but we all want it. The peer-review system that underpins scholarly research across the world is largely run by journals and their editorial teams, and its administration is largely financed by the revenue from the journals. Remove that revenue and the structure of the journals will probably disintegrate. Then there will be no mechanism for the discipline-specific management of peer review. Journals as brands and sources of publishers’ revenues will go. But those same journals are the mechanisms by which peer review is managed. So that will go too. What will we do then? We need to move forward. IRs do exist, and they are slowly being filled with content. Publishers cannot wish them away. But journals also exist, and advocates of IRs should not simply assume that the peer review paid for by those journals will continue even if the journals don’t. Nor can any of us assume that a lack of evidence means anything. The risk is too great. We all need to adopt a precautionary principle – not just publishers, but everyone in the scholarly information chain. Yes, we want progress and improvement. But we have to balance the potential reward against the risk. And the risk that IRs will destroy the peer-review system is real. Any attempts to improve or modify the scholarly communication system should surely start from the principle of not harming peer review. If we can achieve improvements in dissemination (and thus in research productivity) without harming peer review, then we should seek to do so. And in fact this is what publishers have done, and continue to do, through offering big deals to libraries and consortia, through programmes like HINARI and INASP, and through allowing deposit of articles some time after publication. Big deals have doubled the number of journals available in UK higher education libraries.2 But where a proposed improvement might harm peer review, or where we are not certain that it will not harm peer review, then we need to proceed with caution. IRs and mandatory deposit represent just such a proposed improvement. Given that IRs are not going to go away, we need an urgent debate about how to ensure the continued survival of peer review. That debate needs to include the proponents of IRs and 154 Personal View