Clare replies to my reply, here.

My remarks were of course provocative. As a former journal editor myself I agree very strongly that it matters a great deal who reads the articles and my statement: ‘who cares who reads the articles’ was perhaps too ambiguously ironic. It was intended to criticise certain institutional interests in quantity at the expense of any interest in the actual content of what is being written.

What worries me is academic writing that is published but not read. In this case one could argue that this work is not serving its social function and that forms of quasi censorship are in operation by requiring too hermetic a form. Some forms of journal article writing can close down the impact of academic work, rather than opening it up In this sense they operate like a closed club – or in Foucault’s terms ‘a society of discourse’ whose rules are only accessible to a very restricted few.

Of course one needs specialist discussion to advance knowledge – but if this is not being read even by other specialists then there is a problem.

Editors certainly care who reads material, because no-one wants to be the editor of a journal that no-one reads. But how can we tell? We try to keep an eye on what measures we have that can give an indication of that. Two of these are the number of times articles are downloaded, and the number of citations they receive. Neither of course is an indication that they have actually been read, though you’d hope they had been in the last case at least. But I think that the same argument about how editors try to gauge how much stuff is read is more broadly shared. For promotion and hiring decisions, citation rates can actually play quite a big role. Numbers of articles does matter, but not (I think) at the expense of what is in them. Where they are published is much more important. This is often driven by the impact factor of journals, itself a product of citation rates. (There are problems with this of course, but they are rather different from those Clare raises.)

There has been a struggle in the British system, certainly in the humanities and social sciences, not to move to a wholly metric driven system for evaluation of research outputs. The last Research Assessment Exercise panels put quite an emphasis on the reading of the submitted material before its grading. This is surely the right way to continue – rather than grading on the basis of where it was published, citation rates etc. For those outside the UK, the RAE – now replaced by the REF (Research Excellence Framework) – is basically the means by which research funding, other than grants for specific projects, is determined. What this is requires is that over a specific period (the last one ran from 2001-07) all submitted academics have to list the four pieces they have produced in that time that they/their department think represent their best work. Those pieces can be books, articles or chapters. Now, I’m not defending every aspect of this system, but four pieces in that time period puts the emphasis squarely on quality, not quantity. The model of splitting the research into more and more pieces and thereby spreading it thin to increase the volume actually counts against the academic and their department.