Tuesday, April 23, 2013

Are Linguistic Journals Tombstones?

Paul Krugman notes (here) that within economics, dissemination of research is  is not via the journals.  Rather these act like repositories (he calls them 'tombstones') of stale information and is used mainly for "validating your work" and citable "when seeking tenure" but not the place to go to "keep up with what's happening now." I've tended to regard ling journals in the same way: their prime function is not to lead the investigative pack, but to bless work that has been done.  However, given the time and effort generated to reviewing, editing and publishing journals of record, it behooves us to consider whether or not all the effort is worth it. If the aim is to support research, it seems that far less formal venues are already doing most of the heavy lifting.  How important are journals in this context and are they worth the time, money and effort demanded? Are our journals mainly a service for deans and university presidents?  Would their importance go to zero were tenure to disappear?  Just asking.


  1. It does seem like less-formal venues that are widely read (lingbuzz comes to mind) can take the place of journals in terms of disseminating research, as they allow papers to appear and interested researchers to become aware of them and comment on them. The main utility I notice in the blessed forms that eventually come out in journals is that they tend to be cleaner and easier to follow than the originals. In that sense, the journals are useful because they generate a more beautiful final product. Also, there's currently some sense (I think) that if it comes out in a journal, it's been sufficiently vetted by someone knowledgeable.

    On the other hand, the subjectivity that's involved in the selection of which papers to publish at all in journals seems more problematic - my personal experience has pretty much always been that if you try to publish something interesting and/or interdisciplinary, the debates that ought to happen in print happen behind the scenes, often resulting in rejection. (Though maybe this is a sign that you shouldn't try to publish anything really interesting or interdisciplinary until you're got sufficient standing in the field to exert your influence over the review process. My recent experience at Lingua has certainly taught me that one.)

    Computer science seems to have a model where conference proceedings take the place of journal publications for the most part - they're still peer reviewed, but they definitely have the feel of working papers most of the time. Very few journal publications are actually made, and so tenure isn't based on them. On the other hand, I think there's still bean-counting for tenure in the form of the money you bring in - so it's a trade-off.

    My feeling is that if tenure disappeared (and also, if grant awarding agencies respected non-journal publications), we'd have more venues like lingbuzz taking over, though. I think the more-beautiful final product isn't worth the lag time for most researchers.

  2. I find it interesting that you found the final products more readable and well crafted. In my experience (or more likely in my recollections of my experience) I have found that often Nels proceedings that later became papers were not improved by the process. There is a tendency to CYA in the journal published version, no doubt partly in response to captious reviewer comments. However, one effect of this is to obscure the main insight (if there is one) as the process ends up burying any such in prolix technicalia or reservations. This is too bad. However, this may not be the rule, and clearly it has not been for you. Interesting comments.

  3. Having been editor at Syntax for 7 years, I think I became convinced that the review process really did improve a great many of the papers, and a lot of authors said as much to me (although obviously not everyone was happy!). My own experience of publishing varies by journal, so I've seen my own papers really improved by the process in a number of cases, but not in others. Lingbuzz etc are great resources for distribution (although I find the top ten hit parade a bit useless and a bit annoying), but don't get the improvement that can emerge from a journal editorial process.

    A big issue is how much time editors have to really engage with papers, though and to put the effort into innovations (say, online open review). I stopped at Syntax partly because it was just too time intensive to do properly, for me at least (but partly because I was losing faith in the publishing model), and since its just an extra you add on to an already intensive job, it ends up being deprioritized. But I do think that journals can have a place in improving the quality of the work produced in at least some sub disciplines of linguistics.

    I do agree with Lisa that more interdisciplinary work is harder though. I had a recent experience with Language of trying to publish a piece with some sociolinguistics colleagues, showing how bringing together formal and quantitative models could improve our insights into particular phenomena, and the reviewers and editors struggled with this mix, so the quantitatively trained reviewers thought that the 'feature checking' stuff added nothing, while the editor, not a sociolinguist, thought that our stats weren't up to scratch. If we had stripped out the formal stuff, and sent it to a socio journal, it would have been fine, I think, or it we had taken the phenomenon outside of its social context and looked at it purely formally, it would have been quite publishable, at least a one of three case studies, in a more general journal. Hey ho.

  4. In the functional programming community, there's a tendency to publish things online on ones webpage long before publishing it in print. It lets people give lots of feedback. There are also a number of purely digital non-peer-reviewed "journals" which act as little more than a periodic, type-set, bundled distribution method.

    My general view is that we'd do well to move beyond journals, probably towards something like lingbuzz and arxiv, but supplemented with some capacity for feedback. We wouldn't want to eliminate the typesetting benefits of full-fledged documents by moving to blogs, at least not until the internet has better typesetting features. But we do probably want to make distribution as easy as possible, as make interaction between reader and writer more immediate. We also probably want to retain some amount of peer-review, since it acts as a barrier between the cranks and the serious scientists, but in an internet setting, this can probably be done by letting the participating academic parts of the community as a whole review the work.

  5. @David, is the mixed socio-/formal paper on lingbuzz?

    I think a large part of professional training consists of learning to ignore the enormous variety of attractive ways to approach problems that don't actually work in favor of the few that have produced some kind of result in the past, so anything interdisciplinary will normally get nixed by all the relevant filters at once.