Monday, March 25, 2013

Generative Grammar's Bad Press

David Pesetsky gave a talk at the most recent LSA in Boston that addressed the lack of sympathy for Generative Grammar in the popular press.  Indeed, it extends beyond that. Neither Science nor Nature are well know outlets of generative research.  At the end of the talk, I put in my two cents. I suggested that a large obstacle to generous treatment was the fact that the most influential and prolific living linguist (Guess who? A hint: initials are NC) was also one of the leading political dissidents in the West. Much as our media love to laud dissidents over there they have little taste for those that highlight the shortcomings of our own cultural, political and financial elite.  At any rate, suffice it to say that Chomsky's political writings have irritated a lot of people.  These people thus have an interest in poisoning this well and one strategy is to show that even in his area of professional competence his insights have proven false and his research overthrown.  This is an especially urgent task, it seems, for the liberal media like the New Yorker (see here). Glenn Greenwald has an excellent discussion of how character assassination is intended to work.  Sadly, similar BS is not absent even within linguistics.  I need not name names: we all know who the offenders are. So why is generative linguistics a hard sell? Well, part of the answer has nothing to do with the results of the field. What to do about it? In our small part of the world, we can refrain from character assassination and stick to debating the issues.  Ideas are fair game, personalities are not.


  1. Much of the liberal left finds any sort of innateness a non-starter, for reasons that Steven Pinker has discussed quite a bit.

  2. Some possible reasons include

    1) not made here syndrome - the generative approach owes a lot more to Ancient India than to Ancient Greece.

    2) a fair number of exaggerated initial claims (I think we're much better than most biomedical research, but our mistakes are easier to discover, & we can't promise longer lives and freedom from disease, so need to be far better in order not to be perceived as worse). Binarity of merge would be the latest, according to me.

    3) the apparent fact that a large part of the population appears to either be 'structure blind', or to reject the inference from structure of behavior to structure of its causal mechanism. E.g. the fact that NP structure tends to be repeated all over the place including inside itself at least to recursion depth 1, and sometimes, in writing, up to 3 (Fred Karlsson) does not seem to them to justify any conclusion about what lies behind this kind of behavior. (This problem might have been exacerbated by the fact that in phonology especially, history provides alternative explanations which have not always been considered carefully enough. But what to do about it is unclear.)

    4) Too much reliance on weak poverty of the stimulus arguments, leading to many of the exciting claims of the early 80s disintegrating, such as Kayne's Generalization and Taraldsen's explanation of the that-trace constraint.

    1. To my recollection, none of these were discussed by the press that panned generative grammar. The world would be a far more interesting place were these the relevant parameters of disagreement. They are not. Too bad.

    2. I'm trying to identify possible causes. With referees, for example, my experience is often that their actual complaints are wrong-headed, but there's usually something in the general neighborhood that can be improved.

    3. It may come as a surprise but I agree that many of the linguistic issues that would be interesting for 'the general public' are absent from debates in the popular press. I find this quite puzzling because as linguists you are incredibly fortunate to have a voice as well known as Chomsky. He has more interview requests than he can handle, he can be selective about to whom he wants to talk and he can ensure the important ideas get the best possible exposure - so why is he not raising some of the issues you think should be shaping the public debates? You hardly can blame the 'press' for ignoring them when he is not even touching upon them.

    4. I don't understand why Chomsky does much of what he does, and see no likelihood that he will start doing things differently, e.g. give a better explanation of i-language, or a more careful and nuanced account of the role of statistics, so there's no point in waiting around for him to do it. To him, we probably appear to be rather dense, and perhaps he just can't figure out how to say things that would be more meaningful to (many of) us.

  3. I am entirely ignorant about the situation in the circles around Science and Nature. Yet I’ll dare to express what possible reasons have come up to my ignorant mind: (1) Norbert’s point (see below a piece of evidence of its universality), (2) in the context of linguistics, the field is on one hand (one of) the most sophisticated and on the other it produces very little to none beautiful colored diagrams such as those of neural networks, brain images and the like, (3) there is little potential to generate useful apps, and (4) perhaps the 2002 somewhat carelessly written HCF Science paper and the turmoil of (des)interpretations that followed it.

    Ad (1) The Czech and Slovak most prestigious (if not the only one) popular science magazine Vesmir recently published a series of articles on linguistics by Ludek Hrebicek, a member of the then solitary group around Germany’s Gabriel Altmann who call themselves quantitative linguists or quantiativists. Hrebicek, referring to the second half of the 20th century, wrote:

    “What was (and here and there still is) typical for the pseudolinguistics is the big humbug around the work by Noam Chomsky, a linguistic operator at the Massachusetts Institute of Technology. The famous Institute had made famous a man, who led linguistics, with his pride and self-esteem, to the dead end. Moreover, he, allegedly, managed to ironize the quantiativists.” (Hrebicek, Vesmir 2008, 87, 489-90); my translation)

    This is really a piece of language fitting Greenwald’s collection. Yet, no one of those who read the unbelievable article in due time found it desirable to respond to it. Or perhaps they all believed it.(?)

  4. I think refraining from character assassination within linguistics is an excellent idea, but character assassination is just the grand guignol tip of a more serious iceberg for the field: an increasingly sloppy willingness to blur the boundaries between scientific achievement and personality. Maybe it's a disease of small fields where we know each other too well, or maybe it's not as specific to linguistics as I fear, but the problem is all around us -- to such an extent that we don't even notice the harm that it does. One of the most important lessons we have yet to learn, evidently, is the simplest: that good people can do bad work, bad people can do good work, and the other two combinations of feature values are possible too. (I vividly remember being complimented at an LSA Institute several decades ago with the line "you're surprisingly nice for a GB person". I mean, thank you, but what a strange presupposition!)

    In a similar vein, I've been struck again and again over the years by the casual appropriation of the rhetoric of social liberation movements by academic groupings that feel they are not getting the attention or support that they merit. The general idea is that "we who have worked hard developing linguistic Framework X are an embattled minority, discriminated against for illegitimate reasons by an immoral power structure dominated by School-of-thought Y" -- when what "we" actually are is a miscellaneous collection of researchers who have so far failed to convince some other group of colleagues for rational (though quite possibly wrong) reasons. And before the accusations of Chomskyan hegemonism come rolling in, this is an equal-opportunity concern of mine, not an apologia for one school or another. What school of thought has not played both the role of oppressor and oppressed in some linguist's fantasy life? And if we are honest with ourselves, how many of us have never yielded to the comforting tug of illegitimate personalization and politicization?

    Of course, it is easy to see that this kind of behavior is unconstructive. (Harder to keep it under control, of course.) Personalizing and politicizing our disagreements insulates us from ideas that we should engage with. Necessary discussions do not hapen, and interesting questions remain unanswered. That would be bad enough, but there is worse. The further we go down that road as a a field, the more fodder we supply to those outside linguistics who wish to conclude, as the the Chronicle of Higher Education did last Spring, that "linguistics is populated by a deeply factionalized group of scholars who can't agree on what they're arguing about and who tend to dismiss their opponents as morons or frauds or both". Now in an on-line comment on that remark, I described this as not the field I know -- which is basically true, thank goodness! -- but that doesn't mean I totally fail to see where such an opinion is coming from. Thank you, Norbert, for raising the concern here.

  5. This comment has been removed by a blog administrator.

  6. Outside of the US, Generative Grammar also receives bad press, and the association between it and Noam's political work can't be pointed out as the reason. On the contrary, GG is seen as some form of American scientific colonialism.