Comments

Monday, July 8, 2013

Two Cheat Sheets on Computation for Minimalist Syntacticians

NOTE: Apparently the first posting of this had a broken link making the second paper unavailable. I've fixed these so now both PDFs are downloadable. Sorry.


One of the pleasures of being at the LSA summer institute is the chance to interact with people that you've always wanted to know better. I had a couple of personal targets this summer, one of which being Rick Lewis, a computational psychologist here at U Mich.  I've met Rick once before when he came to UMD to give a talk. However, I really got to know about him through his work. He has become a minor celebrity among the UMD processing crowd (I am an informal syntax theory consultant to this group) and several theses have deployed the processing model that he developed (see here for papers. The 2005 with Vasisth is a minor classic where I come from). The proposal provides an account of the observation going back to Miller and Chomsky regarding the processing difficulty of self-embedded sentences (e.g. That that that Bill kissed is surprising is silly is evident). The same model appears to explain some fascinating data Ted Gibson discovered as well. At any rate, the students at UMD have had a field day exploring, criticizing and extending this proposal and I have had a lot of fun listening to their efforts.

This, however, does not exhaust Rick's interests. He is also one of that rare breed, a psychologist cum computer scientist that thinks that Chomsky should be taken seriously. Moreover, Rick has become interested in Minimalism and the idea that one might try to consider grammatical theories from the view point of computational efficiency. Indeed, he has written a couple of papers on the topic that I asked him if I could post. One is a very useful glossary of terms (here). The other (here) is a more extended disquisition on how one may go about thinking of optimal computation in a minimalist setting. I especially like the distinction between 'simplicity' and 'efficiency' and ways they may be tied together (if at all).  Questions of computational efficiency are both important and, sadly, obscure. These papers may help you get some bearings on these issues. They helped me.

2 comments:

  1. Hi Norbert. I'm afraid the links in the second paragraph don't work. They are working relative to the URL of this blog,not relative to Rick's homepage. (First paragraph link is OK, though.)

    Paragraph 2, link 1: http://www-personal.umich.edu/~rickl/pubs/lewis-2010-optimality-notes.pdf

    Paragraph 2, link 2: http://www-personal.umich.edu/~rickl/pubs/lewis-2010-formalizing-mt.pdf

    ReplyDelete
  2. I am always enormously skeptical of claims to computational efficiency that come without any well defined computational system behind them.

    ReplyDelete