Uriel Cohen Priva



WCCFL was fun (hey, it was in LA)

Some people are doing really interesting stuff, which I’ll probably write more about later. If I were you, I’d check Colin Wilson’s page for updates, he keeps presenting interesting stuff in phonology, using algorithms that I have only heard of in NLP context (actually, Sharon Goldwater often uses these terms to discuss linguistics, so you should go to her web page too).

I really liked the responses my talk got. It seems that people are really excited about using information theoretic approaches in linguistics, and the fact that the explanations I demonstrate reach from the functional to the “arbitrary” appeals to people who would like to know “how language really works” (Bruce Hayes seemed amused when I said this is what I want to know).

Speaking of Bruce Hayes, he suggested that I should not be apologetic in future talks, and make sure people get the fine details of the computational models I use, rather than show them that I’m not opposing standard views. He also thinks I should give more attention to dealing with the what my theory has to say about the acquisition of grammar, and the competence-performance split. I guess I should.