Mary Dalrymple

LFG is a not-exactly-Chomskyan formulation of linguistic theory. Part of its machinery is similar to the formalism of attribute grammars used in computer science, although I doubt that the linguists are aware of this. It always seemed to me that relations between people in computational linguistics and people in theoretical linguistics were quite asymmetric: The former were always willing to apply ideas of the latter, but the latter stayed oblivious to efforts of the former.