Office: 3103K Campbell Hall
Winter 2017 Office Hours: Monday, 10am to 12pm
timhunter at ucla dot edu
I am interested in developing new ways of testing grammatical hypotheses by integrating them into "larger-scale" theories of things like language processing or language acquisition. The two papers below provided most of the raw ingredients for my current thinking on this topic. The way I envisage putting the pieces together was presented in my course at ESSLLI in 2015, and I have an in-progress manuscript on this too; email me if you are interested.
My theoretical work has recently veered, somewhat unexpectedly, in the direction of ellipsis (as seems to happen for many people at one point or another). One interesting common thread to the discoveries in both of the papers below is that they involve sluicing-like constructions that are sensitive to quite global properties of the antecedent of ellipsis, not just the immediate local surroundings of the remnant phrase.
My starting point for this line of research was an interest in the argument/adjunct distinction and the extent to which explanations for puzzling syntactic patterns might derive from the ways arguments and adjuncts contribute to neo-Davidsonian logical forms. This was the focus of my dissertation. The system I developed there turns out to make interesting connections with other topics such as remnant movement, redundancies between merge and move, and extraposition.
Along with Paul Pietroski, Jeff Lidz, and Justin Halberda, I'm studying the semantics of quantificational words like 'most'. Based on careful consideration of how linguistic meaning interacts with other cognitive systems (e.g. psychology of number, visual perception), results of our studies indicate that linguistic meaning is not entirely divorced from verification procedures in the way that is implicitly assumed by much standard practice in semantics.
I've also done some experiments, with Jeff Lidz and others, looking at four- and five-year-olds' learning of determiner meanings. The aim of this work is to see how closely the learner's hypothesis space of available determiner meanings reflects existing typological generalisations, such as the restriction to conservative determiner meanings.