Shravan Vasishth: Locality in human sentence comprehension: Empirical and computational investigations

Locality in sentence processing is the assumption that integrating dependents and heads involves a processing cost that is some function of the distance between them. Locality stands at the heart of a major open problem in sentence processing: how are words assembled in real time by the human parsing mechanism to form syntactic parses? Cross-linguistic experimental research on locality has led to fairly diverse theoretical positions, but the final picture is still far from clear. In my talk, I present a series of self-paced reading and eyetracking experimental studies of English, Chinese, German, and Hindi that my collaborators and I have recently conducted, and relate these to the different predictions of three distinct computational models of locality–dependency locality theory (Gibson 2000), surprisal (Hale 2001), and cue-based parsing (Lewis and Vasishth 2005, Vasishth and Lewis 2007).

Toc of the proceedings and download