C3: Information and Encoding Density in Comprehension and Production
Project C3 evaluates the assumptions of the UID hypothesis that language use is sensitive to the information density of alternative syntactic encodings, both in comprehension and production. Specifically, the project examines comprehension difficulty of complex, but semantically equivalent, alternative encodings as a function of their syntactic suprisal as well as the surprisal of the individual words within the encoding, as determined by their ordering. A series of production studies examines the extent to which these complex alternative encodings are indeed deployed by speakers in a manner predicted by UID. Finally, the project seeks to extend the information theoretic approach to encoding preferences in situated communication. The project will combine corpus-based analyses with psycholinguistic experimentation.