Research

Why do empirically attested human languages tend to use certain kinds of structures more than other ones?

Since at least the work of Zipf almost a century ago, researchers in linguistics, psychology, and cognitive science have explored the idea that the structures in human languages and their attested typology reflect pressure for successful communication. More recent work in cognitive science has examined the hypothesis that linguistic structures and their typology are shaped by pressure for learnability or cultural transmission and how this plays out in populations over large time scales.

My work seeks

to clarify what is predicted about the structures of natural languages by the hypotheses that they are shaped by pressures for communication and/or learning.
to evaluate evidence for or against either hypothesis against large datasets of richly annotated linguistic structures.
to identify the implications of these results for linguistic theory.

To accomplish these goals, I draw on the detailed descriptions and formal analyses of structure and typology provided by linguistics, models of inference in individuals and cultural evolution in populations from computational cognitive science, and tools for analyzing models of probabilistic reasoning and communication from information theory. The primary model system I have examined to date is the lexicon, its phonology, and its morphology.