Historical linguists have long suggested that language learning plays a role in language change. Developments in the computational modeling of learning of grammars and of “iterated learning” now allow formally explicit proposals about the role of learning in change, and about the impact of learning biases on typology. In this course we will attempt to push this line of research forward by not only studying current proposals in this area, but also by careful examination of the related historical literature, and by doing original modeling work.

Participants are not required to have special background in either computational modeling of learning or diachronic linguistics, but must be motivated to gain expertise in both over the course of the semester. Students will present results of modeling and readings throughout the semester, culminating with a final modeling project.

We will start with a discussion of some foundational material on modeling of change and then move on to work on phonetic and phonological change, morphological change, and finally syntactic and semantic change (an initial collection of readings on modeling change can be found here).