Computational Approaches to the Study of Language Change

The article reviews computational studies of language change. Computer models of change are helpful because of the complexity of the behavior involved: an entire population of complex, interacting agents must be accounted for. Computational studies frequently bring to light hidden implications of theories, which make them relevant to the theoretical development of both acquisition and change. Studies of language change have focused on discovering mathematical properties of dynamical systems, or on simulating populations of speakers that interact with one another and change their internal states as a result. Models of lexical (including phonological) and syntactic change are considered. Computational models of change have proved useful tools for testing theories of language change, and will prove more useful as the field matures to include more systematic studies of the effects of varying model parameters in complex simulations.