Rate variation in language change: Toward distributional phylogenetic modeling

Since the advent of phylogenetic linguistics, researchers have used a large number of phylogenetic comparative methods adapted from computational biology to model and analyze the dynamics of change of a wide range of linguistic features. Models of this sort vary in complexity; the simplest models of change assume homogeneity of transition rates within families, while state-of-the-art models of heterotachy allow transition rates to vary across lineages within a family. In this contribution, I review a range of applications of biological models of rate variation to questions in diachronic linguistics and highlight some models from computational biology that have remained largely overlooked by linguists.Building off of these and other biological models, I sketch out a program for what I term DISTRIBUTIONAL PHYLOGENETIC MODELING, inspired by an analogousrecently proposed family of hierarchical Bayesian models. I report the results of some work in progress carried out within this framework and present a casestudy illustrating the flexibility of the approach.