Reconstructing phylogenies in historical linguistics
The Oxford English Dictionary defines a phylogeny as "the pattern of historical relationships between species or other groups resulting from divergence during evolution." Large-scale computational phylogenetic reconstruction methods, typically using DNA sequence data and many of them likelihood-based, have been wildly successful in biology. Historical linguists have attempted to develop analogous methods to reconstruct the historical relationships between present-day and extinct languages within language families such as Indo-European. The data in this latter case are very different to DNA sequences and there are significant issues around what, if any, stochastic models are appropriate and useful for describing such data. I will survey some of the work in this area, particularly research that I have done over the last 20 years with two computer scientists, Tandy Warnow and Luay Nakhleh, and a historical linguist, Don Ringe, among others.