The diversification of the ancient Greek dialects remains one of the most recalcitrant problems in Greek linguistics. Debates persist over a number of fundamental issues, including the methodology of historical inference, the divergence times of the subgroups, and the distribution of the dialects in the second millennium BCE. Over the past twenty years, the advent of Bayesian methods has revolutionized the investigation of such issues. We not only have new ways of approaching old debates, but can also pursue questions that were previously out of reach. In this talk, I present the results of Bayesian phylogenetic analyses carried out on a newly curated dataset. My results bear on the diversification of the ancient Greek dialects as well as the methodology of phylogenetic inference more generally.