Bayesian analysis has come to be widely used in archaeological chronologies and has been a regular feature of recent articles in Antiquity. Its application to linguistic prehistory, however, has proved controversial, in particular on the issue of Indo-European origins. Dating and mapping language distributions back into prehistory has an inevitable fascination, but has remained fraught with difficulty. This review of recent studies highlights the potential of increasingly sophisticated Bayesian phylogenetic models, while also identifying areas of concern, and ways in which the models might be refined to address them. Notwithstanding these remaining limitations, in the Indo-European case the results from Bayesian phylogenetics continue to reinforce the argument for an Anatolian rather than a Steppe origin.