Phylogenetic models for Semitic vocabulary

Kitchen et al. (2009) analyze a data set of lexical trait data for twenty five Semitic languages, including ancient languages Hebrew, Aramaic and Akkadian, modern South Arabian and Arabic languages and fifteen ethiosemitic languages. They estimate a phylogenetic tree for the diversification of lexical traits using tree and trait models and methods set up for genetic sequence data. We reanalyze the data in a homplasy-free model for lexical trait data. We use a prior on phylogenies which is non-informative with respect to some of the key scientific hypotheses (concerning topology and root time). Our results are in broad agreement with those of Kitchen et al. (2009), though our 95% HPD for the root of the Semitic tree (the branching of Akkadian) is [4400, 5100]BP and we place Moroccan and Ogaden Arabic in the Modern South Arabian Group.