Item

Abstract

We present an analysis of the patterns of lexical diversity amongst the Aslian languages of Malaysia and Thailand. Burenhult, Kruspe and Becker have compiled comparative wordlists for 22 language dialects and geographically separated varieties, and have independently coded these for cognacy according to explicitly agreed criteria. Dunn has applied bioinformatic methods to this coded database (i) to investigate how 'treelike' the data is, and (ii) to compare the phylogenetic signal in the data with published subgrouping hypotheses. A network analysis has been carried out using the NeighborNet method. This method produces clustering based on surface similarity, and does not presuppose a hierarchical family tree of relationships. Nevertheless, this analysis shows geographically and diachronically plausible groupings, corresponding in some cases to proposed subgroups. Evidence is found for a linguistic divide between hunter-gatherer and agricultural groups which cuts across the subgroup boundaries established by the comparative method. More sophisticated phylogenetic analyses will be shown (generated using a Monte Carlo Markov chain Bayesian tree inference method), which reconstruct a likely family tree and precisely quantify the confidence with which different aspects of the family tree can be reconstructed.