Automated phylogeny of Palaung dialects

Abstract

Improved methods in automatic cognate detection have recently been used by historical linguists to help determine the subgrouping of a clade of languages or dialects, capitalizing on the efficiency of computers when handling substantial amounts of data. In this paper, 16 Palaung dialects are examined using various methods of automatic cognate detection. Partial and whole cognate detection are used together with Lexstat and Sound Class Alignment to generate a phylogenetic tree of these dialects. The results of these methods are compared to the results using cognate detection decided by human experts (Deepadung et al., 2015). These results are substantially similar, suggesting that automatic cognate and phylogeny detection using algorithms is a viable complement to historical linguistic research. Accompanying this paper is a tutorial for the automated cognate detection and phylogeny procedure that was used. By following the steps, users can create results based on segmenting the morphemes of Palaung dialects differently.