Creating a Comparative Dictionary of Totonac-Tepehua

Abstract

We apply algorithms for the identification of cognates and recurrent sound correspondences proposed by Kondrak (2002) to the Totonac-Tepehua family of indigenous languages in Mexico. We show that by combining expert linguistic knowledge with computational analysis, it is possible to quickly identify a large number of cognate sets within the family. Our objective is to provide tools for rapid construction of comparative dictionaries for relatively unfamiliar language families.