Similarity Dependent Chinese Restaurant Process for Cognate Identification in Multilingual Wordlists

Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018), pages 271–281 Brussels, Belgium, October 31 - November 1, 2018. c⃝2018 Association for Computational Linguistics 271 Similarity dependent Chinese Restaurant Process for Cognate Identification in Multilingual Wordlists Taraka Rama Department of Informatics University of Oslo, Norway tarakark@ifi.uio.no Abstract We present and evaluate two similarity depen- dent Chinese Restaurant Process (sd-CRP) al- gorithms at the task of automated cognate de- tection. The sd-CRP clustering algorithms do not require any predefined threshold for de- tecting cognate sets in a multilingual word list. We evaluate the performance of the algorithms on six language families (more than 750 lan- guages) and find that both the sd-CRP variants performs as well as InfoMap and better than UPGMA at the task of inferring cognate clus- ters. The algorithms presented in this paper are family agnostic and can be applied to any linguistically under-studied language family. 1 Introduction Cognates are related words across languages that have descended from a common ancestral lan- guage. Identification of cognates is an important step in historical linguistics while establishing ge- netic relations between languages that are hypoth- esized to have descended from a single language that existed in the past. For instance, English hound and German Hund “dog” are cognates that go back to the Proto-Germanic stage. Cognate identification requires great amount of scholarly effort and is available for some language families such as Indo-European, Dravidian, Austronesian, and Uralic which have a long tradition of compara- tive linguistic research that involves decades (Dra- vidian family) to centuries (Indo-European fam- ily) of scholarly effort. Automatic detection of cognates with high accuracy is very much desired for reducing the effort required in analyzing under- studied language families of the world. Typically, expert annotated cognate sets are em- ployed to infer phylogenetic trees showing lan- guage relationships that can be used to test hy- potheses about temporal and spatial evolution of language families (Bouckaert et al., 2012; Chang et al., 2015), linguistic reconstruction of ancestral states on a tree (J¨ager and List, 2017), or lexical re- construction (Bouchard-Cˆot´e et al., 2013). Rama et al. (2018) showed that cognates inferred from automated methods of cognate detection can be used to infer high quality phylogenetic trees. The authors noted that there is a need for more research towards developing highly accurate cognate iden- tification methods that can be applied to the data of not so well-studied language families which will be of assistance to historical linguists to automate parts if not the whole of the comparative method. The last decades have seen a large amount of computational effort towards automatizing the process of cognate identification since the work of Covington (1996) and Kondrak (2002). The computational effort involved devising new se- quence alignment algorithms (Kondrak, 2005, 2009), novel sound transition matrices which are linguistically guided (Kondrak, 2001; List, 2012b) or data-driven (J¨ager, 2013; Rama et al., 2013, 2017; List, 2012a), and machine learning approaches (Hauer and Kondrak, 2011; Rama, 2015, 2016; J¨ager et al., …

Content truncated to 500 words