BayesVarbrul: a unified multidimensional analysis of language change in a speaker community. Issue 1 (19th April 2022)

• Record Type: Journal Article
• Authors: Hua, Xia
• Is Part Of: Journal of language evolution. Volume 7:Issue 1(2022)
• Page Start: 40
• Page End: 52
• Publication Date: 2022-04-19
• Subjects: gene–environment association -- sociolinguistics -- Gurindji Kriol -- multivariate analyses -- Wright–Fisher model; Linguistic change -- Periodicals; Historical linguistics -- Periodicals; Linguistics -- Periodicals; 417.705
• DOI: 10.1093/jole/lzac004
• Languages: English
• ISSNs: 2058-4571

Abstract

Abstract: Exchange in ideas between language evolution and biological evolution has a long history, due to a shared theoretical foundation between language and biology as two evolving systems. Both systems evolve in terms of the frequency of a variant in a population for each of a large number of variables, that is how often a particular variant of a language variable is used in a speaker community and how many individuals in a biological population carry a particular variant of a gene. The way these frequencies change has been modelled under a similar mathematical framework. Here, I show how we can use concepts from genome wide association studies that identify the source of natural selection and the genes under selection in a biological population to study how social factors affect the usage of language variables in a speaker community or how some social groups use some language variables differently from other groups. Using the Gurindji Kriol language as a case study, I show how this approach unifies existing mathematical and statistical tools in studying language evolution over a large number of speakers and a large number of language variables, which provides a promising link between micro- and macro-evolution in language. The approach is named BayesVarbrul and is ready to apply to datasets other than the Gurindji Kriol dataset, including existing corpus data. The code and the instructions are available at https://github.com/huaxia1985/BayesVarbrul .