[2E4-OS-1a] OS-1 (1)

[2E4-OS-1a-05] Quantifying cultural evolution of language using large-scale corpora

〇Shimpei Okuda¹, Michio Hosaka², Kazutoshi Sasahara^1,3 (1. Graduate School of Infomatics, Nagoya University, 2. College of Humanities and Sciences, Nihon University, 3. JST PRESTO)

Keywords:large-scale social data analysis, text mining, evolution linguistics

Both directed selection and stochastic drift are the driving forces of biological and cultural evolution, and this is also true for language evolution. The recent argument presented by Newberry et al. (2017) is that drift cannot be rejected and stochasticity has an under-appreciated role in grammatical changes in English. In this paper, we focus on the evolution of the English perfect construction (be/have+PP) and aim to detect signatures of selection and drift working there. We used three English Corpora--EEBO, COHA, and Google Books. From these corpora, we computed the longitudinal frequency changes of be/have+PP forms in 19 target verbs. The results of our analysis show that this auxiliary selection is dependent on the nature and grammatical usage of verbs and suggest that frequency changes from be+PP to have+PP are unlikely due to random drift in these verbs.