David Inman | Modeling Obstruent Evolution

We will be meeting for Week 6 on Monday, 3rd November between 6-8pm in Dodd 248 and on Zoom at this link:

https://ucla.zoom.us/my/rehanmuh

This week, we are happy to welcome David Inman, Assistant Professor of Language Science at UC Irvine, who will be giving a talk titled “Modeling Obstruent Evolution“.

David Inman has kindly shared an abstract of his talk:

Many researchers have noticed large-scale typological patterns in the world’s languages, which has led to several proposals of deep-time stability for particular phenomena, such as head- vs dependent-marking or clusivity distinctions. However, until now, statistical research into the stability of non-lexical features has yielded mixed results. The main three studies to date suggest that, when typological features are taken as a whole, rates of evolutionary change vary significantly by family and region, and that structural properties of language typically change more rapidly than the lexicon. While this does not disprove hypotheses about some structural phenomena holding signal beyond the limit of the lexical comparative method, neither do they provide good support for this. To date, this research has been limited to databases designed for typological purposes (such as WALS) and has taken pre-existing typological parameters as their starting point. This research takes a different approach, by creating a new typological parameter for which there is independent reason to believe in its long-term stability: the structure of obstruent phonemes in a language. Based on observations of how consonant inventories change in well-documented language families like Indo-European, I hypothesize that the number of manner and voicing contrasts in a phonological inventory tends to change slowly over time. Using phonemic inventory data from PHOIBLE and an Ornstein-Uhlenbeck process to model evolutionary change, we test this hypothesis across the world’s languages, and find that indeed the number of obstruent series present in a language holds strong phylogenetic signal and changes at least as slowly as the lexicon.