Last month, one of our graduate students presented a paper at the Pacific Asia Conference on Language, Information, and Computation (PACLIC 38) in Tokyo, Japan.
The title of the talk was “Revisiting Leti metathesis: a use case for boolean monadic recursive schemes” and the abstract can be found below.
Abstract
“This paper demonstrates how Boolean monadic recursive schemes (BMRS), a computational method of modeling phonological processes (as proposed in Chandlee & Jardine 2021, characterized in Bhaskar et al. 2020), can model metathesis in Leti, a Timoric language spoken primarily on the island of Leti in the Maluku archipelago. In this language, metathesis—when two segments switch linear position—is morphologically productive and phonologically conditioned. Using data and analyses by Hume (1998) as a starting point, I build on the idea that metathesis is a process that simultaneously deletes a segment and inserts it in a new place, modeling this process using BMRS. In the case of Leti, in certain environments, a wordedge consonant deletes and inserts itself right before its preceding vowel. In contrast with Hume’s optimality and correspondence theory-based analysis, however, BMRS can intuitively account for the environments and opaque phonological interactions driving Leti metathesis without having to appeal to linearity constraints and syllable-level representations, showing that Leti metathesis is a local process that applies to segments.”