Skip to main content

James Kirby

University of Edinburgh

Monday, March 4th at 2:00 p.m., James Kirby will give a colloquium in the UCSD Linguistics Department, in AP&M 4301.

Divergence and stability in phonetic change


The 'phonologization' of intrinsic phonetic variation is often invoked as a mechanism to explain the emergence of novel phonological contrasts. Phoneticians have considered how phonologization might occur at the level of the individual, such as failure to compensate for coarticulation (e.g. Ohala, 1981) as well as constraints on the kinds of variation that are likely to be phonologized (e.g. Blevins, 2004). However, the problem of actuation - why variation should (or shouldn't) become phonologized in a particular language at a particular time - has received rather less attention in the phonetic tradition. Meanwhile, researchers from a socio-historical background (e.g. Weinrich et al., 1968) have carefully considered many aspects of the constraints and actuation problems, but have primarily focused on the acquisition and transmission of discrete representations (grammars), rather than continuous phonetic features.

In this talk, I will attempt to combine the insights of both perspectives, in the interest of providing an 'end-to-end' account of phonologization. First, I will present an empirical study of an ongoing sound change in two dialects of Khmer (Cambodian), suggesting how the differential treatment of a common phonetic precursor might explain their divergence. Next, I will present a statistical model of how phonetic change could take place at the level of the individual speaker, and assess the notion that divergence may be predicted by success or failure in compensation for coarticulation. Finally, I turn to consider how assumptions about population structure and learning bias impact the plausibility of such a learning strategy at the population level.