Aini Li successfully defends her dissertation!

On April 12 Aini Li successfully defended her PhD dissertation entitled "Inferring dynamics of sociolinguistic variation in speech perception”. She was supervised by Meredith Tamminga, who was joined in her dissertation committee by David Embick and Laurel MacKenzie (NYU).



This dissertation examines whether and how  psycholinguistic priming, and social knowledge are integrated in the identification of sociolinguistic variants. Using the English variable (ING), the alternation between -ing and -in' (e.g. thinking vs. thinkin') as a testing ground, this dissertation probes whether and how individuals utilize constraints of different types when they perceive variation in real time. I combine existing experimental paradigms in a novel way to probe how listeners make inferences about the identity of sociolinguistic variants under circumstances of uncertainty through six perception experiments. Listeners hear synthesized stimuli in which there is ambiguity between two sociolinguistic variants, -ing and -in', and are placed in situations that require them to resolve this ambiguity through categorization.

In Chapter 3, I demonstrate the effectiveness of the methods I will use to introduce uncertainty at the word level and the sentence level. I show in Chapter4 that phonological variant identification in perception is subject to psycholinguistic priming. All else being equal, hearing a clear -ing makes listeners more likely to choose -ing again, given an ambiguous target for categorization. This phonological variant priming effect, however, decays rapidly over time, after only one monosyllabic word, suggesting that phonological variant priming is activation-based. In Chapter 5, I further investigate whether phonological variant priming is sensitive to social expectations. Results show that psycholinguistic priming and talker accent both come into play when listeners categorize ambiguous variants, and crucially, they interact by way of prime variant relative frequency, suggesting that social and linguistic unexpectedness jointly modulate priming. In Chapter 6, I establish that listeners possess and make use of dynamic social factors such as stylistic covariation during phonological variant identification. Additionally, target whole-word frequency can be revealing of the perceptual consequences of different types of s-conditioning. Finally, Chapter 7 discusses implications of these empirical results in the context of the typology of conditioning on variation in individuals.

Overall, this dissertation establishes  that psycholinguistic, social and linguistic factors all play a role when listeners perceive variation. However, different factors and processes are not integrated in the same way, suggesting that individuals have sophisticated knowledge of how variation is conditioned but how this knowledge is used is context-dependent. By combining the framework of variationist sociolinguistics with the methods and theories of psycholinguistics, the results of my dissertation shed light on how sociolinguistic variation is processed in real-time language use. This ultimately has the potential to develop a better understanding of the structure and systematicity of language at the community level.