Integration of social-indexical cues in linguistic processing
In spoken language, linguistic meaning and information about a talker's social identity are closely interwoven. Listeners may find an interlocutor's speech more or less intelligible depending on their opinions of the speaker, and may hear different phonemes in the same ambiguous signal depending on the speaker's perceived identity. This work connects speech perception with social cognition: how is social-indexical information represented and associated with linguistic variants, and how do listeners use these associations to form expectations and interpret their interlocutors' speech?
The acoustic realization of a phoneme can vary substantially with the linguistic environment, acoustic surroundings, social context, speaking style, speech rate, talker identity, and lexical properties. Listeners can adjust their phoneme category boundaries in response to coarticulatory, phonotactic, lexical, and semantic contexts. They can also learn talker-specific phonetic distributions and generalize what they learn to new talkers. This research investigates the linguistic and social structures that shape perceptual compensation, adaptation, and generalization.
Much research has examined the acoustics of tonal contrasts and the typology of tone systems, but tone variability remains less explored. Tone is primarily realized through fundamental frequency (F₀), which varies with a talker's sex and vocal state. My work investigates how talker sex, vocal effort, and tonal contrasts are jointly encoded in F₀, and how listeners resolve the ambiguity caused by acoustic overlap. I am also interested in how lexical and prosodic structures shape tonal variation, how to model that variation, and how neutral tones change over time across lexical items.
I am interested in the typology, phonologization, acoustics, perception, and modeling of intonation and prosody. One strand of my research examines how to measure prosodic parameters to reveal the underlying structure of prosody while separating it from unrelated sources of F₀ variation. Another uses perception to investigate how prosodic patterns give rise to contrasting interpretations of an utterance.