Current Research
The PRIDE speech corpus
The PRIDE corpus is a collection of scripted and unscripted speech utterances by individuals with a variety of gender identities including agender, non-binary, transgender man, and transgender woman in English and Spanish. The goal of this project is to create a rich speech database from which to draw for a variety of experiments examining both the projection and perception of gender in speech.
Perceptual representation of speaker gender
We are examining speaker gender representation in listeners with varying identities and experiences. The goal is to determine how listeners perceptually organize diverse voices and the acoustic features that underlie this organization.
Segmental versus suprasegmental features in speaker gender perception
The relative perceptual importance of segmental (e.g.,, articulation) and suprasegmental (e.g., intonation) speech features in listeners' perception of a speaker's gender is not well understood. This project aims to clarify how listeners use these different speech features using approaches such as auditory-perceptual ratings and acoustic distance
Gender projection and perception cross-linguistically
Nuances of vocal gender vary by language and culture. This project aims to clarify differences in the production and perception of gender between American English and Mexican Spanish, with the goal to inform development of linguistically and culturally responsive gender affirming speech therapy for Spanish speakers.
Word gender perception
Despite the lack of grammatical gender in English, listeners still demonstrate conceptual gender of words (e.g. flower is likely to be associated with femininity more so than hammer). The perceived vocal gender of the speaker, however, may influence listeners’ conceptual gender of words. The degree of this affect may depend on listeners’ prior experiences and beliefs. This project aims to determine how listeners’ own gender and sexuality modulates the effect of vocal gender on the conceptual gender of words.