Predictive Processing of Disfluent Speech in Bilinguals

Language is one of the most fundamental cognitive processes, yet we know relatively little about how bilinguals learn, understand, and use language. This is a critical knowledge gap, as around 20% of Canadians grow up in bilingual homes. Furthermore, what we do know about bilingual language acquisition has focused on fluent language. Speech disfluencies (e.g., uh and um) can aid language processing in adults, as experienced listeners harness an association between disfluency and hard-to-produce words to make predictions for what a speaker will say next. Here, we investigate how age and exposure to a second language impact disfluent language processing. We emphasize the role of prediction making, hypothesizing that older children with more exposure to a target language will use speech disfluencies to anticipate what a speaker will say next. We also make a methodological contribution by testing and refining methodological tools, focusing on how eye-tracking can be used to measure prediction making. This project unites two bilingual language acquisition labs at Concordia University and the University of Wisconsin. Through this project, the two labs will be able to investigate bilingualism across a diverse set of participants and develop analytical tools that will support the ongoing work of both labs.

Faculty Supervisor:

Krista Byers-Heinlein

Student:

Partner:

University of Wisconsin-Madison

Discipline:

Sociology

Sector:

Education

University:

Concordia University

Program: