Language processing, production, and adaptation as inference based on past input

Since I’ll be visiting Stockholm University this semester, I hope to use this presentation as a way of introduction. I’ll cover a few studies I’ve been involved in, and use them to describe the type of questions and approaches that drive my research. I’ll focus on studies that relate to language learning:

• second language learning (Schepens, van Hout, & Jaeger, 2020-Cognition)
• adaptation to unexpected pronunciations, and generalization thereof across talkers (Xie, Liu, & Jaeger, 2021-JEP:General; Burchill & Jaeger, 2023-manuscript; Tan & Jaeger, 2023-manuscript)
• second language production, and the question whether L2 speakers are more variable in their articulation than L1 speakers (Xie & Jaeger, 2020-JASA).

I’ll try to illustrate why I have found it helpful to think of all of these questions as a form of implicit probabilistic inference about the present input based on the representations learned from previously experienced input. I also illustrate how I use computational models of perception, production, and adaptation to guide my research and experimentation.

Florian Jaeger

About the Higher seminar in Research on Bilingualism