“Usage-based linguistic models and understanding human behavior”

The past three decades of research in phonetics and psycholinguistics have led to great advances in our understanding of language, representation, and the relationship between language and other cognitive domains. While debates certainly still exist, we can take as established that how often and in what context different speech patterns occur influence both memory and processing. The question now is what we do with this rich foundation.

In this talk, I present a few, short examples of how usage-based approaches to phonetics and psycholinguistics help us understand social biases and human behavior. I provide some evidence showing that phonetically-cued talker information (e.g., emotion, gender) directly activates lexical items, providing us with some insights into the timing and availability of this
information. The purpose of this first part is to illuminate the complexity of experiencing linguistic events from the perspective of a listener.

For the remainder of the talk, I move away from phonetics, taking the basic insights from the studies initially presented (e.g., that we are pattern recognizers) to question assumptions about language use and experience and ask how our understanding of language use, semantic associations and culture can inform society at large. Specifically, I spend the last large chunk of
this talk investigating how we can understand the refugee experience through the lens of spoken language comprehension.

Meghan Sumner is an associate professor of Linguistics at Stanford University.