Date of Degree

9-2022

Document Type

Dissertation

Degree Name

Ph.D.

Program

Computer Science

Advisor

Rebecca Levitan

Committee Members

Elena Filatova

Julia Hirschberg

Michael Mandel

Subject Categories

Computational Linguistics | Other Computer Sciences | Psycholinguistics and Neurolinguistics | Social Psychology and Interaction

Keywords

Entrainment, Prosody, Variation, Accommodation, Personality

Abstract

Entrainment refers to the tendency of human speakers to adapt to their interlocutors to become more similar to them. This affects various dimensions and occurs in many contexts, allowing for rich applications in human-computer interaction. However, it is not exhibited by every speaker in every conversation but varies widely across features, speakers, and contexts, hindering broad application. This variation, whose guiding principles are poorly understood even after decades of entrainment research, is the subject of this thesis. We begin with a comprehensive literature review that serves as the foundation of our own work and provides a reference to guide future research. Then we demonstrate the extent of variation in entrainment through analyses of several corpora, including the first broad investigation of entrainment in Hebrew. Our results challenge the assumption, implicitly made by theoretical accounts of entrainment, that it is a single behavior or structured collection of behaviors. They also show that differences cannot consistently be attributed to gender or native language. In a second part, we present a newly designed, implemented, and partially collected corpus for the purpose of studying variation in the entrainment behavior of the same subjects towards different partners in both task-oriented and free conversations. Preliminary analyses indicate limited generalizability of the impact of personality and suggest the intriguing possibility that speaker states might have greater impact on entrainment than speaker traits. Lastly, we present original, neural entrainment measures which aim to address shortcomings of existing approaches and yield promising preliminary results.

Share

COinS