Emotion Prediction Across Naturalistic Relationship Contexts

Abstract

Emotions change over time, and the temporal dynamics of emotions vary considerably from individual to individual and from group to group. Successful social interactions require people to track this moving target in order to anticipate the thoughts, feelings, and actions of others. In this dissertation, I combine behavioral studies and computational modeling to study how people make emotion predictions that are both accurate and specific to individual partners across three naturalistic relationship contexts. In Chapter 1, I establish that people can indeed make accurate, emotion predictions at a partner-specific level. They do so by relying on specialized knowledge about their partner to account for individual-level emotional idiosyncrasies. Furthermore, I show that how well people make partner-specific emotion predictions is associated with their social success. In Chapter 2, I explore the inferential strategies that support emotion prediction when people encounter new social partners. I show that people can use self-knowledge to effectively approximate other's emotion dynamics. I also show that people are attuned to the utility of their self-knowledge, such that those who are more typical relied on their self-knowledge to a greater extent. In Chapter 3, I investigate how partner-specific emotion prediction can rapidly emerge as new social partners start accumulating specialized knowledge about each other. I demonstrate that people understand the latent structure of how emotion dynamics might vary across individuals in the population. They exploit this structure, and use induction-like mechanisms to quickly individuate their emotion predictions, after learning only a small amount of information about a novel social partner. This work highlights that in order to gain a fuller understanding of the mechanisms underlying predictive social cognition, it is imperative to study it in conjunction with its naturalistic contexts and constraints.