The More Similar, the Better? Associations between Latent Semantic Similarity and Emotional Experiences Differ across Conversation Contexts

TL;DR

This study investigates how the relationship between latent semantic similarity and emotional experiences varies across conversation contexts, finding that dissimilar exchanges in pleasant talks enhance positive emotions, challenging prior assumptions.

Contribution

It demonstrates that the association between semantic similarity and emotions depends on conversation type, using NLP tools to test social psychology theories.

Findings

Greater positive emotions occur with more dissimilar exchanges in pleasant conversations.

LSS's emotional impact varies by context, supporting the broaden-and-build theory.

Context influences the semantic-emotional relationship in social interactions.

Abstract

Latent semantic similarity (LSS) is a measure of the similarity of information exchanges in a conversation. Challenging the assumption that higher LSS bears more positive psychological meaning, we propose that this association might depend on the type of conversation people have. On the one hand, the share-mind perspective would predict that higher LSS should be associated with more positive emotional experiences across the board. The broaden-and-build theory, on the other hand, would predict that higher LSS should be inversely associated with more positive emotional experiences specifically in pleasant conversations. Linear mixed modeling based on conversations among 50 long-term married couples supported the latter prediction. That is, partners experienced greater positive emotions when their overall information exchanges were more dissimilar in pleasant (but not conflict) conversations. This work highlights the importance of context in understanding the emotional correlates of LSS and exemplifies how modern natural language processing tools can be used to evaluate competing theory-driven hypotheses in social psychology.