Decoding Emotional Experiences in Dyadic Conversations of Married Couples: Leveraging Semantic Similarity through Sentence Embedding

2023-09-22

AI Summary

This study challenges the conventional assumption that higher latent semantic similarity (LSS) between interlocutors correlates with more positive emotional experience, investigating whether dialogue valence—specifically pleasant versus conflictual interactions—moderates the LSS–affect relationship. Method: Leveraging naturalistic dialogue data from 50 married couples, we compute sentence-level cosine similarity between consecutive utterances using GTE-Large embeddings, and employ this metric to predict dyadic affect in real-time interaction. Results: Contrary to expectations, lower LSS significantly predicted higher positive affect in pleasant dialogues (p < 0.01), revealing a counterintuitive inverse relationship between semantic synchrony and emotional responsiveness. The metric robustly captures key interactional dynamics—including coordination patterns and feedback latency. This work establishes the affective valence of LSS as context-dependent, thereby advancing computational social science by introducing a novel, interaction-sensitive paradigm for semantic–affective modeling.

Abstract

Recent advancements in Natural Language Processing (NLP) have highlighted the potential of sentence embeddings in measuring semantic similarity (hereafter similarity). Yet, whether this approach can be used to analyze real-world dyadic interactions and predict people's emotional experiences in response to these interactions remains largely uncharted. To bridge this gap, the present study analyzes verbal conversations of 50 married couples who engage in naturalistic 10-minute conflict and 10-minute positive conversations. Transformer-based model General Text Embeddings-Large is employed to obtain the embeddings of the utterances from each speaker. The overall similarity of the conversations is then quantified by the average cosine similarity between the embeddings of adjacent utterances. Results show that lower similarity is associated with greater positive emotional experiences in the positive (but not conflict) conversation. Follow-up analyses show that (a) findings remain stable when controlling for marital satisfaction and the number of utterance pairs and (b) the similarity measure is valid in capturing critical features of a dyadic conversation. The present study underscores the potency of sentence embeddings in understanding links between interpersonal dynamics and individual emotional experiences, paving the way for innovative applications of NLP tools in affective and relationship science.

Problem

• Examines how latent semantic similarity affects emotional experiences differently across conversation contexts
• Challenges assumption that higher semantic similarity always leads to more positive emotions
• Tests competing theories about emotional impacts of similarity in pleasant vs conflict conversations

Innovation

• Uses latent semantic similarity (LSS) analysis
• Applies linear mixed modeling techniques
• Tests theory-driven hypotheses with NLP tools