Shaping Event Backstories to Estimate Potential Emotion Contexts

2025-08-13

AI Summary

Emotion annotation ambiguity frequently arises from insufficient event contextualization. Prior explanations based on annotator characteristics overlook this critical cause. To address it, we propose generating contextualized prehistories to disambiguate emotional interpretations. We introduce the first emotion analysis framework grounded in contextualized narrative generation, leveraging conditional event-chain modeling and short-story generation to automatically synthesize coherent, multi-emotion-oriented prehistories that enrich event descriptions. Experiments demonstrate substantial improvements in inter-annotator agreement (Krippendorff’s α increases by 18.3%) and enhanced interpretability of specific emotions. Our core contributions are threefold: (1) a systematic identification of missing background information as a primary source of emotional ambiguity; (2) empirical validation that contextualized narrative generation effectively mitigates such ambiguity; and (3) demonstration that this approach significantly improves both the reliability and explainability of human annotation in emotion analysis.

Technology Category

• Natural Language Processing: Sentiment Analysis, Stylistic Analysis, and Argument Mining
• Knowledge Representation and Reasoning: Preferences
• Humans and AI: Emotional Intelligence

Application Category

• Web Mining and Content Analysis: Sentiment analysis and opinion mining
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Emotion analysis is an inherently ambiguous task. Previous work studied annotator properties to explain disagreement, but this overlooks the possibility that ambiguity may stem from missing information about the context of events. In this paper, we propose a novel approach that adds reasonable contexts to event descriptions, which may better explain a particular situation. Our goal is to understand whether these enriched contexts enable human annotators to annotate emotions more reliably. We disambiguate a target event description by automatically generating multiple event chains conditioned on differing emotions. By combining techniques from short story generation in various settings, we achieve coherent narratives that result in a specialized dataset for the first comprehensive and systematic examination of contextualized emotion analysis. Through automatic and human evaluation, we find that contextual narratives enhance the interpretation of specific emotions and support annotators in producing more consistent annotations.

Problem

• Addressing ambiguity in emotion analysis by adding context
• Generating event chains to disambiguate emotion interpretations
• Enhancing annotation consistency with contextualized event narratives

Innovation

• Generates multiple emotion-conditioned event chains
• Combines short story generation techniques
• Creates contextual narratives for reliable annotations