$(RSA)^2$: A Rhetorical-Strategy-Aware Rational Speech Act Framework for Figurative Language Understanding

2025-06-10

AI Summary

Existing Rational Speech Act (RSA) frameworks struggle to model figurative language—such as irony, hyperbole, and meiosis—because they rely on literal semantics and context-specific speaker motivations, limiting generalization to implicit meaning. This work introduces $(RSA)^2$, the first RSA-based probabilistic pragmatic framework that explicitly incorporates rhetorical strategies into its inference architecture, enabling motivation-agnostic figurative intent recognition. Unlike prior approaches, $(RSA)^2$ does not presuppose speaker affect or communicative goals, aligning more closely with human pragmatic intuition. By integrating large language models (LLMs) for both inference and generation, $(RSA)^2$ achieves state-of-the-art performance on the newly curated PragMega+ irony subset. Empirical results demonstrate substantial improvements in human-compatible interpretation of non-literal utterances, advancing the modeling of pragmatic reasoning beyond literal semantics.

Technology Category

• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Knowledge Representation and Reasoning: Common-Sense Reasoning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Figurative language (e.g., irony, hyperbole, understatement) is ubiquitous in human communication, resulting in utterances where the literal and the intended meanings do not match. The Rational Speech Act (RSA) framework, which explicitly models speaker intentions, is the most widespread theory of probabilistic pragmatics, but existing implementations are either unable to account for figurative expressions or require modeling the implicit motivations for using figurative language (e.g., to express joy or annoyance) in a setting-specific way. In this paper, we introduce the Rhetorical-Strategy-Aware RSA $(RSA)^2$ framework which models figurative language use by considering a speaker's employed rhetorical strategy. We show that $(RSA)^2$ enables human-compatible interpretations of non-literal utterances without modeling a speaker's motivations for being non-literal. Combined with LLMs, it achieves state-of-the-art performance on the ironic split of PragMega+, a new irony interpretation dataset introduced in this study.

Problem

• Understanding figurative language in human communication
• Improving Rational Speech Act framework for non-literal meanings
• Modeling rhetorical strategies without speaker motivation analysis

Innovation

• Rhetorical-Strategy-Aware RSA framework
• Models figurative language via rhetorical strategies
• Integrates LLMs for irony interpretation