Gricean Norms as a Basis for Effective Collaboration

TL;DR

The paper tackles instruction ambiguity in human-AI collaboration by proposing a normative framework that fuses Gricean maxims with cognitive theories (common ground, relevance theory, theory of mind) for LLM-based agents called Lamoids. It leverages Few-shot Chain-of-Thought prompting to embed pragmatics into action selection within a grid-world task (Doors, Keys, and Gems). Empirical results show that norm-enabled Lamoids achieve higher task accuracy and produce clearer, more relevant responses than norm-free variants, demonstrating the value of pragmatics in cooperative AI. The work advances practical, context-aware human-AI collaboration with potential extensions to spatial grounding, external tools, and broader pragmatic cues across dynamic environments.

Abstract

Effective human-AI collaboration hinges not only on the AI agent's ability to follow explicit instructions but also on its capacity to navigate ambiguity, incompleteness, invalidity, and irrelevance in communication. Gricean conversational and inference norms facilitate collaboration by aligning unclear instructions with cooperative principles. We propose a normative framework that integrates Gricean norms and cognitive frameworks -- common ground, relevance theory, and theory of mind -- into large language model (LLM) based agents. The normative framework adopts the Gricean maxims of quantity, quality, relation, and manner, along with inference, as Gricean norms to interpret unclear instructions, which are: ambiguous, incomplete, invalid, or irrelevant. Within this framework, we introduce Lamoids, GPT-4 powered agents designed to collaborate with humans. To assess the influence of Gricean norms in human-AI collaboration, we evaluate two versions of a Lamoid: one with norms and one without. In our experiments, a Lamoid collaborates with a human to achieve shared goals in a grid world (Doors, Keys, and Gems) by interpreting both clear and unclear natural language instructions. Our results reveal that the Lamoid with Gricean norms achieves higher task accuracy and generates clearer, more accurate, and contextually relevant responses than the Lamoid without norms. This improvement stems from the normative framework, which enhances the agent's pragmatic reasoning, fostering effective human-AI collaboration and enabling context-aware communication in LLM-based agents.

Figures (11)

• Figure 1: Doors, Keys, and Gems grid world Zhi-Xuan+20:rational-bayesian-goal.
• Figure 2: Third component of the prompt: norm-driven vs. non-norm-driven response generation by a Lamoid.
• Figure 3: Fourth component of the prompt: few-shot CoT exemplars with and without norm.
• Figure 4: Responses of Lamoid with norms vs without norms.
• Figure 5: Response for Lamoid with norms vs without norms.