On Emergent Social World Models -- Evidence for Functional Integration of Theory of Mind and Pragmatic Reasoning in Language Models

2026-02-10

AI Summary

This study investigates whether large language models possess a reusable “social world model”—specifically, whether their theory of mind (ToM) and pragmatic reasoning capabilities are underpinned by a unified computational mechanism. Through behavioral evaluation, functional localization, hypothesis-driven statistical testing, and causal mechanistic analysis, the research systematically examines the functional integration hypothesis across seven distinct ToM tasks. Leveraging a newly constructed large-scale Chinese ToM dataset, the findings provide preliminary evidence supporting this hypothesis, suggesting that models’ social cognitive abilities emerge as an integrated, internally connected system rather than isolated modules. This work contributes a novel benchmark dataset, a methodological framework for probing social intelligence, and empirical insights into the computational foundations of social cognition in large language models.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• 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

This paper investigates whether LMs recruit shared computational mechanisms for general Theory of Mind (ToM) and language-specific pragmatic reasoning in order to contribute to the general question of whether LMs may be said to have emergent"social world models", i.e., representations of mental states that are repurposed across tasks (the functional integration hypothesis). Using behavioral evaluations and causal-mechanistic experiments via functional localization methods inspired by cognitive neuroscience, we analyze LMs'performance across seven subcategories of ToM abilities (Beaudoin et al., 2020) on a substantially larger localizer dataset than used in prior like-minded work. Results from stringent hypothesis-driven statistical testing offer suggestive evidence for the functional integration hypothesis, indicating that LMs may develop interconnected"social world models"rather than isolated competencies. This work contributes novel ToM localizer data, methodological refinements to functional localization techniques, and empirical insights into the emergence of social cognition in artificial systems.

Problem

• Theory of Mind
• pragmatic reasoning
• social world models
• functional integration
• language models

Innovation

• Theory of Mind
• Pragmatic Reasoning
• Functional Integration
• Social World Models
• Functional Localization