Grounding Language about Belief in a Bayesian Theory-of-Mind

TL;DR

This paper proposes a Bayesian theory-of-mind framework to explain how humans interpret and attribute beliefs to others, linking mental states with observable actions through probabilistic reasoning.

Contribution

It introduces a novel Bayesian model that grounds belief semantics in goal and plan inference, improving understanding of human belief attribution compared to previous methods.

Findings

Model better fits human belief attributions than baselines

Highlights importance of instrumental reasoning in belief attribution

Demonstrates graded and compositional nature of human beliefs

Abstract

Despite the fact that beliefs are mental states that cannot be directly observed, humans talk about each others' beliefs on a regular basis, often using rich compositional language to describe what others think and know. What explains this capacity to interpret the hidden epistemic content of other minds? In this paper, we take a step towards an answer by grounding the semantics of belief statements in a Bayesian theory-of-mind: By modeling how humans jointly infer coherent sets of goals, beliefs, and plans that explain an agent's actions, then evaluating statements about the agent's beliefs against these inferences via epistemic logic, our framework provides a conceptual role semantics for belief, explaining the gradedness and compositionality of human belief attributions, as well as their intimate connection with goals and plans. We evaluate this framework by studying how humans attribute goals and beliefs while watching an agent solve a doors-and-keys gridworld puzzle that requires instrumental reasoning about hidden objects. In contrast to pure logical deduction, non-mentalizing baselines, and mentalizing that ignores the role of instrumental plans, our model provides a much better fit to human goal and belief attributions, demonstrating the importance of theory-of-mind for a semantics of belief.