Humans possess a remarkable ability to understand what is and is not being said by conservational partners. An important class of models hypothesize that listeners decode the intended meaning of 15 an utterance by assuming speakers speak cooperatively, simulating the speaker's rational choice process and inverting this process for recovering the speaker's most probable meaning. We investigated whether and how rational simulations of speakers are represented in the listener's brain, when subjects participated in a referential communication game inside fMRI. In three experiments, we show that listener's ventromedial prefrontal cortex encodes the probabilistic 20 inference of what a cooperative speaker should say given a communicative goal and context. The listener's striatum responds to the amount of update on the intended meaning, consistent with inverting a simulated mental model. These findings suggest a neural generative mechanism subserved by the frontal-striatal circuits that underlies our ability to understand communicative and, more generally, social actions. 25 3
Main TextA cornerstone of effective communication is our ability to read between the lines, or to recognize the intended meaning of a speaker, even when the meaning is not coded in the utterance directly. 30 The process of disambiguating the intended meaning of a speaker, often known as pragmatic interpretation, has long been hypothesized to rely on cooperation between communicators: A speaker tailors an utterance to help a listener recognize a meaning, and a listener recovers the intended meaning by assuming that the speaker spoke to be understood (1). The hypothesized role of cooperation has inspired a wealth of philosophical inquiries (2-5) and, more recently, empirical 35 (6-8) and computational (9, 10) investigations into human communication, but little is known about the link between the key computational principles and underlying neural mechanisms.Computationally, pragmatic interpretation requires a listener to identify the speaker's underlying intention that motivated the choice of the utterance. According to an important class of models, 40 this can be achieved through an internal generative process similar to how the brain translates sensations into perceptions (10-13). Decades of work suggests that the brain infers sensory causes (e.g., an object) from their bodily effects (e.g., a retinal image) by modeling the sensationgenerating process and then inverting this model to derive the most probable cause of the sensation (11,14,15). In pragmatic interpretation, a similar strategy for a listener entails modeling the 45 speaker's decision-making process, that is, determining how an interacting web of causes-intention, context, and knowledge--give rise to the choice of an utterance. It has been proposed that a listener simulates speaker behavior using a rational, goal-directed choice model (10). Speakers are expected to compare candidate expressions to make a choice for best helping the audience recognize the intended meaning in a given c...