Animacy semantic network supports causal inferences about illness.

Inferring the causes of illness is a culturally universal example of causal thinking. We tested the hypothesis that making causal inferences about biological processes (e.g. illness) depends on the animacy semantic network. Participants (n=20) undergoing fMRI read two-sentence vignettes that elicited implicit causal inferences across sentences, either about the emergence of illness or about the mechanical breakdown of inanimate objects, in addition to noncausal control vignettes. All vignettes were about people and were linguistically matched. The same participants performed localizer tasks: language, logical reasoning, and mentalizing. Inferring illness causes, relative to all control conditions, selectively engaged a portion of the precuneus (PC) previously implicated in the semantic representation of animates (e.g. people, animals). Neural responses to causal inferences about illness were adjacent to but distinct from responses to mental state inferences, suggesting a neural mind/body distinction. We failed to find evidence for domain-general responses to causal inference. Causal inference is supported by content-specific semantic networks that encode causal knowledge.