It was long assumed that the capacity to represent false beliefs did not emerge until age 4 as evidenced by children’s performance on elicited-response tasks. However, recent evidence that infants can succeed on non-elicited-response tasks has led many to conclude that this capacity may be present much earlier than previously thought. Some have argued that this conclusion is inconsistent with well-established associations between social factors (e.g. mental-state language) and preschooler’s performance on elicited-response tasks. If infants understand false beliefs, then why would social input predict preschooler’s performance on elicited-response false-belief tasks?

I began to address this question in the current dissertation. I suggest that rather than promoting the emergence of the ability to represent beliefs, social input fosters an individual’s ability to use such representations in appropriate situations. Individuals who frequently hear and use mental-state language may more readily attend to, encode, and retrieve information related to agents’ mental states. Here, I tested three specific predictions that followed from this claim.

Experiment 1 showed that parental mental-state talk, which predicts toddlers’ performance on a verbal spontaneous-response false-belief task, also predicts 2.5-year-olds’ performance on a non-verbal spontaneous-response task. These findings suggest that parental talk does not simply predict children’s ability to follow a verbal story, but rather that parental mental-state language relates to children’s false-belief understanding. Moreover, parental mental-state language did not predict children’s physical-reasoning abilities, suggesting that such talk relates to children’s understanding of mental states rather than children’s general cognitive abilities.

Experiment 2 replicated the finding that parental use of mental-state language predicts children’s spontaneous-response false-belief performance, extended these findings to a new task, and showed that 3-year-olds’ personal use of mental-state language also predicts their false-belief performance.

Building on this finding, Experiment 3 demonstrated that the relationship between personal use of mental-state language and false-belief performance is also present in adulthood: adults who used more mental-state terms when describing individuals were better at quickly and accurately predicting the actions of a mistaken agent.

Together, these findings provide evidence that social factors support individuals’ ability to use their false-belief understanding in a variety of different situations across the lifespan.