Investigating 18-month-olds’ association-based inferences in an interactive unexpected-identity paradigm

Abstract

In the helping task a deceptive object is introduced in an agent’s presence with its true identity revealed only later when she is absent. Toddlers’ handing of the appearance instead of reality object to the agent has been taken as evidence for false belief understanding. In this study, we consider the alternative interpretation that toddlers are able to do this because they associate the agent’s presence and absence respectively with the appearance and reality identity of the object. In Experiment 1, we showed that 18-month-olds’ helping behavior was apparently guided by their understanding of what the agent thought the object was. Yet a similar pattern was found in Experiment 2 in which the agent was always present, and the appearance and reality identities of the deceptive object were instead paired with the ons and offs of a light cue involving no mind states. In Experiment 3 the toddlers did not show a preference for passing both the first- and second-revealed identity probe to the agent whom had been exposed to both identities, compared to one identity, of a dual-identities object. After the agent had been exposed to both identities, the toddlers were more likely to help by giving her the second- than first-revealed identity probe. The present results could be explained by a “lean” associationist account in which the agent’s presence and absence are used to form simple associations, not to derive complex epistemic states.