Successful social interactions rely on flexibly tracking and revising others’ beliefs. These can be revised prospectively, new events leading to new beliefs, or retrospectively, when realizing that an attribution may have been incorrect. However, whether infants are capable of such belief revisions is an open question. We tested whether 18-month-olds can revise an attributed FB into a TB when they learn that a person may have witnessed an event that they initially thought she could not see. Infants first observed Experimenter 1 (E1) hiding two objects into two boxes. Then E1 left the room, and the locations of the objects were swapped. Infants then accompanied Experimenter 2 (E2) to the adjacent room. In the FB-revised-to-TB condition, infants observed E1 peeking into the experimental room through a one-way mirror, whereas in the FB-stays-FB condition, they observed E1 reading a book. After returning to the experimental room E1 requested an object by pointing to one of the boxes. In the FB-stays-FB condition, most infants chose the non-referred box, congruently with the agent’s FB. However, in the FB-revised-to-TB condition, most infants chose the other, referred box. Thus, 18-month-olds revised an already attributed FB after receiving evidence that this attribution might have been wrong.

Pretend play has been extensively studied in developmental science, nevertheless important questions remain about how children engage in and navigate between pretend episodes. In this proposal, we scrutinize childhood pretense from a social cognitive developmental point of view. First, we review previous theories of pretend play structured around important questions that pinpoint some attributes of pretend episodes, such as their transient and socially defined nature. In these sections, evidence is also reviewed about children’s understanding of these attributes. Following this, we describe a novel proposal of pretend play which extends recent accounts of (pretend) play (Wyman & Rakoczy, 2011 ; Chu & Schulz, 2020a ) by exploiting the importance of social interactions in pretense. We contend that engaging in shared pretending can be considered a manifestation of and support for children’s ability to participate in and set up arbitrary contextual boundaries with others. These claims are discussed with regards to how pretend play may figure into social development, its potential implications for intra- as well as intercultural variation, as well as future research.