Preschoolers’ recognition of and inferences from observed imitation

Six experiments explored the apparent discontinuity between infants’ success at inferring affiliation from observed imitation (Powell & Spelke, 2018), and 4-year-olds’ failure on the same inference (Over & Carpenter, 2015). We first tested whether children younger than age 4 (i.e., 3-year-olds) have an explicit, verbalizable, concept of imitation. Three-year-olds observed acts of imitation and could identify the copier (Exp. 1) and who the copier was “copying” (Exp. 2), thus demonstrating a lexicalized concept of imitation. Adults spontaneously inferred that the copier’s affiliation lies with the model over the non-model and that she will continue to copy the same model in the future (Exp. 3), but even five-year-olds failed at these two inferences and preschoolers showed no improvement from ages three to five (Exp. 4). However, preschoolers succeeded at making correct inferences when we primed the concept of imitation either lexically by use of the word “copy” (Exps. 2 and 6), or non-lexically by drawing attention to the fact that the copier had done something “like” one of the models (Exps. 5 and 6). Taken together, the findings provide evidence of continuity between the representations of imitation from infancy through the preschool years to adulthood. In the context of the present experiments, however, young children do not always spontaneously encode the central character's preferential imitation of one model over another.