Accepting and exempting the unexpected: 30-month-olds' generalization of event knowledge

In 2 experiments, we used elicited imitation to test the effects of the amount and type of change to a previously experienced event on 2.5-year-olds' generalization of event knowledge. In Experiment 1, children were shown and then enacted event sequences. At two subsequent visits the props used to enact some of the sequences were completely or partially replaced by functionally equivalent props. Children used the new props to enact the events, thereby demonstrating spontaneous generalization. Nevertheless, as in previous research (Bauer & Fivush, 1992), there were decrements in performance associated with the prop-change manipulation. In Experiment 2, we examined in more detail the determinants of disruption to generalization. Results indicated that an interaction of the location and amount of change within pairs of actions joined by enabling relations (i.e., change to both members of a pair of actions occurring early in the event sequence) negatively affected generalization. Across locations, there was a greater negative effect on generalization when change was made to an antecedent rather than a consequent member of an enabling pair. We suggest that influences on the accommodation of change within events can be understood by considering the organizational role played by the elements subject to change.