Young children’s default induction strategy
Our study directly pitted category choices against perceptual choices in order to provide a rigorous test of young children’s induction biases. Because our youngest age groups showed a perceptual bias rather than a category bias, our findings make an essentialist bias toward nonobvious category infor- mation in early induction unlikely. This contrasts with Gelman’s (2003) argument that children have an early bias to attend to hidden nonobvious properties and, thus, make generalizations based on kind information over appearance. Similarly, Goswami (1992, 2001) argued that young children’s natural default is for relational interpretations. Once children understand the specific relations relevant to the task, their decision will be based on these. Bulloch and Opfer (2009) also argued that young chil- dren are capable of using relational information when they recognize this as a reliable predictor. In Experiment 1a, the target had a very strong relationship to the category choice. It not only was the same type of bug but also was the very same bug grown up. The connection between the juvenile (cat- egory choice) and adult target was demonstrated prior to every trial when children witnessed the juvenile bug being transformed into the adult target bug. In addition, children’s understanding