Naive theory and transfer of learning: when less is more and more is less.

In this research, we examine how young children perform induction in the domain of biology. According to the essentialist view, even young children have a priori knowledge that some predictors of biological properties are central, whereas others are peripheral. According to the flexible-learning view, the roles of predictors are not fixed, and young children flexibly learn the predictive value of each predictor. In the experiment reported, 4- to 5-year-olds and 11- to 12-year-old preadolescents were taught the predictive values of features that were (by the essentialist account) either central or peripheral for predicting the biological properties of animals. Retention of the learned predictors was then tested immediately after learning, 1 h afterleaning, and 2.5-3.5 months later. The results indicate that 2.5-3.5 months after learning, young children retained learned predictors well regardless of their putative centrality, whereas preadolescents were likely to retain and recall only central, but not peripheral, predictors. These results do not support the essentialist claim that young children differentiate between central (or essential) and peripheral predictors in the domain of biology and, thus, challenge the naive-theory view of induction in young children.