Babies and Bayes Nets: Mechanisms of Theory-Formation in Young Children

Research suggests that by the age of five, children have extensive causal knowledge, in the form of intuitive theories. The critical question is how young children are able to learn causal structure from evidence. Recently, researchers in computer science and statistics have developed representations (causal Bayes nets) and learning algorithms to infer causal structure from evidence. Here I will present evidence suggesting that children have the prerequisites for making causal inferences consistent with causal Bayes net learning algorithms. Specifically, in a series of studies we demonstrate children's ability to learn from evidence in the form of conditional probabilities, interventions and combinations of the two, in a way that is consistent with the formalism. Children in our studies could infer causal strength, causal direction and more complex causal structure from information about conditional probabilities and interventions, including both the interventions of themselves and others. Moreover, they used this information to override prior assumptions about causal mechanisms, and to design effective novel interventions themselves. Causal Bayes net-like learning algorithms may play an important role in intuitive theory formation and change.