Funded Grants

Learning in Early Childhood: A Computational Cognitive Developmental Approach

Many adults find exposure to information that challenges existing understanding to be aversive, even painful. Children, in contrast, are anything but epistemically indifferent; they are curious, eager, active learners. Indeed, in the course of development, children repeatedly refine their beliefs. Sometimes learning is fast and flexible, other times it is protracted and rigid. Sometimes learning is most efficient during self-guided play, other times direct instruction is more effective. And, sometimes children (and adults) eagerly explore to push the boundaries of their knowledge, while other times they are untroubled by their relative ignorance. Studying learning in early childhood provides a unique opportunity to resolve these dichotomies. To this end, my research has focused on the epistemic stance that young learners take, asking: (1) How evidence interacts with prior beliefs in causal learning; (2) How evidence is interpreted in different contexts and shapes play in these contexts; (3) What past experiences drive particular individuals in certain contexts to want to seek out more evidence and, more generally, to be curious.

The “theory theory” provides an initial framework for addressing children's causal learning. Causal beliefs can be thought of as naïve “theories” that support explanation, prediction, and exploration, and that are revisable in light of new evidence (Carey 1985; Gopnik & Meltzof, 1997; Keil, 1992; Wellman & Gelman, 1992). This framework suggests that children may be particularly well adapted to attend to evidence, but it has failed to explain why learning can be fast and flexible or protracted and rigid. Furthermore, the role of naïve theories in guiding children’s play, and the role of play in shaping theories, has been underspecified. Although researchers have long agreed that children can and do learn through play (e.g., Bruner, Jolly, & Sylva, 1976), little is known about why learning is sometimes more efficient during self-guided play, and at other times during direct instruction. Finally, although theorists have long appreciated the importance of human curiosity in learning (Berlyn, 1960), few studies have attempted to specify how past experiences may lead to individual differences in this drive.

My research program aims to resolve these dichotomies and formally to characterize the interactions between children’s learning, play, and curiosity. Doing so requires precise, explicit theoretical assumptions that produce empirically testable predictions. Thus, in addition to studies of children's learning, I use computational modeling to formalize theoretical intuitions. Understanding the generalities as well as the heterogeneities of learning throughout the lifespan also requires a “domain-diverse” and “age-diverse” program. My research draws on multiple domains of causal learning, investigating children’s acquisition of biological, psychological, and physical knowledge. Discovery of general principles also requires diversity of sampling; I investigate learning across the entire range of development. My primary goal is to understand how learning develops, identifying both general principles of learning as well as specific constraints that may influence individual differences in epistemic stance.