Doug Markant, Max Planck Institute for Human Development, Germany
Constructing rules through active exploration: Evidence from human category learning
Abstract: It has long been recognized in both psychology and education that people are active learners: they rely on uncertainty-driven action to explore their environment and discover how the world works. Although the cognitive foundations of uncertainty-driven exploration are increasingly understood, it remains largely unknown how such mechanisms support and/or constrain the acquisition of abstract conceptual knowledge. For example, research on human category learning has predominantly focused on passive forms of training in which learner control is minimized. I will present research showing that active exploration allows people to acquire categorical rules more efficiently than in passive conditions, but this benefit depends on the complexity of the to-be-learned rule. I will argue that this pattern arises from a latent process of sequential hypothesis generation that governs how people “construct” rule-based representations and which guides information search. Understanding this underlying learning process is a critical step toward predicting the kinds of concepts people are able to discover through independent exploration.