The development of problem solving: Integrating perception, cognition, and action in real time.

This paper reviews research on problem solving from infancy through childhood, with particular attention to tool use, means-end behavior, planning, and action selection. Across classic and recent studies, we show that developmental progress is not captured fully by whether children succeed on a task, but also by how they solve it in real time. Findings from behavioral, kinematic, eye-tracking, neural, and computational work suggest that successful problem solving depends on the timing and coordination of multiple processes, including information gathering, motor preparation, and online adjustment. The chapter also examines how exploration, variability, and low-cost error support learning and generalization across contexts. We conclude by arguing for a more integrated developmental science of problem solving that combines structured experiments, naturalistic observation, and embodied computational models to explain how flexible problem-solving skills emerge over development and in everyday settings.