Proceedings of the Annual Meeting of the Cognitive Science Society

Cognitive scientists have long sought to explain the human capability to solve complex and often ill-defined problems, ranging from puzzle games to the challenges of everyday life. Early work cast problem solving as search: given a fixed representation of states, actions, and transitions, an agent must find a sequence of actions that carries it from an initial state to a goal state, often guided by heuristic search. This framework proved successful in explaining human behavior in lab-controlled puzzle tasks, where experimenters could hand-craft fixed problem representations, allowing models to focus analysis on the downstream search process. Yet, it left a more fundamental question comparatively unexplored: how do humans construct the problem representations that make search possible in the first place?