Computational and cognitive complexity in mathematical problem solving

Colloquium lecture by Markus Pantsar (Helsinki), organized by Grolog

The prevalent paradigm in modelling human cognitive capacities focuses on the computational level of analysis. When it comes to mathematical cognition, this approach draws heavily from computational complexity theory, which is concerned with optimal algorithms for solving problems. For human cognizers, however, many cognitive tools (e.g. constructing diagrams) that are computationally suboptimal can in fact be integral to the problem solving process. In this talk, I propose a contrast between optimal and humanly optimal algorithms, which gives us a better-informed framework in which to study the cognitive complexity of mathematical problem solving.