Abstract

Jean is a computational account of several aspects of cognitive development, specifically, a kind of Piagetian schema-driven learning. Jean’s schemas amalgamate Piagetian sensorimotor schemas with Image Schemas and with Dynamic Maps. These schemas are implemented in our Image Schema Language. Using our Experimental State Splitting algorithm, Jean learns by composing schemas into gists and by differentiating schemas. Jean works in a simple dynamical environment, and we have started to demonstrate transfer of schemas and gists between similar scenarios in the environment, as well as between different environments.

Date

March 16, 2026

Authors

Yu-Han Chang, Clayton T Morrison, Wesley Kerr, Aram Galstyan, Paul R Cohen, Carole Beal, Robert St Amant, Tim Oates

Journal

Proceedings of the Fifth International Conference on Development and Learning (ICDL 2006)