| We describe a new framework for understanding the perception of symmetry, repetition, and other forms of regularity. We begin with the claim that symmetry is like analogy: that the processes that govern symmetry detection may be the same as those governing analogy, similarity, and metaphor. Drawing on research in these areas, we show how symmetry detection can be modeled as a structure-mapping process that aligns similar systems of visual and conceptual relations. An implementation of this model, MAGI, is described. This model has significant implications: it provides a more robust method for inexact symmetry detection, a better understanding of how complex structural symmetry is detected, and an explanation for the relationship between perceptual and conceptual symmetry. As part of this third theme, we explore the use of symmetry and repetition in diagrammatic reasoning, showing how diagrams use parallel visual structure to encourage particular comparisons or contrasts. As a prerequisite to our description of our symmetry model, MAGI, we also describe the design and implementation of our model of perceptual structure and spatial reasoning, GeoRep. |
