ABSTRACT
This chapter aims to apply a 'face space' model of the representation of faces to an analysis of the effect of race on face recognition. It reviews a number of experiments which have been guided by the theoretical framework. The chapter argues that the ability to recognize faces requires a process of perceptual learning during which an observer learns which aspects of faces provide optimal information to distinguish individual faces within a population. It also argues that a multidimensional 'face space' model of face processing can provide a theoretical framework for understanding both the own-race bias and the contact hypothesis. The main assumption of the Multidimensional Space framework is that a location in the space provides metaphor for the representation of a face in memory. The dimensions of the space represent the physiognomic features that are used to encode and remember faces.
