ABSTRACT
Earlier studies of repetition priming using faces have been interpreted as indicating that such effects are confined to the processing of known faces. The experiment reported employed eight rather than the more usual two presentation trials and required subjects to make gender decisions to both familiar and unfamiliar faces. Theoretical models of face recognition distinguish between the sets of processes needed to extract basic information common to all faces and the processes required to recognise and retrieve the information about known faces. Ellis et al. argue that the existing face priming data support what they term to be structural theories of face repetition priming, in which response time speed-up on the second viewing of a known face is a reflection of the structural change in the activation unit threshold, which is lowered by the initial presentation of the face.
