Abstract
We quickly and automatically categorize faces, for example, as Asian or Caucasian, or as male or female. At the same time we are also good at recognizing individual faces. One issue is whether face categorization and recognition are parallel or serial processes. Here we investigated whether the presence of other faces of the same category interacts with the recognition of individual faces. More specifically: Do participants encode a more robust face representation when there is more than one face of the same race as the encoded face? To investigate this question we used sets of six faces which had different numbers of Caucasian and Asian faces (1-5, 3-3, and 5-1 ). On each trial, Caucasian participants performed an ancillary task to insure that they looked at all faces during an encoding stage. Subsequently, their recognition performance was tested in a same/different task on a single target face in each set. The results of two experiments showed that Caucasian participants were more accurate at recognizing Caucasian targets when five same-race faces were present instead of a single same-race face, while this effect was not evident for Asian targets. Surprisingly, the participants were significantly better at recognizing Asian than Caucasian targets in 3-3 sets. A similar series of experiments using novel objects instead of faces tested possible alternative explanations such as similarity and expertise . Overall, the effect of encoding context suggests that face categorization and recognition processes interact, pointing to a high degree of interdependency of these two processes.