Abstract

The term configural information refers to the interrelationship between the different facial features: the shape and position of each facial feature (e.g. mouth, eyes etc.), in relation to the shape and position of the others. In this study, by using the "composite paradigm" (Young, Hellawell e Hay, 1987), it was investigated the effect of the availability of configural information on the identification of familiar persons appearing in composite and non composite faces at different orientations (0º, 45º, 90º, 135º and 180º). Composite faces were formed by aligning the top part of a familiar face with the bottom part of an unfamiliar face. Under this condition it was expected that the top and bottom parts of the face would have generated a perceptually novel configuration, which would have interfered with the identification of the familiar person. Non composite faces were formed by using the same top and bottom parts, which however were joined in a misaligned fashion. Under this condition it was expected that no novel configuration would have emerged, and the identification of the familiar person would have been easier. As expected, at the orientation of 0º RTs were slower for the identification conducted in composite, rather than non composite, stimuli. The same finding emerged at 45º and 90º of orientation, while at 135º no difference was observed between the two kinds of stimuli. Unexpectedly, at 180º RTs for composite faces were still slower than those for non composite faces, indicating that it was still possible to encode configural information. RTs decreased almost linearly with the increasing of the orientation angle for composite faces, but not for non composite faces, suggesting that the inversion effect is due to a quantitative difference in the ability to encode configural information.