Department or Program


Primary Wellesley Thesis Advisor

Ellen Hildreth

Additional Advisor(s)

Kami Koldewyn

Additional Advisor

Bevil Conway


It has been proposed that the processing of faces is a highly specialized function that is separate from the processing of other types of objects. Evidence for this appears with inversion, where it has been observed that face recognition is more affected than the recognition of other types of stimuli (Diamond and Carey, 1986). From these observations it can be inferred that holistic processing of faces makes face analysis unique. This thesis focuses on the question of whether configural information encoded in a face can be used quickly to determine the animacy of a face, or if a more detailed analysis of face features is necessary. By animacy we mean whether the image portrays a real human face or that of a “human-like” inanimate object such as a doll or a manikin. In the first part of this thesis face morphs between doll and human faces were inverted to see if animacy perception of faces would be affected with less configural information. In the second part of the thesis, the time-course of animacy perception in relation to face recognition is analyzed. The final part of the thesis examines the role of high and low spatial frequency ranges in encoding information used for animacy judgment. The results from the first experiment suggest that holistic processing of the face is necessary for animacy perception. From the temporal experiment it can be deduced that animacy perception occurs more quickly than face recognition, and may require less detailed face analysis. The results from the spatial frequency pilot experiment suggest that removing the high spatial frequencies or low spatial frequencies from the face images has no effect on animacy perception, however, more data is needed in order to form a strong conclusion.