An investigation into the cognitive processes that mediate face perception /

A set of empirical studies is presented that examines the relationship between face perception, the modular hypothesis of cognitive function proposed by Fodor (1983), and attention. In the first study, two different manipulations were used to examine whether faces automatically trigger holistic processing operations as measured by the composite effect. The results support a modular account of face perception.The second study introduces a novel rivalry phenomenon produced by overlapped upright tilted faces. The results indicate that this effect is dependent upon orientation with overlapped inverted faces being perceived as ambiguous in a majority of trials. The third study further examined the factors underlying this rivalry effect. It was found that contrast reversal did not influence the rivalry effect produced by overlapped upright faces and that overlapped houses did not produce rivalry. Results from both studies were taken as evidence that faces are more readily processed as Gestalts compared to other complex objects and therefore engage domain specific operations. The results also suggest that fast operations underlie perception of a face as a Gestalt. Finally, it was suggested that the rivalry effect produced by overlapped faces may illustrate informational encapsulation in face perception.In the fourth study, faces were used to investigate the relationship between attention and modular functions. Three separate experiments showed that faces and houses compete for attention. This finding suggests that the face perception module does not have its own dedicated attentional resources but rather shares a common pool with other visual processes. Results from one experiment also suggested an advantage for faces in the allocation of attention at very short presentation times. This advantage was postulated to arise from two interacting mechanisms that is, faces capture attention over other objects and faces are more automatically encoded than other objects. Together, these studies indicate that a modular conceptualization of face processing is both appropriate and useful. They also demonstrate the utility of faces for investigating cognitive mechanisms that mediate modular functions