ABSTRACT
We have seen that caricatures can facilitate the recognition of complex multidimensional stimuli like faces, as well as much simpler stimuli. Indeed the power of exaggerated or supernormal stimuli can be seen in many natural communication systems, as well as in the laboratory. I turn now to the question of how such distorted images can be so effective. The peak shift studies described in Chapter 5 offer an important clue as to how to approach this question. They showed that questions about the power of caricatures are intimately bound up with questions about mental representation. In this chapter I treat the two issues jointly, asking what our mental representations of faces are like and why caricatures are so effective.
