Composite faces are not processed holistically: evidence from the Garner and redundant target paradigms.

Composite faces fuse the top half of one face with the bottom half of another. These stimuli inflict a strong illusion of a novel face on their viewers, and are often considered to be processed holistically. The current study challenges this holistic view. Here I present provocative evidence from various classic attention paradigms such as the Garner (1974) and the redundant target (Miller, Cognitive Psychology, 14, 247-279, 1982; Townsend & Nozawa, Journal of Mathematical Psychology, 39, 321-359, 1995) tasks showing that face parts are perceived and processed in an analytic rather than holistic fashion. In Experiment 1, composite faces failed to exhibit Garner interference, indicating perfect selective attention to the constituent parts. In Experiments 2 and 3, composite faces failed to exhibit super-capacity with same-identity composites, demonstrating limited or unlimited capacity. This pattern is consistent with analytic perception. Taken together, the results cast serious doubts on the validity of the holistic processing approach. In many respects, the study proposes disillusionment from the composite face illusion. In addition, the study highlights the importance of converging operations, model testability, and individual differences in the study of faces.