|Faces Studied as Parts are Processed as Wholes|
All faces are made up of the same features (eyes, nose, mouth) arranged in the same general configuration, yet, we can learn new faces everyday. This ability is thought to depend on a “holistic” processing strategy. That is, we process faces based on the whole, not in terms of the individual parts. One demonstration of holistic processing is an inability to consider only part of a face while ignoring another part – the face is processed as a whole, even when we try to attend to a single part. A common explanation for such holistic processing effects is that we encode faces to fit a “face template”. In other words, parts of faces are not represented, only the whole. A recent study by Jennifer Richler and Isabel Gauthier at Vanderbilt University (part of the NSF-funded Temporal Dynamics Learning Center) in collaboration with James Tanaka at the University of Victoria reports a surprising finding: Even faces studied in disjoint, non-face-like configurations (see Figure) still led to holistic effects. This suggests that, in contrast to a common assumption, holistic processing does not require that face parts are encoded together. In the real world, when we see a change in one face half, it is usually because the identity of the face has changed, so the other face half should change as well. Therefore holistic processing might arise because face parts that are encoded separately are not treated independently when judgments are made about them.
Richler, J.J., Tanaka, J.W., Brown, D.D., & Gauthier, I. (in press). Why does selective attention to parts fail in face processing? Journal of Experimental Psychology: Learning, Memory and Cognition.