Different Spatial Scales for Different Face Categorisations

When people categorise complex stimuli such as faces, they might flexibly use the perceptual information available from the visual input. Three experiments were run to test this hypothesis with two different categorisations (gender and expression) of identical face stimuli. Stimuli were hybrids (Schyns and Oliva, 1994 Psychological Science5 195 – 200): they combined either a man or a woman with a particular expression at a coarse spatial scale with a face of the opposite gender with a different expression at the fine spatial scale. In experiment 1 we tested whether a gender vs an expression categorisation task tapped preferentially into a different spatial scale of the hybrids. Results showed that expression was biased to the fine scale, but that gender was not biased. In experiment 2 the same task was replicated, following a learning of the identity of the faces. It was then found that gender also became biased to the fine scale. In experiment 3 the expression task was changed to an identification of each expression to establish whether this could revert the scale biases observed in experiments 1 and 2. Results suggest that different categorisations of identical faces use different perceptual cues. This suggests that the nature of a task changes the representation of a stimulus.