Reading the Mind in the Face: A Cross-cultural and Developmental Study

Previous work suggests only a small set of facial expressions can be interpreted universally with any reliability. This set is confined to the six basic emotions. On the other hand, "cognitive" mental states (such as distrust, recognize, etc.) are held to be private, unobservable, and therefore not available from facial expression. In Experiment 1, we report a study that challenges the notion of the lack of expressive facial correlates to cognitive mental states and tests whether a larger range of mental states are not only detectable, but universally so. Paintings and drawings of faces by one contemporary British artist (Hockney) and one seventeenth-century Spanish artist (Velazquez) were shown to subjects, who were given a forced choice of two semantic opposites as descriptors and asked to select the word that best described the character's mental state. Subjects from three different cultures (Britain, Spain, and Japan) showed good agreement in their judgement of 8 out of 11 mental states from facial information: recognize, threaten, regret, astonished, worried, distrust, contempt, and revenge. Cultural differences were only found on 3 out of the 11 mental states (scheme, wary, and guilt). The seventeenth-century faces were easily interpreted, though the contemporaryfaces were marginally easier. This suggests that a large range of subtle and complex mental states may be universally read in the face, highlighting the role that perception plays in our theory of mind. In Experiment 2, we tested two possible confounding factors: (a) that subjects might have been using "basic emotion" categories to make their judgements about complex mental states, or (b) that they might simply have been making positive/negative judgements. Subjects (from the UK only) were therefore given the same stimuli as in Experiment 1, but this time their forced choice was between the target word and a distractor that was a term describing a basic emotion or a distractor that was another complex mental state term of the same valence (positive or negative) as the target term. The results remained virtually unchanged relative to Experiment 1, suggesting that these judgements are quite subtle. In Experiment 3, we administered a modified version of this test to British children, aged 8-11, in order to see whether there was any development in the ability to recognize these same mental states. Results showed no developmental change during this period, in that even the 8-year-old British child