Files
Abstract
Human adults recognize faces extremely well; however, the way in which this expertise develops remains unknown. One developmental model of face recognition is perceptual narrowing, which contends that at birth perception is broadly tuned, but narrows as a function of experience. According to this model, infants transition from being face generalistsbroadly discriminating facial identity for numerous speciesto being face specialists, who are experts in discriminating human faces. To test this model, facial identity discrimination was measured in 4- to 6-month-olds, 9- to 12-month-olds, and adults in two experiments: one using natural faces, and one using systematically varied faces. The first study revealed that 4- to 6-month-olds (n = 26) and adults (n = 27) discriminate sheep (Ovis aries) faces (ps < .05), while 9- to 11-month-olds (n = 26) showed no evidence of discrimination (p > .05). The fact that young infants discriminate sheep faces is consistent with the perceptual narrowing model, which posits that perception is broad early in the first year of life and narrows with age. However, the ability to discriminate some types of animal faces does not disappear, as we found adults could also discriminate sheep faces. In the second study, we examined the properties of faces viewers can use for discrimination by systematically varying faces of humans, capuchin monkeys (Cebus apella), and sheep (Ovis aries) in their outer features (contour, hair), inner features (eyes and mouths), or spacing among inner features (distance between eyes, vertical position of eyes and mouth). Results indicated that 4- to 6-month-olds (n = 61) showed no evidence of discriminating systematically varied faces. In contrast, 9- to 12-month-olds (n = 41) discriminated human and monkey faces that varied in either outer features or inner features (ps < .05). Adults (n = 71) discriminated all three species using inner and outer features, and additionally discriminated human and monkey faces that varied in their spacing (ps < .05). This finding suggests that adults may use configural processing for primate, but not nonprimate faces. Taken together, findings from both experiments differ from predictions derived from a perceptual narrowing model of development. An alternative modellearned attentionis discussed.