Response to critiques

Abstract

In my target paper, I attempted to build on the work of a number of excellent investigators, several of whom were invited to critique my paper. It is a privilege and an honour to respond to my critics, from whom I have learned so much. In their thoughtful commentary, Slater and Quinn cite a body of work that collectively suggests that newborns can discriminate among different faces, and show preferences for the mother’s face and for attractive faces. In all three sets of findings, the authors initially argue for a rapid learning mechanism. Although I do not disagree with their conclusion, I am troubled by the observations they cite. For example, to discriminate between a mother and a stranger, or to prefer an attractive versus an unattractive face must require access to subtle differences between/among faces; in addition, it is likely that such information is coded in higher spatial frequencies. Given the remarkably limited visual ability of the newborn infant (including the very impoverished contrast sensitivity function; see Dannemiller, 2001, for review), how are to we account for these findings? One possibility is that infants are basing their discrimination/ preferences on different cues than the adult, and/or employing different neural and perceptual mechanisms. If true, it may well be that the phenomena observed in newborns is qualitatively different than what is observed in the older infant or adult. Furthermore, we know that, even in the adult, there are individual differences in facial preferences. For example, Perrett et al. have demonstrated that the phase of a woman’s menstrual cycle will influence her judgment of male face preferences (Perrett et al., 1998; Penton-Voak et al., 1999). My point is that the mechanisms that guide newborn facial preferences may be fundamentally different than what guides such preferences in the older infant, no less the adult. Therefore, it may be misleading to suggest that newborn preferences for certain faces is a result of a face ‘module’ per se. Of course, a relatively simple way to evaluate whether faces represent a special class of objects is to see if they show comparable preferences for some non-face objects over others. If they do, would we conclude that there is an object module? Most likely, not. Rather, we would attempt to account for what distinguishes (perhaps psychophysically) one object from another. Nonetheless, it would be a slippery slope indeed if we began to add modules to the brain every time one object was found to be preferred over another. Would we, as Charlie Gross once reflected, have a module for recognizing our grandmother’s face? Slater and Quinn go on to conclude that infants ‘enter the world with a detailed representation of the human face’. Thus, despite their previous arguments about learning, here they appear to argue for a strong nativist position.