The developmental origins of naïve psychology in infancy.

Research interest in children's understanding of the mind goes back as far as Piaget's claim that children are cognitively egocentric (Flavell, 2000). Many years later, research on the understanding of the mind was revived in a paper that sought evidence for a theory of mind, not for children but for chimpanzees (Premack & Woodruff, 1978). The researchers claimed that chimpanzees' ability to predict what a human actor will do to achieve certain goals implies that the animal attributes mental states to the actor. This seminal paper generated a flurry of studies on theory of mind in nonhuman primates. A review of this research based on several different experimental paradigms concluded that chimpanzees understand others in terms of a perception-goal psychology (i.e., they can perceive what the other's goal is but not understand the mental states associated with the goal), as opposed to a full-fledged, human-like belief-desire psychology (Call & Tomasello, 2008). Around the same time, research on children's understanding of the mind was revived in a landmark paper by Wimmer and Perner (1983) and by other developmentalists (Bretherton, McNew, & Beegly-Smith. 1981). In line with the research on nonhuman primates, part of the progress that has been made in recent years is a recognition that theory of mind knowledge is acquired in an extended series of developmental milestones and that this development is based on a rich set of socio-cognitive abilities that develop in infancy (Wellman, 2002). The evidence outlined in the sections of this chapter suggests that infants possess a nascent understanding of mental states that older children use in explaining and predicting human behavior. Researchers have learned a great deal about the developmental origins of naive psychology in infancy. Nevertheless, the depth of infants' understanding of human behavior is still a controversial issue. For example, a popular paradigm in naive psychology is violation of expectancy. In false-belief tasks, infants look longer at a scene.in which a protagonist searches for an object in a location she does not know than at a scene in which the protagonist searches for an object in a location where she has previously seen the object disappear. The fact that no active behavioral response is required makes many researchers doubt that an infants' looking pattern reflects a deep level of understanding. Looking pattern may simply reflect the infants' detection that something in the scene is novel (e.g., protagonist looks at a location different than the one infants last saw her look at). Indeed this interpretation may account for the conflicting results in recent studies (e.g., Poulin-Dubois et al., 2007; Onishi & Baillargeon, 2005; Surian et al., 2007). Poulin-Dubois et al. (2007) recently reported that the ability to distinguish between knowledge and ignorance (true belief) is absent at 14 months of age and still fragile at 18 months in a violation-of-expectancy task depicting videotaped human actors. In contrast, false-belief attribution to a computer animated caterpillar has been reported in 13-month-old infants (Surian et al., 2007). Given that infants have had more experience with humans looking at objects than with a caterpillar's looking behavior, the current evidence for an implicit understanding of advanced mental states such as false belief should be interpreted with caution. As is the case for nonhuman primate research, infants' mind-reading success might be accounted for by a simple behavior-reading explanation. According to some researchers, primates' (and infants') successful performance in theory of mind tasks can be explained by a sophisticated form of behavior reading. Under this view, infants perform well in such tasks because they are adept at calculating the statistical likelihood that some aspects of people's observable features (e.g., gaze) will be linked to future actions (e.g., search at a location). Distinguishing between a mentalistic and rule-based account is very difficult (Povinelli & Vonk, 2004). One way to address this debate would be to design training studies that provide infants with first-person experience of mental states and to use more active behavioral measures. In terms of training, there is some evidence that infants' performance on goal and visual perception attribution tasks is improved if they received training of relevant skills (e.g., wearing a blindfold, reaching with a "sticky mitten": Meltzoff & Brooks, 2007: Sommerville & Woodward, 2004). Furthermore, longitudinal research using more active measures revealed links between goal detection as measured with the violation of expectancy paradigm at 10 months of age and the ability to infer intended goals in an imitation task at 14 months (Olineck & Poulin-Dubois, 2007b). Developmental changes in the scope of infants' concept of intentional agent also will require more attention from researchers. According to some, infants' attributions of intentional behavior are activated whenever infants recognize an object as a psychological agent, based on an evolutionary designed system which is sensitive to certain cues such as self-propulsion, contingent reactivity or equifinal variation of the action (Baron-Cohen, 1995; Gergely & Csibra, 2003; Johnson, 2000; Leslie, deficient in theory of mind. One may hope that nonverbal theory of mind tasks that reliably predict later theory mind skills will be adapted for use with this population and eventually used for the early detection of autism. In sum, the numerous studies reported here show that by the end of the second year of life, infants have developed ways to predict human actions The review also makes clear that we do not yet fully understand how deep infants' insight into the mind really is. Nonetheless, there appears to be some consensus that infants, like chimpanzees, understand the goals, intentions, perception, and knowledge of others. This provides the foundations for the full-fledged adult-like naive psychology that develops gradually in early childhood.