“ Theory of Mind ” as a Mechanism of Selective Attention

A key component of human intelligence is our ability to think about each other’s mental states. This ability provides an interesting challenge for cognitive neuroscience attempts to understand the nature of abstract concepts and how the brain acquires them. Research over the past 15 years has shown that very young children and children of extremely limited intellectual ability can acquire mental state concepts with ease. Children with Kanner’s syndrome have severe difficulty using these concepts, despite relatively great experience and ability. These discoveries have led to the development of the first information processing models of belief-desire reasoning. The term “theory of mind” was coined by David Premack (Premack and Woodruff, 1978) to refer to our ability to explain, predict, and interpret behavior in terms of mental states, like wanting, believing, and pretending. Because the behavior of complex organisms is a result of their cognitive properties— their perceptions, goals, internal information structures, and so on—it may have been adaptive for our species to develop some sensitivity to these properties. The capacity to attend to mental state properties is probably based on a specialized representational system and is evident even in young children. The term “theory of mind” is potentially misleading. It might suggest that the child really has a theory or that the child has a theory of mind as such. Although there are some writers who hold such views (Perner, 1991; Gopnik and Meltzoff, 1997; Gopnik and Wellman, 1995), I assume simply that the child is endowed with a representational system that captures cognitive properties underlying behavior. To better see what is meant by “theory of mind” ability, consider the following scenario (figure 85.1). Sally has a marble that she places in a basket and covers, and then departs. While she is gone, Ann removes the marble from the basket and places it in the box. A child to whom this scenario is presented then is asked to predict where Sally will look for her marble when she returns. To correctly predict Sally’s behavior, it is necessary to take into account both Sally’s desire for the marble and Sally’s belief concerning the location of the marble. In this scenario, Sally’s belief is ALAN M. LESLIE Department of Psychology and Center for Cognitive Science, Rutgers University, Piscataway, NJ. rendered false by Ann’s tampering. Therefore, to succeed on this task, the child must attribute to Sally a belief that, from the attributer’s point of view, is false. There have been two major discoveries concerning the false-belief problem in figure 85.1. First, Wimmer and Perner (1983), using a somewhat more complex version of the task, found that the majority of 6-year-olds already could pass it, whereas Baron-Cohen and associates (1985), using the version depicted, found that the majority of 4-year-olds could succeed. Subsequently, a large number of studies have confirmed this finding: Whether predicting behavior or reporting where Sally thinks the object is, normally developing children typically solve the problem shortly after the fourth birthday. The second major finding is that autistic children typically fail to solve this task despite mental ages (MAs) well in excess of 4 years, whereas other disabled children—for example, those with Down syndrome—can succeed (Baron-Cohen, Leslie, and Frith, 1985). These two findings raise the following deeply challenging problem for the theorist of cognitive development. How is the young brain able to attend to mental states when mental states cannot be FIGURE 85.1 Illustration of the standard “Sally and Ann” false-belief task given to children to test their ability to attribute beliefs to other people and to calculate the contents of those beliefs correctly. (After Baron-Cohen, Leslie, and Frith, 1985.) 85