Abstract

Thirty-years research seemed to reveal that there is a U-shape development in children’s theory-of-mind abilities: infants have the competence to attribute false beliefs properly when measured by looking time and anticipatory eye gaze, while children younger than four systematically fail the standard false belief tasks measuring their voluntary responses. Why is it, and why does the infants’ implicit belief reasoning seem to be free from the inhibition and selection requirements? Are there really two systems, one explicit measured by verbal tasks and one implicit shown in non-verbal context, involved in theory-of-mind reasoning? We investigated preschoolers’ and adults’ belief attribution with the anticipatory looking method, in both a low-demand and high-demand false-belief reasoning tasks. Subjects’ nonverbal performances turned out to be better in the low-demand condition: they showed strong bias to look at the locations congruent with the actor’s false belief. However, neither adults’ nor children’s eye gaze in the high-demand condition showed belief-congruent anticipations. Besides, adults’ predictions were better with verbal responses than with nonverbal anticipatory eye gaze. Together, these results suggested that there are dual systems involved in belief reasoning, and the two systems share similar property in terms of their sensitivity to the task demands.