Abstract
This longitudinal study assessed the role of children’s self-concept in the relation between maternal mental state talk and children’s later social understanding as measured by their mental state vocabulary and performance on social understanding tasks. At all three time points (21, 27 and 34 months), mothers (N = 70) described pictures to children and maternal talk was coded for mental state content. Children’s self-concept was measured using the classic rouge task, use of the term ‘mine’ and use of second-person pronouns. Results indicated that Time 1 mothers’ ‘current’ references to children’s desires at 21 months predicted unique variance in children’s subsequent acquisition of mental state vocabulary at 27 months, but this effect was exclusive to children who did not demonstrate a basic ‘self-concept’ as measured by their mirror self-recognition and use of the term ‘mine’ at the earlier time point. In contrast, maternal talk about her own thoughts and knowledge at 27 months predicted unique variance in children’s subsequent acquisition of mental state vocabulary and performance on objective measures of social understanding at 34 months, but this effect was greater for those children whose self-concept was present but not yet firmly in place as measured by their use of first-person but not second-person pronouns. Working within a social constructivist framework and drawing upon Vygotsky’s principle of a zone of proximal development, I argue that the specific type of maternal mental state talk that is most likely to facilitate advancements in children’s social understanding changes over time and depends on: 1) the child’s current level of social understanding and 2) the child’s understanding of the distinction between self and other.