Theory of mind and communication in autism

Recent work investigating the biological and cognitive nature of autism is reviewed. The hypothesis that autistic individuals suffer from a specific impairment in theory of mind, and the relevance of this notion to the diagnosis of Asperger's syndrome, is discussed. Theory of mind explanations of autism must account for the minority of autistic subjects who pass false belief tasks. Two possible accounts are considered; delay versus 'hacking'. A battery of theory of mind tasks was given to able autistic subjects, and to mentally handicapped and young normal controls. The results suggested the existence of subgroups: while some autistic subjects performed inconsistently (perhaps using task-specific strategies), others performed consistently well at first-order theory of mind tasks, and some passed an array of second-order tasks. Relevance theory suggests that understanding mental states is vital in communication. Communication tasks were therefore given, to explore the theory of mind abilities of the autistic subjects. Relevance theory predicted that subjects unable to attribute mental states would show literal understanding, subjects with first-order theory of mind would comprehend metaphor, but second-order theory of mind would be necessary to understand irony. These predictions were tested and confirmed with autistic subjects and normal children. The theory of mind hypothesis cannot explain the persistent real-life handicaps of autistics who understand mental states, nor account for the nonsocial peculiarities seen in autism. A deficit in central coherence is suggested as the source of these features. Results from a preliminary test of this hypothesis suggested that even autistic subjects who develop theory of mind ability are impaired at extracting context-dependent meaning. The implications of these findings for the diagnosis of Asperger's syndrome, the relations between the social and nonsocial impairments, and suggestions for further research are discussed