Metarepresentation in autistic language use: A test case for cognitive theories of autism

A recent theory about autism, the Theory of Mind hypothesis, predicts that autistic people will have problems in metarepresenting, i.e. in thinking about their own and other people's thoughts. It is argued that this will show in how they talk about their own and other people's thoughts.;In this study I have tested this prediction by having a closer look at a number of epistemic expressions and descriptions of the process of thinking as used by four high-functioning autistic adults in spontaneous conversations. Contrary to what is suggested in the literature, three of the four autistic adults turn out to be able to use these linguistic expressions appropriately and can be assumed to have a normal understanding of the underlying concepts. One subject uses some but not all expression forms in an adequate way, which is probably due to the fact that she has only partially acquired the ability to metarepresent. Given that this subject happens to be the subject with the lowest level of general intelligence, I have suggested that level of general intelligence is correlated with the ability to develop metarepresentational skills in autism.;What differentiates even my three highest functioning autistic subjects from non-autistic speakers, however, is their extremely frequent use of epistemic modal auxiliaries. I have suggested that this results from their tendency to perseverate. Perseveration is also what characterizes my subjects' conversational abilities in general, and is probably due to an underlying executive function problem.;Current cognitive theories of autism cannot adequately explain the linguistic profiles of my subjects. An alternative (cognitive) hypothesis is therefore suggested.