Relations between intensionality, theory of mind and complex syntax in autism spectrum conditions and typical development

False belief (FB) reasoning emerges around 4-5 years of age in typical development (TD), but has an extended developmental course in children with autism spectrum conditions (ASC). A body of work suggests that this capacity is related to the comprehension of embedded clauses, as in Children think/say that fairies exist. The development of other metarepresentational capacities, such as thinking about the same objects under different descriptions ('intensionality'), has been much less studied, particularly in ASC, and its links both to embedding and FB remain to be determined. Here we aimed to study belief-reasoning, intensionality and embedded clause comprehension in conjunction. An additional aim was to compare performance across different types of embedded clauses, specifically relative clauses and three types of complement clauses (embedded under 'says that', 'sees that' and 'seems that', respectively), and to test their predictive relations to metarepresentational abilities. Twenty-five children and early adolescents with ASC (mean age 9;4) as well as a group of twenty-five children with TD, matched on both chronological and verbal mental age (VMA, mean 9;1), were recruited. A maximally analogous picturematching design was used across all tasks. Results of a series of logistic mixed effect models revealed greater vulnerability in ASC on both metarepresentational tasks (ToM and intensionality) indiscriminately, and within the domain of embedding, on 'seems that'. Scores on embedding predicted performance on metarepresentation, as did those on Seems that. A pervasive main effect of VMA was seen in all models. These results show that metarepresentational impairments in ASC extend to intensionality and relate to language development, where syntactic construction types referencing appearances prove to be particularly challenging in ASC and TD alike.