An investigation into the semantic-syntactic interface in typically developing children and children with grammatical specific language impairment.

This thesis presents an initial investigation of the semantic-syntactic interface in a sub-group of children with Specific Language Impairment (SLI): the Grammatical (G)-SLI group. Previous research of this group has largely focused on their morpho-syntactic abilities and more recently on the morpho-phonological interface. In contrast, the semantic ability of children with SLI in general is an area that has received little investigation and the question remains as to whether their deficit extends to the semantic-syntactic interface. The thesis focuses on aspect. Understanding of lexical and grammatical aspect is investigated using grammatical judgement and sentence-picture matching tasks. The impact of aspect on passives, a grammatical construction that children with G-SLI have difficulty with is also investigated. The group of children with G-SLI who participated were aged between 11 and 16 and their performance was compared to 3 groups of typically developing children matched on either grammar (LAI, mean age 6) or vocabulary (LA2, LA3 mean ages 7 and 9 respectively) ability. The investigations revealed that children with G-SLI were impaired in their understanding of the past progressive. Their difficulty in reconciling an ongoing event with something that is happening in the past provides evidence of morpho-syntax impacting on semantic interpretation. They also found descriptions of events that derived their telic interpretation compositionally, from the addition of a PP, to be harder to understand those that did not, providing further evidence of syntax impacting on semantics. The results are consistent with van der Lely's CGC hypothesis which emphasises the cumulative and interactive effects that deficits in grammar can have on linguistic constructions. Specifically the thesis shows how syntactic context affects understanding and performance in semantic aspect. It also indicates that the interaction and/or impact of language components on each other should be more fully considered in both theoretical and clinical contexts