Type of publicationArticle ( AbstractBackground: It is well documented that children with specific language impairment (SLI)

A B S T R A C TAlthough sentence repetition is considered a reliable measure of children's grammatical knowledge, few studies have directly compared children's sentence repetition performance with their understanding of grammatical structures. The current study aimed to compare children's performance on these two assessment measures, using a multiple-choice picture-matching sentence comprehension task and a sentence repetition task. Thirty-three typically developing children completed both assessments, which included relative clauses representing a range of syntactic roles. Results revealed a similar order of difficulty of constructions on both measures but little agreement between them when evaluating individual differences. Interestingly, repetition was the easier of the two measures, with children showing the ability to repeat sentences

Purpose This study investigated the relationship between 2 components of memory—phonological short-term memory (pSTM) and working memory (WM)—and the control of relative clause constructions in children with specific language impairment (SLI). Method Children with SLI and 2 control groups—an age-matched and a younger group of children with typical development—repeated sentences, including relative clauses, representing 5 syntactic roles and 2 levels of matrix clause complexity. The Working Memory Test Battery for Children was administered. Results All 3 groups showed significant associations between pSTM and both types of matrix clause construction. For children with SLI, significant associations emerged between (a) WM and more complex matrix clause constructions, (b) WM and relative clauses including a range of syntactic roles, and (c) pSTM and the least difficult syntactic role. In contrast, the age-matched control group could repeat almost all syntactic roles without invoking the use of either memory component. Conclusions The role of pSTM and WM in the production of relative clauses by children with SLI is influenced by the degree of difficulty of the structure to be recalled. In therapy, the effect of WM limitations can be minimized by approaching each structure within the context of a simple matrix clause.

Studies examining productive syntax have used varying elicitation methods and have tended to focus on either young children or adolescents/adults, so we lack an account of syntactic development throughout middle childhood. We describe here the results of an analysis of clause complexity in narratives produced by 354 speakers aged from four years to adulthood using the Expressive, Receptive, and Recall of Narrative Instrument (ERRNI). We show that the number of clauses per utterance increased steadily through this age range. However, the distribution of clause types depended on which of two stories was narrated, even though both stories were designed to have a similar story structure. In addition, clausal complexity was remarkably similar regardless of whether the speaker described a narrative from pictures, or whether the same narrative was recalled from memory. Finally, our findings with the youngest children showed that the task of generating a narrative from pictures may underestimate syntactic competence in those aged below five years.