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.
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.