Retour vers liste

Détail de la contribution

Auteur: Karen SMITH-LOCK

Co-Auteur(s): Anna NOTLEY, Macquarie University, Australia Louise DELANE, University of Western Australia Stephen CRAIN, Macquarie University

Titre:
The independence of relative clause comprehension in SLI


Abstract/Résumé: Typically developing children and children with specific language impairment (SLI) have difficulty comprehending sentences with relative clauses (e.g., Hamburger & Crain, 1984; Friedmann & Novogrodsky, 2004; Adani, 2011). To determine what leads to difficulty, previous studies have manipulated grammatical and non-grammatical features of the task, but few studies have directly measured other linguistic or cognitive abilities that might affect performance. The present study examines the relationship between the comprehension of relative clauses and other cognitive-linguistic abilities, in 43 6-year-olds with SLI. Children were asked to act out relative clause structures, as in (1), in three conditions. In condition 1, there were three animate noun phrases. In condition 2, the final NP was inanimate. In condition 3, there were two toys corresponding to the head of the relative clause, and before children acted out the test sentence, they acted out the event mentioned in the subordinate clause with one toy, and they performed a different action with the second toy. This satisfied the felicity conditions for the use of a restrictive relative clause, and it distinguished one parrot from the other. (1) The koala pushed the parrot that kicked the dingo. NP1 NP2 NP3 Children were also tested on an omnibus test of language (CELF-IV), non-verbal IQ (WNV), non-verbal memory (Leiter) and phonological memory (CTOPP). Response types were scores as mature (NP1 acted on NP2; NP2 acted on NP3), immature (NP1 acted on NP2 and NP3), or 'other'. One analysis included 16 children with receptive SLI. Responses were spread over three types: 25% gave mature answers in condition 1; 25% made errors in condition 1 but improved in conditions 2 and/or 3; and 50% made errors in condition 1 but did not improve in conditions 2 or 3. All children except two produced at least two age-appropriate interpretations. Children with poor phonological memory were divided equally between the groups. A second analysis included all 43 children. The pattern was similar. 21% gave all mature responses in condition 1. Half of the remaining children produced correct responses in condition 2 or condition 3. All but five children produced at least one mature response. Performance was unrelated to scores on the CELF, non-verbal IQ, phonological or verbal memory. In summary, in a sample of 43 6- to 7-year old children with SLI, 90% were able to comprehend some relative clauses under certain experimental conditions. Successful comprehension was independent of phonological memory, general language skill, non-verbal memory and non-verbal IQ.