Abstract

Noun clauses are grammatical constructions that are of relevance both to typical language development and impaired language development. These clauses have been part of published techniques for the clinical analysis of language samples, and computer software for the automated analysis of clinical language samples has attempted to identify noun clauses, with limited success. The present study examined the development and clinical use of noun clauses as well as the automated identification of these clauses. Two sets of language samples were examined. One set consisted of 10 children with specific language impairment (SLI) whose age ranged from 7;6 to 11;1 (years;months), 10 peers matched for language development equivalence, and 10 peers matched for chronological age. The second set of samples were from 30 children considered to be typically developing, who ranged in age from 2;6 to 7;11. Language sample utterances were manually coded for the presence of noun clauses (including wh- noun clauses, that- noun clauses, and gerunds.) Samples were then automatically tagged using software. Results were tabulated and compared for accuracy. ANCOVA revealed that differences in the frequencies of WH-infinitive noun clauses and gerunds were significant between the matched groups. "Zero that clauses" (that-noun clauses containing no subordinator that) and gerunds were significantly correlated with age. Kappa levels revealed that agreement between manual and automated coding was high on WH-infinitive clauses, gerunds, and finite wh-noun clauses.

Degree

College and Department

David O. McKay School of Education; Communication Disorders

Rights

http://lib.byu.edu/about/copyright/