| Sentence repetition tasks are widely adopted as an efficient clinical marker fordiagnosing children with specific language impairment (SLI) in many othercountries. Nevertheless, there is no such relevant study on sentence repetition inMandarin-speaking children with SLI. The present study focuses on the sensitivityand specificity of sentence repetition tasks, aiming at examining whether sentencerepetition tasks could be used as useful clinical markers for Mandarin-speakingchildren with SLI. Meanwhile, deficits on Mandarin-speaking children with SLIhaven’t been disclosed in any paper. Therefore this thesis will not only focus on thediagnostic accuracy of sentence repetition tasks, but also reveal the nature of relativeclauses and double object constructions which are proved to be more difficult forchildren with SLI to acquire.This study is conducted on three different groups of children in Changsha,Guangzhou, Hefei and Xinxiang, with the total number of36, including a SLI group,a typically-developing age-matched (TDA) group and a typically-developingyounger (TDY) group. Each group consists of12children. Grammatical items testedin the sentence repetition task involve twelve different items: classifiers, passives,negations, pivotal constructions, pronouns, relative clauses, aspect marker le,topicalizations, double object constructions, aspect marker zai, singular-plural formsand ba constructions. Children of the three groups are required to repeat thesesentences of each grammatical item respectively.Based on the results and analyses of the experiment, we find that there aresignificant differences between SLI and their TDA peers in all the grammatical itemsexcept the following four items: aspect le, aspect zai, ba constructions andsingular-plural forms. However, to be an efficient clinical marker, analyses should bemade on both the sensitivity and specificity of the rest of eight grammatical items.Meanwhile, a threshold score should be used as a cut-off point for predicting potential clinical markers, which means that in order to be considered as a “good”marker, the behavior in question must be present in individuals who have thedisorder and absent in those who do not. Therefore cut-off points are needed to beselected for each test first, with children scoring at or below the cut-off defined asimpaired and children scoring above the cut-off classified as non-impaired.According to the data analyses, we find that the performance on relative clauses at16thpercentile and double object constructions at25thpercentile results in the bestidentification with both sensitivity and specificity exceeding80%.In the study, we have made a tentative analysis about why relative clauses anddouble object constructions could be good potential clinical markers mainly from theperspective of syntactic knowledge. There are two types of relative clauses involved.We find that subject relative clauses seem more difficult for children to repeat thanobject relative clauses, the causes of which are analyzed in terms of word order,Structural Distance Hypothesis, Dependency Locality Theory and RelativizedMinimality to account for the asymmetry. In addition, double object constructionsalso present a big challenge for children to repeat, which is largely due to thecomplexity of the inner structures.The present study provides more evidence for the efficiency of sentencerepetition tasks as a clinical marker, which will facilitate the identification ofchildren with SLI and thus provide additional tools available for intervention andtreatment of Mandarin-speaking children with SLI. If Mandarin-speaking childrenwith SLI are shown to have deficits in sentence repetition tasks as in other languages,then we would be one step closer to uncovering the cognitive mechanisms thatcontribute to language impairments, which can also testify some relevant theoreticalhypotheses made by scholars from home and abroad and help us find out theunderlying causes of SLI children. |
PDF Full Text Request