Developmental perspectives on Specific Language Impairment: Evidence from the production of wh-questions by Greek SLI children over time

Abstract

This study investigates the performance of seven Greek SLI children on the production of wh-questions at two points in their linguistic development. In particular, it compares and discusses the results of two experiments on the production of subject and object wh-questions; the second experiment was carried out 5 years after the first one. Despite the SLI children's significant improvement in the production of subject and object wh-questions, the results indicated persistent problems with specific grammatical operations. Adopting a developmental perspective, issues concerning the linguistic deficits in SLI are discussed.

Keywords:

• SLI
• language development
• wh-questions

Notes

1 See also Thomas (2005) for a current review of the PDH and its predictions for compensation in SLI.

2 This is the mean non-verbal IQ performance of the seven SLI children who participated in the first and second experiment. See Stavrakaki (2001) for the non-verbal IQ performance of all children who participated in the first experiment.

3 The matching was made on the basis of raw scores (and not on the basis of standardised scores) because this test has not been standardised.

4 The child was visually and orally presented the story and she/he had to check whether a silly puppet was able to understand what was going on in the story by asking the puppet a question; in this respect the target wh-question makes sense in the context.

5 See van der Lely and Batell (2003) for such a distinction between grammatical and non-grammatical errors.

6 The parametric t-test was used because the Kolmogorov-Smirnov test showed that data from SLI children were distributed normally.

7 The non-parametric Mann-Whitney test was used because the Kolmogorov-Smirnov test showed that data from TD children did not follow normal distribution.

8 It should be noted that there were a few instances of highly semantically related verbs which were produced instead of the verb that was used in the story (e.g., esprokse: pushed-3s instead of htipise beat-3s). These responses were not thought to be incorrect. Children might interpret the action of “hitting” as “pushing” in the scenarios acted out with toys.

9 See Tsimpli and Stavrakaki (1999) for a brief discussion on bare singular nouns in object position in Greek.

10 This is a tentative suggestion. The reasoning behind this suggestion is that since non-grammatical errors cannot be necessarily attributed to deficient grammatical operations, then other cognitive mechanisms and/or abilities might be involved in the significant decrease of those errors. However, except for the general cognitive IQ measures that have been used in both experiments and a preliminary assessment of the SLI children's short-term memory and theory of mind in the first experiment (see Stavrakaki, 2001), no-systematic investigation of the cognitive abilities of the SLI children has been made. It is, therefore, unclear whether decrease in the non-grammatical errors can be attributed to the general cognitive capacity or other more specialized cognitive systems. I must thank an anonymous reviewer that drew my attention to that point.

11 The same children participated in Stavrakaki's studies (Stavrakaki, 2001, 2002); some of those children have been included in other studies on Greek SLI (Dalalakis, 1996; Clahsen & Dalalakis, 1999; Tsimpli, 2001; Varlokosta, 2002).

12 DP-splitting is possible with adjectives, quantifiers, and demonstratives, as well as with tinos (whose).

13 However, Greek preschool children with SLI show difficulties with case marking among other problems with grammatical features (Stavrakaki, 1996; Tsimpli, 2001, among others).

14 See also Eisenbeiss, Bartke, and Clahsen (2005/2006) who studied the system of case marking in German speaking children with SLI. They have shown that the SLI children do not have any problems with structural case marking: by contrast, they overuse structural cases in contexts where lexical case marking is required. Based on this evidence, Eisenbeiss et al. suggest that the SLI children show increased grammatical sensitivity.

15 This proposal does not implicate generalized problems with feature-checking operations in SLI children. Although possible in SLI (and other cases of language disorders), problems with other feature-checking operations (e.g., tense and /or agreement) have not been attested at these particular developmental stages of the SLI children (see Arabatzi & Edwards, 2002, who report deficits in feature-checking operations in English-speaking patients with aphasia).