- Wh-Movement in Children with Grammatical Sli: A Test of the Rddr Hypothesis
This article presents a test of the proposal that a subgroup of children with grammatical-specific language impairment (G-SLI) have optional movement (the representational deficit for dependent relations (RDDR) account, van der Lely 1998) by investigating wh-movement in fifteen G-SLI subjects and twenty-four younger children matched on language abilities (LA controls). The RDDR/optional movement account predicts that G-SLI subjects would have deficits with both wh-operator and Q-feature movement and therefore would have particular problems producing object questions. We elicited 36 questions balanced for subject and object questions and wh-words (who, which, what). The G-SLI subjects were significantly impaired in producing wh-questions, showing particular difficulties with object questions in relation to the control children. The majority of G-SLI subjects (80%) evinced both wh-operator and T/Q-feature movement errors whereas only one control child (4%) did so, yet on occasion all the G-SLI subjects used appropriate movement operations to satisfy the wh-criterion. We conclude that the RDDR account whereby movement is optional is consistent with the findings of correct and incorrect wh-question formation. Thus, this first test of the RDDR account of G-SLI is supported by the findings. We discuss the possible underlying nature of a grammar that could cause such optionality, the implications for normal and impaired language acquisition, and the generalizability of the findings to other groups of children with SLI. We propose that in the face of no movement, the wh-word and, on occasion, do are merged in situ in the CP, and function as an interrogative adjunct.*
Specific language impairment (SLI) is a heterogeneous disorder of language acquisition in children who have no other apparent cognitive, neurological, or environmental impairment that can account for their deficit (Menyuk 1964). SLI affects around 7% of children and can persist into adulthood (Leonard 1998). Evidence from diverse sources indicates that there is a significant genetic contribution to the disorder (Fisher et al. 1998, Bishop et al. 1995, Hurst et al. 1990, SLI Consortium 2002, van der Lely & Stollwerck 1996). Investigations of subjects with SLI not only inform us about the nature of their disorder but also provide theoretical insight into the innate basis of language and the development of specialized cognitive systems, like the grammatical system.
Towards this end, van der Lely and colleagues have investigated a subgroup of Grammatical (G)-SLI children, who exhibit a relatively pure, domain-specific grammatical impairment (van der Lely 1996b, 1998, van der Lely & Christian 2000, van der Lely et al. 1998). Previous investigations into G-SLI subjects’ syntactic abilities show that they inconsistently manipulate core aspects of syntax, including tense and agreement marking, the assignment of thematic roles to noun phrases, and the assignment of coreference to pronouns and anaphors in sentences when only syntactic cues are [End Page 153] available, as well as comprehension and production of embedded phrases and clauses (van der Lely 1996b, van der Lely & Dewart 1986, van der Lely & Hennessey 1999, van der Lely & Stollwerck 1997). The representational deficit for dependent relations (RDDR) hypothesis, developed to account for G-SLI grammar, identifies the source of the linguistic deficit in the computational syntactic system, that is, in the syntax proper (van der Lely 1994, 1998). The RDDR contends that the core deficit responsible for G-SLI children’s grammar is in ‘Movement’ (Chomsky 1995), and more specifically, that whereas the basic operation/rule ‘Move’ in normal grammar is (by definition) obligatory, in G-SLI grammar it is optional (van der Lely 1998). This would account for the broad range of language deficits and a core feature of G-SLI grammar—the inconsistent use of certain grammatical rules (van der Lely 1998). Further characterization of the RDDR hypothesis is provided below. This study tests the RDDR proposal by investigating subject and object question formation in G-SLI subjects (aged 11: 3–18:2 [years:months]), exploring their ability to compute wh-movement in sentences, and comparing their performance with that of younger, language-matched control children (aged...