The production of wh-questions in Italian-speaking children with SLI

ABSTRACT

We aim at determining whether 7-year-old Italian-speaking children with Specific Language Impairment (SLI): (1) have problems with the production of wh- questions; (2) display a subject/object asymmetry in producing which- and who questions; (3) attempt to simplify questions, especially which- questions; (4) have difficulties with movement and verbal agreement in wh- questions. We elicited subject and object who and which NP questions in 10 children with SLI (M = 7;2), in 10 chronological age (CA)-matched controls (M = 7;2) and 10 language-matched controls (M = 5;2). Results showed that (1) children with SLI produced fewer questions than both control groups; (2) a subject/object asymmetry was observed in who questions but not in which NP questions; (3) which NP questions were more problematic than who questions; (4) children with SLI produced more agreement errors and resorted to simplification strategies to avoid wh- question production. Results point to a grammatical deficit due to the computation of complex grammatical relations and suggest that there is a misalignment among pieces of linguistic competence needed to form Italian wh- questions (wh- movement and agreement computation). Outcomes have implications for clinical assessment recommending the production of wh- questions to be considered in the evaluation of SLI in Italian.

KEYWORDS:

• Wh-question
• agreement
• SLI
• developmental language disorders

Acknowledgements

The authors thank all the children who participated in the study, the speech therapists, Mirta Vernice and Chiara Branchini and the students who helped in data collection. Fabrizio Arosio takes responsibility for the design of the study, for the statistical analysis, and for the following sections: Production and comprehension of Subject and Object questions in children with SLI, The present study, Participants, Materials, Procedure, Scoring, Reliability and Error Coding, Results, Discussion. Maria Teresa Guasti takes responsibility for the following sections: Introduction, Italian subject and object wh-questions, Production and comprehension of Subject and Object questions in Italian-speaking children with typical development.

Declaration of interest

The authors have no conflict of interest to declare.

Notes

¹ (E) in its object reading.

² We can alternatively assume that the post-verbal subject is right adjoined to the VP (Guasti, 1996) in a position that is not the position were focalized (or new information) subjects are located in a declarative sentence (Belletti, 2001) and is anaphoric to a pro in its canonical subject position. See Cardinaletti (2007) and references found there for alterative accounts of the adjacency requirement.

³ We chose to match children on comprehension LAs rather than on productive abilities, because there is no standardized quantitative test measuring productive abilities and adequate for the age under consideration. MLU, which is typically used to match children of younger ages, is inadequate at 7 years. In fact, MLU is not a measure of syntactic complexity, but just of average number of words in a sentence. Children with SLI may provide sentences with adequate length for their age. However, these sentences are likely to be structurally simpler than those of control children. A piece of evidence in this direction is offered by the production of clitic pronouns. It has been shown that failure to produce these pronouns is a clinical marker of SLI in Italian at 5 years (Bortolini et al., 2006) and at 7 years (Arosio, Branchini, Vernice, Barbieri and Guasti, 2014). Seven-year-old children with SLI fail to produce clitic pronouns consistently and obtain scores similar to those of 2.6-year old TD Italian-speaking children. Interestingly, they do not omit clitics at this age, but replace them with the corresponding nouns. Thus, their sentences have the same length or are longer than those of control children, who provide clitic pronouns, but are structurally simpler. In this task, SLI children would even resemble CA mates in terms of their MLU, but under a more careful analysis they are severely impaired.

Additional information

Funding

This work was supported by the LLP grant “Crosslinguistic Language Diagnosis” TRA [grant number STUCOR 2007 – 1992/001 – 001. 135295-LLP-2007-UK-Ka1scr to MTG].