Subject–verb agreement, object clitics and wh-questions in bilingual French–Greek SLI: the case study of a French–Greek-speaking child with SLI

Abstract

In this study we investigate the linguistic performance of a school age French–Greek simultaneous bilingual boy with specific language impairment (SLI) on the production of subject–verb agreement, object clitic pronouns and wh-questions. In addition, we compare his performance on these linguistic structures with that of two French–Greek bilingual children with typical development matched on language age. Furthermore, we discuss the performance of the child with SLI in the light of published data from monolingual French and Greek school-age children with SLI, as reported in the relevant literature. The results indicated that the performance of the bilingual French–Greek-speaking child with SLI was highly comparable to that of bilingual French–Greek typically developing children and monolingual children with SLI speaking French or Greek. In addition, the results revealed a language effect on the performance of the bilingual child with SLI on subject–verb agreement, object clitics and wh-questions. We adopt a comparative linguistic approach to interpret the results and argue that the specific linguistic properties of Greek and French were significant determinants for the manifestation of SLI in each language. We suggest that bilingualism per se did not aggravate the language deficits in this child with SLI. Finally, we discuss the clinical implications of our study for language intervention in bilingual individuals with SLI.

• specific language impairment
• bilingualism
• French language
• Greek language

Acknowledgements

We thank all children and their families for collaborating with us. We are grateful to two anonymous reviewers for comments and suggestions. We also thank the Director of the French School in Thessaloniki for helping us getting access to the French–Greek bilingual children. Many thanks go to Francoise Avgeris and Antonios Tsopanoglou for supporting our research. Furthermore, we thank Christian Dunez, Lucile Gravanis, Rinetta Kigitsioglou, Liliane Sprenger-Charolles and Laurie Tuller for discussing with us the French data. Thanks also go to Ioanna Talli for calculating the French verb lemma frequency for the subject–verb agreement task. In addition, we thank Polyxeni Konstantinopoulou, Vasiliki Koukoulioti, Ioanna Talli and Angeliki Zachou for comments on previous versions of this paper.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper.

Notes

1. With respect to the terminology concerning dual language learning, we follow Paradis et al. (2003: 2–3) who take the term ‘bilingualism’ to refer only to simultaneous acquisition of two languages from birth and the term ‘second language learning’ to refer to sequential L2 acquisition.

2. The term ‘simultaneous’ is used for these bilingual children who have been exposed to two languages from birth.

3. We should point out that the non-null status of French is a controversial issue in French linguistics. On the one hand, Legendre, Culbertson, Barriere, Nazzi, and Goyet (2010) point out that French is a null subject language as clitic subjects can be analysed as agreement markers. On the other hand, De Cat (2005) claims that subject clitics are not agreement suffixes and, thus, French is a non-null subject language. While the linguistic characterization of French as a null or non-null subject language is beyond the aims of this article, in our view, French seems to be a non-null subject language with weak inflection. This is so because our results indicate that the linguistic properties of French should not be identical with those of Greek (a typical null subject language with rich inflection) in the domain of inflection. Indeed, higher correctness scores for subject–verb agreement were achieved by two out of three participants of our study for Greek than for French (see Table VII).

4. However, as pointed out by an anonymous reviewer, in many varieties of spoken French clitic doubling is grammatical (see, Auger, 1994, and references therein).

5. Conevey (2010) refers to the radio corpus of Behnstedt (1973). This corpus was created from 1960 to 1970. It consists of oral reading of written texts.

6. An anonymous reviewer points out that morphological case marking for pronouns and NPs could provide a significant facilitation cue for the comprehension of wh-questions although morphological case errors are possible in production. In this study, we only present data from the production of wh-questions. A previous study on the comprehension of subject and object wh-questions by children with SLI and TD showed that children with SLI performed significantly lower than their LA peers only on who-object questions (not on which-object questions and wh-subject questions; Stavrakaki, 2001a). While it is interesting to investigate the role of case marking for the production versus comprehension of wh-questions, this particular investigation is beyond the aims of our study.

7. The attested dissociation between subject–verb agreement and tense marking in French SLI is consistent with the idea that the acquisition of tense and agreement proceed independently in child French (Legendre, Hagstrom, Vainikka, and Todorova, 2000).

8. For French the term ‘Verbal IQ’ corresponds to the ‘Verbal Comprehension Index’ and the term ‘Non-Verbal IQ’ corresponds to the ‘Perceptual Reasoning Index’.

9. Despite the wide use of the elicited production tasks in the acquisition literature, many researchers point out that these tasks sometimes involve high demands related to the task per se, which may impede linguistic performance (see, Eisenbeiss, 2010).

10. For this particular task, the verbs are widely used in everyday speech (write, read, kiss, wash, eat, go, get off, finish, drink, see). For French, we consulted a grade-level lexical database from French elementary school readers, MANULEX (Lété, Sprenger-Charolles, and Colé, 2004). MANULEX provides grade-based frequency lists of the 1.9 million words found in first-grade, second-grade and third- to fifth-grade French elementary school readers. The database consists of 23,812 lemmatized entries. For Greek, we have consulted a lemma and word frequency database, the IEL corpus (http://hnc.ilsp.gr/statistics.asp) which constitutes 47,013,924 words coming mainly from texts in newspapers. This corpus is regularly updated. The means (and SDs) of the verb lemmas we used in our task are provided in Appendix 1. We would like to point that, according to our view, the Greek database does not provide an objective calculation of word frequency for child language, as all words coming from written texts are mainly for adults. We, thus, believe that the Greek database is not comparable to the French one.

11. This is so because, as pointed out by an anonymous reviewer, the use of clitics in Greek and French is pragmatically driven.

12. The p-value is higher than the expected one presumably because of the small number of items. However, the difference is numerically strong (50% vs.100%).

13. These responses were produced in obligatory object wh-question contexts and could be interpreted either as subject or object wh-questions. For example, when the children had to produce an object wh-question, they produced a response that could be interpreted as a subject wh-question or an object wh-question:

• Child with SLI: Qui a embrassé l’ éléphant?

•                Who has-3SG kissed-PAST.PART.–the elephant

•                ‘Who did the elephant kiss?’

   While the above response is ambiguous, an anonymous reviewer suggests that its default interpretation is that of a subject wh-question.

14. As pointed out by an anonymous reviewer, this is a structure with non-canonical word order. Apparently this error type indicates problems with computing the correct word order for a wh-question.

15. Alternatively, as suggested by an anonymous reviewer, the child might not be (or become) aware of the pronoun or clitic requirement but rather of the necessity of a complement to the verb.

16. It should be noted that according to some linguists clitics undergo movement in Greek (see Alexiadou and Anagnostopoulou, 2000, for a review of the literature; Philippaki-Warburton, Varlokosta, Georgiafentis, and Kotzoglou, 2004). However, we follow the non-movement analysis for clitics in this article for two reasons. First, we believe that such an analysis can better explain the co-occurrence of DPs with clitics in clitic-doubling constructions; second, we claim that such an analysis can provide an adequate explanation for cross-linguistic differences in the acquisition of clitics between clitic-doubling and non-doubling languages, as is the case for the data of our study.

17. These forms underspecified for (morphological) case are legal words in Greek: they are forms in neuter gender instead of the target ones (marked for nominative or accusative case and feminine or masculine gender).

18. Remarkably, English children with SLI are reported to produce non-nominative subjects to a greater extent that TD children (Loeb and Leonard, 1991; Wexler, Schutze, and Rice, 1998; Radford and Ramos, 2001).

19. An anonymous reviewer points out that our suggestion that overt morphological case marking for NPs in wh-questions cause significant difficulties to the children with SLI contradicts with the claim that the highly inflectional properties of Greek facilitate the acquisition of subject–verb agreement. With respect to the monolingual school-age Greek children with SLI, our view is that that case per se does not cause problems to these children, who selectively fail with case assignment to post-verbal subjects in object wh-questions (see Stavrakaki, 2006); these difficulties are not attributed to a deficit in the feature of case per se but they are interpreted as effect of syntactic complexity. With respect to the child of the present study, we observed that there is a particular delay in case acquisition per se as this child behaves like monolingual preschool children with SLI with respect to case marking (the preschool children with SLI produce NPs underspecified for case). While our interpenetration is that there must be a delay in the acquisition of the feature of case by this child, we point out that most difficulties with case retrieval occur in the complex syntactic contexts of object wh-questions with non-canonical OVS word order; therefore, there is an effect of the syntactic context on morphological feature retrieval. An alternative suggestion is that it is quite possible that the acquisition of nominal and verbal domain by the children with SLI do not proceed simultaneously, as it might be affected by a number of factors, for example, the feature interpretability (see Clahsen, Bartke, and Gollner, 1997; Tsimpli, 2001). While case is a non-interpretable feature, the number and person features in Greek (and other pro-drop languages with strong inflection) can be considered as interpretable features (cf. Alexiadou and Anagnostopoulou, 1998). As different interpretations for the difficulties related to case marking in Greek are possible, we point out that further research is required to investigate the cause of these difficulties in Greek children with SLI.