Errors and repairs in French language use of Turkish-French bilingual children and teenagers

Errors and repairs were subject matter of many studies in a developmental perspective (Clark & Andersen, 1979 ; Clark, 1985 ; Ochs, 1985 ; Levelt, 1983 ; Karmiloff-Smith, 1986, 1993). Ochs (1985: 785) defines the error as “a) a deviation from either a socially variable or a categorial norm and b) warrant negative feed-back.” Like Blanche-Benveniste (1997), Wigglesworth (1990: 121) says that repairs can “occur (...) as a result of the speaker\'s own monitoring of his or her speech.” Clark & Andersen (1979 divide repairs into two groups: those concerning the code and those intended for the interlocutor, to make the message understood. From these definitions it can be assumed that errors precedes repairs. A qualitative but also quantitative study of these two phenomena can shed some light on the developmental process of the mastery of the language to the bilingual children (Ochs, 1985). Indeed, the errors are indications of an incomplete knowledge of the considered domain; their analysis, a means to know the parts of the system which are not still completely automated, while the repairs can help to understand the processes and the current acquisition. This paper aims to provide evidence for the ways both errors and repairs develop and show the progressive mastery of French acquired by Turkish-French bilingual children and teenagers living in the Turkish immigrant community in France. The data is composed of narratives elicited by using a picture book task Frog, where are you? (Mayer, 1969). The informants are grouped into six groups: one group of nursery school children (5-year-olds), two groups of primary school students (7 and 10 years of age), two groups of secondary school adolescents (12 and 14 years of age) and one group of high school students (17 year-old). Each group consists of at least 10 subjects. Such a study turned out relevant because, on the one hand, differences appeared in the other domains; we already observed notably a delay which tends to disappear with age for example in the study of the macrostructure to the bilingual with regard to the monolingual (Akinci, Jisa Kern, 2001) or that on the temporality (Akinci & Kern, 1998, Akinci 2001). On the other hand, certain theories on bilingualism place the children stemming from the immigration in the category of semilinguals (Skutnabb-Kangas & Toukomaa, 1976), who not only confuse and mix both languages but also share common points with the learners of a second language among which one of the most important would be the instability of the knowledge. Leaving of this theoretical background, we make the following hypotheses: Hypothesis (1): as far as the acquisition of a language is made at a child\'s in a progressive way, going from a stage where the child begins his acquisition until reach the target system (that of the adult), we predict a decrease with the age not only of the number of errors, but also of their variety. As for the repairs, we think that repairs are going to increase. Hypothesis (2): the comparison of our results with those of monolingual French children (Bange & Kern, 1996) is going to inform us, at first, on specific errors and repairs of bilingual subjects, because of their membership in a double linguistic system, and, then, it is going to show the decrease, with the age, the distance which one can observe between both populations. The results show that the development of the mastery of French of bilingual is at the same moment in keeping with that of the results brought reported on the learners children of a second language (Cummins, 1991 ; Snow & Hoefnagel-Hohle, 1978) and with that of monolingual French