Patrons sociolinguistiques chez trois générations de locuteurs acadiens

Verbs in Acadian French mark third-person-plural subject-verb agreement with the traditional suffix -ont, that is realized by the nasal vowel /õ/ (les enfants i-jou-ont “the children play”), as well as by the zero morpheme that is generally found in contemporary French and that is represented in writing as -ent (les enfants i-jou-ent, les enfants jou-ent). This study reports on variation in the use of these forms by three generations of speakers from northeastern New Brunswick (Canada) and addresses Labov’s “transmission problem”: how children learn to talk differently from their parents, and yet in the same direction, over several generations. Data from a stratified corpus of 16 adult speakers show an ongoing change: the zero morpheme is replacing the traditional -ont suffix. The main external factor that conditions this variation is social network: speakers with closed affi liation networks tend to conserve the traditional variant while those who have open networks use the zero morpheme almost exclusively. Among closed network speakers there is a significant age-by-gender interaction: older males have the highest frequencies of occurrence of the traditional -ont variant followed by younger females; younger males and older females have the lowest frequencies. The main internal conditioning factor is verb class, where classes are arranged according to the number of bases or stems. Verb classes with a small number of bases (for example, arriver “to arrive”, appeler “to call”) are more likely to be associated with conservation of the traditional suffix, while those with larger numbers of stems (including verbs such as avoir “to have”, aller “to go”, faire “to do” that have suppletive forms) are less likely. The corpus of children’s data includes recordings made with 24 speakers in three age groups (3–5, 7–9 and 10–12 years of age). The distribution of the traditional -ont suffix with respect to social network is the same as that found among adults: as early as 3 to 5 years of age, children from families with closed networks use the traditional form almost exclusively while those from open network families use it infrequently, favouring the zero morpheme. However, among all children from closed network families, the frequency of use of the traditional variant is considerably higher than among adult speakers. Furthermore, unlike the pattern observed for adult speakers, there are no significant gender differences at any age level. With respect to the “transmission problem” the results show that the verb class constraint that is formulated by children is not identical with adults’ patterns, but it does resemble more closely the model of adult women speakers than that of adult male speakers. Interestingly, the process of re-structuring the internal constraint on the traditional –ont suffix variation begins relatively late (in the 10–12 year group) when compared with ages reported for the acquisition of constraints on phonological variation. The results of this study provide evidence for a complex picture of the acquisition of sociolinguistic competence; the acquisition of social and linguistic constraints on variation does not follow a clear linear order where social patterning is acquired before (or after) linguistic constraints.