Différences linguistiques et dialectales dans la mise en place des procédures de segmentation de la parole

Abstract

This paper presents a review of recent studies investigating the issue of the early segmentation of continuous speech into words, a step in language acquisition that is a prerequisite for lexical acquisition. After having underlined the importance of this issue, we present studies having explored young infants' use of two major segmentation cues: distributional cues and rhythmic unit cues. The first cue is considered to be non-specific to the language spoken in the infant's environment, while the second cue differs across languages. The first cue thus predicts similar developmental trajectories for segmentation across languages, while the second cue predicts different types of developmental trajectories according to the rhythmic type of the language in acquisition. It was found that segmentation abilities emerge around 8 months of age and develop during the months that follow, and that the weight of the different cues vary across languages, according to the developmental period, and probably as a function of dialectal differences within a given language. We then discuss the fact that word form segmentation requires in all likeliness the combined use of different segmentation cues from the youngest age. We conclude by delineating some pending issues to be addressed in future research.