CCIL Psychology: Thierry Nazzi (CNRS-Univ. Paris Descartes)

Syllabic segmentation in French: behavioral and ERP evidence

Abstract:

We will present data on the early development of speech segmentation abilities in French-learning infants growing up either in Paris or in Montreal. The importance of studying word segmentation in French comes from the finding that English-learning infants use various word boundary cues, many of which are language-specific. One of the most important cues used by young English infants is a prosodic/rhythmic cue: these infants segment fluent speech into trochaic units, which correspond to the rhythmic unit of English. This trochaic-based segmentation procedure would not be useful for French as lexical accentuation is weakly marked and word-final in French, and the rhythmic unit in French is the syllable. Thus, the development of word segmentation in French should follow a different trajectory in which individual syllables play a more important role.

Previous studies have found apparently diverging results for Parisian and Canadian French infants. Nazzi et al. (2006) provided evidence of initial syllabic segmentation: Parisian infants started segmenting syllables between 8 and 12 months of age, and whole bisyllabic words between 12 and 16 months of age. On the other hand, Polka and Sundara (2003) found that Canadian infants segment (whole?) bisyllabic words at 8 months when presented with Canadian or Parisian French stimuli.

We will present new studies further exploring the issue of early syllabic segmentation in French. First, we will present a series of experiments exploring the differences in outcome found between the Parisian and Canadian infants. Indeed, because the stimuli used in both studies were different, the differences found could either come from the fact that infants were learning two different dialects, or more prosaically from the fact that the stimuli had different characteristics. We will present data actually showing a complex interaction between stimuli specificities and dialectal effects.

Second, we will present the results of two ERP experiments with Parisian French- learning 12-month-olds re-evaluating the issue of syllabic segmentation using ERPs. Our results show that segmentation effects can be observed for both the initial and non-initial syllables of bisyllabic words. They also show different ERP signatures for the two syllabic positions, raising the possibility (to be explored in the future) of infants’ sensitivity to the difference between initial and non-initial syllables.