A cross-language investigation of word segmentation by bilinguals with varying degrees of proficiency: Preliminary results.

Author(s): Annie C. Gilbert, Inbal Itzhak and Shari Baum


An extensive body of research on word segmentation has shown that different languages rely on different cues and strategies to segment meaningful units from the speech stream. These cross-language differences make segmentation difficult for L2 learners, and some previous work showed that bilingual speakers tend to keep applying their L1 segmentation cues to the L2. But bilingual experience varies a great deal, even within bilingual communities, so one might ask if such a pattern applies across all bilinguals regardless of language proficiency, dominance, or everyday use. To investigate this, we designed a cross-modal priming task in which a wide range of English-French bilinguals listened to English and French sentences with ambiguous syllable strings containing either two monosyllabic words (e.g. key we) or one bisyllabic word (e.g. kiwi), produced with context-specific natural prosody. A picture prompt representing either the first monosyllabic word (a key), or the bisyllabic word (a kiwi) was presented at the offset of the first syllable of the ambiguous region. Each sentence was presented paired with each picture. Preliminary analyses of a subgroup of English dominant participants show that they process French and English ambiguous strings differently, and that their segmentation schemes seem to vary with L2 proficiency.