Retour vers liste

Détail de la contribution

Auteur: Natalie BOLL-AVETISYAN

Co-Auteur(s): Anjali BHATARA, Université Paris Descartes, CNRS, France Annika UNGER, Universität Potsdam, Germany Barbara HÖHLE, Universität Potsdam, Germany Thierry NAZZI, Université Paris Descartes, CNRS, France

Titre:
Can stress ‘deafness’ be countered after extensive exposure to an L2?


Abstract/Résumé: In this study, we assess how so-called stress "deaf" French advanced learners of an L2 that has lexical stress group rhythmic patterns. Native speakers of French - a language without lexical stress - are less sensitive to stress than speakers of languages with lexical stress (e.g., Dupoux et al. 1997). This relative stress "deafness" persists even with L2 knowledge of a language that uses word stress (Dupoux et al. 2008). This suggests that French speakers' stress perception is filtered by their L1 independent of L2 experience. The persistence of stress "deafness" may surprise as rhythmic perception is thought to be guided by an auditory bias, the iambic-trochaic law (ITL; Hayes 1995). The ITL is based on the observation that humans group sound stretches varying in intensity as trochees (i.e. strong-weak), whereas they group sound stretches varying in duration as iambs (i.e. weak-strong; e.g. Woodrow 1909), though L1 prosodic structure can affect grouping preferences (Iversen et al. 2008). This raises the question of whether knowledge of an L2 with word stress may affect rhythmic grouping preferences in a stress "deaf" population. Participants were 40 French and 40 German monolinguals and 40 French advanced learners of German - a language with lexical, mainly trochaic stress. Participants were exposed to three types of nonsense streams (e.g. ...bezilemo...) alternating in duration, intensity, or in neither. Participants indicated whether the streams consisted of strong-weak or weak-strong pairs. A questionnaire was used to identify potential predictors of attainment. Results showed that all groups perceived trochees if intensity was varied and iambs when duration was varied (both p < .001). This effect was stronger in the monolingual Germans than in both French groups (p < .01). L2 learners and Germans grouped the control streams as trochaic, while the French monolinguals had no preference. The L2 learners differed significantly from the German (p < .01) but not from the French monolinguals. Adding factors of L2 experience to the model revealed that L2 learners with more L2 exposure gave more trochaic responses throughout, including streams in which duration was varied. This suggests that experience with L2 stress does not enhance sensitivity to stress cues for rhythmic grouping in stress "deaf" L2 learners. However, the development of a trochaic bias through L2 exposure indicates that learners internalize the fact that their L2 is mostly trochaic, which suggests a weakening of the L1 filter.