Peperkamp, S. 1 & Brazeal, J. 2
1 Laboratoire de Sciences Cognitives et Psycholinguistique, ENS, Paris, France
2 Department of French and Italian, University of Texas at Austin, Austin, USA
Native speakers of French, a language without contrastive stress, have difficulty perceiving stress contrasts. Using a pretest-posttest design with 10 trainees and 10 controls, we examined whether French listeners can improve their stress perception with auditory training. We used naturally spoken, phonetically varied stimuli in a sequence recall task, in which participants have to recall sequences of two auditorily presented non-words that differ either in the position of stress (test condition) or in a phoneme (control condition). At the end of six 30-minute training sessions on stress contrasts, the trainees showed no improvement in their perception of stress: an ANOVA with factors Session (Pretest/Posttest), Group (Trainees/Controls) and Contrast (Phoneme/Stress) yielded an effect of Contrast only (F(1,18)=53.9, p<.0001), with worse performance on the stress contrast. This result contrasts with previous findings that listeners can be effectively trained to improve their perception of non-native contrasts. We argue that the lack of a training effect is task-specific. In particular, contrary to previous studies that used a 2AFC identification task, in the sequence recall task participants can neither use a low-level acoustic response strategy nor rehearse the stimuli subvocally. We discuss the consequences for theories of phonological learning.