Acoustic Analysis of a Canadian French Speakers Production of English Stress

Abstract

Lexical stress in English plays an important role in native speakersperception and processing of speech. French learners of English have a fixed final-stress system, which differs from the free variable-stress system of English. This difference in stress systems between the two languages may result in interference from the native language stress system on the production of English stress by French learners of English. This paper reports a phonetic experiment that sets out to test the hypothesis of native language interference on stress production. One female native English speaker and one female French learner of English produced 50 initially-stressed, trisyllabic English words of French etymology (balanced for word frequency) twice in a carrier phrase I didnt say ____, but you said ____ yesterday.In the carrier phrase, the first instance of the target word functions as new information and occurs on an intonational phrase boundary, and the second occurrence of the target word serves as old information and occurs in the intonational phrase medial position. The carrier phrase is designed this way in order to avoid the boundary effect of high tone that is often observed due to the French prosodic structure, and to investigate the effect of information structure of new and old information. A 2 (language) × 3 (syllable position) repeated-measures ANOVA was conducted on the data for each parameter of stress: fundamental frequency, amplitude and duration. The results supported the hypothesis of native language interference for the French learner of English. Further analysis was performed for the target word in the prosodic structure and word frequency. Unlike the native English speaker who showed an effect of frequency and prosodic structure, neither had an effect on the overall results for French learners of English.