Perception of expressive prosodic speech acts performed in USA English by L1 and L2 speakers
Résumé
Attitudes have been described for different languages, with varying labels or contexts of occurrence for same labels. It renders cross-cultural comparison uncertain. A corpus was designed to bypass these limitations. This paper focuses on USA English produced by L1 and L2 speakers. The best performances in 9 attitudes are used in a forced-choice test, in both audio and visual modalities. Results show that 6 categories group the presented attitudes in coherent sets. The cultural origin affects marginally the categorisation of the expressions. An acoustic analysis of the fundamental frequency and intensity allows to test the predictions of two theoretical propositions the Frequency code and the Effort code. It concludes to a main coherence of cross-language expressivity, and discusses differences. For negative expressions of imposition, L1 speakers follow the Frequency code and L1 listeners expect this; L2 speakers use the Effort code in the same situations, leading to confusions in the audio-only modality. Differences for seduction and irony are also discussed.