Retour vers liste

Détail de la contribution

Auteur: Natalia KARTUSHINA

Co-Auteur(s): Uli FRAUENFELDER, Université de Genève, Suisse

Native language transfer revisited: L2 phonological development in advanced learners

Abstract/Résumé: The Full Transfer/Full Access (FT/FA) framework (Schwartz and Sprouse, 1996) makes two major claims. First, the native language (L1) constitutes the initial state of second language (L2) acquisition. Thus, at the phonological level, the L1 inventory of phonemes is used via transfer (FT) to process L2 segments (Archibald, 1998). Second, if during learning the L1 system cannot be used to assign a representation to L2, then the learner will restructure the phonological inventory drawing from Universal Grammar (UG) options (FA). Consequently, based on UG principles, new features or phonemic categories that are inexistent in L1 may be created for use in L2 (Escudero and Boersma, 2004). In this study we examine how well this (FT/FA) framework accounts for the phonological development of French (FR) L2 vowel production (PR) in experienced adolescent Spanish (SP) learners. Compared to FR, SP has fewer vowels and lacks the mi-open and mi-closed vowel contrast that is basically marked by F1 (degree of openness) differences. This learning situation is particularly adapted to testing this framework since to master FR vowels SP learners will have to go beyond the FT of their phonological space and expand it from 5 to 10 vowels and to create 3 height contrastive pairs. Twenty native SP speakers learning FR as their L2 (mean 4.4 years) in a SP high school were tested on PR of 1) ten FR oral vowels (naming and repetition tasks to minimize potential orthographical bias, Bassetti, 2008) and 2) five SP vowels (reading). Our predictions are the following. If FT is still functional at this stage, then we expect the transfer and use of SP vowels in FR PR. Thus, 1) acoustically similar SP vowels /i, e, a, o, u/ will be used in L2 PR (Flege, 1995) and 2) contrastive FR vowels will be assimilated to SP categories (Best, 1995). If, on the other hand, L2 SP speakers have FA to L2 phonology, they should develop an L1-like mechanism to distinguish FR height-contrastive vowels on the basis of the F1 value: the higher F1, the lower the perceived vowel (Ménard et al., 2002). Therefore, we expect SP PR in two tasks to differ from FR native and to be marked essentially by differences in F1 value: high-low. The results revealed that with regard to the FT predictions 1) SP speakers employ L1 categories to pronounce L2 sounds, only /a/ vowel was significantly reset to new L2 value (more open); 2) only /e/-/ε/ contrast is assimilated. As for the FA predictions, L2 speakers have ‘partial’ FA to the vowel height contrast that is task and contrast dependent. In sum, we observe a developmental phonological shift: partial FA within persistent FT.