I[z] dat [s]o ? Productie van stemassimilatie in het Nederlands van Franstalige vreemdetaalleerders

This study examines the production of voice assimilation in Belgian Dutch and French and in the interlanguage of Belgian French-speaking learners of Dutch. Recent research has paid much attention to the production of voicing and devoicing assimilation by native speakers of different languages, but far less is known about the production of voice assimilation by foreign language learners. Besides, the few interlanguage studies of voice assimilation deal exclusively with learner varieties of English (Altenberg & Vago 1983; Cebrian 2000; Lew 2002; Simon 2010b). This study therefore fills a gap in scholarly publications on interlanguage phonetics and phonology. \nThe aim of this study is twofold. On the one hand it aims to provide a better understanding of voice assimilation processes in the interlanguage of French-speaking learners of Dutch. Particular emphasis is placed on transfer phenomena and on the interaction between language-specific patterns of the learners’ native language and universal principles of articulatory phonetics. On the other hand, the study proposes a detailed phonetic description of the realisation of obstruents and obstruent sequences in Dutch, in French and in the interlanguage of French-speaking learners of Dutch. Particular attention is paid to the possible influence of phonological factors – namely the underlying voice specification and the manner and place of articulation of obstruents – on the production of voicing and devoicing and on assimilation. \nThe material for this study was collected by means of a shadowing experiment. The target sounds were masked with noise and inserted in short sentences that the participants had to repeat simultaneously. The subjects were 15 Belgian French-speaking learners of Dutch and 15 Belgian native speakers of Dutch. The French-speaking participants completed the experiment both in French and in Dutch, and the native speakers of Dutch served as a control group. The data were segmented orthographically and phonemically and were analysed acoustically. The following acoustical measures were performed: presence of f0 at regular intervals in the obstruent, duration of the obstruent, duration of the previous and following vowel and f0 value of the neighbouring vowels. \nThe results show a strong tendency to obstruent devoicing both in the native and interlanguage Dutch data. In final position, obstruent devoicing in the interlanguage of the French-speaking informants results either from positive transfer of the L1 patterns(*) or from the phonetically-motivated universal tendency to devoice final obstruents in natural languages. In initial position, devoicing mainly occurs in fricatives and is attributed to articulatory constraints on the production of voiced fricatives. The phonetically-motivated tendency to devoice fricative further appears to have a facilitative effect on the learning of progressive assimilation in sequences of the type C+/v,z,ɣ/. However, French-learners of Dutch are still influenced by their native language when producing such sequences: they negatively transfer the regressive assimilation pattern of French in 19% of the cases. In the three language varieties, initial /b,d/ remain voiced when following /ə/, but tend to be devoiced by progressive assimilation in the native and interlanguage Dutch data when following a voiceless obstruent. This suggests that the assimilation rules of Dutch are inadequate, as they predict that regressive assimilation should occur in obstruents before /b,d/ (Booij 1995; Trommelen & Zonneveld 1979; Van den Berg 1972). \nRegarding the role of phonological factors in the production of voicing phenomena in the three language varieties, the results vary according to the phonological context. In final position, there is an effect of the underlying voice specification and of the manner of articulation of obstruents in French, with more voicing in /+voice/ than in /-voice/ obstruents, and in plosives than in fricatives. In the native and interlanguage Dutch data, final obstruents are invariantly voiceless. In initial position, the results show an effect of the manner of articulation of obstruents and of the place of articulation of fricatives, with more voicing in /b,d/ than in /v,z,ɣ/ and in front fricatives than in back fricatives. The effect of articulation place is also found in the interlanguage and French data when fricatives follow another obstruent. The further back in the mouth a fricative is articulated, the less voicing it shows. In obstruent sequences of the type C+/b,d/, an effect of the manner of articulation of the first consonant was found in the interlanguage data. The manner of articulation affects both voicing patterns and voice assimilation: plosives are more voiced and undergo more regressive voicing assimilation than fricatives. When following an obstruent, the voicing status of /b,d/ appears to be conditioned by the articulation place of the plosive in the three language varieties: /b/ is more voiced than /d/, which is formed further back in the mouth.\n(*) While the standard pronunciation of French does not show final devoicing, it has been demonstrated that final obstruents often devoice in Belgian French (Hambye 2005).