The Pro2F project (Pustka, Vienna) will provide the fundamental research needed to improve the pronunciation skills of learners of French as a foreign language. For this purpose, a new corpus with language data from 100 Viennese pupils will be built, the first corpus of learner French in Austria. The speakers range from 0 to B1 (CEFR).

Our analysis of the corpus focuses on two essential phenomena in French phonology: schwa and liaison. Both consist of alternations: in the case of schwa, a vowel/zero alternation (e.g. s(e)ra [səʁa]~[sʁa]), and in the case of liaison, a consonant/zero alternation (e.g. les [le], amis [ami], but les amis [lezami]). In each case, the affected segment has an equivalent in orthography (in this case <e> and <s> respectively), which is usually the starting point in foreign language learning; however, this segment can remain unrealised.

The resulting discrepancy between the pronunciation and the orthography causes learners to feel almost ʻshockedʼ when exposed to the target language in an authentic environment: being incapable of recognizing words, they have considerable difficulties understanding. In the face of these perception problems, the production of schwa and liaison could be fairly easy to learn: As opposed to segmental phonology, phonetics and prosody in which interferences from the L1 will only disappear after a lot of training, schwa and liaison can essentially be taught through grammar and vocabulary, which makes it easy to integrate them into foreign language teaching.

Aside from the correct realization of categorical liaison by beginners, the socio-pragmatic conditioning of variable liaison may challenge advanced learners. Our analysis will demonstrate to what extent this variation is implicitly acquired on the basis of the limited input in the classroom and to what extent larger exposure to authentic material and explicit instruction would be beneficial.

Apart from its descriptive results, the project also contributes to the theoretical discussion in phonology. The data from our learners will be compared to L1-speakers in order to determine if the cognitive representations and processes are globally the same or quite different and if differences exist between the pathways of learning and acquisition. These findings will contribute to a better understanding of the role of speakers of different ages in language change.

In order to allow the results of our research to flow back to the learners, the project will be accompanied by several collaborations: Namely, we will provide individual feedback for the participants of the study as well as a project website for the general public. Since pronunciation training has, for a long time, been very unpopular in language classes, we are, therefore, organizing university courses and teacher development courses in order to increase (future) teacher awareness of the importance of pronunciation and provide them with new approaches to reduce foreign accent.