The aquisition of sonorant voicing.

In the view of Exemplar Theory phonetic variants of words are stored by speakers in their memory as a set of exemplars organized into clusters and grow stronger or decay over time depending on language experience (Bybee, 2002). In Polish and French sonorant consonants in clusters are devoiced word finally, which does not occur in German and English because of phonotactical constraints. A study based on automatic speech alignment and voicing profile extraction from Polish, French, German, and English speech corpuses aims at defining voicing sonorant temporal structure and voicing profile acquisition. The hypothesis is that Polish and French native speakers transfer their L1 exemplars during L2 acquisition of English and German and thus devoice sonorant. After extracting voicing profiles from recordings of native Polish and French speakers speaking English and German as an L2/L3, comparison of voicing profiles of English and German native speakers will be made in order to define the exemplar transfer.