Abstract
A methodological shortcoming in previous second language (L2) acquisition studies has been that researchers have assumed an overly homogenous first language (L1) ignoring dialect differences. In the current study English and German vowel production data were collected from 72 English-speaking learners of German from three distinct North American English dialect regions – the Inland North, North Central, and Western Canada. Following Flege (e.g., 1995), who proposed that L2 segments with L1 counterparts would be more difficult to perceive and produce than new L2 segments, we show that subjects did not transfer their L1 /u/ to German but rather produced the German counterpart in a manner expected in neither German nor English. Instead, this was reflected in terms of formant (F1, F2 or F3) values that varied according to the L1 dialect of the learner. In particular, learners from the North Central dialect region whose English /u/ was produced with the lowest F2 values – though not significantly different from the Inland North learners' /u/ – produced the German /u:/ with the highest F2 values of all three dialect regions. Speakers from all dialect regions were also able to manipulate their acoustic space to allow for the addition of the new German segment /y:/; however, they differed in how they ultimately established the German /u:/–/y:/ contrast. Learners from the two American dialect regions contrasted these vowels according to F2 values, while learners from Western Canada made the contrast utilizing F3. Based on these results, we conclude that L2 vowel formant values differ by dialect region even when the learners' L1 dialects differ only subtly. Lastly, results provide further evidence that this influence is not simply the result of direct L1 transfer.
©Walter de Gruyter
