Galts, T. , Braun, B. & Kabak, B.
Department of Linguistics, University of Konstanz, Germany
This paper investigates the influence of existing native language (L1) suprasegmental contrasts on the lexical encoding of novel contrasts. While previous studies mostly focused on novel segmental contrasts (Hayes-Harb & Masuda, 2008), here we expand this line of research to the L2 acquisition of suprasegmentals and explore the extent to which the stress typology of the L1 may lead to differences in the lexical encoding of novel tonal contrasts. In a cross-modal perception study, native speakers of Russian, German (languages with lexical stress), French (with no lexical stress) and Mandarin Chinese controls (n=32 total), were tested on their ability to lexically encode four Mandarin tones on segmentally distinct sets of disyllabic non-words. The participants were asked to decide if a given auditory word learned in a previous training phase matches the visual stimulus in three different conditions: CompleteMatch, SegmentalMatch-TonalMismatch, and TonalMatch-SegmentalMismatch. While in the CompleteMatch condition, we found no differences among the groups, there was a main effect of language in both accuracy and RTs for the other two conditions. In particular, Russian and German participants were equally able to distinguish tonal minimal pairs with higher accuracy than the French (French: 1.3% vs. Russian: 18.8%, German: 30.5%, p<0.001). All three groups were however significantly worse than the Mandarin controls (86.7%, p<0.001). Surprisingly, Mandarin Chinese controls did significantly worse than other groups in the TonalMatch-SegmentalMismatch condition (p<0.01 for all comparisons), suggesting the primacy of tonal information over segmental information in tone languages (cf. Ye and Connine, 1999; Lee, 2007; Braun and Johnson, 2010). Overall, we argue that learners with lexical stress in their L1 are more efficient in lexicalizing other types of suprasegmental information, i.e., lexical tones, than those with no lexical stress.