Lexical stress perception : a case study on Tamil-English bilinguals

Past research has shown that one’s L1 phonological system influences subsequent perception and learning of a second language, including the domain of stress. Speakers from a non–stress language background often make a considerable number of stress errors when learning an L2 with a stress system. Interestingly, this difficulty has been shown to persist in some cases with bilinguals where one of the native languages contained stress (Doupoux et al., 2010). This paper explores word stress perception in relation to Tamil–English bilinguals who have acquired English, a language with contrastive stress; and Tamil as a co–L1 without contrastive stress. It seeks to determine whether such bilinguals have difficulty in distinguishing lexical stress and if knowledge of phonological generalizations, i.e. implicit syllable structure rules, plays a role in the perception of stress. The study followed an experiment paradigm suggested by Doupoux et al. (2001) for a robust method of determining a “stress–deafness” effect. Participants included Tamil–English bilinguals (N=10) and a control group of Hindi–English bilinguals (N=10), selected as both Hindi and English contain contrastive stress. Random aural sequences containing contrastive phonemes, five bisyllabic nonword tokens per sequence, were presented aurally in a baseline task where participants were to replicate sequences. Their performance was contrasted with an experimental task of replicating aural contrastive stress sequences. Results revealed no significant differences in stress perceptions across groups. The data also did not reveal significant differences across structure types. However, a trend of stress perception difficulty among Hindi-English bilinguals was observed, and language dominance identified as an influencing factor. This unexpected finding is discussed in relation to a few theoretical models, including Cutler’s bimodality theory (1992).Bachelor of Art