Abstract/Summary

The study investigates the acquisition of Setswana speech rhythm and the penultimate syllable vowel length by early sequential Setswana-English bilingual children aged 6-7 years old growing up in Botswana, a country with a diglossic setting, where English is the dominant high-status language in educational and public contexts. For this group of children, taught fulltime in English from the age of three years, the second language (L2) becomes their dominant language through exposure to English-medium education. The speech rhythm and the penultimate syllable vowel length patterns of the Setswana-English bilingual children are compared with those of age-matched Setswana monolingual children educated in public schools for whom English is a learner language. The aim was to ascertain if the prosodic patterns of the bilingual children reflect those of monolingual children or if the high-status English has an effect on these prosodic features in comparison with monolingual children. In view of the on-going debates over perceptions and production of speech prosody, it is valuable to consider monolingual and bilingual speech acquisition to determine the extent to which high exposure to L2 input contributes to foreign accent and divergent speech prosody in L1. Previous studies have reported inconsistent results regarding the rhythmic pattern of bilingual children of 5 years of age and older. While other studies demonstrated that this group of bilingual children keep the rhythmic pattern of their two languages separate (e.g., Bunta & Ingram, 2007) - i.e., they maintain first/second language-specific syllabic stress or prosody patterns during parallel or sequential acquisition of the two languages - others have shown an interaction of the two languages (Kehoe, 2002; Mok, 2011; Whitworth, 2002). The research presented in this thesis tests these claims. The study primarily uses spontaneous speech from twenty participants based on the telling of the wordless picture story Frog where are you? (Mayer (1969). Praat (Boersma & Weenink, 2002) was used to generate waveforms and spectrograms where sound files were segmentally labelled into syllables and vowels. A Praat script was also used to extract the duration of the vowels. The nPVI-V and the Varco V rhythm metrics were utilised to examine the speech rhythm of the children. The results showed that the bilingual group’s L1 prosodic pattern diverged from that of the non-bilingual group. The evidence in this population, of evident transfer effects from English bilingualism on L1 Setswana speech prosody, challenges the assumption that speech prosody is established early in life, especially when the language is a less marked, syllable-timed language like Setswana.