Bilingual language contexts : variable language switching costs and phonetic production

Bilinguals are generally adept at segregating their two competing languages and switching between them when contextually appropriate, although it has been shown that switching languages incurs a reaction time delay, or switch cost (Kolers, 1966). These switch costs are modulated by language dominance, with bilinguals evidencing greater delays when switching into their dominant language relative to their non-dominant language (e.g. Meuter & Allport, 1999). While these asymmetrical switch costs have formed the basis for theories of bilingual language separation and selection, the key factor of language context, the degree to which each language is employed in a given paradigm or conversation, has yet to be considered. In addition, previous research and subsequent theories of language selection have focused exclusively on the lexical level, yet given the distinct phonetic categories in a bilingual’s two languages (Caramazza et al., 1973), selection must also occur at the phonetic level.

Addressing these gaps in the literature, this dissertation investigates the language switching costs and phonetic production of Spanish-English bilinguals in two experimental paradigms: a cued picture-naming task and an oral production task. In both studies, bilinguals (English-dominant, Spanish-dominant, and balanced bilinguals) produced language switches in varying language contexts, from monolingual to bilingual. Analyses focus on switch costs, error rates, and phonetic production, as a means to further the understanding of the language switching mechanism at the lexical and phonetic levels.

Drawing on results from the two experimental paradigms, this dissertation makes several major contributions to the ongoing discussion regarding bilingual language selection. First, this study provides evidence for a gradient nature of the language switching mechanism at the lexical level. Second, it contributes an examination of the effects of language switching at the phonetic level, demonstrating asymmetrical phonetic transfer. And third, parallels are drawn between the underlying effects of language switching and the phonetic realizations produced in connected speech. Implications are considered for theories of bilingual language selection, and a gradient account of the Inhibitory Control Model (Green, 1986) is proposed at both the lexical and phonetic levels.