Abstract/Summary

Bilingual lexical access models assume that semantic representations activate the lexical nodes simultaneously in both languages. The question is whether the activated lexical nodes from both languages compete for selection or not. Three hypotheses have been proposed to test the competitive lexical selection mechanism: language-specific lexical selection – which argues that competition is restricted to within language lexical nodes. Language non-specific lexical selection- supportive of cross-language competition and the dynamic view- is a combination of language-specific and language-non-specific views. This dynamic view is based on the language context hypothesis (Grosjean, 2001), which states that bilinguals can represent themselves anywhere on the continuum, depending on the activation level of non-target language. Bilingual in a monolingual context may lead to language-specific selection due to activation of the target language to greater levels, and when in a bilingual context, it can lead to language non-specific due to activation of both the languages and non-target language activation exceeding that of a monolingual context. Previous studies on bilingual lexical selection have shown lexical selection to be language-non-specific in bilinguals and language-specific at times. However, most of the studies have been conducted in highly proficient bilinguals speaking closely related languages, which has resulted in cross-language interference effects. These observed effects may be due to high proficiency and proximity between the languages. On the other hand, studies which investigated bilingual lexical selection in distant language bilinguals have yielded inconsistent and inconclusive results. In addition, they have been investigated in highly proficient bilinguals, and only a handful of studies have been carried out on different L2 proficiency levels. Moreover, previous studies have mainly studied lexical selection in one language direction, i.e., naming in L2 with L1 or L2 distractors. The present thesis investigates bilingual lexical selection in moderately proficient distant language Kannada-English bilinguals in both language directions. In addition, multiple factors have been shown to influence lexical selection in bilinguals. Specifically, proficiency, language dominance, language use and inhibitory control factors have been shown to modulate bilingual lexical selection. It is assumed that one way to resolve the competition from the activated lexical nodes in both languages is through interference suppression using inhibitory control mechanisms. The second aim of this study was to investigate the relationship between bilingual (proficiency and dominance) and inhibitory control factors and their influence on the lexical selection process. A Kannada picture and word stimuli set was developed in the first study; in the second study, four picture-word interference paradigms were developed with manipulations of distractor type, distractor language and language context to investigate the three lexical selection hypotheses, which included naming the pictures in Kannada with Kannada and English distractors and naming in English with English and Kannada distractors. A within-participant design was used with moderately proficient sequential bilinguals to investigate lexical selection in both language directions within the same bilinguals. The results showed an absence of distractor effects and the presence of distractor language and language context effects. Interaction between the distractor language and language context with bilingual factors and inhibitory control factors on the speed of lexical retrieval were observed. The results from the current thesis suggest that lexical retrieval is dynamic in that distractor language and language context effects are observed, which interact with bilingual and inhibitory factors.