Abstract/Summary

Word production is an essential feature of successful communication where semantic information (meaning) of a word is activated first, and this representation then activates the corresponding phonological form followed by the articulation of the target word. However, the production of words becomes effortful and impaired following neurological impairments (e.g., aphasia). The nature of word production impairments in aphasia is poorly understood and inadequately treated. In healthy monolingual speakers, word production involves selection of target word from competing lexical items within the target language. The situation becomes complicated for bilinguals with two sets of lexical systems leading to enhanced lexical competition. Research has shown different executive control processes are involved while resolving lexical competition. There is currently not a consensus in the literature as to whether this lexical competition is resolved in the same way by monolingual and bilingual speakers. Moreover, research on the nature of word production deficits in bilinguals with aphasia (BWA) and their relationship to executive control mechanism is not established in the literature, especially in Indian languages. In this project, we investigate the relationship between word production and executive control in a systematic and stepwise exploration in two phases (Phase I in UK and Phase II in India) by using different participant groups, wide range of linguistic measures, and separate executive control measures. Participants in Phase I were 25 healthy Bengali-English bilinguals and English monolinguals who were matched on age, gender, years of education, non-verbal intelligence, and vocabulary. Participants completed two linguistic experimental tasks (verbal fluency in Chapter 2 and blockedcyclic naming in Chapter 3) in English and three executive control tasks (inhibitory control: Stroop task, mental-set shifting: colour-shape switch task, working memory: backward digit span test). Results revealed bilinguals performed at par with the monolinguals in some linguistic measures (semantic fluency and heterogenous context in blocked-cyclic naming) and outperformed monolinguals in certain linguistic measures (e.g. letter fluency and homogenous context in blockedcyclic naming). Therefore, bilingual disadvantage in the linguistic domain can be negated if vocabulary is controlled for. Also, bilingual advantage in the non-linguistic domain can be extended to the linguistic domain if the linguistic tasks were made more challenging by increasing the executive control demands. Bilingual Participants in Phase II were eight Bengali-English BWA and eight Bengali-English bilingual healthy adults (BHA) who were matched on age, gender, years of education, and measures of bilingualism. Participants completed two linguistic tasks (verbal fluency in Chapter 4 and picture naming involving cognates and non-cognates in Chapter 5) in both languages and same executive control tasks as in Phase I, except for mental-set shifting (Trail Making Test). As expected, we found evidence of linguistic and executive control impairments at the group level for BWA individuals. Similar to the monolingual group in Phase I, we found BWA had more difficulty in the linguistic measures (e.g. fluency difference score, number of switches, between-cluster pause) where executive control demands were higher. However, the underlying executive control deficits in the linguistic tasks may not be visible with the usual analysis techniques. Therefore, we argued in favor of including a more fine-grained analysis of linguistic tasks. In terms of cross-linguistic impairment following a stroke, our results showed similar impairment in both the languages and the post-stroke language ability (e.g. better performance in Bengali) mirrored their pre-stroke language ability (Bengali dominant). Findings from the linguistic tasks revealed that despite showing deficits in lexical access, BWA still mirrored the BHA in terms of the underlying language processing mechanism which is required to perform in the linguistic tasks. In summary, performance differences on the linguistic measures were mediated by various factors such as the participant groups, nature of the task, language proficiency, and executive control abilities. We emphasise the importance of characterising the BWA participants in terms of their linguistic impairments in both languages, bilingualism related variables, separate linguistic and executive control measures and involved analysis approaches. We provide a foundation for future research on understanding the interplay of linguistic and executive control processes during word production in healthy bilinguals as well as in BWA population.