UC San Diego

Bilinguals have reduced verbal fluency compared to monolinguals and this has been attributed to cross language interference (Rosselli et al., 2000 ; Gollan et al., 2002). To manage interference, bilinguals may rely on executive control mechanisms to suppress the non-target language. We would therefore expect the bilingual disadvantage to increase with aging and Alzheimer's disease, which are associated with declines in executive control (Perry and Hodges, 1999). To test this account, we examined bilinguals' verbal fluency with analyses of a) number of correct responses, b) within language errors (c) cross-language intrusions and two measures of executive functioning (Stroop test and Attentional Network Task ; ANT, Fan et al., 2002). In Experiment 1a, we compared matched groups of 10 young and 10 older Spanish-English bilinguals on 18 fluency categories (5 semantic and 4 phonemic categories in Spanish and English) and we examined correlations between age and older bilinguals' (N= 18) performances on both executive control and response measures. In Experiment 1b, we compared a group of older (n=15) and young English monolinguals (n=36) on the same categories as in Experiment 1a. In Experiment 2, we compared matched groups of bilinguals with AD (n=10) with normal bilinguals (n=13), on the same categories as in Experiment 1a. Supporting an interference account, age and cross language intrusions were correlated such that older-old bilinguals produced more intrusions than younger -old bilinguals and cross-language intrusion rates were positively correlated with error rates in the ANT. Also, bilinguals with AD produced more cross language errors in semantic fluency than controls particularly in the non- dominant language. Challenges for the interference model included low rates of cross-language intrusions, even in older bilinguals and in bilinguals with AD. There was little evidence suggesting that production of a nondominant language becomes more difficult in aging and AD. We propose that executive control is important for language selection and monitoring, but after language selection, there is either (a) limited competition for selection between lexical representations across languages, or (b) a specialized mechanism for controlling competition between lexical representations that is less susceptible to cognitive decline. Thus, bilingualism is mostly maintained in aging and AD