Abstract/Summary

Bilingualism has been found to modulate executive functions because bilinguals constantly manage language selection through inhibition and conflict-monitoring (Bialystok et al., 2012). However, some studies failed to replicate these findings (Paap & Greenberg, 2013). To explain these inconsistencies, research needs to pinpoint which bilingual practices modulate which aspects of the executive system (Bak, 2016). This project is novel in that it investigated the differential impact of three code-switching types (Muysken, 2000) on executive functions: (1) Alternation of structurally independent stretches from each language, (2) Insertion of lexical items from one language into the grammar of the other, (3) Dense code-switching mixing languages at the lexical and grammatical level. Processing models of code-switching suggest that code-switching types involving greater levels of language separation (Alternation) train inhibition, whilst code-switching types maintaining high levels of linguistic co-activation (Dense code-switching) train conflict-monitoring (Treffers-Daller, 2009; Green & Wei, 2014). To test this prediction, German-English bilinguals' relative usage frequency of code-switching types was assessed and correlated with their performance in executive tasks teasing apart inhibition and conflict-monitoring (Costa et aI., 2009). Study 1 compared two groups (N=22) differing in Dense code-switching frequency. As predicted, frequent Dense code-switchers excelled in the condition challenging conflict-monitoring. Study 2 investigated individual differences in immersed adult bilinguals (N=43). In line with existing models, greater usage frequency of Dense code-switching correlated positively with performance in the condition challenging conflict monitoring. Alternation predicted inhibitory performance in the low-monitoring condition, in which bilinguals also outperformed an English monolingual baseline group (N=4I). In Study 3, bilinguals (N=29) were administered a flanker task inducing different language modes (Wu & Thierry, 2013). The Ll-dominant bilinguals performed best at inhibition in the task block inducing an L2-monolingual mode. This was explained by an increase in inhibitory activation required to suppress the dominant Ll (Meuter & Allport, 1999) and suggests that dominance patterns modulated executive functioning (Treffers-Daller, 2016). Moreover, Dense code-switching was a negative predictor of inhibitory performance in the monolingual mode, suggesting that monolingual modes require more global (Guo et al., 2011) forms of inhibition than Dense code-switching. To summarise, the results provide key novel evidence that different code-switching types modulate different aspects of cognition. Moreover, the findings support existing processing models of code-switching and highlight the importance of controlling for sociolinguistic practices in bilingualism research.