We examined how bilinguals process language switches between their first (L1) and second language (L2). Language switching costs (slower responses to language switch than nonswitch trials) appear to arise more systematically in production than in comprehension, possibly because the latter context might sometimes elicit less language coactivation (Declerck et al., 2019). This might reduce language competition and in turn the need for bilinguals to apply language control when processing language switches. Yet even in comprehension, language coactivation may vary depending on variables such as the accent of the speaker (e.g., whether the L2 words are pronounced with an L1 or L2 accent) and input modality (spoken or written). In three experiments conducted during 2021-2022, we tested how unbalanced Mandarin-English bilinguals processed language switches during comprehension and the potential influence of a speaker's accent and input modality. Overall, across settings, participants experienced significant language switching costs. In some conditions, switching costs were larger to L1-Mandarin than to L2-English, an asymmetry consistent with the participants' dominance in L1-Mandarin and the application of language control. However, manipulating accent and input modality did not influence language switches, suggesting they did not impact language coactivation sufficiently to modulate language control.
Public Significance StatementThis study investigates how bilinguals process language switches between their first and their second language. The results suggest that, across a variety of experimental settings, Mandarin-English bilinguals experience language switching costs when comprehending words. That is, being exposed to language switches induces a processing cost and increases the time needed to process incoming language input. However, processing of language switches did not seem to be modulated by variation in speaker accent (i.e., whether the words the bilinguals heard were pronounced by a native or nonnative speaker of the target language) and input modality (i.e., whether the bilinguals heard or read the words).