Project Summary Bilinguals have remarkable control abilities to prevent interference between languages in production and comprehension, two modalities that have different goals and go through different pathways. However, it still remains unclear if and to what extent language control mechanisms are shared or different across these two modalities. The proposed project will systematically compare language control mechanism in production versus comprehension at different processing levels. We will combine ERP (Event-Related Potential in the brain) measures and the mixed-language paragraph reading paradigm to reveal the neuro-cognitive processes of language switching in sentence context, modulating syntactic and phonological constraints. Bilinguals will read the same mixed-language paragraphs in production and comprehension tasks, with the only difference being if articulation is required. ERP measures will provide detailed information about how language control works as the language-switching event unfolds in real time. My Specific Aims are: Aim 1 (K99): to investigate the role of syntactic constraints in language control during production vs. comprehension, by measuring bilinguals’ ERPs on language switching with content vs. function words, which carry more semantic vs. syntactic information, respectively. In production, I expect larger costs switching function words, as these need more inhibition and monitoring, revealed by N2 and ERN respectively. If syntactic constraints equally affect production and comprehension, function switches should be more costly than content switches also in comprehension, revealed by P600/LPC, or even N400 as well. Alternatively, content switches might be more costly in comprehension, revealed by P600/LPC and N400, due to vague syntactic analysis and the priority to integrate semantic information. Aim 2 (R00): to investigate how phonological constraints influence language control during production vs. comprehension, by measuring bilinguals’ ERPs on intra-sentential language switches with cognates vs. noncognates. Cognates are words with cross-language phonological overlap (e.g., coffee-café). I expect larger costs with cognate than noncognate switches in production, as more inhibition and monitoring are needed for cognates, revealed by N2 and ERN components respectively. For comprehension, if language selection is also critical, switch costs with cognates may be larger than noncognates, revealed by N400 and/or P600/LPC. Otherwise, switch costs may be smaller with cognates than noncognates due to easier semantic integration with cognates, revealed at least by N400, or even by P600/LPC as well. My training in the K99 phase will include structured mentorship by my advisory committee on both theoretical framework (language production and comprehension, as well as bilingual language control) and the training of new methodology (ERP measures). It will also include formal course work, workshops, and a program of career transi...