Unmasking translation priming with masked priming: The role of language direction, task, prime duration, individual differences, and word frequency

Abstract The structure of the bilingual mind has been the focus of many studies, especially regarding how different languages are stored and interact. Using the same sample pool of 101 Turkish–English bilinguals across eight experiments involving the masked priming paradigm, this extensive study investigated non-cognate translation priming, along with the potential effects of language direction (L1-to-L2 vs. L2-to-L1), task type (lexical decision task vs. semantic categorization task), prime duration (50 ms vs. 83 ms), individual differences (L2 proficiency, L2 immersion, and L2 dominance), and word frequency. The results revealed a prominent translation effect for both language directions across tasks and prime durations, with semantic categorization tasks eliciting more priming. Since 83-ms primes were largely visible, the findings reveal important differences between pre-conscious (50-ms primes) and conscious (83-ms primes) word processing. Overall, the findings support the quantitative models, Multilink in particular, highlighting that native-like L2 word processing can emerge in late bilinguals.