<p>Cross-language similarity in form and meaning between the first language (L1) and the second language (L2) benefits bilingual language processing. The facilitative role of congruency is well-established in bilingual lexical (single word) processing, such as cognate facilitation effects. Recent research has suggested that the processing advantage afforded by congruency also extends to units beyond the word level, to multi-word expressions (MWEs). Congruent L2 MWEs that have an equivalent form in the L1 are likely to be processed faster and more accurately than incongruent L2-only MWEs that have no equivalent in the L1. However, it is still unclear what mechanisms underpin the congruency effect in the processing of L2 MWEs, and whether congruency can also affect L1 MWE processing. To further understand cross-language influences in the processing of MWEs, the present study investigated cross-language influences in the processing of binomial expressions (knife and fork) in the L1→L2 and L2→L1 directions. Two groups of unbalanced bilinguals (Chinese-English and English-Chinese) and a control group of English monolinguals performed a visual lexical decision task that incorporated unmasked priming. To assess cross-language influences, I used three types of expressions: congruent binomials (English binomials that have translation equivalents in Chinese), English-only binomials, and Chinese-only binomials translated into English. Lexical decision latencies to the last word (fork) in a binomial (knife and fork) were compared with response latencies to the same word in a matched control phrase (spoon and fork). I found that (1) Chinese-English bilinguals in Experiment 1a showed a significant priming effect for congruent binomials but no facilitation for English-only binomials, (2) English monolingual controls in Experiment 1b showed comparable priming for congruent and English-only binomials, (3) English-Chinese bilinguals in Experiment 2 showed a trend toward priming for congruent binomials, which did not reach statistical significance, and no priming for English-only binomials. None of the three participant groups showed priming for translated Chinese-only binomial over controls. These findings suggest that L1 influences the processing of L2 binomials, and that there may be some cross-linguistic influence in the opposite direction, i.e., from L2 to L1, although to a lesser extent. For the Chinese-English bilinguals, the facilitation for congruent binomials is probably not due to the automatic activation of L1 translation equivalents of the L2 binomials. I conclude that exposure to binomial phrases in the L2 is needed for the congruency effect to occur. For the English-Chinese bilinguals, they showed no priming for English-only binomials. I propose that their L1 may be inhibited in L2 learning and immersion contexts and switching back to L1 may come at a cost. However, the English-Chinese bilinguals showed a weak priming in the processing of congruent binomials, providing initial evidence that crosslinguistic influence can occur from the non-dominant L2 to the dominant L1, even in an entirely within-L1 task.</p>