Understanding code-mixed phrases: Evidence from Chinese–English bilinguals

Aims and Objectives/Purpose/Research Questions: The cost of language switching is a matter of continuous debate. While previous research on language switching has mainly focused on manipulated switching in the lab, this study examined how bilinguals process natural code-mixed phrases to enlighten comprehension of language switching in real life. Design/Methodology/Approach: L1-dominant Chinese–English bilinguals completed two experiments. The stimuli were two-word phrases with a fixed structure: either both words were Chinese characters (monolingual phrases), or the first was Chinese and the second English (code-mixed phrases). In Experiment 1, participants performed a lexical decision task targeting lexical access to the second (English) word within code-mixed phrases. In Experiment 2, participants performed a phrase decision task, judging the acceptability of entire phrases as integral units. Both experiments contained the same four critical conditions: Chinese–English code-mixed real phrases, code-mixed non-phrases, Chinese (monolingual) real phrases, and Chinese non-phrases. Data and Analysis: We conducted linear and generalized linear mixed-effects models for the reaction time and accuracy data. We also correlated participants’ L2 proficiency/code-switching experience and their RTs. Findings/Conclusions: In Experiment 1, lexical decisions for the target words were quicker and more accurate within real phrases compared with non-phrases. Code-mixed L2 target words elicited slower, less accurate responses than their L1 counterparts. In Experiment 2, however, the code-mixed phrases, as a whole, were processed as accurately as the non-mixed phrases. Specifically, greater code-switching experience was linked to shorter response times and a tendency toward smaller switching costs in accuracy rates. Originality: This study provides original evidence that processing code-mixed collocations may not be more effortful than processing non-mixed ones, even for non-habitual code-switchers. Significance/Implications: The findings highlight the importance of ecological validity in the study of bilingual code-switching and demonstrate that everyday code-mixing experience critically shapes the processing of code-mixed expressions. Specifically, bilinguals’ habitual exposure to naturally occurring code-mixes could modulate processing costs in language switching tasks.