Language-Switching Cost in Chinese-English Bilingual Reading Comprehension: Evidence from Eye-Tracking

Abstract Code-switching, or language-switching, which is the use of elements from two or more languages in the same utterance or in the same stretch of conversation (Paradis et al., 2011), is prevalent among bilingual and multilingual speakers. According to the parallel activation hypothesis of language comprehension, both languages of a bilingual are activated during sentence processing, even when there is only monolingual input (Spivey & Marian, 1999). However, studies have also shown that such language mixing can slow down processing (Macnamara & Kushnir, 1971; Soares & Grosjean, 1984). This cost has been attributed to either the difficulty of switching languages or the appearance of an unexpected word or wordform. This eye-tracking experiment investigated the cognitive cost brought by code-switching in Chinese-English bilinguals. Specifically, we sought to determine the source of the baseline cost of language-switching in inter-sentential code-switching. Results suggest that narrative consistency and the amount of previous information had longer-lasting effects on bilinguals’ total reading times, along with a brief disruption caused by unexpected wordforms. Taken together, the results indicate that the cognitive cost of language-switching arises more from the surprisal of encountering an unpredicted word or wordform, rather than encountering an unexpected language.