The time cost of mixed-language processing: an investigation

Does it take bilinguals longer to process mixed-language information? This study explores, in two reaction time experiments, the hypothesis that there is only a cost to language switching when the switch is unexpected in the context. Prior to the experiments, an on-line language test and a linguistic background questionnaire were employed to select the bilingual participants. In experiment 1, the subjects, who were Mandarin Chinese-Taiwanese bilinguals, were led to think that Mandarin Chinese was the contextually appropriate language, and were slower to respond on a language switch trial. In experiment 2, the participants were led to think that both languages were contextually appropriate, and were not slower to respond on a language switch trial. The data support Grosjean's language mode hypothesis on the mixed-language processing cost, and show that the cost is a result of context rather than an inevitable consequence of a switch.