Triggered codeswitching: An experimental test of the triggering hypothesis

A hallmark of bilingual language processing is codeswitching, the use of two languages in one utterance. Though this switching between two languages seems effortless, as self-reports of bilinguals confirm, behavioral and ERP-studies report switching costs in comprehension and production studies. How does a codeswitch (CS) come about? The triggering hypothesis (Clyne, 2003) states that cognates can cause a codeswitch within a sentence. Corpus studies (Broersma & de Bot, 2006) have shown that cognates can function as a trigger for a CS. We tested whether different types of triggers can affect the perception of a CS. Both behavioral and ERP-paradigms were used. In the first study, Dutch-English bilinguals read sentences containing a CS. The sentences were either in Dutch (with English CSs) or in English (with Dutch CSs). Furthermore, the sentences contained either a cognate trigger or a non-cognate control, which was presented in the same basic clause as the CS. Example cognate trigger: The well-known supermarket also sells speelgoed to its costumers. Example non-trigger: The well-known store also sells speelgoed to its costumers. In the examples cited above, the CS is ‘speelgoed’ [toys]. The cognate trigger, supermarket [supermarkt], is a cognate with Dutch, while the non-trigger, store [winkel], is not. In the second study, we extended the triggering hypothesis to see whether a socio-contextual trigger would affect the processing of a CS. Participants first read a sentence with a socio-contextual trigger (congruent with the language of the CS) or a socio-contextual non-trigger (congruent with the language of the sentence). Example sentences in the English-Dutch condition are below. Example socio-contextual trigger: For your daily groceries you can shop at Albert Heijn. This well-known store also sells speelgoed to its costumers. Example non-trigger: For your daily groceries you can shop at Wal-Mart. This well-known store also sells speelgoed to its costumers. Again, the CS is ‘speelgoed’ [toys]. The name of a Dutch supermarket (Albert Heijn) is a socio-contextual trigger for the Dutch CS ‘speelgoed’ [toys]. The American supermarket (Wal-Mart) is a non-trigger. We found that CSs were read significantly faster when they were preceded by a cognate trigger than when preceded by a non-trigger, but only when switching from L2 to L1. Socio-contextual triggering showed a marginally significant effect on the reading time of the CS for the behavioral studies, and a small modulation on the N400-component related to reading the codeswitch in the ERP-study, but only when switching from L1 to L2. In the discussion, we will relate triggered codeswitching to current (neuro)cognitive models on bilingual processing. References Broersma, M. & De Bot, K. (2006). Triggered codeswitching: A corpus-based evaluation of the original triggering hypothesis and a new alternative. Bilingualism: Language and Cognition, 9, 1-13. Clyne, M. (2003). Dynamics of language contact. Cambridge: Cambridge University Press