Switching direction modulates the engagement of cognitive control in bilingual reading comprehension: An ERP study

Abstract

While switching costs in production have been explained in terms of top-down cognitive control, researchers do not agree whether switching costs in comprehension should be interpreted in the same way. Within the BIA + model, it has been claimed that the comprehension of code-switches can be explained sufficiently in terms of bottom-up activation of lexical representations. In the current electrophysiological study, L1 speakers of Dutch with high proficiency in L2 English (n = 63) completed a Flanker task in which they intermittently read sentences with or without an alternational code-switch. With this ‘conflict-adaptation’ paradigm we examined whether reading a code-switch engages cognitive control that influences performance on a subsequent Flanker trial. Half of the participants were presented with Dutch sentences and Dutch to English code-switches, while the other half were presented with English sentences and English to Dutch code-switches. On the P300 component, we found a traditional Flanker effect, with larger amplitudes for congruent than for incongruent trials. The effect was modulated by a preceding code-switch and the direction of this modulation depended on the switching direction: the Flanker effect was smaller after a code-switch from L1 to L2 than after a monolingual L1 sentence, but larger after a code-switch from L2 to L1 than after a monolingual L2 sentence. This suggests that the L1 needs to be inhibited when reading a code-switch to the L2, while inhibition needs to be released upon encountering a code-switch from L2 to L1. These results thus show that reading code-switched sentences engages a domain-general cognitive control mechanism external to the lexicon.