Constructing A Discourse Model Across Modalities In Bilinguals

The present study aimed to explore whether discourse integration is affected by differences in modalities, and if the magnitude of the effect is modulated by language dominance for bilingual readers. Spanish-English bilinguals read and listened to pairs of passages describing two fake science facts, one of which was repeated across passages and the other revised in the second passage. The pair of passages were either in the dominant or less dominant language. The modality (audio or text) of each member of the passage pair was manipulated across four within-subject conditions. After the second passage of each pair, participants verbally recalled the content and then answered follow-up, multiple-choice comprehension questions. Overall greater recall and higher accuracy scores were observed for dominant language passage pairs relative to non-dominant language pairs. This was even more evident when the revised information was presented in the audio modality, suggesting there is an added challenge of updating a discourse model if the information is learned auditorily. There was no effect of modality on accuracy or recall for the consistent facts. These findings are discussed according to the transfer-appropriate processing and structure-building framework.