Zentrum für Mehrsprachigkeit

Jasmijn Bosch (University Milano Bicocca): Predictive processing and cross-linguistic influence in bilingual children

The Multilingual Mind: lecture series on multilingualism across disciplines

12.1.2021

Tuesday, 17.00 - 18.30 (CET/UTC+01)

Zoom room: zoom.us/j/94531600895

Abstract

Listeners process speech rapidly and incrementally, and monolingual children are known to use morphosyntactic cues to anticipate upcoming words from a very young age. Using a visual world eye-tracking paradigm, the present study investigated linguistic prediction in bilingual children, whose online processing may be different even when offline comprehension is on target. Specifically, we examined whether children could anticipate upcoming nouns on the basis of gender and number agreement on the preceding article. By comparing different groups of bilingual children and monolingual Italian children, we aimed to test the effects of cross-linguistic influence and language proficiency on predictive processing.

In Experiment 1, we tested anticipation based on grammatical gender in German-Italian bilingual children (aged 6 to 9) living either in Italy or in Germany. The results showed that children processed sentences fast and efficiently by relying on predictive mechanisms. Furthermore, in an Italian task (but not in a German task), we found that children exhibited a ‘gender congruency effect’, i.e., they experienced cross-linguistic influence when the grammatical gender of the two languages did not overlap, leading to delayed anticipation. Both the efficiency of linguistic predictions and the likelihood of a gender congruency effect were related to children’s relative language proficiency. In Experiment 2 we tested anticipation based on grammatical gender and number in Mandarin-Italian bilingual children as compared to monolingual Italian children (aged 8 to 12), to investigate cross-linguistic influence in languages with greater typological distance. We found efficient prediction based on number in both groups, whereas processing of grammatical gender was significantly delayed for Mandarin-Italian bilinguals. One interpretation is that the discrepancy between gender and number was due to transfer, since Mandarin does not have grammatical gender while it does have a conceptual notion of number. Alternatively, the difference may have been caused by the fact that gender is an arbitrary property that requires lexical knowledge, while number is concretely linked to the referential context. Therefore, the arbitrariness of grammatical gender may be especially difficult for L2 speakers with lower proficiency levels, including bilingual children.

Altogether, our results suggest that bilingual children are able to efficiently process speech in their L2 by making use of predictive mechanisms, even when they have to rely on a feature which is expressed differently in their L1. However, depending on their language proficiency, they may experience processing delays related to cross-linguistic influence.