Slide Slam G4
Non-native noun phrase production: An ERP study on the role of language similarity
Sarah von Grebmer zu Wolfsthurn1,2, Leticia Pablos1,2, Niels O. Schiller1,2; 1Leiden University, 2Leiden Institute for Brain and Cognition
Language similarity is a critical component in multilingual language acquisition and processing. For example, in native speakers or high proficient learners, languages from the same family have shown comparable brain activation patterns and faster response times compared to languages from different language families. In lower proficient multilinguals, these neural and behavioural effects were emphasized. At the root of language similarity effects is cross-linguistic influence (CLI), the interaction of the languages within a multilingual system. Here, the question arises whether CLI effects are more pronounced in linguistically similar vs. dissimilar language pairs in late language learners. In the current study, we explored CLI effects via the gender congruency effect and the cognate facilitation effect in two groups of late language learners of Spanish whose native language belonged to the same (Italian) or different (German) language family. We placed a special emphasis on the neural correlates of language similarity effects. More specifically, we explored the P300 effect as an index for conflict monitoring and inhibitory control, and the N400 as an index for language co-activation and CLI effects. We investigated naming latencies and EEG signal modulation during non-native noun phrase production (determiner + noun [la flor] “the flower”) in a “linguistically similar” group of thirty-three Italian late learners of Spanish and a “linguistically dissimilar” group of thirty-three German late learners of Spanish using an overt picture-naming task. First, we predicted faster naming latencies for congruent and cognate nouns compared to incongruent and non-cognate nouns across both groups, reflecting the interaction of the syntactic and phonological systems of the two languages. Next, we also predicted a P300 effect and an N400 effect to reflect the cognitive mechanisms underlying the mitigation of CLI effects. More specifically, for both groups we predicted less positive P300 amplitudes and less negative N400 amplitudes for congruent and cognate nouns compared to incongruent and non-cognate nouns. Second, for an effect of language similarity, we predicted larger CLI effects for the linguistically similar Italian-Spanish group compared to the German-Spanish group. Accordingly, at the neural level, we predicted larger P300 and N400 effects for the Italian-Spanish group compared to the German-Spanish group. Across both groups, we observed a gender-congruency effect and a cognate facilitation effect in the naming latencies. Language similarity, however, did not appear to influence naming latencies. In EEG terms, we found evidence for a P300 effect, but not an N400 effect in both groups. However, while P300 amplitudes were modulated by gender congruency and cognate status for the German-Spanish group, this was not the case for the Italian-Spanish speakers. This suggests an effect of language similarity reflected by larger P300 effects for the German-Spanish group compared to the Italian-Spanish group. Overall, our results suggest traceable CLI effects in late learners. Moreover, we found distinct neural signatures of mitigating CLI effects on the basis of language similarity. Our study has important theoretical implications for the role of linguistic similarity in non-native language production and for our understanding of non-native acquisition and production models in late language learners.