Hierarchical levels of representation in language prediction: The influence of first language acquisition in highly proficient bilinguals

Abstract

Language comprehension is largely supported by predictive mechanisms that account for the ease and speed with which communication unfolds. Both native and proficient non-native speakers can efficiently handle contextual cues to generate reliable linguistic expectations. However, the link between the variability of the linguistic background of the speaker and the hierarchical format of the representations predicted is still not clear. We here investigate whether native language exposure to typologically highly diverse languages (Spanish and Basque) affects the way early balanced bilingual speakers carry out language predictions. During Spanish sentence comprehension, participants developed predictions of words the form of which (noun ending) could be either diagnostic of grammatical gender values (transparent) or totally ambiguous (opaque). We measured electrophysiological prediction effects time-locked both to the target word and to its determiner, with the former being expected or unexpected. Event-related (N200-N400) and oscillatory activity in the low beta-band (15-17Hz) frequency channel showed that both Spanish and Basque natives optimally carry out lexical predictions independently of word transparency. Crucially, in contrast to Spanish natives, Basque natives displayed visual word form predictions for transparent words, in consistency with the relevance that noun endings (post-nominal suffixes) play in their native language. We conclude that early language exposure largely shapes prediction mechanisms, so that bilinguals reading in their second language rely on the distributional regularities that are highly relevant in their first language. More importantly, we show that individual linguistic experience hierarchically modulates the format of the predicted representation.

The power of neural oscillations to inform sentence comprehension: A linguistic perspective

“…Additionally, Li et al (2017) found significantly decreased beta power in highly compared to weakly predictive sentence contexts in the same time window. Another piece of evidence consistent with the reviewed findings comes from Molinaro et al (2017) who compared native and L2 speakers of Spanish (L2 were native speakers of Basque) in their processing of expected and unexpected sentence endings, as in I just left home and I do not remember if I closed the fem door fem (Expected)/the masc closet masc (Unexpected) when I left. Critically, since in Spanish, all nouns and determiners are marked for grammatical gender, if a specific prediction is made about the upcoming noun and if the presented determiner mismatches in gender with that noun, it must be clear already at this early stage-at the determiner-that the prediction was violated.…”

Reanalyzing language expectations: Native language knowledge modulates the sensitivity to intervening cues during anticipatory processing

“…In addition, the anticipation effects that we found are observable independently of the kind of experimental setting, supporting the idea that expectation-related top-down mechanisms are able to quickly and flexibly adjust to the different tasks a language user has to go through. Furthermore, in the present study, the target words were not visually presented (as in Molinaro et al, 2017) but auditorily, providing critical evidence that this type of predictive processing is modality independent.…”

“…2 It is interesting to note that the results from the WordMT emerged using a completely different task, as here we did not use highly constraining contexts in order to elicit topdown language processing, as in the previous experiment, but instead we used a WordMT where the expectation was evoked by a single word on a screen. This is evidence that the kind of representation preactivated in both Molinaro et al (2017) and in the WordMT is strictly word related and does not reflect language-independent semantic knowledge. In addition, the anticipation effects that we found are observable independently of the kind of experimental setting, supporting the idea that expectation-related top-down mechanisms are able to quickly and flexibly adjust to the different tasks a language user has to go through.…”

“…In Molinaro et al (2017), not only did expected and unexpected nouns differ in gender, but transparency was also taken into account. Participants read sentences that could end with a transparent or an opaque noun, such as Acabo de salir de la casa y no recuerdo si he cerrado la puerta/el armario ("I just got out of the house, and I don't remember whether I closed the [ ERPs recorded on the determiners preceding the unexpected nouns (in comparison with the ones preceding the expected) revealed an ERP negative effect starting 300 ms in both groups (identified as an N400 ERP component; Kutas & Hillyard, 1984), independently of the transparency of the preactivated target noun.…”

“…Given the specifics of the PictureMT, we decided to approach and discuss its results, taking advantage of the parallel flow of processing that has been suggested to take place for language production and prediction (Dell & Chang, 2014;Federmeier, 2007;Pickering & Garrod, 2007). In our experiment, participants initially perceive the image stimuli and, to perform the task (and ultimately preactivate the 2 It is interesting that the effect in the reading experiment (Molinaro et al, 2017) was more negative for the mismatching determiner, while in this listening paradigm it is more positive. In our opinion, this is due to the large difference between the two experimental paradigms and deserves more attention in future research.…”