A neural network simulation of event-related potentials: Bilinguals’ response to syntactic violations differing in cross-language similarity
Keywords
Loading...
Authors
Issue Date
2024-02-01
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Event-related potentials (ERPs) are used to study how language is
processed in the brain, including differences between native (L1) and
second-language (L2) processing. One specific use of ERPs is measuring
native-like processing in bilinguals: when the same ERP appears for the
same condition in L2 as in L1, it is considered native-like. A P600 effect
can be measured in proficient L2 learners in response to an L2 syntac tic violation, indicating native-like processing. Cross-language similarity
seems to be a factor that modulates P600 effect size. This manifests in
a reduced P600 effect in response to a syntactic feature in the L2, where
its syntactic constructions is different from the syntactic construction in
the L1. The precise functional interpretation of ERPs remains a matter
of debate. Fitz and Chang (2019) proposed a theory where ERPs re flect learning signals that arise from mismatches in predictive processing.
These signals are propagated across the language system to make future
predictions more accurate. We test if this theory can account for the men tioned P600 effect reduction in late bilinguals, by implementing a model
capable of simulating the P600. We perform an experiment containing
three types of syntactic constructions differing in cross-language similar ity, designed to elicit a P600 effect in simulated L2 learners progress ing through learning stages. Simulated English-Spanish participants dis played a P600 when encountering constructions similar in cross-language
similarity. Conversely, simulated English-Spanish participants displayed a
reduced P600 when encountering constructions different in cross-language
similarity. The difference between these ERP responses of our simulated
participants is similar to the difference between ERP responses of par ticipants in human ERP studies. Simulated participants did not however
show a clear P600 in response to constructions that were unique to the L2,
which is due to the model not being sensitive to the specific construction
used, namely a violation in grammatical gender. Our findings partially
further support the viability of error propagation as an account of ERP
effects, and brought an inability of our model to light.
Keywords: Event-related potential; cross-language similarity; P600;
prediction error; bilingualism.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen