ABSTRACT
Event-related potentials (ERPs) have become widespread in second language acquisition (SLA) research and a growing body of literature has been produced in recent years. We surveyed 61 SLA papers that use ERPs to study L2 sentence processing in healthy late learners. Our main aim was to provide a critical summary of findings from the decade 2010-2020. The qualitative review reveals that proficiency plays a major role in determining ERP components, but its effect is modulated by language similarity and individual differences. The statistical analysis (a multinomial logistic regression) suggests that ERP components are uniquely predicted by learners’ proficiency level and the linguistic phenomenon at issue, while no effect of language distance is found. We also made a cursive methodological overview, which evidences several gaps in the literature and raises some concerns on the way proficiency is factorized across studies.
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Supplementraty Information
Supplemental data for this article can be accessed on the publisher’s website and at https://osf.io/94k6w/?view_only=03e1cdffd74146f9844a9880a4c3cf59.
Supplemental data
Supplemental data for this article can be accessed online at https://doi.org/10.1080/10489223.2022.2141633
Notes
1 Latency is defined by three timepoints, i.e., onset, peak, and offset, corresponding to the emergence, maximal amplitude, and fading of the component, respectively. Amplitude is the differential potential (measured in volts (V)) between each (active) electrode and the reference electrode. Scalp distribution is the area on the scalp where the potential is most reliably detected by the electrode(s) (Kappenman & Luck Citation2011).
2 Following Steinhauer & Drury (Citation2012), the spill-over effect obtains when the effects of a pre-target word prolong after the target word, whereas the offset effect is produced by a noisy baseline in which the experimental condition is more positive than in the control condition, thus resulting in a polarity shift (a sustained negativity) after baseline correction.
4 Note that a similar positive shift is also observed at clefted nouns in Reichle & Birdsong (Citation2014).
7 Starred (*) papers are part of the qualitative review but are excluded from the distributional and statistical analyses.