Second Language Acquisition Laboratory, Department of Linguistics, University of Kansas.
Centre for Research on Bilingualism, Department of Swedish Language and Multilingualism, Stockholm University.
J Exp Psychol Learn Mem Cogn. 2021 Jul;47(7):1106-1140. doi: 10.1037/xlm0000983. Epub 2021 Jan 28.
The present study examines both properties of the language and properties of the learner to better understand variability at the earliest stages of second language (L2) acquisition. We used event-related potentials, an oral production task, and a battery of individual differences measures to examine the processing of number and gender agreement in two groups of low-proficiency English-speaking learners of Spanish who were tested in multiple sessions. The results showed an advantage for number, the feature also instantiated in the native language, as both groups showed a native-like P600 response to subject-verb and noun-adjective number violations across sessions. The more advanced group showed larger effects for number and marginal sensitivity to gender violations. These results suggest that native-like processing of shared features is possible even for novice learners, contrary to proposals suggesting that all morphosyntactic dependencies are initially processed in a non-native manner. Working memory (WM) was a predictor of P600 effects for number and also for gender (where the effect was marginal), suggesting that similar abilities may capture variability in the processing of both shared and unique features despite differences in overall sensitivity. Furthermore, whereas WM predicted performance on online tasks (P600 effects/oral production), verbal aptitude predicted performance on tasks examining morphosyntactic accuracy (grammaticality judgment task/oral production). Our results show that the linguistic properties of the L2, the individual characteristics of the learner, and the nature of the task at hand all play an important role in capturing the variability often observed in the L2 processing of agreement. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
本研究通过考察语言特征和学习者特征，以期在第二语言（L2）习得的早期阶段更好地理解变异性。我们使用事件相关电位（ERP）、口语产出任务和一系列个体差异测量来检验两组低水平西班牙语英语学习者对数字和性别一致性的处理，这些学习者在多次测试中接受了测试。结果表明，在两个受试组中，数词特征在母语中也有体现，数词的处理存在优势，表现出类似于母语者的 P600 反应，即对主语-动词和名词-形容词的数词一致性错误有反应。高水平组在数词上的效应更大，对性别一致性错误的敏感性略高。这些结果表明，即使是新手学习者，也有可能对共享特征进行类似母语的处理，这与一些观点相反，后者认为所有形态句法依存关系最初都是以非母语的方式进行处理的。工作记忆（WM）是 P600 效应（数词）和性别（有边缘效应）的预测因子，这表明尽管整体敏感性存在差异，但类似的能力可能会捕捉到共享和独特特征处理中的变异性。此外，虽然 WM 预测了在线任务（P600 效应/口语产出）的表现，但言语能力预测了形态句法准确性任务（语法判断任务/口语产出）的表现。我们的研究结果表明，L2 的语言特征、学习者的个体特征以及手头任务的性质都在捕捉一致性的 L2 处理中经常观察到的变异性方面起着重要作用。（PsycInfo 数据库记录（c）2021 APA，保留所有权利）。
