|Title:||The Neurocognition of Adult Second Language Learning: An fMRI Study||Add Dissertation|
|Author:||Kaitlyn Tagarelli||Update Dissertation|
|Email:||click here to access email|
|Institution:||Georgetown University, Department of Linguistics|
|Linguistic Subfield(s):||Applied Linguistics; Psycholinguistics; Neurolinguistics; Cognitive Science; Language Acquisition;|
|Abstract:||This study investigates the neural structures and cognitive processes involved in adult second language (L2) learning, and whether and how they change as a function of increasing exposure and proficiency. By combining behavioral and functional magnetic resonance imaging (fMRI) measures, this study aims to address fundamental questions in Second Language Acquisition research that cannot be fully explored with behavioral methods alone.
Research to date suggests that the mechanisms underlying L2 may change during acquisition, though in different ways for lexical/semantics and grammar. However, it is logistically impossible to longitudinally investigate the course of learning a natural language from initial exposure to advanced proficiency. This has left major gaps in this research. Studies on the learning of artificial linguistic systems (e.g., artificial grammars) have begun to address this issue, but their generalizability to natural languages has been questioned. The current study aims to bridge the gap between artificial linguistic systems and natural languages by longitudinally examining the learning of a reduced natural language, or “mini-language.”
Fifteen native speakers of English were trained on a subset of Basque, from initial exposure to high proficiency. Behavioral and fMRI measures were continuously acquired during all grammar training.
Learners achieved very high proficiency in vocabulary and reasonably high proficiency in grammar, though morphosyntactic agreement was difficult to master. FMRI activation was found in areas associated with first language (L1) processing (e.g., BA45/47, and parietal cortex for lexical/semantics, and BA44 and 6 for grammar), suggesting that late-L2 learners have access to L1 regions. Additional areas were engaged, suggesting that L1 mechanisms are not sufficient for L2 learning and processing. At early stages of learning, hippocampal activation was found for both vocabulary and grammar. At later stages, basal ganglia activation was observed for grammar, particularly in the caudate nucleus. The findings suggest that early word and grammar learning relies on declarative memory (and more explicit processes), but that grammar later relies on procedural memory (and more implicit processes). These results highlight the utility of a mini-language model, have implications for neurocognitive theories of L2, and demonstrate the importance of integrating neural and behavioral methods in L2 research.