Introduction This study investigates the overgeneralization of noun phrase (NP) attachment rules in the L2 acquisition of Japanese sound-symbolic words, focusing on psychomime verbs that describe physical and psychological sensations. Methods An experiment using an NP attachment judgement task was conducted with 88 Chinese-speaking learners of Japanese (CSLJs) and 76 Japanese native speakers (JNSs) across NP attachment (Class 1), suppression (Class 3), and free conditions (Class 2). Results The results showed that overgeneralization was category-specific. CSLJs performed with high accuracy on Class 1 items, where both Chinese and Japanese require NP attachment, but struggled with Class 3 items, where Japanese requires NP suppression. Class 2, occupying an intermediate position between concrete bodily sensation and abstract psychological states, imposed the greatest cognitive demands and emerged as the strongest predictor of learners’ overall NP attachment preference. This overgeneralization was shaped by both L1 transfer and L2 grammatical proficiency. Learners initially relied on L1-derived processing strategies due to properties of Mandarin, yet this reliance diminished with increasing proficiency, indicating that L1 transfer is a dynamic rather than static constraint. A threshold effect was observed, with only high-proficiency learners able to suppress the overgeneralized NP attachment rule and perform in a manner similar to JNSs, whereas middle- and low-proficiency learners continued to overgeneralize. Proficiency, when entered as a continuous variable, did not significantly predict NP attachment preference, confirming that proficiency’s influence is non-linear. Discussion These findings collectively demonstrate that the resolution of overgeneralization is structure-specific and depends on the interaction between L2 proficiency, structural markedness, and L1-L2 overlap. The study contributes to a more general understanding of L2 development by elucidating the dynamic interaction between syntactic form and semantic interpretation at the syntax–semantics interface.
No takes yet. Share an insight, caveat, or question.
Feng et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: