Controversy exists about to whether language productivity emerges from a rule-governed system (dual-mechanism model) or an exemplar-based, associative system (single-mechanism model). Results of previous studies using tests of native speakers of Japanese (JNS) with made-up verbs suggest the nonexistence of a rigid 'rule' representation for the Japanese verb conjugation paradigm, which is much more complicated than the English equivalent. This study investigated the acquisition of Japanese verbal inflection by 41 learners of Japanese (L1; Mongol, Chinese, and Korean) using the same nonce verb tests as those used by Klafehn (2003) and real verb tests. Results showed significantly higher scores for real verbs than for nonce verbs. Moreover, as the group members' proficiency increased, their scores increased. Higher-level learners showed native-like performance in real verb tests (94%) and outperformed Klafehn's JNS in nonce tests (77% vs. 56%). The lower-level group showed the worst score in real verb tests (80%) and showed the closest score to JNS in nonce items (59%). The results suggest that L2 development in verb inflection differs from that of JNS and that it is affected by both rote learning and explicit rule learning.
Selected Proceedings of the 2009 Second Language Research Forum: Diverse Contributions to SLA
edited by Luke Plonsky and Maren Schierloh
Table of contents