Limited L1 transfer and efficiency-driven strategies in L2 syntax–semantics processing: An ERP study

Cross-linguistic transfer accounts predict that second language (L2) learners benefit from structural overlap between their native language (L1) and L2, while the accessibility–complexity–efficiency model proposes that L2 learners rely on accessible cues to maintain efficiency when processing non-canonical structures with less accessible cues. Here, we tested these accounts using Mandarin Chinese non-patient-object structures, where verbs take non-patient arguments as objects. Event-related potentials were recorded while Mandarin Chinese L1 speakers and Vietnamese learners of L2 Chinese read sentences and made plausibility judgments, with patient-role priming manipulated in the preceding context. Results revealed that L1 speakers showed comparable N400 responses for patient and non-patient objects. Non-patient objects in the patient-priming context elicited increased P600 effects, suggesting greater semantic integration effort. In contrast, L2 learners showed increased N400 and late negativity amplitudes for non-patient objects compared to patient objects, indicating retrieval difficulties when thematic roles conflict with expected verb–patient mappings. Patient-priming contexts reduced these effects, producing patterns similar to those for patient objects, suggesting that priming patient cues facilitates processing. These findings suggest that despite the L1–L2 structural overlap, L2 learners rely on canonical verb–patient structures to cope with thematic role assignment challenges, reflecting efficiency-driven processing in response to syntax–semantics interface complexity.