Advanced second language performance and influencing factors from the perspective of feature reassembly

This meta-analysis examines advanced second language (L2) performance from the perspective of feature reassembly, investigating how reassembly requirements, overtness, and methodological factors influence acquisition difficulty. We synthesized 25 empirical studies yielding 38 effect sizes ( N = 1,306). Results show no statistically significant difference ( g = −0.092, p = .305) between advanced L2 learners and native speakers. Reassembly requirements do not directly modulate difficulty, but task type does: while learners achieve native-like performance in Form-Focused tasks ( g = 0.144, p = .348), a non-significant negative trend is observed in Appropriateness ( g = −0.323, p = .200), and significant deficits emerge in Meaning-Mapping tasks ( g = −0.679, p = .003). Furthermore, exploratory analyses reveal the transition from L1 overt to L2 non-overt features remains the sole path exhibiting significant deficits ( p = .033). We interpret these results through a multi-dimensional framework involving interface-level complexity, communicative competence, typological distance, and the cognitive costs of feature recognition and inhibition.