Answers are grounded in this paper and cite their sources.
Investigates the influence of statistical properties on L2 verb constructions, suggesting complex patterns of influence based on proficiency.
This study explores how statistical properties of verb–preposition and verb–particle constructions (VPCs) and their interplay in the target language shape second language (L2) production across L2 proficiency levels. We compared phrasal frequency and three contingency measures—mutual information (MI), forward delta P (Δ P ), and backward Δ P —of VPCs in a self‐built L2 corpus of 649 Chinese learners with those in the British National Corpus (BNC). Mixed‐effects models showed that usage‐based distributional statistics in the target language predicted L2 production with complex patterns: (a) Statistical information was selectively encoded, with backward association patterns failing to transfer to L2 output; (b) the acquisition of statistical information was heterogeneous yet interdependent: One target‐language statistical feature influencing another in L2 production, although the specific patterns varied across contingency measures; (c) word‐level statistical and linguistic properties and VPC type influenced L2 production; (d) contingency measures showed stronger associations with language proficiency than frequency; (e) language proficiency moderates the effects of word‐level properties and VPC type.
No takes yet. Share an insight, caveat, or question.
Yanlu Zhong (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: