Adverb placement by Chinese EFL learners: A MuPDAR(F) approach

Abstract

This study investigates stance adverb placement in academic argumentative writing by Chinese EFL learners compared to native English speakers. Previous work has mainly examined overall distribution or misplacement, with little focus on learners’ positional preferences. Traditional statistical methods assuming data independence also limit analysis of the multifactorial nature of adverb placement. To address this, the study applies Multifactorial Prediction and Deviation Analysis using Regression or Random Forests (MuPDAR(F)) to the placement of epistemic and style stance adverbs, by integrating syntactic, semantic, and discourse-level variables such as semantic coherence and sentiment value. The results show systematic learner deviations, most evident in M3 (the position following the main lexical verb), influenced by −ly suffixation, auxiliary presence, and L1 transfer from Chinese. Learners also overuse clause-initial high-certainty adverbs (e.g., absolutely), indicating less syntactic flexibility. Positive sentiment promotes more native-like placement, especially in M3 contexts. Findings highlight how syntactic and discourse factors interact in shaping stance adverb placement, refine second language (L2) morphosyntactic models, and suggest pedagogical strategies for mitigating L1 transfer and raising pragmatic awareness.