Abstract：
Abstract: This article reports on a corpus-based multifactorial analysis of relativizer omission by Chinese EFL learners and native speakers of English. The findings reveal that a) the usage of relativizer omission by both groups is constrained by a variety of linguistic features, thus displaying a prototypical and probabilistic distribution; b) when confronted with complex relative clauses, Chinese EFL learners, compared to native speakers of English, demonstrates a heavier reliance on explicit relativizer to alleviate processing difficulty; c) different from exemplar-based probabilistic approach in L1 acquisition, Chinese EFL learners apply probabilistic approach and abstract grammatical rules respectively to relative clauses with low/high cognitive complexity. This study has identified the global similarities and nuanced discrepancies, with regard to relativizer omission, between Chinese EFL learners and native speakers of English.
Key words: relativizer omission; binary logistic regression; overgeneralization; probabilistic prediction
