Predictive Effects of Phraseological Complexity on the Quality of Chinese EFL Learners’ L2 Oral Production

Abstract

Phraseological complexity measures have demonstrated significant predictive power for L2 proficiency, predominantly in written contexts. However, empirical evidence regarding their role in L2 oral production remains limited. This study investigates the predictive validity of phraseological complexity measures on L2 oral performance and examines their capacity to differentiate learners across proficiency levels. Analyzing oral performances of 90 test takers in the TEM8-Oral (test for English majors band eight oral test), results revealed that phraseological complexity exhibited stronger explanatory power for oral performance compared to lexical and syntactic complexity, with advmod_RTTR (‘adverb + adjective/adverb/verb’ Root Type-Token Ratio) emerging as the most potent predictor. Significant differences were observed in amod_RTTR (‘adjective + noun’ Root Type-Token Ratio), dobj_RTTR (‘verb + direct object’ Root Type-Token Ratio), and advmod_RTTR across proficiency levels. Notably, dobj_RTTR and advmod_RTTR uniquely distinguished between all adjacent levels, while other features differentiated only specific levels or showed no clear trends. This study expands on the research on the relationship between phraseological complexity and L2 speaking ability in a Chinese context and provides implications for oral English teaching and refinement of rating for English speaking assessment.