Type and token frequency of conventional linguistic units in Extensive Graded Reading


  • Kunitaro Mizuno Edogawa University


graded readers, extensive graded reading, incidental vocabulary learning from reading, usage-based model, conventional linguistic units


Previous studies on incidental vocabulary learning from reading were based on the idea of repetition (e.g., Horst, 2005; Pigada & Schmitt, 2006; Waring & Takaki, 2003; Webb, Newton, & Chang, 2013) which has its roots in the comprehensible input hypothesis (Krashen, 1985). This study demystifies the fundamental idea of repetition in terms of learning conventional linguistic units (Langacker, 2008) from extensive graded reading (EGR). In order to explore the frequency effects of conventional linguistic units in EGR, a corpus of 60 graded readers was constructed and analyzed from a perspective of the usage-based model (Tomasello, 2003). The results show that the token frequency of conventional linguistic units is low in the corpus, and indicate that EGR practitioners need to contrive ways to have learners pay more attention to them in the contexts of stories and effectively entrench them in memory.