Computational Linguistics & Reading Comprehension Text Complexity: Implications for the EFL Teacher

This paper reports on a 1-year longitudinal study that aimed at investigating the relationship between intermediate EFL learners’ development of reading and writing proficiency while delineating a range of linguistic features present in a set of 30 reading comprehension texts and 350 written essays. Data (reading comprehension exam scores and written essays) were collected from a junior high school in Athens, Greece and 70 EFL students aged 14-15 years old took part in the study. By making use of advanced Computational Linguistics and Automated Machine Learning systems an attempt has been made to find the relationship, if any, between a set of 135 text variables, students’ L2 lexical growth and the readability level of pertinent texts. The rationale of this study is closely linked to Weir (2005: 292) and Alderson et al. (2004: 11) words of caution that “although the CEFR attempted to describe language proficiency through a group of scales composed of ascending level descriptors, it does not contain any guidance, even at a general level, of what might be simple in terms of structures, lexis or any other linguistic level". In fact, the argument that the CEFR is intended to be applicable to a wide range of different languages offers little comfort to the EFL teacher, who has to select texts or activities uncertain as to the lexical breadth of knowledge required at a particular level within the CEFR. The findings of the present research provide practical guidance to EFL teachers, test designers and material developers as to the range of lexicogrammatical features an EFL learner of an expected level of language ability might be able to handle for a successful reading comprehension performance along with the linguistic strategies s/he might have developed for the production of more elaborate written texts.