Towards building a usable corpus collection for the ELT classroom
Kiyomi Chujo, Masao Utiyama, Chikako Nishigaki · 2007
As tantalizing as the potential for corpus application is in second language acquisition, we educators seem to stumble over how to make concordancing lines understandable for learners. This study explores various criteria for evaluating text samples by difficulty level in order to provide a collection of rated parallel English-Japanese corpus texts which educators can use in ELT classrooms, and provides the rating tools and methodology so that educators can evaluate their own classroom material. Data was collected from two English-Japanese parallel corpora, and seven indices (readability scores, average word length, Japanese school textbook vocabulary coverage, BNC text coverage, Japanese vocabulary ratio, sentence length, and kanji character ratios) were applied to measure the linguistic difficulty of both English and Japanese text samples. It was noted that most of the texts in the collection were advanced level, and that there is a shortage of copyright available e-text data at the beginner level. Nevertheless, this study identifies several applicable indices, provides a rated collection of titles at varying levels of difficulty, and takes corpus usage one step closer to its ideal application.