81006698

open access

Version of Record

A Machine Learning Approach to the Effects of Writing Task Prompts

English (英語)

Learner Corpus Studies in Asia and the World

2

163-175

神戸大学国際コミュニケーションセンター

2014-05-31

2014-06-25

In analyzing learner language use, it is important to understand the influence of prompts given for writing tasks. The study reported upon in the present paper aimed to investigate the effects of essay prompts on language use in essays written by English as a foreign language students, using a machine learning approach. The study compared the essays of East-Asian learners of English in terms of various linguistic features (e.g., parts-of-speech, grammar, and discourse features) proposed by Biber (1988), The essays were drawn from the International Corpus Network of Asian Learners of English (Ishikawa, 2011), which is considered to be the largest East Asian database of written compositions. The database contains essays written in response to two prompts, namely (a) It is important for college students to have a part time job, and (b) Smoking should be completely banned at all the restaurants in the country. The results of the present study indicate that the language used in the prompts significantly influenced the learners' language production. The findings suggest that it is crucial to consider essay prompts in terms of parts-of-speech, grammar, and discourse features, and to investigate the influence of such essay prompts on written compositions when conducting learner corpus studies.

紀要論文