Developing a corpus-based paraphrase tool to improve EFL learners' writing skills

Abstract

Paraphrasing, or restating information using different words, is critical to successful writing. However, EFL learners have difficulty in making paraphrases to meet their writing demands, and there has been little research on developing automatic reference tools to assist these learners' paraphrasing skills for better writing quality. In this study, we developed PREFER, an online corpus-based paraphrasing assistance system. Allowing multi-word input and returning promptly with a list of paraphrases in English and Chinese, along with usage patterns and example sentences, PREFER provides substantial support for EFL learners to vary their expressions during writing. An assessment study of the effectiveness of PREFER was conducted with 55 Chinese-speaking EFL college freshmen in an Asian country. The results indicated that PREFER offered the most benefits to students' writing performance (with an after-use improvement of 38.2%), compared with an online dictionary and an online thesaurus (−31.6% and −6.2%, respectively). Further investigation revealed that the less proficient, more motivated, and more conservative students showed more significant progress in the paraphrasing task with the help of PREFER. In the meantime, nearly 90% of the students expressed satisfaction with the paraphrases generated by PREFER, and its functions, and another 75% of them acknowledged that PREFER benefits their writing task.

Keywords:

• paraphrasing competence
• paraphrase e-reference tool
• corpus-based
• writing performance

Acknowledgements

This paper was support by the project “Corpora and NLP for Digital Learning of English (CANDLE): Academic Writing and Speaking” funded by National Science Council in Taiwan (project number: NSC 100-2511-S-007 -005 -MY3) and was also supported by the Institute of Linguistics Fellowship for Cross-Disciplinary Doctoral Candidates from Academia Sinica, Taiwan. We would like to thank the participating students involved in the assessment phase. Thanks are also extended for the valuable comments on the manuscript from anonymous reviewers.

Additional information

Notes on contributors

M.-H. Chen

Mei-Hua Chen is currently a Ph.D. candidate at the Institute of Information Systems and Applications at National Tsing Hua University, Taiwan. Her research interests involve computer-assisted language learning (CALL), TESOL, computational linguistics (CL), natural language processing (NLP), and second language acquisition (SLA).

S.-T. Huang

Shih-Ting Huang is a graduate student at the Department of Computer Science at National Tsing Hua University. He specializes in the field of natural language processing.

J.S. Chang

Jason S. Chang specializes in the field of natural language processing and teaches at the Department of Computer Science at National Tsing Hua University.

H.-C. Liou

Hsien-Chin Liou teaches foreign languages and literature at National Tsing Hua University with a research focus on CALL.