Online machine translation for L2 writing across languages and proficiency levels

Abstract

Using machine translation (MT) tools for language learning has become a common practice among language students in recent years. Studies have investigated how students use MT, how students and teachers perceive its benefits and drawbacks and how helpful it is for language learning. These studies indicate that students think MT tools are helpful in L2 writing due to their quick and easy access and use them in many aspects of L2 writing, such as vocabulary search, grammar checking, and writing revisions. However, concerns for the accuracy of outputs, the effectiveness of MT for language learn-ing and academic integrity are shared among students and teachers. This present study is based on a survey of 12 teachers and 150 students across five different languages and three proficiency levels at a tertiary institution in New Zealand. The quantitative and qualitative data were analysed to compare MT use and perceptions among proficiency levels and languages as well as between teachers and stu-dents. The findings reveal patterns that indicate different practices and perceptions between students of non-alphabet-based and alphabet-based languages. The analysis also demonstrates correlations between advanced-level students and more sophisticated and critical use of MT. With the proliferation of MT tools with neural machine translation (NMT) systems and improvement in their accuracy, the findings of this study contribute to the understanding of MT literacy to ensure effective and critical use of MT with an awareness of unique contexts and expectations of L2 writing in different language courses.