Abstract:

This study reports the experimental results of the effectiveness of machine translation as a tool in learning of English as a foreign language (EFL). The primary research question is whether reading English sentences with machine translation put heavier burden on EFL learners than reading English sentences without machine translation. Here, machine translation refers to sentences translated by a machine translation system that is accessible on the internet, and the direction of translation is from English to Japanese, a native language of the EFL learners in this study.
The background of this study is the recent application of machine translation in EFL classes. Various studies proposed the use of machine translation in EFL classes, and investigated the effectiveness of machine translation as a tool in EFL learning (Anderson 1995, Fuji et al. 2002, Kumar 2012, Eskenazi 2013). These studies follow recent development of machine translation systems that achieved a high performance and a high accessibility on the internet. For instance, Fuji et al. (2002) reported that the use of machine translation surely assisted EFL learners, because EFL learners had higher scores in an English test written with both English sentences and machine translation sentences than in an English test only with English sentences. These studies showed the effectiveness of machine translation for EFL learners in principle. However, it is still unknown whether the use of machine translation will increase the effectiveness of EFL learners’ language use in practice.

From practical viewpoint, the use of machine translation raises a time-consuming problem, because EFL learners have to read not only English sentences but also translated sentences. Literally speaking, the number of sentences that EFL learners have to read becomes twice. Therefore, it might need longer time to complete reading English sentences with machine translation than reading without machine translation. That is, the use of machine translation might put heavier burden on EFL learners.

In an experiment, approximately 80 EFL learners read two types of reading materials. One had only English sentences, and the other had both English sentences and the machine translation. The experimental result showed no statistic significance in the reading speed between reading materials with and without machine translation (p < 0.01). This result suggests that machine translation does not put heavy burden on EFL learners. Moreover, some of the EFL learners read English sentences with machine translation faster. This shows the reduction of burden on these EFL learners. These results provide a piece of evidence for the practical advantage of machine translation as a tool in EFL learning.
Although the practical advantage of machine translation was confirmed, an important question remained unanswered. That is a question whether machine translation is effective for advanced EFL learners, because the experiment was carried out for EFL learners at beginner and intermediate levels.

Keywords:

English as a Foreign Language, Machine Translation.