Many Japanese high schools have class sizes of 40 or more students, making it difficult to deal with the details. In addition, the Common University Test only measures reading and listening comprehension. Because of these and other factors, English learning in high schools is often focused on reading and listening to texts and achieving high scores on entrance exams and certification tests. As a result, according to a survey of high school seniors, only less than 20% reached the A2 level in written English.
In recent years, the accuracy of Machine Translation (MT) has improved, and some MTs have voice input/output capabilities. The authors decided to aim for a paradigm shift in English language instruction for first-year university students. Teachers will use MT to improve the efficiency of time-consuming English correction, while students will use MT on their own for speech training while giving speeches and debates in class.
First, in 2021, the authors investigated how the use of MT changes the attitudes of first-year university students at the A2 level through the teaching of English essays. The first survey revealed that many students felt guilty about using MT. This is because their experiences in high school was that they used MTs to read given English texts instead of reading them on their own. In addition, many learners were not able to use MT appropriately. Therefore, in the class, the authors taught them how to use MT appropriately as a learning machine for writing essays. Then, many students had a positive attitude toward the use of MT.
Then, in 2022, the authors taught the students how to correct English sentences by themselves using MT and how to do voice training by themselves using the voice input/output function of English. First, as for the correction of English sentences using MT, there were quite a few essays submitted by the students that showed that some of them did not understand the structural differences between the Japanese and English sentences. Therefore, the author instructed the students to make the Japanese sentences to be machine-translated into English well and to use another MT to back-translate the translated English sentences into Japanese to check whether the Japanese sentences were translated as they were meant to be. As a result, the quality of the students' English writing gradually improved. Next, for voice training, the students used the voice input/output function of Google Translate. First, the students type their own English sentences into Google Translate, which outputs them as spoken English, and then they repeat the English sentences over and over to practice pronouncing the sentences. Regarding the checking of pronunciation, the students were instructed that they would pass the test if their pronunciation entered into Google Translate was accurately recognized by it.
Thus, by using MT as a language learning machine, the authors were able to shift the paradigm to a class where students wrote English sentences that effectively persuaded their opponents in debates, and spoke while efficiently learning correct pronunciation in speeches.
