Making Positive Use of Machine Translation for Writing Essays

Writer(s): 
Daniel Hougham, Hiroshima Jogakuin University

Although they remain controversial, machine translation (MT) tools such as Google Translate have become much more accurate and popular in recent years. The switch from “phrase-based” translation systems to “neural” algorithms was a milestone in quality improvements for Google Translate (Le & Schuster, 2016; Caswell & Liang, 2020). A major difference between phrase-based and neural systems is explained by Google research scientists:

Whereas Phrase-Based Machine Translation (PBMT) breaks an input sentence into words and phrases to be translated largely independently, Neural Machine Translation (NMT) considers the entire input sentence as a unit for translation (Le & Schuster, 2016).

Compared to the previous phrase-based systems, Google asserts that neural machine systems produce translations that are immensely improved, reducing translation errors by more than 55%–85% on some major language pairs (Le & Schuster, 2016). Similarly, a recent independent reevaluation study looking at 51 of the 103 languages available reported an increase in accuracy of 34% over an original evaluation study carried out eight years earlier (Aiken, 2019). Even in some languages (such as Japanese and Korean) which are known to be notoriously difficult to translate, there have been discernible improvements.

The controversy is that MT tools might allow students to complete assigned writing tasks without thinking about the language and producing it themselves (and therefore unlikely to learn much or improve their writing skills in the process). However, some university teachers have begun to see the benefits of MT and are seeking ways to encourage students to make positive use of it while avoiding its pitfalls (Lee, 2020; Oda, 2020). The purpose of the present article is twofold: First, to summarize the benefits and drawbacks of MT, and second, to introduce some guiding principles and an awareness-raising activity that will help university students make positive use of MT when researching and writing essays.

 

Benefits and Drawbacks of MT

Research suggests that MT can help improve EFL students’ writing by raising awareness of their lexical and grammatical errors and assisting students in developing positive writing strategies (Lee, 2020). It can be satisfying for EFL students to use Google Translate in their English writing, especially in finding vocabulary items and in helping complete assigned writing tasks (Tsai, 2019). Oda (2020) found that when it comes to writing speeches in English, MT offers other advantages such as subject-verb agreement, verb and object combination, and translating numbers, making MT more useful than dictionaries for novice and intermediate students.

Students learning English in Japanese universities are often asked to submit essays and speeches. These types of assignments require background research and expressing one’s ideas in the form of sentences or paragraphs. MT is arguably an invaluable tool when students research an essay or speech topic, especially when reading source materials in their first language. Students can use it to translate sentences or paragraphs, or entire web pages in real-time. If they use the Google Translate Chrome Extension tool, they can hover their mouse over a chunk of selected text, click a pop-up button, and the translation results will appear in real-time.

One main drawback of MT is that the translation results are often far from perfect, depending on various factors (e.g., similarity of language pairs and complexity of sentences inputted). One crucial problem is that students may be tempted to copy large chunks of machine-translated text and simply paste it into their essay with barely any effort to paraphrase it. Students might think that, since the translation came from Google Translate, they can just use it as is (without paraphrasing it) while citing the source. They also might think that, since the original text was written in a different language, the teacher might not take the time to check the original source. My experience checking students’ references (and reading the original texts in their first language) revealed this problem existed in several of my intermediate-level students’ essays, so it seems possible that this use of Google Translate is a writing strategy employed by some students. This problem seems to stem from the difficulty of paraphrasing in general and in a second language in particular.

 

Tips and Rules for Making Positive Use of MT

Oda (2020) offered some tips and rules to help university students in Japan use MT to write speech drafts in English effectively. In short, her Golden Rules for students using MT when writing speeches are: Adopt a translation result only if:

  • you understand it,
  • you can memorize it (or, at least, read it smoothly) for your speech, and
  • you bear responsibility for what you say.     

Students who follow these rules while using MT to aid in their speech-writing activities will be more likely to choose English suitable for their proficiency level and create a speech draft that they can deliver successfully. Although these rules are beneficial for writing speeches, they leave much to be desired when it comes to using MT for other purposes such as researching and writing essays in English.

Thus, when introducing MT as a possible tool for researching and writing essays, it is essential to establish a few ground rules to help students steer clear of pitfalls. I suggest that teachers encourage students to follow what are herein called the Guiding Principles for using MT in researching and writing essays.    

 

The Guiding Principles for Using MT in Researching and Writing Essays

If you use MT to translate someone else’s writing, you will need to paraphrase and cite the translated output. In other words, if you did not write the sentences that you inputted into Google Translate, then you will need to paraphrase the translation results before you can use them (with a citation) in your essay. If you do not paraphrase the translation results, you could be at risk of committing plagiarism.   

If you use machine translation to translate your own writing, it might not be necessary to paraphrase the translation results, depending on the accuracy of the translation results. In other words, if you wrote the sentences that you inputted into Google Translate, then you might not have to do any paraphrasing of the translation results because the input was your original writing. The output will be a translation of your original sentences. However, the translation results will likely need editing to make it read accurately and smoothly, especially if you wrote and translated long sentences. Do use a good dictionary to check the accuracy of individual words or expressions from the translated output.   

You bear responsibility for what you write. You are responsible for choosing carefully what you write in your essay. This means that, if you are careless about the use of MT in your writing, you will bear the consequences (e.g., you may get a low or zero score on your essay or course, depending on how serious the misuse is and how strict your teacher is).

Students will benefit from learning how to paraphrase the translation results. Paraphrasing means rewriting someone else’s ideas or sentences using your own words without changing the original text’s meaning. In academic writing, paraphrasing is an alternative to quoting. It is usually considered better than quoting because when students can rewrite the ideas in their own words, it shows that they understand the concepts and makes their writing more original. According to Gahan (2018), five easy steps to paraphrasing are:

  • Read the passage several times to fully understand the meaning.
  • Make a note of key concepts.
  • Write your version of the text without looking at the original.
  • Compare your paraphrased text with the original passage and make minor adjustments to phrases that remain too similar.
  • Cite the source of the ideas you are using.

These steps can be tricky for EFL learners, so teachers need to give them support and tips on paraphrasing effectively. Here are four clever ways that can help students to paraphrase effectively (adapted from Gahan, 2018):

  • Start your sentence differently from that of the original source.
  • Use synonyms (words or phrases that mean exactly or nearly the same thing as another word or phrase in the same language).
  • Change the sentence structure (e.g., from active to passive voice).
  • Separate (or combine) the information into more (or fewer) sentences.

In order to raise awareness of the above steps, I suggest trying an activity that will (hopefully) appeal to students as digital natives: Ask students to watch a video on YouTube, read an online article and do a quiz (via Google Forms) that is designed to give them immediate feedback in a self-study manner. Here is a link to a copy of such a quiz I created containing an embedded YouTube video and a link to an online article titled “How to paraphrase sources” (Gahan, 2018) that students can read before answering some questions to self-check their understanding: https://forms.gle/fj9VnJr7oskjozxU6

This awareness-raising approach and the affixed materials can be implemented face-to-face or via remote learning.

It is essential that students are also taught how to check their paraphrasing using good dictionaries so that MT does not appear to be a complete replacement for dictionaries. No matter how useful MT becomes, dictionaries still have an essential role in checking for accuracy.            

 

Conclusion

This article has summarized some advantages and drawbacks of using MT to research and write essays. It has suggested some guiding principles that students can follow to avoid the pitfalls of MT. It has also shared an awareness-raising activity with digital materials that can be used face-to-face and via distance learning.      

As the use of MT becomes more widespread and inevitable, teachers have an essential role in guiding students and training them in the responsible use of MT, online dictionaries and other resources for second language writing.

 

Acknowledgement

I would like to thank my colleagues George Higginbotham and Robert Dormer for their invaluable comments on an earlier draft of this article. I also wish to express my thanks to the copy editor Alexandra Terashima for her constructive suggestions.    

 

References

Aiken, M. (2019). An updated evaluation of Google Translate accuracy. Studies in Linguistics and Literature, 3(3), 253–260. https://doi.org/10.22158/sll.v3n3p253

Caswell, I., & Liang, B. (2020). Recent advances in Google Translate. Google AI blog. https://ai.googleblog.com/2020/06/recent-advances-in-google-translate.html

Gahan, C. (2018). How to paraphrase sources. Scribbr. https://www.scribbr.com/citing-sources/how-to-paraphrase/

Le, Q. V., & Schuster, M. (2016). A neural network for machine translation, at production scale. Google AI blog. https://ai.googleblog.com/2016/09/a-neural-network-for-machine.html

Lee, S.-M. (2020). The impact of using machine translation on EFL students’ writing, Computer Assisted Language Learning, 33(3), 157-175, http://dx.doi.org/10.1080/09588221.2018.1553186

Oda, T. (2020). How to make positive use of machine translation. The Language Teacher, 44(2), 30–32. https://doi.org/10.37546/JALTTLT44.2

Tsai, S.-C. (2019). Using google translate in EFL drafts: a preliminary investigation. Computer Assisted Language Learning, 32(5-6), 510–526. http://dx.doi.org/10.1080/09588221.2018.1527361