Does use of computer translation as an aide when writing a paper require a citation or disclosure?

Question

Related (not duplicates):

• Is it appropriate to cite a paper in a language I don't understand?
• Would it be ethical to hire a proofreader for theses and academic articles?

One of the things that I have found extremely helpful when writing in a non-native language is to use tools such as Google Translate, where I will first write what I want to say in my native language, put it into the translator, and then revise the translated output based on my actual knowledge of the language. Specifically, I am often able to recognize the correctness of a passage in that language (e.g. word usage, false friends, correct verb tenses, spelling errors, appropriate levels of politeness, etc.) better than I can write fluently from scratch.

When doing such a thing in an academic paper, is it generally necessary to either:

• Disclose that an automated translation tool was used to assist in the composition of the paper?
• Formally cite the translation tool (e.g. Google Translate)?

I'm not asking about any specific situation (contact your instructor or editor), but about best practices and general understandings. I'm not asking whether it's acceptable to copypaste directly from the output of an online translator straight into a manuscript, thesis, or course essay (that would be crazy), but whether the fact that one has used a translator in a non-trivial way can be considered unethical. Would the answer differ based on context? For example, my instinct tells me that this would probably not be ok for coursework in which mastery of the target language is a main focus (e.g. you are taking Advanced English as a Second Language and are assigned to write a short story), but would be more acceptable in a context (coursework or journal submission) in which mastery of the target language is not the focus (but rather the medium of demonstrating mastery over or research findings regarding some other thing), such as writing a paper on astrophysics for a journal that requires that all papers be in English (the journal is really focused on your astrophysics findings, and the language is just a standardized medium of information exchange, i.e. a tool to do so).

It's already well-accepted that using general language reference tools such as dictionaries, thesauri, grammars, and style guides are not unethical, but it's not clear if automated translation tools are truly analogous to them.

In a nutshell, the process I'm speaking of is as follows:

• Write my original composition, findings, thoughts, analyses, etc. in my native language, without doing anything remotely resembling plagiarism.
• Send what I wrote through a translator.
• Use the output of the translator as a framework or draft, going through it and making appropriate corrections (conjugations, false friends, incorrect synonym chosen (e.g. translating "hot" as in temperature as "spicy" or "sexy"), awkward phraseology, archaic language, inappropriate diction, blatant translation errors, etc.), but without recomposing the entire paper from scratch in the target language.

Answer 1

My opinion (note: opinion) is that you don't need to do a citation here as the words are yours originally and the meaning/essence isn't intended to be modified in any way.

However, my opinion, also, is that it might be useful to note that auto translation was done. This is just so that a reader fluent in the translated language might be forewarned that any anomalies of language might be explained by the use of the translator.

I think that the workflow in such a case is to tweak the translation in any case as such things are still far from perfect. In this case, especially, it is clear that the translator is just a tool. You state, in fact, that this is your workflow, of course.

Answer 2

In my opinion there are two situations that need to be addressed:

1. Writing an academic paper to be published in a journal. I personally would not cite a translation service that I used to guide my writing. This is similar to how I would not provide a citation for standard uses of C++ or Python in a journal article. It would sound silly to say "I am indebted to Google Chrome and Google Scholar for their assistance in searching the Internet for resources for this paper." The focus of a journal article is not on the writing of the paper as much as it is on what is being presented on the topic.

2. Writing a dissertation/thesis. In my thesis and dissertation I cited every resource that I used. This includes citations for several programming languages and LaTeX.

   This thesis was typeset using LaTeX [citation]. All models in Chapter 4 were fit using Python 3.6 [citation]. All models in Chapter 5 were fit using R [citation].

   I feel that the writing of a thesis or dissertation is of larger emphasis than that of a journal article. The purpose of a dissertation is to, quite literally, dissert. Obviously the content needs to be strong and purposeful. But you are allowed much more space for providing every detail from the research process.

Answer 3

To me it sounds like what your describing is the verification [or re-verification] of what your already know. Obviously, you are not actually relying on google translate in a significant respect, for if you were you would probably paste your whole English text and spit out your answer.

This is similar to what I do all the time while chatting in foreign languages (spoken or written). I have a grasp of the language to a good enough degree, but I can not afford to look up each question which may arise. However, most of the time I would be able to recognize right from wrong. An easy example would be forgetting an irregular verb conjugation: I don’t remember, Google does. But more than that, I can easily forget many idioms which I can verify my recollection through Google, which has access to the internet as a whole. Also other services such as Reverso which specializes in this.

In summary, the machine is not doing anything you wouldn’t do if you had more time or energy.