Editors and Machine Translations into English

I posted this on LinkedIn on August 22, 2024, in three parts, and I have made some minor additions.
 
How best can editors work on text that has been machine translated into English? While some of my clients do use machine translations, some of my clients do not and advise against using machine translations.

When a client uses machine translations into English before hiring me as an editor for the English publication, I find that editing often goes beyond the usual scope for a copyedit (as defined by Editors Canada), and I need to understand the potential benefits and pitfalls of machine translations to do my edit well.

Here are 9 practices I recommend for editors facing the same situation:

1.   Check out what machine translation involves, possible applications, and risks. Here are some sources.
·     Government of Canada’s “Using new language technologies and machine translation,” subsection “To use or not to use machine translation, that is the question” <https://lnkd.in/eBjCuEtP>
·     Slate article “Death by Machine Translation?” https://lnkd.in/ecbcBm6k
·     “5 Risks of Machine Translation” < https://lnkd.in/eJqibMqg>

2.   Ask what post-machine translation process the text has gone through before it lands on your desk. For example, after the machine translation, was there a post-machine translation edit of the text by a human translator? Or did the multilingual authors and subject experts review the text to identify errors in the translation and address those errors before the text was sent to the English editor?

3.   Consider the client’s style guide, other references, and process. Does the client answer questions about many publications, refer to precedents, and refer other questions to the author team?

4.   Consider multiple Englishes. Likely the machine translation defaults to US or UK English, but what does the intended audience know and prefer? What does the client require? What national, regional, and local considerations are factors?

5.   Use high-value resources and use them well–e.g., a specialist glossary of terms in English for the subject, a wealth of online and hard copy dictionaries (Oxford, Merriam-Webster Collegiate 11th edition, Collins, the online Oxford English Dictionary, and the online Oxford Reference).

6.   Consider the process for the edit in English. Will the edit go through at least two rounds of edits: first, an edit with tracked changes and queries or comments to authors; then the authors’ careful review of the text to check the edit and answer all questions; and then at least a second edit round to resolve all text changes and clean up the text?

7.   Review for key words and phrases, concepts, and terms of art that are used and expected in the subject area. Are there areas of text that are vague or might lack nuance? Does the text use many different words or phrases for the same thing? Does the text seem to oversimplify a term of art because of a limited vocabulary in machine translation?

8.   Keep a style sheet and query the subject expert/author team about specific wordings and patterns of language. Does the text use words and phrases that will be familiar to the audience and explain those that are unfamiliar but necessary in the context? Does it use those words and phrases clearly, consistently, and with precision? Does the text vary language in a way that might be unhelpful for the target audience; does it look like the writers dug into a thesaurus too many times? This might result from machine translation translating each word and phrase separately and without regard for the context and consistency. For example, I have often queried early in a manuscript that three similar phrases are used repeatedly but in a way that is not clear and distinct to me, so what distinctions are being made and are the uses clear and appropriate in each case?

9.   Review for tone and style. What tone is wanted for the intended audience? Is the tone and style consistently formal and academic, consistently informal, or a combination? Is the language vague? What style is clear, correct, and precise for the intended audience?

Writers, editors, translators, publishers, and other organizations communicating in multiple languages, please share what you have learned in your work and about machine translations. Thanks!

© Laura Edlund 2010