2 min readDec 4, 2020

Idiotmatic translation - human or not?

Idiomaticness is often cited as the core element of a good translation. However there are many examples of writing that is idiomatic but still bad. You simply cannot translate sentences like this and NOT be wrong: "The discussion about the implementation of these measures was concluded by the board during a joint alignment meeting between the individual members which took place recently."

Before you start translating this grammatically and idiomatically flawless gobbledygook, you first have to paraphrase it in your head, otherwise the translation just perpetuates the evil. (Like all evil, it spreads easiest when we turn off our brains.)

By paraphrasing you sort this information and ask who actually did what.
You'll also ask yourself:
- Does it matter when the meeting took place?
- Does it matter that the board had a meeting and that this meeting has a fancy name?
- Do you need to mention "discussion" and "implementation"? Or are the measures your main object? Or the decision?

Some of these questions answer themselves, others only the author can answer - if they're even human, who knows...
So you'll have to make decisions based on your incomplete information.

Translating this straight into German will produce the same mess, no matter if you're using MT or not:
"Die Diskussion über die Umsetzung dieser Maßnahmen wurde vom Vorstand während eines kürzlich abgehaltenen gemeinsamen Angleichungstreffens zwischen den einzelnen Mitgliedern abgeschlossen."

The only way to make any sense out of this is to cut it down. "The board recently passed the measures." is semantically equivalent to the original sentence and doesn't contain a word too much. It's also easy to translate, both for humans and machines: "Der Vorstand hat kürzlich die Maßnahmen verabschiedet."

This doesn't solve the wider moral issues - most texts should never have been written - but at least the readers can judge quickly whether or not the text contains anything of value. At least you'll have saved someone somewhere some time.

This is where both human and machine translators fail most often: they don't process the input enough, reproduce the author's failures and leave it up to the user to sift through the rubble.

Have you been struggling with crap in, crap out translations? I'd love to hear your thoughts!


Translation, localization, multilingual content, technical communication