Worst translation errors by humans and machines

Worst human translation errors

“To err is human, to forgive divine” and that is the spirit we will adopt for this week’s blog post, which delves into some decidedly second-rate translations. We all have bad days, but some of the examples of human translation errors here crop up sufficiently often or are so clunky, you wonder what kind of quality control was going on before the final translation was approved. In the spirit of forgiveness, mistakes are there to be learned from, so let’s resolve to weed out awkward phrasing of the kind described below.

Overly literal translation

Overly literal translations are probably the number one culprit in poor quality translations. The fact that they are not hard to find suggests either a lack of careful proofreading or a tin ear for unidiomatic phrasing on the part of the proofreader. German clothes are sometimes described as having “hervorragende Trageeigenschaften”, and it is not hard to find translations rendering this simply as “excellent wearing properties”. Without wishing to be too harsh, let us just say this is not a phrase a native speaker would ever naturally use to describe the quality of a pair of trousers. The problem is that the German noun Trageeigenschaft does not have a direct equivalent in English. The meaning is roughly a “characteristic offered by a garment when it is worn”. This usually relates to how comfortable the clothing is or how well it fits, so a more elegant English translation might refer to the item offering “exceptional fit and comfort”.

Nouns can be troublesome to translate

Nouns are not the only troublesome parts of speech that can trip up the translator. If we remain in the world of fashion, German often uses “überzeugt mit … ” to describe impressive features offered by a garment. Once again, the cumbersome literal translation “this garment convinces with …” is all too common. Finally, in a feat that combines poor translation choices for both noun and verb, we have the dubious wording “make a good figure” for the German “eine gute Figur machen”. A native speaker might be more likely to express this simply by saying something “looks fantastic” or is “very stylish”.

Fashion translation is perhaps particularly prone to producing some of the worst translation errors because it often involves bulk product descriptions, some of which may even have been done using machine translation. Good post-editing should weed out the kind of examples described above, so either the human checking stage of the process is being skipped altogether, or the quality control standards are insufficiently stringent. But fashion is not the only domain susceptible to such flaws. I recently came across a translation for a wine producer, in which the German “Edler Tropfen” was translated as “noble drop” (rather than the more idiomatic “fine wine”). Indeed, the Linguee online dictionary, which uses webcrawlers to identify bilingual texts, suggests “noble drop” alongside “fine wines” in its definitions, giving this phrase a status it certainly does not deserve. This definition seems to have slipped through the Linguee QC system offered by its human-trained machine-learning algorithm (which estimates the quality of the paired translations) and the manual ratings submitted by users.

Worst machine translation errors

So far, we have gone on a tour of some rather shameful translations that seem to have either been created by a human or, at the very least, slipped past whoever was responsible for the quality-control process in post-editing a machine translation. To cheer ourselves up, we will now look at some errors that can be blamed on machines, including several that are uniquely the result of not having a human involved in the process. Some of these machine translation errors boost the spirits of any translator involved in PEMT tasks because they are so easy to spot, and the source of considerable entertainment.

Mistranslations of abbreviations

Top of the list is the magnificent machine translation describing the dimensions of a dinner plate, where the German abbreviation “BH” representing “Breite x Höhe” (width and height) has been confused with the same abbreviation meaning “Büstenhalter”, a bra. So the phrase “Dinnerteller BH 27 x 2 cm” is translated as “Dinner plate bra 27 x 2 cm”. Context really is important when figuring out abbreviations!

Errors in the source text can lead to translation errors

This blog has previously tackled the way spelling mistakes in the source text can lead the unwitting translator astray. In general, however, humans are much better at coping with spelling mistakes than computers because of our ability to contextualise and accurately guess what the intended text should say. I came across a great example of this recently, where the typo in the source text would undoubtedly have been spotted by a human translator but had flummoxed the MT system. The German text was rendered as “Italienischer Obestoff”, with a crucial ‘r’ missing from Oberstoff, which means outer fabric or similar. Lacking the ability to make this sensible correction, the MT system had translated this as “Italian obesity fabric”!

Unusual words are another stumbling block for MT systems, which sometimes even fail to translate the same word consistently within a single text. A recent PEMT task involving the word Samtschrecken, contained one sentence where this was translated as a velvet grasshopper, and another where the computer had settled on the delightful choice of velvet scarecrow. The human proofreader should always be suspicious of such unlikely collocations when trying to spot machine translation errors. Indeed, the intended reference here was to stick insects (specifically the species Peruphasma schultei).

Machine translation often lacks context cues

In another demonstration of the superior contextual interpretation offered in human translation, a PEMT project included a description of a proposal of marriage, where the MT system had simply plucked out the dictionary definition ‘genuflection’ for the German Kniefall, resulting in an entirely inappropriate description of the romantic moment where someone gets down on one knee to pop the question. The same text rendered the word Antrag as an ‘application’, making it sound more like a soulless job hunt than a romantic proposal of eternal partnership.

Let us finish revelling in the shortcomings of machine translation with the splendid technical translation of Bauchfrei as ‘belly free’ in reference to a cropped top! Human involvement in translation is not totally redundant yet.

About the Author

Alison Tunley

Alison is a seasoned freelance translator with over 15 years of experience, specialising in translating from German to English. Originally from Wales, she has been a Londoner for some time, and she holds a PhD in Phonetics and an MPhil in Linguistics from the University of Cambridge, where she also completed her First Class BA degree in German and Spanish… Read Full Bio