a white man with blond hair looking at a computer with a look of deep frustration. The man is in his twenties.

5 translation pitfalls in the narrative social sciences

Historians, social anthropologists, and qualitative sociologists often work across different cultural and linguistic contexts. For example, you might do your research in German or Croatian, but publish some of it in English.

All of these endeavours involve a kind of cultural translation. They often include the translation of interview excerpts, field observations, and academic texts too.

However, one problem is that many researchers have never trained professionally in translation and often improvise.

As an editor, translator, and recovering academic, I have noticed several common problems that crop up. Learning about them is the first step to finding a solution, so read on!

Problem one: direct translation and undertranslation

Yes, I know – these are two separate problems! But they quite often come together. The most common issue I see is direct translation.

This is when the translated text maps too closely on to the original. Usually, this is because individual parts of sentences have been translated literally or directly, instead of translating each sentence or the entire text more holistically. The result often sounds clunky, retaining elements of the source-language syntax.

Let’s take a look at an example of a bad translation. Here, I’ve doctored and exaggerated a version of my own early work, translating in 2016 as a full-time researcher with good Croatian skills, but with little translation experience and no training.


  • The text in red highlights certain phrases that have been translated literally or sound awkward
  • The text in orange highlights a phrase that needs more explanation to an English-language audience. Not explaining the phrase results here in undertranslation in the bad translation, where too much information from the source text is lost
  • Finally, the text in blue is a small mistranslation in the bad translation
Source text: Na primjeru Srbije i obrazovanja na manjinskom jeziku (hrvatskom) vidjet ćemo da problemi izlaze izvan okvira jezičnih prava, ali da jezik jest značajan jer je jedan od uvjeta postojanja nastave na manjinskom jeziku.
Bad translation: On the example of Serbia and minority (Croatian) language teaching, we will see that that the problems go beyond the scope of the language-rights framework, and that the language is significant, as it is a condition for the existence of minority-language teaching.
Improved translation: By considering the example of Serbia and education in a minority language (Croatian), we will see that the problems come from outside the language-rights framework, but that Croatian’s classification as a language is important, because it is one of the conditions for there being minority-language teaching.
Source sentence shared with the author’s permission.

Can you feel the difference? Of course, direct translation and undertranslation often occur for good reasons: the researcher wants the translation to map faithfully onto the original. The researcher has often been careful, but the result reads poorly and includes constructions that simply don’t work in the target language. On a deeper level, the researcher hasn’t adequately considered the new audience for the translation.

Finding the sweet spot

Now, how much you choose to depart from the original depends on the kind of text. If you are translating a refugee testimony for legal purposes, conveying the precise wording is more important than translating style features of the discourse. Equally, if translating a creative or literary text, you want the translation to read well, with appropriate solutions for style and tone.

The writing context is important here. However, even a translation deliberately close to the original should not include clunky, imprecise sentences that confuse the reader and don’t convey the meaning accurately.

This brings us on to a related issue:

Problem two: mistranslations

Mistranslations are another frequent problem. The more serious of these result from false friends. Take a look at these examples:

Source phraseMistranslationPossible translation
modni detaljifashion detailsfashion accessories
emisijeemissions(tv) programmes
dubinski intervjuideep interviewsin-depth interviews

You can easily make such slip-ups when you are working between two or more languages.

Of course, language also does this kind of borrowing all the time. Indeed, when a literal word-for-word translation takes root in another language (e.g. beer garden [from Biergarten]), linguists call it a calque. In a way, this is just an extension of a natural thing that our brains do. In turn, when writing for publication, we need to make sure we’re intelligible to our audience.

Problem three: overtranslation

Overtranslation is the opposite of undertranslation. Here, the translator takes liberties, usually unintentionally. They just translate the loose gist of the text, adding extra details that are unnecessary.

I have also come across this, but much more rarely than direct translation. Sometimes, this results from sloppiness, while on other occasions it’s about being very creative combined with a lack of awareness of the translation process.

Problem four: translation inaccuracies

Professional translators working on complex texts will normally triple check each sentence they translate. It’s easy for words to be omitted or for extra words to be inserted by mistake. This happens a lot with words like ‘nevertheless’, ‘though’ etc. Indeed, a judgement call must be made where certain words are used for emphasis, where that emphasis could be conveyed in another way, e.g. through a change in syntax. In Croatian, I see this with words like upravo and naime.

These inaccuracies are often omissions or insertions, and I have discussed them in more detail here. Sometimes you have to insert extra words or sentences to explain a context. For instance, an audience in the UK or US will need a military term like Operation Storm to be explained to them, while a Croatian or Serbian audience will immediately know what this refers to.

Problem five: an overreliance on source-language words and phrases

Anthropological texts often contain a lot of words and phrases from the languages used in fieldwork. This partly expresses the anthropologist’s position between these different environments. Some of these words (e.g. veze in Croatian/Serbian/Bosnian/Montenegrin) cannot be succinctly translated directly, and they form part of the anthropologist’s conceptual toolbox.

However, I have noticed that some researchers, including an earlier version of myself, would over-insert such words to enhance their ethnographic authority in the text.

Always think carefully about what words you choose to use in which language (variety), with your audience in mind. If that audience includes people who will not know all the languages drawn from, have a read through of your text before publishing to see which words should stay and which could cause problems for the reader.

Photo by Sebastian Herrmann on Unsplash

Leave a Comment

Your email address will not be published.