What makes a good translation?

In today’s blog post I give my answer to these two questions:

  • What skills does a translator need?
  • Can academic researchers translate texts in their field well?

A few years ago, I came across a journal offering a prize for (unpaid) quality translation of famous academic texts. This led to an argument with a friend over whether research academics were qualified to work as translators. The journal’s rationale was that researchers best know (sub-)disciplinary conventions.

However, my friend argued that most researchers are not trained and qualified translators, or even linguists. They often don’t understand the art and mechanics of translation. And this is why they won’t do a good job. My position was somewhere in the middle. I felt the journal’s approach was arrogant. I also hated the idea of the journal promoting unpaid translation work (which only suits the minority of professors who have a long-term contract in a highly precarised field). Yet I didn’t think a qualification in translation (or similar) was necessary to do it well.

In today’s blog post, I give three replies based on the insights I gained while training for the CIOL Diploma in Translation. I trained here and the course was based around feedback on several German–English translations completed. My replies are based on the marking scheme for the CIOL Diploma in Translation. It doesn’t cover *everything*. Cultural knowledge is also key, but then a lack of such knowledge would lead to accuracy and comprehension problems. Some projects require extensive research and fact-checking too.

(1) A good translation is well-comprehended

Comprehension errors are the most serious errors a translator can make. If you have not completely understood the text content, you can’t do your job as a translator. Dictionaries and glossaries can help, but the translator really needs top-level language skills (a C2 reading level) in the source language. This doesn’t mean that a translator will know every word in the source text. Rather, with these resources, the meanings in context will be clear.

Academic researcher–translators: Your key strength here is knowing your field well. Comprehension skills are not only general, but also specific to topics and fields, each with their different conventions. If you have a good command of the source language and know the topic well, you are much less likely to make these serious errors.

(2) A good translation is accurate

Various accuracy errors creep into all translations. Indeed, translation always involves information loss, and no translation is perfect. These accuracy errors can be categorised, as in the table below. You can find a more comprehensive list here. Apparently, the most common kind of accuracy error that experienced translators make is with terminology, and so special tools have been developed for handling terminology in translation projects.

What accuracy means also depends on the writing purpose. Getting the tone and style right is key for literary translation, whereas precise terminology could even be a matter of life or death in a medical translation. Every translation loses and gains something when it is taken out of one language context and placed in another.

Accuracy does not mean translating a text word for word. It’s fine to move parts of sentences around to fit the syntax of the target language. It’s also OK if word classes and structures change. For example, English likes to use verbs in places where other languages favour nouns, for example. The key is that the text reads idiomatically for the target audience and that you preserve the author’s voice as much as possible.

In working directly with academic authors, this is one of the areas where obstacles occasionally arise when asking the author to “check” the translation. Why? Because they may not know the expectations of an Anglo-American target audience or think that a “faithful” word-for-word translation sounds better to them as it looks more like the source text. This depends on the author and the ideas they have about language, and also their vision of what the target text should look like.

Academic researcher–translators: Accuracy is one of the key areas that improves with extensive translation practice. Except for terminology (which you are likely to get right), accuracy is perhaps the main area in which an academic translating a text will slip up. For this reason, I would recommend that academic researchers serious about translation take some kind of training in translation practice, e.g. by preparing for the Diploma in Translation.

(3) A good translation is well-written (fit for purpose)

Writing skills in the target language need to be top-notch. Many translators – including myself – therefore only translate into their first (or dominant) language. Different levels of skill are required, though. A text translated for informative purposes, such as a financial report, need not be as well-written as a text translated for publication in a book or museum exhibition. This means that a text translated for informative purposes could reasonably be translated into a person’s second language – but even so, translating in this direction is harder work and the results are often poorer.

For the CIOL Diploma in Translation, you must translate semi-specialised texts in one direction (typically into your first or dominant language). All the texts in genres I have picked (general translation, social science, and literary translation) focus on texts of publishable quality.

Finally, the quality of a professional translation is quite often better than the original, as translators spend their entire working lives writing, whereas it may just be one part of a job that the author of the source text does.

Academic researcher–translators: Those of you who enjoy writing will likely have great writing skills and be able to produce a good quality finished product (although it may still need to be copyedited). Knowing the conventions of your subdiscipline will help you craft sentences nicely and your awareness of the text’s audience is invaluable here.

Type of errorMore details/examples
Consistency errors (writing quality)Using alternative spellings in the same text (e.g. travelled/traveled, competencies/competences). Switching between UK/US usage.
Mistranslations (accuracy)A big mistranslation could be a comprehension error. A small one could be an accuracy error, e.g. getting a tense wrong, mistaking an adjective for a comparative adjective. tenses, comparatives. Calques can be applied too often when translating quickly (e.g. kapitulacija>capitulation, when “surrender” may fit better, or translating Formulierung as formulation when “way of describing” might be better).
Nuance and emphasis (accuracy)Sometime a small word (e.g. upravo) is used for emphasis. On other occasions, sentence structure creates emphasis (“It is not you who has failed” versus “You have not failed”).
Omissions (accuracy)Missing small words out like jedoch, gesamt, upravo, tek, ipak.
Over and under translations (accuracy)überrascht as astonished, when surprised may be better.
Readability errors (writing quality)Compare: “It is precisely older people…”/”Older people, in particular,…”
Register errors (accuracy) E.g. saying “good relations reigned between the two parties” in a conversational, informal text.
Terminology errors (accuracy)Translating Bundestagswahl as “general elections” rather than “federal elections”.
Usage errors, i.e. not following target language rules (writing quality)Incorrectly placed commas, spelling mistakes, not translating punctuation if you use it differently in the source language.

2020 Round up: experimenting and specialising

This year has been my first complete calendar year of running an editing and translation business full-time, despite having run a part-time business for over five years. Over the year, I figured out my longer-term approach, started to see good results from my marketing, and refocused my website on my specialist areas.

Yet back in January, I hadn’t fully shaken off habits picked up in my former career as an academic researcher, and hadn’t got my head around pricing – especially as I’d lived in countries with very different prices.

Translation was more of learning curve this year than editing. Here’s what happened over the year:


I finished my PTC proofreading training (passed with merit) and continued working as an academic editor, becoming a professional member of the CIEP in November. I normally work with authors before submission to a mainstream publisher, but in spring, I copyedited several books for publishers too, which helped me better understand the editorial process. Next year I plan to finish my copyediting training and continue working directly with academic authors on book projects and on texts for publishers. I realised I enjoy line editing most of all (transforming sentences to improve style and clarity).


As my background is in linguistics and anthropology rather than translation, I decided to do a preparation course for the CIOL Diploma in Translation. Feedback on translations I completed in my favourite genres (social science and literary translation) and general translation was invaluable. This helped me massively in reviewing my own work (e.g. noticing kinds of errors – readability, omissions, register errors etc.) and learning how to respond to client feedback. If an issue arises, I now know whether the problem lies with me or with the author’s understanding of the process.

This kind of training is not available for Croatian, but I have set up an initiative with other translators to organise workshops and similar online training in future. I also realised that the field was much more hierarchical for German than Croatian – people had very specific niches, whereas with Croatian people were more likely to do a bigger variety of jobs, as the language and pool of translators was small.

By the autumn, I had also realised that non-specialist translation agencies weren’t a good fit for me, for the following reasons:

  • Requests for small and general jobs were rarely very well paid and always urgent (e.g. send me this press release/certificate/document by tomorrow)
  • My prices were generally too high (living in the UK) for Croatian and Serbian agencies
  • As there are few HR>EN native speaker translators, I found the reviewing process (key to learning and improving my craft) was often skipped

Working with agencies was a better fit for German translation work, though. And it’s highly likely that in future I will receive some more specialist requests from agencies that are a great fit. While I see such agency work (like publisher work for copyediting) as useful on the side, and for keeping up to speed with changes in the industry, my preference for large non-urgent projects means I am better suited to direct work with universities and authors. Finally, I noticed that the most interesting work came from my former academic and research networks. This makes sense as I have a personal connection to the topic.


By July I knew my focus would be mostly in publishing and would include book translation and different kinds of editing and writing. One afternoon, I was reading Louise Harnby’s book on marketing for editorial freelancers. She says having a specialism is important when you start out, as you will bring your networks and clients from your previous job. Once you have run a full-time business for several years, you can then diversify e.g. by moving into a new kind of editorial field, like fiction or literary translation.

This helped me reorganise my marketing around my niches. Once this was clear, it was easy to rework parts of my website to make it more compelling. And things got really niche in places…

Translation specialismsEditing specialisms
Partisan art, media and politicssocial sciences and humanities
Emotions in anthropologypsychotherapy
Museums, marketing and tourismCentral Europe/South East Europe

In the autumn I started to receive a much larger amount of demand for my services. This meant I could pick and choose to avoid problem clients (I also realised a general rule of thumb – ego often comes to the fore when the quality of writing is bad).

What about 2021?

The next year or two will be focused on getting more titles in my specialist areas under my belt, and on finishing my training in copyediting and then line editing.

Finally, I plan on finishing a copywriting course I started, as this fits my interests in published texts that communicate with large audiences. Translation and blog writing also benefit from good copywriting skills, especially when written for big audiences. I have also started to learn a bit about content design, content marketing and UX design.

Yes, I got to the end without mentioning the C-word!

COVID-19 has not had a big positive or negative impact on my work so far, apart from making it difficult to concentrate at times. The spring was slightly quieter than the autumn, but this was a period when I needed to focus on my CPD, translator and proofreading training, and get the basic experience with agencies and publishers needed to confirm my work was up to industry standards. Things are looking particularly good right now compared with a year ago!

Academic Translation from Start to Finish

Translating can seem rather opaque. The author/client will often simply send a file and receive a translated version back. Even with a decent level of language proficiency, it can be difficult to gauge whether a translation into a foreign language is average, good or excellent. In today’s blog, I seek to demystify this process. I do so by discussing the “workflow” I use for academic translation when dealing with direct clients. Translation agencies may divide the tasks up differently.

Understanding authorial style is key to academic translation in the social sciences and humanities. This is because there is more diversity in writing styles than in, say, legal translation. These styles are often intimately tied up with the arguments that authors wish to make and they often link to wider intellectual schools and projects too. Special attention must therefore be paid to accurately translating not only concepts and ideas accurately, but also authorial style and voice.

In my experience, the best academic translations emerge from a direct collaboration with the author, in which there is space to discuss the interventions made to a text. This is because the author knows her style best and is often better versed in the relevant academic literature and jargon than the translator. If the translator has subject knowledge, this is a great help, but translation experience and training are also crucial.

What, then, do I do exactly?

Part 1 – translation

After I receive a text, I use translation software (SDL Trados) to work on the file (see below). This software breaks the text up into manageable segments. It also reformats the translated segments, so I don’t have to mess around with fonts, styles and footnotes in MS Word. I can also highlight specific jargon in the text. E.g. kulturni kapital (HR) – cultural capital (EN). This is then stored in a terminology database and suggested as an accurate translation next time I come across the same term. Note: the following screenshot is of a non-confidential translation.

Trados is also useful because it lets me store a databank of translated phrases: a translation memory. Translation memories are useful for academic translation but indispensable in legal and technical translation where there is even more repetition and an even greater need for conceptual precision.

If I am unsure of a word or phrase, I insert an author query, as the author often has a deeper knowledge, through reading in English, of subject-specific jargon and conventions. It is worth remembering that the translator’s knowledge and expertise is in language and general academic conventions, not in the minute details of each academic sub-discipline. This is why collaboration is so important! I base my schedule on (comfortably) translating 2000 words a day.

Part 2 – bilingual revision

After the first round is completed, I then crosscheck the translated segments in Trados with the original ones, correcting any errors. This is sometimes called translation revision. After the first round, some translated sentences are still ordered in a “Croatian” or “German” style. For example, quite a few Croatian sentences begin with the construction “Zbog toga” (lit. Because of that). This is comprehensible in English, but it sounds clumsy and unwieldy. While this is an obvious example, there are more subtle sentences that can pass through undetected. After this round, I send the text to the author to review. At this point, I stress that the author should focus on the precise translation of concepts, ideas and style: not on a literal, precise translation of every element in each sentence.

Part 3 – monolingual editing

Finally, assuming the text will be sent to a journal or academic publisher for preparation, after receiving the corrected text back from the author, I do a final edit (typically in MS Word) of the translated text for consistency and remaining syntax issues a week or more later. This break gives me a fresh view on the text. All the consistency issues that copyeditors focus on (e.g. -ize/-ise endings) can also be tackled at this stage. I use editing software: PerfectIt and The Editor’s Toolkit to ensure a high level of consistency. This is not achieved during the translation phase as it is impossible to focus on all the different levels of language at the same time. It can also be useful to have another round of copyediting/proofreading completed by someone else before final publication.