How DeepL is your love?

Some time ago I wrote in a magazine article that you should “stop feeding DeepL and Google with your content and intellectual property”. Harsh words, I admit, and they warrant further explanation because it's a bit more complicated than that.

These days, it seems everyone's a language expert. Actually, I don't think it's ever been any different. Everyone has their own idea of "good" English or German or whatever, so setting yourself up as a language professional is like managing your national football team: Whatever you say, there will be a clamour of differing opinions, not all of them informed. And, let's be honest, nearly all of us feel a sense of indignation when our language usage is questioned. But working under pressure or redrafting complex passages can, and does, lead to the occasional slip-up. After all, we are not machines.

How lucky it is, then, that there are now machines that not only check what we've written for spelling and punctuation but actually translate whole swathes of text into the desired language.

DeepL, Google Translate, Microsoft Translator and the like offer tools that provide instant translations free of charge. Anyone can enter their source text into one of these engines and come up with a translation that, depending on the subject and target group, may already be fit for purpose. However, with neural MT packages there is an increased risk of error, since they often fail to recognise context so that a word like the German "springen" comes out as "show jumping" in a text about skiing. No matter how many sources you connect to your source text the translation is likely to need a fair amount of tidying up before it's ready to use. Consequently, we need quality control tools with a substantial terminology database in the background.

In view of the available technology, it's natural that more and more people are producing copy or translations in a second language. Of course, we as language professionals would prefer to handle the complete writing or translating process, but if we are offered a job that has already been started in the target language (either as copy or as a translation, possibly created using MT), we are happy to help you arrive at a top-quality result.

A further point to bear in mind is that content fed into free versions of MT engines is generally stored and may be made available to others. If data security is an issue, use of free MT tools should be considered very carefully. Nevertheless, MT and CAT are now vital elements in a language professional's toolkit since they can significantly expedite translation turnaround, optimise term consistency and reduce costs.

At WORDWORKS, we like to use systems that we have developed together with you. These generally consist of licensed machine translation services in a controlled translation environment with CAT tool integration, so the source text and translation are saved in our database and not on a third-party server. No sensitive content leaves your company, ensuring that your ideas, strategies and commercial assets are protected. Additionally, a terminology database is built up over time that is precisely tailored to your products or services and your CI. Although such a workflow will not be suitable in every case it does offer the type of quality control mentioned above.

It's also important to remember what translation actually is, or rather, what it isn't. It is not simply replacing one word with another. Tone, emotion and localisation are crucial elements. Your English copy is your international shop window: it goes without saying that the grammar, spelling and syntax should be error-free, but it should also sound natural to other speakers of English and display the creativity and (cultural) sensitivity that only human translators can provide.

As billionaire hedge fund manager and Bridgewater Associates founder Ray Dalio puts it in his book “Principles”: “While the computer was much better than our brains in many ways, it didn’t have the imagination, understanding and logic that we did. That’s why our brains working with the computer made such a good partnership.”