Machines vs. Humans: Are people even needed for translation anymore?

This question has been circulating for the past decade, once machine translation really took off. It seems that tools like Google Translate and DeepL have come quite far since the first attempt at machine translation in the 1950s. Nevertheless, classical translators are still needed. In the following, we will show you the advantages of human translators.

At the beginning of machine translation, texts were translated word for word (without regard to the fact that many words have multiple meanings) and sentences were strung together in the same grammatical format as the source language. It is hard to call this process translation because the output didn’t make much sense to a target text reader. Yet after decades of work on machine translation, and many different ideas on how to go about the process, machine translation isn’t actually that inaccurate nowadays. If you input an entire paragraph in Google Translate, the syntax and word order of the output will mostly make sense in the target language. There might be a few word choices that don’t quite fit, but overall the translation is readable and sometimes better than what a rookie translator could produce.

So, why don’t we cut out the middleman translators and just use machine translation for everything?

For companies doing business internationally, there is no doubt that it would save them time and money to be able to use a machine to translate the material that they want to use in another country. Yet the quality of their texts would suffer if they went through this route.

Why?

The main reason that machine translation doesn’t always work is because of CULTURE. Texts that are written in one cultural context will sometimes not make any sense in another cultural context, even if the word order and grammar are absolutely perfect in the target text language. What machine translation is unable to understand is culturally specific language. For example, it won’t be able to detect if a sentence from the source text will sound vulgar or inappropriate in the target culture or if the target text readers will recognize a celebrity reference who is famous in the source language culture. Idiomatic expressions, cultural-based examples, and word connotations cannot be learned by a machine.

In English for example, the three words meticulous, selective, and picky all mean about the same thing, but each word has a different connotation. Being meticulous is positive. It means that you are very careful and precise when choosing something because you want the best end result. Being selective is neutral. It means that you are simply careful when making a choice. Being picky is negative. It means that you are hard to please and are fussy when making choices. These slight differences are hard to distinguish, which is why machine translation tends to lose certain aspects pertaining to connotation when producing target texts.

Other disadvantages regarding machine translation

There is also an aspect of the audience to which machine translation cannot adapt. For example, when translating an informational pamphlet for what to do after surgery for patients who may not have a high school education, the text cannot be filled with technical jargon or the patients will not understand what they need to do after surgery. In translation, stylistic changes need to be made according to the audience. Only humans will have the ability to translate texts for the right audience and context.

So, when talking about machine translation and the idea that in a couple of years translators won’t have jobs, we must remember that culture and audience play a big role in translation and that machines can’t correctly transfer these concepts into the target languages sometimes. Localisation and post-editing machine translation are the next big things for translators. Localising content means adapting a text for a certain country or region so that the content is relative to target readers. This may mean changing culture-specific references or statistics that will make sense in the target culture or simply adapting the level of formality according to the audience. Post-editing machine translation is taking a text that was put through a machine translation tool and editing what doesn’t work or what doesn’t make sense so that the text is coherent and adapted for the target audience.

It is true that with new technologies, the translation field is changing, and translators must adapt if they want to continue in this profession, but humans will still be needed in the translation process for many reasons. Machines won’t be taking over this domain anytime soon. It is therefore important to consult experts so that they can either do the entire translation or later only make cultural adjustments and proofread the already machine-translated text.

For further information, please contact us by e-mail at contact@smylingua.com or call us on +33 1 76 43 32 76.

 

For more information you are welcome to read our other articles about 

Subscribe to our Newsletter and stay up to date on international news with Smylingua 🙂

* indicates required

/( mm / dd )

Cette publication est également disponible en : English

CALL ME
+
Call me!