Why machine translation isn’t enough

Lately, machine translation has given us a lot to talk about. New inventions, improvements in the applications we already know…
Without a doubt, these extraordinary breakthroughs will, if they haven’t already, result in great improvement in the practice of translation. However, will they ever be able to completely replace the human translator? Can a machine handle all aspects of a specific translation job?

Machine translation

We commonly call machine translation those software programs that translate a text from one language to another. These systems do not rely on human help and/or proofreading, and they are based on a series of linguistic or statistical rules.

Without any doubt, the most popular machine translation is Google Translate, although there are others, such as Reverso or DeepL, that will soon start to sound familiar. The latter, created by Linguee, is one of the most recent (launched in 2017) and complete ones.

 

Traducir, Translate, teclado, tecla

 

 

Machine translation v.s professional translation

In light of these developments, some people suggest that machine translation is already capable of completely replacing the role of the translator. Nothing is further from the truth.
As we have already mentioned, these complementary tools are, indeed, becoming more and more effective… although they are still complementary. The human factor remains key throughout the whole translation process to ensure that the function and the intentions of the original text are respected.

The work of a translator goes beyond replacing words with equivalents. A professional translator should be able to convert the functionality of the text and the linguistic elements, as well as the culture. If these elements are ignored, the consequences could be fatal.

If you don’t believe me, here are some examples of misunderstandings created by machine translation tools:

 

This extract from a Hungarian article has been translated by Google Translate. The English version, as we can see, has no grammatical and/or lexical meaning.

 


 

We have to be especially careful when translating national or regional expressions. In the picture, the machine translation DeepL tries to translate a Spanish expression that means “to be distracted”. It is never a good option to literally translate expressions.

 

Here is an example from Reverso.
In this sentence, the word “venga” (a conjugation of the verb “to come“) refers to “el señor Uribe” (“Mr. Uribe“). It is, therefore, a conjugation of the pronoun “you“. Spanish, however, has a tendency to omit pronouns. A translator, based on his training, his knowledge of this language and the context of the sentence, would have understood. A machine translation, no. That is why Reverso changes the meaning of the sentence, referring to a third person instead of “Mr. Uribe“.

 

 

At Eazylang, we believe that nothing can compete with a quality translation produced by a trained professional. Every translation is a delicate process and requires personal attention. It is the only way to ensure that the message of the source text is transmitted perfectly in the target text.
For a quality translation, visit www.eazylang.com 😉.

One Response to “Why machine translation isn’t enough

  • Interesting article. May I offer a curious example in Ecuador “UNASUR en la Mitad del Mundo” “UNASUR in the Half of the World. Another one “have a fine day” “Que tengas un día de multa” Paul Williams

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: