Language is much more than simple words: it’s the way people use to communicate and express their feelings, and it brings a lot of history and culture.
That’s why translation doesn’t just involve a mere change of terms from source to target, but it is rather about conveying the same message and evoking the same emotions.
Given that machines are both heartless and uncultured, it takes human translators to fulfill this task at best.
Even though at times MT can be effective, as it can help people to understand the general meaning of a text, it often sounds unnatural and mechanic: translation is also a matter of choices and there can be plenty of ways to deliver the same concept, nevertheless, a machine can’t distinguish among them.
When translating, there are lots of factors to take into account, such as audience, circumstance, purpose and cultural background. MT is unable to do so, unlike human translators, who have an overall vision of the language and its aspects. In fact, they are aware of different linguistic styles and how to adjust them according to whom they address.
MT usually provides word-for-word translations, which can lead to serious mistakes, as a single term can have different meanings depending on the context. A human translator generally knows that the connotation of the same unit of expression may vary and is able to adapt it on the basis of the surrounding text.
Translations are used achieve various purposes and reach various individuals and we are able to play with words in order to meet diverse expectations: if the goal is selling a product, they can recreate slogans, often using different terms from the ones shown in the source language, or even building completely different sentences, maintaining the same content with the equivalent degree of effectiveness. This is what linguists call transcreation, a mixture of translation and creativity, which is a peculiarity of humans: machines can’t be creative and this explains why MT doesn’t work effectively for these kind of assignments.
Another hard nut to crack for MT is idioms and irony: just like for slogans, it is often necessary to change completely the literal meaning from source to target, in order to keep the real essence of the proposition. A literal translation, in most cases, wouldn’t make any sense in the target language, the majority of proverbs have cultural roots and they can’t usually be expressed through the same words.
In view of the above, although Machine Translation can be very helpful in some contexts, it still has lots of limits which are quite difficult to overcome, therefore it can’t clearly replace human translation.
