Some people think that using machines for translating languages can be a cost-saving and effective solution for translation issues. Nonetheless, it is accurate only when the translated result is good, which machine translation still rarely produces.
How Machine Translation Works
The process of machinery translation starts by examining the structural makeup of each word or phrasal element in the source language, then dissecting them to synthesize them into the target language. The problem with this method is that it can affect the general quality of translation because languages have different structures and the majority of machine translators fail to recognize this.
Translations in Vital Situations
Sometimes having a translation app available on a smartphone can result in saving a life. For example, your car won’t go and you are in some northern country in winter. And the only way to not end up freezing at night is to convince some driver of a passing-by car who speaks another language to help you. That’s where an automated tool can be useful, but again – only to some extent. That’s because speech recognition is still a weak point of machine translators, especially with background noise (remember, you are on a windy road amidst a severe northern winter). And typing with freezing fingers is not the best solution.
Let’s look at some more situations. I’m short of breath and have luckily found a doctor who only speaks Russian. Or I’m in a courtroom in Moscow trying to protect myself against a lawsuit. The last thing I want is automated software shooting out awkward sentences that only marginally correspond to what I’m actually saying.
Behind the scenes of a doctor saving an immigrant’s life, you have a certified medical interpreter well-versed in both language and medicine who was absolutely crucial in bringing about that result. Importantly, these people are fluent in both Legalese and at least two languages – English and Russian.
Sometimes, accurate interpreting can literally mean the difference between life and death.
Industry Expertise
Anyone has to go through training to become a certified interpreter / translator. There are theories and techniques you have to learn, such as note-taking if you’re doing consecutive interpreting. Every interpreter puts in hundreds of hours of work into perfecting their craft, reviewing terminology in their field of expertise, making sure that they stay at the top of their game regarding industry developments.
Behind multinational conferences, you have a team of a dozen bedraggled conference interpreters who spent the past eight hours simultaneously interpreting from English to whatever language, in fields such as metaphysics, psychotherapy, or space navigation for the benefit of everyone in the room.
Cultural Differences
Further, people have cultural consciousness which machine translations don’t provide. Ultimately translations are intended for human recipients and one would find it hard to convince anybody other than another human being to understand and interact more closely with a particular target population.
A number of my friends are bilingual Russian – English. The problem in translation they explain between English and Russian is the cultural gap between the two languages. Much of their effort as translators is filling this gap between the speaker and listener. Some things that can be said in Russian are not translated into English. And the reverse is true as well.
Language Intricacies
Machines excel at limited and simple language like in business correspondence or weather reports. Nonetheless, in cases where the material is subtly connoted with a need for an agreeable voice to some extent, there is still a need for a human translator.
Words transmit ideas and feelings and to successfully capture those in the translation of a text can be surprisingly tricky for humans, but especially for machines. I.e.
Example. We often refer to a ‘warm’ person as someone who is loving and compassionate (vs. a ‘cold’ person) which makes no logical sense. That a warmer person (in temperature) would be any more compassionate than someone with a lower body temperature is nonsense. Instead, we’ve crafted this metaphor from the warm embrace of our mother and her love for us as children, making us equate warmth with love.
AI comes across one of these newly created metaphors – it will have no option but to take it literally and thus make an error in translation. And because language is a continuously evolving thing with new creative metaphors springing up every day, AI is always a bit late to the party. Just like a foreigner who doesn’t speak a language that well doesn’t get a metaphor and takes it literally.
Creativity
Indeed, machines are not creative beings. They are not capable of imagining things a programmer has not put into their systems. Humans can.
It means that robots fall short when dealing with texts that have artistic merit like a poem or a novel where there is a lot of ambiguity.
That is especially true for translating lyrics. The thing is that while versing poets not only use metaphors and other stylistic figures and literary tropes. They can also arrange words in an unusual manner – using conversions – to produce a certain emotional effect. These are only a few to mention.
What is more, can you imagine a machine that can both convey the intended meaning accurately and produce rhymed and rhythmed sentences simultaneously? Me either. So, to translate poetry, for example, to Russian, you’ll end up booking professional Russian translation services, choosing companies with a talented poet on their staff.
So, robots won’t replace translators of literature in the foreseeable future because of the artistic aspects.
Conclusions
Summing up, the bulk of people only need translation to understand an overall idea. They don’t need to go into the particulars, and a gist is enough. Machine translation is good for a quick-and-dirty peek. Human translation for everything else.
For non-real-time translation, AI is ready to replace a significant amount of human labour but their work needs to be checked and improved upon by expert humans. So, computers will absorb the scut work as the best humans only have their jobs made easier.
Simply put, automated tools would make the job of translating for speed and formality faster when it is needed. Nevertheless, the presence of machine tools does not mean that human specialists will become jobless soon. Foreign language skills will be in demand for decades to come, indeed. The world of language is so extensive and the amount of work that has to be put into translating and interpreting is so arduous and so human that we still need translators to fulfill a role that machines cannot play. The same is not only true for translations but also for other Russian Language Services.
