Rumors have been in the air for a long time that machine translation will be able to replace the human translator, and sometimes statements like “Human and Google Neural Machine translations are almost indistinguishable” are heard when Google announced the launch of a neural machine translation system (GNMT). Of course, recently neural networks have made a huge step in their development and are becoming more and more integral to everyday life, but is artificial intelligence really so firmly established in the translation arena that it can replace a person?
Yes, time does not stand still. The processes of globalization connect peoples, regions, cities and countries into a single network, where everyone can get information located on another part of the globe (of course, if they paid for the Internet). People are increasingly fascinated by foreign culture, traditions, and in particular, literature, moreover, in the original language; as a rule, people accept such information already processed and translated into an understandable language by the relevant communities or groups of people, publics or news sites. But it also happens that the information gets in its original form, like some volume in the original language, but the trouble is that a person does not always have this volume translated (there is so much new literature that you don’t have time to translate everything, and they translate it into the first turn popular works), and he himself does not have the skills to read and understand what is written in the book. And here it has several ways: wait for an official translation (and if the work is not popular, then it will take a long time), wait for an amateur one (yes, there are such daredevils who undertake such work) or use improvised means, like the same Google translate.
The first two ways are similar, because you rely on human labor, even if the second one is a little more doubtful, but not every official translator is good, so let's conditionally combine it into one. The second way, the way is much less suitable, although some people are already ready to perceive it as a finished and final product, and this poses a greater threat than the qualities of the machine translator itself, which is convenient as a tool designed to facilitate the translator’s routine work, but no more Togo. And in order not to succumb to this "enemy", which is supported, first of all, by people who are condescending to the quality of translation, one must follow the following principles, which will be outlined below.
1. You translate the meaning of the text, not the words. I do not understand - I do not translate
The machine operates according to algorithms. And these are really complex cross-language algorithms using dictionaries and grammar rules, we must pay tribute to this. But! Text translation is not just a translation of words from one language to another, but a much more complex process. A significant drawback of a machine translator is that it cannot understand the meaning of the text.
Therefore, Human Translator, develop your knowledge of the target language, up to the level of catchphrases, proverbs and sayings, phraseological units. Meaning is the main thing and the first thing you have to learn from the text!
2. Learn your native, dear to your heart, native, great and mighty Russian language. The translation must fully comply with the norms of the language into which the translation is being carried out, in our case, Russian
Yes, I think this point is as important as the knowledge of the foreign language from which the translation is being carried out. There are frequent cases when people who take up the profession of an interpreter make mistakes in their own... When disorder and chaos reign in your home, how can you go to someone else's home and teach its owners order? That's right, no way.
In general, I am a supporter of domestication in the translation strategy, and therefore I think that any attempts to represent cultural differences in the text itself through means that are not typical for the Russian language are local forms of *-mania, where instead of an asterisk, you can substitute, for example, gallo- or English-, and so on. Of course, a certain range of words, such as country-specific titles (wali, shah, king, etc.), ways of addressing (mister, sir, lord) can be transformed, but this would be unwise.
Love your language. Treasure them.
And so that experts do not argue about the preservation of the cultural characteristics of the text, the main thing is that the text contains its plot, characters, emotions and meanings, but the cultural environment can be comprehended in other ways, for example, by learning the original language. And then a translator is needed to translate the text into a format accessible to the reader, that is, into the native language.
3. Don't be afraid to transform foreign text
I will not delve into the theory of translation, but there are a number of specific translation transformations of the text. Additional elements can be added to the translated text, omitted, moved - everything is determined based on the analysis of the translated text, but it also implies a good native base. By the way, in this it is precisely the machine translator that lags far behind the human translator. The machine translates "as is", and the person can decide "what's best" and act accordingly.
4. Well, 4th, be patient and diligent
For the translation of a text is a very hard work that requires a lot of effort and time, as well as knowledge, a broad outlook and the ability to adapt.
As for me, I translate from Japanese, and this guarantees me a number of additional barriers, and it does not make life easier for a machine translator, because pattern recognition for Eastern languages plays a very important role. But during the time that I translate foreign texts, I have developed for myself the above four principles, which make the translation a Translation, and not a simple tracing paper from a foreign text, and which, in my opinion, are unique in any case, be it Japanese or English, For example.
And, summing up, after all, in what exactly is the translator not inferior to the machine?
A person will not yield to a machine translator in the ability to understand what is not obvious, Meaning. The machine understands words, word combinations, grammar, vocabulary, sometimes distinguishes between homonyms, but it will definitely not understand the meaning as something integral to the text in the near future. But a person, in order to understand the Meaning of the text, must be proficient in his native language, and the reader should take into account that the result of machine translation can be very far from the true meaning of the text.
Everything else, I believe, does not go beyond ordinary knowledge.
Source: habr.com