5 reasons not to fear machine translation

machine translation

Machine translation using computers and new technology is now something the general public can access: the leading internet search engines make machine translation tools available to their users, with varying degrees of success. At the same time, the combination of computer science with languages has led to the creation of a number of tools that combine word processing with machine translation.

The evolution of these technologies has accelerated rapidly in the past 20 years, even to the extent that it has many fear it will mean an end to translation as a professional activity. However, time has shown that machine translation will not replace human translation. And here’s why:

1. Language is subjective
Machine translation technologies are based on algorithms and artificial intelligence (AI) which are more suited to situations based on objective realities. However, naturally-developed human languages often operate on the subjective interpretation of different situations. Each language is unique to its place and culture, which only adds to its complexity.
Machine translation apps do not yet have the resources to detect the emotional content of a sentence, the cultural references, nuances in tone, accent and style, etc. They work quickly, efficiently and methodically and strictly apply the rules of grammar and the literal meaning of words and phrases, bringing a number of benefits to the translator’s workflow. In the real world, language is more fluid, and human translators have the natural aptitude to accurately interpret the intricacies of language, culture and context.

2. Language is constantly changing
Languages constantly change to incorporate concepts that previously did not exist as the languages evolve. A word or an idiomatic expression can acquire a completely different meaning in the course of 20 or 30 years, and often much quicker than that. Not only does the vocabulary change over time, so does the syntax and grammatical structure of languages. A historical analysis of any language will reveal changes to punctuation, in the use of suffixes, the emergence of new word contractions and abbreviations, etc.
Human translators are much more aware of these changes and their histories and can adapt to them, while “machines” need to be updated and programmed to incorporate these alterations.

3. There is no single machine translation system
A global machine translation system that can serve the whole world is an idea that remains in the world of science fiction. There are thousands of languages spoken in the world. While there are machine translation tools that work for hundreds of languages, much remains to be done. Moreover, the differences in semantic and grammatical structures between languages are often so great (often making it impossible to establish a direct correspondence between words and expressions) that one quickly comes to the conclusion that human translation and mediation are essential.

4. Nowadays, machine translation is the human translator’s ally
Recent developments in language technologies and machine translation have shown that the panic over the “invasion of the machines” was not justified. An analysis published in Wired in 2017 said that artificial intelligence, neural network technologies and data analysis algorithms did, in many cases help increase the productivity of translators. This is because while they have not completely replaced human translators, they have automated many language and writing processes, eliminating many repetitive steps. With the post revision of material generated by machine translation, translators have been able to do more in the same amount of time.

5. Machine translation is not always the best solution
Machine translation tools are very useful when working on technical manuals, software and application interfaces, web pages, or any support containing concise information that is logically structured and which is frequently repetitive.
However, most texts are more creative and fluid in their use of language. Machine translation lacks the human ability to assess social norms and cultural context (which is also changing constantly), and nor can it evaluate the specific goals of the text.
This is not true only in the world of arts, literature and poetry, it is also evident in other technical and professional texts: advertising, marketing and tourism depend on creativity and spontaneity in the use of language; medicine, engineering and computer sciences constantly teem with new ideas and expressions; in the legal profession and institutional communications (for example, within international, organisations, government and diplomacy). texts requiring translation often deal with topics that are so complex that they cannot be left to algorithms with no human involvement.