Past, present and future of translation
Yesterday, changing the dial of the radio fleeing from advertising, I arrived at BBC Radio 2 where they argued that we live in a world of constant and rapid change, where nothing remains stable and where things arrive and then leave. This reflection was made in relation to music, but is directly transferable to many professional areas. I am sure that if you think about your own profession and sector you will feel immediately identified, even more so if you work in a sector that did not exist 10 or 20 years ago.
Índice de contenidos
Index of contents
Index du contenu
Many professions are dying with the digital revolution, just as many died with the industrial revolution. While I was researching to write this entry I found this information: according to a Wikipedia entry, the last typewriter factory was closed in 2011 in India. What would have been the professional translator's preferred working tool for decades had to give way to computers and word processors.
In the professional work of today's translators, we are living a phase of change that may be similar to the passage from the typewriter to the computer, where many translators resist the changes that are and will eventually be implemented, because it is efficiency and productivity that will drive those who do not adapt out of the market.
One of the technologies that revolutionized the way of working in professional translation, especially in the field of technical translation and information technology, was the use of translation memories, i.e. computer programs that offer advanced management of human translation databases. The term used to describe this type of program is itself a marketing slogan. Imagine what a translator would have to do to remember the translation of a technical term from a text translated a year ago before the digital revolution. The translator would have to look at the printed text, turning pages and pages, hoping to find the previous translation. In the digital world and with translation memory software, the translator is a few clicks away to consult any word that has been previously translated.
Life was better in the old days. This phrase seems to summarize the desire that we all have to continue doing things as we did, to continue in our comfort zone. It still surprises me today that there are translators within the field of technical translation, and translating technical manuals, who have been able to survive the advent of translation memories without using them; a technology that began to be used successfully as early as the 1990s. As with the use of word processors with respect to typewriters, the use of translation memories made the translator's job more productive and efficient. If we imagine the translation of a technical manual or a product catalogue, where there may be a high degree of repetition or similar text, we can clearly see how productivity and translation quality can be improved.
You will hear very few translators who speak well of machine translation. Undoubtedly, most translators feel threatened by machine translation and their immediate reflex is to say that one cannot trust machine translation and the very serious mistakes it makes. It is true, machine translation is not 100% reliable today, nor do I think it will ever be, but increasingly machine translation is a technological tool that helps the translator to do a better and more productive job.
I remember my first translation jobs when I was still a student at the faculty of translation, when I got a job from a translation agency in Madrid that sent me the automatically translated text. I had to go beyond my 20-year vision of machine translation because of the first contact I had with it. That translation agency intended to reduce their fee they paid to translators by offering machine translation, but it was an absolutely literal translation and meaningless in most sentences, so it was more productive to translate everything from scratch than to try to fix that mess.
But machine translation has evolved. The programs that existed 20 years ago were based on grammar rules and dictionaries. This was followed by statistical translation based on metadata, which was an important qualitative step forward. For years now, large multinational companies in the technology sector have been using machine translation successfully in order to reduce their costs. These companies have the right conditions to make use of machine translation as they have large volumes of text translated within a specific field and technical texts that are already produced thinking to be easy to translate for the machine. And because they have the IT infrastructure capable of processing such a large amount of data.
Thelatest paradigm that has revolutionised machine translation is neural networks and machine learning. This new paradigm also requires very powerful computers that are capable of serving themselves and analysing large databases, but has led to a significant improvement in the quality of the translations obtained, according to various comparative studies. In fact, the statistical machine translation engine (called Moses), on which most of the developed statistical systems are based, announced in 2017 that it would launch its latest version that year, which seems to confirm the hegemony of this new paradigm.
With this new paradigm, a new actor(DeepL) has emerged on the international scene, as we announced in November 2017, which has achieved automatic translations of higher quality than the omnipotent Google Translator, which seems to demonstrate that it is not only a question of data (no one can compete in this sense with a technological giant), but that programming based on this paradigm leaves room for human ingenuity to overcome the storage and processing capacity of mega-computers.
There are many technological aspects that have progressively changed the way translators work, but undoubtedly the most significant is the advent of post-editing. This is a term that in the jargon of the translation company and professional translators refers to the editing of an automatic translation. As I mentioned earlier, many large companies, especially in the IT sector, have been post-editing for years now, but the improvement in generic machine translation systems, available now to translation companies and professional translators, and the substantial improvement in these, means that, in certain types of translations and language combinations, post-editing has great results and increases productivity. There are already professional translators whose preferred way of working is post-editing, but there is still a great deal of resistance from most translators.
If we look at the past of translation, we can venture to say that its future is called post-editing. In 20 years' time, most of the translations in the technical and legal fields will be post-editings, for the simple reason that it is such a great help to the task of translating that it will end up being imposed.
Josh Gambin holds a 5-year degree in Biology from the University of Valencia (Spain) and a 4-year degree in Translation and Interpreting from the University of Granada (Spain). He has worked as a freelance translator, in-house translator, desktop publisher and project manager. From 2002, he is a founding member of AbroadLlink and currently works as Marketing and Sales Manager.