Contrary to popular belief, the life of a translator or an interpreter is not a bed of roses. Continuous quest for assignments, arduous work in front of the computer, lots of pressure, deadlines… The truth is, linguists or LSPs (language service providers) are indispensable in the multilingual EU. Yet the demand for language professionals is shrinking as more and more newly-minted graduates enter the job market and employers (the most prominent of them being the EU institutions) choose to invest in well-seasoned and qualified employees. On the other side of the spectrum young linguists have to reconcile their skills acquired during the traditional training at university with the rise of new technologies. Understanding them is the key to new opportunities.
Technology on the rise
The long-awaited arrival of the new real-time Skype translator has not escaped the attention of many Internet users. It is a true breakthrough as this tool is about to offer a vast and an uninhibited access to machine interpretation. According to MIT Technology Review and BBC, the Skype translator clearly cannot offer a human-like quality of translation, yet it definitely offers a sufficient rough translation which provides the gist of the source text and has some astounding features like excellent speech recognition skills. A quantum leap forward as regards machine translation (MT) was also taken by Microsoft, when its chief research officer Rick Rashid delivered a speech in English at a conference in Tianjin in October 2012. His presentation was translated live into Mandarin, first as subtitles on overhead video screens and, then, as computer-generated voice. It is worth remarking that this version reflected the characteristic tones of Mr Rashid's own voice. Even earlier, namely in 2009, an indispensable companion of each consecutive/simultaneous interpreter was born: the digital pen (smartpen). It consists of a microphone, a built-in speaker, 3D recording headsets, and an infrared camera. A normal ink cartridge allows to take traditional notes on a micro-chipped paper which captures data and sends it to the computer. In case we forget a word for “chocolate” in German, it suffices to “whisper” it to the pen in order to hear the answer: “Schokolade”. With such discoveries the traditional role of a linguist is definitely changing, nonetheless human interpreters will always be needed when human life is at stake.
Man with a machine, not against it
Last year's proceedings of the XXth Congress of the International Federation of Translators (FIT) in Berlin revealed the feelings of over 1,500 translators and interpreters from over 70 countries across the world: it is no longer viable to speak of the newest trends in terms of man vs. machine. In fact, we should rather think of this paradigm as man with a machine, pointing to the supporting and complementary role of MT software. And this is very much true: nowadays, professional translators are rarely working alone, both because of the low cost-effectiveness of manual translation, and because of the increasingly frequent demands from translation agencies and end clients to use computer-assisted translation (CAT) tools. Successful and effective implementation of MT software can be best illustrated on the example of the EU institutions. Currently, EU translators are using a number of CAT tools, MT interfaces and terminology databases. And what about pure machine translation? It's on the rise as well! EU translators are free to use it and then they can post-edit the end product. This is reported to save about 30% of the working time. Actually, MT services in the EU operate on a similar principle as Google Translate (statistical machine engine operating on a probabilistic principle and using large corpora of texts). There is only one difference: MT corpora in the EU consist exclusively of high-quality human translations. One of the recent MT projects within the EU, which is called MT@EC, is specifically fine-tuned for EU documents and guarantees high security for confidential documents, which is also a contentious issue as more and more translators work in the cloud.
How does it translate to job opportunities?
Machine translation may seem to be taking jobs away from linguists. But that’s not quite the case. In fact, it can be generating new jobs if applied properly. The job market is constantly transforming, and linguists who keep clinging to classic methods will gradually find it difficult to keep up. Going with the current may be a good idea as many linguists turn to new disciplines such as digital humanities or fascinating interdisciplinary sciences like computational psycholinguistics, speech processing and deep learning or data mining. Machine translation software is only a response towards the rising demand for high volumes of translations, which no human being could manage. Still, complex information systems and quality assurance processes require competent specialists who analyse the output, suggest improvements and supervise the overall process.
Find your niche!
It's easier said than done. Nowadays most graduates in languages follow the same path: exchange programs, scholarships, additional courses, qualifications, memberships, conferences, a doctoral degree... Without doubt, the challenges on the job market require lots of creativity and a certain degree of perseverance from young linguists. Still, it's a pity that few universities offer courses in MT and raise awareness of its importance among young students. Classic academic curricula for linguistic training should be restructured in order to offer greater flexibility to undergraduates. This is one of the pillars of my current research.
All in all, the concerns related to the rise of MT can be compared to the fears surrounding the introduction of the steam engine, which replaced manual work. The question is not: should we be afraid? The question is: will we use it to our good in order to improve the efficiency of our work or will we let it disparage the work of translators, which will inevitably lead to the deterioration of translation quality on the market?