“Ask not what the computer can do for you. Ask what you with the aid of the computer can do for the betterment of mankind.” This piece of advice, published in a 1977 issue of New Scientist, is still relevant after 39 years because many times we have overestimated the capabilities of machine translators and underestimated the value humans add to this profession.
A Quick History of Machine Translation (MT)
The history of research in machine translation dates back to the late 1940s. The first experiments were made in 1949 by R. H. Richens and Dr. A. D. Booth on punched-card computers. They wanted to capture the “idea” behind English, French, German, and Russian words. However, their progress was soon checked by two limitations – the ambiguity of human languages and the limited processing power and memory of computers.
Problem One in MT: Ambiguity
Human languages are ambiguous. An innocuous sentence “When John met Dimitry, he hit him.” can be interpreted in two ways: “John hit Dimitry.” or “Dimitry hit John.” Although humans can figure out which is true from context, it is a difficult job for a computer – and with big consequences in a legal situation.
Problem Two in MT: Limited Computing Power
Computers are powerful today but in the 1940s researchers worked with 250-word dictionaries. A computer that could store 100,000 words was still a decade in future. So Dr. Booth and Richens’ experiments did not succeed in eliminating translators. However, the researchers did not give up.
But the Progress Continued
The next challenge came from Dr. Peter Toma of Georgetown University. He developed SYSTRAN, probably the most advanced machine translation system of the pre-Internet era. SYSTRAN was used by the U.S. Air Force, the Atomic Energy Commission in Tennessee, and the Euraatom Center in Italy. Advanced though it was, it couldn’t put human translators out of their job.
And Your Job Still… Remained Secure
Other ambitious projects–including the European Commission’s Eurotra, Moscow University’s BESM, and Hong Kong University’s Chinese University Language Translator (CULT)–led to important insights in computer science and linguistics which led hopeful pronouncements like “machine break the language barrier” as early as 1984 but the hype could not live up to reality. There are still 61,000 professional translators and interpreters in the U.S. alone and their ranks are set to swell by 17,500 in the next 10 years.
The Future is Bright
If you are a professional translator, you can take it for granted that machines aren’t going to put you out of your job any time soon, not even the recently announced Google’s Neural Machine Translation (GNMT) system. However, just because translation systems aren’t yet humans’ equal does not mean nothing will change.
If You Adapt
Translators work differently today from the professionals of 50 years ago. Back then, Internet, computer-aided translation, translation memory (TM) and many other technologies we take for granted were unheard of. In 2016, most agencies and many individual translators rely on them.
Learn to Use CAT Tools
Computer-aided translation is fast. Many clients take their use for guaranteed. You cannot afford to disappoint them. If you want to keep your job, either invest in a CAT tool or join an agency that does.
Surround Yourself with Professionals
Usually it’s the people who know the least about machine translation are the most worried. You don’t have to be like them. Educate yourself. There are several ways to go about it. But in our experience, the most efficient and usually the only one that works–is to join an agency and stay in touch with other professionals. That will give you a platform to tap into a vast pool of knowledge.
Machines can aid humans but they still can’t replace professional translators. That shouldn’t be a reason to behave as if nothing has changed. To stay relevant in the translation business, professionals can invest in computer-based tools and stay in touch with other translators.