Machine Translation Vs. Human Translator

If translators do not become technologically savvy, they may become professionally obsolete.

Machine Translation Vs. Human Translator

(This article was originally published in Linkedin on June 2nd, 2014) 


I am an old-time translator. Real old-timer. From that time in history when we still were using IBM typewriters. I would say that 30% of my time was spent typing and re-typing. You typed a draft of, lets say, 50 pages. You then corrected it by hand. Page full of garbled signs and arrows, scratched out words and additions, margins full of notes, back of page with more handwriting. Second draft typed to incorporate all those changes. Again read, this time probably for grammar and spelling. More changes. Type again. Review one last time. Type again. Not only that, the other 30% of the time you were consulting huge dictionaries and reference materials (I used to have, literally, piles of books all over the house on every subject matter and in different languages). And to seek the answers you did not find in those books, you had to physically go out of your house to the library and spend hours and hours of research to find whatever it was you were looking for. Or find the expert and try to get them on the phone or in person. The actual process of translation would therefore be 30% of my time at best.

In the very early 80's, I was one of the first translators to use the Apple II-e professionally. At that time, most of the translators I knew refused to learn to use computers because that "thing" would "take away" from the process of translation (I don't even remember what they argued it would take away... but I remember thinking that concept was totally out of touch with reality). I later worked with an Apple IIc and a Macintosh and was again criticized heavily for using technology in a field where we considered ourselves "intellectuals" and "artists".

In the late 80's I had the privilege of working with some professors in the faculty of law of a Latin-American university during their testing of the internet for academia in coordination with the National Science Foundation. That gave me a unique perspective on what was to become one of the largest revolutions in communications in the history of mankind. If you think of it, by 1989, the entire world was still divided by geography. Twenty-five years later, that is no longer the case. Just one example. In 1989 if you wanted to study anything, you had to register to physically go to class at a pre-set schedule and in very structured conditions - dictated by the university. Now, just 25 years later, I can take classes on my computer from universities literally located on the other side of the planet, and connect to them asynchronously 24/7 under conditions I set for myself.

Again, when I started to use the internet for business in the early 90's, I was criticized for using technology for a profession that was "above technology". By now, translators had to learn to use computers as working from typewriters would prove to be less efficient. And soon enough would be required to use the internet to send and receive assignments. Later, as a source of consultation of dictionaries and glossaries, and more recently a one-stop research center with the entire knowledge of the world at the tip of our fingers, literally.

The advent of CAT tools was similarly received with staunch repugnance by a very large number of translators, to the point that many, even today, do not use them in their every day work. In many countries, CAT tools are not yet the standard for the industry.

The advent of Machine Translation has started another movement of fierce opposition from many translators who believe that only human are able to produce good quality translation. And again, I think translators who oppose MT are, once again, on the wrong side of history. Just as typewriters were replaced by computers, as big black heavy dictionaries were replaced by online glossaries, just as library cards were replaced by Google search, such like repetitive tasks were replaced by CAT tools, so will a very large percentage of primary translation efforts be replaced by MT. I truly believe that in the span of the next 5 years, most translation work for any translator anywhere in the world will begin as a PEMT task. What I am saying is that that the original task of taking a blank sheet of paper (or screen) and typing in another language a first draft of the content in another language, at least that first step, will be fully automated.

It is my prediction that with the exception maybe of good literature and poetry, most everything else will be first fed to a sophisticated software/machine that will spit out in seconds a preliminary version for post editing by a human. Or not. That is the trick. Because a preliminary version may be all the client wants. The "gist" may be all the client wants. So, they might go to Google translate (or its equivalents), which in a world of very fast technological advancements might actually be pretty good in the near future.

Of course for many others, the Gist will not be enough. That is where the client will be willing to pay for a professional to EDIT that Machine Translation. And there goes my advise to the translator community as a whole. If translators do not learn very fast and soon to become PEMT, others will. Remember that to be a PEMT you really do not need to be that good of a translator. Actually you just need to be a very good writer and editor of machine translation. But some translators oppose the idea of becoming technologically savvy. And technologically savvy means learning a whole new array of things that the translator of the 20th century did not need to know. Just as at some point they thought they did not need to learn to use computers. Or they thought they did not need to learn how to productively use the internet. Or how to use CAT tools.

Becoming a PEMT requires an entirely new set of skills. If you are a very good legal translator and you want to become a very good healthcare translator, the first thing you have to learn is the healthcare terminology. Well, the same happens with PEMT, with the added hurdle that PEMT can only be learned by DOING. By PRACTICE PRACTICE PRACTICE. And practice takes time. It is a skill that you develop. There is no intellectual shortcut. Your fingers need to learn how to carry out the tasks that you need. Your brain needs to learn to hit the connections it requires to go from here to there.

I am very afraid that many translators will loose very good possibilities of working in the translation industry of the future simply because they lack the technical skills to compete with others. We have seen this happen in many other industries, including medicine. Doctors HAD to learn to use computers (at least in most of the Western world) to be able to practice medicine. Obviously the computer did not replace the doctor, but the doctor who does not know how to use computers will be almost unable to function in the modern world (unless he/she has the money to pay another person to do the "computer stuff").

So, I am begging translators to look seriously into PEMT and all the technologies that are coming out of the Machine Translation advancements. There are many new fields that are opening for translators who jump in the bandwagon of technology early on. I do not believe that the MT industry is going to stop because translators do not want to join in it. If we do not join, the MT industry will find other people to replace us. Good bilingual writers who are technologically savvy. Good bilingual editors who are technologically savvy.

The translation industry is at a crossroads. The "big bucks" are out because there is a gigantic market out there. Yes, it has taken over 50 years for machine translation to become commercially viable. But now there are thousands of millions of dollars in investments to make it work. Those investors do not really care who is going to do the post editing or who is going to work on the corpora or who is going to do what. But translators should care. And if translators do not have the technological skills to drive the car of the future of translation, it will be driven by someone else.

A few months ago I asked the technology industry: Will You Help Me Flip It? <> and very shortly afterwards I have seen some specific courses announced for translators to receive the training they need. I believe translators should actively start looking into this field, what is it that you need, where can you get the training you need, and how fast can you become a translator of the 21st Century.

But, the most important thing for me is to create awareness of the need for a change in ATTITUDE by the translation community in general and by some of the translation industry leaders in particular. This is an issue that can make us or break us as a profession, and it is not receiving the attention it deserves. Even if just opening the big debate and making translators aware of their possibilities (or what they will miss out in the future).

Will You Help Me Flip It?

Claudia Brauer

P.S.: Visit a wiki with free information on the subject : BrauerTraining Wiki