Machine Translation

Long time no see, being a MIIS student does not allow for much leisure time.

I had my first real contact with the facts of MT a month ago and I must admit I loved it. Not bad for a "pre-technological" being like me. I also just created language teaching lessons using social media and feel and feel so proud of myself! But that's something I'll comment in another post (I also loved it).

Here are my thoughts about the future of MT, an essay I had to turn in for my Advanced CAT tools course at MIIS, taught by the amazing Adam Wooten.

I am fully convinced that MT will play a good part in the future of translation. My guess is that it will end merging with CAT tools (something like a much more advanced match or a better MT suggestion plug/ feature).

Still, no matter how good the MT product is, there will always be a need for a human editor/ reviewer who has to be a translator as they need to be able to grasp the nuances of both the source and target texts.

This does not have to carry negative implications for translators (human translator doomsday brought by MT): things will be different, easier in some ways, more complex in others.

I think MT can end up providing both the more basic, as in limited, version it is already offering (needing controlled language, the more controlled the better), and a more advanced product (not that restricted and higher quality). It is just a question of time; it also depends on the development of other technologies. Until then, if anything, MT, even without human revision, is the answer for those offering “peanuts” to human translators. By “peanuts” I mean those rates that could be defined as both hilarious and really depressing at the same time, we have all received one of those offers that already costs money for the time invested in reading the email ($0.012/ word and alike). So yes, MT is already doing us a favor, or should be, at least.

The MT engine project made me realize how good MT can be if we abide by the controlled language rules and the many applications it can have. Then, it is not hard to imagine we can work in broadening the borders of that controlled language until it is less limited. However, reaching the goal of unrestricted text might take some more time and effort, but I do believe it is possible if combined with other technological developments.

That said, I do believe some specific areas such as advertising, dubbing, literature and poetry are not candidates for MT, and I do not think they will ever be. However, it may be used for subtitles, for example, so MT will be used along with human translation in different fields and in different ways.

Also, the possibility of integrating MT and interpretation opens up a whole new world for interpreters and could make our jobs a little bit less stressing without taking away any of the fun.

As I see it, MT and all of its applications and potential developments are just the logical evolution of translation. All other aspects of the sciences are changing and evolving, translation, which I perceive as a hybrid of science and arts/ humanities, should not be an exception.