It’s a common theme. In [5, 10, 20] years, machine translation (MT) will be so good that there will be no human translators left. And, indeed, there are some trends that make this idea look tempting. The move towards statistical machine translation has allowed machines to learn from the texts they are given, allowing them to process at higher levels and produce more convincing results. But this won’t mean that they will replace humans, let’s see why.
The first reason that human translators will still have is that human language is slippery. Even if you were to compile a massive database (or “corpus”, to give it its technical name) of all the language used everywhere on the internet today, it would be out of date within 24 hours.
Why? Because as humans we love to play with, subvert and even break our own linguistic rules. Even people who hate languages love to make up new words and repurpose old ones. The biggest corpus in the world can only tell you how people used language yesterday, not how they are using it today and definitely not how they will use it tomorrow.
The basis of Statistical machine translation is that the way language has been used on previous occasions is a good guide as to how it should be used this time. Hence why Google Translate famously translated “le président des Etats-Unis” [the president of the United States] as “George W. Bush” months after President Obama was elected. The logic behind this decision is that if “George W. Bush” was used in that space enough times, it must mean that that phrase can be used all the time – a mistake that no human good human translator would ever make!
Add to this the fact that meanings of words change (something that has been mentioned elsewhere on this blog) and things look much worse for MT. It gets worse though, since language is bound so tightly to culture, “literal” translations are often incredibly misleading.
Here is a really simple example. In English, we have a set number of phrases we use to sign off a formal letter. We might use “Yours sincerely” or “Yours faithfully” or maybe “Kind regards”. In French, formal letter sign-offs are much longer and one of them might literally be translated as “Waiting for your response, I ask you to accept, Sir, the expression of my distinguished salutations”.
Now, statistical machine translation experts will rightly tell you that a good, trained package would not translate this literally but would look for an English equivalent. The problem is that the English “equivalent” would be different for different contexts and would involve looking much wider than MT normally looks. The decision here is linked to the context of the letter (specifically whether or not you know the name of the person you are sending it to) and not to language considerations themselves.
There are lots of translation decisions that are context-based like this one and it is in these kinds of decisions that MT will always flail around helplessly. It is in these kinds of context-based decisions that good human translators will always triumph.
So where might the future lead? Well, just as human translators are becoming more specialised, so will MT engines. Research presented at the recent IPCITI conference showed that there are ways that MT and precisely, post-edited MT can work. Perhaps one area where MT will work is in specialised fields, which use consistent language. Another view is that human translators will be called upon to make more use of their knowledge of the world, which adds justification to universities like Heriot-Watt who train their students in areas like international organisations and research skills alongside their technical training in translation and interpreting.
The future is bright, but the future certainly isn’t Machine Translation taking over completely from humans.
Excellent post Jonathan, you’ve hit the nail on the head!
I’d like to add a point on statistical machine translation, which I believe is the Google Translate’s way (as opposed to rule based machine translation).
I think the human brain is to a machine translation corpus what the ocean is to an aquarium. The point that many translators fail to understand is that statistical machine translation (SMT) does not translate. It does not work with meaning, but with statistics. It does not understand meaning. In fact, it does not “understand” anything at all.
So while algorithms have fewer chances to find uncommon translation patterns, they are likely to find highly recurrent patterns, whether we, human beings, find them literal or not.
From the statistical point of view, the meaning of the words that make up these patterns (plain texts, expressions, idioms, jokes, cultural references, etc.) simply does not exist. Nor does the evaluation of the given translation (literal, native, fluent, foreign, catchy, tacky, etc.).
What exists is the potential match of two strings of characters.
In fact, statistical machine translation can be very effective at translating cultural references, jokes or creative texts, but only if these have become HIGHLY RECURRENT PATTERNS –
That means that Google Translate is actually quite effective at translating clichés 😉
Good point, Pierre. This also might mean that when we start playing with those clichés, the MT would struggle as what we actrually doing is deliberately and forcefully breaking the structure. This also should give us pause when we think of the use of Translation Memories. If the author is deliberately playing with language, previous units might give us an idea of what they are doing but they certainly won’t give us the perfect answer as to how to translate what they are doing.
Absolutely. As technology develops, corpora are less time-bound, thanks to automatic specialized feeding for instance. The same goes for translation memories, which can be network shared or precontextualized. But machines will never be able to deal with meaning because as you perfectly put it, it is slippery by nature 😉
[…] It’s a common theme. In [5, 10, 20] years, machine translation (MT) will be so good that there will be no human translators left. And, indeed, there are some trends that make this idea look temptin… […]
I started as a translator at an aerospace company, some 31 years ago, after doing freelance translations for 5 years. One year into the job, they asked me to evaluate Systran. In order to ensure my neutrality during the evaluation, they guaranteed me that I would be moved to IT in case that this tool replaced all translators, as I was studying Computer Sciences. Thus, I was supposed not to fear a lay-off.
Well, you can imagine the result. I simply made a report of the machine-translated text, comparing it to a real human translation. I also used the tool to translate it first from Spanish to English, then back into Spanish. Nobody even at the company ever spoke about getting rid of the translators, though some idiot said that “in five years the technology will have solved this”. Yes, sure. That was 31 years ago.
Hi Jonathan, nice post. I support your arguments and would like to go one step further: not only will Machine Translation not take away any translator jobs, it will help create many new jobs; it will also help translators become more efficient and as a result will help them make more money. The amount of text that requires translation keeps increasing. Without MT, much of this text would never get translated. As MT systems get better, more and more texts will be translated, resulting in more opportunities for translators and editors. Translators should embrace MT rather than disdain it.
Wow, incredible blog layout! How long have you been blogging for?
you made blogging look easy. The overall look of your
site is wonderful, let alone the content!
Hi Bernice,
LifeinLINCS has been going a little over 2 years now. Some of our contributors have been blogging even longer than that. The design was created by the superb Colin Miller, of Heriot-Watt University. Thank you for your support.
I do a lot of work modifying language. That is, taking complicated English passages, some from a bygone era, and rendering them accessible to a target group that is composed of adults with a lower than national average command of English, and emerging users of English. This begs the question of whether the dial of MT can be set for disparate parameters, say in the (theroretically speaking) case of academic French translated for English primary schoolers. Having a MT option of working from high level to low level English would make my life much easier. However, I somehow feel that this would be more difficult to achieve than translating from one language to another.
[…] The never-ending debate continues and this article concludes, based on linguistic arguments, that machine translation is here to stay—but translators won’t lose their jobs if they embrace the new technology. (http://lifeinlincs.wordpress.com/2013/12/04/machine-translation-will-not-take-your-job-honest/) […]
[…] potential and limits of machine translation are a topic I have written about before. There is simply no sense in either dismissing machine translation (MT) as useless or pretending […]
When someone writes an article he/she reains the
image of a user in his/her brain that how a usr can understand it.
Thus that’s wwhy this piece of writing is outstdanding.
Thanks!