Machine Translation will not take your job, honest!

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(or “corpus”, to give it its technical name) of all the language used everywhere on the internet today, it would be out of date within 24 hours.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.

Putting a Smile on The Public Face of Languages

If you are a translator, interpreter, sociolinguist, anthropological linguist or any other kind of linguist, there is a lot to get annoyed about. Courses are closing, rates are (in places) dropping, respect is on the wane and hardly a day goes by without some newspaper publishing a story about some new gadget that will entirely automate some way we use language. If we didn’t know better, we would say we were all going to be out of our jobs by Christmas.

True, there is a lot to complain about. True, due to something fundamental about the nature of social media (or should that be “human nature”?) negative posts and comments will get more views than positive ones. But does that make it right?

I write this post as either a complete hypocrite or a reformed addict, depending on how you see it. I have written some negative and sarcastic posts in my time. Yet nowadays, I often wonder whether we might actually be shooting ourselves in the foot if we are too quick to react with negativity.

Back near the beginning of the blog, I wrote a post calling for us to start changing the public face of languages. Late last year, I reprised the same theme and appealed for linguists to start demonstrating the value in what we do by showing exactly what we add to society. Today, I want to bring back the same theme again but with what might be a more personal challenge. Are we portraying a positive or negative view of life as a professional linguist?

If you are a 16 or 18 year old choosing what career you will train for, I think you would be looking, at least partially, for a career where people seem positive about what they do. Surely, aside from purely economic calculations, people want to have a job they will enjoy, among people who are helpful and friendly.

Now, to be absolutely fair, I must say at this point that the vast majority of linguists I have worked with, at all stages of my career have been friendly, happy and passionate about their work. Part of what makes this industry so great to work in is that you will come across some truly amazing and inspiring people.

Sadly, that image of careers in the language sectors is not always the one portrayed in the press and even, I hate to say it, by linguists themselves. Often, in our justifiable and even justified need to fight for a cause, educate clients, improve practice, etc, we forget to sell the positives too. Yes, some clients don’t behave the way they should but surely the good ones deserve as much (if not more) publicity as the bad ones. Yes, there are unfair contracts and court interpreters are often mistreated but surely the examples of professional court interpreters giving a superb service are as worthy of collection as instances where interpreters have failed to show up and do a good job.

Perhaps even our most justified campaigns for better treatment, more funding, less course closures and the like would be even more powerful if gave as much space to positive examples as we did to negative ones.  To those who say that closing language courses is “what everyone is doing”, we can say, that, actually no, universities like Heriot-Watt are creating courses and hiring more staff. To those who say that interpreting is a wasteful expense, we can say, no, interpreters have always played a vital role in delivering fair, just trials, and economic growth.

Some of the most powerful cases we can put for the value of languages come from the places where languages have made a positive difference. Some of the greatest arguments for the value of our work come from stories of happy clients, treated patients, and bestselling books. Maybe we need to work even harder at putting a smile on our public face.

Author: Jonathan Downie

Crowdsourcing and the Shrinking Middle

So, with the creation of new tech firm VerbalizeIT, the world has another company that says they can reduce the cost of translation and interpreting. It’s not as if the idea itself is that new. Regular LifeinLINCS readers, will remember our posts on NTT Docomo (among others), who offered a similar service. The difference this time? Well, it’s people. Instead of trusting your important call to the whims of Machine Translation and Voice Recognition, now you are to trust it to other humans.

Sounds a lot better. But wait, there’s a catch. Anyone who has read the ads VerbalizeIT have posted on translation and interpreting job websites will notice something is missing. There is zero mention of experience or qualifications. In the words of their CEO “we want to tap into the one billion people who speak a second language.”

Okay, no prizes for guessing why they think they can hit lower “price points” than their competition. By going for people who “speak another language” as opposed to those with qualifications to prove the point, they are able to get lower rates than you would ever pay for an in-person, qualified and trained professional.

For this reason, much the same can be said about their services than has been said about every other service that has attempted to overturn the industry. It will no doubt do just fine for tourist needs and perhaps (in a pinch) for trips to the pharmacy to buy medicine but I wouldn’t trust it in a doctor’s surgery or hospital. It is very doubtful whether it will make much of a dent in the business or conference markets too.

There is and always will be a need for telephone interpreting and its newer, slicker cousin, skype interpreting. However, for this to be reliable, it needs to be offered by people who actually know what they are doing. Crowdsourcing is all well and good but in places where quality matters, you will want a professional, just as it might be fine to get your Uncle Mick in to change a fuse for you but you would call in a professional to rewire your house.

There will always be a need for professionals and there will always be a need to educate people about the difference between professional translation and interpreting and the kind you can get from “bilinguals”, most of all those who want to enter the profession themselves. For students and those who one day want to go pro, services like VerbalizeIT might provide an insight into what the job involves and some handy cash but they shouldn’t be confused with the high end, quality-driven services that only fully-fledged professionals can offer.

Still, what this new startup reminds us is that there should always be room for language professionals to re-examine their own pricing structure. This might not mean dropping rates but it might mean looking at whether real efficiency savings can be made in how interpreting and translation are provided. There may be occasions where skype is a perfectly acceptable interpreting medium and where post-edited Machine Translation might be all that is required.

Lastly, while the advent of this new startup is not at all a threat or a real disturbance to the industry as a whole, it may be a sign of things to come. It doesn’t take a Nostradamus to predict that the already fragmented language industry will fragment even further, with even larger gaps between the “professional product,” where quality is king and provider-client partnerships rule the seas and the “crowdsourced zone” where price-points and speed hold sway. The middle ground, it seems is growing ever smaller. The question is, are we ready to cope with its loss?

The Importance of Messy Interpreting

It’s a sad fact that interpreting is still not seen as a particularly difficult and useful skill by many members of the public. After all, it’s just like having a walking dictionary, isn’t it? Interpreters hear words in one language and find their equivalents in another. Surely a computer could do the job.

Professionals might laugh at such opinions (in fact, we have laughed at them before) but it is worth pausing a little to figure out why people might have such a simplistic view of our work. True, it could be due to seeing communication between human beings as being similar to communication between computers. You put information in, process it a bit and then output some more information. Interpreters then become machines. Their job is just to find the “right words” in order to give an “accurate translation” of what they have heard.

The quotation marks are very necessary here. Interpret for five minutes and you know that phrases like “right words” and “accurate translation” are loaded and troublesome. There are, of course, many different ways to “accurately” interpret the same sentence depending on context, clients, speed, and a whole host of different factors. It doesn’t take a genius to realise that the vocabulary and phrasing an interpreter might use when consecutively interpreting the cross-examination of a defendant in a court might be very different to the ones they would use when using interpreting the same defendant’s discussions with their barrister.

Life gets even more complicated when you take into account that interpreters in many contexts have to make a variety of ethical decisions as to what to interpret and how to interpret it. (See our interviews with Robyn Dean). Some researchers have pointed out that sometimes the most “accurate” version of what was said might not be the “right” version for a given context.

Andrew Clifford points to a case where, if the interpreter had given the most “accurate” version of what a doctor had said, a patient might not have been able to concentrate on the vital details of how they could be treated. Cases like this might not be found in any textbook but they are the daily realities of interpreting in many settings.

The problem is that, as Ebru Diriker has pointed out in her book, De-/Re-contextualizing Conference Interpreting, on the rare occasions when interpreters get into the public eye, we tend to shy away from discussing the messier aspects of our work. We talk a lot about our language skills, our speaking skills and the importance of our work. We might, very occasionally, talk about the times when we had trouble interpreting or when we needed to be a bit more creative than usual but we quickly reassert that we are still always “accurate” and “trustworthy.”

Faced with such evidence from interpreters themselves, the public have no real choice but to assume that interpreting really is as easy as they thought. If accuracy can be taken for granted then why do interpreters need to be so well paid? If it’s all just a matter of linguistic abilities, why bother with training? If there are never any real decisions to be made, why not let computers do it? In short, if interpreting is just relaying information, why on earth would it be important to have trained, skilled professionals doing it?

Perhaps, in our quest to present ourselves as trustworthy and accurate, we have made it harder to present our work as skilled and worthy of respect. What do you think?

The Interpreters of the Future

… will either be mobile aps or underpaid, under qualified temps. That’s the impression people could easily get from the last month’s worth of news headlines. We already covered the attempt by NTT Docomo to create an interpreting ap and now, wonder of wonders, Microsoft are at it too. Sure, the results are “comical” in places and it just about scrapes by in two languages if it understands your accent but the idea is sound, isn’t it?

And then there is the on-going saga of court and police interpreting in the UK. So far, a government report and two enquiries into the new single-provider contract are uncovering uncomfortable truths such as:
•    The procurement procedure was not up to scratch
•    Advice was ignored or fudged
•    Rates were set without consulting interpreters
•    Not all interpreters working under the agreement were qualified or properly vetted

The end result is that the vast majority of interpreters who are qualified and checked are refusing to work under the new contract and many courts are having to revert back to the old system if they actually want someone reliable and useful. Whoops!

The problem is that the financial logic behind the original move seemed sound enough, at least to those who made the decision. After all, if companies can save money by outsourcing entire functions to a single supplier, so can government department’s right? And, interpreting is just a service like any other right? Surely any good bilingual can interpret, right?

The fault in this logic stems from exactly the issue that this blog covered in the second ever post, over a year ago. The public face of languages and of the language industry needs to be changed. As long as people see interpreting as a financial cost item instead of a worthy investment, spending patterns won’t change. For as long as people associate interpreters with people they don’t want in their country, justifying pay rises (or even pay stability) will be difficult.

The point is that most, if not all, interpreters know the real potential of their work. Not only does interpreting help justice to be served, it helps people to get medical attention, families to cope with trauma, business to conquer new markets and economies to grow. As soon as you trade anything, be it people, products or ideas, outside of the market that speaks your language you need interpreters (and translators).

If the future of interpreting is to be filled with qualified, vetted, reliable professionals, someone will need to make sure that the message gets out that this future and only this future is the one we should be chasing. Someone has to convince government ministers, business people, and the public that interpreting is worth more than it costs. Anyone up for that?

Is it all over for Interpreting?

If you believe the hype, the interpreting profession is on its last legs. NTT Docomo, the biggest mobile phone network has at last managed to provide a service that allows anyone with a smartphone to instantly have their words interpreted. According to this article from the BBC, other companies are working on the same technology. All that remains now is for a few minor tweaks to be done and language barriers are a thing of the past. Cool, eh?

Well, it might be, if things were really that simple. On the accompanying video, BBC presenter, Richard Taylor feeds the software a simple sentence “are there any good restaurants near here?” and an interpreter (mislabelled “translator,” thanks BBC!) responds with something equally as simple. This is the kind of thing that 1st year language students could do in their sleep and even good Machine Translation software could churn out without too much difficulty.

Faced with this simple task, the software performs pretty poorly. Its English-Japanese version is, according to the interpreter, understandable but not much better than that. The Japanese-English, meanwhile, would not exactly be what an interpreter would produce. We are not in minor tweak territory here. We are actually about 10 years behind the current state of the art in machine translation and parsecs away from what one could call “interpreting.”

So, no need to worry, interpreters still have jobs for a few years yet. It is still worth doing training and looking for work.

Even if this technology does work, will it do humans out of work? The answer is, probably not. Just as machine translation has only really cornered the market in getting the “gist” of emails, letters and the like and has actually become another tool for human translators, we can expect something similar to happen here. The only people who might really feel threatened by this technology are phrase book publishers. For simple requests for the direction of the tourist office and panicked searches for the nearest public toilets, a machine translation-enabled smartphone will (eventually) do the trick.

It is difficult to foresee this technology taking over in the court, conference or business markets. Would you really want your perfectly crafted speech to be turned into garbled but minimally understandable googlish? And that might represent the best outcome!

This well-known video illustrates just what the limitations of NTT Docomo’s technology are likely to be. For those with non-standard accents such as Geordies, Aberdonians or even, (shock!) Glaswegians, voice recognition is already struggling. Add in a layer of not quite perfect machine translation and we are looking at comedic breakthroughs rather than technological ones.

Of course, we can never dismiss technology out of hand. It is possible that, just as google managed to create a step change in machine translation by using texts from large international organisations, someone might manage something similar with this. It is just possible that, with years of work and a lot of training, NTT Docomo or someone else might manage to move this from being a gimmick to being of some use. For the foreseeable future, that’s the best that can be hoped for.

So, while reports of the death of interpreting would be greatly exaggerated, phrase books might be pining for the fjords. Once voice activated machine translation can handle those two obscure allusions said in a Glaswegian accent and produce an entertaining version in another language, then we need to start worrying!

How could research help you?

If you could ask translation and interpreting researchers to look into one thing, what would it be? Here is a simple poll to get you started. Please do leave extended answers in the comments.

Please note, the question below should read:

If you could ask translation and interpreting researchers to look into one thing, what would it be? Sadly, the poll cannot be editied now without resetting the result. Apologies for the typo.

[polldaddy poll=6285388]

Best of Frenemies? Machine and Human Translation

It was all so simple in theory. With enough computing power, a handful of smart linguistics boffins, a few computer scientists and enough time, human translators would be as obsolete as punch card programming. No matter the text, no matter the subject, Machine Translation (MT) would give you a word perfect version every time in whatever language you wished.

As with most utopian dreams, it would all end up a bit more complicated than anyone could have predicted. Language, you see, is a funny beast. Just as you think you have it all described and analysed, along comes some creative addition or reuse. Tired of “footprint” just referring to the remnants of last night’s walk along the beach? Strap the word “carbon” onto the front and you have a chic environmental term. Bored with “surfing” only involving foamy waves and sleek hair? Mix things up a bit and add some fishing terminology and you get “surfing the net,” which, oddly enough has everything to do with technology and nothing much to do with the sea.

Even traditional songs take on a whole new slant once people start playing with language. “Don we now our gay apparel” used to mean getting dressed up for a Christmas meal. Now, on the other hand, it has a completely different meaning altogether!

All this was enough to play havoc with translation engines. How on earth could anyone write software based on nice, stable rules when word use and meaning changed all the time? What a mess!

No worries, thought the MT gurus, we have a plan. That plan was to beat creativity at its own game. Instead of trying stable, unchangeable rules, MT software started using existing language as a template for its versions. Give your software a big enough database of language and some nifty statistical algorithms and you could get reliable results, or so it was thought.

Actually, this works fairly well. The much maligned rule-based Babelfish has been taken over by the stats-based Google Translate, which has the bonus of a huge database of UN language to work from. Feed in the right kind of language to Google Translate and you do get passable results. On the other hand, feed in certain kinds of language, translate backwards and forwards a bit and Rick Astley’s “Never gonna give you up” becomes “You’ll never leave,” which is funny and ironic but wrong.

So, for the moment, it looks like MT isn’t quite good enough on its own, especially since it can only ever be as smart as the data humans give it. In fact, it is this partnership between human common sense and machine processing that seems to be the way forward. Now, instead of MT taking over, it has been coupled with its somewhat more respectable cousin Computer Aided Translation (CAT) into flexible translation toolkits. Today’s translators, it seems, are happy to take whatever help they can get and are smart enough to be able to use a variety of tools at once.

More importantly, this partnership suggests that translators’ jobs are safe, at least in the near future. Rather than Machine vs. Human being a battle to the death, it has become more of a growing partnership.

Jonathan Downie