Are Translation Apps at All Useful?

Machine translation has certainly come a long way since its infancy just a few years ago. So is it good enough to provide the basic idea of a message you have received in a language you don't speak? Not quite, it seems.

Several of my family members collaborated on my combined birthday/Christmas gift: an iPad. Having tried for some time now to come up with a justification for buying a tablet computer -- and failed to do so -- I was overjoyed. Browsing through its app store, I came upon a free translation utility, "Free Translator". Once I downloaded it, a small note in the corner said "powered by Google translate". While Google generally can't provide a polished translation, it is often good enough to get the gist of a text.

So I decided to test it with the German sentence from one of my clients giving me the go-ahead for a project and confirming its due date. The German sentence read "Der Auftrag ist erteilt, Lieferung Mitte nächster Woche OK!" The app returned "The order is issued, starting mid-next week OK!"

This does sound like an understandable English sentence, doesn't it? Well, yes, except for the fact that "Lieferung" means "delivery" (i.e., due date), not "starting". Were I to rely on the translated version of this order confirmation, I wouldn't be able to deliver on time (the project involves 6 PDFs of 2 pages each).

The purpose of small, free apps such as this one is precisely for a reader to understand the basic idea in an e-mail that was written in a language he or she doesn't speak. Good English grammar, let alone polished style, is not necessary in that context, but accuracy is. Even if terms are only "sort of" right (e.g., "udder" instead of "breast" in a sentence about a woman's cancer diagnosis), humans can often discern the actual meaning. However, if the translation is simply wrong (as in "starting" instead of "delivering"), there is no way for a person to know that he or she has have been given the wrong information.

If one cannot rely on such apps to provide even the basics of a message, there seems little point in using them. To answer the question in the title, then: apparently not.

Caveat: I did use Google Translate last year to render "Merry Christmas" into Tagalog for my son-in-law's card. He tells me that while the phrase wasn't idiomatic, it was understandable. So sometimes it does work. But how do I know when it does?/p>

When Is a Hot Dog a Boiled Puppy?

Going through back issues of Wired magazine to cull articles worth keeping, a pull-out quote in a March 2010 article on Google's search algorithm caught my eye. It read "The synonym system knew that a dog was similar to a puppy and that boiling water was hot. But it also thought that a hot dog was a boiling puppy."

To be sure, Google has long since fixed that particular search problem, but it does illustrate nicely why translators are unlikely to be supplanted by machines anytime soon. By the way, Google Translate correctly translates hot dog into German, showing both the anglicism "hot dog" and the German variation on a hot dog, "Würstchen" (little sausage). However, its translation of "boiled puppy" into German results in the ungrammatical -- albeit understandable -- "gekocht Welpen" (it should be "gekochter Welpe", since puppy is singular and Welpe is masculine).

All languages contain any number of expression that cannot be taken literally, although it seems to me that English has a greater share of these than some other languages. The IT field, in particular, is rife with expressions which were re-purposed from other contexts (the mouse on my desk is not grey and furry, and when I save this post I'm not rescuing it from drowning or spiriting it away in a vault).

On the other hand, technical instructions are a good candidate for standardized vocabulary, which in turn would be easier for a machine to process. As I have argued in a previous post, we technical translators may find ourselves increasingly editing machine-produced output. However, as I have also argued, this role is only a viable alternative if the way we are paid changes. A per-word rate simply does not take into account the varying amounts of time such editing may take -- independent of whether the text being edited was produced by a human or a machine.

The challenge now is to insist both on a change in the payment structure and a reasonable hourly rate. Unless all (or, at least, most) of us do so, the race to the bottom in terms of pricing will continue and translators producing quality copy will, indeed, be supplanted -- or will have to work for what amounts to minimum wage when viewed on an hourly basis.

Machine Translation Instead of Global English?

I just read an interesting article in the July 2010 issue of Wired magazine. In "No Language Barrier", Clive Thompson talks about how automated machine translation could make the emergence of a lingua franca obsolete since people can interact with each other using their own dominant language, with the machine providing the linguistic, as well as technical, interface. Such an interface could halt the spread of English as the de-facto lingua franca across the world, he argues.

While this sounds like good news to those of us who are concerned about the increasing influx of English into the world's languages and consequently increasing homogenization of cultures, it may not be so. Language and culture are certainly intertwined, but an American movie in, say, Samoan, is still an American movie with American cultural references, values and view of the world. Being able to communicate in one's own language is certainly a step forward in letting people across the globe communicate with each other, whether they know English or not. (Although whether that will entice my non-English-speaking mother to finally learn how to use a computer is another question ...). Technical solutions, such as automatic machine translation, cannot curb the march of U.S. culture across the globe. That can only be achieved by political and economic means.

Mr. Thompson also cautions that "Certainly any activity requiring serious precision -- legal proceedings, business discussions, diplomatic negotiations -- will still need expert human translators." He is certainly right in this assessment, although judging from some of the translations I have encountered and/or been asked to "fix", not all potential translation clients seem to think so. The challenge for us translators now is to educate clients why machine translation may be fine for a chat on Facebook, but isn't sufficient for a contract to build a new factory in Romania (even if face-to-face negotiations for that factory were conducted in less-than-stellar English).

The Changing Role of the Translator

I just read an interesting op-ed by David Bellos in last weekend's New York Times. In "I, Translator", the director of the Princeton translation program offers a brief overview of machine translation, from its infancy in the U.S. at the start of the Cold War to its current Google Translate iteration and concludes that machine translation cannot replace human literary translation.

A point well taken, but literary translators constitute a fraction of the translation community. Most of us don't work on War and Peace but on such prosaic texts as legal contracts, medical documents or technical manuals. Quite a bit of these documents consist of "boilerplate" or otherwise at least semi-standard text -- that's why translation memory tools have become so useful to human translators. Because of this standardization and repetition, such technical text does lend itself to machine translation, just as the original text is a prime candidate for standard content management systems. This move towards further automation may put both translators and technical writers out of business -- unless we adapt.

As Google Translate and similar systems parse ever-increasing quantities of multi-lingual web content -- including my own bilingual website -- their translation engines approximate human translation ever more closely, at least in relatively common language combinations, such as German-English. But while a business contract doesn't need to be good prose, it does need to be extremely accurate -- approximation just isn't good enough. That's also true for medical and technical texts. So while Google Translate can provide a basic idea of what the parties to a contract agreed upon or how to operate a particular piece of machinery, humans still need to ensure that nothing has been mistranslated and that the end result flows smoothly and is easy to understand.

That's where we human translators come in: as editors to verify and refine the somewhat crude machine translation. To some extent we are already performing this role when we edit the "fuzzy matches" provided by our translation memory tools. Post-editing machine-translated text would be similar, just with a much larger translation memory. The question now becomes: how do we come up with a compensation model where we can still make a living, even though this editing will garner a much lower per-word rate than traditional translation?