Remember the days when machine translations were confined to fictional characters such as Star Wars’ C3PO, a droid fluent in over 6 million forms of communication? Even Dr. Who’s flying phone/time travel booth, Tardis, can instantaneously translate any language in the galaxy.
Nowadays, however, automated translation services are virtually everywhere! And it seems like everyone is getting into the game! But are we really better off surrounded by all these translation websites and apps? Are they mending the deep divide when it comes to us understanding one another across our many languages?
Most of us know the biggest players in the machine translation universe and have some experience with them:
- Earlier this year, Skype started offering voice-to-voice translation for seven languages: English, French, German, Italian, Mandarin, Portuguese, and Spanish.
- Instagram recently introduced a “See Translation” link so you can see what a bio or caption says in 24 languages.
- Google Translate claims the ability to translate languages whether you’re offline or online. It even can also instantly translate languages using your cell phone’s camera. Google also offers an app called Word Lens which allows the user to photograph signs and automatically translates the text on their phone.
- Facebook also has their own translation tool to connect its billions of customers.Facebook even allows its bilingual users to give human translations. Then other users vote on the accuracy of the translations. If enough users vote positively on the transaltions accuracy, the automated translations will be replaced with the human translations.
- This year, Waverly Labs launched ‘Pilot’, an earpiece that allows users to talk face to face in real time in different languages in order to create “a world without language barriers.”
No room for Improv
While these websites and apps can be helpful, they really do have their limitations. Often
word-for-word translation incurs the least errors with machine translation, but more complex sentence structure can cause translation technology to falter. Unique concepts, certain subtleties, figurative language all make for machine translation faux pas.
We tried testing Skype Translate earlier this year. And the results were less than stellar. In fact, disastrous and misleading are better words for it. When tested, Skype Translator added inappropriate translations for seemingly innocuous sentences. Our testers, a married couple, were shocked when Skype Translate insisted on translating one of their sentences as an admittance of an affair. At one time, Skype Translator even turned off. Hardly, accurate translations! Thankfully, our tester are still happily married after that experiment!
Even experts who see the Waverly Labs Pilot as the most progressive innovation of all translation technology, warn that “only time will tell if the machine’s functionality is worth the hype.” They say they are only speculating and taking “with a pinch of salt” the claims of the company’s promotional video. They warn that technology will always struggle when it comes to words and phrases that have “no English equivalent or cultural relevance to someone who isn’t from that country.”
One word at a time
Google’s Word Lens ability to translate words from signs has garnered some praise, but it really is only helpful when deciphering separate words. It finds it hard to deduce sentences, phrases and paragraphs.
Translation technology relies heavily on this word-for-word algorithm. Yet, as we all know, words have their many meanings, language its nuance. A speedy app will not always be accurate.
Translation apps don’t know slang, dude.
Literal machine translation doesn’t know the local dialect, latest slang or cultural appropriation. We have all heard the stories of machine translations that fall flat, creating disastrous outcomes for foreign marketing campaigns.
A work in progress
Facebook and Google are constantly working to improve their automatic translations. They try to incorporate colloquialisms and gather feedback from users. Adding a human element to this can only improve tech translation tools in the long run.
But should everyday users be the translators of websites and apps? While there are steps being taken in the right direction, tech translators are still a long way off from being perfect.
Don’t rely solely on them
Translation tech is helpful if you don’t know a language at all. It can help with basic instructions, signs, single words. But if you are looking for meaningful conversations, you need to do more work to learn languages.
Translation technology is not mending the deep language divides. It is more like they are basting a seam or loosely covering a wound. It is a start, but there is a long way to go. And that is job security for translators.
By Ilona Knudson