The Babel fish, Skype translator and the end of the world as we know it…

I’m no futurologist but I have made two apocalyptic predictions in my time, both of which look like they might come true. The first was that I thought it was a very bad idea to try to convince the Chinese that they should adopt a Western lifestyle. Even before we knew what we know now about the effects on the climate made by mankind, it seemed that having one billion Chinese people driving cars and flying round the world was never going to be sustainable and that is proving to be true.

My other prediction was that one day, when technology made the idea of the Babel fish reality our ELT industry would fall apart.

The Babel fish appeared in The Hitchhiker’s Guide to the Galaxy by Douglas Adams:

“The Babel fish is small, yellow, leech-like, and probably the oddest thing in the universe. It feeds on brain wave energy, absorbing all unconscious frequencies and then excreting telepathically a matrix formed from the conscious frequencies and nerve signals picked up from the speech centres of the brain, the practical upshot of which is that if you stick one in your ear, you can instantly understand anything said to you in any form of language: the speech you hear decodes the brain wave matrix.”

One of the first online translators used the name, https://www.babelfish.com/ – it still exists but has been overtaken by Google, like so many other things.

Outside the realms of sci-fi, what I am really talking about is technology that would allow a speaker to speak in one language and for the person they are speaking to hear in another language. For this to work it would require two technical components: speech recognition and instant translation. Both of these are already ubiquitous – a lot of us already have speech recognition in our pockets in the shape of Siri on the latest Apple iPhones and instant translation is offered on many websites.

Fortunately neither of these technologies really works properly… yet.

We probably have all heard amusing stories about Siri failing – I once sent my mother an unrepeatable text message and this video brilliantly shows how speech recognition fails to allow for varieties in different accents.   

You can also test Google Translate to see how it struggles to translate more than just single words by making it translate back and forth between two languages, like this:

V1

Can Google make any sense of this passage whatsoever? Let’s check it out.

O Google pode fazer qualquer sentido desta passagem que seja? Vamos dar uma olhada.

 

V2

Google can make any sense of this passage is that ? Let’s take a look .

Google pode fazer algum sentido desta passagem é isso? Vamos dar uma olhada.

…it doesn’t even notice that a question mark denotes a question.

However, in my opinion, it is only a question of time until both of these technologies do work. Microsoft has already announced that as part of its Skype acquisition, it is developing the ability for phone calls to be translated in real time, which it demoed in this video.

So what does this mean for us? Well, if you are a teacher ask yourself what percentage of your students is learning English because of a genuine love of the language? Probably quite low, isn’t it? Once technology means that you don’t need to know how to speak a foreign language, then that will be the percentage of students left in your class. Or left buying English language materials…

The end of the world as we know it.

 

Luke Baxter

Luke Baxter is the Digital Publisher at Richmond in Oxford. He taught English in Argentina and then Madrid, where he founded a Business English academy. He joined OUP as an Editor before going to Richmond in 2010. Luke has an MA in Latin American Literature from Warwick University.

1 Comment
  • Dennis Warren
    Dennis Warren
    Posted at 14:21h, 01 setembro Responder

    A thought-provoking article, Luke.

    As someone whose job used to be to create systems to convert text to speech in a single language (English), I think that the possibility of creating successful automatic translation from recognition to production is far off.

    In my area, the seemingly simple feat of converting a written text in English to spoken text in the same language was fraught with problems and barriers. We did a pretty good job of it, but the language is so full of exceptions that the use of lookup dictionaries was an incredibly common fallback, making the system slower and less flexible.

    The way forward would seem to be a greater understanding of human processing; from hearing to understanding to translating to producing. And that’s incredibly complex and difficult to get at.

    Experiments with neural nets in the area of language seems to be very promising as a way of creating robust flexible systems, yet I think that for a very long time to come we will see only purely functional systems, whereby the rudimentary meaning is all we get out of them – allowing us to understand the basic message without understanding the subtleties.

    The universal translator is still something that only sci-fi will have for many generations to come. I think your job is safe for while!

Post A Comment