Google carries the barrier of the language with the translating telephone.

Google is developing software for the first capable assembly to translate foreign tongues almost instantly as the fish bebble and that is all, to speak with Scott in the Galaxy….

Huh? This blog is supposed to be about the Google Voice translation announcement last week, but so far I’m making less sense than usual ― this is smelling more like TranslationParty than some big voice translation breakthrough. Maybe the problem is the SDL machine translation I used for the back translation:

Google lleva la barrera del idioma con el teléfono traductor.

Google está desarrollando software para la primera convocatoria capaz de traducir lenguas extranjeras casi instantáneamente como los peces bebble y eso es todo, hablar con Scott en el Galaxy.

¡Ay, caramba! You don’t need to know Spanish to know how bad that is.  Maybe the problem goes back to an earlier step in the voice translation process. The phrase above is a Google translation from a Google Voice transcription and is actually pretty good for Google Translate. Here’s the source text:

Google leads language barrier with translator phone.

Google is developing software for the first call capable of translating foreign languages almost instantly like the bebble fish and that’s it, talk to Scott at Galaxy.

Hey, but that’s not what I said!

Don’t get me wrong. I love Google Voice (super convenient and free), so who’s to complain? I recorded the headline and lede of this story onto my answering machine, and in less than two minutes I got a transcription in my email. That is amazing, but it’s not what I said, which was:

Google leaps language barrier with translator phone

Google is developing software for the first phone capable of translating foreign languages almost instantly — like the Babel Fish in The Hitchhiker’s Guide to the Galaxy.

That’s what I said, reading in Standard English at a normal conversational speed, in a professionally enunciated tone (I’ve taken voice training) ― Scouts honor, I did not try to jam Google Voice!  Maybe the problem is the process, as error accumulates step by step (so something like TranslationParty).

Now if you read the story closely, Google didn’t actually say they had built the software, just that they were going to.  I suppose that as the CEO of a telephone interpretation service, I should be relieved that Google hasn’t put me out of business just yet, but I’m a bit disappointed too, though not surprised at all.

People have been talking about Star-Trek-like universal translators for a long time. And they will be talking about them for a long time to come. Despite the remarkable advances we’ve seen, this latest endeavour is still a long way from seamless technology, even with the vast resources of Google bent to the task. The challenge of increasing accuracy to the level of easy human comprehension, with an acceptable level of delay, is logarithmically more difficult than the current state of the art and is likely to remain so for a long time.

I anticipate two compromises required of users of audio translation systems for many years to come:

1. You will have to pay. It may be cheap, but it won’t be free.
2. Users will have to learn to settle for third-rate technology.  The level of frustration in learning to adapt and effectively use such systems will be high.

So that universal translator will remain on screen for a long time. And I don’t think Google is going to be the one to do it anyway.

And now the screed, which is always the most fun to write. Each time someone promises universal, instant, free translation, they score a headline, and it’s a point against our profession, we the thousands of linguistic squirrels, racing along behind the scenes on our little linguistic hamster wheels, making it seamless for everyone else. So I close with a quote by Apple entrepreneur Steve Jobs on Google, which I think applies in this case too: Bullshit!