I am a big admirer of Nick's, let me say from the start! I am not as confident as Nicholas Ostler is in the future of machine translation as a replacement for a lingua franca. Both speech recognition and translation have to make major steps forward before anything resembling the Star Trek universal translating device (or Douglas Adams's Babel Fish) is anything close to reality. I spend a lot of time translating idiomatic phrases I find in German writing through Google Translate, and while the software is amazingly good sometimes, it is also amazingly bad much of the time. Common German expressions are simply not in the software's ken.
I've read the expression "in ein Fettnäpchen treten" several times since moving to Germany last summer - roughly, "to put your foot in your mouth" (though it means literally "to step in a greasy bowl"), ie., make a faux-pas. Google Translate turns out "take my foot" as the translation. This is unusably confusing - and this is a common expression, to be translated between two closely related languages for which Google must have zillions of worlds of parallel training text. I simply wouldn't begin to rely on GT to translate Chinese into English for anything crucial. Much less would I do so to translate Chinese into Hungarian - as I understand it, this is still a two-step, which goes through English, not surprisingly multiplying the errors. Remembering that MT is now almost 60 years old, and some of the smartest people in the world have worked on it, and a company with the passion, the money and the brains of Google is still at this stage, I see MT making evolutionary, but not revolutionary, progress.
Add the still-developing quality of speech recognition - good when I hold my iPhone to my face, speak slowly, pronounce my punctuation marks ["COMMA"] and make sure not to use any unusual words, much less unusual names. Speech recognition is a million miles away from being able to translate (say) a spontaneous lover's quarrel or a barroom argument over who was the greatest footballer of the 1970s, with all of the starts and stops, half-sentences, garbled words, slang, and so many other problems of real live speech.
These two fields have miles to make up - in tandem - and then they have to be successfully linked together, before they can be reliably used to replace good old language-learning, in my view. The interview I had about renewable energies this morning in a noisy coffee shop in German will not be replaced by a German-English Babel Fish device in less than 25 years.
Let's say, though, that optimistically, this is possible in 25 years.
(I think that's very optimistic.) What else will happen in 25 years?
The world's current lucky lingua franca will have a lock-in in every region on earth. Europeans are increasingly moving to beginning English-teaching at 1st grade. China famously has more English-learners than America has people, and though teaching quality is still poor, it can only improve with teaching technology and ever-increasing Chinese access to native speakers, not to mention the will of the Chinese state to make it so. To be morbid: 25 years' worth of people who never learned English will die, and 25 years' worth of people who never knew a world without the necessity of English will be born and get between 8-12 years of English in their schools in every region on earth.
I think of English as like the QWERTY keyboard: not perfect, but it doesn't need to be - the point is that it has spread so far that to replace English-learning with another system would require a massive upheaval in a system people have already adapted themselves to - in other words, abandoning English will be a costly shift that people won't see the point in. Meanwhile, the learning itself becomes both easier and more attractive as more people have learned English - a virtuous cycle (for English).
I actually think this is *bad* for many reasons - it could accelerate language death; there is sure to be nationalist backlashquarters (as Nick has written); and I just simply get the willies at the thought of English taking over the entire world in two centuries' time.
I am a keen language-learner myself, and would be seeing the obsolescence of my best party trick. But as for the nationalist backlash, the more that non-English-speakers master English, the less English is associated with Britain and America - it just becomes a tool, our QWERTY keyboard, and I think the nationalist element attenuates. At least, I see this as more likely than the Babel fish.
Obviously, things that I can't foresee are things that I can't foresee, so I could be wrong. But that's how I see the trajectory.
Lane Greene and Nicholas Ostler will continue their debate live at the TAUS Industry Leaders Forum in Dublin on June 2 and 3.