Marek Kohn: Why Babel fish will make us value languages more, not less

November 10, 2023

Imagine having something small enough to put in your ear, that would translate what anybody said to you into the language of your choice. The humorous author Douglas Adams imagined it in the form of a living creature, the Babel fish.

That was some forty years ago. Now technology is catching up with Adams’s imagination. It’s already possible to buy earbuds that will translate speech. The technology isn’t yet smart or seamless enough to take over the world, though.

It needs to do several things better.

Artificial translation has to become more accurate, more comprehensive, and more sophisticated. Voice recognition has to improve and cope better with the noise in the background of most conversations. Personal translation devices will have to become more self-sufficient, so they don’t have to rely on connections to remote servers that do the actual translating. They’ll have to provide a continuous stream of translated speech, like the soundtrack of a dubbed movie, so that conversations don’t have to pause while the machines catch up.

These are big challenges. But there don’t seem to be any technological impossibilities or laws of physics in the way. It seems reasonable to suppose that in time, we will be able to step out into the world with artificial Babel fishes in our ears.

For many people, that may come as a relief. No need to learn foreign languages any more, they will think. Another sophisticated cognitive operation will be outsourced to digital assistants. Artificial voice translators will do for language skills what electronic calculators did for arithmetic.

And that illustrates why the better artificial translation gets, the more people will appreciate the value of actually understanding languages.

Having a translator in your ear will be hugely useful at a basic level, just as having a calculator app on your phone is very useful, but only for doing sums.

It will come in very handy if you make short visits to countries whose languages aren’t widely spoken elsewhere, or if you meet people who speak languages you haven’t encountered before. It will help you avoid confusion in shops or bus stations. It may help you grasp what strangers are saying when they accost you in the street.

That is the kind of level at which screen-based text translators already operate: handling simple communications and getting the gist of what is going on. For many people and many situations, it will be enough. But people who can understand a language for themselves will always be better placed than those who have to rely on their devices, just as a person who can do sums in their head will have an advantage over a person who has to tap the numbers into a calculator.

People who understand languages will have an insider’s sense of nuance and tone. They will get the jokes.

They may also be able to understand important aspects of what is being said to them that are inevitably lost in translation. Many languages have formal and informal modes of address, such as the French ‘vous’ and ‘tu’. A translator, whether human or artificial, will render both those forms in English as ‘you’.

That’s correct – but it loses key information about the encounter. Is the person addressing you formally or informally, politely or disrespectfully? As somebody of higher, lower or equal status? As a stranger or a friend? To know that, you need to know the language.

By taking over the mundane levels of language use, machines will divide the value of language into two tiers. While artificial translation will become a commodity product, human language skills will enjoy a premium. It will arise from their power to bring people together and to make them comfortable with each other.

Learning a language takes time and effort. That signifies commitment, which promotes trust. Even a little can go a long way, which is why people appreciate it when travellers make the effort to learn how to say ‘hello’, ‘please’ and ‘thank you’.

Knowing a language may enable you not just to communicate with people, but to become part of their community. By speaking the language of the larger community, migrants can politely assert their place in it. That’s always going to work better than saying ‘I belong here’ through a microphone and a translation gadget.

Insights like that may be the most valuable kind of assistance that the translation technology of the future can give us.

Marek Kohn

Marek Kohn’s Four Words For Friend: Why Using More Than One Language Matters Now More Than Ever is published by Yale University Press.

Learn a language