Language erosion: You don't hear that often...

Tomorrow is World Languages Day, and it seems appropriate to announce a happy but increasingly uncommon event: the discovery of a previously unknown language in the foothills of the Himalayas.

Koro, as the language is called, is spoken by hill tribes living in the northeastern state of India called Arunachal Pradesh, near the borders with China and Burma. Its discovery bucks a trend, since linguists have estimated that at least half of the roughly 7,000 extant human languages will be dead or moribund – meaning that children will not be able to speak them – by 2100. In fact, Koro was first identified by a team of Indian language surveyors in 2003, but its findings were never published. The three linguists who announced their "discovery" of Koro last month travelled to the remote Indian province as part of National Geographic's Enduring Voices project, to record two other, little-known languages belonging to the Tibeto-Burman language family, Aka and Miju, and rediscovered Koro by accident.

At first, Gregory Anderson, K David Harrison and Ganesh Murmu thought they were hearing a dialect of Aka, but they soon realised it was sufficiently different to be called a language in its own right. The sting in the tail of their find, however, is that Koro itself may be dying. Only around 800 people speak it, few of them are younger than 20, and the language has never been written down.

If it dies, says Harrison, who works at Swarthmore College in Pennsylvania, a unique culture will die with it. "[Koro] contains very sophisticated knowledge that these people possess about this valley, the ecosystems, the animals, the plants, how they survived here, how they adapted, so if they switch over to another language a lot of that knowledge will simply be lost," he says.

The language may also contain clues as to how the Koro speakers came to be in that valley, and why they identify themselves so closely with their Aka neighbours, even to the extent of downplaying the considerable differences in their languages. The linguists suspect they may have been brought there as slaves, though they have yet to prove it.

Harrison believes that languages such as Koro must be preserved because of the knowledge they contain, and because the sheer diversity of human languages provides a window on the inner workings of the human brain. But the idea that linguistic diversity itself could represent a valuable scientific tool has only recently gained traction in the linguistic world, which has been more or less dominated since the 1960s by American linguist Noam Chomsky's concept of universal grammar.

Universal grammar refers to the idea that children inherit an innate system of rules from which all languages are generated – a "language instinct" – which is why they learn so quickly. According to this theory, what is interesting about languages is what they have in common, not how they differ. There's no doubt that children are gifted linguists. By the age of three, the average toddler has a good grasp of the syntax and prosody of their mother tongue, and a vocabulary of almost 1,000 words. But some researchers are now suggesting another explanation for that aptitude.

As Mark Pagel, an evolutionary theorist at the University of Reading, sees it, languages evolve just like organisms, and the ones that survive are the ones that are best adapted to the human brain and, hence, easiest to learn. All humans have the same brain, which is why successful languages tend to resemble one another, giving the illusion of a universal grammar. But, Pagel says, they may have arrived at that similarity via different routes, and solved the problem of being easy to learn in different ways.

"What will happen over the next 10 to 20 years is that there will be a series of papers that will show more and more how words have adapted to us," predicts Pagel, whereas if Chomsky was right, no such adaptation would be necessary because words would be generated by the language organ in our brains, already fit for purpose. If Pagel's prediction is borne out, then the differences between languages become much more interesting, because they illuminate the range of strategies the brain can adopt for learning and making sense of the world.

Last year, linguists Nicholas Evans of the Australian National University in Canberra and Stephen Levinson of the Max Planck Institute for Psycholinguistics in Nijmegen, the Netherlands, seemed to hammer another nail into the coffin of universal grammar when they listed exceptions that had been found to long-established linguistic "rules". A rule that was once thought to be inviolable, for example, was that every language distinguishes between nouns and verbs. But Straits Salish, spoken in the Pacific north-west of North America, may blur that line.

Evans and Levinson suggest that languages may even shape the brains of their speakers, to some extent – not so much that speakers of different languages have incompatible world views, but in subtle ways, making them pay attention to orremember different aspects of their environment, for example, or interpret certain instructions differently.

An elegant demonstration of this was published last year by Daniel Haun and Christian Rapold, also of the Max Planck Institute in Nijmegen. German, like English, locates objects in space in an egocentric way – that is, relative to self – but many other, if not most, languages describe the location of objects relative to another object in the scene, or to an absolute reference such as a compass point.

The researchers taught German and Namibian hunter-gatherer children a sequence of hand movements relative to the body that followed the order right, left, right, right. When they asked them to rotate their bodies 180 degrees and do it again, the German children reproduced the same sequence, while the Namibians now produced the sequence left, right, left, left, indicating their reliance on an allocentric or absolute frame of reference.

Universal grammar is by no means dead, though. Tecumseh Fitch, an evolutionary biologist at the University of Vienna in Austria, argues that many of the criticisms that have been levelled at it are based on a misunderstanding. "Universal grammar is neither a grammar nor is it universal," he says. It actually refers to a set of constraints that limits possible grammars, so that different languages make different choices among those possibilities. Understood correctly, Fitch says, the theory is not incompatible with linguistic diversity.

As founders of the Living Tongues Institute for Endangered Languages, Drs Harrison and Anderson are dedicated to preserving and "revitalising" endangered languages in order to protect that diversity. But whether linguists should try to save languages is a contentious issue.

According to Dr David Lightfoot of Georgetown University in Washington DC, linguists argue about the precise definition of a language, but most agree that it is, at least in part, political. "A language is a dialect with an army and a navy" is an aphorism linguists are fond of quoting. Mandarin is considered one language, for example, though native speakers of its major forms cannot understand one another, while mutually intelligible Norwegian and Swedish are two separate languages. The desire of Koro speakers to downplay the differences between them and the more numerous Aka suggests a political dimension to their linguistic heritage too.

Documenting dying languages is important, Dr Lightfoot believes, because of the wealth of information that will otherwise die with them. But saving them is another matter. That, he says, is a political decision, and not one scientists should get involved with.

Meanwhile, Paul Lewis, who edits the world's most comprehensive language catalogue, Ethnologue, says that there are probably other languages like Koro waiting to be discovered – discovered, that is, by everyone apart from their speakers – though they won't tip the balance, and language erosion will continue to be a problem. "There are still many parts of the world that are relatively unexplored linguistically," he says, giving examples of Papua New Guinea, some parts of Indonesia and Nigeria. "So we can expect that more such cases will come along."

Languages in danger

KHOMANI

Spoken by fewer than 10 people whose lands are located in the Kalahari National Park in South Africa, the Khoisan language makes use of various clicking sounds

GUUGU YIMIDHIRR

An Australian Aboriginal language spoken near Cooktown in northern Queensland Guugu Yimidhirr is spoken by around 200 people

KORO

Spoken by around 800 people in Arunachal Pradesh state in India, Koro has only just been discovered but is at risk

TER SAMI

Located on the Kola Peninsula, there are no more than 10 Ter Sami speakers alive

AINU

The Ainu language is extremely complex and is spoken only by a handful of elderly people on the Japanese island of Hokkaido