Forked tongues: The evolution of human languages

A pair of faces

23 March 2012 by Mark Pagel

Humans have the dubious distinction of being perhaps the only species whose members cannot all communicate with each other. Mark Pagel asks what's behind this apparently unhelpful trait, and what the future might hold for the diversity of human language.

There are about 7,000 distinct human languages spoken on Earth - that's more languages for our single mammal species than there are mammal species. This means 7,000 different ways of saying "good morning" or "what's for dinner?" Large as the number of extant human languages is, there were probably as many as 12,000 to 20,000 at the peak of language diversity around 10,000 years ago, before agriculturalists began to spread across the globe and replace many hunter-gatherers' cultures. And 7,000 languages pales in comparison with the possibly hundreds of thousands of different languages we have ever spoken throughout human history.

This diversity means we are perhaps the only species whose members cannot all communicate with each other. Indeed, it is as if the different human-language groups have come to act almost like different biological species. But why would humans have evolved a system of communication that effectively cuts them off from other members of their own species? Over the past few years my research group and I have been studying human languages using ideas drawn from the theory of evolution by natural selection, ecology and biogeography, and those studies are providing some intriguing answers to this question.

Why so many languages?

Modern humans evolved in Africa as early as 200,000 years ago, and by perhaps 60,000 to 80,000 years ago had migrated out of Africa to establish a permanent presence in the rest of the world. We expect new languages to arise as people spread out and occupy new lands, because as soon as groups become isolated from each other their languages begin to drift apart. We see this in modern times in the differences between British, American and Australian English.

Graph showing densities of languages

Numbers (top) and densities (bottom) of languages and mammal species at each degree of latitude in North America. Both trends show the decline in density towards the poles characteristic of Rapaport's rule.

But the real puzzle is that we see the greatest diversity of human languages not where people are most spread out, but where they are most closely packed together. Take the case of Papua New Guinea. That relatively small land mass - only slightly larger than the American state of Texas - hosts 800-1000 distinct languages, or around 15 per cent of all languages spoken on the planet. Incredible as it sounds, there are parts of north-east Papua New Guinea where a new language can be found every few miles. By comparison, in China, which is roughly 12 times the size of New Guinea, only about 90 languages are spoken. Why?

One of the best known trends of biogeography is Rapaport's Rule, which describes the great increase in the diversity of species as you move from the poles to the Equator. Could Rapaport's Rule hold true for human languages too? Anthropologist Ruth Mace and I decided to investigate.

We started by gathering information on the distribution of around 500 different Native American tribes before they had any contact with Europeans. We used this to plot the number of different language groups found at each degree of latitude in North America, from the Equator to polar regions. We found more languages per unit area in lower latitudes with only a handful at extreme northern latitudes.

For comparison, we then plotted the number of different mammal species at each latitude. To our surprise, the two sets of numbers fell almost on top of each other. Human language groups seem to partition the landscape much like biological species do. It might not be surprising that few mammal species and few language groups are found in the polar regions - after all the landscape is harsh. And the large numbers of different species in the tropics might just reflect the variety and richness of resources in that environment. But, unlike other animals, humans are all the same species, so why are there so many different language groups in this region instead of one large and cooperative society?

Punctuational bursts and language 'divorce'

Our research suggests that, throughout history, human societies have had a tendency to divide into smaller competing groups as soon as they can control enough food and other resources to support themselves. In the rich environment of the tropics, smaller numbers of people can form a viable group and this might be why we find so many different societies there. Over time, these smaller groups develop their own customs and rules and after just 500 years of isolation they could be speaking a new language.

The high density of languages in the tropics suggests that language is important for more than just communicating; we also use it to establish and preserve tribal identities, to distinguish ourselves from others and maybe even to prevent eavesdropping.

There are intriguing anthropological accounts of tribes deliberately changing elements of their languages - mostly vocabulary - for exactly these purposes. A group of Selepet speakers in New Guinea met one day and collectively decided to change their word for 'no' from bia to bune, to be distinct from other Selepet speakers in a neighbouring village. The change was made immediately. One can only sympathise with anyone who was away hunting at the time.

To see if this 'language as identity' effect held more generally, we studied large families of related languages, including the Indo-European languages (Western Europe, Iran, Pakistan and much of India), Austronesian languages and Bantu languages of Africa. By building family trees of these languages - called phylogenies - we could see the number of times a contemporary language had split or 'divorced' from a related language throughout its history. Some had a history of many of these splits, others far fewer. We found that the more divorces in a language's history, the more it had departed from its ancestral or ancient language in its vocabulary. It is not just that these languages have a longer history - all of the languages we compared trace their history back to a common ancestor.

It seems, then, that when languages split they experience short episodes during which they change rapidly, or what evolutionary biologists have called 'punctuational' evolution. American English might itself provide an example. Many of the peculiar spelling differences between British and American English - such as the tendency to drop the 'u' in words like colour and honour - arose almost overnight in a punctuational burst of change when Noah Webster introduced his American English Dictionary at the start of the 19th century. Webster deliberately changed these spellings insisting that "As an independent nation, our honour requires us to have a system of our own, in language as well as government."

Future of language diversity

What does the future hold for languages? In short, mass extinction. Even though contemporary languages continue to evolve and diverge from each other, the rate of loss of minority languages now greatly exceeds the production of new languages. Currently around 30-50 languages disappear every year as the young people of many small tribal societies adopt majority languages spoken nearby. This rate of loss equals or exceeds (as a percentage of the total) the loss of biological diversity through loss of habitat and climate change. The increasing interconnectedness of the modern world along with mass communication is having a homogenising effect on language. Already around 15 out of our 7,000 languages account for around 40 per cent of the world's speakers, while the majority of languages have very few speakers at all. This is the effect of people leaving their native languages and heading towards majority ones.

There is no reason to believe that this loss of linguistic diversity means the world is losing unique styles of thought; contrary to a widely held belief, our languages do not determine how we think. But the loss of languages often does coincide with the loss of cultural diversity, so the world is steadily becoming a more culturally uniform place.


Mark Pagel is a fellow of the Royal Society and professor of evolutionary biology at the University of Reading. His book Wired for Culture: Origins of the Human Social Mind was published earlier this year by Penguin (UK) and Norton (US).