May 6, 2003
World's Farmers Sowed Languages as Well as Seeds
he homelands of the Indo-European languages stretch from Dublin to Delhi. But Hadza, a tongue that is one of a kind, is spoken by just 1,000 people near Lake Eyasi in Tanzania. Why do the world's languages have so uneven a distribution pattern?
Two researchers theorize that much of the answer has to do with events that began 10,000 years ago, as crop plants were domesticated in different regions.
The invention of agriculture has long been invoked to explain the spread of the Indo-European languages. Now, Dr. Jared Diamond of the University of California at Los Angeles and Dr. Peter Bellwood of the Australian National University in Canberra have applied the concept to 15 major language families. Their article appeared in the April 25 issue of Science.
The premise is that when humans lived as hunters and gatherers, their populations were small, because wild game and berries can support only so many people. But after an agriculture system was devised, populations expanded, displacing the hunter-gatherers around them and taking their language with them.
On this theory, whatever language happened to be spoken in a region where a crop plant was domesticated expanded along with the farmers who spoke it.
Even if the farmers interbred with the hunter-gatherers whose land they took over, genes can mix, but languages cannot. So the hunter-gatherers would in many cases have adopted the farmers' language. That is why languages "record these processes of demographic expansion more clearly than the genes," Dr. Bellwood said.
One of the clearest expansions, perhaps because it occurred most recently, can be found in the 1,436 languages in the Niger-Congo or Bantu family, the world's largest. About 5,000 years ago, Bantu speakers in western Africa who cultivated the yam started spreading out from their homeland. One group traveled south, the other first east to the Great Lakes and then south. The two migrations spread the Bantu languages through a third of the continent, displacing the Khoisan, or click-language speakers, who were hunter-gatherers.
The agricultural regions of China made up the homelands of three major families, Dr. Diamond and Dr. Bellwood wrote. One was Austro-Asiatic, which includes a swath of languages now spoken in Cambodia, southern China, India, Malaya and Thailand. Another was the Tai group, which includes Lao and Thai. A third was the Sino-Tibetan family.
In the New World, the farmers who domesticated maize and beans in Mexico expanded northward to the area that became the southwestern United States, spreading the Uto-Aztecan family of languages.
Austronesian, a group of 1,236 languages, is the second largest, after the Niger-Congo. The founder language was spoken by rice growers in southern China who colonized Taiwan before 3000 B.C. and spread through Polynesia, reaching New Zealand by A.D. 1200.
The Science article endorses a bold suggestion for the origin of Japanese. The writers say it is derived from the language of rice farmers who arrived from Korea around 400 B.C. and spread their agriculture northward from a southern island, Kyushu. Modern Japanese is not at all like Korean. But Korea had three ancient kingdoms, each with its own language. Modern Korean derived from the ancient Sillan. Japanese may have evolved from another ancient Korean language, Koguryo, the article says.
Just as China was a powerhouse of new language families in the East, the Fertile Crescent, the arc running through Lebanon and through Iraq, was the source of at least three major language families in the West, the authors say.
One was Dravidian, a language family now centered on southern India. A second was the Indo-European family, which includes English, French and German in its Western branch and Iranian and Hindi in its Eastern branches. A third may have been Afro-Asiatic, a family that includes ancient Egyptian and Semitic languages like Arabic and Hebrew.
The best-known movements of people are those of conquering armies like the Mongols, who overran much of Eurasia. But soldiers are often too few in number to impress their language on a population. The Mongols, because of their rulers' harems, left behind many more genes than conquering armies usually do, but their language vanished from their conquered territories.
"These early language spreads were essentially driven by demographic processes, or colonization, to use another word," Dr. Bellwood said.
Dr. Diamond said that agriculture did not drive all language expansions — the Inuit's spread across the Arctic is an example of that — but that "for most of the widespread language families the driving force for the spread has been agricultural."
Dr. Diamond said the new theory also predicted that expansions would occur more easily on an east-west axis than a north-south axis because the crop plants on which an agriculture depends tend to be able to grow only at particular latitudes.
As to one obvious exception, the Bantu expansion southward from West Africa, Dr. Diamond said the Bantu speakers had remained in tropical Africa and had indeed failed to penetrate the temperate zones of southern Africa, meaning that they were not present at the Cape of Good Hope when the Europeans arrived.
Dr. Diamond is a physiologist with an interest in human origins, and Dr. Bellwood is an archaeologist who specializes in southeastern Asia. Their synthesis spans the fields of many specialists who are unlikely to agree with every detail.
Dr. Christopher Ehret of U.C.L.A., an expert in the history of African languages, said the authors had overstated the role of agriculture in explaining the pattern of language distribution.
"In reality, the spread of language families has come about for different reasons in different times and places, but one of the causes has sometimes been the development of agriculture," Dr. Ehret said.
He said he did not agree with Dr. Bellwood that the Indo-European languages had been spread by farming. Linguistic evidence shows the speakers of the ancestral Indo-European tongue knew of wheels and kept horses in years around 4,500-3,500 B.C., but agriculture had spread to Europe at least 2,500 years previously, Dr. Ehret said.
He also disagreed that the Afro-Asiatic languages were spread by farmers from the Fertile Crescent. Afro-Asiatic arose in in northeastern Africa 13,000 years ago. The Semitic branch spread to southwestern Asia 5,000 years ago, Dr. Ehret has written, with two much more recent back migrations, giving rise to the Amharic language of Ethiopia and the Arab languages of North Africa.
But Dr. Colin Renfrew, an archaeologist at Cambridge University, said although he disagreed with Dr. Diamond on some aspects of Indo-European, "I expect that his synthesis will be useful."
And Dr. Merritt Ruhlen of Stanford, an expert on language families, said the two authors had put together a "very useful overview."