Languages including Hindi and English traced to ancient Turkey
Research overturns dominant theory of origin of Indo-European tongues, from English to Hindi
Researchers have concluded that Turkey was the birthplace of hundreds of languages as diverse as Hindi, Russian, Dutch, Albanian, Italian and English, using complex computer models originally designed to map epidemics.
Similarities between hundreds of languages spoken from Iceland to India have led to hot debates over where they originated and what their spread and evolution can tell us about early humans.
The dominant theory had been that the Indo-European languages now spoken by some three billion people came from Bronze Age nomads who used horses and the wheel to spread east and west from the steppes near what is now Ukraine around 5,000 to 6,000 years ago.
Others had argued that it was agriculture - not the horse - that helped spread the language. They traced the origins to Turkey around 8,000 to 9,500 years ago.
This latest study used a massive database of common words, or cognates, both modern and ancient, to trace the roots all the way back to Turkey.
"This is one of the key cases put forward for agriculture being an important force in shaping global linguistic diversity," said lead author Quentin Atkinson, an evolutionary psychologist at the University of Auckland in New Zealand.
The results build on archaeological and genetic research which suggest that early human migration helped spur the spread of agriculture.
"It wasn't just that all the hunter-gatherers were in Europe and looked over the fence and saw [that] their neighbours were cultivating, and started doing it themselves. There was a real movement of people," he said.
"The languages suggest this is a movement of culture as well. The hunter-gatherers weren't just picking up a plough, they were also adopting culture and the language."
Methods originally designed by epidemiologists to trace disease could be applied to language because of the similarities between the evolution of living creatures and living languages, Atkinson said.
"Darwin talks about it in
The Origin of Species and
The Descent of Man. 'These curious parallels', he calls them," Atkinson added.
Biologists tracing the roots of a global pandemic will take samples in multiple locations, sequence the DNA and map how the virus has evolved through time by looking at how its genes have been modified.
"Once they've got the family tree … they can trace back along the branches of the tree all the way back to the origin," Atkinson said. "What we did was apply the same kind of approach to languages."
The team built a database of cognates such as mother, which is
moeder in Dutch,
madre in Spanish,
mat in Russian,
mitera in Greek and
mam in Hindi.
They then set about building a family tree for the languages which would capture them in space and time, and account for the gains or losses of cognates.
"This is a major breakthrough," archaeologist Colin Renfrew of the University of Cambridge in Britain said in an accompanying article in the journal
Not everyone was convinced, however. "There is so much about this paper that is arbitrary," Victor Mair, a Chinese language expert at the University of Pennsylvania, told
The Atkinson model relies on logical leaps about the rates of language change and how languages diffuse, Mair says, while the steppe hypothesis "is based heavily on archaeological data such as burial patterns, which are directly tied to datable materials".