Published online 13 October 2010 | Nature | doi:10.1038/news.2010.537


Societies evolve in steps

Political complexity increases gradually — but can decline rapidly.

'Review of the War Galleys at Tahiti' oil on panel by William Hodges. Dated 1776A study of different cultures in South East Asia and the Pacific (including Tahiti pictured above) looks for patterns in the way societies change over time.National Maritime Museum/ Ministry of Defence Art Collection

Human societies progress in small steps just as biological evolution does, according to a study of the structure and language of societies in South East Asia and the Pacific Ocean.

"One of the big debates in anthropology has been whether there are any recurring patterns or processes in the way societies change over time," says Tom Currie of University College London, who led the study published in Nature today1. He and his team wanted to know whether societies increase in complexity through a limited number of different forms — from tribe, to chiefdom, state and empire — or whether different societies each have their own pattern.

Their analysis, which uses quantitative methods borrowed from genetics, supports a popular model of political evolution which suggests that societies show a gradual increase in complexity. But the data also back up another theory — societies can decrease in complexity, too, either by the same pattern of small steps or by bigger crashes.

Anthropologists have traditionally looked at human societies by conducting field studies of different groups, interviewing people and observing them, and using archaeological data to get an idea of how cultures change through time.

But these methods have their limitations — the archaeological record doesn't preserve information about social and political structure very well, and field studies don't tell you anything about the past. "A lot of the debates in this field have been purely verbal debates and descriptions," Currie says.

Language phylogeny

Currie's team borrowed an empirical approach from phylogenetics — the study of evolutionary relatedness. But instead of looking at genes, they used a recently created family tree2 of 400 languages in South East Asia and the Pacific (also called Austronesia). This region stretches from Taiwan in the north to New Zealand in the south, and from Madagascar in the west to Easter Island in the east. The area includes provinces such as Bali and Java, as well as small, autonomous groups such as the Iban speakers living in Malaysia.

The tree shows relationships between the languages over time. Two languages with many differences would be placed on distant branches, just as two species with the most genetic divergence would sit at opposite ends of a phylogenetic tree.

The team then noted the types of society that are found in the region today, and superimposed that information onto the tips of the tree. Working back through the branches, they estimated how societies had changed and evolved over time, using language as a proxy measure.

"Once I read the paper, it was so obvious that this was the best method to solve questions about the evolution of political complexity — but it had never been done before," says Jared Diamond of the University of California, Los Angeles, who has written widely on the evolution of human societies and is author of an accompanying News & Views article on the research.

"If one is trying to work out phylogenetic relationships, obviously the best way to do it is by a quantitative analysis, rather than hand-waving and pulling out examples," he says.

Radical approach

But others in the anthropology community are more hesitant. "Their objective is anthropological. But their method is entirely different to anything we've ever used," says Robert Carneiro, an anthropologist at the American Museum of Natural History in New York.

Gary Feinman, an anthropologist at the University of Illinois at Chicago, points out that analyses of political evolution usually look at changes to the political structures of one region over a given time period.

But the data used by Currie and his team relate instead to the spreading out of people and their languages over a wide geographical area — so the ways in which societies might change structure in that situation might be rather different. For example, a small group breaking away from the mainland is a different case to a large state falling apart into several smaller groups, each of which remain in the same area.

Some anthropologists might also lack familiarity with these heavily statistical methods taken from genetics. "Even supposing I knew these statistical techniques, just how you get from a linguistic phylogeny to political evolution, I don't know," Carneiro says.


Diamond defends the method. "The languages are not used to derive any results at all about societies, but to work out the phylogenetic tree. And once that's worked out, you can use that tree to study — in this case — political evolution. So the only question would be 'are languages a good way to work out relationships between societies?'" In general, Diamond says, languages do fit the bill.

But there are some problems, says Feinman. "There can be other reasons why language families may be similar," he says. Words can be borrowed by one society from another without necessarily having a common origin, for example.

"The general conclusions are reasonable, and the approach is certainly innovative," says Feinman. "But they need to have a little better feel for the socio-cultural issues they're trying to speak to."

Currie is keen that previous work in anthropology and sociology is not overlooked. "We don't want to throw that all out of the window," he says. "But we want to be more rigorous about how we test [these theories] and more explicit about what we're talking about." 

  • References

    1. Currie, T. E., Greenhill, S. J., Gray, R. D., Hasegawa, T. & Mace, R. Nature 467, 801-804 (2010). | Article
    2. Gray, R. D., Drummond, A. J. & Greenhill, S. J. Science 323, 479-483 (2009). | Article | ChemPort |


If you find something abusive or inappropriate or which does not otherwise comply with our Terms or Community Guidelines, please select the relevant 'Report this comment' link.

Comments on this thread are vetted after posting.

  • #60886

    I think this study misses the point completely by focusing on humans and human societies. There is far too much social, psychological, etc baggage associated with such a debate and it ultimately conveys little beyond the obvious.

Commenting is now closed.