A Blog by Ed Yong

Bacteria and languages reveal how people spread through the Pacific

The area collectively known as Austronesia covers half the globe. It stretches from South-East Asia and Taiwan, across New Guinea and New Zealand, to the hundreds of small islands dotted around the Pacific. Today, it is home to about 400 million people.

They are the descendants of early humans who spread throughout the Pacific in prehistoric times. These forebears are long dead but they left several unexpectedly important legacies that are evident in their modern descendants. The languages they used evolved and splintered into over 1,200 tongues spoken by modern Austronesians. The bacteria in their bodies did the same, giving rise to distinct strains in different parts of the region.

Two new studies have used these very different hitchhikers – one cultural and one biological – to piece together the routes of this ancient mass migration. And they have come to the same general conclusion.

The Austronesian people originated in Taiwan some 5,000 years ago. After a few centuries of settlement, they started a massive pulse of migration, spreading southwards and eastwards. They moved to the Philippines, dispersed across South-East Asia, and spread as far west as Madagascar and as far east as the Micronesian islands. They reached Fiji and other islands in Western Polynesia about 3,000 years ago and there, they paused again. About 1,500 years ago, they started a second big migration pulse that took them east across the Pacific all the way to Easter Island.


This “pulse-pause model” is one of two major theories for the peopling of the Pacific. The other – the “slow-boat scenario” – puts the origin of the Austronesians in the eastern Indonesian islands about 15,000 years ago. As rising sea levels started to submerge South-East Asia, these prehistoric people moved north into the Philippines and Taiwan and east into the Pacific islands.

These two routes have similarities but they obviously start at different points and at different times. Both archaeological and genetic studies have faced difficulties in working out which theory is correct – for example, geneticists have had trouble separating ancient genetic mingling with more recent colonial gene flow. So Russell Gray and colleagues at the University of Auckland went for a different approach – they studied the languages of modern Austronesians to piece together the travel plans of their ancestors.

Gray’s group have built a massive database of Austronesian basic vocabulary – words for common things like relatives, colours, numbers and simple verbs. Words like these (the Austronesian equivalents of “red”, “mother” or “two”) change very little and are unlikely to be borrowed from other languages, making them perfect for understanding the evolution of languages over time.

The database stores around 210 of these words per language; Gray’s team focused on 400 languages in total and pulled out sets of equivalent words, or “cognates”, from each of them. With this massive set of data, they built a family tree of the various languages. At its base lay the Formosan languages of Taiwan – they were the earliest languages of the Austronesian group. The tongues of the Philippines, Borneo, the Malay peninsula, New Guinea and the Pacific islands branched off from the main tree in that order.

This tree gives the basic sequence in which the Austronesian languages developed, and therefore the order in which the different islands and landmasses were colonised. Gray also managed to use archaeological evidence to put dates on tree’s branches. She found that it has its roots in Taiwan about 5,200 years ago and proceeded to branch with a set of rapid pulses and languid pauses. These dates, sequence of events and pace of change exactly match the predictions of the pulse-pause model and provide three strikes against the increasingly unlikely slow-boat scenario.

The pulse-pause model is an “innovationist” one – it suggests that the big expansions were driven by big advances in technology, or social organisation. In this case,  Gray suggests that the first big expansion from Taiwan was made possible by the invention of the outrigger canoe and effective sailing. These innovations turned a land-locked society into a sea-faring one capable of crossing into the Philippines and beyond. The second pulse into Eastern Polynesia involved longer journeys – perhaps it needed a better knowledge of astronomy for navigation, the creation of more stable double-hulled canoes, or the skills needed to sail across prevailing winds.

In all of these journeys, the early Austronesians didn’t travel alone. They were accompanied by the hosts of bacteria and other parasites hitching a global ride within their bodies. It’s these bugs that Yoshan Moodley from the Max-Planck Institute used on to tell the story of the Austronesian migrations. He focused on one bacterium in particular, Helicobacter pylori, a stomach bug that causes ulcers and in some cases, stomach cancer.

H.pylori is an old human companion and has followed us right from our initial forays out of Africa. Since then, different lineages have diverged from one another and evolved in relative independence due to the barriers of geography. Skip to the present, and different continents have their own distinct strains of H.pylori. With that in mind, Moodley reasoned that he might be able to piece together the relationships between modern Austronesians by looking at the differences between their H.pylori passengers.

His team took bacterial samples of 212 people from all over Austronesia and found a strain of H.pylori called hspMaori that is unique to people from this area. By analysing the genes of these bacteria, Moodley built his own family tree, which bore a striking resemblance to the one that Gray’s group built using languages. Again, Taiwan stood as the source of the Austronesian expansion and to date, it harbours the greatest diversity of hspMaori. And again, the sequence of migration progresses through the Philippines and New Guinea, before heading out into Polynesia.

On the surface, the two family trees – one bacterial and one linguistic – are very similar and it will be interesting to see if they are as closely matched at a greater level of detail. But for now, the fact that two such different approaches have converged on the same result is fascinating. In an accompanying editorial, Colin Renfrew puts it best when he says:

“[These] trees show that the past is still “within us” today. Our past is within us… when the vocabularies of specific modern languages are the basis for historical analysis [and] in a very literal way when the early history of humankind is reconstructed based on the bacterial flora in our guts. The convergence between the approaches suggests that a synthesis between linguistic and genetic interpretations of human history may soon be possible on a worldwide basis.”

References: R. D. Gray, A. J. Drummond, S. J. Greenhill (2009). Language Phylogenies Reveal Expansion Pulses and Pauses in Pacific Settlement Science, 323 (5913), 479-483 DOI: 10.1126/science.1166858

Y. Moodley, B. Linz, Y. Yamaoka, H. M. Windsor, S. Breurec, J.-Y. Wu, A. Maady, S. Bernhoft, J.-M. Thiberge, S. Phuanukoonnon, G. Jobb, P. Siba, D. Y. Graham, B. J. Marshall, M. Achtman (2009). The Peopling of the Pacific from a Bacterial Perspective Science, 323 (5913), 527-530 DOI: 10.1126/science.1166083

More on language evolution: 

More on gut bacteria: 




15 thoughts on “Bacteria and languages reveal how people spread through the Pacific

  1. This is really cool. It is very satisfying when two separate lines of evidence converge to support the same hypothesis (even if it is one that I was not familiar with until now). I wonder how easily H.pylori transfers laterally between hosts?
    I wonder if the linguists have any trouble with biblical literalists?

  2. What I don’t understand is why we have very different looking people (micro/macronesia) versus polynesia with a supposed single origin in Tawian.
    Doesn’t make sense.

  3. and spread as far west as Madagascar
    Really? I’m not familiar with marine migrations, but while trotting around the Pacific Ocean, East & South China Sea and Indonesia seems intuitively to make sense, Madagascar seems unexpected to me. Were the islands between Indonesia and Madagascar examined?

  4. Nick: As the Austronesians spread along the coast of New Guinea and into Near Oceania, they integrated with the existing populations there. Once they entered Polynesia (say beyond Vanuatu and Fiji) they stopped admixing. This is why the Polynesian/Micronesians look quite different to the “Melanesians” in Near Oceania, and why both of these groups look very different to those in Island South East Asia. This is discussed in our language paper, as well as in other papers focusing on the genetic evidence.
    BioinfoTools: Yes, the Austronesians definitely settled Madagascar, although this is quite a bit later – around 2,000 years ago. There’s some information available on Wikipedia, and this is absolutely not controversial.

  5. The Madagascar thing is absolutely right, and absolutely mind-boggling. This is what Jared Diamond had to say about it (taken from Wikipedia):

    “These Austronesians, with their Austronesian language and modified Austronesian culture, were already established on Madagascar by the time it was first visited by Europeans, in 1500. This strikes me as the single most astonishing fact of human geography for the entire world. It’s as if Columbus, on reaching Cuba, had found it occupied by blue-eyed, blond-haired Scandinavians speaking a language close to Swedish, even though the nearby North American continent was inhabited by Native Americans speaking Amerindian languages. How on earth could prehistoric people from Borneo, presumably voyaging on boats without maps or compasses, end up in Madagascar?”

    Also, have a look at the Wikipedia page while you can – some idiot’s left a hilarious fragment at the end of Diamond’s quote and inadvertently Godwinned the page.

  6. Thanks for that. Just in case you think I’m being lazy for not looking it up myself, I’m madly working on a grant application that is getting knocked around by dumb software issues. (Enough said about that…)
    I have to admit I’d have preferred an original source to Diamond, as I’m always a little wary of material from books taking on the large scale: it easy for the authors to wander off what they can judge well and I’ve seen it a few too many times and all that. But it seems it’s “known stuff” by the sounds of it. I have to admit that kind of canoe trip is something to think twice about. I’ve done a little off shore sailing myself, nothing terribly exciting, but enough to say that 3,700-odd km in a canoe is something else!
    I wonder who the wit is that added that bit to the Diamond quote. Bizarre.
    Are there any birds that migrate between Africa and Indonesia? I’m thinking along lines of a clue that “there is land over that way”. Admittedly going 3,700+ km on that basis is still a long shot, eh!

  7. I think the shocking thing is not just the fact that someone made the long trip from South-East Asia, but that no one apparently made the much shorter trip from Africa!

  8. And by the way, a massive thanks to Simon for showing up and answering questions from readers. Very rare that scientists take the time to do that, and I can only think it should be encouraged.

  9. Simon: Thanks for that, when I get time, I might look it up. Say hi to Alexei for me.
    Ed: Perhaps once they’d gone that far, any kid with the bright spark of a idea of more long-distance sailing was promptly told to go a dig the next toilet pit?! (Just kidding, obviously. If they didn’t have toilet pits, hungis)

  10. Hi, in my opinion such convergence requires one condition: the transfer of language and bacteria between two individuals should be correlated. Probably these people lived in closed, tribal communities, so the correlation was obvious; if they lived together, they shared language and bacteria; if they lived separately, they shared none of these.
    But we might think about another situation: groups of people living together (so sharing bacteria), but keeping their own languages, for example in master-slave coexistence after conquerring one nation by another).
    In such a case convergence will not appear.
    What do you think?
    – Krzysiek

Leave a Reply

Your email address will not be published. Required fields are marked *