Austronesian-speaking peoples inhabit vast swathes of the Indian and Pacific oceans, from Madagascar and island Southeast Asia to Tahiti and New Zealand. The remarkable feats of navigation that brought them to these islands occurred far back in time, such that languages across the islands had diverged into related but recognizably different forms, and distant Islanders had lost memories of each other.
One of the most influential hypotheses on the spread of Austronesian-speaking peoples was proposed and popularized by the anthropologist Peter Bellwood through the closing decades of the last century. Drawing on linguistic – primarily the great diversity of Austronesian languages spoken by Indigenous Taiwanese peoples – and archaeological evidence, Bellwood hypothesized that Austronesian-speaking peoples in Taiwan began experiencing major population growth around 5,000 years ago which, over the next few millennia, drove migration to island Southeast Asia and Oceania. Proponents of Bellwood's Austronesian expansion hypothesis often link this population growth to highly productive wet rice agriculture taking hold in mainland Asia and subsequently Taiwan.
In recent years, a simple “Out-of-Taiwan” expansion driven by agricultural productivity has been challenged on various fronts. Studies on pigs, bananas, and other domesticates of the Pacific seem to imply complex, multidirectional networks of cultural exchange among Taiwan, island Southeast Asia, and the islands of Oceania. At the same time, archaeologists have argued that wet rice agriculture is unlikely to have been a driver of migrations and in fact, probably promoted population packing.
Against this backdrop of debates surrounding the “Out-of-Taiwan” model, we saw an opportunity to ask whether rice did come to island Southeast Asia from Taiwan. We were well-placed to undertake such an investigation as our lab had just published a study in 2020 that reconstructed how rice spread across Asia based largely on whole genome re-sequencing data. While we had found that japonica rice had begun to diversify and spread by about 4,000-5,000 years ago, a lack of Indigenous Taiwanese rice among the analyzed samples, however, left open the question of the role of Taiwan in the spread of rice to island Southeast Asia.
The seeds of the study had been sown even before the 2020 study was published. In November 2019, one of us (Michael) gave a talk in Taipei where he reported on our then unpublished work on how japonica rice spread in Asia. After the talk he was approached by Yue-ie (Caroline) Hsing of the Academia Sinica and her student Cheng-chieh (David) Wu, who told him that they had actually sequenced 24 japonica rice landraces that were collected from the Indigenous peoples inhabiting the central mountain range of Taiwan. This was the missing piece.
Together, Michael and Caroline agreed to a collaboration. Michael also remembered that several years back he had hosted Karen Hicks as a sabbatical visitor from Kenyon College, and she had sequenced some traditional rice varieties from the Cordillera peoples of Northern Luzon in the Philippines. Contacting Karen, she immediately agreed to integrate her data with those from Taiwan.
One of us (Ornob) joined the lab in the summer of 2020 and was assigned the task of answering this question as the first project for his dissertation. Under the tutelage of soon-to-depart postdoc Rafal Gutaker (now a Group Leader at Royal Botanical Gardens, Kew), he embarked on his first foray into population genetics.
Combining our new samples with those from the 2020 study, we retraced the route japonica rice took to island Southeast Asia using admixture graph and demographic modeling. Strikingly, we found that Indigenous Taiwanese rice had split off from northeast Asian temperate rice about 2600 years ago. Subsequent gene flow between Taiwan and the Philippines was primarily in the direction of Taiwan and had started before about 1300 years ago, indicating that Taiwan appears to have been a contact zone between the north and the south.
What we also found was that the rice of the northern Philippines did not originate from Taiwan but instead likely came up from the south (Borneo and the Indonesian archipelago). Indeed, our studies suggest that japonica rice in the Philippines may have split off from the Indonesia populations as early as 3500 years ago.
When thinking over such timescales, archaeological data can support or undermine claims. Our long-time collaborator, the archaeologist Dorian Fuller at University College London (UCL), compiled extensive archaeological data to show that movement of japonica rice from the Philippines to Taiwan between 2600 and 1300 years ago would have been within the bounds of known trade networks exchanging bronze and jade between parts of Taiwan and Island and mainland Southeast Asia. Cristina Castillo and Stephen Acabado, both archaeologists that focused on Southeast Asia, also helped us interpret our results. Two other collaborators - Jade d’Alpoim Guedes at the University of California at San Diego and Kyle Bocinsky at Crow Canyon Archaeological Center – helped us look at paleoclimates in Taiwan during the period we were studying.
Overall, our findings reveal complexity in the movement of rice to Taiwan that is rarely considered in mainstream thinking on the Austronesian expansion.
As our paper was being reviewed, two studies came out further contradicting a linear, farmer-driven “Out-of-Taiwan” model for the spread of Austronesian-speaking peoples. The first showed that Philippine Cordillerans had begun branching off from the Indigenous Taiwanese by at least 8,000 years ago. These were likely hunter-gatherers that arrived long before the emergence of agriculture in the Philippines.
The second of the studies showed that the ancestors of Austronesian-speaking Pacific Islanders likely split off from the Indigenous Taiwanese before 5,000 years ago. This predates their dates of arrival in Oceanian islands based on archaeological evidence, indicating that the Philippines, or some other part of island Southeast Asia, may have been the proximal source of these migrations rather than Taiwan.
Both studies, in their conclusions of earlier dispersal from Taiwan, echo older work on mitochondrial genomes, that had suggested a split between island Southeast Asians and Indigenous Taiwanese mitochondrial lineages before the timeline of the hypothesized “Out-of-Taiwan” event. They argued instead that climate change and rising sea levels around the beginning of the Holocene may have driven population splits and migrations between ~15,000 and ~5,000 years ago.
These findings from recent and older studies on human genomes suggest that the spread of Austronesian languages and cultural elements like pottery may represent the expansion of a dominant culture among peoples who were already connected by shared ancestry and trade or migration networks. Our study provides further evidence against an agriculturalist expansion, and points to the existence of networks of movement and exchange – likely with deep historical roots – along the islands off mainland Asia and into the Pacific and Indian Oceans.