Stephen Oppenheimer (Independent Scholar, Oxford, UK)

The question of the origins of the Polynesians has, for over 200 years, been the subject of adventure science. Since Captain Cook’s first speculations on these isolated Pacific islanders, their language affiliations have been seen by many, in particular migrationist scholars, as the solution. The geographic and numeric centre of gravity of the Austronesian language family is in island Southeast Asia, which was therefore originally seen as their dispersal homeland. However, another view has held sway for fifteen years, the ‘out of Taiwan’ model, popularly known as the ‘express train to Polynesia’. This model, based on the combined evidence of archaeology and linguistics, proposes a common origin for all Austronesian-speaking populations, in an expansion of rice agriculturalists from south China/Taiwan beginning around 6,000 years ago. However, it is becoming clear that there is in fact rather little supporting evidence in favour of this view. Alternative models suggest that the ancestors of the Polynesians achieved their maritime skills and horticultural Neolithic somewhere between island Southeast Asia and Melanesia, at an earlier date. Recent advances in human genetics now allow for an independent test of these models, lending support to the latter view rather than the former. Although local gene flow occurring between the two bio-geographic regions may have been the means for the dramatic cultural spread out to the Pacific, the immediate genetic substrate for the Polynesian expansion came not from Taiwan, but from east of the Wallace line, probably in Wallacea itself.


Like most authors I look forward to constructive criticism. I recently received an offprint of a review, written by Peter Bellwood, of my book Eden in the East. The book was published three years ago in 1998 and the review in 2000. Had I seen the review before it went to press, I would have been glad to correct the numerous mis-readings of my text and bibliography that it contained. I hope to set the record straight here.

Bellwood’s book review was given as part of a lecture to the Japanese Society for Oceanic Studies meeting in Japan 1999 and then published as a paper. The paper starts with a three page re-statement of Bellwood’s own particular view of Oceanic and in particular Austronesian prehistory, set as eight questions and answers. There is then a bridge. This first compares various scholarly approaches to Oceanic prehistory – grand-scale interpretive versus small-scale detailed investigation of a few islands; cultural phylogeny seen as an extension of migrationism versus analysis of cultural evolution as a result of interchange between reticulated spheres of interaction. Bellwood states his preference, at least in the case of Pacific prehistory, for the former approach in each of these two comparisons.

Having admitted to perhaps unfashionable tendencies himself, Bellwood then draws a line between such scholarly revisionism and the use, in other peoples’ grand interpretations, of what he implies to be poor data. It becomes clear later that some of these ‘data’ may include publications by archaeologists with whom he does not agree (e.g. Pookajorn, Meacham, Chazine, Chia and Solheim). He then introduces “Eden in the East” as an example of such mis-use with an explicit link to Atlantis books: “…claims from the archaeological literature to postulate the presence of an Atlantis-like culture… with an…agricultural economy” The introduction of Atlantis claims sets the tone of the review.

As author of Eden in the East I chose to avoid making any such claims (location of the mythical Atlantis or the presence of a super-culture) in order, vainly as I now see, to escape such ‘critical’ tactics. As far as horticulture is concerned I did question the assumption that absence of evidence for root-crop domestication in Island Southeast Asia (ISEA) before Bellwood’s hypothetical date of agriculture movement out of China into Taiwan (and thence to ISEA) was evidence for its absence. Work on phytoliths of taro and of other root crops may resolve this question soon for ISEA. The reviewer himself acknowledges the difficulty of finding such evidence (end p.10). I also cite Solheim with a similar view of early root-crop horticulture on Mainland Southeast Asia, and also Solheim’s view that rice spread from Borneo to the Philippines rather than the other way round.[2] I mention this also in view of the argument for early root-crop horticulture in Papua New Guinea from 9,000 BP (Eden p.71). As an aside on the issue of data quality, I gather that, unlike Bellwood, not all archaeologists even accept the evidence for the latter at Kuk in the New Guinea highlands.

While on the lack agricultural evidence, I am not clear quite why Bellwood feels that I am using Surin Pookajorn’s evidence for early rice use on mainland Southeast Asia as the major archaeological weapon for my own thesis. On the contrary, I devote considerable space to questioning the relevance of rice agriculture to the bulk of the Austronesian language expansions. The point I am trying to make with the Sakai cave evidence and with the pre-Philippines rice dates from Sarawak caves is that the existing evidence is not strong enough to pin down the date of introduction of rice domestication in Southeast Asia to only 4000 BP and via the Philippines. Such evidence would be a necessary proof for the Bellwood hypothesis that a rice-growing people invaded from the Philippines and swamped a purely hunter-gatherer population in ISEA.

Putting aside the opinions of other archaeologists, with whom he does not agree, the reviewer flatters me with more archaeological re-interpretation than I suggest on my own part. Far from trying to claim a firm archaeological reconstruction of ISEA prehistory I repeatedly point out the very paucity of material cultural remains of any description on the rump Sunda shelf before 5500 BP, which was the high-water point of the post-glacial sea-level rise (e.g. see p.80). Maybe that is why post-glacial early Holocene ISEA prehistory is so poorly characterised; but the same limitation of cultural remains applies to any archaeological reconstruction of this period, including Bellwood’s. For this reason, much of the book deals with other sources of evidence of cultural and demic dispersals in the region including that most direct measure of human migration: genetics.

Following the ‘Atlantis’ shot, the reviewer states “I do not wish here to argue in detail against Oppenheimer’s linguistic and genetic hypothesis; I am neither a linguist nor a geneticist and would prefer to leave these issues to those of my colleagues who are better equipped in these disciplines than me.” This humble protestation seems disingenuous in view of Bellwood’s huge published output over the past 30 years using evidence from both of these as well as many other disciplines. He has shown no such coyness to discuss other disciplinary evidence, from what I have read of his work. I would certainly not challenge his right to cross into other people’s turf, carefully assess their arguments (rather than technique) and then make his synthesis. On that basis, I do not believe he should object to others treading that same contentious grass.

I find Bellwood’s unwillingness to confront my specific criticisms of his own model rather frustrating and puzzling. In my book I state my sincere admiration for his massive contribution to Austronesian studies, but register specific and extensive disagreement with his own grand interpretation. As a foremost authority on Southeast Asian prehistory, Bellwood has done me a great honour in devoting most of one published paper and an international address, to critically reviewing a book, explicitly written for a general audience, as if it were an academic text. This implies that he takes the challenge seriously. But you cannot review a film on the basis of the credits at the end. He seems more concerned to pick up perceived details of factual inaccuracy in archaeological citation than to defend his own grand interpretive synthesis against my specific criticisms. These concerns are also shared by academics far more versed in this field than I am. I cite a number of other critics in the endnotes to my book.

The reviewer speculates on page 12 whether he would ever change his views and register any agreement with the thesis presented in Eden in the East. There is now substantial academic rejection of his own theory of a rice-driven demic expansion out of Taiwan to occupy and swamp the islands of Southeast Asia and the Pacific, popularly labelled “the express train to Polynesia”. Such is clear from a recent review in Science,3 for which both of us, and a number of linguists, geneticists, anthropologists and archaeologists were interviewed. I quote several excerpts:

‘“I have to write a review myself of the spread of early farmers, and it’s very difficult,” says archaeologist Peter Bellwood of the Australian National University in Canberra. “It’s the genetics that is causing headaches”’
‘“I don’t think there’s any question that the Austronesian expansion comes out of island Southeast Asia,” says archaeologist Patrick Kirch of the University of California, Berkeley. “The danger is getting too specific about Taiwan when we don’t know enough archaeologically about the coastal China area, Taiwan, or the Philippines’”

‘At this point, it’s hard to find any archaeologist who admits to riding the express train, including Bellwood, whose name is tied most often to that model. … “I don’t believe in express trains,” says Bellwood’ (Gibbons 2001)[3].

Notes on Bellwood’s review

For the record, I attach below some responses to points raised at the end of the review:

P.11 Para 4. Quote: “It is not possible, using available evidence, to push Proto-Austronesian back to 8000 BP and to locate it on the equatorial Sunda shelf.” My view is that it is equally impossible to disprove it, language dating being as imprecise as it is (Dixon 1997)[4].

P11. Para 5. Bellwood makes it clear in this paragraph and in the phrase “…movement of native speakers…” that, in spite of genetic evidence to the contrary given in Eden in the East (and extensively augmented subsequently[5],[6],[7]), he still believes this was a mass migration spreading from Taiwan to Polynesia, with little biological mixing, rather than a mainly cultural spread with local acculturation.

P12. Last sentence before ‘Notes’. Bellwood’s point that:

“pots don’t speak or carry genes. If they did, we probably wouldn’t have very much to argue about, at least not in terms of ethnolinguistic history” seems a strange concluding remark for someone who lays such store by a marriage of archaeology and historical linguistics. It is strangely similar to my own statement in Chapter 2 of Eden in the East, except that it misses the point that such a marriage is in danger of incest:

Because pots cannot talk and language splits are difficult to place and date, much of the argument relating to Austronesian origins depends on mutual underpinning by linguistics and archaeology. It is very important, therefore, that each piece of evidence in the structure be independent and not reciprocally linked to the other discipline as Blust himself has pointed out (Eden in the East p. 66).

P12. Note 5: Note 5 contains a number of bulleted sub-notes nearly all related to details on citations given in Eden in the East as endnotes. I answer them all below, in note-form, with pagination, in order of bullet points. The first four of these sub-notes complain of lack of references yet they refer to sections in my Prologue and Introduction. A glance at the endnotes would have revealed that these two introductory sections (pp. 1-21) have no endnotes. This was at the request of my scientific editor, since introductions, like abstracts, are not generally referenced. In most cases where the reviewer fails to find a reference, the reference is there. My editor was (probably rightly) horrified at the number of references (originally 100 pages) and they were cut by 30% generally by merging multiple citations to the same text. Other academic authors writing in the same non-academic genre have been more radical, some even dispensing with a bibliography altogether.[8]

· P.4 Wild taro in Indonesia. The quotation was in the Prologue, which was an unreferenced section.

· P.5 Ban Chiang dates, and bronze. I make my differing interpretation quite clear here and in the later reference below. I do not mis-cite Joyce White.

· P.18 Pottery appears in China from 9th Mill. BC and in the Spirit Cave in Thailand from 8th Mill. BC Again in this introductory section there are no references, but an easily accessible summary of early East Asian pottery dates may be found.[9]

· P.19 No reference was given for Solomon Island grindstones with wild cereal starch residues for the reason given above. In any case, Bellwood should know the original paper on the cave, occupied during the Pleistocene on Buka Island, since it was co-authored by Professor Matthew Spriggs who checked his review and is in the same department. The reference was in fact given later in the book – see below.

· P.67 Bellwood, betel nuts and Austronesians: see Bellwood, P. (1997) pp. 108, 111, 135, 152. [10]

· Sakai cave rice dates: see above.

· P.69 Ban Chiang & Joyce White: see above

· P.69 Bellwood and Hemudu. I did not say that Bellwood states that Austronesian cultures originated in Hemudu. The word I used was believed. Having re-read the relevant section in his book (1997: 208-214 ), I still feel anyone could be forgiven for thinking Bellwood believed Hemudu rice cultures to be the original proximate source of his reconstructed AN rice culture in Taiwan. He bemoans the lack of references here. Again, my editor reduced the total number of references to Bellwood’s own book alone, since it was the most commonly cited source in this section.

· P.70 The Yangtze and optimal regions for rice growing. I was quoting from the map printed in Bellwood’s own book (1997:243 Fig 7.17) and making the point that the natural homeland areas for rice cultivation are centred around Indo-China not the Yangtze valley. Incidentally, the period Bellwood has chosen in his review (6-8,000 BC) is not the period I referred to in this paragraph.

· P. 77 Prehistoric cultural contact between Ur and East Asia. This last page of a chapter is explicitly interpretive and speculative and not a statement of fact or observation. It is misleading for the reviewer to imply otherwise.

· P.86 The abandonment of Madai cave around 7,000BP. First of all in spite of Bellwood’s denial I did give a reference for the inexplicable abandonment of the cave – from his own book. I did not give one for the ‘cause of abandonment’ since this was explicitly my own speculation. Several other points should be noted. Bellwood may be unaware of research into the destructive nature of giant tsunamis which can (and have been recorded to) ascend to some altitude (500m) up mountains and to penetrate far inland (discussed in chapter 1 of Eden in the East). In Bellwood’s book, the people who used the Madai caves were also “…visiting the encroaching coast fairly frequently.” Unless they were all perched on the limestone massif there is every chance they could have been in trouble from such tidal waves.

· P.88 In spite of Bellwood’s denial, a reference was given for cord-marked pottery in East Kalimantan dated to 5500 BP – Chapter 3, endnotes 22 & 23. I did and do not claim the same date for the Kain Hitam paintings in spite of the reviewer’s claim that I did.

· P.89 Sanskrit, Chams and Austronesian. The citation is, nearly verbatim from the introductory sentence of Ian Glover et al.:

‘a number of stone inscriptions written in Sanskrit and in an Austronesian language ancestral to that used by the Cham people’[11][12] Bellwood was the editor of this publication.
It is possible that the literal meaning could have changed in my transcription from the original by the omission of ‘and’; but this is not relevant to, and does not alter the point being made in my text.

· P. 96 Pig bones in the New Guinea Highlands between 5000-8000 BP. I actually cited 5000BP (Kirch, 1997; p.43),[13] not the upper margin. In any case, the fact that as Bellwood states, he personally disagrees with such claims in the archaeological literature and published his disagreement at the same time my book was published, hardly seems to warrant an inclusion in this litany of supposed errors and distortions. The same applies to the other archaeologists whom I cite and he disputes.

· P.96 Matenkupkum pottery. The reference was in fact given.

· P.97 Pre-Lapita pottery in northern New Guinea. I cite Pamela Swadling for this in my book.. Presumably here is yet another archaeologist Bellwood disagrees with.

· P.91 Date of Austronesian settlement of southern Vietnam. I referred specifically to the linguistic and not the archaeological chronology here. This is as given in Blust (1985)[14]. Bellwood’s own archaeological timeline and detail of events for Southeast Asian prehistory is, I realise different in a number of respects from Blust’s linguistic one. In any case Bellwood’s 1997 edition of his Indo-Malaysian book was unavailable to me at the time of writing10.

· P.91 Bukit Tengkorak in Sabah and the date of obsidian imported from Melanesia. I do realise that Bellwood was the original excavator of the site and that he disagrees with a subsequent report by Malaysian archaeologist Stephen Chia on fresh excavations from the same site. I was, however, citing Service (Science, vol. 274, 1996, pp. 2012-3) on Chia’s unpublished findings correctly and in good faith. At the time of writing, Service’s secondary report was the only one available to me.

· P.94 Settlement of the North Solomons from New Caledonia. As Bellwood suggests there is a typo. New Caledonia should read New Ireland. In spite of Bellwood’s repeated incredulity for such an early colonisation of the North Solomons, reference was however already made on P.92 (in full in endnote 34). Since the reviewer has already complained earlier (see above) that this finding was unreferenced I give it again here.[15]

· P.96 Jomon pottery in Vanuatu. It is not clear what Bellwood is trying to say here, unless it is that I am citing an unpublished rumour. He is wrong saying that I am citing a paper published in 1999 a year after my book was published. I wish, but like the most of us I do not have the power of foresight; I was instead correctly quoting from a paper given at a conference in Vanuatu in 1996.[16] In any case I can hardly see why my citation should be regarded as ‘in error’.

· P.130 The reviewer attributes me with the remark that “Most Laotians speak Austro-Asiatic languages”, whereas I say “Austro-Asiatic languages are spoken by most Vietnamese and …” The ‘most’ refers to Vietnamese.

It is difficult to see the motive for such highly-spun and detailed objections to my bibliography, when my objections to Bellwood’s own synthesis received no comment. I do hope the reason was not that my book was perceived as a personal attack, with the review being returned in the same vein. This would be sad; because if he had read my endnote 46 in Chapter 2 the reviewer would have seen that this was expressly not my intention.

[1] Bellwood, P. 2000 Some Thoughts on Understanding the Human Colonisation of the Pacific People and Culture in Oceania 16: 5-17

[2] See Solheim, W. 1994 Southeast Asia and Korea: from the beginnings of food production to the first states, in UNESCO The History of Humanity: Scientific and Cultural Development, vol. 1: Prehistory and the Beginnings of Civilisation, Routledge, London, pp. 468-481

[3] Gibbons, A. 2001 The Peopling of the Pacific Science 291: 1735-7

[4] Dixon R.M.W 1997 The rise and fall of languages CUP, pp.46-9

[5] Richards, M., Oppenheimer, S. and B. Sykes 1998 mtDNA suggests Polynesian origins in eastern Indonesia. American Journal of Human Genetics 63:1234-6

[6] Oppenheimer S.J. and M. Richards 2001 Polynesian Origins: Slow boat to Melanesia? Nature 410: 166-7

[7] Oppenheimer, S.J. and M. Richards 2001 Fast trains, slow boats, and the ancestry of the Polynesian islanders. Science Progress 84 (3):157-181

[8] e.g. Sykes, B. 2001 The Seven Daughters of Eve. Bantam, London

[9] Scarre, C. (ed.) 1995 The Times Atlas of Archaeology. Times Books, London, pp. 12, 101

[10] Bellwood, P. 1997 Prehistory of the Indo-Malaysian Archipelago (rev ed.), University of Hawai’i Press, Honolulu

[12] Glover I.C., Yamagata, M. and W. Southworth 1993 The Cham, Sa Huynh and Han in early Vietnam: Excavations at Buu Chau Hill, Tra Kieu. Bulletin of the Indo-Pacific Prehistory Association 15:166-76

[13] Kirch, P.V. 1997 The Lapita Peoples: Ancestors of the Oceanic World, Blackwell, Oxford

[14] Blust, R.1985 The Austronesian Homeland: A Linguistic Perspective Asian Perspectives, 26(1):85

[15] Loy T.H. et al. 1992 Direct evidence for human use of plants 28,000 years ago: starch residues on stone artefacts from the northern Solomon Islands Antiquity 66:898-912.; see also original report on the cave on Buka Island: Wickler S., and M. Spriggs 1988 Pleistocene Human Occupation of the Solomon Islands, Melanesia. Antiquity 62:703-6

[16] Was There a Pre‑Lapita, Japanese Jomon, Cord‑marked Pottery Occupation in Vanuatu? Sinoto, Y.H., Shutler, R. Jr., Dickinson, W.R., Shutler, M.E., Garanger, J. and T.M. Teska. paper presented to The Vanuatu National Museum ANU Conference: The Western Pacific, 5000-2000 BP; Colonisations and Transformations”, Vanuatu National Museum, Port Vila, August 1996