Sep 2, 2026
When writing Peopling of Indonesia I realised some of the gaps in my understanding of the prevailing theories, particularly about the Austronesian Expansion. In an effort to fix this I've been doing a lot of reading, and it seems like about time to attempt to consolidate this into an article. Writing it down is a good way to test I have at least some coherent grasp of the issues, and hopefully it may be of some slight interest to others.
The original Austronesian Expansion model, as proposed by Bellwood and Blust, among others, explained the spread of Austronesian culture as being a process of "demic diffusion", movement of people along with their technology and language. So according to the theory, people in coastal Southern China moved to Taiwan, taking rice and millet agriculture with them. After establishing themselves in Taiwan, some fraction of the population migrated southwards about 5000 years ago, first to the Philippines, and then eventually to all parts of Island Southeast Asia (ISEA) and beyond. Because the local populations in the places they migrated to were mostly hunter-gatherers, the incoming Taiwanese migrant farmers could compete with them successfully and the locals either adopted agriculture, and in the process also adopted other aspects of Austronesian culture, or were replaced. Beyond agriculture, the Austronesian success is explained by a "package" of culture and technology, including pottery and sailing in addition to rice, millet, chickens, pigs and dogs.
This is a simple description of the theory - there are variations, including whether the impetus for migration was population pressure caused by the increased population density enabled by agriculture, or cultural factors such as the desire to achieve status as a founder of a new community.
This theory gained widespread acceptance because it does explain the remarkable spread of Austronesian languages through ISEA, the Pacific, and as far as Madagascar, which is a phenomenon that needs explaining. It has however been under sustained attack for a number of years and it seems that at least some of the assailants have obtained some traction. The objections to the theory as stated above fall into a few categories. I'll cover the general ideas first, then dive into the particular papers where they are expounded.
At first it seemed that the genetics of ISEA and Pacific peoples were consistent with the simple Austronesian expansion. People had substantial ancestry that was clearly related to modern-day Austronesian-speaking Taiwanese people, so that seemed to back up the idea of people migrating from Taiwan. However, more recently the ability to resolve the genetics more precisely led some to conclude that the apparent Austronesian signal in ISEA and Pacific genetics came from a branching from mainland Asia at about 8kya, and not from a branch from Taiwanese populations at about the 4-5kya mark that would be needed to match the AN expansion. The mainland Asians who were the ancestors of ISEA people were not the people who lived on Taiwan. There is also a clear signal from other parts of Asia, which in my earlier article I called "Austro-Asiatic" following one particular paper but subsequent research suggests it wasn't quite that simple. Unless new developments show this genetic analysis to be incorrect, it appears to me that a substantial migration of people from Taiwan and replacement of existing populations just didn't happen.
I touched on this in the earlier article - there is actually little to no archeological evidence of Austronesian migrants cultivating rice in ISEA until much later than the AN expansion dates. No agricultural implements have been found in the right date range, and what little evidence there is of actual rice is always associated with pottery, leading to the suggestion that people may have imported the rice from elsewhere, rather than growing it themselves. Archeological evidence is by its nature sparse - there is always the possibility something new will be discovered tomorrow that changes the whole picture - but based on what we know at the moment, as far as I can tell we just don't see anything that backs up the idea that Taiwanese Austronesian migrants brought rice cultivation to ISEA in the right time period for the Austronesian Expansion.
The idea that rice cultivation would lead to sufficient expansion in population density to drive outwards migration has been challenged. A study in mainland China showed places where population density became quite high, but the population responded by further intensifying their rice cultivation and not by migrating to surrounding areas. In ancient times population densities were very low by modern standards - it's easy to forget that - so there would generally have been scope to increase the population further locally as long as agriculture was successful. In deference to Peter Bellwood we should note that not everyone agrees that this matters.
The classic Austronesian sailing vessel is the outrigger canoe, either double or single outrigger, with lash-bound planking and wooden dowels. Go to any seaside village in ISEA and you will see fishing boats constructed along these lines still. I had a ride in one from Amed, in Bali, recently, and made sure to expound to everyone within earshot how amazingly ancient and successful this design was. However, from archeology it looks like there were no such vessels made in Taiwan at that time - while people already living in ISEA had developed that technology, and anyone migrating there learned it in situ, rather than bringing the technology with them.
With some notable exceptions, the vast majority of languages spoken in ISEA are Austronesian. The exceptions are Papua and some other areas of Eastern Indonesia where Papuan languages are spoken (although from my limited understanding, Papuan languages are extremely diverse so calling them all Papuan is really just saying they are not Austronesian, rather than calling out any essential similarity between them). The other exceptions are the Negrito peoples in Peninsular Malaysia, who speak Austro-Asiatic languages distantly related to Vietnamese and Khmer. Apart from this, the other languages in ISEA, including such important languages as Malay, Indonesian, Javanese, Sundanese, Tagalog and Visayan, are all Austronesian.
This near-uniformity of language group is remarkable and needs explanation. When we get into the details things become interesting. The Austronesian language family has ten subgroups, and nine of those ten are spoken on Taiwan. The languages of ISEA and the Pacific are all members of the tenth subgroup, Malayo-Polynesian. This by itself is one of the main pieces of evidence in favour of a Taiwanese origin for the Austronesian languages - in a place where the languages have been spoken the longest, you expect more variation because more time has elapsed in which local variations can accumulate, and once you have enough local variations you eventually have separate languages. If you want a very rough but hopefully enlightening analogy, consider the variety of regional dialects and accents in English within the British Isles, compared with the relative uniformity of English in either Australia or the USA. In Australia you need to be a linguist to reliably point out regional accents, and in the USA a non-expert could probably name fewer than half a dozen different accents, but the variety in England, where the language originally developed, is immense and apparent to anyone who has met people from different parts of England.
However, getting further into the detail raises a few puzzles. The words in Malayo-Polynesian that correspond to everything in the so-called Austronesian cultural package do not correspond to words in the other Austronesian language subgroups. It looks like Malayo-Polynesian speakers developed new vocabulary for agriculture and sailing after they left Taiwan, probably by borrowing words from languages already spoken in ISEA. The implication is that this cultural package was not something the Austronesians brought with them from Taiwan. This weakens the idea that the Austronesian cultural package was responsible for their success.
This attempt at a literature survey is of necessity incomplete. I've followed the trail of who cites who a few steps and come up with that looks to me like a reasonable collection of relevant papers, but of course I'm no expert, and I don't have access to the text of all the papers, not being affiliated with any university and not interested in paying the exorbitant rates the academic publishers charge.
Vol 26, No 1 of the journal Asian Perspectives of 1984-1985 was a major event for those interested in Austronesians. There were three major papers, two of which formed the basis for the Austronesian Expansion theory, and one of which was a rearguard action against it.
Peter Bellwood, Asian Perspectives Vol. 26, No. 1 (1984-1985), pp. 107-117
Peter Bellwood is currently Emeritus Professor of Archeology at the Australian National University. This is as far as I can see the most cited paper on the Austronesian Expansion, the one considered the definitive statement of the theory. Here Bellwood helpfully puts together an assembly of points he had made in earlier papers, to produce a coherent whole. Interestingly, he spends a lot of words in trying to hose down some of the ideas that, according to him, people had incorrectly inferred from his earlier work. He also devotes some effort to debating a paper written by Meacham in the same volume of the journal (Meacham did not respond in that paper, so I presume Bellwood saw Meacham's paper before publication but not the other way around).
He states the linguistic argument for Taiwan as the Austronesian homeland, and emphasises that the linguistic argument is the strongest evidence and the most reliable available. Culturally and genetically, the Austronesians have merged to varying degrees with other cultures and developed locally in relative isolation in various places, so archeological findings about material culture and genetic analysis can't be relied on to produce an unambiguous signal of the Austronesian expansion - the best one can hope for is to find evidence consistent with the theory.
He makes the point that Proto-Austronesian seems to have some relationship to the Tai-Kadai languages of Thailand and Laos, which suggests an ultimate ancestry for the Austronesian languages somewhere on the South Chinese mainland.
He claims that rice and millet agriculture and the resultant high population density was the main advantage the Austronesians had to expand into ISEA and gradually replace indigenous hunter-gatherers.
He emphasizes that he is talking about a gradual expansion and assimilation, and not anything like a conquering army sweeping all before it.
A common interpretation of the AN expansion theory is that population pressure was a driver of migration - ie that once the population density in a particular place becomes high, some of the people will look to migrate elsewhere because it has become harder to make a good living in the original location. Bellwood clarifies that this is not his point - instead he argues for a strong founder culture, that for example the younger sons of a chieftain, who will not inherit their father's status, will look for ways to establish their own status elsewhere by leading a group of people to settle somewhere else. Population density itself enabled the success of the expansion, but was not the motivating force behind it.
His argument with Meacham mostly boils down to archeology verses linguistics - Meacham thinks archeology trumps linguistics, Bellwood the reverse. Let's move to a discussion of Meacham's paper itself.
William Meacham, Asian Perspectives Vol. 26 No. 1, 1984-1985, pp 89-106
Meacham is an American archeologist working in Hong Kong. He starts with the point that the model of Chinese prehistory had been recently revised. The earlier idea that rice cultivation and other aspects of Neolithic technology had arisen in Northern China and then spread by population migration to Southern China was no longer considered correct, and in its place a model where people in Southern China developed their own agriculture had become accepted. Meacham considered that Bellwood was making the same mistake that had been corrected in the case of China - that of assuming a central point for dispersal of culture, rather than crediting people with the ability to develop locally. In particular he thought it ironic that Southern China should be considered the ultimate point of origin for an expansion given that until recently it had been considered a technological backwater only able to develop after an influx of people from further north.
Meacham's argument about linguistics is that languages develop at variable rates, and are subject to the influence of neighbouring cultures that we generally know little about, so it is hard to conclude anything concrete. Bellwood's reply is that linguists are aware of the limitations, linguists know what they are doing, and that Meacham is under-estimating them. It's interesting that Bellwood's argument for linguistics over archeology is a fairly precise mirror of Meacham's for the opposite stance.
Robert Blust, Asian Perspectives Vol. 26 No. 1, 1984-1985, pp 45-67
Blust, who is sadly no longer with us, died 2022, was a linguist at the University of Hawaii. He stated the linguistic case for a Taiwanese homeland of the Austronesians. After a quick survey of early efforts to understand where Austronesian languages may have come from, going back more than a hundred years, he moved to a discussion of reconstructable vocabularies of different Austronesian languages and what could be deduced from them. If a language has a word for a concept or thing, you can safely assume the speakers of that language were familiar with that concept. If two Austronesian languages have a very similar word for a concept, you can assume the two languages can trace a descent from some other language that also had a similar word for that concept. The other possibility of course is that one language borrowed the word from the other - linguists tell us they can tell the difference between the two. By comparing common words with the known geographic extent of certain concepts (eg plants that thrive at particular latitudes but not others) you can make deductions about where the vocabulary in question must have arisen. His case is thus stronger than the point I made earlier, about the greater variety of Austronesian languages occurring in Taiwan then elsewhere - he also relies on vocabulary reconstruction compared with geographically specific concepts.
Interestingly, he also addresses, at such an early stage, an issue I mentioned above, about the vocabulary for nautical concepts in current Taiwanese Austronesian languages being apparently unrelated to the plentiful nautical terminology of Malayo-Polynesian-derived languages. His point was that Taiwan was settled by Chinese speakers beginning a few centuries ago, and the newcomers mostly settled near the coast. The Austronesian speakers retreated inland, and thus lost contact with the sea. With the sea no longer being a part of everyday life, the corresponding vocabulary could have been lost, or could have been replaced by that of Chinese-speaking sailors.
Mark Donohue and Tim Denham, Current Anthropology Volume 51, Number 2, April 2010
This paper was a full-frontal assault on the Austronesian Expansion - the authors were open about that being the objective. Mark Donohue is a linguist at the Australian National University, Tim Denham an archeologist at Monash University. I imagine ANU departmental seminars must have been lively if Peter Bellwood was also in attendance. By combining the efforts of an archeologist and a linguist they hoped to avoid the "stay in your lane" response that practitioners of one of those disciplines sometimes meet when wandering into the other.
The editors of Current Anthropology, to their credit I think, invited and published comment from a substantial number of experts in the field, including of course Bellwood. The whole thing is fascinating to read.
The paper starts with a statement of the particular AN expansion theory they take issue with - the movement of peoples with their language and culture, enabled by agriculture. The picture is a confluence of language, culture and genes, all moving across the landscape together. Bellwood had already pre-rebutted this a little by stating that linguistics is the main thing, and culture and genetics can only provide supporting evidence, and may not necessarily correlate perfectly.
The paper attacks all three of these, one by one. On genetics, their main point is that there is no clear signal in ISEA genetics of a Taiwanese origin for any substantial contribution to the ancestry of the modern population. There was sign of ancestry related to that of Taiwanese Austronesian people, but that could as easily be explained as having a common ancestry with people in southern coastal China, there was nothing to tie it specifically to Taiwan. In 2010, with the resolution of genetic analysis then available, they could say no clear signal, or "no genetic distinctiveness". As we will see later, more recent work has tightened this up to a clear signal for origins from elsewhere, not Taiwan.
On linguistics, they make a few different points. If our model is people migrating into the Philippines, and from there moving further afield into other parts of ISEA, you would expect to be able to construct a hierarchy of languages splitting from each other, and you would expect to be able to trace that back to the Philippines. That's not what is observed - the whole of ISEA seems to have a very flat hierarchy of Austronesian languages. The lack of any sign of directionality implies the picture must be more complex than a simple wave of expansion south from Taiwan.
The closest thing they identify to a directionality is that the closer languages are to Papua, the more original Austronesian vocabulary they seem to have lost. The implication is that Papuan languages had an influence on Austronesian-speakers - or alternatively, an Austronesian language was adopted by people already in place who spoke some other language and that language remained as a "substratum" (a term linguists like to use when they identify some features of a language as coming from some other language that may have been spoken by the ancestors of the speakers of the first language).
They also point out the lack of Taiwanese vocabulary related to maritime and agricultural matters mentioned above. On seafaring, Blust had already offered an answer to this in advance with his suggestion that the arrival of Chinese-speakers in Taiwan in coastal areas may have led to loss of maritime terminology from the Taiwanese Austronesian languages, but as far as I can see the hole in agricultural vocabulary was unaddressed.
There is a great deal of diversity in grammar among Austronesian languages, much more so than in vocabulary. For example, conventional word order varies widely. Tagalog is verb-subject-object, Indonesian is subject-verb-object, and verb-object-subject, verb-subject-object and object-subject-verb also occur. For an outsider learning Indonesian, one distinctive thing you notice is the fondness Indonesians have for passive voice, ie thing1 was verbed by thing2, which seems to be a less-than-wholehearted commitment to SVO and might indicate some remaining preference for a different word order. The authors suggest this variety is due to contact with pre-Austronesian languages. The relevance of this is that it makes the Austronesian languages much less of a monolithic block, and it downplays the extent to which the Austronesian languages dominated ISEA. Most of the languages spoken there today are indeed Austronesian, but languages spoken earlier are still there in some form, and it is clear the language transmission wasn't all in one direction.
On archeology they begin with pointing out that precise dating of archeological finds is difficult and that errors can and do occur, then move on to the lack of evidence of Taiwanese-style agriculture in the correct time period. They make the point mentioned above, that the little evidence we have for the presence of rice is all associated with pottery, not cultivation, and could indicate that rice was a traded commodity. Local people in ISEA might not have been growing their own rice at the time period that Austronesian Expansion would require. Firm evidence of rice cultivation comes much later and could have been imported direct from mainland Asia - there is no evidence the technology came from Taiwan, even though rice cultivation on Taiwan certainly was established before the Austronesian Expansion time period.
There is also no sign of the sort of environmental disruption you would expect to see with the introduction of agriculture. There are such signs in the archeological record in Papua, as well as Borneo and Sumatra, about 7kya, but not in a time period that coincides with the presumed spread of Austronesian languages.
The other aspect of culture you would expect to find in the archeological record is pottery, and here Taiwan does seem to be the origin for the "red-slipped" pottery style that did become widespread throughout ISEA in the correct time period. There was also a paddle-impressed pottery style that appears to have spread from mainland Asia at an earlier date, however this aspect at least of the Austronesian Expansion idea still seems to be on solid ground and the authors don't dispute it.
Other technologies that have been claimed as Austronesian, such as certain artefacts made from shells, have been found to have predated the Austronesian Expansion, so don't support the theory. There is also evidence that ISEA pig rearing was introduced from mainland Southeast Asia, not brought from Taiwan, and similarly for chickens. Of the so-called Austronesian cultural package, only language and red-slipped pottery remains.
The authors go into some detail about crops common in ISEA that were originally domesticated in Papua. Their intention seems to be to emphasize that agriculture wasn't something introduced from Taiwan, but was already long-established before the Austronesian time period.
The authors conclude with their own picture of what they think happened. Language, genetics and culture need to be de-coupled, the model where they move together isn't sustainable. Instead they see ISEA as having a maritime trade culture established long before the Austronesian Expansion time period where peoples in different places were in contact with each other and technologies and languages could be transmitted in all directions.
The responses published after the paper were many and varied. Bellwood was generally unimpressed - the most positive thing he could bring himself to say was that the paper was "nuanced", and he might have been sarcastic. He accused them of attacking a straw man version of the Austronesian Expansion theory, in particular the emphasis on genetics and the idea of population pressure, which he had never advocated. He thought their genetic arguments were fragile, based on technically unreliable methods, which you could argue was fair criticism at the time, but more recent work has put things on a firmer footing and has made the argument stronger. The linguistic arguments were not novel - the Malayo-Polynesian language tree was what he called a "rake", ie one trunk with a multitude of spikes, not multiply-branching lines, due to the rapid spread of Austronesian culture. He saw no evidence of early maritime trade networks. At least in the northern Philippines he claimed there was clear evidence pigs were introduced from Taiwan and not elsewhere in Asia. He saw the lack of evidence of rice cultivation as due to lack of the necessary archeological techniques having been applied to find it, and absence of evidence is not evidence of absence. He concluded by standing by his argument that population movement was the main reason for Austronesian language dominance in ISEA.
Murray Cox and Stephen Lansing, of Massey University NZ and University of Arizona, thought the authors' proposed model was too vague to be useful, and didn't offer enough testable predictions. They also pointed out an example, Sumba, in Nusa Tenggara Timur, where genetics and language do have good correspondence, and suggested the decoupling of the two proposed by the authors was unnecessary.
Waruno Mahdi, at the Fritz Haber Institute, Berlin, gave a survey of linguistic evidence, got into some details of rice cultivation and which varieties of rice thrive at which latitudes, and then concluded that the authors were correct to seek to modify the farming/language hypothesis.
Stephen Oppenheimer, of Oxford, was generally supportive but thought they should have cited more of his work, which was consistent with some of the arguments in this paper.
Victor Paz of the University of the Philippines was interested in the implication of the paper that there was already a significant inter-island culture in place in ISEA before the time period of the Austronesian expansion, and he thought this was probably correct. He pointed out that the spread of Austronesian culture is often identified as the arrival of the Neolithic in ISEA. He considered that the definition of Neolithic and its identification with Austronesian culture needed a rethink and this paper was a contribution in that direction.
Matthew Spriggs of Australian National University thinks the lack of observed directionality of the spread of Austronesian languages can be explained by frequent trade contacts and doesn't require us to throw out the expansion from Taiwan theory. He makes the absence of evidence is not evidence of absence point about archeological evidence for rice cultivation, and then adds there is a lack of evidence for a pre-Austronesian trade network that the authors suggest could have existed. He also points out that just because crops spread west from Papua, one can't infer the existence of a large volume of trade - occasional voyages over a long period would have been sufficient to spread seedlings and agricultural knowledge.
John Edward Terrell of the Field Museum, Chicago was generally pleased that the authors agreed with what he had been saying decades ago, that connections within ISEA and the Pacific must have been long-standing and a single migration event is insufficient. It was clear that his main focus of attention was on the Pacific, and the Polynesians, rather than ISEA.
Robin Torrence of the Australian Museum, Sydney started with the interesting point that the original reason people got interested in the Polynesian expansion into the Pacific and hence traced that back to Austronesians was colonial racism. Europeans were impressed by the Polynesians as they had lighter skin and fairly well structured societies, whereas the Melanesians had just a bit too much melanin and their societies tended to be more consensual, which to Europeans meant chaotic. He quoted James Cook's comments on the subject as one example. It was therefore important to find an origin for the Polynesians that didn't involve Melanesia. I don't think he meant to imply that modern researchers are necessarily following a racist agenda, but he thought it important to be aware of the origins of this line of research. He also pointed out that the apparently stark divide between Papua and the rest of ISEA was probably an imposed colonial distinction, and that contact and trade between the two was probably quite extensive before the Dutch, Portuguese and British started creating borders and controlling trade. On the spread of material culture, he pointed out that some people actively seek out innovations to take home with them when they travel, and that material culture can spread without the necessity for population movement - in this he approved of the paper in seeking to separate population movement from material culture. He found the idea of an inter-island cultural exchange network that built up slowly long before Austronesian times to be persuasive.
The authors were given the opportunity to reply to all the replies. They addressed the responses that pointed out prior art by readily conceding they were not the first to explore many of the ideas mentioned in the paper - though they did claim their approach was different to previous work. They re-iterated that they had no dispute with the idea of Taiwanese origin of the Austronesian languages, though they did dispute that we can be certain that the same is true for genetics and material culture. They were not impressed by the replies that point out linkages between archeology and language, and found plenty to quibble with about the Sumba example brought up by Cox and Lansing. They addressed Bellwood by just doubling down on their claim of lack of rice cultivation at the right time in the archeological record. They cited a paper by Paz that found plentiful evidence of rice cultivation at the right time in mainland Asia, despite the lack thereof in ISEA, so there is no intrinsic reason why archeologists could not find such evidence if it existed.
I think it's safe to say this paper was enormously important. It didn't kill the Austronesian Expansion idea, but it weakened it.
Ling Qin and Dorian Q. Fuller, {C. Wu and B. V. Rolett (eds.)}, Prehistoric Maritime Cultures and Seafaring in East Asia, The Archaeology of Asia-Pacific Navigation 1, ch 9 2019
This paper challenged the idea that rice cultivation itself could be a driver of population movement. Using many examples from China, the authors found that increasing population density leads to increasing cultivation intensity, rather than migration. This opposes some versions of the Austronesian Expansion idea, that population pressure led to migration from Taiwan, although Bellwood had disassociated himself from that idea and claimed that founder status, rather than population pressure, was the important driver of migration.
Larena et al (large number of co-authors), PNAS 2021 Vol 118 No 13 e2026132118
This paper seems to knock the genetic basis for the Austronesian Expansion on the head. They obtained a large genetic sample from all over the Philippines, and compared it with ancient Chinese DNA samples. Their genetic study showed that what had previously been identified as the Austronesian contribution to Philippine genetics was actually a population that branched from the Taiwanese Austronesian population 8kya, thousands of years before the Austronesian expansion and before the ancestors of Taiwanese Austronesians left southern China. They see no evidence of any substantial migration of population south from Taiwan in the 4-5kya time period.
Their analysis showed evidence of five different "waves" of migration to the Philippines. The initial settlement before 50kya shows up in the "negrito" populations of the Philippines. Sometime after the Papuan-Australian split, there was some movement of population from Papua north to the Philippines. There was an East Asian migration about 12kya, which they call Manobo because the genetic signal is clearest among the modern Manobo ethnic group in Mindanao. The Manobo split from the "basal" East Asian population about 15kya. A migration from mainland southeast Asia which they call Mlabri as they appear to be related to the Austro-Asiatic ethnic group of that name in Thailand and Laos occurred between 8kya and 12kya. The Mlabri ancestry split from that of the Manobo about 12kya. Then there was what they call a Cordilleran migration about 7-8kya. The Cordillerans are an ethnic group in the northern Philippines who have a low level of admixture with outside groups due to their remote location in mountainous regions and their resistance to colonisation and outside religions. This makes them good subjects for study if one hopes to understand ancient populations. Modern Cordilleran people show the ancestry from this last wave very clearly.
The Sama-Bajau (or Bajo) people who live in the southwest Philippines, as well as Malaysian Borneo and Eastern Indonesia, mostly trace their descent from the Mlabri migration.
The paper of course focussed on the Philippines, so one has to address how this affects our understanding of the population of other parts of ISEA. We could reasonably assume that the waves of population they identify for the Philippines also sent people to other parts of ISEA. All but the last wave they identify was, according to the authors, via mainland southeast Asia, so would of necessity have involved at least the Malay Peninsula and Borneo - only the Cordilleran wave was direct from coastal Southern China. Certainly their results show that a substantial Austronesian migration south from Taiwan to the Philippines 4-5kya did not happen.
It's worth repeating the caveat that I'm no expert, I'm just poring through the academic literature and trying to make sense of it. The latest paper I've found was published in 2021. Quite likely there have been new findings since then that I haven't read about, and it is definitely my plan to keep digging and reading on this subject. With all that stated, let's try to summarise what appears to be the current state of play as regards Austronesians and ISEA.
A substantial migration of people south from Taiwan to the Philippines about 4-5kya appears to just not be the case. In addition to the original migration more than 50kya, there have been ongoing migrations from further north in mainland Asia for at least the last 12 thousand years, and quite likely longer than that. Of course this isn't to say that no migration south from Taiwan occurred - it's just that any such migration did not provide the bulk of the ancestry of the modern inhabitants of ISEA.
The Austronesian "cultural package" doesn't appear to have been important - the two aspects of Austronesian culture that appear to have been propagated throughout ISEA are red-slipped pottery and language. The spread of the pottery style needs explanation - the point raised by Robin Torrence in the response to Donahue and Denham, that innovations can be spread by individual travellers without the need for large population movements, could be one way to explain it.
Agriculture doesn't seem to have arrived in ISEA with Austronesian migrants from Taiwan - instead it appears to have arrived via a number of routes, both from Papua and mainland Asia.
The spread of Austronesian languages is the surviving aspect of the Austronesian Expansion hypothesis that remains undeniable. It was the initial piece of evidence that led to this whole endeavour, and still survives as something remarkable that warrants explanation. The suggestion of Donahue and Denham, that language could spread via a trading network that would have included Taiwan, comparably to how Malay spread as a language of trade throughout ISEA and became the national languages of Indonesia, Malaysia, Singapore and Brunei, seems plausible. However it also seems fair that some of their critics point out that a plausible story by itself isn't enough, we need actual evidence to be confident of it, and we don't seem to have that evidence.
This has been yet another example of my lifelong experience where you learn a bit about a subject, to the point of thinking you might actually understand it, and then you learn a bit more and realise what you thought you know is probably wrong. No doubt the cycle will repeat once I learn a bit more.