Chapter 5. Austronesian Prehistory in Southeast Asia: Homeland, Expansion and Transformation

This paper will commence by focusing on the question of where the immediate ancestor of Proto-Austronesian was located, and when. Proto-Austronesian is the hypothesized linguistic entity, perhaps a single language or perhaps a dialect network (see Pawley and Ross, this volume), ancestral to all subsequent and existing Austronesian languages. But like all languages it also had an ancestor, prior to the budding of the Austronesian (henceforth An) family as a linguistic taxon with its own unique history.

An observation relevant for this question, one particularly intriguing in terms of its relevance for world prehistory, is that the general homeland regions of many of the major language families which have had long histories of association with agriculture seem to be geographically correlative with regions of primary (i.e. indigenously-generated) agricultural origins (Bellwood 1990b, 1991, in press b). In the Old World such language families include Indo-European, Elamo-Dravidian, Afro-Asiatic, Niger-Kordofanian, Nilo-Saharan, Sino-Tibetan, Austroasiatic, Thai-Kadai (or Daic of Ruhlen 1987) and Austronesian. The first two of these families, and possibly Afro-Asiatic, have arguable homelands in or closely adjacent to southwest Asia, the second two in northern sub-Saharan Africa, and the last three (with Sino-Tibetan being uncertain) in central and southern China. These three geographical regions are known from archaeological data to have witnessed major local developments of plant and animal domestication, in each case well before any such developments in intervening regions of the Old World (MacNeish 1992). Because of these widespread correlations between early centres of agriculture and major language family homelands (Renfrew 1992), one may posit a process whereby demographically-expanding agricultural populations moved outwards from primary agricultural homeland regions, perhaps slowly but certainly inexorably occupying lands previously occupied by foragers (as suggested for Europe by Ammerman and Cavalli-Sforza 1984; Renfrew 1987).

If one examines the geographical distributions of these major language families and the geography of diversity between and within them, one sees that areas of agricultural origin reveal both a larger-than-average number of different language families and high levels of internal language family diversity, as revealed by the close proximity of subgroups which have long histories of separation. This is true for the three areas listed, and also for the early agricultural homeland regions of central Mexico, the northern Andes of Peru and Ecuador, and New Guinea. Directly relevant for this paper, the Austronesian, Thai-Kadai, Hmong-Mien and Austroasiatic language families seem to have arisen by a process of dispersal out of subtropical southern China and northern Mainland Southeast Asia, a zone lying between the Yangzi and northern Thailand/Indochina,^[1] where the cultivation of rice and other crops developed widely between about 6000 and 3000 BC. Some of the Papuan language families of New Guinea are also associated with an early and primary centre of agriculture, although in this case the result seems to have been population maintenance and increase in situ rather than actual dispersal into new territories.

The Neolithic “revolutions” of China evidently occurred in two culturally-connected regions. The first, in the basin of the Yellow River, led to the domestication of foxtail and broomcorn millet by 6000 BC. The second, in the middle and lower Yangzi basin, led to the domestication of rice by about the same time (Yan 1991). However, although Chinese archaeologists tend to regard the Yellow and Yangzi basins as supporting unrelated Neolithic cultures, we perhaps need now to regard central and eastern China as one single centre for the early development of Asian monsoon agriculture.

Both rice and foxtail millet have been found in Chinese early Neolithic sites in storage pits and habitation layers, in quantities sufficiently large to suggest that they rapidly attained a major dietary importance.^[2] There is little doubt that they would have fuelled an increase in population numbers which was perhaps quite rapid, given the archaeological appearance of the oldest Neolithic cultures across huge areas of China by about 5000 BC (Chang 1986). One result of this would have been an outward expansion of those populations involved into areas which hitherto had been inhabited entirely by foragers.

By 5000 BC, settlements of rice cultivators were in existence down the eastern coastline of China to as far south as Guangdong, and by perhaps 4000 BC in northern Vietnam and Thailand. In their archaeological remains are found assemblages of artefacts which leave no doubt about the overwhelming impact of the new lifestyle. For instance, the 7000-year-old village of pile dwellings at Hemudu, near the southern shore of Hangzhou Bay in Zhejiang Province, has yielded pottery, stone adzes, wooden and bone agricultural tools, evidence for carpentry and boatbuilding, paddles, spindle whorls for weaving, matting, rope, and large quantities of harvested rice. In addition, the site produced the bones of domesticated pigs, dogs, chickens and possibly domesticated cattle and water buffalo. Such a large village settlement and such a range of tools suggest a fundamental and drastic shift from the presumed mobile foraging lifestyle of the East Asian Late Palaeolithic. The inhabitants of Hemudu and other contemporary Chinese agricultural settlements lived during an episode of cultural evolution which was ultimately to have repercussions over the whole of temperate and tropical eastern Asia and the Pacific. One of these repercussions, albeit one which developed its first momentum a thousand kilometres or more south of Hemudu, was the ultimately-phenomenal expansion of the speakers of Austronesian languages.