The Austronesian expansion ultimately extended well over half way round the world, mainly in tropical latitudes, from Madagascar to Easter Island. In its western regions it intruded into a far more complex pre-existing social landscape than did those pioneer navigators who settled the empty islands of Remote Oceania beyond the Solomons (Green 1991). There is a large quantity of linguistic and archaeological information that relates to early Austronesian expansion, and readers are referred to fuller sources (Bellwood 1985, 1989, 1991, 1992; Blust 1988; Ross 1989; Spriggs 1989). I list some major relevant observations.

Dated archaeological assemblages that can be related to Austronesian proto-language inventories (Blust 1976; Pawley and Green 1984; Bellwood 1985:102-129) suggest that Austronesian expansion moved from Taiwan, through the coastal Philippines, into Sulawesi and towards coastal New Guinea between about 3000 and 2000 BC. This represents a hypothetical average rate of “as the crow flies” colonization of perhaps three km per annum, or 75 km per 25-year generation. The rate after about 1500 BC within Melanesia and western Polynesia was apparently much faster; the Lapita culture achieved a distribution from the Admiralties to Samoa within about 300 years (Kirch and Hunt 1988), representing an average of about 13 km per annum, or 325 km per 25-year generation. Obviously the real rate of coastal colonization was much slower than that across sea, but these average rates still give some idea of relative speeds in different areas. The spread into Polynesia might have been equally as rapid as that of Lapita according to the navigational calculations of Irwin (1989), although central Polynesian island groups such as the Cooks and Societies have not yet produced direct archaeological evidence for settlement contemporary with Lapita.

The overall archaeological and linguistic records also suggest that the process of Austronesian expansion can be divided into three roughly-defined geographical units in terms of the pre-existing social landscapes with which Austronesian colonists had to contend. These are a) from Taiwan through the Philippines and Indonesia — pre-existing low density foragers; b) western Melanesia, Vietnam and Peninsular Malaysia — pre-existing agriculturalists or intensive foragers with high population densities; and c) Remote Oceania and Madagascar — no attested pre-existing human populations at all. Since Melanesia is central to the issues discussed in the presence here of an apparently independent transition to agriculture, especially in the New Guinea Highlands (Golson 1981), but with presumed lowland repercussions, is worthy of emphasis. The rapidity of the Austronesian (Lapita) spread from the Bismarck Archipelago into Remote Oceania may thus be related to the presence of high density prior occupation in many parts of western Melanesia.

Early Austronesian expansion, like that of the Indo-European and Bantoid language speakers, was thus quite a rapid phenomenon when placed in perspective against overall human ethnolinguistic prehistory. In the Austronesian case the archaeological record for Island Southeast Asia and Oceanic Lapita suggests strongly that founders moved onwards to new regions after good coastal locations were occupied, before any major attempts to colonize island interiors. Population pressure alone is thus not an acceptable sole reason for early Austronesian expansion. More feasible ecological and techno-economic reasons which might have contributed to the expansion, especially in its earlier phases, include continuous population growth and fissioning due to the possession of an agricultural economic base; the inherent ability of an agricultural economy to be transported and to support colonizing founder populations; the possession of skilled canoe and navigational technology; a search for prestige goods;^[8] and predilections for special environments such as swamplands for rice and taro or good fishing lagoons.^[9] One purpose here, however, is to suggest that this list, while of undisputed validity, probably lacks an essential component in the sphere of ideology.