1 Polynesians Chapter 22 Polynesians: Devolved Taiwanese Rice Farmers or Wallacean Maritime Traders with Fishing, Foraging and Horticultural Skills within the past 8000 years, this does not have to imply they sprang suddenly into life de novo at that time. As an example, Dixon (1997, 46–8) has sug- gested that present evidence does not preclude Proto- Indo-European dating back 10,500 BP, or perhaps even 12,000 years ago. If this were the case, language and demic expansion congruity could also be fitted to the major post-glacial climatic amelioration fol- lowing the Younger Dryas event c. 11,500 BP (Dansgaard et al. 1993). This raises the possibility that demic expansion occurred as much as a result of climate improvement as of the fecundity of agricul- tural societies (e.g. Adams & Otte 1999). Since accu- mulating genetic evidence from both mitochondrial DNA (mtDNA) (Richards et al. 2000) and the Y chro- mosome (Semino et al. 2000) in Europe have sug- gested that post-glacial expansions and recolonizations had a much greater impact on gene pools than Neolithic immigration, this seems to be a possibility worth considering for Asia as well. Polynesian origins There are two main clusters of homeland hypotheses attempting to explain Austronesian and/or Polynesian origins and dispersals: mainland farming dispersal (Bellwood 1997; Blust 1984–85), or offshore, non-farm- ing dependent dispersal (Dyen 1965; 1971; Terrell & Welsch 1997; Terrell et al. 2001; see also Meacham 1994– 85; Solheim 1994; 1996 for arguments for an Austro- nesian homeland in Island Southeast Asia). Both mainland and offshore models have relied on combinations of linguistic palaeontology and ar- chaeology. This approach has been criticized, how- ever, on grounds of the relative mobility of cultural and linguistic markers when compared with the large Stephen Oppenheimer & Martin Richards Most linguists and archaeologists support a migra- tion model for the origins of Polynesians, but until recently there was little relevant genetic evidence with which to test this view. Archaeologists may reconstruct material culture and date it but, in the absence of written texts and skeletal remains, they have less success identifying the origins of the manu- facturers of the assemblages. Similarly, linguists can reconstruct the branching history of their languages, but have no direct evidence for the origins of the speakers of those languages. Here, we discuss the new genetic evidence on human dispersals into the Pacific. At the same time, we question the mutually dependent structure of evidence from archaeology and linguistics that is used to support a farming- fuelled range expansion all the way from South China out into the Pacific. Because languages change at a variable rate, both Indo-European and Austronesian farming dis- persal models derive their dates and cultural hori- zons primarily from archaeology. However, linguistic palaeontology remains extremely influential in the reconstruction of prehistoric dispersals (Mallory 1989; Bellwood 1997), even though this practice has been condemned by Renfrew (1987, 77–86). Even the ghost of glottochronology seems to linger. For example, comparative linguists often argue that, because of lexical decay, language families cannot be traced back more than about 6000–8000 years; in other words, not before the Neolithic transition (Trask 1996, 377). Bellwood (2001) has recently argued that this figure may not have arisen by chance, but may after all be telling us that the major language families were spread by farming dispersals. But although the great modern language families may have only become recognizable in the modern reconstructed record I can't find figure references in the text. Please could you insert them where appropriate. - ed.