Late Colonization of East Polynesia
Spriggs, Matthew, Anderson, Atholl, Antiquity
The settlement of East Polynesia, a vast region (20 million sq. km), stretching between Hawaii, New Zealand and Easter Island and containing little and scattered land (0.29 million sq. km), raises intriguing questions about prehistoric development of maritime skills, cultural change and adaptation to island environments, but hypotheses about these depend importantly upon the chronology of colonization. An older model had initial settlement of the Marquesas about AD 300, then dispersal to Easter Island about AD 400, Hawaii about AD 750 and the Societies and New Zealand about AD 800 (Sinoto 1970; 1983a; Bellwood 1978a; Jennings 1979). A more recent model argues for initial colonization of the Marquesas during the 1st millennium BC, both Hawaii and Easter Island by AD 400 or earlier, and attributes scarcity of early evidence in central East Polynesia to sampling error (Kirch 1986). A derived argument has New Zealand colonized before AD 500 (Sutton 1987). The difference between the models cannot be attributed substantially to new data. Rather, old dates were reconsidered and differentially approved. In our view, however, there remains a significant lack of critical rigour in doing so.
A recent study of dates for early Hawaiian settlement (Hunt & Holsen 1991), for example, takes reported radiocarbon ages en masse to examine the pattern of dates, virtually without regard to problems such as suitability of dated material and stratigraphic inconsistency of some early dates. The 'old wood' problem is referred to, but only as a question for future study. Hunt & Holsen's conclusion is that the dates 'might be suggestive of a human presence as early as the first century A.D.' (1991: 158). The justification for the blanket acceptance of the Hawaiian radiocarbon corpus is threefold (Hunt & Holsen 1991: 157-8). First, that natural fires as a source for charcoal are rare in Hawaii and the products of vulcanism should be easy to distinguish, leaving a human origin for the dated charcoal almost certain even when the charcoal is in secondary deposition. Second, that the earliest sites have had the most time to be destroyed and are the hardest to locate, and also that the bulk of dated sites are in leeward environments likely to have been settled later than the more fertile windward regions. Third, that 'in spite of the real potential for erroneously early dates, we must be careful that current predilections about what is "too early" do not seriously bias our interpretations' (1991: 158).
The case for the rarity of natural fire in Hawaii, particularly in dry, leeward areas, has not been convincingly established. Fires caused by volcanic flows certainly have the potential to spread considerably beyond the areas affected by vulcanism. The claim that almost any carbonized particles dated in Hawaii will relate to human occupation is unverifiable and likely to lead to serious misinterpretation in what amounts to an 'anything goes' approach to dating (Hunt & Holsen 1991: 158):
The fact that in some cases carbonized particles are transported and deposited anew means that human activity is dated, but not the activity necessarily associated with the excavated context. In light of this consideration, even dates seemingly 'out of context' may be telling us something about the presence of people and their use of the surrounding landscape.
That the earliest sites have had the greatest time to be destroyed is not in doubt, although as the New Zealand case shows (Anderson 1991: 792) it does not necessarily mean they are likely to be invisible. The assertion that leeward zones were occupied later than the wetter windward areas is open to dispute, however: more benign climate, productive and sheltered fisheries, perhaps a greater variety of avian fauna occurring in more open situations where hunting was easier and ease of clearing dry forests for agriculture might well have made particular leeward areas most attractive as early settlement sites. …