Investigations on small Polynesian outliers have illustrated how difficult it can be to identify archaeological evidence of intrusion, or to interpret the effect of any intrusion on the resident populations. In Samoa, the still meagre amount of artefactual and faunal remains from archaeological excavations adds to these problems. A review of the known Samoan archaeological sequence finds little or no evidence of intrusion, apart from a probable post-settlement introduction of pigs and dogs. This need not mean that Samoa was ever isolated from contacts with other islands.
Keywords: Triple-I model, Polynesian outliers, Lapita, subsistence, material culture, monuments
In a recent paper, Addison and Matisoo-Smith (2010) proposed a "Triple-I Model" of intrusion, integration and innovation for the Samoan sequence. They suggested a possible arrival about 1500 BP of new people, who introduced new lineages of rats, dogs and chickens, new plants, new material culture, and new ideas, and tentatively proposed a route from the west through the low islands of Micronesia. Their paper stimulated me to think about the difficulties archaeologists face in identifying and interpreting evidence of intrusion and replacement, a subject that has long concerned me (Davidson 1970, 1974a). The present paper briefly considers problems in interpreting archaeological evidence of intrusion (or lack thereof) in several Polynesian outliers, and then reviews current evidence about aspects of the archaeological sequence in Samoa.
Archaeological research on Polynesian outliers has been driven, not surprisingly, by the fact that these small islands in the geographical regions of Micronesia and Melanesia are today inhabited by people who speak Polynesian languages. Identifying the arrival of Polynesian speakers has been a major objective, which has, however, proved very difficult to achieve, as the following examples show.
Nukuoro is the northernmost of the known Polynesian outliers. It is a small atoll between New Britain and Pohnpei, with only a few hundred inhabitants. In the 1870s, the German ethnographer, Kubary, recorded traditionally remembered canoe arrivals from some 17 different islands stretching from Yap to Rotuma. Some of the arrivals stayed and intermarried, some were killed, some were banished, some introduced new dances; those banished had introduced a new kind of murder (Kubary 1910: 6-8).
Despite these historical accounts, the known archaeological sequence on the atoll, beginning about 1200 years ago (Davidson 1971, 1992) shows no evidence of intrusion or new arrivals apart from the late appearance of the serrated-edged pearl shell coconut grater. This archaeological sequence falls within the timeframe within which linguistic models would expect the present Polynesian language and its immediate ancestor to have been spoken on the atoll. Although Kirch (2000: 179-180) considered that "there is no reason why the Nukuoro sequence should not be regarded as 'Polynesian' from bottom to top", my own conclusion was that if Nukuoro had been uninhabited at the time of European contact it would never have been recognised as a Polynesian outlier.
Leach and Ward (1981) faced a similar situation on the nearby Polynesian outlier of Kapingamarangi. Although Emory (1965: 1-2) considered it exclusively Polynesian, earlier German ethnographers had found what they considered Melanesian and Micronesian as well as Polynesian influences (Eilers 1934: 155). Leach and Ward (1981: 93-97) had difficulty in suggesting a likely origin for the people. They found no evidence of intrusion in the archaeological sequence, although they pointed to ethnographic evidence of a type of food preservation and a method of roof thatching that suggested contacts, perhaps drift voyages, from Kiribati or the Marshalls.
The Polynesian outliers in the Santa Cruz group have much longer archaeological sequences than Nukuoro and Kapingamarangi, all beginning with ceramic occupations early in the first millennium BC. …