The greatest impetus to Polynesian archaeology, however, occurred in 1920 when geologist Herbert E. Gregory acceded to the directorship of the Bishop Museum in Honolulu, convened the first international Pan-Pacific Science Conference, and proclaimed that the study of Polynesian archaeology and anthropology should be a major research priority (Kirch 2000, 20–24). Gregory, who continued to hold a professorship at Yale University, had important connections with the East Coast establishment in the United States and was able to secure major funding for a series of research expeditions to several Polynesian archipelagoes. The Bayard Dominick Expeditions of the Bishop Museum, from 1920 to 1922, were designed to implement the emerging Americanist vision of a holistic anthropology, combining multiple lines of evidence from ethnography, archaeology, ethnobotany, and physical anthropology (somatology). Research teams combining these disciplines were dispatched to Tonga, the Austral Islands, the Marquesas, and Hawai’i to carry out parallel investigations designed to address the overarching problem of Polynesian origins.

In retrospect, however, archaeology played a subordinate role in the Bayard Dominick Expeditions, leaving the field to be dominated by the comparative ethnologists. In the Tongan expedition, for example, archaeologist William C. McKern (later to become famous for his work on North American ceramic taxonomy) focused most of his efforts on the mapping of large stone monument sites, making only limited excavations in a few caves and kitchen middens (McKern 1929). McKern did recover an elaborately decorated form of pottery in these excavations, but lacking any method for direct dating, he interpreted this as a late-prehistoric variant of Fijian trade ceramics. Only decades later would his shards be properly recognized as part of the Lapita cultural complex, dating to the early part of the first millennium b.c. and associated with the first human settlement of Polynesia.

In the Marquesas Islands Ralph Linton (best known for his later ethnographic work in Africa) directed the archaeology but completely failed to recognize the potential for stratigraphic excavations. Without even bothering to test excavate, Linton simply concluded that “no opportunity was afforded for the gradual accumulation of stratified deposits” and that “no kitchen midden or shell heaps exist in the islands” (1925, 3). Thus he focused entirely on the mapping of late-prehistoric and early-postcontact monumental structures, interpreting these strictly within the context of a static, ethnographic reconstruction of “traditional” Marquesan culture. Even the possibility of cultural change or time depth (long periods of time) was thereby eliminated.

Consequently, despite a renewed emphasis on modern, scientific methods of archaeological survey (and sometimes excavation), the interpretation of Polynesian prehistory from the 1920s until World War II was largely dominated by comparative ethnology. The failure of archaeology to take hold partially resulted from the absence of any evident method for direct (or even relative) dating. Radiocarbon dating was still a thing of the future, and the methods of seriation being developed in North America and elsewhere for generating relative chronologies were seen as not applicable in Polynesia, given the general absence of pottery. Thus the stone structures and stone tools mapped and recorded by archaeologists in Polynesia were fitted into a largely static, ethnographic reconstruction and subsumed under the rubric of “material culture.” To be sure, a great deal of fundamental survey work was carried out during these years, such as that done by Kenneth P. Emory in the Society Islands (Emory 1933) and elsewhere or that of Wendell C. Bennett (1931) on Kaua’i Island (Bennett would later make his mark in Andean archaeology). This was archaeology, but it was not prehistory.

Migration Theories of the Early Twentieth Century

With the failure of archaeology to provide a real temporal framework for culture change and culture-history in Polynesia, the interpretive field was left to the comparative ethnologists, who adduced the new archaeological survey data only rarely. Dominant among this group of scholars was Edward S. Craighill Handy, who had led the Marquesas party of the Bayard Dominick