INTERDISCIPLINARY RESEARCH ON POLYNESIAN ORIGINS

Abstract

Introduction : The Modern Problem of Polynesian Origins Recent archaeological work by Beyer in the Philippines, Suggs and Golson in Polynesia and others in intermediate regions, has raised in a new form the old question of the origins of the Polynesians. Linguistics has long had an obvious interest in the question. Linguists have sought to claim autonomy for their own discipline. In the Pacific field the result has been chaotic : quot homines, tot sententice remains the verdict. In view of the work being done by archaeologists, showing better approach to unanimity, it may now be time for linguistics to enter the field afresh by approaching the problem in a somewhat new way, as participant in an interdisciplinary study. There are many facets to this approach. The first is a new analysis of the reconstructed Austronesian vocabulary provided by Otto Dempwolff,2 and this paper begins from those results. In what follows, however, no attempt has been made to assess the phonetic correctness of Dempwolff's reconstructions. That is part of the research needed. Dyen3 has laid the foundations of this revision. Nor have other words been added to Dempwolff's lists of reconstructions, though it is certain that such additions can validly be made : the author did this himself in his Linguistic Position of South-eastern Papua (Sydney, 1943). In Dempwolff's vocabulary, a total of 724 words out of some 2,000 reach eastern Oceania, representing approximately 36% of the vocabulary. Of these words 308 are recorded by Dempwolff as not occurring east of Fiji. These may therefore be regarded as in a special sense Melanesian (MN). The remaining 416 are divided into two sub-groups. The first 141 (33-9%) are recorded as not appearing in Melanesia. They may be called specifically Polynesian (PN). The second subgroup contains a larger number of words 275 (66 -i%) which are found by Dempwolff in Sa'a (South-eastern Solomon Islands) and/or (Bau) Fijian as well as in his PN. The total of the 416 are therefore PN. Of course, it must be realized that knowledge of all these languages is still imperfect, especially of the Melanesian languages. These figures might be altered