Volume 111 2002 > Volume 111, No. 2 > Special review, p 171-176
                                                                                             Previous | Next   

- 171
KIRCH, Patrick Vinton and Roger C. Green: Hawaiki, Ancestral Polynesia: An Essay in Historical Anthropology Cambridge: Cambridge University Press, 2001. xvii + 375 pp. bib., figs, glossary, index, maps, photos, tables. Price: £47.50 (cloth), £17.95 (paper).
SIMON BEST University of Auckland

Ten years of work have gone into this book, which explores the relationship between Archaeology, Ethnography and Linguistics in the deep past of Polynesia. The authors have struck out into new territory, and, like all explorers, have returned with tales of wonders that those who have stayed behind may wonder at, or even find hard to believe. They appear to have been well fed in their travels, with the first paragraph of the preface reading like Mills and Boon meets Cordon Bleu. Once past this purple prose, the book is stuffed with data from all three disciplines, with some convincing (and some not-so-convincing) attempts to get the three legs to fall into step. This review looks at a sample of the archaeological input to the “triangulation method” that Kirch and Green have employed, and assesses whether their conclusions are realistic.

The first, and major, criticism concerns the place that Fiji has been afforded in the work, which is no place at all. These islands have been given short shrift in the early Polynesian equation on two grounds—projection back from contact criteria and a strange interpretation of the early archaeological data. Using characteristics such as domestic architecture, canoe types, exchange systems, and status and kinship organisation (pp.65-70), the authors use phrases such as “striking differences ethnographically” and “valid phylogenic unit[s]” to emphasise how Fiji differs from Polynesia and to justify that their “concern in this book is only with Polynesia” (p.65). While these regional differences are certainly true of the immediately pre-contact era, the question of when they started to arise is still unanswered.

At the other end of the time scale, the boundary between Fiji and Polynesia is said to be archaeologically visible from around 500 B.C. (pp.75, 79, 91), with the Ancestral Polynesian homeland at that time restricted to Tonga, Samoa, Futuna and Uvea, and specifically excluding Fiji (p.89). The unfortunately-named Polynesian Plainware ceramics are regarded as markers of an Ancestral Polynesian stage (although they are present during the same period in Fiji for several hundred years), and so is a distinctive Polynesian adze kit (p.78).

There is, however, no archaeological evidence for the kind of cultural divergence between Fiji and places east about which Kirch and Green write until at least 1000 years after the settlement of both areas. The break at that time is not so much the genesis of a Polynesian identity in Tonga and Samoa as some radical rearrangement of the Fijian system—Polynesia became Polynesia because Fiji became Fiji. Any - 172 differences between the ceramics of Fiji and Western Polynesia are no greater than the regional differentiation noted by Kirch and Green within their Ancestral Polynesian homeland (p.170). The triangular section adze, the flagship of the Polynesian adze kit, has yet to be found earlier than c.2000 years ago, although Kirch and Green project its probable manufacture into the first millennium B.C. (p.178).

Archaeology, then, suggests that, until about the 1st century or so A.D., the entire area—Fiji, Tonga, Samoa, Uvea and Futuna—could be said to have been the Ancestral Polynesian homeland, irrespective of how different Fiji appears from the rest 2000 years later.

The timing of the move away from the Western Polynesia area, both to the northern atolls and into Central East Polynesia, is put at between 2200-1900 B.P. (p.79). There is, however, little archaeological evidence as yet for this, and in the case of Tokelau such dates would double the length of known occupation, that is if the atolls had even surfaced by that time. The inference is also made that the occasional potsherd found in the northern atolls (Tokelau and Tuvalu) were remnants from the original expansion at that time (p.168), but it has been shown that Fiji is the most likely source for both of these, and the Tokelau examples are at least 1000 years younger.

Some of the archaeological evidence for the reconstructed linguistic categories is questionable. Foremost among these is the identification of pits for fermenting breadfruit, where “[d]irect archaeological evidence backs up the linguistic hypothesis for fermentation and pit ensilage … [with] … appropriately sized and shaped pits [present] in many Ancestral Polynesian sites” (p.160). For Niuatoputapu, however, these pits are described as large, with straight sides and flat bases, and for Samoa as shallow and rounded (p.160). Given this range of variation, almost any feature that is not a post-hole or an oven could qualify as a fermentation pit.

The octopus lure caps, shell sinkers, shell scrapers etc. are all artefact types with fuzzy brackets, and it is probable that many of these are actually food waste.

A couple of quibbles: Early Eastern Lapita and Eastern Lapita appear to be used interchangeably (the inclusion of Niuatoputapu in the first is difficult to sustain), and the nut-cracking hammerstone (p.148) is surely an anvil? And does not Manu'a have a resident boa (p.109)?

With Hawaiki Kirch and Green have attempted an extraordinarily difficult task, with archaeology perhaps the weakest part of the triangulation method, by nature of its mainly equivocal data, which permits, if not demands, multiple interpretations. The value of the book, which is a significant contribution to the anthropological and archaeological literature, lies not so much in whether they have actually succeeded in some or even any of their attempts to marry the three disciplines, but in that they have opened up new horizons, and provided a synthesis of data that will stimulate research and debate for some time to come.

One suspects, however, that if this intrepid duo explore much further in this direction, they may disappear into the jungle out there—to be found later, perhaps, in some sun-dappled clearing, surrounded by the remains of their last meal of roast Petaluma duck, and still clasping a decent Cabernet.

- 173
PAUL GERAGHTY University of the South Pacific

The rationale for this book is simple. It is an attempt to reconstruct the way of life of the Ancestral Polynesians who lived in Western Polynesia over 2000 years ago. Archaeology, the authors tell us, can only do so much. Even in the domain of material culture, some 82 percent of Polynesian artefacts are made of perishable materials and so leave no trace. In other domains, such as kinship and religion, there is practically no tangible evidence. So appeal must he made to comparative Ethnography and Linguistics. It is the use of comparative linguistics in this book that will be the primary focus of my review.

The major linguistic flaw in this work is the uncritical use of POLLEX, the file of Proto Polynesian reconstructions assembled by the late Bruce Biggs and colleagues at the University of Auckland. One point that the authors fail to realise is that, monumental though POLLEX is, it is far from complete. It is ongoing and has hitherto been based largely on internal comparison; even if only existing dictionaries of extra-Polynesian languages were to be searched thoroughly, it would at least double the number of Proto Polynesian reconstructions.

More crucially, POLLEX is simply a collection of apparently related forms in Polynesian languages: the compilers do not claim that Ancestral Polynesians used the reconstructions. The list includes words, such as *kumala ‘sweet potato’, which are known to be intrusive in Polynesia. The main value of such lists in the reconstruction of prehistory is not the “proto-forms” themselves, many of which may be totally spurious such as *kumala, but the detection of borrowing that is made possible by the study of sound correspondences among basic vocabulary.

While the authors point out that linguistics has procedures for detecting borrowings (p.87), they do not mention that certain conditions need to obtain, including the presence of diagnostic phonemes, the absence of bilingualism, and a full understanding of the relevant historical phonology. As a result, they greatly understate the amount of borrowing in Polynesian prehistory. Evidence from Tupaia's chart, the list of islands known to the Tongans recorded by Cook, oral traditions, and Polynesian loanwords in many parts of Micronesia and Melanesia leaves no doubt as to the extent of inter-island voyaging that would have facilitated diffusion, both internally and externally. The number of Polynesian loanwords in Rotuman and Kiribati is comparable to the French component of the English lexicon—hardly the result of a few small bands of wandering aristocrats, as the authors would have us believe (p.86). Nor was the traffic one-way: of the supposedly Ancestral Polynesian items mentioned in this volume, many are clearly later arrivals: *fasu ‘uterine nephew’, *faasua ‘giant clam’, *ngasau ‘arrow’, *saka ‘boil food’ and *wī ‘Spondias dulcis’ from Fiji; *mei ‘breadfruit’ and *pulaka ‘swamp taro’ from Micronesia; and *pusa ‘box’ and possibly * kumete ‘wooden bowl’ from Dutch; while * tautahi ‘master fisherman’, * matau ‘axe’, *aveloa ‘breadfruit cultivar’ and * maasoaqa ‘arrowroot’ represent recent Samoan and Tongan coinages. No doubt many more spurious PPN reconstructions will be discovered, including perhaps a few more plant names from South America.

- 174

Seen in this light, it is difficult to decide whether the misrepresentation of my 1993 Pulotu paper in JPS (102:343-84) is deliberate or just sloppiness. I am said (p.242) to have reconstructed *pulotu as Proto Central Pacific for the ancestral homeland of the Polynesians (for which I am chided), and to have claimed that the Proto Polynesians originated from Matuku in Eastern Fiji (p.96). What I actually proposed was that a group left Togaviti in Matuku in approximately 1000 B.P., long after the break-up of Proto Central Pacific, and settled many parts of Polynesia, introducing the concept of Pulotu and the names Togafiti and Fanuakula. Either the authors simply never read my paper, or they chose to misrepresent its findings as being too diffusionist for their liking.

When the authors leave the security of POLLEX to embark upon their own linguistic reconstructions, their work is often too slipshod to be trusted. In an attempt to find external cognates, they cite Wilkes (p.311) as having recorded Fijian kelekele ‘covered in earth’, when what Wilkes actually recorded was vulai kelekele ‘digging yams’; the mistake in the final vowel is typical for English-speakers—Horatio Hale, the linguist who was with Wilkes, recorded it correctly as vulai kelikeli. They also cite a number of Tongan forms with missing glottal stops and others with ‘r’, a sound not found in Tongan, and claim cognacy for Tongan moui and Samoan muli, PPN *mana and Anuta mank, and PPN *fatu and Samoan lātū (given incorrectly as latu). Other instances of sloppiness unexpected in authors of their standing are the recording of the scientific name of the palolo seaworm (Eunice viridis) as Nereis on p.100, Neiris on pp.271 and 372, and Neris in the index, the lack of cross-referencing between Nereis and palolo in the index, and the listing of *ango as *tingo in the Index of Reconstructions. While they mention linguistic procedures for detecting loans, they sometimes choose not to apply them, stating, for instance, that Rotuman kainaga is related genetically to PPN *kainaga when it is clearly a loan, as are the alleged Micronesian cognates of this term and also Rotuman fono and tano'a. They also fail to understand that semantic reconstruction is based on agreement among highest-order subgroups, not on a show of hands (pp.146, 291).

The book is further marred by bouts of conceit, such as the claim that the authors are pioneers of cross-disciplinary prehistoric reconstruction that they call “triangulation” (a term I used with considerably less fanfare in my 1993 Pulotu article). They go on (p.249) to attribute to this “innovative” method their “discovery” that PPN *qariki meant not only ‘chief’ but also ‘priest’—an observation that was made by Horatio Hale in 1846. And just because you have reconstructed a word meaning ‘good-tasting’ and another meaning ‘bad-tasting’, I think it is a little premature to talk of having discovered “an Ancestral Polynesian theory of taste” (p.146).

Since Hawaiki is dedicated to the late Professor Bruce Biggs, and since the linguistic aspects are based largely on POLLEX, it would be appropriate to cite a particularly apposite piece of Biggsian wisdom from 1972, alluding to the use of linguistic data by people who do not understand it: “If we want to use the other fellow's last we had better know exactly what it is good for.”

Linguistics apart, I enjoyed reading this book. It is on the whole well written, - 175 methodically organised, professionally produced and edited (I noted only about 20 typos and miscellaneous mistakes), and packed full of references, albeit selectively chosen, and interesting comparative data. But I am far from convinced of the reality of the authors' reconstruction of the way of life of the Ancestral Polynesians; and, at least from the linguistic point of view, it has to be said that the authors have borrowed Bruce's last—and produced a load of cobblers.

- 176 Page is blank