Idol talk and flights of fantasy

Flipping through the pages of this handsome hardback, one is struck, sometimes awed, by photographs and illustrations showing Gargantuan ruins on tiny islands lost in the ocean. A revelation. The easy, catchy style of the opening paragraphs of the book's chapters add to this first impression: "The reddish-gold light of early morning streamed laterally through motionless coconut palms... Escorted by perhaps 3,000 canoes, His Britannic Majesty's ships..."

Is this then "a synthesis for those whose curiosity would take them on an intellectual voyage into the Oceanic past", a book accessible to all, the work to replace Peter Bellwood's Man's Conquest of the Pacific (1978) - as claimed by its author Patrick Kirch, a professor of anthropology in Berkeley, California? Is it, as writes Roger Green, emeritus professor of prehistory at Auckland University in New Zealand, "the first book to which scholars and general readers will go to pursue any topic relating to Pacific archaeology"?

Regretfully, no. As early as the introduction, Kirch contrasts the "extremely wide distribution of the Oceanic lanteen [ sic ] sail" in the Pacific to that of the "Oceanic spritsail ". An illustration would have been appropriate here; there is none. Even readers conversant with sailing technology will be left wondering in what ways Oceanic lateen sails may differ from other lateen sails and whether there are really such things as lanteen sails - the terms are unexplained and will never be mentioned again. On the same page, readers unfamiliar with modern genetics are on their own to puzzle over "9-base-pair deletion... in mitochondrial DNA" and just how and how precisely it reflects "a broadly geographical and presumably historical distinction between [non-Austronesian] and [Austronesian]-speaking people". Since the author offers no explanation in this chapter, nor indeed in the rest of the book, the truly curious are reduced to hunting for the lone reference, a 1999 article by D. Merriwether et al in the American Journal of Anthropology .

This is not an isolated occurrence. Kirch typically proceeds by assertions, sometimes backed with a reference or two, more often entirely unsupported. When he does provide justifications or explanations, they are insufficient, sometimes even misleading. For example, the section entitled "Lapita and the Austronesian expansion" starts in the easy tabloid style already quoted, with Kirch and a native helper uncovering a small bone statue on an island of Papua New Guinea. "Simultaneously, we caught our breath... Neither of us spoke, then Bauwa muttered in Melanesian pidgin, 'God belong ol Lapita!'" To the general reader, the helper has spontaneously recognised the figurine as representing a "god of the Lapita people" (god belong ol Lapita). In reality, this is a case of ethnological contamination. For Kirch's helper knows of the "Lapita people" from archaeologists, and that they are excavating the area for evidence of Lapita occupation. This exclamation is therefore misleading to the general reader, and irrelevant to the specialist. Things soon get worse, with flights of groundless speculation taking the place of factual reporting.

Lapita is the name of a site in New Caledonia, 3,000km away, where a distinctive type of pottery was first discovered 50 years ago. In time, similar potteries were discovered over the Pacific, and from this emerged the hypothesis of a prehistoric culture characterised by the use of such pottery, and so the terms Lapita culture and Lapita people were coined. Because early Lapita wares were richly decorated (the earlier, the richer), Kirch hypothesises that they "played a non-utilitarian role in Lapita society". Because some bore - given a great deal of imagination - motifs representing " human faces " (Kirch's emphasis), he further hypothesises that they were representations of ancestors, used in rituals. Finally, he equates the "act of decorating the pots (with) tattooing a human body". On what evidence? "We know that Lapita peoples practiced tattooing, for tattoo chisels have been found in some sites" (my emphasis). On the same line of argument, Mediterranean fishing boats, traditionally decorated with eyes at the prow, must be non-utilitarian craft used in ancestor worship, and the act of decorating those boats is analogous to cattle branding because cattle brands are found in some places around the Mediterranean.

Kirch seems to realise, however dimly, that his speculations are vacuous, and he defends his position thus: "Admittedly, these hypotheses cannot be 'proved' to the satisfaction of hard-line positivists, but they are certainly reasonable." Soft-line sceptics might beg to differ, especially in the next chapter where Kirch identifies a Lapita culture on Aneityum island in Vanuatu, when there is no trace of it whatsoever: "Although no Lapita occupation sites were discovered on Aneityum, a pollen core... exhibits a clear signal of human arrival on the island around 900BC, believed to reflect a Lapita colonisation." Why Lapita, and believed by whom? Readers wishing to dispute Kirch's assertions are cornered into battling windmills constructed in ways they can only guess at, and so Kirch's argument remains unassailable, for no evidence is presented.

Does this book constitute at least a reliable source of information, once stripped of its speculations? As Kirch draws from archaeology, anthropology, genetics, ethnology and linguistics, it would take an Areopagus of experts in each field to tell. But there are alarming signs.

About Easter Island, Kirch writes without any qualifications, as if these were facts established beyond the shadow of a doubt: "The statues represented deified ancestors, and the chiefly elite constructed their houses to be literally under the gaze of the ancestors... The statue cult came to an end and the island plunged into a state of intense inter-tribal raiding and warfare ...from about AD1500 to 1722... Power now fell to the warrior class... [The native hieroglyphic script] encapsulated ritual or sacred recitations, including procreation chants." Those assertions are mere fantasy, unsupported by archaeological, ethnological and linguistic evidence. The rongorongo script on wooden tablets has never been deciphered, although two lines of a tablet obviously constitute a lunar calendar, and a short sequence on another is very likely a genealogy - but there is no evidence of procreation chants or sacred recitations. Again, Kirch's other claims are difficult to refute, for the few facts presented are too imprecise to be unambiguously identified.

For instance, he describes the houses of the "chiefly elite" as having "foundations of cut and dressed basalt". A photograph of those foundations - and there are many available - would have dispelled any doubts that he refers here to the long-houses observed by early navigators and called hare paenga . But as the description stands, without any photograph, one is reduced to guessing. As for those houses being residences of the chiefly elite, there is no evidence of it, even in oral tradition. They were likely communal houses, each sheltering several families, nothing grander. But the commoners, according to Kirch, dwelt in "small round, oval, or rectangular houses with rough stone foundations". This description rather fits the stone buildings called tupa and hare moa , the true functions of which remain unknown - according to oral tradition the former were watch towers, the latter chicken coops (which is the meaning of hare moa ). Whatever they were - peasants' cottages, watch towers or chicken coops - we know that we are in for another flight of fantasy when Kirch goes on to describe them as "visually 'controlling' access to the limited maritime resources". As a whole, his pages on Easter Island are far closer to fantasy than fact.

Turning to the linguistic evidence, Kirch presents only the side that suits him, the views of those comparative linguists who believe (I am not one of them) that human prehistory can be reconstructed reliably by cobbling modern languages into family trees, and that these trees mirror the diaspora of their speakers; let readers with Cornish and Gaelic-speaking ancestors ponder the likely validity of such claims. It is sadly true that wondrous claims are more powerful attractors of funds than humble doubt.

Also true is that Bellwood's Man's Conquest of the Pacific remains a work to which specialist and layman alike may turn, even though, published more than 20 years ago, it is naturally rather out of date. It is better organised, richer and wider than Kirch's book, encompassing the southeast Asian mainland, as it should, which Kirch's work ignores. Its photographs and illustrations, more abundant than Kirch's, seem to have been chosen for their informative power rather than their visual impact. Its bibliography is more representative, as Bellwood does not reserve for himself the lion's share, as does Kirch, who lists no fewer than 76 of his own publications, well ahead of the distant second, Roger Green, with 48 references. Thus, an up-to-date prehistory of the Pacific remains to be written, hopefully by "hard-line positivists", or even soft-line sceptics, without the hard sell so evident here.

Jacques B. M. Guy was formerly an Austronesian comparative linguist with a PhD from the Australian National University and now works in natural-language understanding. He is also an expert on the Easter Island script.