Into the Wind: The Peopling of the Eastern Pacific

AThe settlement of the eastern Pacific is the largest movement of people into empty ocean that we know of, and for two centuries it was argued about with far more confidence than evidence. Until the 1990s the standard chronology allowed some eighteen hundred years for the process. Then a re-audit published in 2011 went back to the raw material: of 1,400 radiocarbon determinations from island sites, only 268 were made on samples whose age at death could be controlled — short-lived seeds, twigs, the bones of birds that ate nothing from the sea. The rest were set aside. What survived compressed the whole eastern expansion into roughly the years AD 1025 to 1290, a span shorter than the errors that had preceded it. The tightening was welcomed as a triumph, and it was one. But a tighter chronology tells us when the canoes arrived; it does not tell us why they left, and the two questions have been run together far too often since.

BMarama Teariki, a maritime archaeologist at the University of Otautahi, has spent twenty years on the second of them. Her argument turns on a single navigational fact: across most of the region the trade winds blow from the east, and the unfound islands lay to the east as well. A canoe that runs downwind into unknown water can come home only by beating back against the very wind that carried it out. Between 2004 and 2019 her team logged 38 experimental passages in replica double-hulled craft, and found that a well-handled canoe could hold a course within 75 degrees of the wind on the outward leg, then retrace that leg in about a third of the time by running home before the same wind. Setting out into the wind, on her reading, was not bravado but its opposite: it turned exploration into what she calls a guaranteed return. Crews went east because going east could be undone.

CAmira Adebayo, a climate historian at Bergen, accepts the sailing arithmetic and rejects the conclusion drawn from it. Wind is not a constant, he objects, and the easterly regime that makes the argument work breaks down for months at a time during strong El Niño events, when westerlies run far out into the tropical Pacific. His group put 100,000 simulated voyages through a model driven by winds reconstructed from coral isotope records rather than by modern averages. Under mean conditions a canoe leaving Mangareva made landfall on Rapa Nui, 2,600 kilometres away across water with nothing in it, in 2 per cent of trials; during the prolonged westerly anomalies his records place between AD 1140 and 1260, the figure rose to 31 per cent. Opportunity and achievement arrive together, he concludes, and no experiment with a replica can prise them apart. What looks like a strategy may be only a register of the years in which the ocean permitted one.

DNeither side disputes that such knowledge could be lost as easily as won. Geochemical sourcing has traced nineteen basalt adzes recovered from one central Pacific atoll to a single quarry some 2,500 kilometres off, and every one of them falls inside a 120-year band; after that band, nothing arrives from anywhere. Comparable sequences turn up across the region, and they carry a condition that is easy to read past: long-distance exchange outlasted two or three generations only where a settlement could draw on more than one source, so that the failure of a single route did not end the traffic altogether. Teariki treats the collapse of voyaging as evidence that it had been a technique, since techniques are the kind of thing communities abandon. Adebayo answers that a closing climate window would produce exactly the same silence, and that this silence is the one signal both of them can read whichever way suits.

EThe provisioning figures unsettle both accounts. A double canoe carrying twenty people needs, by Teariki's own estimate, 60 litres of fresh water a day, and the largest reconstructed hulls hold about eleven days' worth; at an average of 4.5 knots the crossing from Mangareva takes sixteen. Rain and fishing must therefore have covered the shortfall, which they do dependably in some seasons and not at all in others. Then there is the sweet potato, a South American plant present in the eastern islands by about AD 1200 and known there, in several languages, by a word close to the Andean kumara. Teariki reads it as proof of a completed round trip and so of navigation under deliberate control. Adebayo replies that a single arrival need not imply a voyage completed in both directions, since the plant travels well in a hull and the people who carried it may never have gone home.

FTwo features of the quarrel deserve more attention than they get. The first is that the antagonists are not answering the same question. Teariki asks what a crew was capable of doing on purpose; Adebayo asks how often the ocean let anybody try. A result that strengthens one account need not weaken the other, and treating the dispute as a contest has cost the field a decade. The second is the traffic between model and voyage. Replica passages are sailed by crews who know that land lies ahead of them and that a rescue vessel is somewhere behind, while simulations inherit whatever the wind reconstruction has assumed. Adebayo concedes that the evidence behind his reconstruction comes from one archipelago alone, and that a single badly read century in those corals would shift his windows by decades. That admission is worth more than another season of replica sailing.