Collapse: How Societies Choose to Fail or Succeed

by Jared Diamond

  By what route was Easter itself, the Polynesian island farthest east, occupied? Winds and currents would probably have ruled out a direct voyage to Easter from the Marquesas, which supported a large population and do seem to have been the immediate source for Hawaii’s settlement. Instead, the jumping-off points for the colonization of Easter are more likely to have been Mangareva, Pitcairn, and Henderson, which lie about halfway between the Marquesas and Easter, and the fates of whose populations will be the story of our next chapter (Chapter 3). The similarity between Easter’s language and Early Mangarevan, the similarity between a Pitcairn statue and some Easter statues, the resemblances of Easter tool styles to Mangarevan and Pitcairn tool styles, and the correspondence of Easter Island skulls to two Henderson Island skulls even more closely than to Marquesan skulls all suggest use of Mangareva, Pitcairn, and Henderson as stepping-stones. In 1999 the reconstructed Polynesian sailing canoe Hokule’a succeeded in reaching Easter from Mangareva after a voyage of 17 days. To us modern landlubbers, it is literally incredible that canoe voyagers sailing east from Mangareva could have had the good luck to hit an island only nine miles wide from north to south after such a long voyage. However, Polynesians knew how to anticipate an island long before land became visible, from the flocks of nesting seabirds that fly out over a radius of a hundred miles from land to forage. Thus, the effective diameter of Easter (originally home to some of the largest seabird colonies in the whole Pacific) would have been a respectable 200 miles to Polynesian canoe-voyagers, rather than a mere nine.

  Easter Islanders themselves have a tradition that the leader of the expedition to settle their island was a chief named Hotu Matu’a (“the Great Parent”) sailing in one or two large canoes with his wife, six sons, and extended family. (European visitors in the late 1800s and early 1900s recorded many oral traditions from surviving islanders, and those traditions contain much evidently reliable information about life on Easter in the century or so before European arrival, but it is uncertain whether the traditions accurately preserve details about events a thousand years earlier.) We shall see (Chapter 3) that the populations of many other Polynesian islands remained in contact with each other through regular interisland two-way voyaging after their initial discovery and settlement. Might that also have been true of Easter, and might other canoes have arrived after Hotu Matu’a? Archaeologist Roger Green has suggested that possibility for Easter, on the basis of similarities between some Easter tool styles and the styles of Mangarevan tools at a time several centuries after Easter’s settlement. Against that possibility, however, stands Easter’s traditional lack of dogs, pigs, and some typical Polynesian crops that one might have expected subsequent voyagers to have brought if those animals and crops had by chance failed to survive in Hotu Matu’a’s canoe or had died out soon after his arrival. In addition, we shall see in the next chapter that finds of numerous tools made of stone whose chemical composition is distinctive for one island, turning up on another island, unequivocally prove interisland voyaging between the Marquesas, Pitcairn, Henderson, Mangareva, and Societies, but no stone of Easter origin has been found on any other island or vice versa. Thus, Easter Islanders may have remained effectively completely isolated at the end of the world, with no contact with outsiders for the thousand years or so separating Hotu Matu’a’s arrival from Roggeveen’s.

  Given that East Polynesia’s main islands may have been settled around A.D. 600-800, when was Easter itself occupied? There is considerable uncertainty about the date, as there is for the settlement of the main islands. The published literature on Easter Island often mentions possible evidence for settlement at A.D. 300-400, based especially on calculations of language divergence times by the technique known as glottochronology, and on three radiocarbon dates from charcoal in Ahu Te Peu, in the Poike ditch, and in lake sediments indicative of forest clearance. However, specialists on Easter Island history increasingly question these early dates. Glottochronological calculations are considered suspect, especially when applied to languages with as complicated histories as Easter’s (known to us mainly through, and possibly contaminated by, Tahitian and Marquesan informants) and Mangareva’s (apparently secondarily modified by later Marquesan arrivals). All three of the early radiocarbon dates were obtained on single samples dated by older methods now superseded, and there is no proof that the charcoal objects dated were actually associated with humans.

  Instead, what appear to be the most reliable dates for early occupation of Easter are the radiocarbon dates of A.D. 900 that paleontologist David Steadman and archaeologists Claudio Cristino and Patricia Vargas obtained on wood charcoal and on bones of porpoises eaten by people, from the oldest archaeological layers offering evidence of human presence at Easter’s Anakena Beach. Anakena is by far the best canoe landing beach on the island, the obvious site at which the first settlers would have based themselves. The dating of the porpoise bones was done by the modern state-of-the-art radiocarbon method known as AMS (accelerator mass spectrometry), and a so-called marine reservoir correction for radiocarbon dating of bones of marine creatures like porpoises was roughly estimated. These dates are likely to be close to the time of first settlement, because they came from archaeological layers containing bones of native land birds that were exterminated very quickly on Easter and many other Pacific islands, and because canoes to hunt porpoises soon became unavailable. Hence the current best estimate of Easter’s settlement is somewhat before A.D. 900.

  What did the islanders eat, and how many of them were there?

  At the time of European arrival, they subsisted mainly as farmers, growing sweet potatoes, yams, taro, bananas, and sugarcane, plus chickens as their sole domestic animal. Easter’s lack of coral reefs or of a lagoon meant that fish and shellfish made a smaller contribution to the diet than on most other Polynesian islands. Seabirds, land birds, and porpoises were available to the first settlers, but we shall see that they declined or disappeared later. The result was a high-carbohydrate diet, exacerbated by the islanders’ compensating for Easter’s limited sources of fresh water by copiously drinking sugarcane juice. No dentist would be surprised to learn that the islanders ended up with the highest incidence of cavities and tooth decay of any known prehistoric people: many children already had cavities by age 14, and everyone did by their 20s.

  Easter’s population at its peak has been estimated by methods such as counting the number of house foundations, assuming 5 to 15 people per house, and assuming one-third of identified houses to have been occupied simultaneously, or by estimating the numbers of chiefs and their followers from the numbers of platforms or erected statues. The resulting estimates range from a low of 6,000 to a high of 30,000 people, which works out to an average of 90 to 450 people per square mile. Some of the island’s area, such as the Poike Peninsula and the highest elevations, was less suitable for agriculture, so that population densities on the better land would have been somewhat higher, but not much higher because archaeological surveys show that a large fraction of the land surface was utilized.

  As usual anywhere in the world when archaeologists debate rival estimates for prehistoric population densities, those preferring the lower estimates refer to the higher estimates as absurdly high, and vice versa. My own opinion is that the higher estimates are more likely to be correct, in part because those estimates are by the archaeologists with the most extensive recent experience of surveying Easter: Claudio Cristino, Patricia Vargas, Edmundo Edwards, Chris Stevenson, and Jo Anne Van Tilburg. In addition, the earliest reliable estimate of Easter’s population, 2,000 people, was made by missionaries who took up residence in 1864 just after an epidemic of smallpox had killed off most of the population. And that was after the kidnapping of about 1,500 islanders by Peruvian slave ships in 1862-63, after two previous documented smallpox epidemics dating back to 1836, after the virtual certainty of other undocumented epidemics introduced by regular European visitors from 1770 onwards, and after a steep population crash that began in the 1600s and t
hat we shall discuss below. The same ship that brought the third smallpox epidemic to Easter went on to the Marquesas, where the resulting epidemic is known to have killed seven-eighths of the population. For these reasons it seems to me impossible that the 1864 post-smallpox population of 2,000 people represented the residue of a pre-smallpox, pre-kidnapping, pre-other-epidemic, pre-17th-century-crash population of only 6,000 to 8,000 people. Having seen the evidence for intensive prehistoric agriculture on Easter, I find Claudio’s and Edmundo’s “high” estimates of 15,000 or more people unsurprising.

  That evidence for agricultural intensification is of several types. One type consists of stone-lined pits 5 to 8 feet in diameter and up to 4 feet deep that were used as composting pits in which to grow crops, and possibly also as vegetable fermentation pits. Another type of evidence is a pair of stone dams built across the bed of the intermittent stream draining the southeastern slope of Mt. Terevaka, in order to divert water onto broad stone platforms. That water diversion system resembles systems for irrigated taro production elsewhere in Polynesia. Still further evidence for agricultural intensification is numerous stone chicken houses (called hare moa), mostly up to 20 feet long (plus a few 70-foot monsters), 10 feet wide, and 6 feet high, with a small entrance near the ground for chickens to run in and out, and with an adjacent yard ringed by a stone wall to prevent the precious chickens from running away or being stolen. If it were not for the fact that Easter’s abundant big stone hare moa are overshadowed by its even bigger stone platforms and statues, tourists would remember Easter as the island of stone chicken houses. They dominate much of the landscape near the coast, because today the prehistoric stone chicken houses—all 1,233 of them—are much more conspicuous than the prehistoric human houses, which had only stone foundations or patios and no stone walls.

  But the most widespread method adopted to increase agricultural output involved various uses of lava rocks studied by archaeologist Chris Stevenson. Large boulders were stacked as windbreaks to protect plants from being dried out by Easter’s frequent strong winds. Smaller boulders were piled to create protected aboveground or sunken gardens, for growing bananas and also for starting seedlings to be transplanted after they had grown larger. Extensive areas of ground were partly covered by rocks placed at close intervals on the surface, such that plants could come up between the rocks. Other large areas were modified by so-called “lithic mulches,” which means partly filling the soil with rocks down to a depth of a foot, either by carrying rocks from nearby outcrops or else by digging down to and breaking up bedrock. Depressions for planting taro were excavated into natural gravel fields. All of these rock windbreaks and gardens involved a huge effort to construct, because they required moving millions or even billions of rocks. As archaeologist Barry Rolett, who has worked in other parts of Polynesia, commented to me when he and I made our first visit to Easter together, “I have never been to a Polynesian island where people were so desperate, as they were on Easter, that they piled small stones together in a circle to plant a few lousy small taro and protect them against the wind! On the Cook Islands, where they have irrigated taro, people will never stoop to that effort!”

  Indeed, why did farmers go to all that effort on Easter? On farms in the northeastern U.S. where I spent my boyhood summers, farmers exerted themselves to carry stones out of fields, and would have been horrified at the thought of intentionally bringing stones into the fields. What good does it do to have a rocky field?

  The answer has to do with Easter’s windy, dry, cool climate that I already described. Rock garden or lithic mulch agriculture was invented independently by farmers in many other dry parts of the world, such as Israel’s Negev desert, southwestern U.S. deserts, and dry parts of Peru, China, Roman Italy, and Maori New Zealand. Rocks make the soil moister by covering it, reducing evaporative water loss due to sun and wind, and replacing a hard surface crust of soil that would otherwise promote rain runoff. Rocks damp out diurnal fluctuations in soil temperature by absorbing solar heat during the day and releasing it at night; they protect soil against being eroded by splashing rain droplets; dark rocks on lighter soil warm up the soil by absorbing more solar heat; and rocks may also serve as slow-time-release fertilizer pills (analogous to the slow-time-release vitamin pills that some of us take with breakfast), by containing needed minerals that gradually become leached out into the soil. In modern agricultural experiments in the U.S. Southwest designed to understand why the ancient Anasazi (Chapter 4) used lithic mulches, it turned out that the mulches yielded big advantages to farmers. Mulched soils ended up with double the soil moisture content, lower maximum soil temperatures during the day, higher minimum soil temperatures at night, and higher yields for every one of 16 plant species grown—four times higher yields averaged over the 16 species, and 50 times higher yields of the species most benefited by the mulch. Those are enormous advantages.

  Chris Stevenson interprets his surveys as documenting the spread of rock-assisted intensive agriculture on Easter. For about the first 500 years of Polynesian occupation, in his view, farmers remained in the lowlands within a few miles of the coast, in order to be closer to freshwater sources and fishing and shellfishing opportunities. The first evidence for rock gardens that he can discern appears around A.D. 1300, in higher-elevation inland areas that have the advantage of higher rainfall than coastal areas but cooler temperatures (mitigated by the use of dark rocks to raise soil temperatures). Much of Easter’s interior was converted into rock gardens. Interestingly, it seems clear that farmers themselves didn’t live in the interior, because there are remains of only small numbers of commoners’ houses there, lacking chicken houses and with only small ovens and garbage piles. Instead, there are scattered elite-type houses, evidently for resident upper-class managers who ran the extensive rock gardens as large-scale plantations (not as individual family gardens) to produce surplus food for the chiefs’ labor force, while all the peasants continued to live near the coast and walked back and forth several miles inland each day. Roads five yards wide with stone edges, running between the uplands and the coast, may mark the routes of those daily commutes. Probably the upland plantations did not require year-round effort: the peasants just had to march up and plant taro and other root crops in the spring, then return later in the year for the harvest.

  As elsewhere in Polynesia, traditional Easter Island society was divided into chiefs and commoners. To archaeologists today, the difference is obvious from remains of the different houses of the two groups. Chiefs and members of the elite lived in houses termed hare paenga, in the shape of a long and slender upside-down canoe, typically around 40 feet long (in one case, 310 feet), not more than 10 feet wide, and curved at the ends. The house’s walls and roof (corresponding to the canoe’s inverted hull) were of three layers of thatch, but the floor was outlined by neatly cut and fitted foundation stones of basalt. Especially the curved and beveled stones at each end were difficult to make, prized, and stolen back and forth between rival clans. In front of many hare paenga was a stone-paved terrace. Hare paenga were built in the 200-yard-broad coastal strip, 6 to 10 of them at each major site, immediately inland of the site’s platform bearing the statues. In contrast, houses of commoners were relegated to locations farther inland, were smaller, and were associated each with its own chicken house, oven, stone garden circle, and garbage pit—utilitarian structures banned by religious tapu from the coastal zone containing the platforms and the beautiful hare paenga.

  Both oral traditions preserved by the islanders, and archaeological surveys, suggest that Easter’s land surface was divided into about a dozen (either 11 or 12) territories, each belonging to one clan or lineage group, and each starting from the seacoast and extending inland—as if Easter were a pie cut into a dozen radial wedges. Each territory had its own chief and its own major ceremonial platforms supporting statues. The clans competed peacefully by seeking to outdo each other in building platforms and statues, but eventually their competition took the form of fe
rocious fighting. That division into radially sliced territories is typical for Polynesian islands elsewhere in the Pacific. What is unusual in that respect about Easter is that, again according to both oral traditions and archaeological surveys, those competing clan territories were also integrated religiously, and to some extent economically and politically, under the leadership of one paramount chief. In contrast, on both Mangareva and the larger Marquesan islands each major valley was an independent chiefdom locked in chronic fierce warfare against other chiefdoms.

  What might account for Easter’s integration, and how was it detectable archaeologically? It turns out that Easter’s pie does not consist of a dozen identical slices, but that different territories were endowed with different valuable resources. The most obvious example is that Tongariki territory (called Hotu Iti) contained Rano Raraku crater, the island’s only source of the best stone for carving statues, and also a source of moss for caulking canoes. The red stone cylinders on top of some statues all came from Puna Pau quarry in Hanga Poukura territory. Vinapu and Hanga Poukura territories controlled the three major quarries of obsidian, the fine-grained volcanic stone used for making sharp tools, while Vinapu and Tongariki had the best basalt for hare paenga slabs. Anakena on the north coast had the two best beaches for launching canoes, while Heki’i, its neighbor on the same coast, had the third best beach. As a result, artifacts associated with fishing have been found mainly on that coast. But those same north-coast territories have the poorest land for agriculture, the best land being along the south and west coasts. Only five of the dozen territories had extensive areas of interior uplands used for rock-garden plantations. Nesting seabirds eventually became virtually confined to a few offshore islets along the south coast, especially in Vinapu territory. Other resources such as timber, coral for making files, red ochre, and paper mulberry trees (the source of bark pounded into tapa cloth) were also unevenly distributed.