You are on page 1of 16

ANRV287-AN35-04 ARI 13 August 2006 6:41

Archaeology of Overshoot
and Collapse
Joseph A. Tainter
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

Global Institute of Sustainability and School of Human Evolution and Social Change,
Arizona State University, PO Box 873211, Tempe, Arizona 85287-3211;
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

email: Joseph.Tainter@asu.edu

Annu. Rev. Anthropol. 2006. 35:59–74 Key Words


First published online as a Review in intensification, Malthus, resources
Advance on May 10, 2006

The Annual Review of Anthropology is Abstract


online at anthro.annualreviews.org
The literature on sustainability and the human future emphasizes
This article’s doi: the belief that population and/or mass consumption caused resource
10.1146/annurev.anthro.35.081705.123136
degradation and collapse in earlier societies. Archaeological litera-
Copyright  c 2006 by Annual Reviews. ture proposing overshoot and collapse appears in current debates
All rights reserved
over resource conservation versus continued economic growth. The
0084-6570/06/1021-0059$20.00 prominence of this debate, with its national and international di-
mensions, makes it important to assess whether there is evidence in
the archaeological literature for overshoot and collapse brought on
by Malthusian overpopulation and/or mass consumption.

59
ANRV287-AN35-04 ARI 13 August 2006 6:41

Overshoot: the
INTRODUCTION The concept of overshoot clearly depends on
outcome when a that of carrying capacity, which Catton (1980)
Overshoot is a term and concept in the wild.
trajectory is defines as follows:
Like a computer virus in the wild, it prolif-
unsustainable for
erates in the public arena. It mutates like a the maximum population of a given species
environmental,
technical, or social biological virus, assuming altered forms and which a particular habitat can support indef-
reasons new meanings. Overshoot is part of con- initely (under specified technology and or-
Collapse: rapid loss temporary politics, ideology, and public dis- ganization, in the case of the human species)
of an established course. Many believe that humanity has over- (p. 272).
level of social, shot the carrying capacity of some resource or Catton did not seem to realize that this quali-
political, and/or other (e.g., Rees 2004). Conversely, neoclassi-
economic complexity fier may make the concept moot for humans.
cal economists and the politicians they influ- As an ecologist he emphasized population,
ence argue that resources, and the concept of whereas today’s concerns are about both pop-
overshoot, can be left out of economic calcu- ulation and consumption.
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

lations. As a resource becomes scarce, they be- Collapse, too, is in the wild, as shown by
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

lieve market prices will signal that there are re- Diamond’s (2005) popular book Collapse: How
wards to innovation: A new resource or tech- Societies Choose to Fail or Succeed. The term col-
nology will emerge. This debate tinges public lapse has many meanings. Colloquially, col-
life and international relations. It also means lapse means everything from what happened
that academic discourse, including archaeo- to the Soviet Union to what a worker may do
logical and historical research, about over- at the end of a hard day. Societies collapse, but
shoot has political connotations. then so do bridges, levees, and cardiovascu-
The concept of overshoot is often traced lar systems. Academically, the problem is not
to Malthus’ (1798) Essay on the Principle of that definitions vary, but that scholars some-
Population. Malthus believed that whereas times discuss the collapses of societies with-
agricultural production increases linearly, out defining the term. Authors assume that
population tends to grow geometrically. Pop- we know what it means, without individual,
ulation will overshoot food supply. Malthus cultural, or temporal variation.
was influenced by Wallace (1761), who ar- Where authors fail to define collapse,
gued that progress would undermine itself by I use my own definition (Tainter 1988).
filling the world with people. Stimulated by Diamond (2005) does explicitly define col-
Malthus, Jevons (1866 [1865]) wrote The Coal lapse, giving priority to population: “By col-
Question, in which he proposed that Britain’s lapse, I mean a drastic decrease in human pop-
industrial development would outrun the coal ulation size and/or political/economic/social
supply. Ehrlich (1968) brought Malthusian complexity, over a considerable area, for an
overshoot to public attention in The Popula- extended time” (p. 3).
tion Bomb. The concept was systematized by The interest in Diamond’s book suggests
Catton (1980) in Overshoot: The Revolutionary that, in the popular imagination, collapse re-
Basis of Ecological Change. Catton defines over- sults from overshoot. The belief that previ-
shoot as follows: ous collapses did result from overshoot (Rees
(v.) to increase in number so much that the 2004) contributes to the debate over our own
habitat’s carrying capacity is exceeded by the future. While recognizing a wider range of
ecological load, which must in time decrease writings about resources and collapse, I con-
accordingly; (n.) the condition of having ex- centrate on literature pertinent to the con-
ceeded for the time being the permanent car- temporary debate. Is there evidence in his-
rying capacity of the habitat (p. 278). tory or prehistory that population and/or
mass consumption overshot carrying capacity

60 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

and degraded resources, causing societies to must have been introduced from elsewhere.
Intensification:
collapse? By this circuitous reasoning, the Maya over- increase in the inputs
Although we commonly consider over- shot the limits of the environment in which to a productive
shoot to be an excess of population and/or they found themselves, and so collapsed. system to obtain
consumption relative to natural resources, Culbert (1988) argues that the Maya could higher gross outputs
overshoot may take several forms. A com- not achieve equilibrium because they “were
mon theme in the study of early states is a growth system” (p. 77). Population may
whether elite demands for taxes and/or labor have reached 200/km2 . Intensive agriculture
exceeded the peasants’ tolerance or capacity could have supported such dense popula-
to produce (e.g., Lowe 1985). Agrarian tions, but at a cost of grass invasion, fertil-
empires may overshoot a sustainable size ity loss, and erosion. Sustaining such a system
within the constraints of transportation and required high levels of continual labor. Yet
communication—a province too far, so to labor was diverted into monument construc-
speak (e.g., Tainter 1988). Government costs tion and warfare, depriving the agricultural
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

may grow to the point that they undermine sector of necessary maintenance. The re-
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

a state’s ability to respond to crises (Phillips sulting environmental degradation ultimately


1979). More broadly, social and political com- created the conditions for a collapse.
plexity carry costs, which may increase to The most widely accepted archaeological
the level of becoming unsustainable (Tainter cases of overshoot and collapse were proposed
1988). These topics merit their own reviews. by Jacobsen & Adams (1958; Adams 1981). In
ancient southern Mesopotamia, intensive irri-
gation could produce growing prosperity. The
STUDIES IN OVERSHOOT AND Third Dynasty of Ur (ca. 2100 to 2000 b.c.)
COLLAPSE expanded the irrigation system and encour-
In 1931, Cooke proposed that the collapse in aged growth of population and settlement. It
the southern Maya lowlands (Classic period established a bureaucracy to collect taxes and
ca. a.d. 250–830) was caused by soil erosion tribute. Unfortunately, after a few years of
and land scarcity, encroachment of grasses, overirrigating, saline groundwaters rose and
silting of lakes with attendant destruction of destroyed the basis of agricultural productiv-
water transportation, decline in water supply ity. The political system lost its resource base
in dry years, and an increase in mosquito pop- and was destabilized. Large irrigation systems
ulations along with the introduction of, or in- that required central management were use-
crease in, malaria. Thirty years later, Sanders less once the state could not maintain them.
(1962, 1963) conducted an extensive study of In the Early Dynastic period (ca. 2900 to
lowland ecology and reached similar conclu- 2300 b.c.), crop yields per hectare averaged
sions. He argued that swidden agriculture in 2030 liters. Under the Third Dynasty of Ur
this region leads to soil depletion, weeds, and this declined to 1134 liters. Ur III farmers had
savanna formation. Between the times when to plant an average of 55.1 liters per hectare,
Cooke and Sanders wrote, Betty Meggers double the previous rate. The Third Dynasty
(1954) proposed her environmental limitation of Ur pursued intensification that yielded di-
theory: More productive environments can minishing returns.
sustain more complex societies. Cultural com- The Third Dynasty of Ur hung on
plexity is limited by environmental potential, through five kings and then collapsed. The
especially for agriculture. Meggers considered consequences were catastrophic for the pop-
tropical rainforests to have insufficient agri- ulation. By 1700 b.c. yields were down to 718
cultural potential to support much cultural liters per hectare. Of the fields still in pro-
complexity. This raised the problem of the duction, more than one fourth averaged only
Maya. Her solution was that Maya civilization about 370 liters per hectare. The labor to farm

www.annualreviews.org • Archaeology of Overshoot and Collapse 61


ANRV287-AN35-04 ARI 13 August 2006 6:41

a hectare of land was constant, so for equal for centuries in 10,000 square kilometers of
efforts cultivators took in harvests about one Mesopotamia.
third the size of harvests in the previous mil- Chew (2001) has employed the framework
lennium. Soon southern Babylonia was exten- of World Systems Theory to explore ecolog-
sively abandoned. By a millennium or so after ical degradation between 3000 b.c. and the
the Third Dynasty of Ur the number of set- present. Chew believes that the relations be-
tlements was down 40%, and the settled area tween culture and nature are exploitative and
contracted 77%. Population did not rebound that the outcomes include “civilizational col-
to Ur III densities until the first centuries lapse” (p. 1). More broadly, “[i]t is the inter-
a.d. (Adams 1981). play between the limits of Nature and the
The fullest development of Mesopotamian trends and dynamics of social, political, and
agriculture began in the sixth century a.d. economic relations that ultimately defines the
Under the Abbasid Caliphate (750–1258) the historical tendencies of transformation and
needs of the state took precedence over peas- evolution of societal systems” (p. 2). Col-
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

ants’ ability to pay. Tax was fixed whatever the lapse is one kind of societal transformation
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

yield, so peasants had to cultivate intensively. and also an environmental interlude: “[D]ark
Taxes were no longer remitted for crop failure. ages should be appreciated as periods for
Fifty percent of a harvest was owed under the the restoration of the ecological balance . . . ”
Caliph Mahdi (775–785), with many supple- (p. 10). Chew believes that during the Bronze
mental payments. Sometimes taxes were de- Age of Mesopotamia and the Indus Valley,
manded before a harvest, even before the next excess consumption produced environmental
year’s harvest. degradation that led to collapse.
As the irrigation system grew in size and In the Tigris and Euphrates valleys, argues
complexity, maintenance was no longer within Chew (2001, pp. 20–26), wood require-
local capacity. Communities depended on the ments for manufacturing, urban consump-
government, which became unstable. Fields tion, construction, implements, and ship-
again developed problems of salinization. building caused deforestation in northern
Peasants lacked reserves, and revolts were Mesopotamia. Combined with overgrazing,
inevitable. Civil war meant that the hierar- the result was siltation of irrigation systems.
chy could not manage the irrigation system. Canals had to be cleaned regularly. When
Mesopotamia experienced an unprecedented there was unrest, agriculture declined. This
collapse. In the period from 788 to 915 rev- decline was exacerbated by the problem of
enues fell 55%. At the center of the empire, salinization. Overall, core-periphery relations
the Sawad region had supplied 50% of the coupled with high consumption to degrade
government’s revenues. This dropped within the environment (p. 26). This problem was
a few decades to 10%. Most of this loss oc- exacerbated when peripheral elites tried to
curred between the years 845 and 915. In emulate the lifestyles of core elites.
many areas there were revenue losses of 90% But human action was not the only factor.
within a lifetime. State control contracted, Between 3100 and 1200 b.c., Mesopotamia
which diminished any chance to resolve the experienced declining precipitation (Chew
agricultural problems. By the early tenth cen- 2001, pp. 36–39). This decline was combined
tury irrigation weirs were limited to the vicin- with increases in evapotranspiration and salin-
ity of Baghdad. ization. Agriculture was intensified nonethe-
In portions of Mesopotamia the occupied less. In Chew’s assessment, the eventual
area had shrunk by 94% by the 11th cen- decline in agricultural production brought
tury. Population dropped to the lowest level collapse to southern Mesopotamia and a shift
in five millennia. Urban life was eliminated in power to the north.

62 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

The Harappan Civilization of the Indus and perhaps internal rebellions. Mycenaean
Valley and beyond, asserts Chew (2001), had Greece ultimately had to import food and
the same culture-nature relations as southern metals. The Hittite collapse disrupted the
Mesopotamia (pp. 26–36). Demand for wood, trade routes on which Mycenaean imports
including wood needed to manufacture fired depended. Notwithstanding such exogenous
brick, led to “extreme deforestation of the factors, Chew concludes that “[e]cological
Harappan landscape and its hinterland” scarcity required a downscaling of material
(p. 27). Increased aridity between 1800 and and cultural lifestyles” (p. 60).
1500 b.c. stressed the Harappans, as did a Chew (2001) continues his narrative into
decline in Mesopotamian imports. Tectonic the Roman era. Forest loss was “predomi-
shifts diverted water courses. By ca. 1700 b.c., nant all over the Roman Empire” (p. 93).
“overcultivation, overgrazing, salinity, defor- During 400 years of silver smelting in Iberia,
estation, and flooding contributed to the de- some 500 million trees were cut. Morocco lost
cline of the Harappan urban complex” (p. 35). five million hectares of forest in the Roman
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

As southern Mesopotamia declined simulta- period. Yet Chew refrains from linking this
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

neously, Harappan exports to the Gulf fell. loss to the Roman collapse. Other scholars
People migrated to the north and south, un- are less reticent. Hughes (1975) indicts the
til even small trading towns “merged into the Romans for failing to establish harmonious
countryside” (p. 36). relations among economy, society, and the
Chew sees the same processes in the environment. He believes that this was a
Minoan and Mycenaean collapses (2001, major contributor to the Roman collapse.
pp. 41–61). Population on Crete increased Deforestation led to erosion, the most ac-
until ca. 1600 b.c. Large volumes of wood cessible minerals were mined, lands were
were needed for construction and shipbuild- overgrazed, and agriculture declined. Food
ing. Chew argues that deforestation led to shortages and population decline sapped the
erosion and flash flooding. Crete imported Empire’s strength. In later writing, Hughes &
wood from Mycenaean Greece and metals Thirgood (1982) focused particularly on de-
from various places, in turn producing value- forestation as a cause of collapse.
added exports. The environmental problems Deforestation has also been implicated
on the island weakened Cretan manufactur- in the collapse of Cahokia in the American
ing, and also the island’s political and eco- Midwest. A high volume of wood went into
nomic strength, by the 15th century b.c. Cahokia’s construction and occupation, and
Chew postulates that similar transforma- land was cleared for planting. Lopinot &
tions occurred in Mycenaean Greece. Cretan Woods (1993) point out that in the Stirling
wealth was transferred to the mainland in ex- phase (a.d. 1100–1200) wood use came to be
change for Mycenaean wood. But the Myce- increasingly localized and diversified. Intensi-
naeans had to undertake extensive environ- fied local cutting increased runoff and caused
mental engineering to combat erosion and floods to become more frequent, severe, and
siltation caused by deforestation. As with unpredictable. Milner (1990) notes that after
Mesopotamia and Harappa, “ceaseless ac- 1050 the area experienced rapid soil deposi-
cumulation of capital, core-hinterland rela- tion, and he suggests that it was caused by in-
tions, urbanization, population growth, de- creased runoff from cutting bluff-zone forests
forestation, and intensive land use, ultimately (p. 7). It appears that flood levels were in-
led to severe constraints on the contin- creasing and that this increase reduced the
ued expansion of socioeconomic communi- area of bottomland suitable for farming and
ties of the Aegean” (Chew 2001, p. 56). But habitation (p. 7).
other factors were involved: natural catastro- The most wide-ranging attempt to draw
phes, changes in climate, external invasion, contemporary lessons from past ecological

www.annualreviews.org • Archaeology of Overshoot and Collapse 63


ANRV287-AN35-04 ARI 13 August 2006 6:41

crises is Diamond’s (2005) recent work. lands, for example, are small, remote, and
Diamond’s working model is overshoot, fol- lacking critical resources. They could not be
lowed by degradation and collapse: occupied for long without obtaining resources
from elsewhere. The failure of attempts to oc-
It has long been suspected that many of those
cupy them (which I do not consider collapses)
mysterious abandonments were at least
is attributed by Diamond as much to prob-
partly triggered by ecological problems:
lems plaguing trade partners on Mangareva
people inadvertently destroying the envi-
as to anything done by the occupants of
ronmental resources on which their societies
Pitcairn and Henderson islands (pp. 120–35).
depended . . . . Unsustainable practices led to
For comparison, if the resupply of the Inter-
environmental damage . . . . Consequences
national Space Station were to fail, future his-
for society included food shortages, starva-
torians would not wonder at the fate of the
tion, wars among too many people fight-
astronauts, nor draw broader inferences. The
ing over too few resources, and over-
cases of Pitcairn and Henderson islands are
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

throws of governing elites by disillusioned


no more compelling.
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

masses . . . . The risk of such collapses to-


Norse Greenland is not an enduring les-
day is now a matter of increasing concern
son for the same reason: Local resources could
(pp. 6–7).
not indefinitely support a medieval European
Diamond realized early on that collapses are society. Add to this the problem of low-
more complex than simple overshoot. So he frequency events—rare but inevitable adverse
developed a more nuanced framework involv- climate fluctuations—and it is clear that the
ing environment, climate, hostile neighbors, Norse settlements were temporary (although
friendly trade partners, and societal responses. the Norse themselves did not recognize this).
Yet slips of the pen betray his conviction Although the Norse did put the local envi-
that environmental deterioration is really to ronment to use, and thereby degraded it, the
blame. A modern collapse would be “trig- great problem they faced was the Little Ice
gered ultimately by scarcity of environmental Age (e.g., McGovern 1994). As McGovern
resources” (Diamond 2005, p. 7). Environ- noted, the dilemma is not that the Green-
mental problems “undermined preindustrial land Norse went extinct or left, but that they
societies” (p. 35). “The Anasazi and Maya need not have done either. The continued oc-
were . . . undone by water problems” (p. 490). cupation of the area by the Inuit shows that
“Deforestation was a or the major factor in alternative subsistence strategies and ways of
all the collapses of past societies described in life would have allowed the Norse to sur-
this book” (p. 487, emphasis original). To- vive in Greenland. The Norse failure, argues
day’s Third World trouble spots all suffer from McGovern, lay in not adopting Inuit ways.
environmental deterioration: “[I]t’s the prob- Diamond, like McGovern, wonders at the in-
lems of the ancient Maya, Anasazi, and Easter flexibility of the Norse. Whatever the an-
Islanders playing out in the modern world” swer to that conundrum, Norse Greenland
(p. 516). is not a simple example of overshoot and
Diamond discusses six archaeological collapse. It may illustrate a limited kind of
cases: Easter Island, Pitcairn Island, Hen- overshoot—overshoot in respect to extreme
derson Island, the Anasazi (especially Chaco conditions.
Canyon) of the U.S. Southwest, the Maya, and Diamond’s (2005, pp. 136–56) account of
Norse Greenland. Some of these are places the Anasazi is a confused muddle, includ-
where subsistence producers could not have ing Chaco Canyon as “the Anasazi capital,”
survived long-term and are thus unsuitable for a Chaco “mini-empire,” and perhaps “provin-
deriving broader inferences, let alone lessons cial capitals” elsewhere (pp. 148–49). Di-
for the future. Pitcairn and Henderson is- amond alternates between discussions of

64 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

Chaco Canyon, “warfare-related cannibal- surpassing its limits—crashed devastatingly”


ism” (p. 151), and Long House Valley, with (p. 264). Many have come to regard Easter
mention of other places that “also underwent Island as a metaphor for “Spaceship Earth”
collapses, reorganization, or abandonment at (e.g., Flenley & Bahn 2002; Ponting 1991).
various times within the period a.d. 1100– Easter Island now figures in economic the-
1500” (p. 154). This includes the Mimbres, ory. Brandner & Taylor (1998) use the
Mesa Verdeans, Hohokam, Mogollon, and Lotka-Volterra predator-prey model to sim-
others. One wonders what society, over a ulate overshoot and collapse on the is-
period of 400 years, would not experience land, whereas Dalton & Coats (2000) ex-
some process that could be labeled “reor- plore whether markets and a private-property
ganization.” Diamond concludes that Chaco regime could have saved Easter Island.
Canyon was abandoned in the twelfth century Erickson & Gowdy (2000) use Easter Island to
because of human impact and drought. With evaluate the consequences of substituting hu-
centuries of population growth, demands on man for natural capital. Easter Island is in the
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

the environment grew, resources declined, wild.


by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

and people lived precariously. The proximate Islands are useful laboratories for the study
reason was a drought beginning in a.d. 1130, of some ecological processes. Cut off to vary-
which “pushed Chacoans over the edge” (p. ing degrees from the influence of larger land
156). At a lower population density they might masses, islands permit ecologists to study such
have survived it. things as invasion, population growth, and ex-
Diamond (2005) also considers collapses tinctions in a more controlled environment
in the Maya region (pp. 157–77). The Maya than is otherwise possible. For this reason, and
damaged their environment, of course. They because scholars have recently discussed it at
also fought among themselves over farmland, length, Easter Island merits review in some
and, Diamond believes, emphasized “war and detail.
erecting monuments rather than . . . solving We do not know when Easter Island was
underlying problems” (p. 160). But climate first occupied. Flenley & Bahn (2002) cite a
also played a role. A drought beginning date of a.d. 690 ± 130 (p. 77). This is from
around a.d. 760, and peaking around 800, was a large platform, so it cannot record the ear-
“suspiciously associated with the Classic col- liest settlement. Kirch (2000) notes what may
lapse” (p. 174). Diamond concludes that one have been human forest clearance as early as
strand in the Classic Maya collapse was pop- 1630 ± 130 B.P. (p. 271). Initial settlement
ulation overshoot. This contributed to defor- sometime in the first few centuries a.d. seems
estation and erosion. More and more people likely.
fought over fewer and fewer resources. Elites The settlers found an island that was poor
fiddled while the lowlands burned. Drought in resources by the standards of Oceania.
then came to a system at the margins (pp. 176– Its latitude is subtropical, so the climate is
77). Diamond’s analysis of the Maya places cool compared with the rest of Polynesia
them with the Anasazi and Norse Greenland: (Diamond 2005, pp. 83, 86). The ocean is
a society that overshot the level of popula- too cold in winter to support a coral reef, so
tion, consumption, and political complexity the population of marine fauna is depauper-
that could be sustained under rare, extremely ate (McCoy 1979, p. 140). Around the island
adverse circumstances. there are 126 species of fish, compared with
Thus Easter Island is Diamond’s best case 450 species around Hawai’i and 1000 around
of overshoot, resource degradation, and col- Fiji. Sea mammals and turtles may never
lapse. It is not Diamond’s case alone. “Easter have been abundant (Flenley & Bahn 2002,
Island,” wrote Kirch (1984), “is an example of p. 19). As happened elsewhere in Polynesia,
a society which—temporarily but brilliantly the settlers (and the rats that they introduced)

www.annualreviews.org • Archaeology of Overshoot and Collapse 65


ANRV287-AN35-04 ARI 13 August 2006 6:41

destroyed a diverse community of nesting marine mammals and larger fish could be
birds (Steadman 1989, 1995). found (Kirch 1984, pp. 270, 272).
An early faunal assemblage from the Ahu Estimates of maximum population,
Naunau site was analyzed by Steadman et al. reached most likely in the sixteenth century,
(1994). Radiocarbon dates ranged between range from 7000 to 10,000 (Van Tilburg 1994,
660 and 900 B.P. Dolphin was the most com- p. 52), up to 15,000 (Diamond 2005, p. 91).
mon marine species. Bones of the Polyne- Population declined to ∼2000 by 1722
sian rat (introduced for food where Polyne- (Flenley & Bahn 2002, p. 169).
sians settled) were second in abundance, and This decline in population qualifies
chicken bones were outnumbered by those under Diamond’s definition as a collapse
of native birds. Overall, marine mammals, (2005, p. 3). Archaeologists point to concomi-
seabirds, and native land birds were more tant changes in subsistence, and in the so-
common than in later prehistoric sites, and cial, political, and economic spheres. In these
bones of fish and chickens were much rarer. realms, the essence of collapse is a marked
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

The settlers found a forest dominated by a reduction in complexity (Tainter 1988). The
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

now-extinct palm that is related to the Chilean complexity of Easter Island society seems
wine-palm (Flenley & King 1984). Histori- to have changed significantly after ∼1500.
cally this forest no longer exists. Its disap- Labor was no longer organized for corporate
pearance plays a large role in overshoot-and- group projects. Rights to move statues across
collapse interpretations of Easter Island and the island ceased to be exercised, along with
in contemporary imaginings of humanity’s fu- the interclan political and/or ceremonial rela-
ture (e.g., Diamond 2005; Flenley & Bahn tions that this implies. Island-wide exchange
2002). and cooperation declined. The hereditary sta-
Easter Island is known, of course, for its tus system lost influence. Chiefs lost economic
statues. The cessation of statue carving fig- power, including the right to surplus produc-
ures prominently in the idea that Easter Island tion, and a rigid class structure gave way. War-
underwent a collapse. Statues were placed fare became chronic. A warrior class emerged,
on ahu, elongated platforms parallel to the largely supplanting the hereditary rank sys-
shore. Ahu and statues may have been built tem. Late in prehistory the island had no high
over a period of 700 years (Kirch 1984, p. chief. Instead, the island’s descent groups co-
271), starting probably no earlier than a.d. 900 alesced into opposing factions of the east and
(McCoy 1979, p. 152). Many statues were west (Flenley & Bahn 2002; Kirch 1984, 2000;
standing in 1722, when a Dutch ship McCoy 1979). The loss of organizational ca-
commanded by Roggeveen made the first pacity makes it appropriate to label this a
recorded European contact. A Spanish ship collapse.
in 1770 saw much of the island but reported Late in the island’s occupation there was
no fallen statues. Only four years later Cook conflict over land, and defeated people risked
found many statues toppled, and the ahu were dispossession or enslavement. Many lived in
no longer maintained. Subsequent visits con- fortified caves. Late middens have a high fre-
tinued to report toppled statues (Flenley & quency of fractured and charred human re-
Bahn 2002, p. 150). mains, many from juveniles. This is inter-
Society was hierarchical. The island was preted as cannibalism. In time the statues
unified sufficiently to allow transport of stat- were overthrown and many were destroyed
ues from the Rano Raraku quarry to all coasts. (McCoy 1979).
The high-ranking Miru clan occupied the The Easter Islanders themselves ascribe
western and northern parts of the island. As the end of statue construction to the out-
is common in Polynesia, this group had a break of war. Mulloy (1970) proposed that war
monopoly on fishing in deeper waters, where resulted from overpopulation, which led to

66 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

“competition for agricultural land and fishing (2005) sketches the parallels between Easter
rights . . . . With warfare came considerable Island and our own potential future:
cultural decline” (p. 5). In 1983, Flenley &
Polynesian Easter Island was as isolated in
King (1984) cored deposits in the three main
the Pacific Ocean as the Earth is today in
craters and showed what had long been sus-
space . . . . Earthlings have [no] recourse else-
pected: The island was once forested. Defor-
where if our troubles increase. Those are the
estation occurred since ca. 990 ± 70 B.P. The
reasons why people see the collapse of Easter
findings “do not conflict with the hypothesis
Island as a metaphor, a worst-case scenario,
that the decline of megalithic culture was asso-
for what may lie ahead of us in our own
ciated with total deforestation” (p. 50). Mann
future (p. 119).
et al. (2003), undertaking soil analyses, found
that the island’s primeval soils began to erode
severely by a.d. 1200 and that there was for-
est clearance and erosion everywhere between EVALUATING OVERSHOOT
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

1200 and 1650. Deforestation presumably oc- AND COLLAPSE IN


by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

curred as a result of cutting wood to build ARCHAEOLOGY


canoes, using wood for structural elements, The past is contested in many ways, includ-
fashioning wood into implements, cooking, ing whether Malthusian overshoot caused ear-
cremating, perhaps moving and raising stat- lier collapses. We can gather more data, refine
ues, and possibly early swiddening. chronologies, and develop new techniques to
Several authors see the deforestation as the reconstruct paleoenvironments, but none of
start of a cascading process that led to a decline these will resolve the question of whether
in fishing and farming; changes in farming past collapses provide analogues for possible
technology; increases in warfare and insecu- future ones. Fundamental beliefs and short-
rity; changes in settlement patterns; popula- term well-being are at stake. Still, one can try
tion decline; and sociopolitical collapse. For- to make the debate as rational as possible. To
est depletion, in this view, led to a shortage that end, this section evaluates whether those
of wood for canoe construction, and thus to who see overshoot and collapse in past soci-
a decline in the consumption of fish (espe- eties have made their case.
cially deep, pelagic fish) and marine mam- Betty Meggers’ (1954) environmental lim-
mals. Soils eroded, and the fertility of for- itation theory never found widespread favor
est soils was lost. Crops were exposed to and is today rarely discussed. In the Amazon
the winds that blow at Easter Island most of Basin, where Meggers’ own work stimu-
the year. Soils lost moisture. Crop yields de- lated her theory, we now know that some
clined. Springs and streams dried up. Peo- societies had levels of complexity that her
ple responded with agricultural intensifica- theory predicted would be impossible (e.g.,
tion. Stone dams diverted water, while lithic Heckenberger et al. 2003). Meggers did not
mulch was employed to retain soil moisture. consider the role of intensification in raising
Stone-lined pits were created to shelter plants the productivity of land.
from the wind. Rocks were stacked to create Intensification also undercuts the argu-
windbreaks. Stone chicken houses up to 20 ments of Cooke (1931) and Sanders (1962,
meters long were built to prevent theft. Fires 1963) regarding the consequences of swid-
were fueled by herbs and grasses (Ayres 1985; dening in the southern Maya Lowlands. We
Diamond 2005; Flenley & Bahn 2002; Kirch have learned that the southern cities were not
1984, 2000; McCoy 1976, 1979). supported by low-production swiddening, but
Easter Island, in this account, is the by a landscape that managed to give high
paradigmatic case of overshoot and collapse, agricultural yields (e.g., Turner 1974). It is
the prototype Spaceship Earth. Diamond sometimes suggested that this landscape was

www.annualreviews.org • Archaeology of Overshoot and Collapse 67


ANRV287-AN35-04 ARI 13 August 2006 6:41

unsustainable (Culbert 1988), but that ques- which they and their societies depended. We
tion cannot be divorced from the costs and cannot yet answer this question. For the
consequences of sustaining urban centers, Abbasid Caliphate, though, some pertinent
elites, monumental construction, and inter- factors can be discussed.
polity warfare, coupled perhaps with drought. One factor in Abbasid tax policy may have
The Maya collapse cannot be explained by the been the need to fund the frequent wars
simple overshoot of Cooke and Sanders. with the Byzantine Empire. Another factor
Culbert (1988) believes the dramatic was that the Abbasid Caliphate was a costly
population loss of the Maya collapse suggests regime. Under the Abbasids there was un-
agricultural failure. Yet if, as he suggests, agri- precedented urban growth. Baghdad grew to
cultural breakdown came from inept decision- five times the size of tenth-century Con-
making, what ultimately was the problem: stantinople. The capital was moved often, and
population, competition, public construction, each time built anew on a grand scale. The
mismanagement, disaffection, or all these? If Caliph al-Mutasim (833–842) built a new cap-
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

Maya elites had allocated labor better, would ital at Samarra, 120 kilometers upstream from
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

scholars today discuss overshoot? Culbert Baghdad. In 46 years, he and his successors
(p. 100) compares the Maya to Mesopotamia built a city that stretched along the Tigris for
and reaches conclusions similar to those 35 km. It would have dwarfed imperial Rome
of Adams (1981). As with Mesopotamia, (Hodges & Whitehouse 1983, pp. 151–56).
if the southern lowland cities collapsed Overshoot in these cases resulted from
from resource degradation, the main ques- the failure of feedback loops. Negative feed-
tions involve decision-making rather than back, which dampens excesses in a system’s
procreation. This would mean that the Maya behavior, operates by feeding back informa-
collapse is not an example of Malthusian over- tion about the system’s state or behavior to
population but of elite-driven intensification. the control, which responds by altering its
The Third Dynasty of Ur and the own state or behavior. The thermostat is a
Abbasid Caliphate are two of history’s best simple example. The Mesopotamian collapses
candidates for overshoot, degradation, and resulted from a failure of this mechanism.
collapse. Powell (1985) questions the role of Information about the deteriorating state of
salinization in the Ur III collapse on textual agriculture either failed to reach the govern-
grounds. Whether Powell’s objections under- ment or was ignored. As agricultural yields fal-
mine the case, Ur III is poorly known. The tered, these regimes demanded yet more pro-
political situation in Iraq has prevented ar- duction. What should have been a negative
chaeologists from acquiring data to address feedback system, dampening departures from
the questions that the case raises. This consid- sustainability, instead became a positive feed-
eration applies also to the Abbasid Caliphate, back system. The government responded to
but its collapse was more recent and is better agricultural problems in a way that made the
documented. Still, both Mesopotamian cases problems worse. Positive feedback forced the
have limited potential to give a broader under- agricultural system, the regime, indeed
standing of overshoot and collapse. Neither the complexity of the society as a whole, into
is a Malthusian overshoot, nor an overshoot a downward spiral toward collapse.
brought on by excess production for mass The interpretation of overshoot and col-
markets, today’s primary concerns. The agri- lapse in Bronze Age societies advanced by
cultural problems in both cases can be traced Chew (2001) is empirically weak. Chew pro-
to elite mismanagement. The critical ques- vides no data, merely sweeping assertions,
tions then involve not merely the resource, but on the extent of deforestation or erosion
the more fundamental question of why rulers in northern Mesopotamia, the Indus River
would undermine the agricultural regime on watershed, Minoan Crete, and Mycenaean

68 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

Greece. In southern Greece, erosion does not of climate and other factors. These were not
correspond temporally to the Mycenaean col- Malthusian overshoots. In Diamond’s formu-
lapse (van Andel et al. 1990). Paleoenviron- lation, these cases may illustrate overshoot in
mental studies may someday provide further reference to extreme climatic conditions. If
data to test Chew’s ideas. These cases have the extreme conditions had not occurred, the
not yet been demonstrated to be examples of societies in question might not have collapsed.
overshoot. Sheets (1999) has argued that the ability to
Deforestation did not cause the Roman withstand extreme events (volcanism in his
collapse. To the end, Roman mints consumed cases) varies with the complexity of the society.
tons of charcoal to produce millions of coins, This idea suggests that the many archaeolog-
even in places like North Africa. Builders fired ical studies that focus on extreme events such
millions of bricks to build the walls of Rome as drought merit their own review.
in the 270s. There is no evidence of wood Easter Island again requires discussion: It
shortages. Indeed, forests were regrowing in is the last candidate for an overshoot resulting
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

the late empire (Tainter 2000a). from excess population and/or consumption.
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

It is unlikely that deforestation and flood- Diamond (2005) considers Easter Island to be
ing caused the collapse of Cahokia. The rulers “as close as we can get to a ‘pure’ ecological
of Cahokia had demonstrated extraordinary collapse” (p. 20). Yet there are many questions
capacity to move people, relocate settlements, about deforestation, deep-sea voyaging, social
mobilize labor, and build massive earthworks organization, agriculture, and statue produc-
(Dalan 1997). If flooding was a threat to tion in the Easter Island collapse.
Cahokia, the elites could have moved polit- Diamond (2005) writes, “I have often asked
ical centers, villages, agricultural production, myself, ‘What did the Easter Islander who cut
and population to higher ground a short dis- down the last palm tree say while he was do-
tance away. They had already accomplished ing it?’” (p. 114). In fact some scholars suggest
similar feats of organization. that remnants of the island’s forest may have
Diamond’s inclusion of Pitcairn and survived in the volcano craters into the nine-
Henderson islands, and Norse Greenland, teenth century (Flenley and Bahn 2002, pp.
illustrates problems in his conception of col- 86–87; Ponting 1991, p. 5), after which the
lapse. There was no loss of sociopolitical com- last trees may have been killed by sheep and
plexity in these places, with a less complex so- goats (Flenley & Bahn 2002, p. 160). Seeds of
ciety succeeding. The occupants either died or the extinct indigenous palm have been found.
left. Apparently any place that was abandoned These are important: Every seed discovered
is potentially, to Diamond, a collapse. When had been gnawed by the Polynesian rat and
he includes Rwanda in his cases, he equates could not have germinated (Flenley & Bahn
genocide with collapse. There is no logical 2002, pp. 160–61). The rats ate these nuts sys-
consistency among these cases. Diamond’s ap- tematically, and the Polynesians ate the rats.
proach was seemingly to find cases where (a) Once the Polynesian rat was introduced to
bad things happened, and (b) he could con- Easter Island, the palm forest may have been
struct a plausible environmental reason. The doomed. Human use merely sped up the in-
outcomes, however diverse their nature, are evitable. This could explain why practices of
lumped into the category “collapse.” sustainable forest use were not successful, if
Diamond would have liked to have shown they were attempted.
that Chaco Canyon, the Anasazi and other The extinct palm was not uniformly use-
Southwesterners, and the Maya overshot the ful. Palm wood is porous and would have been
capacity of their environments, degraded undesirable for fashioning canoes. The for-
them, and collapsed. In each case, though, est did contain other trees. Even so, “larger
he was confronted with the potential roles trees suitable for canoe construction may have

www.annualreviews.org • Archaeology of Overshoot and Collapse 69


ANRV287-AN35-04 ARI 13 August 2006 6:41

been sparse or lacking” (Kirch 1984, p. 268). posed soils to drying and challenged young
Diamond observes that canoes in the his- plantings. The islanders responded by dig-
toric period were small, flimsy, and leaky. Yet ging pits, erecting small windbreaks, and em-
when Roggeveen approached the island in ploying lithic mulch. Problems of soil fertility
1722, a canoe came nearly 5 kilometers out could have been addressed by shifting cultiva-
to the Dutch ships. The marine zone where tion and/or by use of night soil. In late pre-
the Easter Islanders could obtain delphinids history, chickens were kept in stone structures
and pelagic fish was well within this radius. with small runs. This was reportedly to deter
Moreover, Easter Islanders claimed that they theft, but it would also have the consequence
regularly undertook voyages to the island of making chicken manure easier to collect.
of Salas-y-Gómez, to the east-northeast, a The chicken houses have small entrances and
round-trip of 930 km (Flenley & Bahn 2002, passageways (McCoy 1976, pp. 23–26), but
p. 67). The Easter Islanders have a name not so small that a child could not crawl in-
for this island, and the claim is generally ac- side. Chicken manure could have been used
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

cepted. Salas-y-Gómez is considered Polyne- to enhance soil fertility.


by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

sia’s easternmost tip. Boats could also have Nearly the entire surface of Easter Island
been made of reeds (Van Tilburg 1994, pp. 47– is arable (Kirch 2000, p. 272), yet in recent
48). These facts challenge Diamond’s (2005) centuries only the coast and the interior of
claim that “Lack of timber . . . brought to an Rano Kau crater were intensively cultivated
end . . . the construction of seagoing canoes” (Flenley & Bahn 2002, pp. 96, 158; McCoy
(p. 107). Still, unless the early faunal remains 1976, pp. 78–79, 84). The factor limiting agri-
from Anakena are an unrepresentative sam- culture may have been water (McCoy 1976,
ple, deep-sea catches seem to have declined in p. 142), rather than erosion, wind, or soil
later prehistory (Steadman et al. 1994). fertility. If agricultural productivity declined
Even if lack of wood meant a decline in with deforestation, this could have been com-
the availability of sea mammals and large fish, pensated for by increasing the area under
the entire population would not have been intensive cultivation. Diamond (2005) cites
equally affected. As is common in Polynesia, Cook’s 1774 description of the Easter Is-
the high-ranking Miru clan controlled fishing landers as “small, lean, timid, and miserable”
in deeper, offshore waters. High chiefs had the (p. 109) to support his view that deforestation
right to distribute prestigious fish. The Miru led to starvation. Yet he ignores Roggeveen’s
could place tapu (taboo) on marine resources statement in 1722 that the islanders were “well
from May through October, when only no- proportioned, generally large in stature, very
bles could eat larger fish like tuna (Flenley and sturdy with strong muscles, and extremely
Bahn 2002, p. 100; Kirch 1984, p. 272). If de- strong swimmers” (Flenley & Bahn 2002,
forestation did lead to a decline in the catch of p. 90). When people face agricultural prob-
larger fish and marine mammals, a large part lems, a common response is to intensify pro-
of the population was only minimally affected. duction. This is what the Easter Islanders did
A decline in agriculture would have been (McCoy 1976, pp. 145–146).
more serious than the loss of marine delica- Finally there is the mystery of the end of
cies. Easter Island undoubtedly experienced statue production. Some scholars have sug-
erosion. It is a common mistake, though, to gested that production ceased because there
assume that erosion is always detrimental. was no longer wood to move and erect stat-
Ancient Egypt and Mesopotamia were sus- ues. Others have even suggested that moving
tained by upstream erosion, as other places the statues caused the deforestation. Perhaps
have been. No research has shown that ero- 800 to 1000 statues had been produced, or
sion adversely affected Easter Island agricul- partly produced, when work ceased (Flenley &
ture. Removing forest cover would have ex- Bahn 2002, p. viii). Of these, 324 statues were

70 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

ultimately erected on 245 ahu around the coast of statue production seems to have occurred
(Kirch 2000, p. 272). Some 397 statues re- for reasons other than deforestation.
main in the quarry zone of Rano Raraku crater
(Van Tilburg 1994, p. 21). About 200 statues
lie unfinished within the quarry (Kirch 2000, CONCLUSIONS
p. 272), some of which broke during pro- When the ARA Editorial Committee invited
duction. Ninety-two statues were abandoned me to address the topic “Archaeology of
in transport (Van Tilburg 1994, p. 148). From Overshoot and Collapse,” I assumed I could
these figures it is clear that statue-making con- review only part of a voluminous literature.
tinued at an undiminished rate until produc- Although I have extensively read the col-
tion ceased. It is even possible that the rate lapse literature (Tainter 1988), I was surprised
of production was increasing. When produc- to realize that the literature has produced
tion ceased, it did so abruptly. Had produc- few cases that postulate overshoot of pop-
tion simply tapered off, we would expect to ulation and/or mass consumption, followed
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

see fewer unfinished statues and fewer stat- by degradation and collapse. Writers today
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

ues in the quarry zone. These figures con- prefer to explain collapse by the occurrence
tradict the interpretation that deforestation of extreme events (e.g., Binford et al. 1997).
caused the end of statue production. It is in- Within the small overshoot literature, many
conceiveable that new statues would be com- of the most ardent proponents are outside
missioned at a constant or accelerating pace if archaeology.
the means to transport them was visibly dwin- Some overshoot interpretations are not
dling. Clearly the end of statue production credible (environmental limitations), or have
was linked to some factor other than the end been proven wrong (savannah formation in
of the forest. Even without trees, statues can the Maya Lowlands). Others are untested
be dragged over ground lubricated with sweet (Bronze Age core-periphery systems) or can-
potatoes (Flenley & Bahn 2002, p. 122). This not explain collapse (deforestation in Rome
would take a lot of sweet potatoes, but that and Cahokia). Other cases advanced as over-
resource is more readily replenished than are shoot examples (Anasazi, Maya) prove to
trees. be overshoot in regard to extreme condi-
In short, Easter Island may not qualify as tions. The collapses of the Third Dynasty
a case of overshoot and collapse. The Poly- of Ur and the Abbasid Caliphate seem to be
nesian rat had much to do with the decline of cases of overshoot. Yet the proximate causes
the palm forest, so human use affected the rate were not overpopulation nor mass consump-
of deforestation but not the ultimate outcome. tion, but elite mismanagement and a failure
Diamond (2005) acknowledges that Easter Is- of information feedback. The reasons why
land’s deforestation was due in part to its lat- Mesopotamian elites acted detrimentally in
itude, low rainfall, lack of volcanic ash, lack regard to their long-term interests remain to
of dust from Asia, and small size (pp. 115–18). be determined. This question pertains also
Even without tall trees, Easter Islanders made to the Maya collapse. Easter Island, consid-
water craft capable of long voyages and deep- ered the paradigmatic case of overshoot and
sea fishing. Marine mammals and large fish collapse, is equivocal. Easter Island may be
did decline in the diet, but this decline pri- such a case, but there is contradictory evi-
marily affected the elite segment of the pop- dence. There does not presently appear to be
ulation. Agriculture was probably compro- a confirmed archaeological case of overshoot,
mised by deforestation, but Easter Islanders resource degradation, and collapse brought
responded to this in time-honored fashion: on by overpopulation and/or mass consump-
They intensified production. Finally, the end tion. As a personal aside, I consider myself to

www.annualreviews.org • Archaeology of Overshoot and Collapse 71


ANRV287-AN35-04 ARI 13 August 2006 6:41

be conservation-minded, and I currently fo- ing more land. Wallace, Malthus, Jevons, and
cus my research on sustainability (e.g., Allen Ehrlich have, so far, been wrong.
et al. 2003). I realize the political and ideolog- As an alternative to intensifying, societies
ical implications of this conclusion and how it or institutions may simplify so that they are
might be used. less costly, or people may otherwise reduce
With limited space, I briefly raise two consumption. This was the strategy of the
questions: Why are there no cases of what Byzantine Empire when it lost its wealthiest
Diamond (2005) calls “‘pure’ ecological col- provinces in the seventh century a.d. and re-
lapse” (p. 20)? And does this mean that Space- sponded with what may be history’s only ex-
ship Earth is reprieved? These are rich top- ample of a large, complex society systemati-
ics that cannot be explored here. A few words cally simplifying (Tainter 2000b).
alone are possible, but the fundamental issues The question for our time is whether in-
are quickly drawn. tensification can continue indefinitely. Can we
The concept of overshoot is teleological, forever find some way to escape the Malthu-
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

as if humans could set a target for popula- sian fate? Neoclassical economists assume
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

tion or consumption. Overshoot denies the that, with incentives and unfettered markets,
human capacity for flexible adjustments, in- there will always be new technologies and new
cluding intensifying production. Overshoot resources. Humanity, in this view, need never
and depletion of megafauna, for example, oc- face a crisis of overpopulation or overcon-
curred in many places (e.g., Barnosky et al. sumption. The contrary view is well known:
2004) but seem not to have caused a sin- We must reduce our ecological footprint or
gle collapse. It is usually possible to coax eventually collapse. The neoclassical argu-
more resource production by applying capital ment is based on faith that markets will always
and technology, increasing labor, applying en- work and denial of diminishing returns to in-
ergy subsidies, and making production more novation (Rescher 1978). Should we base our
knowledge-intensive. Irrigation, fertilization, future on faith and denial, or on rational plan-
and mechanization are all ways to increase ning? That is open for debate. It is a question
production, as is putting in more hours or till- in the wild.

ACKNOWLEDGMENTS
I am pleased to express my appreciation to the following colleagues for their suggestions:
Jeanne Arnold, John Bintliff, Jeffrey Dean, Thomas Dillehay, Timothy Kohler, Barbara Mills,
Ben Nelson, Charles Redman, Payson Sheets, and Sander van der Leeuw.

LITERATURE CITED
Adams RM. 1981. Heartland of Cities. Chicago: Aldine
Allen TFH, Tainter JA, Hoekstra TW. 2003. Supply-Side Sustainability. New York: Columbia
Univ. Press
Ayres WS. 1985. Easter Island subsistence. J. Soc. Ocean. 80:103–24
Barnosky AD, Koch PL, Feranec RS, Wing SL, Shabel AB. 2004. Assessing the causes of late
Pleistocene extinctions on the continents. Science 306:70–75
Binford MW, Kolata AL, Brenner M, Janusek JW, Seddon MT, et al. 1997. Climate variation
and the rise and fall of an Andean civilization. Quat. Res. 47:235–48
Brandner JA, Taylor MS. 1998. The simple economics of Easter Island: a Ricardo-Malthus
model of renewable resource use. Am. Econ. Rev. 88:119–38

72 Tainter
ANRV287-AN35-04 ARI 13 August 2006 6:41

Catton WR. 1980. Overshoot: The Ecological Basis of Revolutionary Change. Urbana: Univ. Ill.
Press
Chew SR. 2001. World Ecological Degradation: Accumulation, Urbanization, and Deforestation,
3000 B.C.–A.D. 2000. Walnut Creek, CA: AltaMira
Cooke CW. 1931. Why the Mayan cities of the Peten district, Guatemala, were abandoned. J.
Wash. Acad. Sci. 21(13):283–87
Culbert TP. 1988. The collapse of Classic Maya civilization. In The Collapse of Ancient States
and Civilizations, ed. N Yoffee, GL Cowgill, pp. 69–101. Tucson: Univ. Ariz. Press
Dalan RA. 1997. The construction of Mississippian Cahokia. In Cahokia: Domination and Ideology
in the Mississippian World, ed. TR Pauketat, TE Emerson, pp. 89–102. Lincoln: Univ. Neb.
Press
Dalton TR, Coats RM. 2000. Could institutional reform have saved Easter Is-
land? J. Evol. Econ. 10:489–505. http://springerlink.metapress.com/link.asp?ID=
2G7POXXAU7L8LA9G
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

Diamond JD. 2005. Collapse: How Societies Choose to Fail or Succeed. New York: Viking
by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

Ehrlich PR. 1968. The Population Bomb. New York: Ballantine


Erickson J, Gowdy JM. 2000. Resource use, institutions, and sustainability: a tale of two Pacific
island cultures. Land Econ. 76:345–54
Flenley J, Bahn P. 2002. The Enigmas of Easter Island. Oxford: Oxford Univ. Press. 2nd ed.
Flenley JR, King SM. 1984. Late Quaternary pollen records from Easter Island. Nature 307:47–
50
Heckenberger MJ, Kuikuro A, Kuikuro UT, Russell JC, Schmidt M, et al. 2003. Amazonia
1492: pristine forest or cultural parkland? Science 301:1710–14
Hodges R, Whitehouse D. 1983. Mohammed, Charlemagne and the Origins of Europe. Ithaca,
NY: Cornell Univ. Press
Hughes JD. 1975. Ecology in Ancient Civilizations. Albuquerque: Univ. N.M. Press
Hughes JD, Thirgood JV. 1982. Deforestation in ancient Greece and Rome: a cause of collapse.
Ecologist 12:196–208
Jacobsen T, Adams RM. 1958. Salt and silt in ancient Mesopotamian agriculture. Science
128:1251–58
Jevons WS. 1866. The Coal Question. London: Macmillan. 2nd ed.
Kirch PV. 1984. The Evolution of the Polynesian Chiefdoms. Cambridge, UK: Cambridge Univ.
Press
Kirch PV. 2000. On the Road of the Winds: An Archaeological History of the Pacific Islands. Berkeley:
Univ. Calif. Press
Lopinot N, Woods WI. 1993. Wood overexploitation and the collapse of Cahokia. In Foraging
and Farming in the Eastern Woodlands, ed. CM Scarry, pp. 206–31. Gainesville: Univ. Press
Fla.
Lowe JWG. 1985. The Dynamics of Apocalypse: A Systems Simulation of the Classic Maya Collapse.
Albuquerque: Univ. N.M. Press
McCoy PC. 1976. Easter Island settlement patterns in the late prehistoric and protohistoric
periods. Int. Fund Monum., Easter Island Comm., Bull. 5. New York: Int. Fund Monum
McCoy PC. 1979. Easter Island. In The Prehistory of Polynesia, ed. JD Jennings, pp. 135–66.
Cambridge, MA: Harvard Univ. Press
McGovern TH. 1994. Management for extinction in Norse Greenland. In Historical Ecology:
Cultural Knowledge and Changing Landscapes, ed. CL Crumley, pp. 127–54. Santa Fe, NM:
Sch. Am. Res. Press
Malthus TR. 1798. An Essay on the Principle of Population. London: J. Johnson

www.annualreviews.org • Archaeology of Overshoot and Collapse 73


ANRV287-AN35-04 ARI 13 August 2006 6:41

Mann D, Chase J, Edwards J, Beck W, Reanier R, Mass M. 2003. Prehistoric destruction of the
primeval soils and vegetation of Rapa Nui (Isla de Pascua, Easter Island). In Easter Island:
Scientific Exploration Into the World’s Environmental Problems in Microcosm, ed. J Loret, JT
Tanacredi, pp. 133–53. New York: Kluwer Acad./Plenum
Meggers BJ. 1954. Environmental limitation on the development of culture. Am. Anthropol.
56:801–24
Milner G. 1990. The late prehistoric Cahokia cultural system of the Mississippi River Valley:
foundations, florescence, and fragmentation. J. World Prehist. 4:1–43
Mulloy W. 1970. A speculative reconstruction of techniques of carving, transporting and erect-
ing Easter Island statues. Arch. Phys. Anthropol. Oceania 5:1–23
Phillips DA. 1979. The growth and decline of states in Mesoamerica. J. Steward Anthropol. Soc.
10:137–59
Ponting C. 1991. A Green History of the World. New York: St. Martin’s
Powell MA. 1985. Salt, seed, and yields in Sumerian agriculture. A critique of the theory of
Annu. Rev. Anthropol. 2006.35:59-74. Downloaded from www.annualreviews.org

progressive salinization. Zeitschrift für Assyriologie und Vorderasiatische Archaeologie 75:7–38


by JOHNS HOPKINS UNIVERSITY on 10/21/10. For personal use only.

Rees WE. 2004. Carrying capacity and sustainability: waking Malthus’ ghost. In Introduction to
Sustainable Development, ed. DVJ Bell. In Encyclopedia of Life Support Systems. Oxford, UK:
EOLSS. http://www.eolss.net
Rescher N. 1978. Scientific Progress: A Philosophical Essay on the Economics of Research in Natural
Science. Pittsburgh: Univ. Pitts. Press
Sanders WT. 1962. Cultural ecology of the Maya Lowlands, part I. Estudios de Cultura Maya
2:79–121
Sanders WT. 1963. Cultural ecology of the Maya Lowlands, part II. Estudios de Cultura Maya
3:203–41
Sheets PD. 1999. The effects of explosive volcanism on ancient egalitarian, ranked, and strati-
fied societies in Middle America. In The Angry Earth: Disaster in Anthropological Perspective,
ed. A. Oliver-Smith, SM Hoffman, pp. 36–58. New York: Routledge
Steadman DW. 1989. Extinction of birds in eastern Polynesia: a review of the record, and
comparison with other Pacific island groups. J. Archaeol. Sci. 16:177–205
Steadman DW. 1995. Prehistoric extinction of Pacific island birds: biodiversity meets Zooar-
chaeology. Science 267:1123–31
Steadman DW, Casanova PV, Ferrando CC. 1994. Stratigraphy, chronology, and cultural
context of an early faunal assemblage, Easter Island. Asian Perspect. 33:79–96
Tainter JA. 1988. The Collapse of Complex Societies. Cambridge, UK: Cambridge Univ. Press.
Tainter JA. 2000a. Global change, history, and sustainability. In The Way the Wind Blows:
Climate, History, and Human Action, ed. RJ McIntosh, JA Tainter, SK McIntosh, pp. 331–
56. New York: Columbia Univ. Press
Tainter JA. 2000b. Problem solving: complexity, history, sustainability. Popul. Environ. 22:3–41
Turner BL. 1974. Prehistoric intensive agriculture in the Mayan Lowlands. Science 185:118–24
van Andel TH, Zangger E, Demitrack A. 1990. Land use and soil erosion in prehistoric and
historical Greece. J. Field Archaeol. 17:379–96
Van Tilburg J. 1994. Easter Island: Archaeology, Ecology and Culture. London: Br. Mus. Press
Wallace R. 1761. Various Prospects of Mankind, Nature, and Providence. London: A. Millar

74 Tainter