Journal of World-Systems Research, VI, 1, Spring 2000


67

Rise, Fall, and Semiperipheral Development 
in the Andean World-System

journal of world-systems research, vol vi, 1, spring 2000, 67-98
http://jwsr.ucr.edu/
issn 1076-156x 
© 2000 Darrell La Lone 

This paper examines how different power strategies have been played over the course of the three-thousand year Andean civilization. In 
broad terms, the power strategies center on ideological, economic, and mili-
tary power (Mann 1986; Earle 1997). The case materials show that what we 
call “states” are formations that succeed, albeit temporarily, in combining strat-
egies of ideological, economic, and military power. Nonetheless, following 
Randall Collins (1981:71), I would argue that coercive power is the sine qua 
non for state development. If we start from the power to coerce as a foun-
dation, varying strategies of manipulating economic, ideological, and mili-
tary power will produce a number of different pathways to varying forms of 
organization. Some may be more centralized than others, some may be less 
steeply stratifi ed than others, some may have more specialized institutions 
than others, some may look more like environmental management systems, 
and others may look more like well-organized predatory protection rackets. 
Those which were less successful in manipulating these strategies were more 
vulnerable to collapse, opening the way for new formations to emerge. As 
world-systems approaches would predict, this has often happened on the 
peripheries of former core formations.

In common with many discussions of precapitalist world-systems the 
central problem I address is that of cycles of rise and fall in Andean civili-
zations. Such discussion of rise and fall is now well-established in world-
systems theory (Chase-Dunn and Hall 1997), but it has an even longer 
ancestry in archaeology and anthropology, where we have a rich literature 
on state formation and collapse (Yoffee 1995; Haas, Pozorski and Pozorski 
1987; Wright and Johnson 1975; Jones and Kautz 1981; Culbert 1973; 
Marcus 1989; Millon 1988; Toland 1987; Carneiro 1970; Service 1975; 
Tainter 1988; Yoffee and Cowgill 1988). 

Darrell La Lone

http://jwsr.ucr.edu/


Darrell La Lone68 Rise, Fall, and Semiperipheral Development in the Andean World-System 69

Before moving toward some of the particulars of Andean Civilization, 
it may fi rst be useful to preview some of the larger lessons and warnings. 
First, our search for origins of “the” state has led us through a number of 
hypotheses. One of the most promising of these was the hydraulic hypoth-
esis (Steward-Wittfogel), which made immensely great sense as logic and as 
an interpretation of Chinese history in particular (Wittfogel 1957). How-
ever, the hydraulic hypothesis has not fared well in empirical archaeological 
explorations of state origins (see, for example, (Adams 1966)). This is not 
to say that we should trash everything gained from it, but rather that it does 
not seem to be an accurate account of the early phases in state development. 
Nonetheless, it remains useful as a way of understanding some of the power 
strategies available to rulers. 

Its most promising successor seemed to be Carneiro’s model for popula-
tion expansion and internecine confl ict in circumscribed spaces. As was the 
case with the hydraulic hypothesis, Carneiro’s model has considerable logi-
cal power. I comment on it in part because it seems to have been developed 
as a way of explaining state formation on the Peruvian coast. What may 
not be immediately apparent, though, is that it was not rooted in actual 
archaeological data from that region. When we examine the archaeological 
data from the Peruvian coast, we may well come back with reasons to ques-
tion and challenge our confi dence in the circumscription model. This is one 
example of how archaeological data may well be relevant to understanding 
long term change, but the lesson to be drawn is, perhaps, “Not So Fast”! 
This then leads us toward further exploration of a more complex relation-
ship between warfare and state formation.

A further issue in state formation and collapse studies in archaeology 
underscores the appropriateness of world-systems analysis of rise and fall. A 
growing consensus in archaeology is that states are far more fragile than we 
might have thought after so much discussion of states as centralized man-
agement systems. The attainment of signifi cant degrees of centralization is 
gained at considerable cost and against resistance from many sides. 

Second, portrayal of states as management systems has been consistent 
with an emphasis on the connections between population expansion, inten-
sifi cation of production, environmental degradation and potential collapse, 
and pressures toward innovation in productive technologies that may sustain 
continuing population growth. Despite the apparent success of ecological 

systems approaches in terms of the elegance of their logic and their fruit-
fulness in generating research questions, archaeological evidence suggests 
a need to reconsider. For example, intensifi cation of production may have 
less to do with the subsistence needs of a growing population of common-
ers than it does with the demands for prestige and wealth items from elites. 
Population expansion may not necessarily be a precursor to warfare but a 
consequence of warfare as a result of pressure to produce more orders on 
the part of elites, and a felt need on the part of commoners to replace sons. 
Conceptualizing states as management systems or problem solving organi-
zations masks the inequalities characteristic especially in archaic states.

Third, my discussion of cycles of rise and fall in Andean polities draws 
upon insights from world-systems analysis on the importance of semipe-
ripheries, and an important part of the overall dynamic is best understood 
from enduring principles of geopolitics. However, I would also acknowledge 
extensive independent discussions among archaeologists of cycles of rise 
and collapse around the world (see, for example, (Marcus 1998)). But, more 
specifi cally in regard to the Andean evidence, I would underscore another 
reason to be cautious. Many of our discussions of strategies in statecraft are 
based on the premise that states control contiguous territories. However, 
it has long been the case that Andean polities attempted to organize non-
contiguous territories. 

Ups and Downs in the Andes

Fifty years ago in his seminal paper on “Cultural Causality and Law,” 
Julian Steward noted cycles in civilizations of the Americas which fl uc-
tuated between small regional polities and much larger social formations 
which we have come to call “Horizon Styles” (Steward 1949). This paper 
should be particularly intriguing for world-systems scholars because of its 
efforts to outline similar patterns of rise and fall in ancient civilizations 
around the world. More recently, Gordon Willey (1991) has revisited this 
theme in his paper on “Horizontal Integration and Regional Diversity: 
an Alternating Process in the Rise of Civilizations.” In this essay, Willey 
observes that both in Mesoamerica and in the Andes we fi nd alternations 
between horizontal integration and regional diversity. He offers as a thesis



Darrell La Lone70 Rise, Fall, and Semiperipheral Development in the Andean World-System 71

…that such an alternation of horizontal integration and regional diversity is 
a vital, perhaps even a necessary, process in the rise to that social, cultural, 
ideological, economic, and political complexity which, taken as a whole, we 
refer to as civilization” (Willey:197). 

However, Willey’s discussion of his thesis is largely descriptive, and 
invites other scholars to fi ll in more of the causal explanation for why such 
alternations occur and why they are such vital processes in civilizations. 

The following pages offers a brief outline of these alternating cycles 
in Andean civilization. As the discussion is largely for the benefi t of non-
Andeanists, I will to avoid some of the highly specifi c and arcane archeologi-
cal debates between Andeanists. Instead I will attempt to point out some of 
the more intriguing points of interest for the world-systems perspective.

the late pre-ceramic (3000–1800 b.c.e.)

The late pre-ceramic is our starting point for the origins of complex 
societies in the Andes. During this period along the Peruvian coast we 

fi nd our fi rst evidence of monumental architecture. Some of these fi rst 
monumental structures date back to 2500 B.C. E., so that they are roughly 
contemporaneous with the building of the pyramids in Egypt. One of the 
largest pre-ceramic sites is Aspero, in the Supe Valley on the north coast of 
Peru. Starting from about 2800 B.C.E., people at the Aspero site began to 
construct a number of large mounds, the largest of which, known as Huaca 
de los Idolos and Huaca de los Sacrifi cios measure some 30 by 50 meters. 
Walls were made of large blocks of basaltic rock, and a number of rooms 
were constructed on the summits (Morris and Von Hagen 1993:19). El 
Paraiso, the largest pre-ceramic site, in the Chillon Valley, dates to about 
1800 B.C. E.. Here platform mounds and stone buildings were constructed 
out of more than 100,000 tons of quarried stone (Morris and Von Hagen:
39).

What we fi nd particularly interesting in these data is fi rst, that these 
very substantial monumental constructions are not associated with sub-
stantial residential debris. Second, although subsistence was based on some 
farming, there is no evidence for irrigation systems at this time. Third, 
although construction on this scale demonstrates signifi cant labor mobi-
lization, there is no evidence that labor mobilization was undertaken in 
order to intensify production in response to population increase. Fourth, 
although we have substantial labor mobilization, we have little evidence for 
social stratifi cation in this era. And, fi nally, we are led to conclude that these 
large structures seem best interpreted as dedicated to public ceremonial pur-
poses. 

Some question the usefulness not simply of archaeological data for 
understanding long-term change, but also of the value of archaeological 
thinking. An old archaeological joke holds that when we don’t know what 
something is, we simply say it must be ceremonial. However, what we fi nd 
from excavations at Aspero is revealing evidence about construction tech-
niques indicative of labor mobilization for public purposes. Buildings at 
Aspero were constructed with the use of bagged fi ll, a characteristic tech-
nique in sites of this era. Rubble was packed into “large open mesh satchels 
called shicra, made of coarse sturdy reeds and capable of holding as much as 
26 kg of stone” (Moseley 1992:110). This promoted uniform units of con-
struction in two senses: (1) structures were built from a known number of 
physical units, and (2) labor contributions could also be calculated in terms 

Table 1 – A Synoptic Chronology for the Central Andes
Dates Period Core Politics World-Systems Interactions
1438-1532 Late Horizon Inca Pan-Andean Information,

prestige-goods, political-military,
bulk-goods and labor networks

1000-1438 Late Intermediate Chimu (North
Coast), Highland
"hill-fort
chiefdoms" (e.g.
Wanka, Chanka,
Quechua)

Regional Information, prestige-
goods, political-military, and
bulk-goods networks

600-1000 Middle Horizon Wari, Tiwanaku Pan-Highland (and some coastal)
information, prestige-goods,
political-military, bulk-goods
and labor networks

100-600 Early Intermediate Moche, Nazca Multi-valley coastal (Moche)
information, prestige-goods,
political-military, and bulk-
goods networks

800-100
CE

Early Horizon Chavin Highland-based Pan-Andean
information and prestige-goods
networks

1800-800 Initial Coastal Valleys Coastal intra-valley information
and bulk-goods networks

3000-1800
BCE

Late Preceramic Coastal Valleys Coastal intra-valley information
networks



Darrell La Lone72 Rise, Fall, and Semiperipheral Development in the Andean World-System 73

of such standard units. Although this and similar construction techniques 
characterize subsequent Andean public architecture, the bagged fi ll tech-
nique is not found in domestic architecture (Morris and Von Hagen 1993:
38).

The fi rst traces of Andean civilization take the form of ceremonial mon-
uments. The labor force necessary to construct the larger platform mounds 
must in many cases have been much larger than the nearby residential pop-
ulations subsisting on sea resources and small-scale farming. People must 
have gathered from a wide area to pool their labor in the service of con-
structing a monument. And what, we might well ask, is the value of that? 
As Earle comments, “one factor common to all monuments… is their abil-
ity to be experienced simultaneously by large numbers of individuals” (Earle 
1997:155).

As Earle argues, it is such common experiences that shape culture and 
culture change, rather than the ideas invoked mystically in the mentalist con-
cept of culture that has for so long plagued anthropology. Monuments, and 
the common experiences they inspire, are examples of materialization. Again, 
as Earle explains,

Materialization is the transformation of ideas, values, stories, myths, and the 
like into physical reality that can take the form of ceremonial events, symbolic 
objects, monuments, and writing. It is the process by which culture is created, 
codifi ed, and contained. Ideas and objects unite and are inseparable; ideas, 
unconnected to the object of world, have no means of being communicated, 
experienced, use, and owned. Ideas must be materialized to become social, to 
become cultural things. Our ideas are private and powerful for ourselves, but 
their materialization brings them into the public arena. (Earle 1997:151)

In the highlands also we fi nd conspicuous Preceramic Period ceremo-
nial sites such as Kotosh, La Galgada, and Huaricoto. Here, as on the coast, 
we fi nd public architecture with numerous ritual chambers and remains of 
offerings. As on the coast, we have no evidence for dense settlement or social 
stratifi cation. 

initial period (1800–800 b.c.e.)

The Initial Period (also known as Early Ceramic) is marked by the 
appearance of ceramics, signifi cant changes in settlement patterns, and the 
development of irrigated agriculture. Settlements moved from the early 
coastal settings into valleys allowing for irrigation agriculture. Pottery and 

cloth production on a large scale suggests craft specialization. For the most 
part, though, it appears that the gains from agricultural intensifi cation were 
invested in bigger and better ceremonial structures. During this period we 
fi nd the development of large U-shaped platform mounds that became 
characteristic along much of the Peruvian coast. Huge U-Shaped temple 
mounds were constructed at sites such as La Florida (in the Rimac Valley), 
Garagay (in the Chillon Valley), Cardal (in the Lurin Valley), Las Haldas 
(south of the Casma Valley) and Moxeke-Pampa de las Llamas (in the 
Casma Valley). At Sechin Alto, the platform mound measures 300 by 250 
meters and rises 44 meters high (Morris and Von Hagen 1993:50). 

The overall picture suggests that agricultural and labor intensifi cation 
was invested in increasingly large ceremonial monuments. We have no evi-
dence for political centralization and only scattered hints of social stratifi ca-
tion suggested by a small number of burial offerings (Pozorski and Pozorski 
1986). Overall, we fi nd little evidence for trade, warfare, or political interac-
tions between coastal valleys. Each valley appears to be relatively self-suffi -
cient, and the only signifi cant changes in settlement pattern are movements 
from the coastline to higher points in the middle and upper valleys. The 
spread of the large U-shaped platform mounds along much of the central 
and north-central coast shows no indication of political, military, or eco-
nomic interactions between valleys. Nor do we have reason to suspect ideo-
logical interactions, since each center gives every appearance of being a local 
cult site.

 How, then, might we account for the spread of a similar architectural 
form along the Peruvian coast? Two reasons come to mind. One, already 
discussed above, is that monuments are important because of “their ability 
to be experienced simultaneously by large numbers of individuals” (Earle 
1997:155). In this way, monuments provide a focus for shared identity link-
ing numerous small settlements scattered throughout each coastal valley. 
A corollary to such identity markers is that they also mark who does not 
belong. A highly visible monument also makes an important signal to out-
siders concerning its makers’ ability to mobilize substantial amounts of 
labor. Again, as Earle puts it,

Monuments convey a simple message of power and wealth (Trigger 1990). 
This elemental message comes across regardless of the viewer’s language, age, 
gender, or cultural affi liation (Earle 1997:156)



Darrell La Lone74 Rise, Fall, and Semiperipheral Development in the Andean World-System 75

However, if the monuments of this era convey messages of power and 
wealth, we should not assume that power and wealth were politically or 
economically centralized. Apart from public monuments, we have none of 
the archaeological markers of state societies, and little evidence for the emer-
gence of political or economic elites.

early horizon (800 b.c.e.–100 c.e.)

Toward the beginning of the fi rst millennium B.C.E., many of the 
coastal settlements seem to have collapsed. Between 900 and 700 B.C.E., 
construction was halted at the U-shaped complexes in the Lurin Valley and 
in many of the other ceremonial complexes of the central coast as well as 
along the north-central coast (Burger 1995:184-190). Although our current 
evidence concerning the causes of the collapse allows little more than specu-
lation, I am inclined to the argument that demand both for subsistence and 
ceremonial construction pushed agricultural intensifi cation toward increas-
ingly higher yields, but coastal production was highly vulnerable to catas-
trophes (see Wilson 1999:352ff; Wilson 1981). In Peru catastrophes tend 
to be customary, taking such dramatically predictable forms as earthquakes, 
volcanic eruptions, landslides, massive fl oods, and of course catastrophes we 
term El Niño events. The Early Horizon marks a shift in focus of Andean 
civilization from the coast to the highlands, and from localized ceremonial 
centers to a pan-Andean cult. And it appears that this fi rst major transfor-
mation in the development of Andean civilization emerges from a semipe-
ripheral region. 

The Chavin cult is named for its north-central highland type site, Chavin 
de Huantar, located at an altitude of over 3000 meters. This site extends for 
some 42 hectares, including the 5 hectare temple sector (Morris and Von 
Hagen 1993:56). Chavin seems clearly to have been a ceremonial center, 
though on a grander scale than its coastal precursors. A resident population 
of several thousand people built and served the center, but it seems clear that 
the center was not built to serve them. Unlike the coastal platform mounds 
which could “be experienced simultaneously by large numbers of individu-
als” (Earle 1997:155), Chavin offered a different form of cult experience. The 
Old Temple is a maze of rooms and passageways allowing access only to 
limited numbers of people at a time. Deep within its galleries are monolithic 
representations of the principal deities. Access to the cult images, and pre-

sumably rituals, was exclusive rather than inclusive. It gives every appearance 
of a location for exclusive rituals, initiations, and oracular consultations. Yet 
this very closed cult became the fi rst to span almost the entire length of the 
Andes and to reach down to the coast as well. 

Chavin would certainly seem, nestled in its Andean heights, peripheral 
to the coastal valley communities. However, I believe the term semiperiphery 
is appropriate on several grounds. My usage of the term follows the broad 
outline of the terms offered in Chase-Dunn and Hall’s Rise and Demise:

1. A semiperipheral region may be one that mixes both core and peripheral 
forms of organization.

2. A semiperipheral region may be spatially located between core and peripheral 
regions.

3. A semiperipheral region may be spatially located between two or more com-
peting core regions.

4. Mediating activities between core and peripheral areas may be carried out in 
semiperipheral regions.

5. A semiperipheral area may be one in which institutional features are in some 
ways intermediate between those forms found in core and periphery (Chase-
Dunn and Hall 1997:37).

The structures at Chavin, including the U-shaped design of the Old 
Temple, would certainly seem familiar to any coastal visitor. However, the 
images and iconography of the temples refl ect Amazonian tropical forest 
animals and stylistic motifs (defi nition 1). Chavin de Huantar is located 
high in the Andes between the coastal core region to the west and the 
peripheral tropical forest regions to the east (defi nition 2). The rituals per-
formed at Chavin and then spread into core regions may well have been of 
tropical forest inspiration if not origin and very likely included use of tropi-
cal forest hallucinogens (defi nition 4).

Soon the images and stylistic conventions seen at Chavin de Huantar 
would spread throughout much of ancient Peru. Felines, caymans, the 
Chavin Staff God, and characteristic Chavin stylistic conventions were 
represented on ceramics, textiles, stone, and goldwork found in sites extend-
ing from the northern highlands to the south central coast. Our belief that 
this was a temple cult and not a military/political/economic movement is 
founded on the fact that the characteristic Chavin iconography is devoid 
of explicit references to political, military, or economic exploits of any sort.  
Chavin presence is expressed in ceramic and textile designs, but not in 



Darrell La Lone76 Rise, Fall, and Semiperipheral Development in the Andean World-System 77

alteration of settlement patterns or any other indicator of invasion from an 
expansionistic polity. Andeanists of all theoretical persuasions share con-
sensus that Chavin was a cult.

The expansion of the cult from the highlands to the coast does not take 
the form of monumental structures in Chavin style, but instead its presence 
is most often shown in portable objects representing Chavin iconography, 
and 

these objects, judging from their form and decoration, appear to have func-
tioned in ceremonial contexts as ritual paraphernalia or as part of the 
costumes of the offi ciants. Several classes of portable item are believed 
to be involved in the preparation and ingestion of hallucinogens (Burger 
1988:124).

Apart from their association with a new iconographic tradition, many of 
these objects also demonstrate technological innovation. Both textiles and 
goldwork were produced in what some have characterized as “revolutionary” 
techniques. Textile production in the Chavin horizon underwent an

astonishing range of new techniques and materials fi rst utilized as carriers of 
the Chavin designs. In fact, a set of inventions was promulgated whose infl u-
ence obliterated most of the older textile traditions in the coastal areas, and 
which became the foundation for later Peruvian textile evolution (Conklin 
1978:1 as cited in Burger 1988:129). 

Goldworking also took forms never previously seen, and as the leading 
scholar of Andean metallurgy explains the new technologies,

It may be that in the Central Andes gold was used earlier than any other 
metal and that its prominence in the archaeological record during the period 
of spread and infl uence of the Chavin cult…refl ects a religious and ceremo-
nial bias of the cult for that metal…My feeling is that gold may have had 
special symbolic signifi cance for the cult and that certain religious values 
ordoctrine were expressed through its use (Lechtman 1980 as cited in Burger 
1988:130).

The Early Horizon then represents major transformations in the shape 
of Andean civilization. In the Preceramic and Initial periods, settlements 
and monumental ceremonial structures developed primarily on the coast. 
Little interaction was evident between valleys, but some exchange occurred 
between coastal valley settlements and their adjacent highlands. Thus, it 
seems appropriate to describe highland regions as semiperipheral. Then, 
following the collapse of the coastal communities at the end of the Initial 

Period, the Chavin cult took off from its semiperipheral base to become 
the fi rst nearly Pan-Andean core. In what was semiperiphery, we see revolu-
tionary innovations in technology and ideology, and these innovations soon 
spread over much of the Peruvian Andes.

Again, I would emphasize that this expansion did not follow from the 
expansionistic moves of a centralized polity, nor did it seem to be driven 
by economic forces. If, as Andeanists argue, it was a cult, then we might 
ask why it had such wide appeal. Why would peoples of the coast or of the 
highlands trade their local cults for a foreign cult? Materialists have perhaps 
too often been misled by overly stubborn insistence on treating ideology as 
“mere” superstructure. As Earle points out,

Ideology is a system of beliefs and ideas presented publicly in ceremonies 
and other occasions. It is created and manipulated strategically by social seg-
ments, most importantly the ruling elite, to establish and maintain positions 
of social power. (Earle 1997:149)

Whatever led to the collapse of so many coastal communities, it seems 
reasonable to suspect that the local cults were unsuccessful in forestalling 
catastrophe. And, 

…if the legitimacy of a center’s cosmology was called into question, so too 
would the non-coercive authority of the leaders associated with this sacred 
knowledge. Ultimately, this would have undermined the very basis of the 
social relations that underlay the productive systems, and the functioning of 
these societies as viable units (Burger 1995:189). 

Under these circumstances, perhaps a wonderful new cult might help 
ease people through times of crisis. However, we can fi nd a better reason for 
the appeal of a new foreign cult, and that reason connects directly with the 
attempts of social segments to manipulate ideology to establish and main-
tain positions of social power. Over time, the ceremonies and rituals of the 
local cult would no longer serve communities in an egalitarian manner, if 
they ever did. If nothing else, with the establishment of new settlements 
at a greater distance from the ceremonial monument, people would be less 
immediately connected. Inevitably, some people would become disaffected, 
and it is precisely to such increasingly marginalized groups that a new for-
eign cult would appeal. As Elizabeth Colson notes,

The cult with universal claims then serves as a further court of appeals 
when a local cult does not satisfy its adherents…This was true in ancient 



Darrell La Lone78 Rise, Fall, and Semiperipheral Development in the Andean World-System 79

Greece where each small city state had its own cult, associated with its fate as 
ahuman community, but delegations went to consult the Oracle of Delphi,or 
perhaps Dodona, which served the entire classical world for manycenturies. 
(Colson 1977:119 as quoted in Burger 1988:114). 

Colson’s reference to oracles such as Delphi and Dodona is not in the 
least far-fetched in regard to Andean civilization. One of the most ancient 
pilgrimage shrines in Andean civilization was the Oracle of Pachacamac, 
near Lima. Since it was still functioning at the time of the Spanish Conquest 
of Peru, we have ample ethnohistorical material on this Andean oracle. It is 
entirely reasonable to see Pachacamac as a parallel to Chavin.

The Chavin cult then offered a new basis for making claims to power. 
We may speculate that people who had made a pilgrimage to Chavin 
claimed special status. However, I think it is particularly signifi cant that so 
many Chavin objects are portable, and that the major technological innova-
tions of Chavin were in textiles and goldworking. This allowed the move-
ment of high value/low bulk items over considerable distances, and the 
display of Chavin status items may have been an integral part of new claims 
to status and power.

It is, in fact, during the Early Horizon that we begin to see evidence for 
social stratifi cation and wealth differentials, and

Signifi cantly, many of the portable examples of Chavin religious art have been 
discovered in the rich burials of the upper social stratum. It is likely that these 
individuals used the ideology of the Chavin cult to legitimize and naturalize 
their wealth and positions of authority; indeed, the iconography of the cult 
frequently appears on jewelry and garments that these people wore. (Burger 
1995:203) 

Where this leaves us, then, is that the rise of a formerly semiperipheral 
area to the central core of a pan-Andean cult produced signifi cant transfor-
mations in Andean civilization. New technologies did not increase subsis-
tence production, but were directed toward production of prestige items. 
Such prestige items were important status markers and promoted compe-
tition among rival factions. This ideologically rooted movement initiated 
the fi rst extensive long-distance exchange connections in the Andes and 
intensifi ed the earliest movements toward control of political and economic 
power.

the early intermediate period (100–750 c.e.)

By the end of the Early Horizon, Chavin de Huantar was no longer an 
important ritual center. Following its “collapse,” the focus of Andean civiliza-
tion shifted again to the coast. Although reasons for Chavin’s decline may be 
pondered at length, I see the transition from Early Horizon to Early Inter-
mediate as an unsurprising instance of pulsation in world-systems. Prior to 
Chavin’s rise, coastal communities showed little stratifi cation and appeared 
to be surprisingly egalitarian in view of their ability to mobilize substan-
tial amounts of labor for community projects such as the large U-shaped 
mounds. These were in fact community projects and not directed toward pri-
vate or elite interests. The rise of the Chavin cult and its impact on securing 
status markers heated up incipient factional rivalries. However, claims to 
power that rely primarily on status markers and prestige displays are fragile 
when not backed up by military and economic power strategies.

 The end of the Early Horizon and the beginning of the Early Interme-
diate mark our fi rst evidence for consolidation of political, military, and eco-
nomic power as well as ideological power. It is during this period that we 
see the emergence of complex chiefdoms and archaic states. These polities 
were localized in the coastal valleys, until the Moche state established the 
fi rst multi-valley (conquest) state in the Andes. In any event, Chavin opened 
the way toward the development of strong local and regional polities along 
the coast, and thereby returned the highlands to semiperipheral status. 

Chavin and Chavin-inspired coastal communities had not established 
coercive force over their regional populations. Relying upon status alone is 
not suffi cient to compel others to compliance. However, during the late Early 
Horizon and in the Early Intermediate, there exists evidence of increasing 
inequality in power and wealth in several settings along the Peruvian coast. 
In fact, the Peruvian coast became the model for one of our most infl uen-
tial accounts of origins of “the” state—Robert Carneiro’s circumscription 
model.

Carneiro’s model begins by hypothesizing a Peruvian coastal valley 
populated by small, dispersed, and autonomous villages. As population 
grew, villagers adapted by intensifying production and by fi ssioning into 
new locations. In time, the landscape became fi lled, yet population contin-
ued to grow. Since the desert boxed in fertile lands, villages then began to 
fi ght their neighbors over access to land. Formerly egalitarian and autono-



Darrell La Lone80 Rise, Fall, and Semiperipheral Development in the Andean World-System 81

mous villages would then become subordinate to the victors and chiefdoms 
would form. As population continued to grow along with military prowess, 
confl ict between neighboring chiefdoms would persist until a single polity 
dominated the entire valley. States then emerged from such internecine 
confl ict, and further growth would lead to warfare with neighboring valleys, 
ultimately leading to a multivalley state (Carneiro 1970). 

Perhaps one reason the model seems so clear and elegant is that it was 
constructed in deductive style, and does not in fact come from actual archae-
ological evidence:

The reconstruction of these events that I present is admittedly inferential, but 
I think it is consistent with the archeological evidence (Carneiro 1970 :735)

A test of Carneiro’s model would require thorough settlement survey 
archaeology of Peruvian valleys. Such survey work has in fact been carried 
out, and David Wilson has presented his evaluation of Carneiro’s model in 
a number of publications (Wilson 1983; 1987; 1988; 1989; 1997; 1999).

Although Wilson’s work confi rms the importance of warfare in the 
development of complex society in the Santa Valley, it disputes the model’s 
implications for settlement pattern and the development of valley polities. 
He found that the earliest agricultural villages in the Santa were not widely 
and thinly distributed, but rather grouped in four major aggregations, sepa-
rated from one another by several kilometers, in the upper and middle parts 
of the valley. Each cluster appeared to be autonomous. So far, this would not 
seem a major deviation from Carneiro’s model, so we might then expect to 
fi nd warfare between sites or between clusters. 

However, sites within each cluster are so immediately adjacent that war-
fare between them would be improbable. Between-cluster warfare would 
satisfy Carneiro’s model of internecine confl ict, but this too seems improb-
able in view of the data. First, clusters differ so greatly in population that his 
Cluster Two should have disposed of rivals rapidly. Second, both Cluster 
Two and Cluster One did not have suffi cient irrigable land to support their 
populations, and each needed access to valley fl oor land not only between 
clusters Two and Three, but also farther downstream. And, as Wilson 
argues, “this clearly would have required cooperation among the clusters, not 
just among the sites that constituted them” (Wilson 1999:363). Wilson then 
concludes that if warfare was occurring during the Cayhuamarca Period 
(1000-350 B.C.E.), and the presence of a number of fortifi ed sites suggest 

it was, then it was most likely not internecine warfare, but rather warfare 
between valleys. 

On the Andean coast, I would fi nd a model of inter-valley rather than 
internecine warfare plausible also because of a persistent tradition of appar-
ently co-operative connections within valleys, as evidenced by the early 
construction of monumental works in the absence of any indication of 
hierarchy, and by the later tradition of ethnic groups defi ning themselves in 
terms of valleys. 

The appearance of fortifi cations along with increasing valley-wide inte-
gration of settlements is one indication that warfare from this point forward 
will open the way toward power strategies built on coercion rather than 
prestige. This development reaches its culmination in this period with the 
emergence of what would seem to be the fi rst indisputable state society in 
the Andes—the Moche state.

The most conspicuous site of the Moche Valley is at Cerro Blanco, 
where we fi nd the gigantic adobe pyramid platforms known as the Huaca 
del Sol and the Huaca de la Luna. The Huaca del Sol measures 340 by 160 
meters at its base and rises 40 meters high. Between the two huge platform 
mounds, recent excavations have shown dense residential rubble extending 
over at least one square kilometer. The summit of the Huaca del Sol also 
has extensive occupation debris, and a number of elite burials have also been 
found on lower terraces. Smaller amounts of occupation debris were found 
on Huaca de la Luna, but it was the site of even richer tombs than Huaca 
del Sol. Platforms and courts surrounded by high walls on its summit, and 
the remains of sacrifi ced individuals also indicate that it may have been the 
site of important rituals as well as elite residences (Wilson 1999:386-387).

With the Moche state, we are able to make more than simply good infer-
ences about social and political organization based on settlement pattern 
data. Moche ceramics offer extraordinary portrayals of every aspect of life: 
sexual positions craft workshops, functionaries at work, warfare, sacrifi cial 
rituals, deities. Moche iconography is particularly rich and explicit in its 
portrayal of warriors, battle scenes, the taking of captives, and the presenta-
tion of prisoners before offi cials or for sacrifi ce. We are left with no doubt 
whatsoever that Moche elites commanded formidable coercive power in the 
form of military force.

Economic power strategies are suggested in the extensive irrigation 



Darrell La Lone82 Rise, Fall, and Semiperipheral Development in the Andean World-System 83

systems that may well have been extended in this time. It is, however, dif-
fi cult for us to be sure of the dating of the irrigation works, and we should 
exercise caution also by avoiding the assumption that such works required 
centralized management. But another instance of economic power might 
be seen in the massive labor mobilization required to construct the gigantic 
pyramids. Alfred Louis Kroeber noticed back in the 1920’s that the Huaca 
del Sol was constructed by placing the adobe bricks into vertical columns, 
and he suggested that different communities were responsible for a part of 
the construction, “each contingent of a community building its own wall or 
column” (Kroeber 1930:61 as quoted in Wilson 1999:387).  Such a con-
struction technique follows, albeit on a vaster scale, the shicra technique 
noted in structures of the Early Preceramic.  

 Economic power strategies may be seen even more vividly in Moche 
crafts. As Wilson notes, “the superb quality of Moche weavings, pottery ves-
sels, and metallurgy strongly suggests that state artisans were engaged full-
time in carrying out their respective craft specialties” (Wilson 1999:391). 
Full-time craft specialization is indicated not only by the quality of the 
products, but also from representations in Moche art. These include por-
trayals of numbers of women weaving on backstrap looms, under the appar-
ent supervision of men in elaborate costumes doing no physical labor and 
represented on a scale signifi cantly larger than the other fi gures in the scene 
(see Wilson 1999:392, fi gure 9.17). This makes a good case for state-spon-
sored craft specialization dedicated to production of wealth or status items. 

As we saw earlier with the rise of the Chavin cult, new technologies 
were developed not to increase subsistence production or to sustain growing 
populations, but rather to provide status markers to support growing social 
inequalities. The elaborate clothing and jewels portrayed in Moche art, and 
found in Moche tombs, are vivid indicators of a highly stratifi ed society.

Ideological power in the Moche state was not limited to display of exotic 
ritual symbols. Perhaps one of the most awe-inspiring rituals portrayed in 
Moche (or anywhere!) was the Moche Sacrifi ce Ceremony. Naked and 
bound men are taken to attendants who slit their throats and collect the 
blood in goblets. The blood-fi lled goblets are then presented for drinking 
to a fi gure identifi ed from his military gear as the Warrior Priest. Perhaps 
we might dismiss such a blood-curdling scene as a nightmarish fantasy or a 
hallucinogenic vision (remember the likelihood of hallucinogens in Andean 

rituals). However, when Walter Alva and his colleagues excavated an intact 
tomb at the Sipan site, they found costumes and ritual equipment match-
ing in detail the portrayals of the Warrior Priest. And, yes, they found the 
goblets as well (Alva 1988; Alva and Donnan 1993). We then have reason to 
believe that the Moche Sacrifi ce Ceremony was not simply art, but a ritual 
that was actually practiced. Certainly such stirring demonstrations of the 
power of elites served not only to impress the locals, but also were no doubt 
performed in valleys the Moche state conquered. 

The combined ideological, economic, and, above all, military power of 
the Moche produced the fi rst multi-valley state in the Andes. At its height, 
the Moche state (some have called it an empire) incorporated ten valleys 
with a combined population estimated on the order of 650,000 persons 
(Wilson 1999:390). Wilson’s survey of the Santa Valley identifi ed a number 
of features indicating Moche conquest:

• pottery molds for mass production of ceramics in the Moche style
• a pyramidal mound with occupation debris on the summit similar to the 

Huaca del Sol, and made with similar construction techniques
• a shift in settlement pattern so that sites were moved away from high, defen-

sible positions down to the valley fl oor, and fortifi cations were entirely elimi-
nated

• irrigation canals were extended into the most productive downvalley areas 
allowing greater agricultural intensifi cation and production of surpluses pre-
sumably for tribute

• a large mural depicting Moche military motifs (Wilson 1999:393-397)

Building upon traditions of valley-wide co-operation established since 
the Early Pre-Ceramic, traditions of manipulating materializations of sacred 
ideology that became even more elaborated in the Chavin cult, and tradi-
tions of inter-valley warfare seen in the late Early Horizon and Early Inter-
mediate, the Moche State combined ideological, economic, and military 
power strategies to build a powerful regional state that represented the core 
of Andean civilization.

The Late Early Horizon and Early Intermediate then culminated with 
the emergence of a polity with coercive power over a vast region of coastal 
Peru. But this too came to an end, and once again we see a shift toward a 
new power emerging from the highland semiperiphery. We may speculate 
about reasons for the Moche collapse. One plausible hypothesis invokes 
the all-too-frequent environmental catastrophes that loom in the Andes. 



Darrell La Lone84 Rise, Fall, and Semiperipheral Development in the Andean World-System 85

Michael Moseley notes evidence of fl ood damage at the Huaca del Sol 
and Huaca de la Luna that may be the result of torrential El Nino rains, 
which were then followed by a thirty-year drought. Ice core samples from 
the Quelccaya ice cap indicate a severe drought from 562 to 594 C.E.. Sand 
then choked fi elds as well as irrigation canals (Moseley 1992:211; Morris 
and Von Hagen 1993:84). 

Invoking an Andean environmental catastrophe may seem too easy a 
solution to the problem of collapse of the coastal state, opening the way for 
the emergence of the fi rst highland conquest state. Speculation aside, I think 
it is suffi cient to suggest that, despite its ability to organize military, eco-
nomic, and ideological power in ways not previously seen in Andean civiliza-
tion, the Moche State remained vulnerable. As argued above, the course of 
Andean civilization led to increasing integration within each of the coastal 
valleys. Labor mobilization for the construction of sometimes gigantic plat-
form mounds does not seem to have been done under the command of 
elites, but rather seems to be done under communitarian organization. 
When warfare began to play a visible role in shaping polities, it did not 
follow the Carneiro model of internecine warfare within valleys, but rather 
was rivalry and confl ict between valleys. Any crisis in Moche State power 
strategies would provide opportunity for the state to fragment into its for-
merly autonomous provinces.

middle horizon (600–1000 c.e.)

The breakdown of the coastal Moche State (out of consideration for 
non-Andeanists, I have omitted discussion of smaller coastal states contem-
porary with Moche along the central and south central coast), opened the 
way yet again for another historic shift. Once again, the highland semipe-
riphery became the core of Andean civilization as large areas of the Andes 
were incorporated into what has been described as the fi rst great empire 
of the Andes, the Wari-Tiwanaku “Empire.” The dual name refers to the 
two central cities of this era: Wari, 2800 meters high near the modern city 
of Ayacucho in Peru’s south central highlands, and Tiwanaku, 3850 meters 
high in the Andes near the southern end of Lake Titicaca in what is now 
Bolivia. 

Tiwanaku was a massive city, extending 2.8 by 1.6 kilometers accord-
ing to shard scatter, and probably much larger at its height, with an overall 

population estimated at 30,000-40,000 (Kolata 1993:174). The central 
structures at its ceremonial core are monumental platforms and enclosures 
incorporating the largest monolithic stone sculpture in the Andes. Stone 
for the sculptures and construction was quarried as far away as the Copaca-
bana Peninsula on Lake Titicaca and transported by reed boat and overland 
skids (Morris and Von Hagen 1993:102). The gigantic stone sculptures and 
gateways and the vast patios and enclosures seem designed to mark vast 
ceremonial spaces.

Although we may be jumping ahead of our story, the ceremonial spaces 
at Tiwanaku may be ancestral to the later Inca use of space to project ideo-
logical power. As Morris and Von Hagen argue,

…intriguing parallels exist between the public architecture of Tiwanaku and 
that of the Inka a few centuries later. Both emphasized gateways that marked 
boundaries between one space and another. The Inka had great concern with 
such boundaries and with the movement of people across them. The ceremo-
nial passage from one space or one patio to another may have marked impor-
tant social and political transformations, and gateways themselves may have 
symbolized and legitimized such changes. It seems evident that the public 
architecture of Tiwanaku is ancestral to that of the Inka, and thus it is reason-
able to suggest some of the same emphasis on ceremonial paths punctuated 
by gateways, shrines, and other stopping points along the way. Like the Inka, 
Tiwanaku rulers had to convince the people of their power and to guide the 
populace through a transformation establishing a new order of elites whose 
instructions were to be obeyed (Morris and Von Hagen 1993:103).

Further evidence of the importance of ceremony at Tiwanaku is that the 
artistic motifs represented on the massive gateways and sculptures became 
so widespread through the Andean highlands as well as the coast that it was 
argued that Tiwanaku must have been an expansionistic state, an empire. 
Alan Kolata, for example, refers to Tiwanaku as a “predatory state” (Kolata 
1993:174) and as an expansionistic empire (Kolata 1993:243ff ). However, 
to justify such an inference, we might ask what evidence we fi nd for Tiwan-
aku impact on distant regions.

Tiwanaku presence in the northern Chilean desert oasis of San Pedro 
de Atacama is indicated by ceramics. The Tiwanaku materials are decorated 
with religious images characteristic of Tiwanaku’s gateways. The materials 
also include drug paraphernalia such as the snuff trays, inhaling tubes, and 
spatulas found at Tiwanaku (Kolata 1993:139).

The Omo site, at Moquegua in southern Peru some 300 kilometers 



Darrell La Lone86 Rise, Fall, and Semiperipheral Development in the Andean World-System 87

away from Tiwanaku, was the fi rst clear instance of Tiwanaku ceremonial 
architecture at any distance from Tiwanaku itself (Goldstein 1989; Kolata 
1993:266ff; Morris and Von Hagen 1993:108; Moseley 1992:224ff ). Here 
Paul Goldstein found a large platform mound 120 meters long, but con-
structed of adobe rather than stone-faced earth. Stepped platforms and a 
square sunken enclosure clearly identify the site as Tiwanaku. According 
to Kolata, “the designers of the Omo ceremonial complex were well familiar 
with the architectural and sculptural tradition of Tiwanaku, and drew their 
inspiration directly from Tiwanaku’s central monuments” (Kolata 1993:
268). A large residential area surrounded the platform, and cemeteries 
further out indicate a substantial population of likely Tiwanaku origin, as 
evidenced by the cranial deformation favored at Tiwanaku (Moseley 1992:
226-227). Kolata characterizes the Omo community as “an agricultural 
colony of altiplano origins” (Kolata 1993:263). Other Tiwanaku sites also 
“…… seem to have been constructed in sparsely occupied areas, and is very 
likely that they were built by peoples from the vicinity of Tiwanaku itself ” 
(Morris and Von Hagen 1993:109). 

However, I would suggest that such sites are not necessarily evidence 
for a Tiwanaku expansionist state, but rather that we are seeing archaeologi-
cal evidence for Tiwanaku colonies. Ethnohistorical records from the Andes 
have shown that a number of Andean highland polities established colonies 
in distant environmental zones. (Murra [1972] is the classic discussion of 
what he called “vertical archipelagoes”).  Through establishing colonies at 
distances as far as fi fteen days walk, the home community was able to obtain 
desired resources without having to resort to trade, and without taking 
political control over intervening territories. Ethnohistorical discussion has 
contested whether such colonies were established late in Andean civiliza-
tion, or whether, as Murra argued, they were in fact quite ancient. The 
archaeological materials that some cite as evidence for an expansionist Tiwa-
naku state seem to me to be more likely to be some of our earliest evidence 
for long-distance colonies.

Evidence for the “aggressive and predatory” character of the Tiwanaku 
polity seems to me rather thin. Iconography at the ceremonial center of 
Tiwanaku does include portrayals of trophy heads and sculpture apparently 
representing decapitation, and this, apparently, is what Kolata refers to as 
“aggressive, martial themes” (1993:134). However, I would ask whether in 

fact we should characterize trophy head themes or headhunting as necessar-
ily “martial.” Headhunting and trophy taking are characteristic of many vil-
lage societies in the Amazon and in the Andes as well. I would be inclined to 
characterize such activities as “raiding,” and I would suggest that headhunt-
ing is not a state activity. War, on the other hand, is a state activity, based on 
quite different logics and logistics. 

Kolata seems to contradict his own characterization of Tiwanaku as an 
aggressive predatory state (1993:243) almost immediately after making it:

But, apart from a few exceptions, the expansion of Tiwanaku was not the 
product of a militaristic grand strategy worked out self-consciously by the 
lords of Tiwanaku and implemented through force-of-arms. Tiwanaku’s 
empire was not of the same order of magnitude as that of the Inca, or the 
classical empires of the Old World, nor was its expansion primarily the work 
of warrior-kings with huge, standing armies ready at an instant to conquer, 
intimidate, and oppress local populations. Military stratagems and conscious 
efforts to conquer territory played a relatively small role in the movement 
of Tiwanaku into foreign territories. Rather, we must think of the state as 
a dynamic mosaic of populations linked at times imperfectly by a mosaic 
of strategic policies and political relationships devised by Tiwanaku’s elite 
interest groups in an almost ad hoc fashion. These political stratagems 
responded to local cultural diversity and specifi c historical contingencies 
(Kolata 1993:243)

In other words, Tiwanaku did not project its power through military 
means, and at numerous points Kolata argues that coercive force was not 
necessary, but rather that “Tiwanaku state religion and imperial ideology 
performed much the same work as military conquest, but at signifi cantly 
lower cost” (1993:245, 246, 252, 82, 85). 

 The absence of military motifs, military architecture, or occupation of 
previously occupied territories with consequent changes in settlement pat-
tern also speaks to me of something other than a conquest state. What it 
does suggest, in my view, is another widespread Andean cult in many ways 
reminiscent of the Early Horizon Chavin cult. Tiwanaku iconography in 
many cases resembles Chavin, and, as Isbell notes, “there can be little doubt 
that Tiahuanacoid iconography belongs to a mythical tradition with its roots 
in late Chavin culture” (Isbell 1988:180). The prominence of ceremonial 
space and even the spread of drug paraphernalia also seem to echo Chavin. 
Again, as Kolata argues, “an essential element of Tiwanaku imperial policy 
was the exportation of state cults” (1993:248). Apart from dissemination of 



Darrell La Lone88 Rise, Fall, and Semiperipheral Development in the Andean World-System 89

images of the Gateway God, Tiwanaku craft items were mass-produced for 
export to colonies as prestige items and status markers associated with the 
religious ideology of Tiwanaku.  

Where Tiwanaku also seems to have built upon post-Chavin core civili-
zation is in a mobilization of economic power beyond anything seen in the 
Early Horizon. Apart from the massive scale of Tiwanaku, so vast that it 
must be characterized as a city rather than simply a ceremonial center, the 
Tiwanaku polity transformed agricultural production in a manner compa-
rable to the great coastal irrigation works. 

Tiwanaku is situated at a higher altitude than any ancient city in the 
world, and its population seems to have been far too large to have been sus-
tained by agricultural techniques typically available at such altitude. How-
ever, Tiwanaku relied on extraordinary agricultural techniques to intensify 
production. Along the shore of Lake Titicaca, 75 square kilometers of raised 
fi elds constructed on platforms from 5 to 15 meters wide and up to 200 
meters long allowed the high yields necessary to sustain a city of Tiwanaku’s 
magnitude (Kolata 1993:183ff presents a detailed accounting of Tiwanaku 
agricultural technology; see also Moseley 1992:229;Morris and Von Hagen 
1993:105).

In sum, I would argue that Tiwanaku built upon the forms of ideological 
power spread under the Chavin cult during its time as a highland core region, 
and it built upon strategies of economic power expanded in the coastal core 
regions. What seems less convincing is that it exercised military power, and 
for that reason I am disinclined to speak of a Tiwanaku “empire.” None-
theless, this discussion of Middle Horizon began with reference to the 
Wari-Tiwanaku “Empire,” and I believe that we do fi nd evidence of the com-
bination of ideological, economic, and military power strategies in Wari.

Wari’s urban core, on the eastern edge of the Ayacucho Valley, sprawled 
over an area of 3-4 square kilometers with a population on the order of 
20,000-30,000 (Moseley 1992:217). Massive stone walls divide Wari into 
sectors containing numerous building compounds ranging in size from 40 
meters to 100 meters in length, and often reaching two or three stories high 
(Isbell 1988:168-169). Such a large and dense population was made possible 
because of a very large irrigation system and extensive terracing of steep 
slopes to increase the available productive lands. By putting steep slopes into 
production for the fi rst time, and by drawing irrigation water from elevated 

springs or streams less vulnerable to irregular rainfall patterns (Moseley 
1992:218), Wari economic power strategies represent yet another example of 
new technologies appearing in formerly semiperipheral regions.

Ideological power was materialized in Wari’s adoption of some of the 
most important religious symbols from Tiwanaku, such as the Gateway 
God (which itself may have been derived from the Chavin Staff God). Wari 
iconography shares so much with Tiwanaku that they were once thought 
to be dual capitals of a single empire. However, one fi nds a number of sig-
nifi cant differences between Tiwanaku and Wari iconography. Tiwanaku 
iconography was depicted largely on monumental stone objects, while Wari 
iconography was represented primarily on portable objects that could be 
taken out to the wider world (Schreiber 1992:280). In contrast to Tiwanaku, 
where vast open ceremonial spaces seemed to invite pilgrims to come into 
the city, Wari architecture is closed and in many instances baffl ingly inacces-
sible. The message appears to be “You need not come to Wari, but Wari will 
come to you!”

Wari’s propensity to extend itself outward is seen in a string of a dozen 
major administrative centers outside the Wari heartland, stretching some 
1000 kilometers from Cuzco to Cajamarca. Three of these Wari provincial 
centers have been excavated suffi ciently to give some picture of how Wari 
outliers worked. The smallest, Jincamocco, covers about four hectares in 
the Carhuarazo Valley, six days away from Wari. Schreiber argues that it 
was a major administrative center, and three smaller centers were built on 
the other side of the valley. As she writes, “The Wari occupation of this 
valley resulted in an alteration of local lifestyles so revolutionary that some 
of those changes still affect life there today” (Schreiber 1992:259).  It appears 
to have been situated along a road extending through Jincamocco to the 
coast (Schreiber 1992:261) (Isbell 1988:185). 

Two larger centers suggest a forceful Wari presence at great distance 
from Wari itself. Viracochapampa, in the highlands near Cajamarca some 
700 kilometers north of Wari, is a vast rectangular compound measuring 
more than half a kilometer on a side. Within the compound walls enclosing 
a space larger than 30 hectares are numerous courts, patios, and structures 
surrounding a central plaza (Moseley 1992:223). It is not clear that Viraco-
chapampa can be taken as evidence of military conquest so far north of Wari, 
but it does demonstrate Wari’s farthest reach northward (Topic1991). 



Darrell La Lone90 Rise, Fall, and Semiperipheral Development in the Andean World-System 91

Pikillaqta, in the Cuzco Valley some 250 kilometers southeast of Wari, 
is the southernmost and largest of the Wari centers. In fact, the extent of 
Pikillaqta is as large as the core of Inca Cuzco (McEwan 1991:93). Within 
this space we fi nd enormous rectangular compounds and vast warrens of 
rooms that appear to have been storage facilities and barracks-like quarters 
(Morris and Von Hagen 1993:116; Isbell 1988:185). Following some of the 
most recent investigations at Pikillaqta, McEwan concludes:

…Pikillacta represents a very sophisticated, complex product of a highly orga-
nized society. It was probably built to house a political and religious elite who 
administered the southern highland part of the Huari empire, and to con-
centrate in one place the economic and political functions together with a 
resident local population in order to centralize this provincial administration 
(McEwan 1991:118).

Although Wari compounds, particularly Pikillaqta, have been charac-
terized as garrisons , the outer walls at Pikillaqta “lack parapets, ascending 
steps, and other features characteristic of fortifi cations” (Isbell 1988:185) 
(for discussion of how we may identify military architecture in the Andes, 
see Topic and Topic 1987). Nonetheless, Wari structures were such formi-
dable strongpoints that it seems implausible that they would be vulnerable 
to anything less than massive force. 

Our most direct view of Wari military organization comes from Cerro 
Baul, a 600 meter high mesa in the Moquegua Valley, where we have earlier 
encountered the Tiwanaku Omo colony situated some 20 kilometers away. 
The route up Cerro Baul was fortifi ed with walls and parapets, and its 
summit was covered with a large complex of Wari buildings (Moseley 
1992:221). The construction of such a formidable fortifi ed site gives us 
some idea of what Wari was capable of when it moved into territory con-
tested by a potentially powerful adversary (Moseley, et al. 1991). Formidable 
fortifi cations in the Moquegua Valley zone where Wari and Tiwanaku seem 
to have come in direct confrontation might expectably differ from what we 
might see in Wari provincial sites between Pikillaqta and Viracochapampa 
where there were no rivals so potentially powerful as Tiwanaku. As Sch-
reiber evaluates the thousand-kilometer string of Wari highland sites and 
evidence for sometimes “revolutionary” changes in local lifestyles, she con-
cludes that

…available data are consistent with the interpretation that Wari was an 
imperialist state. It controlled a very large territory, from Cuzco to Cajamarca 
in the highlands, and possibly the coast as well. It was a territory that was 
ethnically and ecologically diverse. It expanded very rapidly, and lasted per-
haps only 150 years. Above all, there is good evidence of direct Wari political 
control over a series of discrete regions;… (Schreiber 1992 :267) 

Whatever we may make of evidence for Wari as a conquest state, I 
believe that Wari demonstrated its economic power by reorganizing agricul-
tural production in its provincial sites through terracing, irrigation, and 
promotion of maize agriculture. The substantial storage facilities in Wari 
sites also indicate formidable surplus production. Wari also projected ideo-
logical power, though primarily in the form of imagery taken from Tiwanaku. 
Finally, we have some indication that Wari was able to plant sites over a dis-
tance on the order of a thousand kilometers, and this in itself suggests that 
it was able to handle any objections from the locals. But Tiwanaku and Wari 
world collapse.

“After approximately 700 years of growth and colonial expansion, the 
Tiwanaku state disappeared as a regional political force in the south-central 
Andes between A.D. 1000 and 1100.” (Kolata 1993:284) Collapse of Tiwa-
naku seems to be closely correlated with yet another Andean environmen-
tal disaster. Data from the Quelccaya glacier indicate that a major drought 
began around 1000 C.E. and lasted for decades. During the drought, the 
canal-based irrigation in the Moquegua Valley and the groundwater-based 
irrigation agriculture on the Peruvian and Chilean coasts failed (Kolata 
1993:294; Clement and Moseley 1989). The raised ridged fi elds in the 
Tiwanaku heartland were initially less vulnerable to degradation from the 
drought, but in time the lowering water table had a devastating impact 
there as well. Abandonment of many of the Tiwanaku outliers followed col-
lapse of Tiwanaku’s agricultural base, and cities and urban civilization disap-
peared in the Titicaca Basin for another 400 years (Kolata 1993:299). The 
successors to Tiwanaku retreated into sites, often fortifi ed, where access to 
water could be defended.

For the Wari collapse, I would propose another scenario. Wari had been 
unable to move toward the Titicaca Basin because of Tiwanaku’s presence 
to the south. Wari did try to challenge Tiwanaku in the Moquegua Valley, 
as we see at Cerro Baul, but apparently withdrew. Blocked to the south, 
Wari moved as far as it could into the Cuzco Basin and as far south from 



Darrell La Lone92 Rise, Fall, and Semiperipheral Development in the Andean World-System 93

there as Sicuani, where it met the northern most extension of Tiwanaku 
(evidence: Canas and Canchis were formerly Aymara speakers, according to 
Cieza). From the Cuzco region, Wari extended northward along the sierra 
as far north as the Cajamarca region. A one thousand kilometer long stretch 
would be diffi cult to sustain without military superiority and good supply 
lines. Evidence suggests that Wari built a road system and established stor-
age and rest facilities along the route.

However, already overextended, Wari would also be vulnerable to any 
environmental catastrophe. The collapse of Wari, as well as that of Tiwan-
aku, and, for that matter, of all previous Andean civilizations did not mean 
a return to scattered tribal villages. The successor polities continued to 
exercise economic, military, and, presumably, ideological power, albeit on a 
smaller scale. The small successor polities to the Middle Horizon regional 
polities often took the form of hill-fort chiefdoms (see Earle 1997:121), 
and such polities were characteristic in the Andes in the Late Intermediate 
Period.

late intermediate and the emergence of the inca state

The period following the collapse of the Middle Horizon is known as 
the Late Intermediate Period. Once again, the core polities, Wari and Tiwa-
naku, collapsed and what had been their semiperipheral provinces broke 
away and formed what Earle refers to as hill-fort chiefdoms all throughout 
the highlands. The Late Intermediate chiefdoms are conspicuous archaeo-
logically for their hilltop fortifi cations, and the oral traditions carried into 
the ethnohistorical literature reinforce the argument that their power was all 
but exclusively military. Each chiefdom was led by a sinchi, and documents 
from throughout the Andean highlands make it clear that a sinchi was a war-
rior chief.

In the absence of demographic disparities, and given the simplicity of 
the weapons of war, what we would expect to fi nd is a situation of military 
stand-off. However, it is out of this situation and this era that we see the 
rapid rise of what would become the greatest empire of the Americas, and 
certainly one of the most extensive of all ancient empires. Discussion of the 
Late Intermediate and the rise of the Inca state is not to be undertaken as a 
few concluding pages of a hasty survey of Andean civilization. It is a monu-
mental story, and I wouldn’t dare even to outline it here. However, I would 

conclude with a speculative prospectus for the next part of the story.
Toward the end of the Middle Horizon, Wari, unable to move south, 

had moved north and was dangerously overextended. At any point between 
Cuzco and Cajamarca, a belligerent chiefdom may have found opportunity 
to challenge the Wari presence. One such warrior group who dominated 
parts of Ayacucho, Huancavilca, and Andahuaylas were known as the 
Chancas. In the scenario I propose, the Chancas were the ones whose chal-
lenge dismantled the Wari state. Following victory over Wari, the Chancas 
then moved, as had Wari before at Pikillaqta, toward the Cuzco Valley.

This, however, proved a fatal error, or perhaps only bad luck in histori-
cal contingency, since, certainly against Chanca expectations and those of 
many in Cuzco as well, the Cuzco forces defeated the Chancas. The military 
success of Cuzco is the defi ning moment for the emergence of Cuzco as an 
expansionistic state. Following consolidation of the Cuzco region, perhaps 
it should not be surprising that the new Inca state would soon move toward 
the south. Without the opposition of Wari or Tiwanaku states, Cuzco could 
control the entire length of the Vilcanota Valley, and move to take the alti-
plano regions that were formerly the Tiwanaku heartland.

And what were they after? We have no evidence suggesting that the Inca 
restored the vast ancient raised fi elds. What the Titicaca Basin meant for the 
Inca was a vast pastureland and vast herds of camelids. With llamas carry-
ing the loads in supply caravans, the Inca were able to expand upon Wari 
techniques of road extension with waystations as points for storage, labor 
mobilization, military garrison, and state displays.

From the semiperiphery, the Cuzco Valley, then rose the greatest Empire 
of the ancient Americas. Its economic power transformed the Andes both in 
the highlands and on the coast as it reorganized agricultural production, 
mobilized labor throughout its entire reach, and constructed a vast net-
work of state storage facilities. It skillfully manipulated ideological power as 
it imposed new state cults and co-opted local deities. And, most conspicu-
ously, it was able to organize military power on a scale without precedent in 
the Andes. As Earle puts it,

The changes accompanying the rapid Inka expansion …involved weaving a 
tapestry of power that united the powers of military, economy, and ideology. 
Warfare became effective as it became supported by a state fi nancial system 
and institutionalized by a state ideology. The destructive power of military 
might was fashioned into an effective political source of centralizing power in 



Darrell La Lone94 Rise, Fall, and Semiperipheral Development in the Andean World-System 95

Peru only when it was harnessed and controlled through its articulation with 
the other sources of power (Earle 1997:122)   

summary and conclusions

Collapse and succession in Andean world-systems follow patterns and 
rhythms common to other ancient world-systems. Although they were not 
the fi rst to do so, anthropologists and archaeologists have returned to this 
problem over the past fi fty years (e.g. Steward 1949, Willey 1991, Marcus 
1998, Chase-Dunn and Hall 1997).

Particularly powerful hypotheses for regularities in collapse and succes-
sion come from an understanding of geopolitical principles. I have argued 
in this paper that effective exercise of power requires organization of eco-
nomic, ideological, and military power. But, following Collins, “the state con-
sists ultimately of military control over a territory” (Collins 1981:71). To 
elaborate on this rather stark statement of necessary conditions for a suc-
cessful state, Collins draws our attention to “the organizational resources 
that make up the bases of military power, and the territorial confi gurations 
in which this power is exercised” (Collins 1981:71).

Discussion of the organizational resources underlying military power 
is fundamental in current archaeological and anthropological treatments of 
complex society (e.g., Earle 1977). What is missing in many works, how-
ever, is study of the territorial confi gurations in which power is exercised. 
Once states have fi lled landscapes, a number of geopolitical principles come 
into play without regard to cultural particularities. Some of the issues Col-
lins directs us toward include, weponry, heartland size and wealth, march-
land advantages, overxpansion, and technological diffusion (1981:75, 80, 98, 
101)

I believe archaeologists and anthropologists may fi nd richer under-
standing of the organization of power through closer study of what we 
have learned from geopolitics and military history. As in chess, the better 
we understand the territory, the more likely we are to fi nd meaningful pat-
terns and strategies in the middle and end games. Archaeology, apart from 
a fl irtation with central place theory, has not deeply explored the geography 
of power. As Collins argues, our task is to understand “the organizational 
resources that make up the bases of military power, and the territorial con-
fi gurations in which this power is exercised” (Collins 1981:71). Although 

I have agreed with the argument that without military power a polity has 
little chance of success, military power is not the whole story. Leaving the 
last word to Earle,

A strong central political institution depends less on any one source of power 
than on the interrelationships among the power sources and on its ultimate 
grounding in the political economy. (Earle 1997:192)

REFERENCES

Adams, Robert Mc. (1966). The Evolution of Urban Society: Early Mesopotamia and 
Prehispanic Mexico. Chicago: Aldine.

Alva, Walter (1988). “Discovering the New World’s Richest Tomb,” National Geographic 
174(4):510-523.

Alva, Walter, and Christopher B. Donnan (1993). Royal Tombs of Sipan. Los Angeles: 
Fowler Museum of Cultural History.

Blanton, Richard E. (1998). “Beyond Centralization, Steps Toward a Theory of 
Egalitarian Behavior in Archaic States,” in Archaic States. Gary M. Feinman and 
Joyce Marcus, eds. Santa Fe: School of American Research Press.

Burger, Richard L. (1988). “Unity and Heterogeneity Within the Chavin Horizon.” in 
Peruvian Prehistory. Richard W. Keatinge, ed. Pp. 99-144. Cambridge: Cambridge 
University Press.

Burger, Richard L. (1995). Chavin and the Origins of Andean Civilization. New York: 
Thames and Hudson.

Carneiro, Robert (1970). A Theory of the Origin of the State. Science 169:733-738.
Chase-Dunn, Christopher, and Thomas D. Hall (1997). Rise and Demise, Comparing 

World-Systems. Boulder: Westview Press.
Clement, C., and M. Moseley (1989). “Agricultural Dynamics in the Andes,” in Ecology, 

Settlement, and History in the Osmore Drainage, Peru. D. Rice, C. Stanish and P. 
Scarr, eds. Pp. 435-455. Oxford: British Archaeological Reports International 
Series 545(ii).

Collins, Randall (1981). “Long-Term Social Change and the Territorial Power of States.” 
in Sociology Since Midcentury: Essays in Theory Cumulation. Randall Collins, ed. 
New York: Academic Press.

Colson, Elizabeth (1977). “A Continuing Dialogue:Prophets and Local Shrines Among 
the Tonga of Zambia”. in Regional Cults. R.P. Werbner, ed. Pp. 119-139. New 
York: Academic Press.

Conklin, W. J. (1978). The Revolutionary Weaving Inventions of the Early Horizon. Nawpa 
Pacha 16:1-12.

Culbert, T. Patrick, ed. (1973). The Classic Maya Collapse. Albuquerque: University of 
New Mexico Press.

Earle, Timothy (1997). How Chiefs Come to Power. Stanford: Stanford University Press.
Goldstein, Paul (1989). Omo, a Tiwanaku Provincial Center in Moquegua, Peru. Ph.D. 

dissertation, Chicago: The University of Chicago.



Darrell La Lone96 Rise, Fall, and Semiperipheral Development in the Andean World-System 97

Haas, Jonathan, Shelia Pozorski, and Thomas Pozorski (1987). The Origins and 
Development of the Andean State. Cambridge: Cambridge University Press.

Isbell, William (1988). “City and State in Middle Horizon Huari,” in Peruvian Prehistory. 
Richard W. Keatinge, ed. Pp. 164-189. Cambridge: Cambridge University Press.

Johnson, Allen W, and Timothy K. Earle (1987). The Evolution of Human Societies: From 
Foraging Group to Agrarian State. Palo Alto: Stanford University Press.

Jones, Grant D., and Robert R. Kautz (1981). The Transition to Statehood in the New 
World. Cambridge: Cambridge University Press.

Kolata, Alan L. (1993). The Tiwanaku: Portrait of an Andean Civilization. Cambridge, MA: 
Blackwell.

Kroeber, Alfred (1930). “Archaeological Explorations in Peru, Part II: The Northern 
Coast.” Field Museum of Natural History, Anthropology Memoirs 2(2).

Lechtman, Heather (1980). “The Central Andes:Metallurgy Without Iron.” in The 
Coming of the Age of Iron. T Wertime and J. Muhly, eds. Pp. 267-334. New Haven: 
Yale University Press

Mann, Michael (1986). The Sources of Social Power. Vol. I. A History of Power from the 
Beginning to A.D. 1760. Cambridge: Cambridge University Press.

Marcus, Joyce J. (1989) “From Centralized Systems to City-States: Possible Models for 
the Epi-Classic.” in Mesoamerica After the Decline of Teotihuacan, A.D. 700-900. 
Richard A. Diehl and Janet C. Berlo, eds. Pp. 201-208. Washington, D.C.: 
Dumbarton Oaks.

Marcus, Joyce J (1998). “The Peaks and Valleys of Ancient States, An Extension of the 
Dynamic Model,” in Archaic States. Gary M. Feinman and Joyce Marcus, eds. Pp. 
59-94. Santa Fe: School of American Research Press.

Marcus, Joyce, and Gary M. Feinman (1998). “Introduction.” in Archaic States. Gary 
M. Feinman and Joyce Marcus, eds. Pp. 3-13. Santa Fe: School of American 
Research Press.

McEwan, Gordon F. (1991) “Investigations at the Pikillacta Site: A Provincial Huari 
Center in the Valley of Cuzco,” in Huari Administrative Structure. William H. 
Isbell and Gordon F. McEwan, eds. Pp. 93-119. Washington, D.C.: Dumbarton 
Oaks.

Millon, René (1988). “The Last Years of Teotihuacan Dominance.” in The Collapse of 
Ancient States and Civilizations. Norman Yoffee and George L. Cowgill, eds. Pp. 
102-164. Tucson: University of Arizona Press.

Morris, Craig, and Adriana Von Hagen (1993). The Inka Empire and Its Andean Origins. 
New York: Abbeville Press.

Moseley, Michael (1992). The Incas and Their Ancestors. London: Thames and Hudson.
Moseley, Michael, Robert A. Feldman, Paul Goldstein, and Luis Watanabe (1991) 

“Colonies and Conquest: Tiahuanaco and Huari in Moquegua,” in Huari 
Administrative Structure. William H. Isbell and Gordon F. McEwan, eds. Pp. 
121-140. Washington, D.C.: Dumbarton Oaks.

Murra, John V. (1972). “El Control Vertical de un Maximo de Pisos Ecologicos en 
la Economia de las Sociedades Andinas.” in Visita de la Provincia de Huanuco, 
Vol. 2. John V. Murra, ed. Pp. 427-476. Huanuco, Peru: Universidad Nacional 
Hermilio Valdizan.

Schreiber, Katharina J. (1987). “From State to Empire: The Expansion of Wari Outside 
the Ayacucho Basin,” in The Origins and Development of the Andean State. Jonathan 
Haas, Shelia Pozorski and Thomas Pozorski, eds. Pp. 91-96. Cambridge: 
Cambridge University Press.

Schreiber, Katharina J. (1992). “Wari Imperialism in Middle Horizon Peru,” University of 
Michigan Museum of Anthropology Anthropological Papers N0. 87.

Service, Elman Rogers (1975). Origins of the State and Civilization: The Process of Cultural 
Evolution. New York: Norton.

Steward, Julian (1949). “Cultural Causality and Law: A Trial Formulation of the 
Development of Early Civilizations,” American Anthropologist 51:1-27.

Steward, Julian (1955). Theory of Culture Change: The Methodology of Multilinear Evolution. 
Urbana: University of Illinois Press.

Tainter, Joseph A. (1988). The Collapse of Complex Societies. Cambridge: Cambridge 
University Press.

Toland, Judith D. (1987). “Discrepancies and Dissolution: Breakdown of the Early Inca 
State,” in Early State Dynamics. Henri J. M. Claessen and van de Velde, eds. Pp. 
138-153. Leiden: Brill.

Topic, John (1991). “Huari and Huamachuco,” in Huari Administrative Structure. William 
H. Isbell and Gordon F. McEwan, eds. Pp. 141-164. Washington, D.C.: 
Dumbarton Oaks.

Topic, John, and Theresa Topic (1987). “The Archaeological Investigation of Andean 
Militarism:Some Cautionary Observations,” in The Origins and Development of the 
Andean State. Jonathan Haas, Shelia Pozorski and Thomas Pozorski, eds. Pp. 
47-55. Cambridge: Cambridge University Press.

Willey, Gordon R. (1991). “Horizontal Integration and Regional Diversity: An 
Alternating Process in the Rise of Civilizations,” American Antiquity 56:197-215.

Wilson, David J. (1981). “Of Maize and Men:A Critique of the Maritime Hypothesis of 
State Origins on the Coast of Peru,” American Anthropologist 83:93-120.

Wilson, David J. (1983). “The Origins and Development of Complex Prehispanic 
Society in the Lower Santa Valley: Implications for Theories of State Origins,” 
Journal of Anthropological Archaeology 2:209-276.

Wilson, David J. (1987). “Reconstructing Patterns of Early Warfare in the Lower Santa 
Valley: New Data on the Role of Confl ict in the Origins of Complex North 
Coast Society,” in Origins and Development of the Andean State. Jonathan Haas, 
Shelia Pozorski and Thomas Pozorski, eds. Pp. 56-69. Cambridge: Cambridge 
University Press.



Darrell La Lone98

Wilson, David J. (1989). “Full-Coverage Survey in the Lower Santa Valley: Implications 
for Regional Settlement Pattern Studies on the Peruvian Coast,” in The 
Archaeology of Regions: The Case for Full-Coverage Regional Survey. Suzanne Fish 
and Steven Kowalewski, eds. Pp. 117-145. Washington, D.C.: Smithsonian 
Institution Press.

Wilson, David J. (1997). “Early State Formation on the North Coast of Peru: A 
Critique of the City-State Model,” in The Archaeology of City-States:Cross-
Cultural Approaches. Deborah Nichols and Thomas Charlton, eds. Pp. 229-244. 
Washington, D.C.: Smithsonian Institution Press.

Wilson, David J. (1999). Indigenous South Americans of the Past and Present. Boulder: 
Westview Press.

Wittfogel, Karl (1957) Oriental Despotism: A Comparative Study of Total Power. New 
Haven: Yale University Press.

Wright, Henry T., and Gregory A. Johnson (1975). “Population, Exchange, and Early 
State Formation in Southwestern Iran,” American Anthropologist 77:267-289.

Yoffee, Norman (1995). “Political Economy in Early Mesopotamian States,” Annual 
Review of Anthropology 24:281-311.

Yoffee, Norman, and George L. Cowgill (1988). The Collapse of Ancient States and 
Civilizations. Tucson: University of Arizona Press.


	La Lone