core, peripheries, and regional realities in middle horizon peru · 2020. 3. 26. · core,...

Journal of Anthropological Archaeology 25 (2006) 346–370

www.elsevier.com/locate/jaa

Core, peripheries, and regional realities in Middle Horizon Peru

Justin Jennings

Department of Anthropology, Franklin and Marshall College, Lancaster, PA 17604, USA

Received 17 June 2005; revision received 18 November 2005Available online 17 February 2006

Abstract

In this article, I argue that the basic structure of the core–periphery models of dependency and world system theories areso deeply embedded in archaeological interpretations of early states and empires that social scientists working inside, out-side, and against these paradigms tend to frame their arguments, often implicitly, around a core–periphery model of radi-ating lines that connect peripheral zones to the core. I suggest that the dominance of this model significantly inhibits ourability to understand the social dynamics of outlying groups in some instances. Using a case study from the Wari state ofthe Pre-Columbian Andes, I demonstrate one example of a situation where a core–periphery model fails to capture theemergent properties of a regional system embedded within the cultural and economic milieu of the expanding Wari state.Without a more serious consideration of regional dynamics, archaeologists may misinterpret the past at all levels of inter-regional interactions involving state societies in the Andes and elsewhere.� 2006 Elsevier Inc. All rights reserved.

Keywords: Core–periphery; World systems; Inter-regional interaction; Early states; Andes; Wari

In the earliest expansionary states from aroundthe world, rapid urbanization created increasingimport demands for people, staple products, andluxury goods. Since the flow of imports underwrotepopulation growth, craft specialization, and elitedisplays, the development and stability of the statebecame dependent, in no small measure, on produc-tion from groups living far removed from the state’surban centers (Algaze, 1993a). Geographic, techno-logical, and organization hurdles to increasing andmaintaining inter-regional exchange networks weredaunting (Stark, 1990, pp. 250–256; Stein, 1999b,pp. 160–165), and state leaders met these challengesby variously establishing colonies, forming trade

0278-4165/$ - see front matter � 2006 Elsevier Inc. All rights reserveddoi:10.1016/j.jaa.2005.12.001

E-mail address: [email protected].

partnership, forging political alliances, developingelite ideologies, and waging war in outlying areas(e.g., Adams, 1984; Algaze, 2001; Goldstein, 2005;Gosden, 2004; Stein, 1999a; Zaccagnini, 1987).The political economies of states were thereforelinked in different ways to the political economiesof other regions, and, in many cases, significantchange occurred in these contacted areas becauseof inter-regional interaction.

Apart from diffusionism, the first sustained con-sideration of the link between early state developmentand peripheral change occurred during the late 1970swhen archaeologists were introduced to dependencyand world systems theory. Over the last thirty years,these theories have had an immense influence on thefield and have spurred a widespread interest ininter-regional interaction (e.g., Algaze, 1993b;

.

mailto:[email protected].

J. Jennings / Journal of Anthropological Archaeology 25 (2006) 346–370 347

Champion, 1989; Cusick, 1998; Gosden, 2004; Halland Chase-Dunn, 1993; Kardulias, 1999; Peregrineand Feinman, 1996; Stein, 1999a). Over this time,the theories have been significantly modified (e.g.,Berdan et al., 1996; Chase-Dunn and Hall, 1992;Rowlands et al., 1987; Schreiber, 1992), cogentlyattacked (e.g., Stein, 1999a; Schortman and Urban,1998), and rejected for rival models (e.g., Cusick,1998; Lyons and Papadopoulos, 2002). Perhapscounter intuitively, I argue that this long debate haspropelled the radial core–periphery model that struc-tures dependency and world systems theory into thediscipline’s practical consciousness.

In this paper, I suggest that recent debates on theinter-regional interactions caused by expandingstates have failed to challenge the underlying archi-tecture of the core–periphery model of dependencyand world systems theory. Instead, conceptions ofstate–local relationships are so strongly influencedby this model that social scientists working inside,outside, and against these paradigms frame theirarguments, often implicitly, around a state core thatconnects to peripheral zones through lines radiatingout of the center. The largely tacit acceptance of theradial model might be because it is the best repre-sentation of a complex reality. I suggest, however,that the radial core–periphery model of expansioniststates often fails to capture many of the realities ofinter-regional exchange and local political develop-ment in outlying areas. I argue that the modelaccentuates core–periphery relationships, assumesstructural links within the system, and neglectsregional networks. To illustrate my argument, Iexamine how core–periphery relationships fail tocompellingly explain the relationship between the

Fig. 1. Graphical representations of dependency and world systems thethrough the flow of goods. In the dependency model, the system is contthe world systems model demonstrate how core control changes in the

Pre-Columbian Wari state and the people of theCotahuasi Valley in southern Peru.

Cores and peripheries in the ancient world

Before the introduction of dependency and worldsystems theory in the 1960s and 1970s, archaeolog-ical rationalization for culture change were largelylimited to explanations of the migration of popula-tions, the diffusion of ideas and technology, or envi-ronmental shifts (Schortman and Urban, 1987,1992). These new theories forced archaeologists tomore fully consider the existence of sustained, sys-tematic links between societies and to explore therelationship of these linkages to culture change.The theories introduced or inspired an analyticalvocabulary of cores, peripheries, hegemonies, directcontrol, and prestige goods that have been very use-ful for modeling the interactions in the past. In thissection, I briefly describe dependency and world sys-tems theory, the engagement that archaeologistshave had with these theories, and the deficienciesof the radial model for describing some cases ofinter-regional interactions.

The emergence of the radial model

Dependency and world systems theory initiallyarose out of a context of increasing concern overthe connection between the growth of the westernworld and the stasis or decline of what would becomeknown as the third world (Fig. 1). Frank (1966, 1967)the key developer of dependency theory, first drewthe outline of the radial core–periphery model bypositing a systematic, structural connection between

ories. The core (C) is connected to peripheral regions (black dots)rolled (represented in gray) by the core. The variations of gray insemi-periphery and periphery.

348 J. Jennings / Journal of Anthropological Archaeology 25 (2006) 346–370

rich and poor nations. He suggested that the thirdworld was locked in an exploitative relationship withthe first world because of an international division oflabor, the extraction and exportation of third worldraw materials, and third world dependency on firstworld high-value, finished goods. The third worldcould not follow the development path of the Westbecause colonialism in these countries had restruc-tured these political economies to meet Westerninterests.

Wallerstein (1974) built on Frank’s dependencywork and created a new world systems model thatnevertheless left the basic radial model of core–pe-riphery relations intact. According to Wallerstein,a world system can be divided into three majorzones, a core, semi-periphery, and periphery, thatare tied together by a world market of bulk com-modities that are necessary for every-day life.Each zone has its own economic structure andsystem of labor control. The core contains stronggovernments and generally supports industry andwage labor, the periphery contains weak statestypified by coerced labor and monoculture, andsemi-strong states with limited industrializationmake up the semi-periphery. Surplus wealth flowsinto the core from peripheral and semi-peripheralnations. The strong states in the core increase andsustain the flow of wealth through extra-economicmeans.

While André Gunder Frank embraced Waller-stein’s elaboration of his ideas, the two scholarsdisagreed on the applicability of the world sys-tems approach in the ancient world. Wallerstein(1974, pp. 15–16) suggested that his model couldnot be applied to understanding the pre-capitalistworld both because ancient states had insufficientcontrol over peripheries and because the systemlacked the technologies necessary to produce asignificant surplus. Nonetheless, he suggested thatat certain points in history, ‘‘world empires’’ hadrisen to hold economic sway over large areas ofthe world. In these cases, the periphery was close-ly integrated to the core through administered,tribute paying provinces (1974, p. 16). Frank,on the other hand, thought that the modernworld system was the end result of thousands ofyears of evolution. In his more recent work, hedirectly connects the Uruk expansion of fourthmillennium BC Mesopotamia to the present daythrough cycles of capital accumulation andcore–periphery relations (Frank, 1993; Frankand Fills, 2000).

The resiliency of the radial model

The scholarship of Frank and Wallerstein wasattractive to archaeologists because it linked poli-tics, economics, and geography in ways that werearchaeologically visible (Stein, 1999a, 2002b).World system analyses became popular in the disci-pline in the 1980s and 1990s, and the model wasbrought to bear across the world and throughouttime on a wide variety of societies (e.g., Chase-Dunnand Hall, 1992; Kardulias, 1999; Rowlands et al.,1987). At one extreme, Chase-Dunn and Halldescribed a ‘‘very small world-system’’ that operat-ed among the Wintu and rival groups in California’sSacramento Valley (1997, pp. 121–189). At theother extreme, the world system approach was usedto understand the relationship of the RomanEmpire to northern Europe, the Middle East, andnorthern Africa (Nash, 1987).

While some scholars hoped to bring the world-systems approach into the ancient world largelyunchanged, most archaeologists were critical ofworld-systems to one degree or another and, asSchortman and Urban (1992, p. 19) suggest, ‘‘thetheory is never accepted without reworking, redefi-nition, or other adjustments.’’ Schneider (1977)offered perhaps the most influential early critiqueof Wallerstein’s work by suggesting that the flowof small amounts of luxury goods could also bindtwo regions together because luxury goods con-ferred prestige to those that acquired the items.Other archaeologists challenged other core aspectsof the theory, such as the existence of a semi-periph-ery and the long-term stability of the core (e.g.,Blanton and Feinman, 1984; Ekholm and Fried-man, 1979; Kohl, 1978, 1987; Upham, 1986). Theseearly critiques led to three major developments inthe ways that archaeologists have conceptualizedthe inter-regional interaction of ancient states.

The first development has been the introductionof new models that are more appropriate for theancient world (Fig. 2). At one extreme, the modelsare only slightly modified versions of dependencyand world-systems theories that downplay coredominance and highlight structure and interdepen-dence (e.g., Chase-Dunn and Hall, 1997, pp.12–15;Denemark et al., 2000, p. xvi; Kardulias, 1999;Peregrine and Feinman, 1996). At the other extremeare new models that have been introduced thatchange the items moving through the system, alterthe flow of goods, deny core dominance over theperiphery, and demonstrate peripheral influence

Fig. 2. Graphical representations of recent conceptualizations of early state interaction systems. (A) Represents a prestige goods economywhere peripheral elites become dependent on core goods, (B) is an interaction model that stresses variability in the ways that peripheralareas are integrated into the state, and (C) represents a focus on peripheral agency and the local effects of state expansion.


over core culture (e.g., Earle and D’Altroy, 1989;Schneider, 1977; Stein, 1999a). Some of the latermodels are offered as counters to world systemsand dependency approaches.

The second development is the increasing recog-nition of variability within each core–periphery sys-tem. In contrast to a monolithic system of direct orindirect control over a periphery, scholars have not-ed that the relationship between states and outlyingareas can be dramatically different from place toplace (e.g., Berdan et al., 1996; Malpass, 1993;Rothman, 2001; Schreiber, 1992). Variations ininteraction have been linked to a wide array of fac-tors, including the political economies of the core,the political complexity of the contacted area, thetransportation costs of reaching the area from thecore and vice versa, the area’s economic resourcepotential, and the ideological importance of thecontacted area. The variability models often com-bine aspects of dependency and world systems mod-els with the newer models introduced for the ancientworld.

The third development in the ways that archaeol-ogists have conceptualized the inter-regional inter-action of ancient states has been a movementdown in scale to study the nuances of local agency,heterogeneous identities, technological innovation,and other aspects of culture change. Coincidingwith postcolonial concerns of the subaltern (Spivak,1994), this development has been particularly pro-nounced in peripheral areas, where archaeologistshave formulated a more nuanced view of the politi-cal, economic, and social consequences of culturecontact by focusing on the interactions betweenstate agents, local populations, and material culture(e.g., Dietler, 1990, 1998; Domı́nguez, 2002; Stein,1999a,b; Wells, 1999). This work has demonstratedthe importance of smaller scale analyses to theunderstanding of inter-regional interactions, and

has demonstrated that core dominance and eco-nomic hegemony was often much weaker than inmodern systems. With the exception of Stein(1999a,b, 2001, 2002a,b), scholars have not general-ly proposed macro-level interaction models basedon their studies of local dynamics.

Over the last three decades, archaeologists andother social scientists have debated the merits ofdependency and world systems based interpreta-tions of the past. Scholars have adapted the core–periphery model outlined in these theories to theancient world by shrinking the system, making thecore less dominant, allowing a more active and var-ied periphery, and diversifying the types of connec-tion between the two zones. These substantial shiftshave been the result of fruitful exchanges betweensupporters and detractors of world systems anddependency theory (especially see Chase-Dunn andHall, 1997; Deagen, 1996, 1998; Frank, 1993; Light-foot, 1995; Stein, 2002a; Schortman and Urban,1998). While this scholarship has significantlyadvanced our understanding of early state dynam-ics, I argue that these ideas are variations on atheme that leaves the radial core–periphery modelembedded in our interpretations.

The deficiencies of the radial model

I suggest that the radial model has become amodel of ancient reality that has been naturalizedover the course of our engagement with dependencyand world systems theories (Bourdieu, 1977, p. 164;Geertz, 1973, p. 93). While archaeologists shift ter-minology, create new models, and explore periphe-ral agency, the basic architecture remainsunchallenged because it is part of our ‘‘practicalconsciousness’’ that is rarely articulated to ourselvesand each other (Giddens, 1984, p. 7). The conse-quence of its non-discursive acceptance is the


imposition of an a priori structure on the relation-ship between a polity and outlying regions. Technol-ogy, ideas, people, and trade items are envisioned asflowing from the core to a peripheral area (and viceversa) along a series of connecting pathways thatare sometimes anchored by colonies and/oradministrative centers. The structure of the modeldetrimentally influences, often implicitly, interpreta-tions of the material record in at least three linkedways (Fig. 3).

First, the model accentuates core–periphery rela-tions by privileging the connections between thestate and outlying areas. For example, a handfulof state style sherds at a peripheral site are givenmuch more weight than ceramics from otherregions, local reactions to state ideology are high-lighted, the amount of state investment in localresources is monitored, and the adoption of periph-eral foods in the core is studied. Understandinginter-regional interaction within and around statesbecomes tantamount to understanding the relation-ship between the core and a peripheral area becausethe structure of the model highlights the importanceof this relationship. Following the model’s logic,studying the relationship between the core and eachperipheral area reveals the nature of the inter-re-gional interaction system. The whole is equal tothe sum of its parts, and the parts are cores andperipheries linked together.

The second deficiency of the radial model is thatit assumes important structural links between thecore and the periphery. In some cases, this link isthrough direct state control or the establishmentof colonies. In other cases, the system is a more den-dritic one in which core and periphery goods flowback and forth between many intermediate tradersthrough unbalanced economic exchanges (Santleyand Alexander, 1992). Despite these conceptualvariations, the underlying model assumes that these

Fig. 3. Deficiencies of the radial core–periphery model.

connecting links exist and that these core–peripherylinks form the defining relationship of the peripheralarea’s inter-regional interaction. Archaeologiststherefore attempt to find the links between theirarea and the core, and use this data to argue for‘‘direct control’’, ‘‘indirect control’’, or ‘‘outsideinfluence’’ of their area by the state. Outside interac-tions are effectively collapsed into a single relation-ship with the core.

The final deficiency, the neglect of regionalism,builds on these first two critiques. By accentuatingthe core and the periphery and assuming core–pe-ripheral structural links, there is little interpretativespace for intra-regional interactions. Contactbetween areas in the periphery is seen as less impor-tant, and tangential to interpretation of inter-re-gional interactions with the state. Connectionsbetween places are often considered as intermediatelinks between the core and a peripheral area, cultur-al change is a result of the dynamics of a core–pe-riphery interaction, and the logic of differentactors is explained by their position within the struc-ture of the radial model. Possible emergent proper-ties of regional interactions are ignored or poorlyinterpreted because the framework of the radialmodel is not conducive to exploring theserelationships.

Core and periphery in the Wari state

The radial model’s architecture of inter-regionalinteraction is deeply embedded in our understand-ing of the Wari state of the ancient Andes. WhileI discuss in another paper the history of Wari schol-arship (Jennings, in press), I only consider in thissection the available data and consider how theradial model informed our current interpretationsof the state’s relationship with outlying areas. Likeother areas of the world, the development of archae-ological views on Wari has played out within thecontext of the discipline’s general engagement withdependency and world systems theory. The nuancesof this engagement over time are less important tous than a general grasp on the Wari data and theways in which these data have been generallyinterpreted.

During the Middle Horizon Period (600–1000AD), the Wari state developed in the AyacuchoValley of the Peruvian Andes (Fig. 4). The Waricapital (also called Wari) quickly grew to reach apopulation size of at least 10,000 people (Isbell,1997, p. 186), and, as the population became more

Fig. 5. Model of the Wari state and its variable interaction withperipheral zones.

Fig. 4. Map of northwestern South America with modernpolitical boundaries marked in bold outline. The area of Waristylistic influence encompassed much of Peru and is shown indark grey. The dots show the locations of places mentioned in thetext: (A) Chavı́n de Huantar, (B) the Wari capital, (C) Cuzco, (D)Cotahuasi, (E) Chuquibamba, (F) Colca, (G) Arequipa, (H) LakeTiticaca, and (I) Tiwanaku.


dense, increasingly specialized and wealthy(Gonzalez Carre, 1981, p. 94; Isbell, 1997, p. 206;Spickard, 1983, pp. 153–154; Von Hagen and Mor-ris, 1998, p. 130), the demand for non-local goodsincreased. The need for agricultural productsappears to have been largely met by intensificationand state regulation of surrounding valleys (Brow-man, 1999; Isbell, 1977; Raymond, 1992; Raymondand Isbell, 1969; Vivanco and Valdez, 1993), whileless bulky and/or non-perishable goods likely cameinto the valley from more distant regions (Jenningsand Craig, 2001).

Inter-regional exchange increased significantlyduring the Middle Horizon throughout Peru. Wariand other cultures during the period benefited fromthe increased flow of prestige goods, staple items,and ideas (Burger et al., 2000; Lechtman, 1980;Shady Solı́s, 1988). Wari was an important, if notthe most important, node in this exchange network.Raw materials and finished products flowed into the

city from across the central Andes, and specialists atthe capital produced textiles, stone figurines, ceram-ic vessels, metal objects and other items that wereexchanged across a wide area of Peru (Isbell, 1997;Pérez Calderón, 1999). The extension of MiddleHorizon exchange networks was paralleled by theestablishment of at least twenty sites with Wari stylearchitecture outside of the Wari heartland. Surveysand excavations at four of these sites have demon-strated that the sites were built and occupied by set-tlers from the Wari heartland, and three of thesesites most likely administered activities in the valleysthat surrounded the sites (Isbell, 1989, 1991; McE-wan, 1987, 1991, 1996; Schreiber, 1992, 1999; Wil-liams, 1997, 2001; Williams and Isla, 2002).

In the predominant model of Wari political econ-omy (Fig. 5), the Wari capital and its surroundingvalleys were the core of the state and all of the siteswith Wari style architecture served as nodes in aWari administrative hierarchy that helped to orga-nize local political economies for the extraction ofdesired resources (Isbell and McEwan, 1991; Isbelland Schreiber, 1978; Jennings and Craig, 2001;McEwan, 1987, 1996, 2005; Schreiber, 1978, 1987,1992, 1999, 2001). The model stresses the variabilityof state–local relationships, with the administrativesites anchoring pockets of direct control within animperial mosaic of direct and indirect control thatwas determined by state interests, logistical con-cerns, and local conditions (Schreiber, 1992, p. 276).

Wari administrative centers were locations where‘‘administrative elites controlled vast resources byexerting authority over the labor of others or influ-ence over local leaders who possessed labor andresources’’ (Nash and Williams, 2004, p. 154). Thestate intensified production of resources valued in

1 From 1999–2000, my field crew and I conducted a settlementsurvey of the upper reaches of the valley between the moderntowns of Puica and Cotahuasi. Due to the valley’s extreme, and attimes dangerous, topography, the survey used a mixed strategy offull coverage survey and judgmental survey based on informants’knowledge and topographical criteria. During this time, we alsoexcavated seven test units at the site of Ancient Alca in order tobuild the local chronology. Unless otherwise indicated, theCotahuasi data discussed in this paper are drawn from the1999–2000 survey and excavations (Jennings, 2002).


the core, collected these resources at state centers,and sent these items to the Wari capital (Schreiber,1999). In return, the state provided prestige goodsand lavish hospitality to local leaders and their cli-ents (Cook and Glowacki, 2003; Nash and Wil-liams, 2004). Regions without administrativecenters, according to this model, were still ‘‘abso-lutely consolidated under the administrative sys-tem’’ through indirect control (Schreiber, 1992, p.31). Wari indirectly controlled regions through col-laboration with local leaders, the re-organization ofeconomic structures, patron-client feasting, thethreat of force, and/or the capture of mummiesand sacred objects (Cook and Glowacki, 2003;Isbell, 1988; McEwan, 1998, 2005; Schreiber, 1992).

This model of Wari is not without its critics.Some scholars have argued that Wari control wasnon-existent, or largely ephemeral, over some areas(Grossman, 1983, p. 85; Hastorf, 1993, p. 46; Lau,2002, p. 300; Mackey, 1982, p. 330; Shady Solı́s,1982, 1988, 1989; Shimada, 1985, p. 361; Wilson,1988, p. 358), and these data have been used to sug-gest that the state’s territorial extent was far morelimited (e.g., Topic, 1991b, p. 244), and, more rare-ly, that Wari was only a minor highland polity (e.g.,Bawden and Conrad, 1982, pp. 31–32; Shady Solı́s,1982, pp. 63–64). Other archaeologists do not neces-sarily dispute the spatial extent of Wari, but suggestthat the evidence points to a more consensual rela-tionship between the state and local populationsor a relationship based more on religious conversionthan on conquest. (e.g., Topic, 1991a, p. 162). Thisopposition to the prevailing model of Wari, howev-er, tends not to challenge the structural assumptionsupon which the model was based. The alternativemodels only change either the variety of core–pe-riphery relations at the local level or alter the geo-graphic spread of the core–periphery system.

While recent scholarship stresses the variable nat-ure of Wari control strategies across Peru, the mod-el used to describe interaction during the MiddleHorizon remains based on a radial core–peripherymodel. The model frames our understanding ofthe period in Peru, and thus shapes the ways inwhich archaeological data are interpreted. Evidencefor agricultural intensification coupled with Wariinfluenced architecture and/or artifacts, for exam-ple, is taken as evidence for Wari extraction of agri-cultural products and the eventual consumption ofmuch of these products in the state’s core. Wariceramics and textiles in a region points to Wariinfluence over an area, and shifts in political organi-

zation and settlement pattern are attributed to Warimanipulation of local populations. In some areas,Wari likely did re-organize groups for resourceextraction in a manner similar to that proposed bythe proponents of the Wari core–periphery model.In other areas, however, this model fails to reflectthe reality of the state’s impact on people’s lives.One area that is poorly served by the core–peripherymodel is the Cotahuasi Valley.

The Cotahuasi Valley and the Wari state

Located at 14� south latitude and 73� west longi-tude, the Cotahuasi Valley lies in the highlands ofsouthern Peru almost 300 km from the Wari capi-tal.1 The valley is a resting point along a naturalcorridor between the sierra and the sea (Trawick,1994, p. 33) and contains rich sources of obsidian,gold, silver, copper, and rock salt (Burger et al.,1998; Canchaya et al., 1995; Zapata Llano, 1904).In part, these features led the later Inca Empire toinvest heavily in the valley by building an adminis-trative center, placing a major road through the can-yon, and creating a ceremonial center out of a localritual site (Jennings, 2003a,b). The valley’s geo-graphic position and natural resources may havebeen equally desirable to the Wari state, and thereis significant evidence for social, political, and eco-nomic change in the valley that correlate to Wariinfluence.

During the Middle Horizon, population signifi-cantly increased, new sites were found, agriculturalproduction expanded, exploitation of the valley’sobsidian, precious metals, and rock salt intensified,and social stratification increased. These changesoccurred at a time of profound Wari influence onlocal ceramics, the possible importation of Wariceramics and textiles, and the establishment of a sitenamed Collota with Wari-style architecture. Thecorrelation of Wari influence with societal changeled previous scholars to suggest that Wari directlyinfluenced the valley (Chávez Chávez, 1982, p. 86;


Trawick, 1994, p. 72). In my own work, I havemoved away from an earlier initial argument ofdirect control (Jennings and Craig, 2001) to onemore recently of Wari indirect control or influenceover the valley (Jennings, 2002; Jennings and YépezÁlvarez, 2001a,b).

Arguments of Wari direct control, indirect con-trol, or influence fit the data to a degree, but theyare limited by the structure of the core–peripherymodel within which they are formed. A closerinspection of the Cotahuasi data leads to a differentvision that de-emphasizes the core’s importance,highlights regionalism, and clarifies local agencyduring the period. I believe that this alternativeinterpretation better captures the reality of MiddleHorizon life in the valley by considering the emer-gent properties of regional interactions during thisperiod. As with any interpretations based largelyon surface remains, one must be aware of post-de-positional changes that can affect the materialrecord. In Cotahuasi, sites tend to be well preservedand significant looting in the area has only begunover the last ten years. Moreover, excavations attwo sites in the valley demonstrate that surfaceremains often correspond closely to excavatedassemblages. Nonetheless, further excavations areneeded to test interpretations based on surfaceremains.

To make my argument against the utility of aradial core–periphery model in this case, I considertwo broad themes in the valley’s material recordfrom the Inchoate Period (900 BC–600 AD)2

through the Middle Horizon that reflect change ininter-regional interaction. First, I look at the flowof goods and ideas between Cotahuasi, the Wariheartland, and other regions. Second, I considerthe role of the Wari state in the socio-politicalchanges that occurred in the valley during the Mid-

2 The Inchoate Period combines the Early Intermediate Period(200 BC–600 AD) and Early Horizon Period (900–200 BC) in thegeneral central Andean sequence developed by Rowe (1962)because there are no stylistic changes that mark the shift betweenthe two periods in Cotahuasi. The Cotahuasi chronologicalsequence is built from a local ceramic and point typology that issequenced and dated by both our own excavations and compar-ison with similar artifacts from other regions (Jennings, 2002).Only seven of the 1340 diagnostic sherds collected during thecourse of the fieldwork can be dated to the Early IntermediatePeriod based on comparisons to ceramics from other regions.Since excavation data from Ancient Alca suggest that there wasno break in occupation during the Early Intermediate Period, Iargue that Cotahuasi’s ceramics changed little during theInchoate Period.

dle Horizon. After each section, I evaluate the fitbetween the data and the radial core–peripherymodel of the Wari state and propose what I thinkis a more appropriate interpretation of the availabledata.

The flow of goods and ideas

The flow of goods into the valley can be tracedprimarily through importation of ceramics and theexportation of natural resources. During the Incho-ate Period, the valley was largely insulated from theoutside world. The pottery style from the period,called the Aicano-style, was part of a broad regionalceramic tradition in Arequipa that includes theHachas, Sopora, and Ayawala styles (Neira Aven-daño and Cardona Rosas, 2001). Pots were oftenundecorated, and when decorations are found onvessels they often include an appliqué of raised linesrunning parallel to the lip of the vessel, applicationsof a modeled human face with circular incision forthe eyes and mouth, or depressions made by theuse of a thin circular stamp on the surface of theceramic (Fig. 6). While Nazca and Early Tiwanakustyles significantly influenced adjacent traditions(Bauer, 1999, pp. 123–141; Neira Avendaño, 1998,pp. 29–33), the Aicano style was remarkably stablewith very few changes in form and decoration fromthe introduction of the styles on the coast around1400 BC to the Middle Horizon. The stability ofCotahuasi’s ceramic style was not a marker of cul-tural isolation, but rather marked a lack of emula-tion. The presence of seven imported sherds in thevalley from undefined styles of the central or south-ern highlands of Peru suggest regional contacts dur-ing this period.

Cotahuasi’s geologic resources flowed out of thevalley during the Inchoate Period. Obsidian fromCotahuasi, known as the Alca type obsidian, wasfirst used on the coast of Arequipa by at least13,000 years ago (Sandweiss et al., 1998, p. 1832).During the Inchoate Period, distribution was gener-ally confined to the south central Andes, but Alcaobsidian was found as far away as Chavı́n de Huán-tar in the northern highlands of Peru (Burger et al.,2000, p. 313). Cotahuasi was the prominent regionalsupplier of volcanic glass for much of what are nowthe departments of Cuzco and Arequipa (e.g., Bur-ger and Asaro, 1979, 1993, p. 33, p. 198; Burgeret al., 2000, p. 314; Cardona Rosas, 2002, p. 49),and a supplier of ‘‘moderate abundance’’ to sitesin the northern Titicaca basin (Burger et al., 2000,

Fig. 6. Aicano style ceramics from the Inchoate Period.


p. 267). Although undocumented archaeologically,small amounts of other locally available resources,such as rock salt and ochre, were likely alsoexchanged during this period.

The Middle Horizon brought sweeping changesto the ceramic assemblages of the Cotahuasi Valley.With the exception of a few utilitarian forms, Aican-o style ceramics were completely replaced by ceram-ics painted in black, white, red, or purple thatemulate Wari styles in form and design (Fig. 7).Based on petrographic temper analysis of the pastes,97% of the sherds collected during our fieldworkwere manufactured locally. Most of these locallymade ceramics (60%, n = 171) were in a Wari influ-enced local style that did not relate to any otherknown Wari state or regional ceramic style. Sixtyof the collected sherds (21%) closely followed thedecorative motifs of the Viñaque style, a commonWari style found throughout the Peruvian highlands(Menzel, 1964, p. 36), and six sherds were painted inloose imitation of the Chakipampa (1.7%, n = 5)and Okros (.03%, n = 1) styles, two styles common

in the state’s heartland (Menzel, 1964, pp. 17, 67).The remaining locally produced Middle Horizonsherds from the valley are in the Ccocscopa style(Neira Avendaño, 1990, p. 129), a regional stylefrom the nearby Chuquibamba Valley (14%,n = 36).

There are only eleven Middle Horizon sherdsmanufactured with pastes unknown to the valleyin previous or subsequent periods. Six of thesesherds are in the Ccoscopa style (Neira Avendaño,1990, p. 129), two sherds are in a local Cuzco relatedstyles (Bauer, 1999) and the three remaining frag-ments are Viñaque sherds. The most likely sourcefor the Viñaque sherds is perhaps the Wari centersof Pikillacta and Huaro in the Cuzco area bothbecause the sites are the closest major Wari centerto Cotahuasi and neutron activation analysis sug-gests that most of the Wari style pottery found atleast at Pikillacta was manufactured locally (Mon-toyo et al., 2001, p. 181). Although Viñaque styleceramics are quite rare at and around Pikillacta(Bauer, 1999, pp. 83–85; Glowacki, 2005, p. 105),

Fig. 7. Wari influenced ceramics from the Middle Horizon.


they are much more prevalent at the site of Huaro(Glowacki and McEwan, 2001).

Geologic resource exploitation in Cotahuasiincreased dramatically in the Middle Horizon. Thetrade volume and distribution range of Alca obsid-ian increased significantly during the Middle Hori-zon (Burger and Asaro, 1979, p. 36; Burger et al.,2000, p. 324). While continuing to be used most

heavily in Cuzco and Arequipa, Alca obsidian wasused at the Wari capital and at several provincialsites scattered across Peru (Burger et al., 2000).The intensification of Alca obsidian exchange maycorrelate with the founding of five small quarriesat the source (Jennings and Glascock, 2002, p.114), but no artifacts found at these sites could beused to securely date them.


The arsenic rich copper ores available in the val-ley were first used in Andean metallurgy during theMiddle Horizon (Lechtman, 1980, 2005, p. 307, p.132; Olchauski and Dávila, 1994), and the only dat-able mine in our survey area is a small copper minethat was used during the Middle Horizon and LateIntermediate Periods.3 Since my field crew andI documented no evidence of smelting activities(i.e., Lechtman and Klein, 1999; Van Buren andMills, 2005), I suggest that miners may have crushedthe ores manually, separated metal-bearing mineralsfrom the rock matrix, and then exported these min-erals (i.e., Van Buren and Mills, 2005, pp. 7, 25). Itis doubtful that the valley’s copper was traded out-side of the region because mines in northern andcentral Peru likely served metalworkers in thoseregions, and Bolivian artists used exclusively tin-rich ores that were not available in Cotahuasi(Lechtman, 2003, 2005, p. 133).

Cotahuasi contains one of the few rock saltsources in Peru (Zapata Llano, 1904). Besides itsuse as a seasoning and an additive in metal smelting,rock salt is seen today as a superior preservative rel-ative to salt brines, and thus is a highly valued com-modity. At present, the importance of rock salt inthe barter economy far outweighs that of the metal-lurgical deposits that remain in the valley (Trawick,1994, p. 86). Although no archaeological remainsare found in or adjacent to the present day rock saltmine, sherds dating to the Middle Horizon and afterwere found at Huarhua, the town adjacent to themine. Oral histories in the village also suggest a pre-hispanic origin of the mine and describe how the saltwas traded extensively deep into the past. The rocksalt is traded today throughout the region.

The flow of goods into and out of Cotahuasi wasparalleled by the flow of ideas. In the Inchoate Peri-od, there seems to have been little cultural exchangebased on ceramic styles. Despite the existence of ahandful of imports, only one inchoate period sherd,an imitation of the Pukara style of the Lake Titicacaregion (Franquemont, 1990), demonstrates a local

3 Pre-Columbian mines are difficult to date because they oftencontinued to be mined after conquest and subsequent miningactivities tended to destroy or bury most traces of earlier mining.The only datable mine that we documented in Cotahausi, CO-20,is found near the modern town of Alca. We date the site based oneleven ceramic shreds found around the exterior of the mine.Since no sites or agricultural terracing are located above or belowthe mine, it is likely that the sherds date to the period of themine’s use. We located two other mines that may also be Pre-Columbian, but contain no dateable ceramics.

potter’s experimentation with outside styles. Thelongevity of the Aicano style likely reflects a lackof desire for innovation in a social context of ceram-ic production that reinforced a local tradition thatstressed community affiliations (i.e., Dietler andHerbich, 1998). This insularity shifted dramaticallyin the Middle Horizon with the profound effect ofWari styles on the ceramic traditions of the valleythat I described earlier. New forms, colors, and ico-nography were introduced that swept aside localstyles that had remained stable over hundreds ofyears. Almost all of the Wari style vessels were man-ufactured locally. The painted designs were oftenpoorly executed relative to true Wari examples,mixed and abbreviated iconographic elements, anddrew from Ccoscopa, Ccoipa, and other Wari influ-enced local style from the region. Unlike the Incho-ate Period, Middle Horizon potters were open toinnovations and experimented with Wari stylesfrom across the region.

Although Wari profoundly influenced Cotahu-asi’s ceramic styles, there were few significantchanges in the ritual activities of the valley duringthe Middle Horizon as measured by funeraryremains, painted tablet offerings, and ceremonialarchitecture. There were at least six cemeteriesspread out across the valley during the InchoatePeriod, and people were buried collectively withinthese cemeteries in tombs underneath boulders andin above-ground burial houses. During the MiddleHorizon, ten additional cemeteries were found, butthe dead continued to be buried in all of the preex-isting cemeteries, using the same types of tombs,and, in a few cases, old tombs appear to have beenreused. At the new cemeteries, the only change wasthe introduction of chullpas (single-storied burialtowers) and tombs set into terraces. The introducedtombs likely reflect increased contact with otherareas of the Arequipan sierra where tombs like thesewere more common (Sobczyk, 2000, p. 53; Wernke,2003, p. 162).

In some cases, burial customs in the valleyoverlapped with a long-standing ritual practiceinvolving stone tablets that were placed in graves,in rock hollows, or deposited directly into theearth. The tablets, likely linked to ideas concern-ing fertility, were almost always painted on onlyone side, and the iconography on the tablets var-ies widely and includes an array of anthropomor-phic, zoomorphic, and geometric patterns (Fig. 8;Jennings, 2003a; Kauffmann-Doig, 1991; LinaresMálaga, 1988). The tablets are often found

Fig. 8. Painted stone tablets from the Cotahuasi Valley.


together in groups, and are occasionally stackedin pairs with the painted surfaces facing eachother. The Cotahuasi tablets are part of a regionaltradition that dates to at least 2000 BC (Jennings,2003a, p. 113; Ravines, 1970). The earliest knowntablets from the valley come from the surface ofWayñuna, a site dating to between 6000 and 900BC. During the Middle Horizon through LateIntermediate Period (1000–1476 AD), the tablettradition flourished in Cotahuasi and throughoutmuch of southern Peru (Linares Málaga, 1988,

p. 56, 62). Middle Horizon tablets appear to havebeen generally more colorful and more diverse rel-ative in its iconography than both earlier and laterexamples. This change, however, does not seem toreflect Wari influence over the medium. On thecontrary, the iconography used on Middle Hori-zon tablets is similar to that used on the region’spainted tablets and in rock art in general since atleast the Middle Archaic Period (6000–4000 BC)(Kauffmann-Doig, 1991, p. 29; Linares Málaga,1978, p. 386).


Although ritual architecture is poorly understoodfor the period, activities in the valley were likely cen-tered on at least three shrines that were adjacent tojutting rock outcrops (Jennings, 2003b). D-shapedstructures were important to Wari religion, andthese structures are found in Wari sites in Ayacuchoand in several sites throughout Peru (Cook, 2001).These structures, at least in the Sondondo Valleyof the central highlands, were important means bywhich local populations were integrated into Wariideology (Schreiber, 2004, p.143). In Cotahuasi,there are two D-shaped buildings at the site ofAncient Alca, but these buildings date to the LateIntermediate Period and worship at the sites wasdeeply intertwined with the painted tablet tradition(Jennings, 2003b, p. 444). Since these buildings arelater in time and are associated with local ritualpractices, they do not reflect any significant Wariinfluence over religious practices during the MiddleHorizon.

Cosmological principles are often key symbolsthat are integral to the structural organization of asociety (Ortner, 1973). These symbols help to formmodels of and for reality that shape a person’sunderstanding of their culture and their cosmos(Geertz, 1973, p. 93) and are often materialized inplaces of ritual importance (Bradley, 1998). If evi-dence from the funerary customs and the tablet tra-dition can serve as testaments to the beliefs of thepeople of Cotahuasi, then we see no evidence forsubstantial cosmological change during the MiddleHorizon. The introduction of new tomb types andthe sharing of tablet iconography during the MiddleHorizon likely reflect increased influences fromneighboring valleys with related traditions (Cardo-na Rosas, 1993; Sciscento, 1989). While Wari styleschanged Cotahuasi ceramics, this influence did notspread into the cosmological realm.

From the Inchoate Period to the Middle Horizon,there was a significant change in Cotahuasi relation-ship with the outside world. More goods flowed intoand out of the valley, and, more importantly, therewas a fundamental shift in Cotahuasi’s social hori-zons. This shift does not appear to have been orches-trated directly by the Wari state. The valley’s ceramic,geologic resource, and ritual data suggest only a ten-uous, usually second hand, exposure to the Wariheartland and its state centers. Three sherds and ahandful of obsidian flakes are the only goods thatmoved from the Wari core to Cotahuasi and vice ver-sa. In accordance with the radial core–peripherymodel, one could use this paucity of data, combined

with the Wari influence over ceramics, to perhapsposit ‘‘indirect control’’ or ‘‘influence’’ of Wari overthe valley. I argue that either attribution, however,would obscure the reasons behind the changes occur-ring in the valley during the Middle Horizon.

Although the Cotahuasinos were no doubt awareof the Wari state, the valley’s social horizon appearsto have largely widened only to encompass the regionin which they lived. Most of the trade in ceramics andgeologic resources occurred within the surroundingarea of Arequipa and Cuzco, most of the ceramicinfluence came from nearby styles, and the influenceson tablet motifs and tomb types came from neighbor-ing valleys. This is not to suggest that Wari did notplay a role in these developments. The intra-regionalrelationships were made possible by the expansion ofthe Wari state and increasing urbanism in the Waricore. The state’s increased demand for non-localproducts, and perhaps its sponsorship of long-dis-tance llama caravans, stimulated the expansion oflong-distance exchange in Cotahuasi and throughoutmuch of Peru.

Wari also stimulated developments in Cotahuasiby the introduction of ceramics styles with iconog-raphy that drew on images of widespread cosmolog-ical importance (Menzel, 1964). During periods ofintensified inter-regional interaction, styles associat-ed with elite, especially ritual, behaviors can quicklybecome widespread. These ‘‘international’’ stylesare often emulated as potters attempt to place theirwork within a shared cultural milieu (Schortmanet al., 2001, p. 325). Objects and their iconographycan often be significantly transformed as they areincorporated into new cultural settings (Dietler,1990; Saunders, 1999), and the lack of change inCotahuasi’s ritual practices suggest that Wari beliefshad no deep impact on local beliefs. Cotahuasi’semerging elites appear to have seized upon the ‘‘in-ternational’’ identity of the Wari state less in adesire to associate with a vaguely understood dis-tant state and more in a bid to forge a shared iden-tity with their neighbors. Wari images and ideas(and more rarely objects) became cultural capitalin the region (i.e., Bourdieu, 1977, pp. 183–184) thatsignaled to followers ones connection to the outsideworld (e.g., Helms, 1988, 1998).

Wari and sociopolitical changes in the Cotahuasi

Valley

Shifts in the exchange of goods and ideas, ofcourse, are only part of the changes that occurred

Fig. 9. Locations of villages during the Inchoate Period.

5 Paul Trawick, a cultural anthropologist, documented nine


in the Cotahuasi Valley during the Middle Horizon.Population also significantly expanded, agriculturalproduction intensified, social stratificationincreased, and a site called Collota was built withWari-style architecture. These changes could alsobe used to argue for ‘‘direct control’’, ‘‘indirect con-trol’’, or ‘‘influence’’ within the Wari state mosaic ofcontrol. Evaluating evidence from the valley’s settle-ment patterns and from Collota, however, demon-strates how the changes that occurred are betterexplained outside of a radial core–periphery model.

My survey crew and I identified 10 small Incho-ate Period villages (1–3 ha) scattered across the val-ley (Fig. 9).4 During the Middle Horizon, thepopulation of the valley substantially increased(Fig. 10). All settlements from the Inchoate Period

4 Since the valley’s extreme topography restricts the amount ofland available to construct villages, many of Cotahuasi’s sitesremained in the same locations over thousands of years. This sitestability makes it difficult to interpret Inchoate Period settlementpatterns because Inchoate Period levels have been obliterated orobscured by subsequent occupations that were larger and denser.Nonetheless, Inchoate Period site size can be tentatively estimat-ed based on the surface scatter of ceramics. Although sizeestimates based on surface sherds can be problematic (Lewarchand O’Brien, 1981), excavation data from Ancient Alca correlat-ed well with surface remains.

at least doubled in size, eight new villages werefound in our survey area, and seven other new Mid-dle Horizon sites are known from other survey workin the lower reaches of the valley (Trawick, 2003,pp. 40–41, 50).5 The location of the Inchoate Periodsettlements suggests that the villagers were agro-pastoralists that cultivated crops in the lower reach-es of the valley and grazed camelids in the valley’supper reaches. The locations chosen for new MiddleHorizon sites were similar to where older sites wereplaced and suggest the maintenance of an agro-pas-toral lifestyle. Nonetheless, the increases in the

prehispanic sites in the lower reaches of the valley during hisfieldwork on the history of Cotahuasi’s water management(Trawick, 1994, 2003, p. 91, p. 50). Eight of these sites fall outsideof our survey area, and seven of these sites were habitation sites.Trawick collected ceramics from each of the sites, and had thesherds analyzed by archaeologists Neira Avendaño and Pablo dela Vera Cruz (Trawick, 1994, p. 71). The sherds that Trawickcollected date from the Middle Horizon to the Late Horizon.Since Inchoate Period ceramics were undefined at the time ofTrawick’s collection, it is possible that some or all of his sites alsodate to this period. With this caution in mind, I include Trawick’ssites in Fig. 10 only to give a preliminary impression of thesettlement pattern in the lower reaches of the valley.

Fig. 10. Location of villages during the Middle Horizon. Sites in white were documented during my survey and sites in black weredocumented by Paul Trawick during his work in the valley.


valley’s population likely required an intensificationof agriculture.

The earliest sherds found on the surface of agri-cultural terraces in the valley date to the MiddleHorizon. A Middle Horizon date for the terraceswould conform to dates for terracing in the nearbyColca Valley (Deneven, 1986), and some scholarssuggest that terracing was first introduced through-out the Andes during this period (Goldstein, 2005,pp. 166–167; Williams, 2002). The introduction ofthis technology would have allowed for large areasof the valley to be cultivated for the first time. Atthe same time, inter-regional interaction likelyincreased the exchange of plants throughout theAndes and led to the introduction of higher yieldingvarieties in Cotahuasi from the coast and otherareas of Peru. If terracing and new crop varietiesled to agricultural intensification, then increases inhousehold fertility rates likely led to the valley’spopulation expansion during the Middle Horizon(i.e., Kramer and Boone, 2002).

The first evidence for emerging social stratifica-tion in Cotahuasi also comes during the MiddleHorizon. Each village that is preserved well enoughto examine its overall layout had a group of build-ings that can be separated by their size and architec-

tural elaboration from the rest of the site. Thebuildings’ form, associated assemblages, and loca-tion on the site suggest that the structures may bethe remains of elite houses. Unfortunately, it is dif-ficult to determine the date of construction for manyof these buildings, and some may date to the subse-quent Late Intermediate Period. Ceramics fromexcavations and wall fill at the site of Ancient Alca,however, suggest that the group of elite houses atthis site dates to the Middle Horizon. Additionalevidence for emerging social stratification alsocomes from elite burials that can be distinguishedfrom others based on the quality of ceramics andtomb architecture. The tombs follow local funerarytraditions and usually contain locally made sherdsthat emulate Viñaque and Ccoscopa styles.

Despite the increase in population and the emerg-ing social hierarchy, there is no evidence for politicalcentralization of the valley. Instead of clusteringaround preexisting villages, new sites generallyspread out to take advantage of under-utilized land.Rank size plots from both periods suggest no polit-ical hierarchy between villages (Fig. 11). In eachperiod, the rank–size curve is convex in respect toits log-normal line, the straight lines moving fromleft-right. The convex rank–size lines indicate a

A

CB

D

Fig. 12. Construction sequence (A–D) for Wari sites.

7 In July of 2005, my colleagues and I began a three-yearexcavation project at Collota. While our work is ongoing, there is

Fig. 11. Rank-size for villages from the Inchoate Period andMiddle Horizon in the Cotahuasi Valley.


fairly even distribution of villages across the land-scape, and by extension, little to no political central-ization (e.g., Johnson, 1977, 1980). The A coefficientfor the rank–size curves, a mathematical character-ization of the shape of the curve that varies between1 and �1 (Drennan and Peterson, 2004), are virtual-ly identical (the A = .701 for the Inchoate Periodand A = .709 for the Middle Horizon), and furthersuggest that there were no significant difference incentralization between the periods.

Of the eight new domestic sites found during theMiddle Horizon in Cotahuasi, Collota6 was uniquebecause of its location along the river and its archi-tectural elements that were derived from the Wariarchitectural cannon. While these idiosyncrasieshave led to interpretations of the site as a Wariadministrative center, closer inspection of data fromthe site suggests that it was more likely a local siteestablished to profit off of inter-regional exchange(Jennings and Yépez Álvarez, 2001a,b). Collota’sarchitecture incorporated elements of the architec-ture typically found at Wari and its satellite centersoutside of Ayacucho. This architecture radicallydeparted from previous building traditions inCotahuasi and elsewhere (Isbell, 1991; Schreiber,1978; Spickard, 1983). The most salient differencewas a high degree of planning that contrasted sharp-ly with the typical architecture of the more organiccommunities in the central Andes. Wari sites were

6 In previous publications (Jennings and Yépez Álvarez,2001a,b), I have separated Collota into two sites, Collota andNetahaha. In this paper, I treat Netahaha as a sector (B) ofCollota.

typically composed of high, rectangular wallsenclosing internal cells made up of a central patiosurrounded by elongated rooms. Structures weregenerally constructed from the outside in—buildersfirst made the enclosure wall, they then divided theinterior into a series of quadrangular cells, and final-ly added walls, benches, platforms, and other fea-tures to each of the cells. Access into Warienclosures was usually limited to only one or twoentrances and movement through the structureand between cells was severely limited (Fig. 12).

Collota was a local site made to look Wari (Jen-nings and Yépez Álvarez, 2001a,b). The site’ssequence of construction, its stochastic variation inroom size, imprecision in measurements, conformityto natural topography, and overall initial layoutcleaves closely to local concepts of site planning.In later phases of construction (Fig. 13), the site’sarchitecture was brought into closer conformitywith Wari cannons by more clearly defining patiosand restricting access through the site.7 The site’shigh exterior enclosure walls, restricted access, andred and white painted interior walls were notabledepartures from local building tradition. These fea-tures may have been of the greatest importance toCollota’s builders because they were likely used asan attempt to signal the group’s privileged knowl-edge of Wari, and thus support an ‘‘international’’identity for the inhabitants.

Collota’s placement on the river bottom was alsounusual. The prevailing settlement logic frombefore, during, and after the Middle Horizon was

clear evidence for later occupations at the site. Some of thechanges in architecture that I identify as a ‘‘Warification’’ of thesite may therefore prove to date to later periods. Until furtherexcavations and analysis, however, I will retain the architecturalsequence detailed in earlier writings (Jennings and Yépez Álvarez,2001a,b).

Fig. 13. Construction sequence of enclosure A at the site of Collota in the Cotahuasi Valley.


to place settlements halfway up the valley’s flanks inorder to exploit spring-fed agricultural terraces inthe lower reaches of the valley and high grasslandgrazing areas in the upper reaches of the valley.From an agro-pastoral perspective, Collota waslocated in an awful place. The land around the sitewas not irrigated until the Inca conquest of theregion (Trawick, 1994, p. 81), the nearest terracedfields were almost 3 km away, and pastoral landwas 1400 m up. Grinding stones and culinary equip-ment attest to on-site food processing and prepara-tion at Collota, but residents likely relied on barterand kinship ties to maintain much of their food sup-ply. Collota was however placed in an excellentposition to monitor inter-regional trade.

Based on inspection of satellite images and con-versations with llama herders, there are presentlythree major entry points into the valley for llamacaravans (Fig. 14). The first two entry points werelikely part of an Inca road through the valley thatconnected Cuzco to the Pacific Ocean (Hyslop,1984; Von Hagen, 1955). The road appears to have

entered the valley above the modern town of Puica,descending the valley along its western side, crossedthe river near the modern town of Cotahuasi, andthen climbed out of the valley near the rock saltmine at Huarhua (Burger et al., 1998, pp. 193–194). In order to use the Inca route through the val-ley, a suspension bridge would have been requiredbecause the Huarcaya River has formed a deep nar-row cleft through this section of the river (seeFig. 14). Since suspension bridge technology wasunavailable during the Middle Horizon (Schreiber,1984, 1992), it is unlikely that this route throughthe valley was available when Collota was built.

The third major entry point into the valley isabove the modern town of Pampamarca. Unlikethe Huarhua route, travelers from Puica can crossthe river further upstream at a point where the riveris braided and more easily bridgeable. If the Puica–Pampamarca route was the major trade route dur-ing the Middle Horizon, then Collota was in an ide-al location. The site is a few hundred metersdownriver from the last bridgeable section of


Huarcaya River and was constructed in a narrowplain between the river’s edge and a cliff face(Fig. 15). Traders moving along the Puica–Pampa-marca route were funneled into this narrow plain,where they were forced to interact with the peopleliving in Collota.

Collota appears to have been constructed to bothassociate the residents with the Wari style and toprofit in some way from inter-regional exchange.A comparison of surface ceramics frequencies, how-ever, suggests that the site failed to fulfill either ofthese goals. As described earlier, Middle Horizonceramics in the Cotahuasi Valley can be divided intothree: Wari, regional, and local. Although mostWari and regional vessels were made locally, a high-er percentage at Collota of these wares would sug-gest a greater degree of contact with the Waristate and the outside world in general. This doesnot seem to be the case (Fig. 16). There are no sta-tistically significant differences between the frequen-cies of Wari and regional surface sherds at Collotaand the mean frequencies for all sites in the surveyarea.8 While statistical analyses of surface remains,of course, must be treated as tentative because oftaphonomic processes, these data suggest not onlythat Collota failed to have a stronger relationshipwith the outside world relative to rival villages,but that people in all of Cotahuasi’s villages wereactively emulating Wari and regional styles.9

The changes that occurred in the valley weremost likely the result largely of endogenous change.

8 Since sites have significantly different sherd counts basedprimarily on their state of preservation (see Fig. 16), I have mademy calculation based on stylistic frequencies. The mean frequencyof Wari sherds to total Middle Horizon sherds at each site in thevalley is 21%; the frequency is 17% at Collota (sites = 11,s = 0.153). The mean frequency of Ccoscopa sherds to totalMiddle Horizon sherds at each site in the valley is 22%; thefrequency is 4% at Collota (sites = 9, s = 0.228). Standarddeviations are large because of the small number of surfaceceramics collected at each site. For each calculation, there weresites that contained either all or no local sherds. These sites weretreated as outliers and eliminated from the frequency calculationsabove. Including the outliers only widens the standard deviationand does not change the general result.

9 Although one might suspect that imported vessels aredisproportionably found in elite burials and therefore more likelyto be looted, my field crew and I found no examples of vesselsimported from outside of the region in local collections.Moreover, we found no such vessels when we excavated threeintact collective tombs from Collota this summer. Of the morethan forty intact or nearly intact vessels recovered, none of thevessels are obvious exports from Wari or from other areas outsideof the region.

The introduction of terracing, and perhaps higheryield crop varieties, allowed for the intensificationof agriculture in the valley. Intensification led toincreased demands on household labor, and thisdemand fueled increased fertility rates. As existingvillages doubled in size, scalar stress increased tothe point where preexisting social mechanisms forgroup cohesion were insufficient (e.g., Friesen,1999). The reaction in the valley was twofold. Onthe one hand, larger villages ‘‘fissioned’’ to decreasesettlement size and new settlements were found inunderutilized areas of the valley (e.g., Bandy,2004). On the other hand, incipient stratificationbegan to occur as village size continued to creepupward (e.g., Johnson, 1982). Wari’s role in thesechanges was significant, but once again was limitedto the state’s role in stimulating the increase in inter-regional exchange of goods and ideas throughoutPeru during this period.

Increasing trade contacts presented opportunitiesfor Cotahuasi’s emerging elites both because of the‘‘international’’ appeal of Wari iconography and thepotential to control or profit on the goods flowingin, out, and through the valley. Wari iconographywere perhaps most important to Cotahuasinos.The styles were transformed into a regional culturalcapital that buttressed an inchoate ‘‘high culture’’(i.e., Baines and Yoffee, 1998, p. 235) that wasformed by constructing a regime of common differ-ence (i.e., Wilk, 2004). In other words, one’s statusbecame defined by one’s knowledge of Wari styles.By trading goods and knowledge with like-mindedindividuals in other valleys either directly orthrough llama caravan leaders, elites throughoutthe regional interaction network helped each othermaintain and increase their social positions. Thecloser knowledge over time of Wari and regionalstyles by potters from Cotahuasi and neighboringvalleys likely reflects increasing contact betweenregional elites (Sciscento, 1989; Wernke, 2003).

Collota was an attempt to create a new power-base that would out compete with rivals from othervillages because of its position along the major traderoute through the valley and its emulation of Wariarchitecture. Local groups across the ancient worldoften emulated state architecture in order toenhance their status (Schortman and Urban, 1998,p. 111; Wells, 1999, pp. 203–204; Whitehouse andWilkins, 1989, p. 109), but in this case Wari stylisticinfluence over Collota was most likely tied to anattempt to symbolize regional connections ratherthan state ones. Despite their efforts, Collota inhab-

0 20 kmN

NASA ASTER Satellite Imagery, Bands 1-3September 2000-May 2004

IncaIncaPre-Inca

A

B

C

D

E

F

Fig. 14. Major trade routes through the Cotahuasi Valley. The routes pass near the following places mentioned in the text: (A) Puica, (B)Ancient Alca, (C) Collota, (D) the town of Cotahuasi, (E) Pampamarca, and (F) Huarhua.

RiverCrossing

0 200 mNPeruvian Air Force Aerial Photograph, July 1955

Collota

Fig. 15. The location of Collota relative to the probable trade route through the valley during the Middle Horizon.


0

5

10

155

200

255

300

WariNon-LocalTotal

SHERD

COUNT

SITE NUMBERS2 6 11 13 19 20 21 22 23 26 27 29 33 36 37 39 45 47 50

Fig. 16. Middle Horizon ceramic style frequencies from eachvillage in the Cotahuasi Valley. Collota is site 45 (marked ingrey).


itants were never able to cultivate close relationshipswith the outside world. Other sites were probablyjust as effective, and perhaps more so, in their exter-nal dealings. Collota’s failings are probably relatedto the low volume of trade through the valley, itslack of a subsistence base, and its distance fromthe obsidian, copper, and rock salt deposits thatplayed an important role in inter-regional exchang-es. Collota was abandoned soon after the MiddleHorizon.10 Ultimately, land, kin, and other tradi-tional bases of power in Cotahuasi proved moreimportant than contacts with the outside world.

Core, periphery, and regional realities in Cotahuasi

Whether one envisions a group under the imperi-al boot, locked into unequal exchange relationships,dependent on prestige goods, or acting as an equaltrading partner, conceptions of state–local relation-ships are often informed by the ingrained geometryof the radial model of core–periphery systems (seeFig. 1). Like many archaeologists, my previous dis-cussions of the relationship between Wari andCotahuasi have weighed the evidence for differentkinds of core–periphery relationships and placedthe valley on a continuum of state direct control,indirect control, and influence (Jennings, 2002; Jen-nings and Craig, 2001; Jennings and Yépez Álvarez,2001a,b). None of these explanations, however, fit

10 Our ongoing excavations at Collota demonstrate that por-tions of the site were occupied during the Late Horizon. There isno clear Late Intermediate Period occupation of the site and itseems that the site was likely abandoned and later re-occupied.Further excavation and analyses, however, may reveal a LateIntermediate occupation of the site.

what happened in Cotahuasi because the radialmodel that structures these arguments significantlydistorts the reality of the regional interaction systemthat was spawned by the expansionary dynamics ofthe Wari state.

The changes in Middle Horizon Cotahuasi can-not be explained without the expansion of the Waristate. Wari’s expansion stimulated the growth ofinter-regional interactions across Peru and createda widely emulated artistic style that could be usedto legitimate elite positions. These developmentsallowed for Cotahuasi’s sociopolitical developmentto occur, but this development was based onincreasing contacts between Cotahuasi’s emergingelites and elites from other neighboring valleys.These contacts led to the exchange of ideas, and per-haps plant seeds, that allowed for agricultural inten-sification. This intensification led to populationincreases in villages that caused scalar stress. Thescalar stress led to greater intensification (as newareas came under cultivation), village fissioning,and the emergence of elites. Elites seized on intra-re-gional interaction as a means to legitimate their sta-tus primarily through the use of Wari influencedstyles as cultural capital that buttressed a systemof common difference. The founders of Collotaattempted to further capitalize on this cultural cap-ital by both situating themselves along the majortrade route through the valley and emulating Wariarchitecture. This bid ultimately failed because thepeople of Collota were too far removed fromagro-pastoral subsistence zones.

A core–periphery model could be applied toCotahuasi. In a sense, Cotahuasi was enmeshed inWari’s inter-regional exchange network, its indirectadministrative structure, and its prestige goodseconomy. It would be a mistake however to concep-tualize the Cotahuasi data in this matter because wewould force an interpretive structure that I argueruns counter to the data (see Fig. 5). For example,we might privilege certain aspects of the ceramicdata to suggest a ‘‘strong’’ relationship with thecore, or privilege other aspects to argue that therelationship was ‘‘weak.’’ We could note the pres-ence of state goods in Cotahuasi and Cotahuasigoods in the core and suggest that the area was ‘‘in-directly controlled’’, or suggest that the low volumeof goods circulated between core and periphery isbest explained by core ‘‘influence’’ over the periph-ery. Using a radial model to interpret Middle Hori-zon Cotahuasi ultimately fails because the model ismeant to highlight a kind of core–periphery


relationship that did not exist in the valley. Instead,Wari expansion stimulated the growth of regionalinteraction system that appears to have operatedlargely independently from the state.

André Gunder Frank, Immanuel Wallerstein,and their colleagues developed core–peripherymodels to understand the structures of integratedinter-regional economic systems in the modernworld (Rowlands, 1987, p. 4). Archaeologists havevigorously debated many aspects of the model,such as the strength of core dominance, the sizeof the system, the types of goods flowing to differ-ent parts of the system, and the relative depen-dence or independence of the periphery. Thesedebates have been fruitful, but have largely playedout within the deeply embedded structure of core–periphery models. While archaeologists have longcommented on the importance of interactionspheres in the emergence of early states (Yoffee,2005, p. 230), there has been far less concernabout regionalism in the expansion of states. Thisneglect is a product of a radial model that leaveslittle space for such intra-regional interactions. Inthis case study of Middle Horizon Cotahuasi, Idemonstrate one instance where the emergentproperties of a regional system embedded withinthe cultural and economic milieu of the Wari statehave gone unrecognized. Without a vigorous reas-sertion of regional dynamics into our interactionmodels, the realities of life in Cotahuasi and other‘‘peripheral’’ places around the world will remainpoorly understood.

Acknowledgments

The ideas in this paper benefited enormouslyfrom conversations with students and facultymembers at the University of California, SantaBarbara and Franklin and Marshall College. Ithank George Herbst, Ian Lindsey, AdrienneRand, and James Tate for reading drafts of thispaper, the people of Cotahuasi for their hospital-ity, and the members of the project (Amelia Argü-elles Talavera, Clarence Bodmer, Forrest Cook,Michael Hendrik, Kelly Knudson, Klarissa More-na, Gregory Mazzeo, Cecilia Quequezana Lucano,Hendrik Van Gijseghem, and Willy Yépez Alva-rez). This research was done with a grant fromthe National Science Foundation (Award No.9903508) and with the permission of the PeruvianInstituto Nacional de Cultura (Resolución Direc-toral Nacional Nro. 977/INC).

References

Adams, W., 1984. The first colonial empire: Egypt in Nubia,3200–1200 BC. Comparative Studies in Society and History26, 36–71.

Algaze, G., 1993a. Expansionary dynamics of some early pristinestates. American Anthropologist 95 (2), 304–333.

Algaze, G., 1993b. The Uruk World System. University ofChicago Press, Chicago.

Algaze, G., 2001. The Prehistory of imperialism: the case of UrukPeriod Mesopotamia. In: Rothman, M.S. (Ed.), Uruk Mes-opotamia and Its Neighbors: Cross Cultural Interactions inthe Era of State Formation. School of American Research,Santa Fe, pp. 27–84.

Baines, J., Yoffee, N., 1998. Order, Legitimacy, and Wealth inAncient Egypt and Mesopotamia. In: Feinman, G., Marcus,J. (Eds.), Archaic States. School of American Research, SantaFe, pp. 199–260.

Bandy, M.S., 2004. Fissioning, scalar stress, and social evolutionin early village societies. American Anthropologist 106 (2),322–333.

Bauer, B.S., 1999. The Early Ceramics of the Inca Heartland.Fieldiana Anthropology, New Series 31. Field Museum ofNatural History, Chicago.

Bawden, G., Conrad, G.W., 1982. The American Heritage.Peabody Museum Press, Cambridge.

Berdan, F.F., Blanton, R.E., Boone, E.H., Hodge, M.G., Smith,M.E., Umberger (Eds.), 1996. Aztec Imperial Strategies.Dumbarton Oaks, Washington.

Blanton, R.E., Feinman, G., 1984. The Mesoamerican worldsystem. American Anthropologist 86, 673–682.

Bourdieu, P., 1977. Outline of a Theory of Practice. CambridgeUniversity Press, Cambridge.

Bradley, R., 1998. The Significance of Monuments: On theShaping of Human Experience in Neolithic and Bronze AgeEurope. Routledge, New York.

Browman, D., 1999. Wari impact on the Upper Mantaro Basin.Paper presented at the 64th Annual Meeting of the Society forAmerican Archaeology.

Burger, R.L., Asaro, F., 1979. Análisis de rasgos significativos enla obsidiana de los Andes centrales. Revista del MuseoNacional 43, 281–326.

Burger, R.L., Asaro, F., 1993. La distribución y procedencia deartefactos de obsidiana durante el periodo inicial y horizontetemprano. In: Burger, R.L. (Ed.), Emergencia de la civiliza-ción en los Andes: ensayos de la interpretación. UniversidadNacional Mayor de San Marcos, Lima, pp. 189–231.

Burger, R.L., Asaro, F., Trawick, P., Stross, F., 1998. The alcaobsidian source: the origin of raw material of cusco typeobsidian. Andean Past 5, 185–202.

Burger, R.L., Mohr Chávez, K.L., Chávez, S.J., 2000. Throughthe glass darkly: prehispanic obsidian procurement andexchange in southern Peru and northern Bolivia. Journal ofWorld Prehistory 14, 267–362.

Canchaya, A., Aranda, A., Guevara, T., 1995. Mapa metalogénicodel Perú. Instituto Geológico Minero y Metalúrgico, Lima.

Cardona Rosas, A. 1993. Caracterı́sticas geográficas del Patrónde asentamiento para el valle de Chuquibamba-Arequipadurante el Perı́odo Comprendido entre el Horizonte Medio, yel Horizonte Tardı́o. Unpublished thesis presented for thelicense of archeology, Universidad Católica Santa Marı́a,Arequipa.


Cardona Rosas, A., 2002. Arqueologı́a de Arequipa: de susalbores a Los Incas. Magazine Impresores, Arequipa.

Champion, T.C. (Ed.), 1989. Center and Periphery: ComparativeStudies in Archaeology. Rouledge, New York.

Chase-Dunn, C., Hall, T.D. (Eds.), 1992. Core/Periphery Rela-tions in Precapitalist Worlds. Westview Press, Boulder.

Chase-Dunn, C., Hall, T.D., 1997. Rise and Demise: ComparingWorld-Systems. Westview Press, Boulder.

Chávez Chávez, J.A., 1982. Evidencias arqueológicas en laCuenca del Rı́o Cotahuasi-Ocona. Unpublished thesis pre-sented for the license of archeology Universidad Nacional deSan Agustin, Arequipa.

Cook, A.G., 2001. Huari D-Shaped Structures, Sacrificial Offer-ings, and Divine Rulership. In: Benson, E.P., Cook, A.G.(Eds.), Ritual sacrifice in Ancient Peru. University of Texas,Austin, pp. 137–163.

Cook, A.G., Glowacki, M., 2003. Pots, politics, and power: huariceramic assemblages and imperial administration. In: Bray,T.L. (Ed.), The Archaeology and Politics of Food andFeasting in Early States and Empires. Kluwer Academic/Plenum, New York, pp. 173–202.

Cusick, J.G. (Ed.), 1998. Studies of Culture Contact: Interaction,Culture Change, and Archaeology. Southern Illinois Univer-sity, Carbondale.

Deagen, K., 1996. Colonial transformations: Euro-Americancultural genesis in the Early Spanish-American colonies.Journal of Anthropological Research 52, 135–160.

Deagen, K., 1998. Transculturation and Spanish Americanethnogenesis: the archaeological legacy of the quincentenary.In: Cusick, J.G. (Ed.), Studies of Culture Contact: Interac-tion, Culture Change, and Archaeology. Southern IllinoisUniversity, Carbondale, pp. 23–43.

Denemark, R.A., Friedman, J., Gills, B.K., Modelski, G., 2000.An introduction to world-system history: towards a socialscience of long-term change. In: Denemark, R.A., Friedman,J., Gills, B.K., Modelski, G. (Eds.), World System History:The Social Science of Long-term Change. Routledge, NewYork, pp. xv–xxii.

Deneven, W. (Ed.), 1986. The Cultural Ecology, Archaeology,and History of Terracing and Terrace Abandonment in theColca Valley of Southern Peru. Department of Geography,University of Wisconsin, Madison.

Dietler, M., 1990. Driven by drink: the role of drinking in thepolitical economy and the case of Early Iron Age France.Journal of Anthropological Anthropology 9, 352–406.

Dietler, M., 1998. Consumption, agency, and cultural entangle-ment: theoretical implications of a Mediterranean colonialencounter. In: Cusick, J.G. (Ed.), Studies of Culture Contact:Interaction, Culture Change, and Archaeology. SouthernIllinois University, Carbondale, pp. 288–315.

Dietler, M., Herbich, I., 1998. Habitus, techniques and style: anintegrated approach to the social understanding of materialculture and boundaries. In: Stark, M.T. (Ed.), The Archae-ology of Social Boundaries. Smithsonian Institution Press,Washington, pp. 232–263.

Domı́nguez, A.J., 2002. Greeks in Iberia: colonialism withoutcolonization. In: Lyons, C.L., Papadopoulos, J.K. (Eds.), TheArchaeology of Colonialism. Getty Research Institute, LosAngeles, pp. 65–95.

Drennan, R.D., Peterson, C.E., 2004. Comparing archaeologicalsettlement systems with rank–size graphs: a measure of shape

and statistical confidence. Journal of Archaeological Science31, 533–549.

Earle, T., D’Altroy, T., 1989. Political economy of the Inkaempire: the archaeology of power and finance. In: Lamberg-Karlovsky, C.C. (Ed.), Archaeological Thought in America.Cambridge University Press, Cambridge, pp. 183–204.

Ekholm, K., Friedman, J., 1979. ‘‘Capital’’ imperialism andexploitation in ancient world systems. In: Larsen, M. (Ed.),Power and Propaganda. Akademish, Copenhagen, pp. 41–58.

Frank, A.G., 1966. The Development of underdevelopment.Monthly Review 18, 17–31.

Frank, A.G., 1967. Capitalism and Underdevelopment in LatinAmerica: Historical Studies of Chile and Brazil. MonthlyReview Press, New York.

Frank, A.G., 1993. The World System: Five Hundred Years orFive Thousand. Routledge, London.

Frank, A.G., Fills, B.K., 2000. The five thousand year worldsystem in theory and praxis. In: Denemark, R.A., Friedman,J., Gills, B.K., Modelski, G. (Eds.), World System History:The Social Science of Long-Term Change. Routledge, NewYork, pp. 3–23.

Franquemont, E.M., 1990. The Ancient Pottery from Pucara,Peru. Ñawpa Pacha 24, 1–30.

Friesen, T.M., 1999. Resource Structure, Scalar Stress, and theDevelopment of Inuit Social Organization. World Archaeol-ogy 31 (1), 21–37.

Geertz, C., 1973. The Interpretation of Cultures: Selected Essays.Basic Books, New York.

Giddens, A., 1984. The Constitution of Society. University ofCalifornia Press, Los Angeles.

Glowacki, M., 2005. Pottery from Pikillacta. In: McEwan, G.F.(Ed.), Pikillacta: The Wari Empire in Cuzco. University ofIowa Press, Iowa City, pp. 101–114.

Glowacki, M., McEwan, G., 2001. Pikillacta, Huaro, y la GranRegión del Cuzco: Nuevas Interpretaciones de la OcupaciónWari de la Sierra Sur. In: Kaulicke, P., Isbell, W.H. (Eds.),Huari y Tiwanaku: Modelos y Evidencias, Secundo Parte.Boletı́n de Arqueologı́a PUCP, No. 5.. Fondo Editorial de laPontificia Universidad Católica del Perú, Lima, pp. 31–49.

Goldstein, P.S., 2005. Andean Diaspora: The Tiwanaku Coloniesand the Origin of South American Empire. University Pressof Florida, Gainesville.

Gonzalez Carre, E., 1981. La Antigua Ciudad de Wari enAyacucho. Boletı́n de Lima 16–18, 83–97.

Gosden, C., 2004. Archaeology and Colonialism: CulturalContact from 5000 BC to the Present. Cambridge UniversityPress, Cambridge.

Grossman, J., 1983. Demographic Change and EconomicTransformation in the South-Central Highlands of Pre-HuariPeru. Ñawpa Pacha 21, 45–126.

Hall, T.D., Chase-Dunn, C., 1993. The world-systems perspectiveand archaeology: forward in to the past. Journal of Archae-ological Research 1 (2), 121–143.

Hastorf, C., 1993. Agriculture and the Onset of PoliticalInequality before the Inka. Cambridge University Press,New York.

Helms, M., 1988. Ulysses’ Sail: An Ethnographic Odyssey ofPower, Knowledge, and Geographical Distance. PrincetonUniversity Press, Princeton.

Helms, M., 1998. Access to Origins: Affines, Ancestors, andAristocrats. University of Texas Press, Austin.


Hyslop, J., 1984. The Inka Road System. Academic Press, NewYork.

Isbell, W.H., 1977. The Rural Foundations of Urbanism.University of Illinois, Chicago.

Isbell, W.H., 1988. City and state in Middle Horizon Wari. In:Keatings, R.W. (Ed.), Peruvian Prehistory. Cambridge Uni-versity Press, Cambridge, pp. 164–189.

Isbell, W.H., 1989. Honcopampa: Was it a Huari administrativecenter? In: Czwarno, R.M., Meddens, F.M., Morgan, A. (Eds.),The Nature of Wari. BAR International Series 525. BritishArchaeological Report Publications, Oxford, pp. 98–114.

Isbell, W.H., 1991. Huari administration and the orthogonalcellular architecture horizon. In: Isbell, W.H., McEwan, G.F.(Eds.), Huari Administrative Structures. Dumbarton Oaks,Washington, DC, pp. 293–316.

Isbell, W.H., 1997. Reconstructing huari: a cultural chronologyfor the capital city. In: Manzanilla, L. (Ed.), Emergence andChange in Early Urban Societies. Plenum Press, New York,pp. 181–227.

Isbell, W.H., McEwan, G.F., 1991. A history of huari studies andintroduction to current interpretations. In: Isbell, W.H.,McEwan, G.F. (Eds.), Huari Administrative Structures.Dumbarton Oaks, Washington, DC, pp. 1–17.

Isbell, W.H., Schreiber, K.J., 1978. Was Huari a state? AmericanAntiquity 43 (3), 372–389.

Jennings, J. 2002. Prehistoric Imperialism and Cultural Devel-opment in the Cotahuasi Valley, Peru. Unpublished Ph.D.Dissertation, University of California, Santa Barbara.

Jennings, J., 2003a. The fragility of imperialist ideology and theend of local traditions, an Inca example. Cambridge Archae-ological Journal 13 (1), 107–120.

Jennings, J., 2003b. Inca imperialism, ritual change, and cosmo-logical continuity in the Cotahuasi Valley, Peru. Journal ofAnthropological Research 59 (4), 433–462.

Jennings, J., in press. Understanding Middle Horizon Peru:Hermeneutic Spirals, Interpretative Traditions, and WariAdministrative Centers. Latin American Antiquity 16.

Jennings, J., Craig, N., 2001. Politywide analysis and imperialpolitical economy: the relationship between valley politicalcomplexity and administrative centers in the Wari Empire ofthe Central Andes. Journal of Anthropological Archaeology20, 479–502.

Jennings, J., Glascock, M., 2002. Description and Method ofExploitation of the Alca Obsidian Source, Peru. LatinAmerican Antiquity 13 (1), 107–117.

Jennings, J., Yépez Álvarez, W., 2001a. Collota, Netahaha, y eldesarrollo del poder Wari en el valle de Cotahuasi, Arequipa,Perú. In: Kaulicke, P., Isbell, W.H. (Eds.), Huari y Tiwanaku:Modelos y Evidencias, Secundo Parte. Boletı́n de Arqueologı́aPUCP, No. 5. Fondo Editorial de la Pontificia UniversidadCatólica del Perú, Lima, pp. 13–29.

Jennings, J., Yépez Álvarez, W., 2001b. Architecture, local elites,and imperial entanglements: the impact of the Wari Empireon the Cotahuasi Valley of Peru. Journal of Field Archaeol-ogy 28 (1/2), 143–159.

Johnson, G.A., 1977. Aspects of regional analysis in archaeology.Annual Review of Anthropology 6, 479–508.

Johnson, G.A., 1980. Rank-size convexity and system integration:a view from archaeology. Economic Geography 56, 234–247.

Johnson, G.A., 1982. Organizational structure and scalar stress.In: Renfrew, C., Rowlands, M., Seagraves, B.A. (Eds.),

Theory and Explanation in Archaeology. Academic Press,New York, pp. 389–421.

Kardulias, P.N. (Ed.), 1999. World Systems Theory in Practice:Leadership, and Exchange. Rowman and Littlefield, NewYork.

Kauffmann-Doig, F., 1991. Pinturas mágicas sobre placas decerámica (Chucu/Condesuyos, Arequipa). Arqueológicas 21.Instituto de Investigaciones Antropológicas, Lima.

Kohl, P., 1978. The balance of trade in southwest Asia in mid-third millennium B.C. Current Anthropology 19, 463–492.

Kohl, P., 1987. The use and abuse of world systems theory: thecase of the ‘‘pristine’’ west Asian state. Advances in Archae-ological Method and Theory 11, 1–35.

Kramer, K.L., Boone, J.L., 2002. Why intensive agriculturistshave higher fertility: a household energy budget approach.Current Anthropology 43 (3), 511–517.

Lau, G.F., 2002. Feasting and ancestor veneration at chinchawas,north highlands of ancash, Peru. Latin American A

core, peripheries, and regional realities in middle horizon peru · 2020. 3. 26. · core,...

Documents