what was tiwanaku?. 2013. charles stanish. in visions of tiwanaku. edited by a. vranich and c....

9

WHAT WAS TIWANAKU?

CHARLES STANISH

151

The culture of Tiwanaku represents one of thegreat civilizations of the ancient world. It iseasily on par in size, complexity, and sophis-

tication with the more well-known civilizations ofthe Near East, Mediterranean, and Asia. In fact, Ti-wanaku eclipses many other famous civilizations ofthe classical world in many aspects of interest here—physical size and architectural complexity of the cap-ital, population size, regional dominance, economicproduction, long-distance exchange, and the com-plexity of its sociopolitical organization.

In spite of this, some scholars still do not recog-nize what is obvious to archaeologists working else-where in the world who review the data—that Tiwa -naku was a state society with all of the attributescommon to the great ancient civilizations of the restof the world. The reasons for these views are manyand all are certainly legitimate, but several stand out.First, the traditions of scholarship in the Andes thatwere set up in the 1960s, as so nicely described byMoseley in his paper in this volume, have inadver-tently come together to diminish the achievementsof Andean civilizations such as Tiwanaku. This is re-inforced by a lack of texts for cultures such as Ti-wanaku. Furthermore, anthropological archaeolo-gists working in the Andes have adopted what I calla straw-man model for premodern states. The expec-tations of this model can never be met because they

are unrealistic and do not match any known state inworld history. As a result, an analysis of Tiwanakuusing the criteria found in this model will alwaysconclude that it was not a complex, archaic state. Fi-nally, until recently we have had few modern archae-ologists of indigenous descent who have worked onTiwanaku. In this paper I will outline the latest datafor Tiwanaku and try to place it in a broader com-parative and intellectual context.

RESEARCH TRADITIONS

Andeanists can understandably be a somewhat insu-lar group, working in an area that in sheer distancealone would touch three areas of Old World first-generation state development (from the Nile to theIndus). As a result, it is not often that Andeanistsrefer in any depth to other cultures around the worldthat are structurally similar to those found in west-ern South America. Most Andeanists are surprised tofind that the iconic “city” center of Mycenae in theancient Greek Peloponnese—the “city of kings”; theplace founded by Perseus, husband of Andromedaand father of Agamemnon and Menelaus; home topalaces, guardhouses, and storehouses; and a princi-pal city in the Trojan War—was, at its height, lessthan 3 ha in size. That is correct—3 ha. The architec-tural core of Mycenae fits inside the area bounded by

the Sunken Court, Akapana, and Kalasasaya in Tiwa - naku. The total, maximum size of the “city” of Myce-nae was no more than 75 ha, including all surround-ing domestic settlement and cemeteries. Aegeanarchaeologist Todd Whitelaw (2001:29) actually es-timates it at a mere 40 ha. The entire settlementcomplex of Mycenae was about half the size of justthe architectural core of Tiwanaku. In short, Tiwa -naku is at least six times larger than the settlementof Mycenae and is comparable in size to the greatMinoan Crete cities of the Bronze Age. Whitelaw(2001:29) estimates the maximum size of Knossos at80 ha, about 7.5 times smaller than that currentlyproposed for Tiwanaku. Tiwanaku, by any compari-son to many other ancient cities, is as architecturallycomplex and is in the general site range of non -imperial archaic states in most areas of the world.1

There are no historical documents available forthe Middle Horizon and only vague references to theLate Intermediate periods. The lack of documentsrepresents a huge difference between the Andes andmany parts of the Old World, particularly the NearEast and Mesopotamia. At Mycenae, for instance, wehave the great Homeric epics to tell us that King Aga -memnon led armies of thousands to Troy. We haveLinear B fragments that describe the administrationof state societies in Bronze Age times. If we had com-parable documents for the culture of Tiwa naku, per-haps one that told of the exploits of the legendaryKing Taypihuanca, who with his gold-studdedscepter led his army and navy to the north, defeatedthe combined forces of the swift-footed Taraceñosand Arapans, and returned with the great trophy ofthe Thunderbolt Stela on a huge triumphal marcharound the lake, we would perhaps be more sympa-thetic to viewing Tiwanaku in the same way we viewcomparable Old World civilizations.

As mentioned, we unfortunately have few schol-ars of indigenous descent who work on Andean ar-chaeology. One can imagine the frustration of some-one who has a genealogical or emotional connectionwith Tiwanaku culture and who has to live with thedizzying array of social and political tensions thatswirl around this magnificent civilization. Here is anancient society with a capital as large as or largerthan any Mycenaean or Minoan Aegean Bronze Agecenter and comparable in size to the city of Londonin the sixteenth century. It has demonstrated influ-ence around a vast area from desert to forest, created

beautiful works of art on a massive scale, carvedhuge stelae with stones dragged from 20 km away,and built roads, temples, and palaces, and in spite ofall this evidence, foreign scholars simply will notgive their ancestors credit for creating a civilizationon par with those of the Mediterranean, Mesoamer-ica, and Mesopotamia. I am firmly convinced that ifmore indigenous peoples were involved in the writ-ing of Tiwanaku history, the archaeological fairy talesof a vast empty ceremonial center would disappearas fast as they were created.

These factors have combined to create a verycurious research tradition in the region that has dis-engaged the study of Tiwanaku and other Andeancultures from comparative anthropological and his-torical analysis. Some scholars think that this is ap-propriate and that “lo Andino” should be the basisfor our studies. In my opinion, such disengagementdoes us a disservice and opens the doors to archae-ological flights of fancy similar to those of the gen-eration of Mayanists who argued for the “peaceful”Mayans who built “empty ceremonial” cities like Qui -rigua and Tikal. Archaeological interpretation mustbe grounded in something real. If it is not groundedin the data of history and ethnography utilizing rigidcriteria of verification, then it will be no more thanjust-so stories that reinforce some particular socialor academic fad at any moment.

Within this tradition of research in the Andes,perhaps the greatest theoretical and methodologicalflaw is to set up a straw-man argument in the formof the traditional state model of complex, coercive,stratified societies and to then conclude that Tiwa -naku did not fit this model. I will argue below thatthis conception of hierarchical states is fatally flawed.It is a straw-man argument that does not representreality anywhere on the planet, anytime in history.Once the straw-man model is rejected, the way isthen open to the most creative models imaginable.Some Andean scholars, for instance, pick and chooseelements from Tiwanaku’s great religious traditionsand architectural feats to argue that there were nopolitical hierarchies. From a comparative anthropo-logical and historical point of view, that is, in all duerespect, an untenable theoretical position. No soci-ety of any size and complexity—certainly one thatwas capable of building a few square kilometers ofplanned urban space out of hewn volcanic rocks—has ever existed without some kind of hierarchical

VISIONS OF TIWANAKU152

political structure that mobilized and organizedhuman labor. Some scholars reduce this great capitalcity of a great ancient state to a place where peacefulpeasants came together to eat, drink, dance, and re-inforce social bonds. The reification of contempo-rary Western spiritual fads in scholarly work is in-deed intriguing from the perspective of intellectualhistory but is at odds with the empirical data fromTiwanaku as well as from systematic comparativeanalysis of other states of similar size and complexityfrom the rest of the ancient world.

In this paper, I will first try to briefly summarizewhat we know, empirically, about Tiwanaku. Second,I will compare some of these empirical facts withother historically documented civilizations fromother areas of the world. Ultimately, I conclude thatwhile Tiwanaku, like all ancient states, had uniquecharacteristics, it is structurally similar to the classi-cal nonimperial civilizations in the Old World as wellas to those in the Americas such as Teotihuacán,Tula, and the classic Maya states. To be sure, it is notan empire like that of the Inca. But it was certainly acentralized and complex political entity that we rec-ognize as a first-generation state with a dominantelite and many other powerful groups that interactedin a myriad of ways.

Tiwanaku, like all ancient states, was an urban-ized polity that created a huge city center for eco-nomic, political, social, religious, and cultural activi-ties and expanded its influence, albeit selectively,over a vast area. Tiwanaku was not just a ritual gath-ering spot, a big place to throw ideologically chargedparties, an empty religious center, an astronomicalobservatory, or a pilgrimage destination. Like everygreat capital among the world’s civilizations, it hadelements of most or all of these, sustained by a polit-ical and economic organization that produced, ex-changed, and created valuable resources, backed by amilitary and religious elite that created Tiwanaku’spolitical ideology and cultural values and offeredthem to, promoted them to, or forced them on peo-ples who lived in an area about the size of the mod-ern U.S. state of California. And it successfully didthis for almost a half millennium.

WHATWE KNOW NOW

Archaeologists working diligently over the past fewgenerations have unearthed the cultures of Tiwa -

naku. Unlike its counterparts in the ancient classicalworld, there are no surviving documents fromTiwa naku times. As a result, we do not have the richnarratives that bring to life the sophistication andcomplexity of these cultures with ease. We have towork much harder to unlock the secrets of ancientTiwa naku using all the tools, both theoretical andmethodological, at our disposal. By combining state-of-the-art fieldwork with sophisticated theoreticalwork, we are able to deduce the broad structure ofTiwanaku society. The results of some of this workcan be summarized into a series of empirical obser-vations:1. Tiwanaku was an urbanized, class-based society,

centered in a huge city located on the altiplanoin the eponymous site in Bolivia.

2. Tiwanaku had a large, permanent resident pop-ulation that numbered at least 20,000 and prob-ably much more. There was an additional ruralpopulation in the Tiwanaku Valley as well aspeople living in quasi-urban sites such as Lukur-mata on the Taraco Peninsula. The total popula-tion for the Tiwanaku polity most likely reachedup to six figures.

3. Tiwanaku maintained colonies on a large scale,the only one seriously studied to date being theOmo complex in Moquegua. Others include theAzapa Valley in northern Chile and Cocha -bamba in Bolivia.

4. Tiwanaku maintained long-term and long-dis-tance trade relationships with autonomous andsemiautonomous groups throughout a vast areain a dizzying array of ecological zones. The mostfamous of these is San Pedro de Atacama, wherelocal elites adopted Tiwanaku accoutrementsand maintained some kind of economic traderelationship.

5. The city of Tiwanaku hosted a class of expertcraft specialists, mainly in architecture and arti-san goods such as pottery, metal, and cloth. Ti-wanaku artisans engaged in commodity produc-tion in pottery and cloth and possibly in otherobjects as well.

6. Tiwanaku artists drew off of a millennium oftradition, borrowing and reinterpreting Chavín,Pucará, and other highland cultures in theAndes. These Tiwanaku artists created works ofunprecedented sophistication and beauty, most

CHAPTER 9: WHAT WAS TIWANAKU? 153

notably in megalithic stone, textiles, and potterybut also in bone and metal.

7. Tiwanaku architects created a planned, urbanspace that embodied a number of ideologicaland/or religious constructs unique to its culture,creating a monument of unprecedented sophis-tication. It was almost certainly a great pilgrim-age destination, bringing in people from aroundthe south-central Andes for short and possiblysemi-permanent residence.

POPULATION ESTIMATES FOR TIWANAKU

The size of the population at Tiwanaku has alwaysbeen a difficult and contentious issue, yet it is onethat we cannot dodge because of its theoretical sig-nificance. The population estimates for Tiwanakuvary widely, based upon the assumptions one uses incalculating densities, longevity of sites, the area cov-ered, and so forth.

The first issue is to define precisely the area thatwe are calling Tiwanaku. We can usefully define fourareas of the Tiwanaku phenomenon that can be usedto propose population levels. First, there is the archi-tectural core itself, comprising a bit over 1 km2. Sec-ond is the surrounding area of a few square kilome-ters where mounds and surface debris indicate somekind of domestic occupation well beyond the core.The third area can be defined as the Tiwanaku Valleyitself, from roughly 2 km east of the Akapana andwest all the way to the lakeshore some 17 km away.The fourth area would be what we can call greaterTiwanaku (Figure 9.1). This would include the Tiwa -naku Valley, the Katari drainages/Pampa Koani, andthe northern Desaguadero area. This area is basedupon a one day’s walk to the city, about a 25-km ra-dius from the Akapana.2

Archaeologists have used a variety of soundmethods to calculate the population of Tiwanaku.One approach is to deduce the carrying capacity ofthe land and to derive minimum and maximum fig-ures for the population at its height. One then makesadjustments for various factors, usually bringing theestimate toward the lower end. A second approach isempirical. We measure the known areas of residenceas determined by survey and excavation and thenproject out to the rest of the area in question. A thirdapproach is to use comparative data from citiesaround the Andes, or even the ancient world, whose

political and economic characteristics are similar tothose of Tiwanaku. Finally, historical data on popu-lations in the region can be used as baselines to makemeaningful projections.

Each of these approaches has strengths andweaknesses. The deductive approach that derivescar rying capacity suffers from what all models areplagued with: slight variations in underlying as-sumptions can lead to very wide variations in modeloutcomes. The empirical approach is plagued byhuge gaps in data, so the projections onto unsur-veyed areas are also subject to wide variation. Thereis also the problem of contemporaneity of sites andmultiple residences. The comparative approach pro-vides good empirical parameters to bracket the esti-mates, but the nature of the political and economicstructure of every culture is different and so is thepopulation density. Finally, historical data are useful,but in the case of Tiwanaku, there is a 500-year gapbetween the earliest census data and the collapse ofTiwanaku.

Archaeologist Jeffrey Parsons (1968) first offeredan estimate for Tiwanaku of between 5,200 and20,000 people based upon his calculation of theurban center (2.4 km2) and using comparisons withMesoamerican cities. In this instance, he appears tohave been estimating just the urban core and imme-diate surrounding area. Carlos Ponce (1981:62) latercalculated an urban area almost twice the size (4.2km2) and an urban population of 42,000. Again, thisnumber referred strictly to the city core and the sur-rounding and adjacent urban area. Both archaeolo-gists were using the figure of around 10,000 peopleper square kilometer as a baseline assumption, anumber that was derived from studies mainly inMesoamerican urban centers.

A little later, Alan Kolata and Carlos Ponce(1992:332) argued that “an estimate of thirty to sixtythousand in not unreasonable” for the “permanentlyresident population” of the city. They also upped theestimate for the size of the city to 6 km2.

In his 1993 book, Kolata (1993:204–205) calcu-lated carrying capacities and marshaled the thenavailable archaeological data for the “metropolitan”Tiwanaku region, which he defined as the PampaKoani–Tiwanaku–Machaca area. This area is quiteclose to what I to refer to as greater Tiwanaku above.He concluded that the population during the peak ofTiwanaku “approached approximately 365,000, dis-


tributed into a concentrated, urbanized componentof some 115,000 and a dispersed rural component of250,000.” This figure was significantly below his cal-culation of the carrying capacity but implied largernumbers than either Ponce or Parsons had sug-gested earlier. Some mistakenly think that Kolata’sfigure of 115,000 meant only the resident popula-tion of Tiwanaku. However, he most likely refers tonot just Tiwanaku but to the other urban centers inthe greater Tiwanaku area, such as Lukurmata. Thiswould leave an estimate for Tiwanaku itself a bit un-clear in this calculation but certainly consistent withthe 30,000 to 60,000 figure published a year earlierwith Ponce.

One of Ponce’s latest articles (1995:32) repeatshis argument that Tiwanaku reached a population ofat least 40,000 in the urban center and up to 100,000total including the “outskirts” (aldeaños). It is notclear what outskirts he refers to, but it is most likelyan area within the 25-km radius of the city center.

In a later publication, Kolata (2003:15) loweredhis population estimate of Tiwanaku to around15,000 to 20,000. This figure, however, refers to justthe urban architectural core and the adjacent sur-rounding residential area. It does not include theTiwa naku Valley and the greater Tiwanaku area. Ko-lata bases this figure on new archaeological data an-alyzed from Proyecto Wila Jawira. John Janusek like-wise (2008:128) estimates the total population ofTiwa naku at 10,000 to 20,000, arguing that abouthalf of the city was empty space.

The work of these scholars is commendable. Wemust have some idea of the size of this city and itssurrounding area to proceed with our research in ameaningful way. However, the simple fact is that wedo not have a good database to determine the size ofTiwanaku with any great precision, and this is evi-denced by the shifting and at times contradictoryestimates. When the data are ambiguous, Tiwanakuscholars tend to rely upon their professional and


FIGURE 9.1. The area of greater Tiwanaku, an approximate 25-km radius from the city center.

individual perception of what Tiwanaku was inthe past. Not surprisingly, those of us who see Tiwa -naku as a classic archaic state tend to estimatehigh, while those on the other side tend to esti-mate low.

In this light, it is possible to briefly review thedifferent modes of assessing Tiwanaku population:

Carrying Capacity

The work of Kolata and his associates mentionedabove suggests that a very high population densitywas possible in the greater Titicaca region. While thisin and of itself does not prove high populations, it in-dicates that such levels are theoretically possible.While one could alter these figures and assumptions,the fact is that the altiplano in that region, withraised fields and more rainfall, has a capacity to sup-port a fairly dense population, at least in the greaterTiwanaku region.

Historical Data

Demographic data from the Toledo Tasa are veryprecise for encomiendas in the Titicaca region, but wedo not know how dispersed these populations were.Places in the northern lake area like Saman, Taraco,and Paucarcolla have around 4,000 to 5,000 total in-habitants in the encomienda lists. The towns them-selves were around 1 km2 in size. Including the entireencomienda tribute list from these towns would stillgive a population figure below the 10,000 per squarekilometer used in earlier estimates for Tiwanaku.However, these were entirely rural towns made up al-most completely of farmers and herders, with noeconomies of scale, production areas, intensive trade,and so forth. We also know that male taxpayerswould leave the area before the census takers came.Many of the males also were meeting tribute obliga-tions, working in the gold-bearing areas of Carabayaor possibly even in the silver mines of Bolivia.

Comparative Analysis

As mentioned above, Tiwanaku is almost an orderof magnitude larger than the icon of Aegean BronzeAge archaeology, Mycenae. It is also larger or aboutthe same scale as the four great Minoan cites ofPhaestos, Knossos, Malia, and Zakros. Whitelaw

(2001:29) estimates Neopalatial Knossos at its heightto have around 14,000 to 18,000 people in an urbanarea of around 80 ha. Averaged over the entire sitearea, he calculates a population density of 200 to 250per hectare. That would give a density substantiallylarger than the 10,000 used in earlier estimates forTiwanaku.

Skipping forward to the medieval period in Eng-land, the town of London in A.D. 1086 had approxi-mately 17,850 people (Barron 2004) in an area sub-stantially smaller than that estimated for Tiwanaku.In a short two centuries, that figure rose to between50,000 and 80,000, falling again in the fourteenthcentury after plagues decimated Britain. Around1300, the largest cities in Europe, Florence and Paris,had about 100,000 inhabitants each.

A famous map of the city of London in 1562 de-picts an urban space of approximately 300 ha, fromthe Tower in the east to Somerset House in the west.The city is bounded on the south by the Thames andon the north by pasture lands. Combined with thecity of Westminster about 2 km to the southwest, theurban space of greater sixteenth-century London isroughly equal to the 500 to 600 ha estimated forTiwa naku. The population density would have beenaround 10,000 to 15,000 people per square kilome-ter, a number consistent with earlier assumptions forTiwanaku. If London had been abandoned in the latesixteenth century, the architectural signature wouldlook similar to that of Tiwanaku—a few stone build-ings in ruins with the bulk of the former city coveredin sod mounds. The analogy is not unwarranted.

Coincidentally, the population density of six-teenth-century London is about the same as that ofmodern Puno, Peru. Puno has about 125,000 peopleliving in about 6 km2 of area; the vast majority aresingle families living in single- or two-story adobe orbrick houses. There are many open spaces in Puno aswell. The same densities hold for Juliaca and othermodern towns in the region. The problem here, ofcourse, is the commensurability of modern townsand pre-Hispanic Tiwanaku.

From a worldwide perspective, estimates forgreater Tiwanaku below 40,000 appear to be quitelow compared to cities of similar geographical sizeand complexity. The 15,000 population range is sim-ilar to that of settlements like Cahokia, a complexsettlement to be sure but hardly a fitting analogy toTiwanaku.


EMPIRICAL DATA FROM SURVEYAND EXCAVATION

What we can say is that there was an indisputablepresence of about 1 km2 of a relatively dense urbanconstruction (ritual constructions, buildings, plazas,alleyways, palaces, temples, and so forth), built incarved and shaped basalt, sandstone, and andesiteblocks.

We can also say that surrounding the core ofTiwa naku comprised several additional square kilo-meters of residential and domestic-use space. Therewas likewise a substantial population of nonurbanvillagers who lived from the edges of Tiwanaku itselfup to the lakeshore, a distance of some 17 km. Thisis based upon the survey work of Albarracin-Jordan(1992, 1996a, 1996b) and Mathews (1992). Duringtheir Tiwa naku IV and V periods, they documentedhundreds of settlements in the region between sub-urban Tiwanaku and the lake edge.

To the north was the Katari Valley and theTaraco Peninsula. Sites such as Lukurmata reachedat least 1.5 km2 in size (Stanish 1989). Dozens ofother Tiwanaku sites populated the Katari Basin andthe Taraco Peninsula. Bandy (2001) surveyed theTaraco Peninsula. His data indicate a populationdrop during the Late Formative 2 period, suggestingan out-migration to Tiwanaku, an observation thatconfirms the settlement pattern dynamics in theTiwa naku Valley. During the Middle Horizon, thepopulation rebounded to near normal levels whileTiwanaku continued to grow at a very high rate.These data strongly suggest that the initial growth ofTiwanaku was a result of a depopulation of the sur-rounding area, but once established as an urban cen-ter, this part of greater Tiwanaku was fully repopu-lated. These data support a population estimate forTiwa naku on the high end.

Archaeologists have also excavated outside thecore area of the Tiwanaku urban zone and havefound fairly dense domestic settlement. The moderntown of Tiwanaku is full of archaeological debrisfrom the late Tiwanaku IV and V periods. We do notknow the extent of the entire settlement complex atTiwanaku. Certainly, there are areas near the archi-tectural core that do not have evidence of settlement.However, there are also areas near the core with veryintense domestic occupations. In particular, the workof Alconini (1995), Couture (2002), Janusek (2008),

Escalante (1997), Portugal (1993), Rivera Casanovas(1994), and many others has uncovered craft pro-duction and residential areas. Tiwanaku artisansmanufactured a great variety of commodities, frommusical instruments to textiles. These commoditieshave been found in distant areas, such as San Pedrode Atacama in northern Chile some several hundredkilometers away.

So how large was Tiwanaku? My inclination isalways to defer first to the empirical data and secondto comparative analysis. We do not have historicalcensus data for Tiwanaku as an urban phenomenon.From this perspective, greater Tiwanaku would havehad a population of about 70,000, with perhaps halfof this concentrated in the urban center at any onetime. There would have been many instances of mul-tiple residences for families, and the total populationof the center would have varied according to the rit-ual and agricultural calendar.

The question that immediately arises is: Doesdemographic size matter? The answer is yes. Non-hierarchical societies simply have structural limitsfor growth. Such a society could not build a city likeTiwanaku. The relationship between population sizeand internal complexity is not linear. But at certain“tipping points” of demographic densities, structuralshifts toward greater complexity are unavoidable.There does not exist a single example in the history ofthe world where a site as large and complex as Tiwa -naku was not constructed by a fairly complex statesociety. In every single instance where we find siteslike Tiwanaku with historical documents, we alsofind that there were unequivocal social and politicalhierarchies, backed by intensive economic produc-tion and exchange.

Looked at from another angle, there does notexist a single historically documented empty pil-grimage center or ceremonial site of even a quarterthe size of Tiwanaku found outside of a state organ-ization. Sites such as Mecca, Delphi, and others allhave large permanent settlements attached to themand functioned within a regional political organiza-tion. From yet another perspective, every historicallydocumented ancient city in the world was home topageant and ceremony. Medieval London was thecenter of royal celebrations on a grand scale. Onecould easily pinpoint empty spaces in this city, and acreative archaeologist could discuss the more than100 documented churches within 4 km2 to argue


that London in the sixteenth century was a huge cer-emonial center built for overly religious peasants. Ofcourse, we know otherwise because we have texts,and also perhaps because we value the image of ourancestral civilizations as something more than pas-sive, praying peasants and disheveled agrarian serfs.

Based upon the data available and analogies tocomparable centers the world over, the city of Tiwa -naku was an architectural monument that was simul-taneously a pilgrimage destination, a political center,an economic powerhouse, and a residential place forelite, specialized labor classes and a large proletariat.Like Jerusalem since the ninth century B.C., perhapsthe quintessential icon of a religious center in theWest, Tiwanaku was first a political and economiccenter that also hosted an array of highly charged re-ligious institutions and events. Anyone who arguesthat Tiwanaku was an empty ceremonial center orjust a pilgrimage destination holds an extraordinar-ily high burden of proof given the lack of any histor-ically verified analogue from any culture on five con-tinents throughout the course of human history.

THE STRAW-MAN MODEL OF THECOERCIVE HIERARCHICAL STATE

The term hierarchy is one of the most widely usedand most poorly developed concepts in the archaeo-logical literature. Scholars in archaeology in generalseem to adopt an idealized concept of hierarchy bestrepresented by those pyramid charts from the late1970s published in the older evolutionary anthropo-logical literature. In this model, there is a paramountor king/queen, with discrete levels of decision mak-ing layered one on top of another. These classes are,to use another term, stratified. The implication isthat the highest decision maker receives and givesinformation through a formal network of subordi-nates. This model also implies that each subordinatelevel is subject to the control of the level above. Intheory at least, each superordinate level has virtuallylife and death power over subordinates. Most signif-icantly, each level has different access to resourcesand wealth (e.g., see Schortman 1989).

In the 1970s, archaeological theorists took thisconcept to a new level, looking at hierarchy not onlyinternally within a society but regionally, over a land-scape. In this model, multitiered settlement patternsrepresented congruent control over people’s lives and

resources. A typical statement is: “Simple chiefdomshave one decision-making level, or control hierarchy,above the local community; complex chiefdoms haveat least two such levels” (Beck 2003:643). States weredifferent in the degree to which force was monopo-lized. Some states became empires, with Rome beinga kind of default model. In this theoretical construc-tion, the state was composed of elite who exertedtheir will by military force and other kinds of coer-cion. Most importantly, there was strong structuralcontinuity between these levels. By that, these differ-ent stages were seen as smooth and evolutionary,with states intensifying the structure of chiefdomsand with empires effectively being hyperstates thatwere simply larger.

This was a good model. It is parsimonious, bold,useful, and testable. Unfortunately, from a historicalperspective, it is wrong; no matter where one looks inthe historical literature, we never find a stratified so-ciety in which power is so elegantly and rigidly dis-tributed. Even the most hierarchical preindustrial so-ciety in the world that we know of—say, the Franceof Louis XIV or Rome under Augustus—imposedlimits on the power of the elite and dispersed controlto various groups and institutions. There are alwaysmultiple sources of power in any society, no matterhow rigid and stratified it may be or, more impor-tantly, how powerful the elite claims to be. Kingsneed senates and parliaments, the non-elite resistsencroachment constantly, middle classes famouslyusurp the power of the aristocracy, religious soci-eties form counter-hegemonic entities that threatenthe state, and so forth.

As anthropological archaeologists searched forthese ideal hierarchies in the dirt, they of coursecame up empty-handed since they did not exist. In-stead they encountered reality—multiple contempo-rary palaces, “corporate” elite strategies, non-elitehouseholds with lots of “elite” goods, humble tombswith priceless objects, ephemeral elite from Teoti-huacán to Harappa, “network” states, nodal commu-nities, and the like. Instead of questioning the empir-ical utility of the traditional hierarchy model, weAndeanists instead came to believe that “real” hierar-chical states were indeed found “over there” (usuallyin the Mediterranean, Aegean, and Mesoamerica,where the documents told us there were indeedkings) and that the cultures that we studied “here”did not in fact have hierarchies. We therefore in-


vented or borrowed new terms like heterarchy, seg-mentary states, and so forth to try to explain the ap-parent absence of centralized, hierarchical states inthe pre-Inca Andes.

The reality is, to repeat, that there is no suchthing in the real world as the rigid hierarchy modelas used in archaeology. The case of London in thelater Middle Ages is a telling example of the com-plexities of power and authority in a premodernstate. England at this time was the quintessential cen-tralized, premodern bureaucratic state with a king, aparliament, armies, navies, police forces, exchequers,sheriffs, taxing bodies, bishops, earls, dukes, royalcourts, a strong state-supported church in which theking was head, embassies, palaces, a feudal aristoc-racy of landlords and knights, vassals, serfs, inden-tured servants, and the like. Yet authority was distrib-uted in anything but a pyramid of power.

The work of Barron (2004) provides an excellentsummary of power and authority in England andLondon from 1200 to 1500. As she puts it, the rela-tionship between the king and London could be re-duced to the fact that “the king needed money andthe Londoners wanted self-government” (Barron2004:9). This fundamental economic and politicalfact underlay much of the tension over centuries.The power of the Crown waxed and waned overtime. In theory, the Crown always retained de jurerights to govern the city. If one were to read officialdocuments from that era, it would appear that therewas a clear-cut political hierarchy with the king ontop. This, indeed, echoes the political ideals set out inother more fragmentary documents, such as LinearB tablets and cuneiform texts. In fact, to the contrary,there was a great deal of negotiation and formal andinformal political restrictions on elite action in Eng-land at this time.

A look at how power actually was distributed inmedieval England is far from the stratified pyramidsthat we assume. As Barron says (2004:10), becausethe king spoke with one voice, the Crown “had a dis-tinct advantage in formulating and carrying out pol-icy but, if push came to shove, the Londoners couldmuster a sizeable military force.” A mayor and acourt of aldermen, institutions that changedthroughout the centuries, governed the city. Therewere 24 wards in the city in the early thirteenth cen-tury (Barron 2004). Under the aldermen were anumber of offices such as ward beadles, rakers, scav-

engers, and constables. These were all answerable tothe aldermen, being sworn in before them and themayor. Mayors were elected early on by the barons ofLondon under royal charter. Over time they wereelected by groups of aldermen from each ward, withwaxing and waning influence by the commoners.The history of the institution shows periods of inter-ference by the Crown, without question, but thechoice of mayor rested in the hands of the lesser eliteand at times the commoner classes (Barron 2004).Sheriffs were likewise selected this way. There weretimes when the king tried to interfere, but the cityprotected its right to select the sheriffs (Barron2004). Other members of the bureaucracy were al-most all appointed by the city and not the king.

Over time a variety of democratic institutionstook hold in London. In anthropological terms, wewould view these as centers of non-elite resistanceto authority. These institutions included those suchas the Folkmoot and the Hustings Court. By the laterfourteenth century, the court of common counciltook over from these earlier institutions (Barron2004). Along with the aldermanic councils, these in-stitutions sought ways to raise money for communalprojects. By the fifteenth century, taxes for theseprojects could be raised only by consent of both thecommoner and aldermanic groups (Barron 2004).These projects included piping in water to the city,the repair of granaries, and the improvement of thecity walls and ditches (Barron 2004). From an ar-chaeological perspective, significant construction ofthe urban area was done virtually outside of royal or“state” control.

Another check on both city and royal powerwas the institution of sanctuaries controlled by thechurch. Effectively the 100 or so churches providedhavens from state authorities for criminals and evenpeople accused of treason. In 1566, though, thesanctuary was abolished, as the combined politicalpower of the city and Crown diminished thechurch’s role (Barron 2004). The king did havepower of life and death over some people, particu-larly political troublemakers. However, this powerwas exercised in the city only sparingly. On theground, other elite had palaces, controlled long-dis-tance exchange, created their own bureaucracies,and the like. Middle-class merchant groups hadconsiderable power over everyday life, a spherewhere the Crown could not interfere, no matter how


much it tried. Even commoners could exact de-mands from the authorities, including the king,mayor, barons, and other minor aristocracy.

A historical anecdote is quite telling. In 1369 Ed-ward sent a royal writ to the city, telling it to clean upthe disgusting filth in the streets left by butchers. Thealdermen effectively ignored the order for more than20 years (Barron 2004). The king of England, theapex of a stratified hierarchical state, while capable ofoccasionally executing a heretic and able to raisearmies and navies at will, could not even force thepeople in his capital city 2 km away to clean up theirgarbage. If this is not heterarchy or “segmentary,” Idon’t know what is.

The oft-cited article by Fritz (1986) about theHindu imperial capital of Vijayanagara falls into thiskind of archaeological theory-building trap. He de-scribes Vijayanagara, following the historian Stein(1980), as a “segmentary state, consisting of relativelyautonomous polities.” Quoting Appadurai (1978:51),he describes this urban capital of more than 10 km2of core architecture as housing a state with “no single,centralized, permanent bureaucratic organization, buta temporary affiliation of local groups, authoritativelyconstituted by, or in the name of the king, and em-powered to make public decisions on specific mat-ters.” These three scholars from three disciplines—archaeology, anthropology, and history—contrast thissegmentary model with apparently that of a nonseg-mentary or stratified hierarchy model. But, as we cansee, their description of the “segmentary” state fitsvery well, almost to a tee, sixteenth-century England.If sixteenth-century England is not a centralized state,then what is? If it is not, then the concept of a coer-cive, hierarchical state has no analytical meaning.

This straw-man hierarchy model is part of ourdiscourse on Tiwanaku. Goldstein (2005) brings inVijayanagara as a kind of analogy for Tiwanaku. Heargues, following Sinopoli and Morrison (1995), thatthe rulers of the southern Indian state did not con-trol domestic relations of production and had to ne-gotiate with local elite. In his view, therefore, thisstate is not an example of “globalist model[s] of coer-cive core-centered hierarchy” but something less hi-erarchical. To Goldstein’s credit, he clearly articulateswhat this globalist model is, unlike far too many ofus who just assume it to be understood. In this view,a state is hierarchical, bureaucratic, and predatory—the end point of an evolutionary process. It controls

production and interferes in the domestic economy.There must be provinces or colonies as part of theexpansion process. These colonies must have admin-istrators and governors.

Goldstein goes on to list several things missingfrom the Tiwanaku state that one would expect fromthis coercive state model. In his view, there are nopalaces at Tiwanaku that would qualify as royalty.Goldstein also states that there is little evidence thatthe Tiwanaku elite interfered with domestic produc-tion. He says that there were no formal roads. Gold-stein, commendably echoing most of our colleaguesin this symposium, draws a very profound conclu-sion of what we should find at a place like Tiwanaku,given the globalist, evolutionary model:

Neoevolutionists thus implicitly assume thathorizontally distinct corporate social groups likeclans or Andean ayllus wither away as their in-creasingly redundant functions are usurped bythe state’s hierarchy. Patterns of authority andgroup identity based on kinship and ethnicityare seen to become socially vestigial, politicallyimpotent, and administratively irrelevant inclass-based societies. (Goldstein 2005:307)

These are fascinating inferences but are not nec-essarily born out by the data from comparative his-tory or anthropology. A quick read of the Iliad showshow important kinship and ethnicity are in structur-ing the state civilizations of the Aegean. The Azteccalpulli are famous horizontal institutions that weredefined by both ethnicity and kinship. The OttomanEmpire had too many ethnic, national, and religiousinstitutions within its midst to even count. In ourother example of medieval London, the town was fullof horizontal institutions that existed outside of di-rect royal authority, including guilds, baronial housesdefined by kinship, merchant barrios, religious insti-tutions, foreign quarters, and the like. There is literallyan unlimited number of examples from history andethnography to draw from in state societies.

In short, the empirical record shows that ayllu-like institutions were found throughout the greatstates and empires of the ancient world. Far from dis-appearing as an inevitable result of state develop-ment, such institutions flourished as a primarymeans to structure society in hierarchical, state mod-els. They are historically varied but structurally verysimilar across space and time. The fact that they ex-


isted in Tiwanaku does not diminish its status as acentralized state; to the contrary, it brings in Tiwa -naku as one of the great states of the ancient world.

There is likewise little evidence that the elite ofmost state societies interfered with domestic produc-tion of the non-elite for their own subsistence. Al-most all states in premodern times tax primary pro-ducers either through corvée labor or direct tribute.It is not in any elite group’s interest to tell farmershow to farm, to tell butchers how to butcher, and soforth. The elite simply take a portion of that produc-tion for their own use. Even the Inca, perhaps one ofthe most economically intrusive states in the ancientworld, did not intervene in local production, prefer-ring instead to use corvée labor to work state landsand installations. Local production was largely un-touched. While Goldstein may be absolutely correctin deducing this feature as a component of state so-cieties out of the theoretical literature, it is not an em-pirically verified component of state societies fromaround the world.

The Tiwanaku peoples had roads, albeit not likethe ones the Inca had. The surveys from the Juli-Po-mata area show that the Tiwanaku sites align alongroads that were then co-opted by the Inca. The sur-veys to the north in the Huancané-Putina area alsoindicate a concentration of settlement on the roadsystem (though these data have not been published).Our (Stanish et al. 2010) recent survey in the area be-tween Desaguadero and Moquegua indicates thatthere were indeed Tiwanaku artifact scatters on sitesalong the road between the lake and the largestcolony. However, it is quite true that this was a moreinformal and noncentralized kind of system than wesee for the Inca. Nevertheless, while we did not findtambos or other kinds of way stations that dated tothe Tiwanaku period, we did find a line of sites thatled to Moquegua, indicating that some kind of ex-change took place along this road.

I disagree about the lack of palaces as well. Callthem what you want, but from a comparative per-spective, if Tiwanaku structures such as the Puma-punku were not elite residences, then surely thepalaces of most of the Bronze Age and early Iron AgeAegean and Mediterranean would not qualify either.Indeed, there are no royal tombs like Moche at Tiwa -naku. But the Inca did not have royal tombs either. Infact, building elaborate tombs for the elite is not, andappears to never have been, a feature of highland An-

dean culture except for some post-Tiwanaku chull-pas found only in the south-central Andean region.They are not found in the Tiwanaku, Pucará,Chavín, or Wari cultures, indicating a highland tra-dition that does not include elaborate tombs for in-dividual dead elite like anything remotely as elabo-rate as among the Moche.3

Goldstein notes that there are no iconographicrepresentations of secular hierarchy in Tiwanaku,unlike among the Moche. There are none in the Incaculture either until the colonial period; nor, for thatmatter, are any represented in Chan Chan, Teoti-huacán, and many other ancient states.4 There are noMoche-like representations of elite in Wari or Pu-cará art either. This also appears to be a highlandtradition (although the images of people wearingpuma headdresses decapitating sacrificial humansin Pucará art might come close).

If the criteria adopted by those testing Tiwa -naku against the straw-man argument were appliedaround the world, then there would be very few an-cient states and no first-generation states at all. Toput it another way, by the criteria proposed above,most ancient empires and all first-generation stateswould be segmentary states. I adamantly believe thatwe have to rethink what a coercive hierarchy reallyis in practice. In light of the fact that the king of latemedieval England could not even get his subjects toclean up their garbage, it is clear that our models ofstates and even empires need to be substantially re-vised. As I have implied, I believe that the modelused by some and attacked by others is one thatnever has existed and that, in fact, all societies theworld over had multiple axes of power and wealth.

By sticking strictly to consistent archaeologicalcriteria of verification, it is virtually impossible todeny that Tiwanaku was an urban, stratified centerof an ancient state more complex than MycenaeanGreece and certainly as complex as twelfth-to-sixteenth-century London, Minoan Crete, and othergreat civilizations of the ancient world.

TIWANAKU COLONIES

Virtually all of the great classical civilizations of an-tiquity in the Old World had colonies. In reviewingthe literature on the Andes, I sense that a similarstraw-man logic is at work here as well. In this flawedconcept, a colony is a product of a reasoned, rational


policy on the part of state administrators to controlan area, politically, economically, and/or militarily. Astate bureaucracy of some sort decides where to putthe colony, and then it amasses military and adminis-trative resources to build an outpost that is incorpo-rated into the political orbit of the home country. Fol-lowing world-systems theory, the peripheries werecompletely subordinate to the core, which extractedresources and dominated the political, economic, andat times even cultural life of the periphery.

This is indeed a model that was proposed bysome system theorists and incorporated into manyprocessualist models in the 1970s, particularly thosethat drew off of world-systems theory. But onceagain, a look at historically documented colonies inthe ancient world of states and empires presents adifferent picture.

Colonies are highly varied. They physically con-tain all sorts of things, including stores, forts, armies,government agents, religious buildings, residences,and specialized production areas. Sometimes the re-lationship between colony and home country isstrong, sometimes weak; and almost always the rela-tionship breaks down after a few generations. Theearliest documented Greek colonization was farfrom this straw-man model of intentional, rationalcolonization by home capital city:

These migration settlements were not colonies inthe usually understood sense of the term. Theywere not organized movements, directed fromand set forth by a particular city, but small bandsof homeless folk dispossessed by the so-calledDorian Invasions. . . . The real colonial move-ments began somewhat later . . . when the Greekcities of the mainland recovered from the . . . de-struction of the Mycenaean centers in the thir-teenth and twelfth centuries. (White 1961:444)

Later Greek colonies, such as “Al-Mina and laterNaukratis were trading posts, occupied with the con-sent of Assyria and her successors in Syria andEgypt. Neither the Syrian coast nor Egypt was openfor the foundation of true colonies, that is, sizeablesettlements of colonists on agricultural land, wherean independent new community with full civic lifecould develop” (White 1961:446).

The most mature Greek colonies were sent outby the home city and do indeed fit aspects of thestraw-man model of colonies: “The mother-city or

metropolis selected the site, appointed the leader ofthe colony, called for volunteers, and organized thecolony. The major portion of the colonists usuallycame from the mother-city, but contingents fromother places sometimes participate” (White 1961:449). However, even the most complex of coloniesquickly rid themselves of their political and eco-nomic links to the home city, and this deviates sub-stantially from the core of the straw-man modelstructure: “Once the new city was founded the tieswhich bound it to the mother-city were those of re-ligion and sentiment only. Colonies were in no sensea colonial empire of the mother-city; they pursuedtheir independent ways, and many soon becamemore prosperous and more famous than the mother-cities” (White 1961:449).

The short of it is this: bureaucratically adminis-tered, formal colonies with politically appointedagents of the state occur virtually only in the mostcomplex of premodern states. And even in these, theties that bind colony and home country are quiteephemeral, weakening or disappearing in two orthree generations. The vast bulk of colonies in thepremodern world are far less formal. Virtually allcolonies eventually develop their own identities andcease functioning as an extension of the home polity.

From this perspective, Tiwanaku likewise hadcolonies. The Tiwanaku colony in Moquegua hasbeen ably described by several scholars, most no-tably Goldstein in a series of publications since theearly 1990s. What we know, empirically, is this.There is a huge amount of Tiwanaku pottery foundin a series of settlements throughout the MoqueguaValley, stretching from the coast in Ilo to up to 3,000m in the high drainages. No other pottery styles ofany iconographic complexity (with the exception ofthe very rare Wari sherd found on occasion) arefound to coexist with the Tiwanaku pottery in thesesites. There is a large site complex called Omo thatcontains an unusually large amount of Tiwanakupottery, even for Moquegua. There is a site calledChen Chen that had literally thousands of cisttombs filled with Tiwanaku-style pottery and otherartifacts. There is a structure on the main site ofOmo that is built in a miniature style similar to thatfound at the Tiwanaku capital. Below Omo are veryextensive fields that up to the present day providesome of the richest agricultural land in the south-central Andes.


The Wari site of Cerro Baúl is also located in theMoquegua drainage and had been ably studied andpublished by Luis Lumbreras, Bertha Vargas, RobertFeldman, Michael Moseley, Donna Nash, RyanWilliams, and others. It is in the same size categoryand is as complex as Mycenae, Tiryns, Gla, and otherwell-known Bronze Age Aegean sites. Moquegua isthe only known place in the Andes where there areboth Tiwanaku and Wari settlements (though somesmaller sites in Moquegua have some pottery fromboth cultures). The Wari site is located on a famouslydefensive massif, once aptly described by MichaelMose ley as the Masada of the Andes, which providesthe highest level of protection available in any pre-modern settlement. Adjacent to Cerro Baúl is the siteof Cerro Mejía, also a Wari-affiliated site. Surround-ing Baúl is a series of Tiwanaku sites that were, forall intents and purposes, contemporary for a sub-stantial period of time with the site of Wari on thesummit above.

There is a very clear ritual/religious componentto Cerro Baúl, as there is at Omo, Tiwanaku, Cuzco,Teotihuacán, Tenochtitlán, and virtually every otherpolitical center in the ancient world. There are alsosome Tiwanaku objects found at the summit as well,indicating that the site was a place where Tiwanakuand Wari peoples probably met, drank, possibly slept,and almost certainly negotiated with one another. InMoseley et al.’s excellent characterization, it was an“embassy-like delegation of nobles and attendant per -son nel that endured for centuries” (Moseley et al.2005:443).

Cerro Baúl was also a defensive site that served tokeep its occupants safe from the surrounding Tiwa -naku settlements. The Tiwanaku-affiliated peoplesdid not need defensive locations, since they vastlyoutnumbered the Wari contingent. People who argueagainst conflict in the Middle Horizon or against adefensive function for Cerro Baúl fail to understandthat historically, competing polities rarely have ac-tual battles more than a mere fraction of the time. Itis very common in the historical literature to notethat a cooperative “live and let live” philosophy is thenorm in human affairs, a norm that is occasionallypunctured by outbursts of organized violence. Theseoutbursts can indeed have enormous political andother consequences, but they are in fact quite rare.The Hundred Years’ War (a total of around 115 years,with 80 of those actually violent) is called what it isprecisely because it is so rare. If we use history as our

guide, it is most likely that 99 percent of the time,the Tiwanaku and Wari peoples in Moquegua wereinteracting in peace for their mutual self-interests.That is not to say that they were not adversaries, andas such they had to maintain defensive postures vis-à-vis each other. But the historical record is repletewith examples of adversaries engaged in simultane-ous conflict and trade. Usually it is the political elitethat promotes the first, and it is the non-elite thatengages in the latter. Nevertheless, conflict and co-operation are not mutually exclusive. Once this is re-alized, the settlement complex around Cerro Baúlmakes eminently good sense and fits in well withour historical understanding of complex societies.

I would also argue that the Wari leadershipmaintained Cerro Mejía to prevent sieges. Thiswalled site is located in a manner that would pre-cisely ensure that the only access to Cerro Baúl couldnot be blocked. Any attempt to do so would put thebesiegers in a tactically impossible position, out-flanked on low ground between the two hills of Baúland Mejía. In premodern military terms, such a po-sition is almost certainly fatal and one to be avoidedat all costs, a point constantly reiterated by militarystrategists from ancient China to modern armymanuals.

I would also argue that the construction of Tiwa -naku sites around the base of Cerro Baúl is strategicin nature as well. It effectively cuts off Wari “expan-sion” outside of Baúl and ensures that an uneasypeace existed between the two peoples. Of course theTiwanaku peoples could have built more settlements,but they did not. I would suggest that, again basedupon historical analogies, the settlement system inthe Cerro Baúl area preserves the outcome of pro-tracted negotiations and understandings between thetwo rivals in the only place in their world where theychose to interact on a formal basis. Wari was welcomein Tiwanaku territory in this one instance because itwas in each side’s interests; but the Wari were appar-ently given their limits, and they accepted these toachieve other goals. If not, we would see many moreWari settlements up and down the valley—we do not.

According to Williams and Nash (2002), there isa Tiwanaku construction on the summit of Baúl.This is reminiscent of the Teotihuacán barrio inKalminaljuyu, the Oaxacan barrio in Teotihuacán,the central Mexican temple in Tikal, and countlessexamples in the history of the Old World. In the


eighteenth-century Ottoman Empire, foreign repre-sentatives and ethnic minorities were required tolive in their own neighborhoods (Göçek 1987:6).The Hanseatic League had a special area in sixteenth-century London called the Steelyard. Throughoutthe historical world in state societies, it was com-mon for larger settlements to host a small section offoreigners, who maintained their cultural attributes.It is found at Baúl, and the work of Janusek at Tiwa -naku may have found some similar barrios.

The Tiwanaku peoples expanded into other areasin the south-central Andes as well. In their heartland,they established colonies or enclaves in the Puno Bay,in the north near Arapa and Huancané, and through-out the lowlands in areas that we barely understand.Many of the Tiwanaku sites in Cochabamba arecolonies that utilized a distinctive architectural stylethat we see in the northern Titicaca Basin.

Some scholars have suggested that the nature ofpottery production in the peripheral Tiwanaku set-tlements indicates a nonhierarchical relationshipbetween colony and core. They argue that the pot-tery is all locally manufactured as opposed to beingimported from the capital. In their view, this arguesfor a less centralized state, or no state control at all. Iwould note that this Tiwanaku pattern is preciselythe pattern of pottery manufacture that we see inthe Inca Empire. The leaders of Tawantinsuyu mili-tarily captured provinces and installed artisan work-shops for pottery, cloth, and metals. Their norm forceramic manufacture was to create workshops inpro vincial territories, where Inca canons were exe-cuted more or less according to some standard set inCuzco, although there was considerable borrowingfrom local pottery traditions. If anything, data fromTiwanaku colonies that indicate that they pursued apolicy similar to the Inca simply reinforce the no-tion that Tiwanaku followed the same norms ofstate building in the Andes as the Inca. Polities suchas Chavín, Wari, and others that did not follow thisnorm most likely were less complex than the Tiwa -naku or Inca, relying instead on earlier practices ofthe direct movement of ceremonial and feasting ob-jects to distant places.

SUMMARY

The model of the coercive, hierarchical state used toassess Tiwanaku is a flawed one—one without em-pirical foundation. The fact is that Tiwanaku is a city

comparable in size and complexity to medieval Lon-don and Bronze Age Knossos, to name but just twoiconic sites of the ancient world. The people of Ti-wanaku built massive agricultural fields, establishedcolonies over hundreds of kilometers, and built hugetemples and palaces adorned with monoliths andgreat art. They established roads and causeways.They created the Kalasasaya and the Akapana, plusnumerous buildings surrounded by a great moat.The city was planned, with components that rangedfrom a prosaic but sophisticated sewer system to ar-chitectural feats that combined centuries of religiousprinciples and an amazing understanding of how tomove people through magnificent space. Tiwanakupeoples massed-produced ceramic art on a scale notseen before in the region. They created sophisticatedmetalworking, textile, and musical instrument in-dustries and much more.

The city of Jerusalem is perhaps a better analogyfor Tiwanaku. It was also the political, economic, anddemographic center of kingdoms and empires fromat least the ninth century B.C., and while only halfthe size of Tiwanaku in area, it held a population ofup to 200,000 in the first century A.D. I see no struc-tural difference in kind between Jerusalem, Tiwa -naku, and any other capital of a great civilization.

NOTES1 I have argued elsewhere (Stanish 2010) that there is a struc-

tural limit of around 100,000 people in Andean cities due to thenature of their political and economic organization. The earlyfirst-generation states of Moche, Wari (possibly Huaro), and Ti-wanaku reached about half this size quickly but did not growmuch beyond that. Even Inca Cuzco did not grow beyond100,000 inhabitants, even though it was the capital of one of themost powerful empires in world history.2 An easy walk is about 5 km per hour with an occasional

rest. At a fast walking pace, one can cover a kilometer in about10 minutes.3 The only exception would be the Late Horizon chullpas at

sites like Sillustani and Tanka Tanka.4 A possible exception being wooden keros, but most of those

are colonial in date.

REFERENCES CITED

Albarracin-Jordan, Juan1992 Prehispanic and Early Colonial Settlement Pat-

terns in the Lower Tiwanaku Valley, Bolivia. Un-published Ph.D. dissertation, Department of


Anthropology, Southern Methodist University,Dallas.

1996a Tiwanaku. Arqueología regional y dinámica seg-mentaria. Editores Plural, La Paz.

1996b Tiwanaku Settlement System: The Integration ofNested Hierarchies in the Lower Tiwanaku Val-ley. Latin American Antiquity 7(3):183–210.

Alconini Mujica, Sonia 1995 Rito, símbolo e historia en la pirámide de Aka-

pana, Tiwanaku: Un análisis de cerámica ceremo-nial prehispánica. Editorial Acción, La Paz.

Appadurai, Arjun1978 Kings, Sects and Temples in South India, 1350–

1700 A.D. In South Indian Temples: An Analyti-cal Reconsideration, edited by Burton Stein, pp.47–73. Vikas, New Delhi.

Bandy, Matthew2001 Population and History in the Ancient Titicaca

Basin. Unpublished Ph.D. dissertation, Depart-ment of Anthropology, University of California,Berkeley.

Barron, Caroline 2004 London in the Later Middle Ages: Government

and People 1200–1500. Oxford University Press,Oxford.

Beck, Robin A., Jr. 2003 Consolidation and Hierarchy: Chiefdom Vari-

ability in the Mississippian Southeast. AmericanAntiquity 68(4):641–662.

Couture, Nicole C. 2002 The Construction of Power: Monumental

Space and Elite Residence at Tiwanaku, Bolivia.Unpublished Ph.D. Dissertation, Departmentof Anthropology, University of Chicago,Chicago.

Escalante Moscoso, Javier 1997 Arquitectura prehispánica en los Andes boli-

vianos. Producciones CIMA, La Paz. 2003 Residential Architecture in La K’arana, Tiwanaku.

In Tiwanaku and Its Hinterland: Archaeologicaland Paleoecological Investigations of an AndeanCivilization. Vol. 2, Urban and Rural Archaeology,edited by Alan L. Kolata, pp. 316–326. Smithson-ian Institution Press, Washington, D.C.

Fritz, John M.1986 Authority and Meaning of a South Indian Impe-

rial Capital. American Anthropologist 88(1):44–55.Göçek, Fatma Müge 1987 East Encounters West. France and the Ottoman

Empire in the Eighteenth Century. Oxford Uni-versity Press, Oxford.

Goldstein, Paul 2005 Andean Diaspora: The Tiwanaku Colonies and

the Origins of South American Empire. Univer-sity Press of Florida, Gainesville.

Janusek, John 2008 Ancient Tiwanaku. Cambridge University Press,

Cambridge.Kolata, Alan1993 The Tiwanaku: Portrait of an Andean Civiliza-

tion. Blackwell, Cambridge, Massachusetts.2003 The Proyecto Wila Jawira Research Program. In

Tiwanaku and Its Hinterland: Archaeology andPaleoecology of an Andean Civilization. Vol. 2,Urban and Rural Archaeology, edited by Alan L.Kolata, pp. 3–17. Smithsonian Institution Press,Washington, D.C.

Kolata, Alan, and Carlos Ponce Sanginés1992 Tiwanaku: The City at the Center. In The An-

cient Americas, Art from Sacred Landscapes, ed-ited by Richard F. Townsend, pp. 317–333. ArtInstitute of Chicago, Chicago.

Mathews, James Edward1992 Prehispanic Settlement and Agriculture in the

Middle Tiwanaku Valley, Bolivia. UnpublishedPh.D. dissertation, Department of Anthropol-ogy, University of Chicago, Chicago.

Moseley, Michael, Donna J. Nash, Patrick Ryan Williams,Susan deFrance, Ana Miranda, and Mario Ruales2005 Burning down the Brewery: Establishing and

Evacuating an Ancient Imperial Colony at CerroBaúl, Perú. Proceedings of the National Academyof Sciences 102(48):17264–17271. Washington,D.C.

Parsons, Jeffrey1968 An Estimate of Size and Population for Middle

Horizon Tiahuanaco, Bolivia. American Antiq-uity 33:243–245.

Ponce Sanginés, Carlos1981 Tiwanaku: Espacio, tiempo y cultura: Ensayo de sín-

tesis arqueológica. Los Amigos del Libro, La Paz.1995 Tiwanaku: 200 Años de investigaciones arque-

ológicas. Producciones CIMA, La Paz. Portugal Ortiz, Max1993 Trabajos arqueológicos de Tiwanaku, Pt. 1. Tex-

tos antropológicos 4:9–50. Rivera Casanovas, Claudia1994 Ch’iji Jawira: Evidencias sobre la producción de

cerámica en Tiwanaku. Unpublished licenciaturathesis, Universidad Mayor de San Andrés, La Paz.

Shortman, Edward 1989 Interregional Interaction in Prehistory: The

Need for a New Perspective. American Antiquity54(1):52–65.

Sinopoli, Carla, and Kathleen Morrison 1995 Dimensions of Imperial Control: The Vijayanagara

Capital. American Anthropologist 97(1):83– 96.Stanish, Charles1989 Tamaño y complejidad de los asentamientos

nucleares de Tiwanaku. In Arqueología de


Lukurmata, Vol. 2, edited by Alan Kolata, pp.41–57. Instituto Nacional de Arqueología, La Paz.

2003 Ancient Titicaca: The Evolution of Complex Soci-ety in Southern Peru and Northern Bolivia. Uni-versity of California Press, Berkeley.

2010 Labor Taxes, Market Systems and Urbanizationin the Prehispanic Andes: A Comparative Per-spective. In Archaeological Approaches to MarketExchange in Ancient Societies, edited by C. Gar-raty and B. Stark, pp. 185–205. University of Col-orado Press, Boulder.

Stanish, C., E. de la Vega, M. Moseley, R. Williams, B.Vining, and K. LaFavre2010 Tiwanaku Trade Patterns in Southern Peru.

Journal of Anthropological Archaeology 29(4):524–532.

Stein, Burton1980 Peasant State and Society in Medieval South

India. Oxford University Press, New Delhi.

White, Mary E. 1961 Greek Colonization. Journal of Economic History

21(4):443–454.Whitelaw, Todd2001 From Sites to Communities: Defining the

Human Dimensions of Minoan Urbanism. InUrbanism in the Aegean Bronze Age, edited by K.Branigan, pp. 15–37. Sheffield Studies in AegeanArchaeology, Sheffield.

Williams, Patrick Ryan, and Donna J. Nash 2002 Imperial Interaction in the Andes: Huari and Ti-

wanaku at Cerro Baúl. In Andean Archaeology.Vol. 1, Variations in Sociopolitical Organization,edited by William H. Isbell and Helaine Silver-man, pp. 243–265. Kluwer Academic/PlenumPublishers, New York.


what was tiwanaku?. 2013. charles stanish. in visions of tiwanaku. edited by a. vranich and c....

Documents