VISIONS OF

T I WA N A K U

COTSEN INSTITUTE OF ARCHAEOLOGY PRESS

MONOGRAPH SERIES
CONTRIBUTIONS IN FIELD RESEARCH AND CURRENT ISSUES
IN ARCHAEOLOGICAL METHOD AND THEORY

Monograph 77

Monograph 76

ADVANCES IN TITICACA
BASIN ARCHAEOOLOGY-2

THE DEAD TELL TALES

Alexei Vranich and

Abigail R. Levine (eds.)

Monograph 75

THE STONES

OF

TIAHUANACO

Jean-Pierre Protzen and

Stella Nair

Monograph 70

CHOTUNA AND CHORNANCAP:

EXCAVATING AN ANCIENT PERUVIAN LEGEND
Christopher B. Donnan

Mara Cecilia Lozada and

Barra ODonnabhain (eds.)

Monograph 71

CRUCIBLE OF PUEBLOS: THE EARLY PUEBLO

PERIOD IN THE NORTHERN SOUTHWEST
Richard H. Wilshusen, Gregson Schachner, and
James R. Allison (eds.)

Monograph 68

THE CHANKA: ARCHAEOLOGICAL RESEARCH

ANDAHUAYLAS (APURIMAC), PERU
Brian S. Bauer, Lucas C. Kellett, and
Miriam Aroz Silva

IN

VISIONS

OF

T I WA N A K U
EditEd by
AlExEi VrAnich And
chArlEs stAnish

MonogrAph 78
cotsEn institutE of ArchAEology prEss
2013

thE cotsEn institutE of ArchAEology prEss is the publishing unit of the cotsen institute of Archaeology at uclA. the cotsen institute is a premier research organization dedicated to the creation, dissemination, and conservation of archaeological knowledge
and heritage. it is home to both the interdepartmental Archaeology graduate program and the uclA/getty Masters program in the
conservation of Archaeological and Ethnographic Materials. the cotsen institute provides a forum for innovative faculty research,
graduate education, and public programs at uclA in an effort to positively impact the academic, local and global communities. Established in 1973, the cotsen institute is at the forefront of archaeological research, education, conservation and publication and is
an active contributor to interdisciplinary research at uclA.
the cotsen institute press specializes in producing high-quality academic volumes in several different series, including Monographs,
World heritage and Monuments, cotsen Advanced seminars, and ideas, debates, and perspectives. the press is committed to making the fruits of archaeological research accessible to professionals, scholars, students, and the general public. We are able to do this
through the generosity of lloyd E. cotsen, longtime institute volunteer and benefactor, who has provided an endowment that allows
us to subsidize our publishing program and produce superb volumes at an affordable price. publishing in nine different series, our
award-winning archaeological publications receive critical acclaim in both the academic and popular communities.
the cotsen institute of Archaeology at uclA
charles stanish, director
gregory Areshian, Assistant director
Willeke Wendrich, Editorial director
randi danforth, publications Manager
EditoriAl boArd of thE cotsEn institutE of ArchAEology:
Willeke Wendrich
Area Editor for Egypt, north, and East Africa
richard g. lesure
Area Editor for south and central America, and Mesoamerica
Jeanne E. Arnold
Area Editor for north America
Aaron burke
Area Editor for southwestern Asia
lothar Von falkenhausen
Area Editor for East and south Asia and Archaeological theory
sarah Morris
Area Editor for the classical World
John papadopoulos
Area Editor for the Mediterranean region
Ex-officio Members:
charles stanish, gregory E. Areshian, and randi danforth
External Members:
Kusimba chapurukha, Joyce Marcus, colin renfrew, and John yellen
copyediting by peg J. goldstein
text and cover design by leyba Associates
Library of Congress Cataloging-in-Publication Data
Visions of tiwanaku / edited by Alexei Vranich and charles stanish.
pages cm
includes bibliographical references and index.
isbn 978-0-917956-09-6 (alk. paper)
1. tiwanaku site (bolivia)--congresses. 2. tiwanaku culture. 3. indians of south America--tiwanaku river Valley (bolivia)--Antiquities--congresses. 4. Excavations (Archaeology)--tiwanaku river Valley (bolivia)--congresses. 5. tiwanaku river Valley (bolivia)--Antiquities--congresses. i. Vranich, Alexei, 1968- author, editor of compilation. ii. stanish, charles, 1956- author, editor of
compilation.
f3319.1.t55V58 2013
984'.01--dc23
2013031790
copyright 2013. regents of the university of california
All rights reserved. printed in the united states of America

C ONTENTS
Authors........................................................................................................................................................................................viii
list of figurEs ...............................................................................................................................................................................ix
list of tAblEs ...............................................................................................................................................................................xii
prEfAcE .........................................................................................................................................................................................xiii
Alexei Vranich
chAptEr 1: Visions of tiWAnAKu .................................................................................................................................................1
Alexei Vranich
chAptEr 2: stylistic VAriAtion And sEriAtion .......................................................................................................................11
Michael Moseley
chAptEr 3: tiWAnAKu: A cult of thE MAssEs.........................................................................................................................27
Patrick Ryan Williams
chAptEr 4: tiWAnAKu And WAri stAtE ExpAnsion:
dEMogrAphic And outpost colonizAtion coMpArEd ....................................................................................41
Paul Goldstein
chAptEr 5: thE culturAl iMplicAtions of tiWAnAKu And huAri tExtilEs......................................................................65
William J Conklin
chAptEr 6: tiWAnAKu influEncE on thE cEntrAl VAllEy of cochAbAMbA ......................................................................87
Karen Anderson
chAptEr 7: tiWAnAKu rituAl And politicAl trAnsforMAtions
in thE corE And pEriphEriEs ..............................................................................................................................113
Matthew T. Seddon
chAptEr 8: tiWAnAKu origins And thE EArly dEVElopMEnt: thE politicAl
And MorAl EconoMy of A hospitAlity stAtE .................................................................................................135
Matthew Bandy
chAptEr 9: WhAt WAs tiWAnAKu? ..........................................................................................................................................151
Charles Stanish
chAptEr 10: nAturE of An AndEAn city: tiWAnAKu And thE production of spEctAclE ............................................167
William H. Isbell
chAptEr 11: sociAl diVErsity, rituAl EncountEr, And thE
contingEnt production of tiWAnAKu ............................................................................................................197
John W. Janusek
chAptEr 12: tiWAnAKus coMing of AgE: rEfining tiME And stylE in thE AltiplAno ................................................211
Patricia J. Knobloch
chAptEr 13: concluding thoughts ......................................................................................................................................235
Alexei Vranich
indEx ............................................................................................................................................................................................239

AUTHORS
Karen Anderson
university of california, santa barbara

patricia J. Knobloch
institute of Andean studies, la Mesa

Matthew bandy
sWcA Environmental consultants, Albuquerque

Michael Moseley
university of florida, gainesville

William J conklin
the textile Museum, Washington, d.c.

charles stanish
university of california, los Angeles
santa fe institute, santa fe, new Mexico

paul goldstein
university of california, san diego

Matthew t. seddon
church divinity school of the pacific, berkeley

William isbell
state university of new york, binghamton

Alexei Vranich
cotsen institute of Archaeology, los Angeles

John W. Janusek
Vanderbilt university, nashville

patrick ryan Williams

the field Museum, chicago

vi

F IGURES
figure 2.1. the two axes of stylistic variation portrayed as perpendicular to one another,
with different variants of the same style represented by different boxed symbols .........................................................19
figure 2.2. interaxial stylistic variation portrayed in a social temporal matrix, with different
variants of the same style represented by different boxed symbols ..................................................................................22
figure 3.1. Moquegua Valley map ..............................................................................................................................................28
figure 3.2. plan of the Wari colony at cerro bal ...................................................................................................................28
figure 3.3. plans of the tiwanaku colonies at (a) chen chen and (b) omo.......................................................................29
figure 3.4. comparative plans of tiwanaku monumental architecture ...............................................................................30
figure 3.5. Monumental architecture on the summit of cerro bal ....................................................................................31
figure 3.6. comparative plans of three scales of residential architecture at cerro bal and at cerro Meja .................33
figure 3.7. Examples of tiwanaku residential architecture at tumilaca la chimba and omo .........................................34
figure 4.1. the south-central Andes, showing vectors of tiwanaku colonization, Waris southernmost
outpost, and the Moquegua study area..................................................................................................................................42
figure 4.2. principal tiwanaku and Wari sites in the osmore drainage, showing the middle Moquegua Valley ..........46
figure 4.3. the indigenous setting. Map of pre-tiwanaku huaracane settlements of the middle Moquegua Valley ...47
figure 4.4. radiocarbon dates for huaracane-, tiwanaku omo, Wari-, and tiwanaku chen chen
style occupations, Moquegua mid-valley (1 calibrated ranges).......................................................................................48
figure 4.5. the tiwanaku colonies. Map of tiwanaku settlements of the middle Moquegua Valley ..............................51
figure 4.6. tiwanaku caravan route geoglyphs, including llamas and keros: (a) chen chen M1; (b) omo M10 .........52
figure 4.7. tiwanaku site group at ro Muerto, showing separation of contemporary omo (M70)- and
chen chen (M43)style sectors.............................................................................................................................................53
figure 4.8. tiwanaku red-slipped serving ware: front-face god keros from isla del sol (AMnh bandelier
collection), ro Muerto M70b cemetery (Moquegua), and copacabana (copacabana Museo del sitio) .................53
figure 4.9 (a) tiwanaku black polished vessels in the form of reclining llamas, ciriapata cemetery,
island of the sun (AMnh bandelier collection) and omo M12, Moquegua; (b)tiwanaku black
polished serving ware:titin uuayani and ciriapata cemeteries(AMnh bandelier collection);
and omo style, omo M12, Moquegua ..................................................................................................................................54
figure 4.10. Wari chicheria patio group, trapiche M7, structure 3 .......................................................................................56
figure 4.11. Wari serving ceramics, trapiche M7, huaracane neckless olla utilitarian vessel...........................................57
figure 4.12. relative surface frequencies of Wari and huaracane ceramics, Middle horizon
sectors, trapiche M7 ................................................................................................................................................................57
figure 5.1. portions of two painted textiles, each of which contains an example of a chavn figure:
left, an apparently low-status chavn figure; right, a high-status figure...........................................................................66
figure 5.2. A breechcloth is shown on this pucar sculptural figure....................................................................................67
figure 5.3. A sculptural stela found at Khonkho .....................................................................................................................67
figure 5.4. the two-dimensional figures incised onto the Kantatayita stone architrave
wear body banding similar to figures in chavn and pucar art but have very different
headdresses and ritual paraphernalia ....................................................................................................................................68
figure 5.5. A fragment of a tiwanaku textile from a burial on Quebrada Vitoria, chile ..................................................69
figure 5.6. A group of huari statues photographed after their discovery and assembly but
before they were transferred to the local museum ..............................................................................................................70
figure 5.7. reconstruction drawing of the Akapana pyramid as it would have been viewed
from the northeast....................................................................................................................................................................71
figure 5.8. the stela bennett, the largest and perhaps best preserved of the many tiwanaku
stelae, was recovered by American archaeologist Wendell bennett. ...............................................................................72
figure 5.9. A reconstruction drawing of a ceremonial huari textile believed to have been found near
ica, peru, but now in the textile Museums collection. .......................................................................................................73

vii

viii

Visions of tiWAnAKu

figure 5.10. left, a digital reconstruction of a tiwanaku tunic from a burial excavated
by padre gustavo lepaige in san pedro de Atacama, chile; right, a huari tunic ...........................................................74
figure 5.11. typical figures used repeatedly in the compositions of the two tunics;
left, tiwanaku; right, huari .....................................................................................................................................................74
figure 5.12. the weft interlocking creates sharp demarcations between colors in this detail
of a huari tunic. cotton warp (vertical), camelid weft (horizontal).................................................................................75
figure 5.13. A tiwanaku tunic textile fragment from a tiwanaku site west of Moquegua,
displays camelid weft interlocking but also uses camelid fiber warp ...............................................................................75
figure 5.14. A tiwanaku loom, illustrating how a tiwanaku tunic from the coyo oriente
burial grounds in san pedro de Atacama, chile, would have appeared in a tiwanaku loom.......................................76
figure 5.15. A huari loom, showing how one part of the tunic would have looked in its loom ......................................76
figure 5.16. the attendant figure in the center is from the carved sun gate; the figure
on the left is a detail from the tiwanaku tunic shown above it. the figure on the right
is from an early huari sleeved tunic, shown above it .........................................................................................................77
figure 5.17. A diagram showing an image of a tiwanaku tunic on a human figure and an image
of the gate of the sun ..............................................................................................................................................................78
figure 5.18. the mythological figures on tiwanaku and huari tunics seem to represent an image
of horizontally flying mythological figures that recede in size and into the apparent distance ...................................79
figure 5.19. A diagram showing an array of the figures characteristic of the arrangement
of tiwanaku and huari tunics as perceived from the wearers point of view .................................................................79
figure 5.20. the back side of the gate of the sun portal .......................................................................................................80
figure 5.21. A checkerboard huari tunic ..................................................................................................................................80
figure 5.22. this khipu, which is quite unlike inca khipu and is attributed to huari, makes
extensive use of multicolor wrapping....................................................................................................................................81
figure 5.23. sun gate undulating band .....................................................................................................................................81
figure 5.24. A comparison of three undulating bands, each with an animated head
at the turning points. bottom, the tiwanaku sun gate band; center, a reconstruction drawing
of the central band from the early textile Museum huari ceremonial textile; top, a characteristic
huari tunic design band ..........................................................................................................................................................82
figure 5.25. left, a tiwanaku square hat; right, a huari square hats ....................................................................................83
figure 5.26. left, detail image of a tiwanaku hat consisting of knotted construction; right,
detail image of a huari hat with pile construction..............................................................................................................83
figure 5.27. A group of cut stones in the site of Wari whose geometry and workmanship
seem remarkably like those found in tiwanaku ..................................................................................................................84
figure 5.28. A reconstruction drawing of a typical inca checkerboard tunic as it would
have appeared during its creation on the loom....................................................................................................................85
figure 6.1. Map of the south-central Andes showing tiwanaku and relative distances to peripheries
cochabamba and Moquegua. ................................................................................................................................................88
figure 6.2. Map of cochabamba valleys and sites mentioned in the text. ..........................................................................89
figure 6.3. cochabamba central Valley chronology and ceramic styles..............................................................................92
figure 6.4: comparison of illataco and piami decorated style frequencies as a percentage
of decorated sherds. .................................................................................................................................................................93
figure 6.5. increasing frequency of decorated ceramics from the late formative through the Middle horizon .........93
figure 6.6: Examples of common cVct-style drinking vessels from piami, including variations
on kero forms and challadores (funnel-shaped cups)..........................................................................................................94
figure 6.7. comparison of illataco and piami decorated style frequencies as a percentage of the
total assemblage ........................................................................................................................................................................95
figure 6.8. late piami rectilinear domestic architecture from piami...............................................................................97
figure 6.9. camelid jaw scraper tools from Middle horizon contexts at piami...............................................................98
figure 6.10. A tiwanaku-style snuff spoon from a late piami context .............................................................................99
figure 6.11. stone trompos from Middle horizon contexts at piami .................................................................................99

list of figurEs

ix

figure 6.12. diagram of various early Middle horizon tomb types at piami: (a) rectangular
stone-lined and stone-covered tomb; (b) rectangular cut pit tomb with a stone rim;
(c) round cut pit tomb with a stone rim and cover; (d) circular cut-walled tomb .......................................................102
figure 6.13. comparison of changes in pitcher and cup forms from the Middle horizon (right) and the
late intermediate period (left). All vessels are from piami ..........................................................................................105
figure 7.1. inca challa offerings................................................................................................................................................118
figure 7.2. An inca king consults with huacas .......................................................................................................................118
figure 7.3. small canals located in the tiwanaku iVperiod occupation of chucaripupata ..........................................122
figure 7.4. A complex of circular pits located in the tiwanaku iVperiod occupation of chucaripupata ..................122
figure 7.5. serving vessel sherds from the tiwanaku iVperiod occupation of chucaripupata....................................123
figure 7.6. circular pits in the tiwanaku iVperiod occupation of chucaripupata, post-excavation ..........................123
figure 7.7. Keros from tombs, late tiwanaku iVperiod occupation of chucaripupata..................................................124
figure 7.8. View of portion of massive outer retaining wall, tiwanaku Vperiod occupation
of chucaripupata....................................................................................................................................................................125
figure 7.9. overview of possible Kalasasaya structure at chucaripupata..........................................................................126
figure 7.10. View of interior storage structures, a possible Kalasasaya structure, chucaripupata .................................126
figure 7.11. plan of a possible Kalasasaya structure at chucaripupata ..............................................................................127
figure 8.1. the ponce Monolith ...............................................................................................................................................137
figure 8.2. left hand of the ponce Monolith..........................................................................................................................138
figure 8.3. right hand of the ponce Monolith.......................................................................................................................138
figure 8.4. Modes of commensality inferred from serving vessels: (a) Middle formative;
(b) late formative and Middle horizon.............................................................................................................................139
figure 8.5. Kalasasaya-style serving bowl, characteristic of the late formative 1............................................................139
figure 8.6. the ponce Monolith left hand; directional indicators marked by arrows......................................................145
figure 8.7. details of the gateway of the sun; directional indicators marked by arrows:
(a) central figure; (b) attendant figure, lower row .............................................................................................................145
figure 8.8. the flow of sami in the gateway frieze ................................................................................................................146
figure 9.1. the area of greater tiwanaku, an approximate 25-km radius from the city center......................................155
figure 10.1. tiwanaku vessel forms .........................................................................................................................................174
figure 10.2. contour map overlaying an aerial picture of the monumental area of tiwanaku......................................180
figure 10.3. photograph from Max uhle showing the flat area to the south of the Akapana ........................................181
figure 10.4. plan view of the palace at puruchuco................................................................................................................182
figure 10.5. comparison of monumental buildings in the Andes that have been interpreted as palaces ....................183
figure 10.6. omo M10 complex in Moquegua ......................................................................................................................185
figure 12.1. Map of tiwanaku..................................................................................................................................................212
figure 12.2. A museum-curated example of a Qeya-style vessel with incised decoration
of a rayed head deity............................................................................................................................................................212
figure 12.3. A museum-curated example of a Qeya-style escudilla with avian mythical figures
on the interior rim..................................................................................................................................................................212
figure 12.4. A museum-curated example of a Qeya-style escudilla with avian mythical figures
on the interior rim..................................................................................................................................................................213
figure 12.5. Map of the putuni palace courtyard area excavated by nicole couture and
Kathryn sampeck, between the Kherikala enclosure to the west and the putuni palace
to the east, indicating late formative 2 and tiwanaku iVperiod occupations
figure 12.6. Map of the Akapana East 1 area excavated by John Janusek..........................................................................214
figure 12.7. A museum-curated example of a Qeya-style incensario with stylized mythical
icons painted on the exterior and a modeled feline head at the scalloped rim.............................................................215
figure 12.8. A museum-curated example of a Qeya-style incensario with stylized mythical
icons painted on the exterior and a modeled feline head at the scalloped rim.............................................................215
figure 12.9. A museum-curated example of a Qeya-style vasija.........................................................................................215
figure 12.10. A museum-curated example of a tiwanaku-style red-slipped tinaja .........................................................216

Visions of tiWAnAKu

figure 12.11. Map of the Akapana A and b sectors in the northwestern terrace area,
excavated by linda Manzanilla and Mara rene baudoin .............................................................................................217
figure 12.12. The Akapana A and B sectors northwestern terrace area with: (a) a view of the top
surface of wall 1, the first terrace, and wall 2, looking southeast; (b) northsouth section of wall 1,
looking northeast...................................................................................................................................................................218
figure 12.13. A museum-curated example of a portrait cup................................................................................................219
figure 12.14. A museum-curated example of a tiwanaku-style kero with a rounded double torus..............................220
figure 12.15. Map of Akapana East 1 Mound area excavated by John Janusek ................................................................223
figure 12.16. Map of putuni palace courtyard area excavated by couture and sampeck,
between the Kherikala enclosure to the west and the putuni palace to the east, indicating
tiwanaku Vperiod occupations .........................................................................................................................................225
figure 12.17. A museum-curated example of a tiwanaku-style kero with a rayed head deity icon ............................228

TABLES
table 4.1. A continuum of Archaic state traits.......................................................................................................................43
table 4.2. radiocarbon dates from the Moquegua Middle Valley........................................................................................49
table 4.3. Moquegua Mid-Valley survey results by cultural Affiliation.............................................................................50
table 6.1. cochabamba Western, Eastern, and far Valleys, compared with Moquegua and tiwanaku Valleys ............91
table 6.2. carbon 14 dates for central Valley sites for the beginning of the Middle horizon ........................................92
table 6.3. tomb construction techniques for seated-flexed and on-the-side-flexed burials by
Middle horizon phase...........................................................................................................................................................101
table 6.4. Vessel forms by style in Early and late burials at piami.................................................................................103
table 12.1 comparison of correlated 14c samples for omo and chen chen styles ......................................................229

P REFACE
AlExEi VrAnich

dozen of the leading scholars in the field with active

research.
in the spring of 2006, the cotsen institute of Archaeology hosted a conference with invited speakers,
with a view to having each presenter allow time for
questions and discussion, which would continue into
break periods and mealtimes. the result of those
presentations and the dynamic discussion during
and after the conference are published in this volume. our rules were simple and straightforward:
present your data set and, from the perspective of
your research, answer the question: What was tiwanaku? the paucity of data is such that there is a danger that tiwanaku becomes an abstractionsimilar
to the role that Atlantis played in platos Republica
platform for scholars to showcase their writing and
rhetorical skills. We did not want the conference and
resulting volume to become an exercise in rhetoric,
with the outcome based on the debating abilities of
each presenter. Every opinion needed to be informed, and consequently the editors have tried to
draw upon a representative sample of scholars working on various aspects and in different regions of the
tiwanaku phenomenon. the contributors could criticize previous and present models, but in the end, all
had to propose and justify their own models.
in the first chapter, Vranich provides a general
overview of the changing views, or visions, of the
meaning and place of tiwanaku over the last half
millennium. thereafter, Michael Moseley describes
how the political climate of the middle of the twentieth century marginalized tiwanaku from mainstream developments across the Andes. tiwanakus
decades-long theoretical and intellectual isolation

esopotamia, the indus Valley, the nile

river Valley, china, the titicaca basin, the
north coast of peru, central America,
lowland Amazonia. these are the places where civilizations developed independently, largely untouched
by any outside influence. they are, in effect, locations
where archaeology can address two major questions
about the human condition: (1) Why, after thousands
of years, did people change from nomadic to sedentary agricultural lifestyles; and (2) why, after thousands of years, did these small, sedentary groups become builders and move into cities? one might
conclude that there would be substantial synergy
among scholars working in these places, since they
seek answers to similar questions and face comparable challenges. furthermore, their publications are
driven by the same termsfor example, complexity,
state, urbanism, hierarchy, and specialization, to
name a few. Exchanges between these scholars are
frequent and often formalized into organized symposia and edited volumes. however, in the end, the
results tend to disappoint, if not confound, when we
realize that were employing a similar vocabulary to
describe radically different data sets. As archaeologists, we first have to decide whether it is possible to
compare on a global scale abstract concepts generated exclusively from material remains, and if the answer is yes, then we have to determine how we can efficaciously incorporate and reconcile the results of
this cross-cultural research short of becoming experts in each respective area. this volume offers at
best a partial yet focused stab at this challenge by taking the aforementioned fifth geographic area, the
titicaca basin, and gathering contributions from a

xi

xii

Visions of tiWAnAKu

has put present-day scholars at a disadvantage. on

the other hand, these lacunae make tiwanaku and
the titicaca basin some of the more exciting places
to work because of the potential to redefine entire
periods of prehistory. Moseley possesses the authority to relate persuasively such a controversial story
about recent and present-day personalities and institutions, not only because of his broad experience in
the Andes but also because his own project, programa contisuyo in the Moquegua Valley, was a result of tension that forced scholars to conduct research outside of the political borders of bolivia.
As a result, areas outside tiwanaku are better
understood than the actual site itself. that situation
is changing quickly with new research, but it would
be fair to say that most of our recent ideas have been
developed outside the immediate heartland. the
order of the following chapters takes the reader from
the farthest-known reaches of the tiwanaku phenomenon to the titicaca basin and then to the ruins
themselves. the continuation the programa contisuyu in the tiwanaku paradise of Moqueguaa
description that refers to its pleasant climate and excellent preservation of archaeological materialshas
created a level of resolution of data sufficiently fine
that it has been possible to make deeper inquiries
into the identity of the people who ascribed to a tiwanaku way of life. paul goldstein and ryan patrick
Williams are two scholars who have continued and
expanded on the research goals of the programa
contisuyo Moquegua. goldstein primarily concentrates on the tiwanaku occupation, while Williams
has been developing research on the impressive
huari outpost on the summit of cerro bal. both
scholars consider defining the relation between the
colonist and indigenous populations of the valley to
reveal the basic organizational structure of these two
complex polities.
the elaborate weavings described and analyzed
by William conklin originated in the dry coastal valleys and oases, areas that are distant from the highland capital. An architect by training, conklin has
published extensively on Andean architecture but
has also found time to become an expert in textiles.
one reason for these seemingly divergent interests is
that conklin sees little difference in the manner in
which cloth and architecture are designed and conceptualized. Early in the development of Andean
studies, the iconographic similarities between ti-

wanaku and huari led some investigators to conflate

both cultures into a single entity. the two sites were
identified and differentiated long ago; since then,
their relationship has been one of the burning questions in Andean prehistory. conklin considers both
the mechanics of weaving and the arrangement of
symbolism to distinguish between tiwanaku and
huari, and he hypothesizes on their social and political structures.
Although the fertile bolivian valley of cochabamba offers neither the research synergy of the
drier valleys of coastal peru nor optimum conditions for preservation, Karen Andersons stratigraphic excavation of one of the few intact mound
sites in the valley complements previous large-scale
surveys of the cochabamba. Anderson notes a
process similar to the one that Matt seddon describes on the island of the sun, where imitation of
key tiwanaku elementssuch as the keroby the
indigenous population preceded any evidence that
could attest to the highland sites direct control.
circling geographically closer to the southern
titicaca basin, Matt seddon presents the results of
his excavation on the sacred island of the sun,
thereby documenting the first confirmed incursions
of the tiwanaku polity outside the heartland. fortunately for tiwanaku specialists, the comparatively
modest tiwanaku platform escaped the notice of
inca architects who monumentalized the nearby
area of exposed bedrock, as well as the notice of later
colonial looters who targeted the same area when
searching for fabled treasures. seddons broad and
deep excavation on this tiwanaku monument presents a relatively intact testament to the role of rite
and ideology in the expansion of a complex polity,
from the initial contact to what seddon interprets as
a forceful intervention of architecture of social and
political control.
charles stanish traces his academic genealogy
to the programa contisuyu. the influence of this
project is evident in his emphasis on directing largescale surveys, followed by selective excavation of
sites guided by specific research goals. in a series of
publications, stanish has presented a case for a selective tiwanaku occupation around the lakeshore
and along routes leading to important lowland and
tropic resources outside the basin. though his research is cited by scholars aiming to temper previous claims of a monolithic and militarized tiwa-

prEfAcE
naku Empire, stanish remains steadfast in his conviction that within much smaller, redefined boundaries, tiwanaku acted with intentionality and, if requisite, with force. stanish regularly visits many of the
cotsen institutes active research projects around
the world. these experiences are evident in his contribution, in which he draws upon analogy to other
premodern civilizations that, like tiwanaku, possessed a social and political infrastructure that one
would associate with a developed state despite small
primary centers and limited territorial extent.
John Wayne Janusek is often cast as the advocate
of a decentralized tiwanaku, in contrast to stanishs
centralized model. however, the reality is far more
complex. though both scholars disagree on key
points related to the foundations of power in tiwanaku society, they overlap in many other areas, including a similar concern about the wide gaps in our
knowledge, where an entrenched opinion is indefensible. Januseks ideas are based on more than a decade
of work at the core of the site of tiwanaku, where he
personally directed and oversaw the excavation of
residential, monumental, and agricultural remains.
the search for the origins of social complexity
took Janusek away from the heavily modified tiwanaku site to the neighboring valley of Jess de
Machaca. the well-preserved and shallow formative-period site of Khonko Wankane provided an
opportunity to expose broad architectural patterns
and would form the basic blueprint for the design of
the mature tiwanaku site. Just as influential to the
model Janusek proposes in this volume was the ethnographic setting for the ruins, where he was able to
observe how the archaeological site became the
venue for yearly reunions of the communities of the
valley.
Matthew bandy directed excavation and analysis on a project new to the southern basin. the
taraco Archaeological project, based on the penin-

xiii

sula of the same name, founded by christine hastorf at the university of california, berkeley, produced a generation of scholars attuned to the value
of the most modest of remains, such as seeds and
fish scales. cold objective quantification would be
the best way to describe bandys previous research
on prehistoric demographics. on the basis of an extensive survey of the taraco peninsula, followed by
a valley-wide resurvey of the valley bottom of the
southern basin, bandy has systematically evaluated
models for the development and growth of the site
of tiwanaku. in this contribution, bandy presents a
more speculative, and more theoretical, approach to
the foundations of political power at a mature tiwanaku from the perspective of the use of space and
the iconography of the monoliths.
the research of William isbell and his colleagues in southern peru was cut short by the violent peruvian civil war of the 1980s, but not before
huari became the archaeological exemplar of expansion and hierarchy. isbells familiarity with the
archaeological correlates of intentional, centralized
planning is the basis for his critique of previous
models of the organization of the site and polity of
tiwanaku. patricia Knobloch systematically reviews
ceramic pottery styles within their excavation context and relative or absolute dates with a view to refining the basic tiwanaku chronology. Knobloch
concludes that the period of tiwanaku expansion
was far shorter than previously thought, lasting approximately 200 years, from A.d. 800 to 1000.
in conclusion, i would like to thank the contributors for their patience and their willingness to review and modify their contributions over the long
period between the conference and the publication
of this volume. i would like to personally thank
lloyd cotsen for his continuing support in the institute and its mission to investigate, preserve, and disseminate our global archaeological heritage.

For Freddy Arce (1948-2011),

friend and dedicated Tiwanaku scholar

V ISIONS

OF

T IWANAKU

ALEXEI VRANICH

publicly acknowledging his gratitude to the indigenous Aymara, who viewed him as one of their own.
The presidential personnel made sure that the press
reported that the staff, historically made of wood
and bands of silver, contained signature material
from each of the provinces of modern Bolivia. Literally, Evo became the leader of all the indigenous
communities in the highlands; symbolically, he held
all of Boliviafrom the indigenous highlanders to
the white colonists of the Amazonin the palm of
his hand.
Even for his detractors, this ceremony was an
awe-inspiring event that appeared for one moment
to unite a country perpetually on the verge of dissolution. As unique as this may have felt to the enraptured audience, similar scenes had been played out
several times in the past. Two hundred years earlier,
a general leading an army seeking independence
from Spain had passed through the ruins and ordered that a notable but much abused piece of sculpture be returned to its upright position. Calling forth
the local inhabitants to witness the event, he declared
the end of foreign tyranny and the start of a new, and
free, republic. Nearly 200 years before that, Spanish
priests discovered that the strange and demoniclooking statues were still being worshiped by the natives. Uprooting and dumping these stone monoliths

n January 21, 2006, in a valley 13,000 feet

above sea level, a former coca leaf farmer
from the subtropics of the eastern slopes of
the Bolivian Andes walked barefoot past shattered
blocks of stone to the summit of an eroding mound.
Donning a colorful poncho, he gave thanks to the
gods for a good year and walked to the edge of the
temple, where his private moment of communion
and gratitude was interrupted by the roar of an estimated 30,000 people. Evo Morales had become the
first indigene elected to the highest office in Bolivia,
and on that day he was celebrating his symbolic inauguration costumed as a pre-Columbian priest. His
message proclaiming the end of 500 years of resistance, the start of 500 years of power, and the beginning of a new year for the original inhabitants of our
land resonated well with the mostly indigenous
crowd. The moment of his symbolic inauguration
was marked by the presentation of the traditional
staff of office, which was ceremoniously handed to
him by the leaders of the Aymara Indians, the indigenous people living in the highlands of Bolivia.
A new and powerful political force, the Aymara
had tumbled the previous two governments and
now had brought Evo, as he is known to one and all
in Bolivia, to power. Thus, by celebrating his inauguration at Tiwanaku in a traditional costume, Evo was

VISIONS OF TIWANAKU

into a prepared pit, the priests called for the natives

of the area to witness the end of pagan darkness and
the start of the enlightened Christian era. And two
centuries before that, even before anyone in Europe
was aware of the existence of an entire continent to
the west, a man held to be the direct descendant of
the sun arrived at the head of the largest army ever
assembled in the land. Seeing the ruins, he ordered
that the most beautiful temple be cleared and its fallen stones reset. Calling forth the local natives, he
proclaimed the return of their legitimate ruler and
the start of a new period of fertility and abundance.
In the following sections I delve in greater detail
into the methods and motivations behind various
interpretations of and interventions on the ruins.
Though the format is chronological, there is clearly
no implication of evolution. This selective review of
half a millennium of interpretations, written on paper, painted on canvas, or physically etched into the
stones on the site, serves to contextualize this publication and its contributions as part of a longer tradition of presenting a vision and the implication of the
ruins of Tiwanaku. Even if we declare neutrality, our
visions exist with a highly charged political and social context, in which archaeology and the archaeologist are elements in a high-stakes game involving
global forces. We do have the advantage of nearly unlimited access to centuries of descriptions and extensive funding for specialized analysis that can source
and date materials in a manner unimaginable just
decades ago. Nevertheless, these are all just tools, and
history has shown us that with enough time, our interpretations, or at least most of our interpretations,
will fade as new data come to the forefront and new
theoretical models make the previous frameworks
seem anachronistic.

tiplano and has provided its residents with a continuous supply of lake and shoreline resources for more
than 10,000 years.
At the near end of a funnel-shaped valley, 10 km
from the shoreline, is the site of Tiwanaku. Most of
the monuments one can see at the site date to a period from A.D. 500 to 1000. Prior to that epoch,
Tiwanaku was one of several villages in the Titicaca
Basin with a small temple and a cluster of associated
houses. After A.D. 500, Tiwanaku evolved into the
largest site in the basin both in terms of population
and monumentality. The stylized icons of the Tiwanaku culture, such as the Staff God and the Decapitator, were extensively reproduced on stone, clay,
wood, textile, and metal. At Tiwanakus apogee, these
artifacts were coveted by inhabitants of the shoreline
of Lake Titicaca and several distant valleys along the
slopes of the basin. The site itself was a collection of
stepped platforms and sunken courtyards in various
stages of construction and decay. Adobe houses concentrated near the monuments but spread out across
the valley bottom as far as the eye could see. Monumental construction and the use of Tiwanaku imagery on artifacts come to an end around A.D. 950.
The basin remains heavily populated and productive, but nothing comparable to Tiwanaku is built
during the pre-Columbian period. Over the next
half millennium, the story of Tiwanakuthe rise
and collapse of a powerful societywas told in the
form of a narrative in which the narrator stood at
the end of history. Visitors would puzzle over the
huge stones and wonder why the ruins appeared to
be sui generis: there was nothing as monumental either before or after Tiwanaku.

SETTING

At its height, the Inca Empire stretched the length of

the Andes, from modern-day Santiago de Chile to
Quito, Ecuador. This remarkable achievement was
reached through a combination of force, alliances,
bribery, and a flexible and forceful use of religion. The
Inca religion readily accepted other deities into its
pantheon, as well asto the confusion of the Spanish,
who attempted to create a universal history of the
Andes of their owna wide range of creation myths
revered by the various groups the Inca conquered.
However, the Inca were uncompromising in the belief
that their own creation myth was supreme and that

The site of Tiwanaku is located in the republic of Bolivia, in an area known as the altiplano, meaning
high plain or high plateau. Set between the two
primary mountain ranges of the Andean continent,
the altiplano is known for its high altitude (average
11,000 feet above sea level) and piercingly cold climate. At the lowest point of the altiplano lies Lake
Titicaca, the worlds highest navigable lake. The presence of this massive body of water moderates some
of the more extreme aspects of the climate of the al-

THE INCA

CHAPTER 1: VISIONS OF TIWANAKU

their status as firstborn to the world gave them the
right to conquer and subjugate. This claim was sorely
tested when Inca armies entered the Titicaca Basin
and stood before the eroding stones and towering
monoliths of a civilization far older than theirs.
With a tweak of mythic history and a great deal
of creative construction, the Inca transformed this
potential liability into a powerful symbolic resource.
Cycles of destruction and creation are common
themes in Andean cosmology, and the Inca ingeniously interpreted the ruins of Tiwanaku as the remains of a previous world destroyed by the creator
god Viracocha. Disillusioned with the behavior of its
inhabitants, Viracocha turned its residents to stone
and then made the sun, moon, and stars and subsequently created the first ancestors of all human nations, with the Inca being the first among first. This
having been accomplished, he left Tiwanaku and
journeyed across the landscape, naming places and
calling out the first couples from their pacarina, or
dawning places, to populate their respective homelands. In the meantime, the first Inca couple traveled
underground to the valley of Cuzco with instructions to conquer and bring all the trappings of civilization to the people of the Andes. Inca rule was
now based on their firstborn status at Tiwanaku,
and, as if to clear up any potential doubt, the sun
himself gave the first Inca, Manco Capac, a specific
mandate of rule over other Andean groups. Having
been descended from the sun, their supreme deity,
the Inca could thus claim special privilege and declare his legitimacy as a pan-Andean ruler.
To make this history tangible, the Inca invested a
huge amount of energy in constructions across the
entire Titicaca Basin, modifying and enhancing important ritual places. Tiwanaku became the terminal
point on an important ritual pilgrimage route that
began in the Inca capital of Cuzco. The best-preserved structure at Tiwanaku, the Pumapunku, was
renovated and refurbished, and a royal and religious
settlement, complete with a palace, bath (essential for
Inca rituals), kitchen, and associated plaza, was built
against this temple. The Incas even intervened in a
more aggressive fashion on the platform, emptying
out fill between the retaining walls to form three
chambers overlooking the plaza. Arriving in Tiwanaku, visitors and pilgrims would be hosted in the
refurbished and Incanized temple, where the mythic
history would be told through ritual performance.

The Inca, then, essentially carried out the first excavation and interpretation-driven restoration project
and, to continue this modern touristic metaphor,
built a complete planned settlement to receive,
house, and entertain visitors.

EUROPEAN INVASION AND COLONIZATION

For the Inca, Tiwanaku was the end of one world
and the start of the next. However, within a few decades, the Spaniards would arrive with the clear intent of razing the Andean world and starting a new
Spanish Catholic world. By the time Cieza de Len,
a young Spanish soldier contracted to write the first
official history of the Andes, arrived in Tiwanaku a
few short years after Spains incursion into the Inca
Empire, the site had been in ruins for nearly half a
millennium. Deriving no answers from the site itself
as to the purpose of the huge eroding pillars and the
strangely cut stone blocks or the identity of the
builders, he interviewed the local Aymara Indians
and asked them if the Inca built these ruins. Laughing, they said no and instead related the story of a
race of giants who moved the huge stones in a single night. Their reason and motivation behind such
a story are of course speculation, but based on my
own personal experience and a well-documented
history of interactions between the Aymara and more
powerful foreigners, I am firm in the belief that
nothing is said in a casual manner. In that light, this
story would serve the Aymara well by divorcing them
from the tangible evidence of a pagan past. They
could then claim, as indigenous writers would in the
coming decades, that they had always been Catholic,
though with a slight variation on the particulars.
More immediate was the partition of the former empire and its wealth. Land that belonged to the Inca
was being confiscated for the Spanish Crown or divided up by the conquistadors. Land that belonged
to those conquered by the Inca (and subsequently
liberated by the Spanish) could ostensibly revert to
its original owners. The largest, most visible claim by
the Inca to the fertile Titicaca Basin was debunked;
consequently, the question of tangible proof of ownership retreated into a nonhuman past. The next of
kin would be the Aymara.
The Spaniards found this explanation compatible with their own agenda. Like the Inca, their mandate to rule was divine, and their justification for the

VISIONS OF TIWANAKU

destruction of the Inca Empire was based on the idea

that they had liberated the continent from tyrants
who were in fact recent interlopers. The ruins of
Tiwanaku were not Inca and thus served to ridicule
their mythic history by showing that other societies,
perhaps even more amazing cultures that had been
destroyed by the tyrant Inca, had existed in the past.
The Spanish priest Sarmiento de Gamboa, commissioned to write an official history of the Incas and to
find a justification for the conquest, would present
this and other information to the remaining Inca nobility and force them sign a document admitting
their crimes. Indigenous writer Guaman Poma would
retort with a 1,000-page illustrated letter to the king
of Spain, presenting a Christianized version of Inca
creation and laying claim to the ruins on the basis
that the Inca emperor had built Tiwanaku after hearing of the buildings of Babylon and the Temple of
Solomon.

AGE OF THE TRAVELER-SCHOLAR

After the first century of the conquest, the site and
what it represented began to fade from the public
imagination. The economic and political center of
the Andes had shifted to the coast and the new city
of Lima; the high Andes and its ruins of stone became a dangerous backwater subject to periodic indigenous rebellions. Even though abandoned for half
a millennium and disassociated (at least officially) by
the local indigenous population, the silent monoliths
and carvings would occasionally offend Catholic
sensibilities and become targets of ritual mutilation
and burial. The ruins were rediscovered by a series of
traveler-scholars in the late eighteenth and the nineteenth century after the colonies achieved their independence from Spain. Though the site had changed
from a sacred location to a convenient quarry for the
haciendas, churches, and civic architecture of the
new colonial order, enough remained to draw wonder and awe. The Tiwanaku had a curious way of
building walls and revetments by placing large pillars
in between sections of ashlars. When all the small
well-dressed stone had been quarried, the remaining
pillars brought to the Western mind familiar places
like Stonehenge and Baalbek. A few of the much
abused monoliths remained, including the celebrated Gateway of the Sun, reset as a noble gesture to
the indigenous people but, unfortunately for archae-

ologists, without note of its original location. Most

curious were two monoliths set directly at the entrance to the church in the town plaza. According to
early travelers, these monoliths were brought from
the ruins by the local priest and commissioned to be
carved into statues of the saints Peter and Paul. One
could interpret this story, if true, as either the resourceful Aymaras way of both preserving a portion
of their past and avoiding the expense of making
from scratch one more set of saints for their colonial
masters, or as proof of the degree of accommodation
that had been reached by the Catholic authorities
and their indigenous charges.
It is of interest to note that Tiwanaku also became known to the outside world through popular
travel books that often highlighted, if not exaggerated, the mysterious and exotic aspect of the wilds of
the Americas. Descriptions of human sacrifices and
drawings of strangely dressed and impoverished natives fed readers appetites for the dangerous and exotic. Notwithstanding, many of the drawings, maps,
and later photographs of these early traveling scholars are invaluable resources for the present-day researcher, since they show many structures that have
been substantially modified or lost. Their other activitiescollecting artifacts for European museums,
elaborating theories on the origins of the Tiwanaku
people based on hyper-diffusion concepts of the
time, and spinning romanticized ideas about the rise
and fall of civilizationswould eventually lead to a
backlash in the midtwentieth century that continues to affect the field today.

BUILDING TIWANAKU
In the turbulent years after emancipation from
Spain, Bolivia lost a series of wars and ceded substantial parts of territory to its stronger neighbors.
The last major warthe tragic War of the Chaco
(19321935)destroyed the country both economically and psychologically. The exhausted soldiers of
the defeated army retained their army-issued firearms and straggled back to their villages, harboring a
deep mistrust for the state that had thrown them into
a useless conflict. Even before that disaster, the loss
of access to the sea in the War of the Pacific (1879
1884) had been etched deeply into the national psyche. To this day, a facsimile of a historic map of Bolivia before the War of the Pacific is displayed in gov-

CHAPTER 1: VISIONS OF TIWANAKU

ernment offices and private homes. The banner of
the lost coastal province of Litoral is carried along
with the colors of the present-day provinces in the
quasi-military parades endemic to every plaza space
in the country.
It was in this environment of defeat, xenophobia,
and mistrust that Carlos Ponce Sangins, an archaeologist and politician, formed an audaciously nationalistic agenda aimed at reconstructing Tiwanaku as
the pre-Columbian symbol of Bolivian national
identity. Whereas the broken beauty of the isolated
and eroding gigantic stones on the windswept plains
might represent the height of romance and mystery
to the foreign traveler and scholar, to Ponce Sangins
the stones must have been symbolic of the impotence of his poor and landlocked country. As excavations began in the late 1950s, rifle fire from the rusty
guns of the Chaco Waralready antiquated in the
1930srang out sporadically from the village in the
direction of the ruins; Ponce Sangins, with a machine gun emplaced on the summit of the largest
eroding building, responded in kind.
From his perspective, Ponce Sangins was firm
in the belief that foreigners had held sway in Tiwanaku for centuries and that the time for a Bolivian
perspective was overdue. Fine Tiwanaku pieces, perhaps even finer than those on display in Bolivia, were
exported under suspicious circumstances and placed
on display in European museums. He dedicated a
large space in his publications to attacking and discrediting previous ideasespecially exaggerated accounts and outlandish ideaswhile promoting
Tiwanaku as just as large and powerful as any ancient equivalent in the Old World. He would draw
upon the Incas as the theoretical analogue for his vision of an aggressive and expansive empire. The
heavily cemented and overly reconstructed pyramids
of Teotihuacn in Mexico would serve as the inspiration for his methods. The problem was that Tiwanaku had neither the magnificent setting of a Machu
Picchu nor the monumentality and size of Teotihuacn. Even more ironic was the fact that the more
he excavated, the less monumental the site appeared
to be. The celebrated standing pillars that seemed to
sprout from the ground were, in reality, part of a retaining wall covered for most part behind thickly
mortared revetments.
Ponce initially respected the archaeological evidence and rebuilt the primary building of the core,

the Kalasasaya, as a low stepped platform. Though

authentic, the effect was unsatisfying. Rather than
instilling a sense of awe, the stepped sides of the
platform became an obstacle course challenging visitors to scramble up the sides. Ponce halted and then
dismantled his initial reconstruction. In the future,
remains that were not sufficiently monumental were
labeled colonial or, worse, Inca and were taken
apart and reused in the reconstructions. What survived were those remains that were considered to be
from the classic or imperial period. The Kalasasaya was transformed from a low, stepped plaza with
multiple points of entry into a fortified platform
bounded by a single wall of monumental pillars. As
the national and international outcry over the reconstructions increased, Ponce withheld the field
notes and promised that soon all the data would be
published in a manner that would redeem his reconstructions.
Instead of containing objective data, Ponces publications were for the most part a forceful discourse
on the despotic and coercive power of Tiwanaku. The
boundaries of this empire, defined by the distribution
of a generously defined Tiwanaku style of artifacts,
corresponded rather well with the present borders of
Bolivia, but with the addition of access to the sea.
Tiwanaku was now an empire like that of the Aztecs
and the Egyptians and accordingly needed a pyramid
in the center of the site. The heavily eroded and damaged Akapana was the best candidate, but for the previous half millennium it had been described as a
fortress or a hill surrounded by rings of stone. Recent
archaeological investigations support the theory that
the Akapana was the most monumental version of a
long tradition of raised stepped plazas. With a terminological sleight of hand, the suffix pyramid was
added and repeated in all publications (books, newspapers, pamphlets) until it became inconceivable to
say Akapana without adding pyramid. In artistic
reconstructions, the summit was reduced to increase its pyramidal form. Bolivia now had a claim to
its original prewar boundaries; the evidence for this
claim was a monumental pre-Columbian capital,
complete with a pyramid.

RISE OF THE AYMARA

The twilight of Ponces tenure at Tiwanaku corresponded with the return of democratically elected

VISIONS OF TIWANAKU

governments to Bolivia (1982), and a process of

political and economic decentralization began to
redirect resources to local governments. Politically
and economically astute, the local Aymara Indians
clearly perceived the mutual value of the ruins as a
powerful symbol and as a source of sustainable income from tourism and research. A near half century
of archaeological work created an experienced labor
force that organized itself into a formal labor syndicate. Foreign and national projects pumped money
into the local economy in the form of wages for work
that involved much less labor than cultivating potatoes or grazing animals over hundreds of kilometers
of grassland. The number of visitors to Tiwanaku increased from several hundred a year around the middle of the twentieth century to an estimated 45,000
in 2008. (European and North American travelers
visit during the summer months of June, July, and
August, whereas national and South American visitors arrive during the rest of the year.) Gradually, the
ruins at Tiwanaku became an iconic stop along the
campaign trail for any aspiring or elected official.
The usual format for such a visit was for the candidate, most likely a mestizo dressed in a Western suit,
to preside over a small Aymara ceremony (a pago, or
gift to Mother Earth for good luck) and then deliver
a speech about the importance of heritage, the proud
history of the Bolivian people, and the need to continue investigations in the form of large-scale excavations that would employ a number of people. Both
sides would see the exchange as political theater, but
this inexpensive and purely symbolic offering would
resonate with both the countryside and the increasingly active urban Aymara.
In recent decades, the management of the site
and its budget was controlled from La Paz by two
government agencies: the Department of Culture
in charge of cultural events that range from a soccer
program for youths to the investigation and preservation of historic sitesand the Institute of Archaeology, a smaller and usually subordinate branch dedicated exclusively to the protection and investigation
of an estimated 30,000 archaeological sites from the
high Andes to the Amazon Basin. It was easy to figure out the primary interest of this institute: located
on Tihuanacu Street, this early twentieth-century
building was covered with brightly colored Tiwanaku imagery both inside and out. Starting in the
early 1990s, local Aymara communities and the mu-

nicipality of Tiwanaku began a series of maneuvers

to gain greater control over the management of and
the profits derived from the increasingly popular
ruins. After several years of escalating acrimonious
negotiations, the defining moment came in 2000
when a coalition of indigenous leaders took over the
ruins and adjacent facilities and expelled all state
employees. Archaeological projects, both national
and foreign, would continue to follow the norms established by the Department of Culture and the Institute of Archaeology, though occasionally a ministry in La Paz would block a foreign research permit
as a means to bait the Aymara back to the negotiating table. Hiring was decided at the local level by a
complex process that divided up the scarce jobs with
incredible accuracy across the 23 different communities in the valley. The most profitable day of ticket
sales, a festival held on the day of the winter solstice,
was completely under local control: traditional leaders of the communities, attired in red ponchos and
wielding staffs of authority, would collect tickets at
the entrance. The funds from this event and from
ticket sales during the year became an important
part of the working budget for the communities of
the valley and a constant source of friction as different communities and local organizations vied for a
larger share.
The roles and responsibilities of the local government, indigenous organizations, the national government, and international agencies would be the
subject of future negotiations and even conflict.
Notwithstanding, there have been moments of authentic collaboration and goodwill between local
and state authorities. The most high-profile collaboration was the repatriation of the largest piece of
Tiwanaku sculpture from its much abused location
in a traffic circle in front of the football stadium in
La Paz. The monolith had been uncovered in the
1930s and was moved several decades later by
Arthur Posnanskyfirst to the main avenue of the
city, then to a mock Tiwanaku sunken temple in
what was then a quiet area slightly outside the main
congested area of the capital. In subsequent years, the
capital expanded exponentially, and the tranquil faux
temple become isolated in a congested traffic circle,
suffering from the effects of urban pollution and
neglect. The return of the monolith was a stirring
event: a movable party carried it along the 72 km
from La Paz to the ruins. Indigenous-based cere-

CHAPTER 1: VISIONS OF TIWANAKU

monies shared the stage along with the usual ribbon
cutting and presentation of plaques and speeches
that are the mainstay of politicians. The moment the
monolith was reset at Tiwanaku was an emotional
one; the Aymara said that their giant had returned
home; the mestizos felt they had done their part to
right centuries of wrong.
The goodwill between the agencies in La Paz and
the local community lasted for several years. At the
time of the writing of this contribution, a critical
evaluation from UNESCO on the quality of excavation and restoration became the catalyst to expel all
the archaeologists of the institute from the ruins. The
new mayor of Tiwanaku, a young Aymara woman
with deep family roots in the town but educated and
socialized in the nearby capital, appointed a close
group of young archaeologistsboth Aymara and
mestizofrom the university in La Paz to assess the
situation. The news of a protracted argument and litigation among the archaeologists and government
agencies and ministries went viral across the Internet. The accusation that a pyramid was being reconstructed reprehensibly struck a chord with the Webbrowsing public; images of the reconstruction of the
pyramid became pictures of the week. Modern
technology assured that nearly identical versions of
the original story were reported across various news
agencies.

EXPERIENCE GLOBALIZED
The ruins must have been a remarkable sight when
they were a singular and difficult destination to
reach rather than one of the dozens of sites that your
average tour package will see in a given week. The
massive stones of Tiwanaku appear even larger to
the traveler who has spent days moving across the
seemingly inhospitable altiplano, seeing little else
than brown grass, brown adobe houses, and a few
roughly shaped stones that pass for an archaeological site. Within the cultural context of the Titicaca
Basin, the achievement of Tiwanaku is nothing less
than incredible. However much affection I feel for
Tiwanaku, I recognize that the ruins themselves can
be a rather underwhelming experience. Misleading
reconstructions dismay the increasingly savvy
tourist, and a failure to control growth has marred
the site and vistas with modern buildings of exposed brick and rebar. Spread out over several

square kilometers, the site has no recognizable natural boundaries apart from a barbed-wire fence that
runs around the main concentration of platforms.
Most tour groups are not even taken to the slightly
isolated ruins of Pumapunku, the temple the Inca
refurbished to commemorate the birth of humanity
that is now widely considered by experts to be the
most remarkable pre-Columbian construction in the
Andes. After two hours in the withering sunlight and
thin air, few have the energy to complain when the
tour bus drives by these remarkable ruins directly to
the highway and back to the capital. Unfortunately,
part of this indifference is the product of modern
travel. Most tourists arrive from the direction of Peru
and by then not only have seen Machu Picchu but
are most likely suffering from what guides call temple fatigue: the inability to appreciate ruins or to differentiate one set from the next. The general feeling
one gets from tour books is that Peru is the place to
go to see ruins; Bolivia is the place to go for natural
wonders and to avoid the crush of tourists in Peru.
The same global forces that thrust Tiwanaku
into an unfair popularity contest with Machu Picchu,
Tikal, and Angkor Wat have also provided access to a
well-developed and lucrative market. Every June 21,
thousands of people brave subzero temperatures to
stand in the early morning hours inside the low platform of the Kalasasaya. As the horizon over the low
mountains to the east begins to lighten, the crowd
becomes quiet; all eyes focus on the silhouetted figure of an important indigenous leader standing on
the lintel of a monumental gateway and waving the
multicolored indigenous flag of the Aymara. At 7:21
a.m. the sun breaks over the horizon several meters
to the north of the gateway. Film crews aligned with
the center of the gateway quickly snap up their tripods and move to the right, lowering their cameras
close to the ground in order to frame the sun with
the jambs of the gateway. The rest of the crowd,
packed shoulder to shoulder along the near-kilometer length of the platform, may not even notice that
the anticipated alignment never takes place. Even
fewer know that the celebration began in 1986, when
a dozen Bolivian archaeologists and their friends
were inspired to revive a solstice festival rumored to
have existed in the past.
The first festival was conducted with no reference to ethnographies and no consultations with
ritual Aymara specialists in the valley. In fact, the

VISIONS OF TIWANAKU

organizers primary sources of information were descriptions written by the Spanish of Inca ceremonies
that occurred 500 years after the collapse of Tiwanaku. The attendees of this first re-creation, numbering no more than 15, were all urban residents with
either a strong interest or formal training in anthropology; they brought with them a basic ritual kit
that can be purchased in the indigenous market for
special events such as the construction of a new
house. The second re-creation, done the following
year, required a bit more preparation; a serious attempt to contextualize the festival within the local
culture was made. An Aymara ritual specialist made
the basic preparations, and other indigenous authorities were invited. With lit torches, the mixed group
of urban mestizos and rural Aymara processed into
the Kalasasaya platform while two photographers
recorded the event. In the emotional moments of the
light of sunrise, two goals were established by the
core group: one was to have the large monolith in La
Paz returned to Tiwanaku, and the second was to revive the solstice ceremony. Twenty years later, a
boisterous and rather intoxicated crowd of 15,000
greeted the solstice sunrise in the Kalasasaya while
the original group either stayed home or stood by the
side with rather mixed feelings about what they had
created.
The irony of the present-day June 21 solstice ceremony is that the sun does actually pass through the
gate, both on sunrise and sunset, during the equinoxes. There are also several other archaeologically
confirmed solar, lunar, and stellar alignments that are
more revealing about Tiwanaku society. The choice
of the solstice sunrise as the important ceremony is
based on some basic rules of event planning: the
other alignments are either too small, are too subtle,
or take too long to be appreciated by a large group of
people. In the case of the March 22 equinox, the solar
alignment is perfect and dramatic, but it takes place
at the lowest moment of tourist visitationthat is,
during the highest period of rain. Sunrise over a desolate landscape 13,000 feet above sea level after a terribly cold and dark night is an impressive sight that
can be appreciated by thousands of people at the
height of the tourist season. The dramatic setting
and the pageantry of the event are part of the draw,
but what is being marketed is the idea that the past
can be experienced without enduring the complications of culture or society, or by going through stacks

of books that are, according to public perception, difficult to understand, often contradictory, and lacking
a satisfactory answer to the questions of how and
why. Outside of a concert or a political rally, most
visitors will have never felt such solidarity with such
a large group of strangers. This model has been developed and tested at other archaeological sites
around the world, such as Cuzco in Peru, Chichn
Itz in Mexico, and Stonehenge in England. Tiwanaku has bought into the successful solstice franchise with a ready-made market and financially has
benefited.

RESTARTING THE AYMARA WORLD

The fact that Evos inauguration was so similar to the
solstice festival comes as no surprise: decades of planning and organizing have provided the local expertise
and a blueprint for managing the large and often unruly crowds that gather on the fragile archaeological
site. The claim, privately whispered, that Evo Morales
would be no one were it not for the ruins of Tiwanaku ignores a series of globalizing forces and a tenacious, well-run campaign that lasted for several turbulent and occasionally violent years. But politics is in
large part both timing and performance, and in this
situation Morales was a consummate performer, making his entrance at the most opportune moment. Setting the stage was a time-consuming process going
back to the midtwentieth century, when the Agrarian Reform Law (1953) swept away vestiges of the
Spanish colonial land tenure system and allowed the
Aymara to vote, own property, and migrate to the city.
This act also had the effect of dissolving centuries-old
bonds between family, community, and land. Several
decades later the Aymara had become a powerful political force of related families and kin groups living
in the city or in the countryside. The festival took on
a special significance in 1992, when it became both a
protest against and an alternative celebration to the
quincentenary of the voyage of Columbus. The solstice would mark the start of the Aymara New Year,
and the date of 5499 was calculated by adding the
number of years since 1492, plus five cycles of 1,000
years. The explanation for this date is hazy at best, but
the implication is clear: the Aymara world had existed
before this date and continued to exist afterward.
By 2005, the year of the election, tens of thousands had attended the event, and braving the cold-

CHAPTER 1: VISIONS OF TIWANAKU

est night of the year was considered to be a rite of
passage and a reaffirmation of indigenous identity.
The entire country had become accustomed to the
image of thousands of people standing around the
ruins waving indigenous Aymara flags with mouths
full of coca leaves. However, after an evening in the
bitter cold and in very basic accommodations, most
urban people were done reconnecting to their traditional rural roots and were ready to return to the
comforts of the city. The rural Aymara, on the other
hand, remained in the countryside and continued to
work in basic if not arduous conditions. Nevertheless, the festival ingrained the idea of deep-seated
identity through a common experience.
Tiwanaku provided the perfect setting for a
major dramatic event, although it lacked the players.
Previous politicos had donned ponchos and appeared among the ruins, conducting Aymara rituals
with severe solemnity. Nevertheless, after the event,
the candidate would take off the poncho, shake the
confetti out of his hair, and, dressed in his Westernstyle suit and tie, return to the backseat of his chauffeured vehicle and be driven back to the capital. For
several reasons Evo could play the role on a national
scale, not the least of which was that he was Aymara

and as a result had inherited a motivated audience

already familiar with the euphoric feelings of unity
and solidarity that a proper performance could induce. After his successful campaign and inauguration, he also became heir to the ruins themselves,
which, unlike the hundreds of other locales that
served as settings for political rallies, would become
a mnemonic reminder for the story of Evo and the
place of the Aymara in the Bolivian republic.
Since the inauguration in 2006, the solstice festival has acquired an even greater symbolic significance: a united Aymara have elected one of their own
to the nations highest office. Morales, dressed in a
modest-looking alpaca sweater, still attends the festivals, but he sits in a spot that affords him a good view
of the sunrise and renders him nearly invisible to the
crowd. Politics require taking the center stage; but
politicians also must possess a degree of humility
and evince recognition; that is to say, there is no need
to be blatant when the obvious will do. Evo and the
Aymara no longer need to dominate center stage to
stake their claim for representation and power; the
mere existence of the ruins serves as a daily reminder of this claimuntil, of course, the next successful reinterpretation.

S TYLISTIC VARIATION
AND S ERIATION
MICHAEL MOSELEY

The one common denominator to visions of

Tiwanaku is that it was a center of some ilk. Such acknowledgment reflects the physical facts that the
scale of corporate construction and the elaboration
of monumental arts at the complex are both unprecedented and unsurpassed in the Andes south of
Cuzco. Consequently, it has long been recognized as
a special place in the Andean landscape and imagination. For the Inca it was a nexus of divine creation
and political authority; its status as a fountainhead of
the Inca ruling elite and their right to rule was invoked by later communities as well, from national
states to small farming communities.
For archaeology its centrality has long been associated with presumptions about core, expansion,
and periphery and linked inseparably with art, style,
and change. Tiwanaku was integral to basic ideas that
arose a century ago in Americanist archaeology. These
persistent concepts go unstated, but they profoundly
influence modern understanding of the past. Elucidating the critical suppositions that have crept into
the foundations of contemporary interpretation can
contribute to a more refined appreciation of what
ancient centers like Tiwanaku may have been.
I therefore review the very curious intellectual
history of what Tiwanaku has been to the discipline
because its trials and tribulations tell an insightful

his essay examines relationships between style

and seriation. It was provoked by the question
at the core of this symposium: What was Tiwanaku? This query has been asked repeatedly over
the last 200 years, and one can imagine that it has
been asked over the course of a millennium since the
demise of the great urban center of Tiwanaku. A long
history of research yields a consensus that the archaeological complex was a unique nexus of powerideological, economic, political, and socialand enduring symbolism. Yet what the sprawling ruins signified
for their creators vanished with their passing long
ago. Subsequent generations have therefore continually interpreted the monument anew as an emblem of
what they wish Tiwanaku was.
For the Inca it was a very special place where the
creator, Viracocha, sculpted the first humans from
Lake Titicaca clay. Yet the Inca encountered the metropolis in ruins, according to Cieza de Len (1976
[15531554]), one of the sites early colonial visitors.
As a kind of monumental Rorschach test, Tiwanaku
has represented diverse things to different people
and to various social and political movements.
These changeable readings of the ruins are well reviewed by Stanish (2002), and further dimensions
will be added as future generations redefine the past
anew (Albarracn-Jordn 1999, 2014).

11

VISIONS OF TIWANAKU

12

story about theoretical approaches to the past that

the present is still tied to. The origin center horizon of civilization concept that Tiwanaku fostered is
reviewed next in the context of how and why phases
of seriated sequences change. Finally, I elucidate the
two different axes of stylistic variation.
Well recognized is the fact that styles change
over time and the diachronic axis of variation is essential to the construction of seriated chronologies.
Ignored is the fact that, as in linguistics, dialects of
style constitute a significant synchronic axis of variation. Schools of fashion differentiate people who
share the same culture. Yet this blatant reality remains unacknowledged by the discipline because
seriation cannot distinguish which axis is the source
of variation. It can only assume that all significant
differences are products of change over time due to
cultural change. Consequently, the methodology
can and has delivered misleading sequences that
obfuscate the dynamics of concurrent social differentiation.

FROM CORE TO PERIPHERY

Of all the great ancient centers of the Andes, Tiwanaku was the first to receive broad international
recognition. Many early visitors published commentary on and illustrations of the ruins (Stanish 2002).
However, it was the German investigator Alphons
Stbel who initiated a vital turn toward systematic
description by methodically measuring, drawing,
and recording the many archaeological remains in
18761877. Upon returning to Berlin, Stbel brought
in a young museum assistant, Max Uhle, to help digest and report on the large compendium of data
that was published as Die Ruinensttten von Tiwanaku in 1892. As the title implies, the monument was
considered a ruined city, and its lavishly illustrated
arts differed from those of the Inca represented by
collections in European museums that Uhle and
other scholars had studied. Furthermore, the style
could be recognized as pre-Inca on the basis of early
Spanish chroniclers observation about the relative
antiquity of the site (Stbel and Uhle 1892). The
well-received volume appeared when the discipline
of archaeology was in its infancy, and it promoted
immediate international recognition of the imposing ruins as a premier New World monument. Consequently, Tiwanaku has long been a privileged cen-

ter of attention, scholarly debate, and intense speculation.

Upon first traveling to South America, Uhle
planned to excavate at Tiwanaku. Visiting the ruins in
1894, he found the military using stelae for target
practice. Horrified, he publicly denounced the destruction and created a scandal in the press. The government then prohibited all excavation. Completely
stalled on receiving a permit, the frustrated investigator moved to Lima the next year (Rowe 1954). Uhle
supported his endeavors by generating collections for
foreign museums. Grave accompaniments provided
high returns, and a grand ceremonial center, Pachacamac, a short trip south of Lima, was long known for
producing numerous tombs, so he excavated there for
10 months in 1896. Deep deposits yielded a fourfold
stratified sequence of mortuary remains that Uhle
was able to sort out on the basis of his prior museum
studies and Bolivian visit. Inca-related remains occurred at the top. They were followed by graves with
local styles of pottery found above and below a layer
of burials with goods that he recognized as Tiwanaku. The research results were written up in 1897
and published in 1903 (Kaulicke 1998). The monograph was a milestone because time-depth, stratigraphy, and seriation were previously unacknowledged
in the New World. With sponsorship from the University of California, Berkeley, Uhle broadened his investigations of coastal cemeteries, and by 1901 his
studies had ranged from the southern Ica drainage
through the northern Moche Valley (e.g., Uhle 1913).
Most remains were not stratified, but local sequences
could be roughly seriated by comparison with the
Pachacamac stratigraphic sequence. This led to the
conclusion that Inca and Tiwanaku materials were
very widespread in comparison with intervening
local styles that varied from region to region.
How could these coastal discoveries be explained? As in any pioneering investigation, variation
and heterogeneity came across as noise, whereas
similarity and consistency commanded his attention.
Focusing upon the latter, Uhle turned to Spanish accounts about Tahuantinsuyu to fathom Tiwanaku
(Kaulicke 1998; Menzel 1977; Rowe 1954). He extracted four interpretative tenets from the chronicles. First, prior to the ascendancy of Cuzco, conditions of barbarism and savagery reigned in the
Andes. Second, civilization was singularly invented at
Cuzco by a creative monarch, Pachacuti. Third, it was

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

then disseminated across the Cordillera as a horizon
of enlightenment by Inca legions within a historically
very rapid period of time. Fourth, the short-lived imperium collapsed abruptly in the wake of Francisco
Pizarros military campaign.
Finding these tenets attractive, by way of analogy
Uhle (1903, 1913) proposed that ancient Tiwanaku
was the origin center and capital of an advanced preHispanic polity that had incorporated much of the
central Andes for a relatively brief period. Furthermore, the far-flung arts issuing from the capital constituted a distinct horizon style, similar to that of the
Inca, and thus the Tiwanaku style became a chronological marker for ordering the local and regional
archaeological sequences. With this bold stroke of interpretation he transformed Tiwanaku from its 1892
Ruinensttten status as a major archaeological city
into a nexus that was the fountainhead of a far-flung
horizon of ancient civilization. Thus the age-old question What was Tiwanaku? could be asked of both
the core and its periphery, as the present volume does.
Uhles vision of Tiwanaku is credited for introducing the horizon concept to American archaeology (Willey and Phillips 1958). Historically, his origin center horizon of civilization vision was also
exceptionally important as the pioneering use of an
indigenous analogue, the Inca, to interpret preColumbian archaeological remains. Uhle could well
have drawn upon classical antiquity, other Old World,
or evolutionary schemes to explain Tiwanaku. Instead, he developed an original Andean model that
was highly appropriate for the region. Unfortunately,
this innovative lead in pursuing indigenous interpretative models was not pursued by subsequent foreign
investigators.
Uhles formulations might have attracted scant
attention due to his dearth of substantive publications other than the 1903 Pachacamac report. Fortunately, Phoebe Apperson Hearst was the patron of
his explorations from 1899 to 1905, and the majority of his important coastal collections from the Ica
through the Moche valleys ended up at the University of California in Berkeley, where A. L. Kroeber
had founded the Department of Anthropology
(Menzel 1977). Kroeber had been skeptical of timedepth in the Americas until 1915, when he collected
and seriated surface sherds from around Zui
Pueblo in the Southwest. This sparked his interest in
working in either Mexico or Peru. The latter won out

13

because of the availability of the Uhle collections

and because Kroeber could have the field virtually
to himself, which he did for several decades. This led
to the establishment of the Berkeley School of Peruvian archaeology, which long dominated the
course of scholarship pursued by U.S. and foreign
investigators.
Uhles legacy came about because Kroeber had a
young student, William Duncan Strong, who needed
support and was put to work on the Peruvian collections in 1922. Other students joined Kroeber in the
undertaking, and by 1927 all of the materials had
been published. The study procedures entailed spatially laying out all the pieces in a particular collection. Those with the most obvious similarities were
then grouped together to create stylistically discrete
entities. Finally, seriation was employed to order the
entities chronologically (Rowe 1962c). The end results largely confirmed and formalized Uhles more
intuitive seriations. Publication was critical in making
the materials clearly intelligible and useful to other
scholars (Rowe 1954). In so doing, it reaffirmed the
contention that Tiwanaku was an ancient cultural
and stylistic wellspring of expansive geographic
scope, broadly analogous to imperial Cuzco.
It is curious that Kroeber followed Uhles lead
down several questionable paths. Provided the opportunity to excavate in Peru in 1925 and 1926, he
focused almost exclusively on the recovery of tombs,
as had his predecessor. Yet in the Nazca Valley he did
open a cut in a refuse deposit using the European
model of excavating in 50-cm arbitrary levels. Unfortunately, change in domestic ceramics was not
readily evident in sequential levels. Nonetheless,
Kroebers protg, William Duncan Strong, later pursued such excavations when he and students Gordon
Willey and John Corbett investigated sites on the
central coast in 19411942. Digging into sloping natural strata at Pachacamac and elsewhere often resulted in the recovery of mixed assemblages. Nonetheless, these assemblages included a far greater
range of ceramic formsparticularly domestic utilitarian waresthan were represented in mortuary assemblages. Yet there were few such excavations.
Confronted with creating a ceramic chronology
from surface sherds for the 1946 Vir Valley Project,
James Ford (1949:31) observed, Indeed, with few exceptions of fairly recent date, it can be said that virtually all published knowledge of the archaeology of

14

VISIONS OF TIWANAKU

the Peruvian-Bolivian areas has dealt with tomb furnishings. This emphasis has an effect on current interpretations of Peruvian prehistory because, as he
noted, the mortuary focus excluded 95-plus percent
of the broader ceramic record and the domestic settlements of common folk associated with it. Kroeber
(1944) may have become aware of the broader archaeological potentials by the time of his last trip to
Peru, in 1942, when he toured and recorded insightful observations on a large number of sites and monuments. Nonetheless, the Berkeley School never deviated from Uhles focus on mortuary remains.
When Uhle worked at Pachacamac and New
World archaeology was devoid of time-depth, it is
understandable that he envisaged Inca and Tiwanaku arts as standardized and relatively homogeneous. These qualities made them regional temporal
markers. Indeed, the essence of horizon styles has
always rested on the unity, not variability, of their
arts. Kroeber and his students accepted these tenets
as the Uhle collections were studied and published.
Given chronological concerns, notions of stylistic
homogeneity spilled over to Ica and Moche fine arts.
With the publication of the Uhle collections, the discipline was well disposed to conceptualize these and
other bodies of mortuary ceramic arts as whole cultural styles, as Kroeber would call them in 1957. Stylistic homogeneity is an essential premise of seriation, and this is the means by which chronologies
have been constructed for a century. What Kroeber
practiced in Peru was curiously divorced from what
he preached theoretically in Style and Civilization.
This 1957 volume defined whole cultural style and
attacked it as a misleading notion that failed to capture variable aesthetic expressions reflecting cultural
diversity and heterogeneity. Curiously, the book was
not listed as a relevant archaeological publication in
Kroebers American Antiquity obituary (Rowe
1962a). The discipline remained content with the vision of style that Kroeber and his students set forth
in publishing the Uhle collections. These publications reaffirmed the sanctity of horizon styles and
cemented the horizon concept into the foundations
of Andean archaeology as foreign investigators have
practiced it.
Beginning with Julio C. Tello, Peruvian scholars
have been far less sanguine about the utility of the
horizon concept than their overseas counterparts
(e.g., Lumbreras 1974). If for no other reason, this

reticence reflects the fact that the rugged topography

of the Cordillera fosters and shelters ethnic, social,
and artistic diversity that conquests, including those
of the Inca and Spanish, failed to homogenize into
tidy epochs of cultural unity. It is my understanding
that Julio Tello viewed the evolution of indigenous
societies as a highly intertwined process with roots
and branches that diverged out but often merged
back together (Burger 2014; Tello 1960). This vision
seems remarkably akin to Kroebers (1948) tree of
culture with diverging and converging evolutionary
paths. Yet Tellos views were anathema to his North
American colleagues. These colleagues literally
straightened things out (Willey 1945). Temporal
continuities were linearized into traditions and
Chavn de Huntar was co-opted as the hearth of a
new early horizon. The latter prompted the first
methodological scrutiny of the horizon concept,
particularly by Gordon Willey (1948; Willey et al.
1956). Inspection focused largely upon the horizon
style notion and its characterization. Although there
has never been a horizon without an origin center,
centers went unexamined and the Inca archetype
went unquestioned.
Horizon styles figured in a marvelous 1946 experiment in Andean archaeology: the Vir Valley
Project. This was the first broad-perspective, multidisciplinary undertaking of its kind to focus on longterm human adaptations in a single study. Excavations were undertaken at a number of different sites
by a number of archaeologists, and they brought to
light a wealth of artifacts and data that had received
little prior attention. It also marked the first systematic survey of a broad range of prehistoric sites. The
survey required dating sites from surface sherds,
which James Ford (Ford and Willey 1949) undertook
using innovative methods of quantitative frequency
seriation that brought common utilitarian ceramics
into focus. These controls, in turn, allowed the temporal and spatial distributions of sites to be analyzed
as changing settlement patterns, and this opened an
entirely new vista for archaeological inquiry.
The pathbreaking Vir Valley Project might have
spurred Andean archaeology to move beyond its traditional fixation with seriating mortuary remains.
However, none of the participating archaeologists returned to actively engage in field studies. Rather,
many elected to work elsewhere or to pursue museum studies. This created a profound void in Peru-

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

vian studies. It was shortly filled in 1948 when John
Rowe joined the Berkeley faculty and found that he
had the field to himself, as Kroeber had before him.
Hired to teach both linguistics and archaeology, Rowe
was a brilliant scholar with an impeccable command
of Latin and Greek and deep intellectual roots in classical archaeology (Menzel 2006; Schreiber 2006).
Classical education led him to view ancient objects
individual ceramics and other specimensin the
manner of an art historian, emphasizing style rather
than the typology that characterized the Vir research. Consequently, the course of Peruvian archaeology turned back to Uhles original focus on tombs
and seriation.
Tiwanaku had reigned supreme in Andean archaeology for nearly half a century before the sierra
center of Huari was acknowledged to be the source
of broadly similar arts and influences in Peru that
Uhle and his disciples had erroneously attributed to
Bolivia. The recognition of Huari reconfigured the
spatial scope of the Tiwanaku horizon but did not
eliminate it. Significantly, it raised the very fascinating evolutionary spectrum of two first-generation
archaic states emerging concurrently in the Andes.
Unfortunately, critical research issues surrounding
the emergence of paired states were quickly obfuscated. In 1960 John Rowe advocated a new, unified
chronological system for the central Andes. Previously, foreign investigators had employed rather
idiosyncratic schemes, often invoking evolutionary
stages that blended style and time (Rowe 1945).
Therefore, standardizing procedures to clearly distinguish between cultural and temporal change was
certainly needed.
To this end, Rowe (1960) proposed two interrelated things. The first was that the discipline should
organize dating and interpretation on the basis of a
sequence of temporal units, each strictly defined and
subdivided on the basis of calendrical dates. The
commendable intent was to divorce the components
of chronology from the vicissitudes of cultural
change, thereby avoiding problems inherent in developmental stages (Rowe 1962c). The calendrical
brackets of archaeological time might have been labeled in an entirely neutral framework by tagging
them numerically or alphabetically. Instead they
were labeled horizons and periods. Unfortunately,
this nomenclature did little to divorce the discipline
from thinking along the lines of differences in cul-

15

tural content, much as Uhle and Kroeber had done.

Thus today, when archaeologists talk about the Late
Horizon, they tend to envisage the Inca rather than a
narrow 42-year span of absolute time.
This narrow time span is a product of Rowes
second proposition. It contended that the temporal
units of his pan-Peruvian system were to be defined
and dated on the basis of a master sequence that
Rowe and his Berkeley associates were formulating
for the coastal Ica Valley (e.g., Menzel et al. 1964). To
paraphrase Menzel (2006), the foundations of this
formulation held that archaeological style analysis
dealt with communities of art created in workshops
adhering to standard models of widely acceptable visual communication in a culture. Arts could therefore be analyzed by procedures analogous to those of
linguistics to elucidate the structure and vocabulary
of aesthetic expression and to chart, in detail, their
changes over time. Indeed, these procedures produced a very fine-grained sequence for Ica that was
essential to the pan-Peruvian calendrical system it
supported. The systems Late Horizon began in 1476,
when the Inca conquered the valley, and ended with
its Spanish incorporation in 1534. This 42-year precision dating came not from seriation but from
Rowes interpretation of ethnohistoric documents.
What, then, is the dating precision of earlier periods
and horizons of the calendrical superstructure? They
were derived from 14C assays. Because the assays and
the master sequence did not always agree, there was
very little incentive for the Berkeley School to invest
in statistically significant suites of radiometric constraints. Indeed, because few absolute dates were run,
a pick-and-choose philosophy prevailed, and the
only acceptable 14C dates where those that agreed
with the master sequence.
That Uhles Ica grave lots underwrote the longest
detailed sequence in the Andes is a scholarly monument to Rowe. Rowe was not an excavator except on
rare occasion. He disdained cultural evolution and
was highly suspicious of the chronological efficacy of
stratigraphy because excavation by arbitrary levels
had characterized most U.S.-led investigations in the
1940s and often resulted in mixed assemblages
(Rowe 1961). His penchant lay with the nuances of
seriationspecifically qualitative versus quantitative
seriationand he believed that ancient styles could
be dissected into very short divisions spanning a
generation or two (Schreiber 2006). Consequently,

16

VISIONS OF TIWANAKU

the Ica master sequence was based upon seriated

mortuary ceramics in collections made by Uhle and
collections of looted pieces assembled by others.
Although the chronological framework proposed by Rowe is now prevalent, it did not arise out
of scholarly deliberation and consensus. It was delivered by fiat. And it was imposed because Rowe was
the dominant foreign archaeologist working in Peru
at the time. Furthermore, Berkeley was the principal
center for U.S. graduate studies in Andean archaeology, and it produced cadres of disciples. Expectably,
Peruvian archaeologists working in the tradition of
Tello did not embrace the scheme. However, foreign
investigators were strongly compelled to do it the
Berkeley way if they contemplated working in Peru.
Rowes chronological system effectively purged
Tiwanaku from the purview of the Middle Horizon
by recognizing Huari as the origin center of stylistic
influence reaching Ica and the other coastal valleys
where Uhle had collected. This left Tiwanaku looming large but literally beyond the horizon. Indeed,
out of sight, out of mind was the case, because the
Bolivian center is not acknowledged on any of Rowes
chronological charts.
Methodologically, it can be asked if the Berkeley horizon scheme could possibly accommodate
two concurrent centers of regional influence. Obviously not if one, Huari, was 250 km away and the
other, Tiwanaku, was 750 km away from the Ica
master sequence. In theory, banishing the altiplano
center to the periphery was not important because
Huari manifestations in Ica marked a temporal
horizon that could be projected to the highlands of
Bolivia. In reality, this presumption never worked
due to cross-dating problems. Radiocarbon assays
for the Ica sequence are not always internally consistent. Furthermore, sequences in other Andean
areas are not always well calibrated. Thus cross-correlations made on the basis of absolute dates can be
tenuous, particularly as distance increases. Crosscorrelations made on the basis of trade items or
items exhibiting stylistic influence become more
tenuous with increasing distance from the master
sequence. In part, this is because the small Ica Valley did not regularly export art to adjacent
drainages or exert regional influence. Consequently, temporal horizons are defined by foreign
influence arriving in Ica, not by Ica influence flowing to other centers. Again, distance is a critical

variable. Cuzco and Huari are relatively close to the

coastal drainage, and their stylistic influence is
quite evident. This cannot be said of Chavn de
Huntar some 600 km away. Its stylistic influence in
Ica requires a greater stretch of the imagination.
Obviously, things would be organized very differently if the master sequence lay in a Pacific drainage
north of Lima or south of Arequipa. Consequently,
the temporal framework of Peruvian archaeology,
as practiced by foreigners, is tied by historical accident to where Uhle chose to excavate and to which
of his Berkeley collections were amenable to detailed seriation. Serendipity, not rational research
design, continues to structure interpretation of the
past.
No objections were raised to marginalizing Tiwanaku, for various reasons. Rowe had great disdain for
Carlos Ponce, a nationalist political figure who controlled Bolivian archaeology and excluded foreign
investigation in the country. Thus there was no
scholarly constituency to analyze the implications of
Rowes chronological system or to argue against the
purge. The net result was that Tiwanaku and its entire Titicaca hinterland were marginalized from the
prehistoric Peruvian scheme of things. This did not
diminish the importance of Tiwanaku in the eyes of
Argentinian, Bolivian, or Chilean archaeologists. Indeed, creative insights and provocative proposals
about the altiplano center thrived when it was free of
Berkeley School constraints. Thus, when barriers to
outside research in the Titicaca region ameliorated
and the pace of investigation accelerated, there was
no allure for alignment with the distant Ica master
sequence. Consequently, a very different chronological and organizational framework emerged. It is
young, vibrant, and finding its way, as the contributors to this volume illustrate.

CENTRALITY AND VARIATION

Having weathered the slings and arrows of archaeological interpretation, Tiwanaku is not yet liberated
from suppositions that accompany the origin center
horizon of civilization concept. Presumptions
about art, style, and variation are embedded in the
concept, and they extend to other ancient centers as
well. Uhles Inca archetype envisaged Cuzco as the
singular fountainhead of imperial-style art and architecture, which was then spread throughout the

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

political realm. Implementing this interpretive
model entailed two presumptions. First, Cuzco was
stylistically homogeneous and encapsulated the pristine expression of the Inca imperial whole culture
style. And second, as this imperial horizon style radiated across the conquest landscape, more variable
expressions transpired in hinterlands, where the arts
were not always faithfully duplicated. Faltering hinterland purity of origin-center fashion has long been
a problem in accurately identifying horizon-style expressions in peripheral areas.
The problem plagued Chavn when U.S. investigators pronounced it a horizon-style origin center.
Recognizing that fashion, like beauty, lies in the eye of
the beholder, Willey (1951) cogently proposed that
the Chavn horizon style should be strictly defined by
the art and iconography carved in stone at the origin
center of Chavn de Huntar. Reflecting the supposition that the most pristine expression of the horizon
style occurred at the origin center, Willeys proposal
remains a highly viable procedure for operationalizing the origin center horizon of civilization concept. Yet it won few adherents in a profession lacking
self-discipline. Investigators instead pursued more
idiosyncratic characterizations, to the point that the
term Chavnoid arose as a catchall for a morass of
supposedly early materials. This definitional problem
persists, and it afflicts all archaeological centers and
their associated styles, including Tiwanaku.
Ensuing scrutiny of the monumental arts and
architecture at the Chavn origin center revealed stylistic variation. This did not conform to the Cuzco
archetype of presumed homogeneity. How, then, were
graphic differences at the early center to be explained? After visiting the ruins and examining their
construction, stone carvings, and other arts, John
Rowe (1962b) produced a widely acclaimed seriation
of the Chavn style that divided the monumental
complex into an old temple and a new temple annex
and the arts into four chronological phases characterized by different stylistic expressions. The ramifications for the early horizon were far reaching.
Equally important, Rowes Chavn work set important precedents. The first was that whole culture
styles were encapsulated at their origin centers. The
second was that if stylistic change at the origin centers could be identified and calibrated, then parallel
change was expectable in the hinterland areas of influence.

17

Phasing
This interpretive framework structured my vision of
the Huacas del Sol and de la Luna during investigations in the early 1970s. At the time, this Moche complex was considered the center and capital of the first
archaic state to arise in the Cordillera. Its art, considered a unified style, had been previously seriated into
a fivefold sequence by Rafael Larco (19381939,
1948), and this was a very highly respected sequence.
Although the manner of vessel painting often differed between phases, Larcos (1948) sequence relied
principally upon form differences in the spouts of
mold-made stirrup-spout vessels evident in his private collections from the Moche and Chicama valleys. Rowe immediately checked the proposed sequence against the Uhle Moche collection of 31
grave lots with 600-plus objects and pronounced it a
viable one (Donnan 1965). Indeed, in many ways the
Larco seriation was an ideal one, with simple phase
distinctions of broad geographic scope.
Common to all such schemes, Larcos presumed
rapid transitions from one phase to the next. Indeed,
without short, crisp transformations, seriated units
would overlap and lack chronological utility. At Sol
and Luna, I assumed that stylistic differences at the
center were products of change over time that then
radiated out to the broader Moche realm. Yet I was
very perplexed by questions about how and why the
seriated phases changed. Archaeological literature
provided little guidance because this was not something Andeanists worried about. The technological
basis for implementing Moche stylistic change was
not a problem. Yet how this was done mystified me. I
wondered if a great gong struck atop Huaca del Sol
resonated as a grand proclamation rang out to drop
your pots and change your spouts? Whatever form
it took, some sort of proclamation seemed essential.
Without an authoritative declaration, how could
crisp phase transition transpire rapidly and uniformly in any style sequence? If mandates for punctuated change were absent, then style transformations would lapse into the sphere of artistic drift and
temporal ambiguity. If authorities at Sol and Luna
ordered style change, then did the capitals artisans
quickly fashion new spouts? Did they produce new
molds and create appropriate new painting rules?
Once formulated at the origin center, was the newphase production package then shipped to all Moche

18

VISIONS OF TIWANAKU

provinces, thereby facilitating concurrent transformation throughout the realm? This scenario seemed
politically and technologically plausible. Yet ensuring
hinterland phase-change compliance would have
been a conundrum for centralized states and even
more problematic for confederated polities. I wondered if style police were sent forth to inspect and
standardize rural workshop production by Chavn,
Moche, Huari, Tiwanaku, and other centers. Did style
enforcers also bust old folks who clung to bygone
motifs, thereby ensuring concurrent phase changes
in seriations? If my former musings now seem amusing, they simply reflect the fact that the profession
has been seriating styles for a century without explaining how rapid phase changes in Andean art
transpired.
If how the style changed is perplexing, why is
even more so. The Spanish and Inca conquests provide one model of why artistic transformations transpired. Uhle and his followers applied it to what they
considered coastal Tiwanaku. The subjugation model
defines the Late and Middle horizons in the Ica sequence. In the absence of conquest, the Early Horizon
is defined by more shadowy foreign influence from
a distant center some 600 km away. Yet all seriations
have numerous phase transformations that are not
readily laid at the feet of foreign interlopers.
What, then, were the cultural conditions that occasioned the sharp transitions in local style that
make successive phases of chronological utility?
Working at Sol and Luna, I was much vexed by this
question. If foreign influence was not particularly evident in the Larco sequence and if artistic drift over
time did not result in punctuated change, then what
triggered style transformations? Certainly it was not
nature! Kroeber defined culture as a super-organic
phenomenon, and archaeologists continue to conceptualize Andean societies as floating above and beyond their disaster-prone habitats. Therefore, all triggers for artistic change had to be cultural. For Moche,
transformations in religious ideology were possible
activators, and the thematic content of iconography
did differ between phases. Yet the gods made them
do it seemed a tenuous explanation for why artistic
fashion changed. My guess was that there was a correlation with political transformations, such as the
ascendancy of new governing dynasties. This could
not be demonstrated for Moche, but at origin centers
such as Chavn, Huari, and Tiwanaku, the dynamics

of governance seemed the best explanatory stimulus

for the dynamics of style. Nonetheless, ranging from
the Ica and Nazca valleys to many other study areas,
there are seriated chronologies for societies that lack
evidence of primary centers, focal governance, and
centralized production of arts. Why should crisp
phase transitions be expected in these contexts, or
indeed elsewhere? Artistic expression certainly
changes over time, and seriation can gauge this. Yet
its abject century-long failure to identify the how
and why of aesthetic transformations in the Andes is
cause for reflection. The discipline seems to have
embraced a methodology that works with no attention to why and how it might work, or might be
a house of cards.

THE AXES OF STYLE VARIATION

Andean settlement patterns are characterized by relatively frequent shifts in residential foci. Deep stratified deposits that superimpose millennia of continuous occupation are rare, as are site formations akin
to Near Eastern tells with very long records of settlement. Consequently, Andean archaeological
chronologies are heavily dependent upon seriated sequences of ancient remains. Seriation is remarkably
versatile. It will distinguish between three randomly
drawn handfuls of red, white, and blue marbles extracted from a pot full of 3,999 marbles, with each
color equally represented. Three grab samples have
a very low probability of being the same or reflecting
equal numerical representations of the marble population they were drawn from. In archaeology, the luck
of the draw presumes diachronic distinctions. Seriation is a self-contained methodology and a flexible
one that can order unprovenienced artifacts, grave
lots, excavated assemblages, or surface collections. In
the Andes, techniques of seriation have been generally similar for the last century, particularly in the
Berkeley School. A whole community or cultural
style is presumed when dealing with fine arts and
grave goods such as those of Ica, Moche, Nazca,
Huari, and Tiwanaku. Other than distinctions attributable to function or to social status, all seriations assume that notable differences in aesthetic expressions
are chronologically significant. All presume that variation in style is the exclusive product of change over
time. All presuppose rapid phase transformations in
order to have chronological utility.

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

Seriation has formed the backbone of Andean inquiry since the days of Uhle. Therefore, to assert that
the methodology can and has produced flawed results
might seem heretical if it were not for a very simple
truth. This reality is that the method can deliver valid
chronological results only when style is shared homogeneously among all contemporary members of a society. Artistic unity and presumptions of homogeneous style are critical to the starting premise of
seriation, which holds that variation in style is the exclusive product of change over time. Synchronic artistic heterogeneity within a style is anathema to this
critical premise. Yet diversity in fashion can transpire
if members of a culture or a society use distinctions in
aesthetic expression to identify social distinctions
among themselves. This is not uncommon. Today
there are clothing distinctions between and within
Aymara- and Quechua-speaking populations as well
as between rural and urban populations. Different
fashion dialects tend to have different spatial distributions. However, they can co-occur in profusion at cosmopolitan centers in the highlands and altiplano. At
Cuzco, Inti Raymi reenactment is marked by a colorful influx of many different indigenous groups distinguished by different attire.
Methodologically, synchronic aesthetic heterogeneity is very important. It constitutes a second,
nontemporal axis of graphic variation: stylistic differences paralleling cultural ones among contemporary
people who share similar art. This is diagrammed in
Figure 2.1 as two perpendicular axes. In the figure,
variants of a single style are represented by different
patterns that could align along either axis. In linguistics, patterned variations in shared speech constitute
dialects. They are critical elements of linguistic analysis, and analogous variation is expectable in other aspects of cultural behavior. The history of European
painting is replete with synchronic schools of art.
Their supposed absence in the New World cannot be
assumed on a priori grounds. The human genome
has not identified indigenous people as possessing a
gene that predisposed adherence to whole cultural
styles. As with linguistic dialects, learning accounts
for most of the variation in aesthetic expression that
marks social distinctions in the Andes. Yet aesthetic
taste could be physically imposed by one generation
on the next in the case of cranial deformation. Different deformation patterns characterized different
cemeteries surrounding the Tiwanaku Omo M10

19

complex in Moquegua (Blom et al. 1998; Goldstein

2004; Moseley et al. 1991). That these were kin-based
cemeteries of socially distinct M10 residential groups
is a reasonable supposition.
Historically, the concept of whole cultural styles
was implanted at a time when archaeologists viewed
cultures as relatively homogeneous wholes. Excavating at Moche decades ago, I presumed a high degree
of cultural unity throughout the geographic realm of
Moche arts. Seriation was a logical fit with this perspective. However, the old perspective has constructively shifted toward a more nuanced view that involves ancestor veneration, descent groups, and
kinship organization as the cornerstones of indigenous cultures. There are differing views about such
organization (e.g., Albarracn-Jordn 1996; Isbell
1997; Moseley 2001). Nonetheless, there is now more
attention to differences in the archaeological record
than there was twenty years ago. Consequently, synchronic variation in style is currently more expectable, and it can be understood as an indigenous
reality used to identify and distinguish diverse
groups of people within a broader populace sharing
commonalities. Schools of visual communication
and dialects of verbal communication enunciate social distinctiveness and social differentiation.
Distinct dialects or schools of fashion are a fact of
Andean life, but they have gone unacknowledged in
seriation. The reason is that in and of itself, seriation
is incapable of distinguishing between the social and

FIGURE 2.1. The two axes of stylistic variation portrayed

as perpendicular to each other, with different variants of
the same style represented by different boxed symbols.

20

VISIONS OF TIWANAKU

temporal axes of style variation. It scrutinizes differences in style, but it has no internal means of segregating ethnic or social distinctions from temporal or
chronological ones. Powerless to differentiate heterogeneous from homogeneous contexts, it proceeds
from the premise that variance in style is the exclusive
product of change over time. Methodologically, this
is its Achilles heel. In a context of heterogeneity, the
methodology orders two very different phenomena
along the single axis of temporal change. Consequently, the chronological efficacy of seriated sequences varies relative to the amount of noise from
concurrent social variation in style that is phased
into them.
How much noise might the social axis of aesthetic distinction inject into Andean chronologies? If
the Inca Empire was divided into about 80 provinces,
each with multiple linguistically, socially, ethnically,
and genealogically distinct populations, then heterogeneity was pronounced (Morris 1995). If political
formations were characterized by recursive, nested
hierarchies of kin-based descent groups, then heterarchy was pronounced. If indigenous distinctions in
style of cranial formation, costume, paraphernalia,
accoutrements, utensils, housing, and other aspects of
material culture defined distinctions in cultural, political, social, and genealogical identity, then variation
in fashion was rife among contemporary people in
the ancient Andes. Consequently, seriation has been
applied to a heterogeneous (not homogeneous) cultural landscape. Because seriation is a self-contained
but not self-correcting methodology, the chronological ramifications may prove far-reaching.
The Social Axis
Identifying the social axis of style variation requires
dating constraints that are independent of seriation.
They can include stratified assemblages, but most
commonly they come from radiocarbon (14C) assays.
Synchronic social noise in seriated sequences likely
varies. When it is present, identification generally requires a minimum of 10 assays for each phase or unit
of a seriation. This requirement is rarely met. The reasons are, in part, historical. Incongruities between seriated sequences and associated 14C assays have been
recognized for a long time, even though the samples of
dates have been relatively small. An early precedent
was set for dealing with inconsistencies. The Berkeley

School held seriations to be valid and inconsistent carbon dates to be in error (Rowe 1967). Unfortunately,
this position transformed seriation into an end in and
of itself rather than an independently verifiable means
to an end. It reaffirmed the chronological efficacy of
seriation and the unstated presumptions of stylistic
homogeneity that continue to structure archaeological interpretation. Furthermore, it disparaged the fundamental need for large-scale independent dating of
seriated sequences. Consequently, 14C controls have
emerged only in areas where there is a great deal of
ongoing excavation and fieldwork, and I will touch
briefly on several examples.
Independent dating has not been kind to the
most venerable of Andean seriations: Larcos (1948)
five-phase Moche sequence. Within the last decade,
the sequence has largely collapsed, with current investigators using a loose terminology of early (I and
II spouts), middle (III and IV spouts), and late (V
spouts) Moche. The current terminology is still beset
by overlapping 14C dates and, in some cases, elite
grave lots containing early and middle stylistic accompaniments. The five different spout forms certainly identify different schools of fashion, but their
social correlates remain elusive. Some have varying
degrees of geographical focus and may call out social
distinctions among groups of people with ties to different areas. Yet who these self-segregating groups
were, other than ruling elites, is not understood.
Therefore, Moche sociopolitical organization is ill defined. The consequences of a seriated chronology
gone awry can have discipline-wide ramifications. In
the early 1990s, consensus held that Moche was the
first archaic state to arise in the Andes. Now Moche
seems to have been a federated elite ideology without
marked political centrality.
A much quicker metamorphosis in the axes of
stylistic variation transpired in the arid sierra of the
Moquegua Valley of southern Peru. Research that
began in the early 1980s identified three stylistically
distinct expressions of Tiwanaku, each with spatially
distinct residence patterns. Methodologically, the first
impulse was the traditional oneto seriate the occupations: an altiplano folk colonization, an imperial intrusion from Tiwanaku, and then a collapse-driven
reversion to fragmented local expressions of the former order (Goldstein 1989). A significant cadre of
14C assays has negated this interpretation. Populations pursuing different schools of fashion, different

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

forms of cranial deformation, and different patterns
of settlement and residential organization overlap in
time (Goldstein this volume; Owen and Goldstein
2001). They also co-occur to varying degrees with the
Huari imperial colonization of Cerro Bal in the Moquegua headwaters (Williams this volume). In overview, the importance of this transformation from the
temporal to the social axis of variation is that it brings
archaeological interpretation into the theater of Andean actualities and what Murra (1972) identified as
multiethnic ecological zones.
I conclude these considerations with a case that
is still evolving. Rowes (1962b) seriation of the art
and architecture of Chavn de Huntar confronts
challenges from ongoing field studies at the center
by Stanford University investigators. Their work
demonstrates that the old and new temples were not
built sequentially. Rather, both exhibit numerous
construction stages with interspersed episodes of
use and building activities that were often concurrent in both areas of the monument before it assumed final form (Kembel 2001). Thus the seriated
architectural sequence of an old edifice followed by
a new one is not substantiated. Because the two were
ornamented with stone artwork assigned to sequential phases, the temporal validity of the seriation is
now open to scrutiny. Assessment of the situation
awaits an emerging corpus of 14C dates.
Chavn de Huntar may or may not prove to be
a harbinger of things to come. Nonetheless, it raises
issues worthy of consideration. The origin center
horizon of civilization concept presumes stylistic homogeneity at Andean nexuses of power. Yet if power
was heterogeneous or segmented, and centers were
cosmopolitan, then expectations of standardized
fashion may be unrealistic.
Sociopolitical differentiation is physically stamped
on ancient capitals, such as Cuzco, where high-walled
residential compounds forcefully seclude and exclude
different cadres of ruling elites. At Chan Chan each
major compound was distinguished from all others by
the specific motifs and compositions of its wall friezes
(Pillsbury 1993, 1996, 2009). Although built sequentially, multiple compounds were in use at the same
time. Therefore distinctions in compound ornamentation called out social distinctions at the apex of rule.
Differences in the ornamentation of palatial quarters
suggest that elite residents also differentiated themselves with elements of costume and material goods. If

21

distinctions in fashion defined who was who, then

they would be very important at the top of the sociopolitical hierarchy. If particular areas of foreign conquests or rural estates supported particular compounds at Cuzco (Zuidema 1986) or Chan Chan, then
might not the specific stylistic stamps of the proprietors be expected on their hinterland holdings? If
origin centers were heterarchical, why should archaeologists expect cosmopolitan cores to export homogeneous horizon styles? Horizon styles have been
presumed to be homogeneous styles. Yet they are most
likely abstractions similar to whole culture styles.
Interaxial Variation
I have portrayed the temporal and social axes of aesthetic variation as perpendicular to one another (Figure 2.1). In doing this and in critiquing seriation, I do
not wish to imply that art fails to change over time. It
certainly does. Temporal dynamics may have been
the dominant trajectory in many cases. But in others,
social dynamics were in play. This is expectable in
complex cultures embracing assorted populations.
Contemporary populations employ stylistic distinctions for many conscious and unconscious reasons.
Identifying and analyzing these peculiarities in expression is no less important in studying art than in
studying language.
Stylistic distinctions were played out on an interaxial grid framed by temporal and social trajectories
of variation. This is illustrated schematically in Figure
2.2, with different variants of a single style represented
by different patterns. The figure attempts to show that
each variant has its own cultural and chronological
history. Some stylistic dialects may have been old, conservative, and enduring. Others may have been innovative and expansive but short-lived. Intermediate
mixes can be expected. The challenge is to identify
where each dialect falls on the interaxial grid and
then explain its cultural and chronological history.
Seriation has pioneered inquiry into the identification of stylistic distinctions. The next step is to independently contextualize their temporal and social parameters. This is entirely tractable with good data
sets. For example, it is now possible to distinguish
substyles within the arts that Larco (1948) called
Moche phase V. Furthermore, within the San Jose de
Moro substyle, it is feasible to recognize and define the
works of individual artisans (McClelland et al. 2007).

VISIONS OF TIWANAKU

22

FIGURE 2.2. Interaxial stylistic variation portrayed in a

social temporal matrix, with different variants of the
same style represented by different boxed symbols.

CONCLUSIONS
Tiwanaku is a remarkable monument, commanding
admiration as an ancient central place for centuries.
For archaeologists, it became the primordial keystone for the origin center horizon of civilization
concept formulated by Max Uhle. This formulation
was very original for its time because it was constructively based on a New World analogue, the Inca,
rather than on an Old World template. Historically,
Uhles interpretive model of Tiwanaku was vital to
foreign scholars of the twentieth century, who were
concerned with the daunting task of sorting out ancient cultural successions in the Andes. Although
variation in fashion was common in Uhles Inca analogue, artistic differences represented background
noise to be ignored in the quest for horizon styles
that could temporally align prehistory in different
regions of the Cordillera.
Originating with Tiwanaku, the horizon style
concept reinforced an enduring reliance upon seriation as a method for identifying change and ordering chronologies. Seriation has been applied to many
types of remains. However, its dominant and most
detailed focus in Peruvian studies has been on mortuary accompaniments, above all those excavated by
Uhle and deposited at the University of California. In
the quest for relative dating, the Berkeley School em-

phasized ceramic grave goods as being particularly

sensitive to temporal transformation. Consequently,
ensuing interpretations of culture history have been
time sensitive but relatively myopic due to exclusion
of the far broader potential of the archaeological
record for addressing issues ranging from everyday
life through economic adaptations.
Beginning with Uhles pioneering orderings of
mortuary remains, seriation was refined over time.
Kroeber and his students used typology to cluster
ceramic grave accompaniments before ordering
them chronologically. Rowe turned back to individual grave lots as the primary units for analyzing associations. Rejecting typology, he emphasized qualitative analysis of stylistic expression as the basis for
developing very nuanced seriation. His methodology was explicated in a masterful series of very insightful articles of contemporary relevance. Drawing upon Uhles collections from the Ica Valley, Rowes
approach was manifested in a very long, very detailed seriation of local ceramics. Indeed, some
phases reputedly spanned no more than a generation or two, and this remains by far the most comprehensive seriation in the Cordillera.
Rowe considered his rigorous analysis of style to
be similar to the analysis of language. Confident in
this vision, he proposed two things. The first was
that the Ica relative chronology served as the master
sequence for a system of calendar-defined horizons
and intermediate periods applicable to all of the
central Andes. This scheme has structured archaeological interpretation by foreigners for decades.
Rowes second proposal held that if relative dating
by Berkeleys method of seriation disagreed with absolute dating by 14C, then the latter was in error.
This pronouncement transformed seriation into an
end in itself, immune to assessment by other means!
Consequently, there are few 14C constraints on the
master sequence. With scant independent verification, the disciplines adherence to the sequences
overarching scheme of calendrical horizons and periods is largely faith based.
There is a long-standing faith in seriation because it delivers relative chronologies through attention to variation in material remains. This is consistent with the fact that arts and artifacts are not static
but change, as the discipline of art history attests. As
practiced in the Andes, seriation rests upon two keystone premises. The first is that styles are broadly

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

homogeneous synchronically. This, in turn, sustains
the second premise, which holds that stylistic variation is the exclusive product of change diachronically. Without these keystones, the methodology
loses viability. Maintaining the first premise may explain why practitioners of seriation tend to treat ancient arts as uniform whole cultural styles. Preserving the second premise may explain why the
discipline has not recognized that stylistic variation
occurs along two axes. One is temporal and diachronic, as the methodology requires. The other axis,
however, is social and synchronic. Seriation distinguishes variation, yet in and of itself, it cannot distinguish if the difference falls on the synchronic or
diachronic axis or both.
The social axis of variation is important in both
visual and oral communication. Dialects of speech
and dialects of style are nodal road maps that set the
stage for appropriate interaction among populations
that are different yet also similar in terms of their
shared behavior. Social variation in fashion has long
been indigenous to complex societies of the Andes.
This condition explains why some seriations have
gone awry, and others may follow when they are assessed against adequate suites of radiometric constraints. However, the significance of synchronic
variation has less to do with chronology than with
culture. Scrutiny of the social axis will tell the discipline who was who, and transform archaeology into
anthropology. Recognizing these neglected potentials
of the prehistoric record now resides with scholars
working in Bolivia and Tiwanaku. This is because
their Peruvian counterparts are wedded to the faithbased interpretive structure of horizons and periods
that Tiwanaku inspired more than a century ago.

ACKNOWLEDGMENTS
Assuming responsibility for all errors, I am pleased
to acknowledge highly constructive commentary
from Juan Albarracn-Jordn, William Conklin,
Susan deFrance, Tom Dillehay, Christopher Donnan,
Carol Mackey, Joyce Marcus, Donna Nash, Joanne
Pillsbury, Charles Stanish, and Alexei Vranich.

REFERENCES CITED
Albarracn-Jordn, Juan
1996 Tiwanaku. Arqueologa regional y dinmica segmentaria. Editores Plural, La Paz.

23

1999 The Archaeology of Tiwanaku: The Myths, History and Science of an Ancient Andean Civilization. Publicidad Arte Producciones, La Paz.
2014 Una reflexin histrica sobre Tiwanaku y el marco
institucional indgena en Bolivia. Sus origenes civilizatorios y su posterior desarrollo. Editores Plural, La Paz, in press.
Blom, D. E., B. Hallgrmsson, L. Keng, M. C. Lozada C.,
and J. E. Buikstra
1998 Tiwanaku Colonization: Bioarchaeological Implications for Migration in the Moquegua Valley,
Peru. World Archaeology 30(2):238261.
Burger, Richard (editor)
2014 The Writings of Julio C. Tello. University of Iowa
Press, Iowa City, in press.
Cieza de Len, Pedro de
1976 [15531554] The Incas of Pedro Cieza de Len.
University of Oklahoma Press, Norman.
Donnan, Christopher B.
1965 Moche Ceramic Technology. awpa Pacha 3:
115138.
Ford, James A., and Gordon R. Willey
1949 Surface Survey of the Viru Valley, Peru. Anthropological Papers of the American Museum of
Natural History Vol. 43, Pt. 1. American Museum of Natural History, New York.
Goldstein, Paul
1989 The Tiwanaku Occupation of Moquegua. In
Ecology, Settlement and History in the Osmore
Drainage, Peru, edited by Don S. Rice, Charles
Stanish, and Phillip R. Scarr, pp. 219255. BAR
International Series Vol. 545, No. 2. British Archaeological Reports, Oxford.
2004 Andean Diaspora: The Tiwanaku Colonies and
the Origins of South American Empire. University Press of Florida, Gainesville.
Isbell, William
1997 Mummies and Mortuary Monuments: A Postprocessual Prehistory of Central Andean Social
Organization. University of Texas Press, Austin.
Kaulicke, Peter
1998 Max Uhle y El Per Aantiguo. Pontifica Universidad Catlica del Per, Lima.
Kembel, Silvia Rodriguez
2001 Architectural Sequence and Chronology at
Chavn de Huntar, Peru. Unpublished Ph.D.
dissertation, Department of Anthropology,
Stanford University, Stanford.
Kroeber, A. L.
1925 The Uhle Pottery Collections from Moche. University of California Publications in American Archaeology and Ethnology 21(6):191234. Berkeley.
1944 Peruvian Archaeology in 1942. Viking Fund Publications in Anthropology No. 4. Viking Fund,
New York.

24

VISIONS OF TIWANAKU

1948 Anthropology. Harcourt, Brace and Company,

New York.
1957 Style and Civilization. Cornell University, Ithaca.
Larco Hoyle, Rafael
19381939 Los Mochicas. Casa Editora La Cronica y
Variedades, Lima.
1948 Cronologia arqueolgica del norte del Per. Sociedad Geogrfica Americana, Buenos Aires.
Lumbreras, L. G.
1974 The Peoples and Culture of Ancient Peru. Smithsonian Institution Press, Washington, D.C.
McClelland, Donna, D. McClelland, and C. B. Donnan
2007 Moche Fineline Painting from San Jose de Moro.
Cotsen Institute of Archaeology Press, Los Angeles.
Menzel, Dorothy
1977 The Archaeology of Ancient Peru and the Work of
Max Uhle. R. H. Lowie Museum, Berkeley.
2006 John Rowes Archaeology. awpa Pacha 28:229
231.
Menzel, Dorothy, John H. Rowe, and Lawrence E.
Dawson
1964 The Paracas Pottery of Ica: A Study in Style and
Time. University of California Publications in
American Archaeology and Ethnology Vol. 50.
University of California, Berkeley.
Morris, Craig
1995 Symbols to Power: Styles and Media in the Inka
State. In Style, Society, and Person: Archaeological
and Ethnological Perspectives, edited by C. Carr
and J. Neitzel, pp. 419433. Plenum Press, New
York.
Moseley, Michael E.
2001 The Incas and Their Ancestors: The Archaeology of
Peru. Revised ed. Thames and Hudson, London.
Moseley, M., R. Feldman, P. S. Goldstein, and L.
Watanabe
1991 Colonies and Conquest: Tiahuanaco and Huari in
Moquegua. In Huari Administrative Structure: Prehistoric Monumental Architecture and State Government, edited by W. H. Isbell and G. F. McEwan,
pp. 2140. Dumbarton Oaks, Washington, D.C.
Murra, John V.
1972 El control vertical de un mximo de pisos ecolgicos en la economia de las sociedades andinas. In Visita de la provincia de Len de Hunuco
en 1562 por Inigo Ortiz de Zuiga. Documentos
para la historia y etnologa de Huanuco y la selva
central, Vol. 2, edited by John V. Murra, pp. 427
476. Universidad Nacional Hermilio Valdizn,
Hunuco, Peru.
Owen, Bruce, and Paul Goldstein
2001 Tiwanaku en Moquegua: Interacciones regionales y colapso. In Huari y Tiwanaku: Modelos vs.
evidencias, Pt. 2, edited by Peter Kaulicke and

William H. Isbell, pp. 169188. Boletn de Arqueologa No. 5. Pontificia Universidad Catlica
del Per, Lima.
Pillsbury, Joanne
1993 Sculpted Friezes of the Empire of Chimor. Unpublished Ph.D. dissertation, Columbia University, New York.
1996 The Thorny Oyster and the Origins of Empire:
Implications of Recently Uncovered Spondylus
Imagery from Chan Chan, Peru. Latin American
Antiquity 7:313340.
2009 Reading Art Without Writing: Interpreting
Chim Architectural Sculpture. In Dialogues in
Art History, edited by Elizabeth Cropper. Center
for Advanced Study in the Visual Arts, National
Gallery of Art, Washington, D.C.
Rowe, John H.
1945 Absolute Chronology in the Andean area. American Antiquity 10(3):265284.
1954 Max Uhle, 18561944: A Memoir of the Father of
Peruvian Archaeology. University of California
Publications in American Archaeology and Ethnology Vol. 46, No. 1. University of California,
Berkeley.
1960 Cultural Unity and Diversification in Peruvian
Archaeology. In Men and Cultures, edited by
A. F. C. Wallace, pp. 627631. University of
Pennsylvania Press, Philadelphia.
1961 Stratigraphy and Seriation. American Antiquity
26(3):324330.
1962a Alfred Louis Kroeber (18761969). American
Antiquity 27(3):396415.
1962b Chavin Art: An Inquiry into Its Form and Meaning. Museum of Primitive Art, New York.
1962c Stage and Periods in Archaeological Interpretation. Southwestern Journal of Anthropology 18
(1):4054.
1967 An Interpretation of Radiocarbon Measurements on Archaeological Samples from Peru.
Reprinted in Peruvian Archaeology: Selected
Readings, edited by John H. Rowe and Dorothy
Menzel, pp. 1630. Peek Publications, Palo Alto.
Originally published 1966, Proceedings of the
Sixth International Conference, Radiocarbon and
Tritium Dating. U.S. Atomic Energy Commission, Springfield, Virginia.
Schreiber, Katharina
2006 In Memoriam: John Howland Rowe, 1918
2004. awpa Pacha 28:195202.
Stanish, Charles
2002 Tiwanaku Political Economy. In Andean Archaeology. Vol. 1, Variations in Sociopolitical Organization, edited by William H. Isbell and Helaine
Silverman, pp. 169198. Kluwer Academic/
Plenum Publishers, New York.

CHAPTER 2: STYLISTIC VARIATION AND SERIATION

Stbel, M., and M. Uhle
1892 Die Ruinensttte von Tiahuanaco im Hochlande
des Alten Peru: Eine Kulturqueschichtliche Studie.
Verland von Karl W. Hiersemann, Leipzig.
Tello, J.
1960 Chavn: Cultura matriz de las civilizacines andina. Universidad Nacional San Marcos, Lima.
Uhle, Max
1903 Pachacamac, Report of the William Pepper M.D.,
L.L.D. Peruvian Expedition of 1896. University
of Pennsylvania, Philadelphia.
1913 Ruinen von Moche. Journal de la Socit des
Amricanistes de Paris 10:95117.
Willey, Gordon R.,
1945 Horizon Styles and Pottery Traditions in Peruvian Archaeology. American Antiquity 11:49
56.
1948 Functional Analysis of Horizon Styles in Peruvian Archaeology. In A Reappraisal of Peruvian
Archaeology, Memoir 4, edited by W. C. Bennett,
pp. 815. Society for American Archaeology,
Menasha, Wisconsin.

25

1951 The Chavin Problem: A Review and Critique.

Southwestern Journal of Anthropology 7(2):103
144.
Willey, Gordon R., and Philip Phillips
1958 Method and Theory in American Archaeology.
University of Chicago Press, Chicago.
Willey, Gordon R., John H. Rowe, Charles C. DiPeso,
William A. Ritchie, Irving Rouse, and Donald W. Lathrap
1956 An Archaeological Classification of Culture
Contact Situations. In Seminars in Archaeology:
1955, No. 11, edited by R. Wauchope, pp. 130.
Memoirs of the Society for American Archaeology, Salt Lake City.
Zuidema, R. T.
1986 La Civilisation Inca au Cuzco. Presses Universitaires de France, College de France, Paris.