AN ARCHAEO-HISTORY OF ANDEAN COMMUNITY AND LANDSCAPE:

THE LATE PREHISPANIC AND EARLY COLONIAL COLCA VALLEY, PERU

by

Steven Arlyn Wernke

A dissertation submitted in partial fulfillment of

the requirements for the degree of

Doctor of Philosophy

(Anthropology)

at the

UNIVERSITY OF WISCONSIN-MADISON

2003

i

To my parents, Arlyn and Grace Wernke,

for their unwavering support

ii

CONTENTS List of Figures v

List of Tables xii

Acknowledgements xiii

Abstract xviii

PART I: INTRODUCTION AND BACKGROUND

Chapter 1: Community and Landscape in the Andes 1 Introduction 1 A Political-Ecological Approach to Community and Landscape 6 Alternative Hypotheses and Their Correlates 18 Dissertation Outline 32

Chapter 2: Physiographic, Climatic, and Anthropogenic Characteristics of the

Colca Valley Landscape 34 Location, Geology, and Geomorphology 34 Soils 42 Climate Characteristics and Trends 44 Anthropogenic Landscape Features 54

Chapter 3: The Collagua in Regional and Historical Context: Previous Research 67

Introduction 67 Previous Archaeological Research 67 The Collagua in Colonial Documents 80

Chapter 4: Methodology and Research Operations 92

Introduction: An Archaeo-Historical Approach 92 Overview of Archaeological Methodology 96 Selection of Project Area 98 Archaeological Surveying Techniques in a Montaine Landscape 99 Site Nomenclature 102 Field Mapping Techniques 102 Site Sizes 103 Site Classification: Chronology 105 Artifact Collection Strategy 106 Survey Area Coverage and Total Sites Recorded 107 Basemap Source Data 109 Early Colonial Land Tenure Analysis: Overview of the Documentary Sample 111

iii

The Visitas as Text: Issues of Representation 115 GIS Cartographic Representation and Database Management of

Documentary Data 117

PART II: ARCHAEOLOGICAL FINDINGS

Chapter 5: Early Settlement and Agriculture: The Formative and Middle Horizon 126

Introduction 126 Spatial and Chronological Considerations 128 The Formative Period Settlement Pattern: Overview 129 Site YA-032: Chiquero 135 Agricultural Production and Landscape Modification During the Formative 140 Terminal Archaic and Formative Period Hunting and Pastoralism in the Puna 143 Discussion 146 The Middle Horizon Period Settlement Pattern: Overview 150 Middle Horizon Cemeteries 157 Agro-Mortuary Wall Sites and Associated Settlements 158 Discussion 167 Chapter Summary and Conclusions 169

Chapter 6: The Late Intermediate Period and Late Horizon Occupations 171

Introduction 171 Chronological Considerations 172 The Late Intermediate Period Settlement Pattern: Overview 176 The Late Horizon Settlement Pattern: Overview 182 Settlement Sizes 189 Collagua Domestic Architecture 195 Inka Imperial Architecture 217 Cemeteries, Tombs, and Mortuary Architecture 225 Agro-Pastoral Infrastructure and Production during the LIP and Late Horizon 234 Pukara Fortifications and Evidence for Violent Conflict 251 Descriptions of Principal LIP/Late Horizon Settlements 266

Chapter 7: Spanish Insertions: Early Colonial Settlement and Agro-Pastoral

Production 298 Introduction 298 Spatial and Chronological Considerations 299 The Early Colonial Period Settlement Pattern 302 The Doctrinas of the Colca Valley, 1540-1595 305 Archaeological Evidence for the Doctrinas: Early Colonial Chapels 312 Toledan Dislocations 330 Chapter Summary 343

iv

PART III: ETHNOHISTORICAL RECONSTRUCTION AND ARCHAEO-HISTORICAL SYNTHESIS

Chapter 8: Regional Political-Ecological Syncretism under Inka and

Spanish Rule 344 Introduction 344 Hybrid Communities: Reduccin, Ayllus and Kurakas 345 Structural Grafting: Ayllus and Reducciones 359 A Regional View of Collagua Complemenarity Practices 369 Kurakas, Craft Specialists, and the Provincial Political Economy 385 Chapter Summary and Conclusions 390

Chapter 9: Reconstructing Colonial Land Tenure and Late Prehispanic

Residence Patterns 393 Introduction 393 The Local Crop Mosaic 394 Moiety-Level Land Tenure Patterning 398 Ayllu-Level Land Tenure Patterning 400 Discussion and Conclusion 430

Chapter 10: Summary and Conclusions 435

Review of the Research Problem 435 Early Agriculturalist Settlement in the Central Colca Valley 436 Community Organization and Production during the LIP and Late Horizon 439 Collagua Communities and Complementarities during Inkaic and Early

Colonial Times 443

APPENDICES

Appendix A: Ceramic Sequence 445

Appendix B: Projectile Point Illustrations 538

Appendix C: Archaic Period Sites 541

Appendix D: Survey Artifact Registry 543

Appendix E: Ceramic Distributions by Site 561

Appendix F: Site and Sector Registry 585

v

References Cited 603

vi

List of Figures

Figure 2.1: Location of the Colca valley 37 Figure 2.2: The Colca valley and survey area 38 Figure 2.3: Volcanic features and obsidian sources 39 Figure 2.4: Geomorphic surfaces in the central Colca Valley 41 Figure 2.5: Average monthly temperatures in the Colca valley 46 Figure 2.6: Average annual precipitation, 1955-1982 47 Figure 2.7: Average monthly precipitation, 1951-1982 49 Figure 2.8: Monthly precipitation, Chivay, 1973-1972 49 Figure 2.9: Perspective view of Pampa Finaya 61 Figure 2.10: Chijra/San Antonio/Chilacota area 63 Figure 3.1: Overview of the Province of Collaguas, showing internal divisions 81 Figure 4.1: Survey area and all sites registered (N=169) 108 Figure 4.2: Map of all toponyms with matching names in the visitas 122 Figure 5.1: Map of Formative Period sites in the survey area 131 Figure 5.2: Formative site counts by ecological zone (N=30) 133 Figure 5.3: Formative site counts by altitude (N=30) 133 Figure 5.4: Formative ceramics recovered from YA-032 137 Figure 5.5: Obsidian Type 5D projectile point and scraper from YA-032 138 Figure 5.6: Andesite hoes from YA-032 138 Figure 5.7: YA-032 (Chiquero) and surrounding agricultural landscape features 141 Figure 5.8: sketch map of YA-094 144 Figure 5.9: Site CO-106, showing cliff overhang (rockshelter) and midden areas 145 Figure 5.10: Middle Horizon ceramics 151 Figure 5.11: Painted tablets from YA-169 (Bomboncilla) 152 Figure 5.12: Middle Horizon sites in the survey area 153 Figure 5.13: Middle Horizon site counts by ecological zone (N=37) 155 Figure 5.14: Formative site counts by altitude (N=37) 155 Figure 5.15: Agro-mortuary walls at site YA-014, Sector E 160 Figure 5.16: Detail of agro-mortuary wall 161 Figure 5.17: Ovoid features in agro-mortuary walls at YA-014 164 Figure 5.18: Gallery tomb at CO-148, Sector A 166 Figure 6.1: Collagua ceramic sequence (lot numbers indicated) 175 Figure 6.2: Late Intermediate Period settlement pattern map 181 Figure 6.3: Late Horizon settlement pattern map 186 Figure 6.4: Detail of Late Horizon settlements in the central area of the survey zone 187 Figure 6.5: Late Horizon site size histogram, by area (ha) 190 Figure 6.6: Late Horizon site size histogram, by house count 191 Figure 6.7: Late Intermediate Period rank size graph, by site area 193 Figure 6.8: Late Intermediate Period rank size graph, by house count 194 Figure 6.9: Late Horizon rank size graph, by site area 194 Figure 6.10: Circular houses at the site of Laiqa Laiqa, near Tuti 200 Figure 6.11: Small house of Type 1 masonry with preserved doorway 202

vii

Figure 6.12: Small house of Type 1 masonry with preserved doorway 202 Figure 6.13: Detail of doorway 202 Figure 6.14: Detail of doorway/facade 203 Figure 6.15: Interior northwest corner of Structure 154 at YA-050 203 Figure 6.16: Detail of central niche on north wall, YA-050, Structure 154 203 Figure 6.17: Detail of plaster, YA-050, Structure 154 203 Figure 6.18: Small house of Type 1 masonry, from the east. CO-164, Structure 28 207 Figure 6.19: House of Type 2 masonry, from the southeast. YA-050, Structure 107 207 Figure 6.20: Southwest corner of house of Type 3 masonry. YA-050, Structure 114 208 Figure 6.21: Facade of large house of Type 4 masonry, from the east.

YA-050, Structure 104. 208 Figure 6.22: Detail of split-river boulders of Type 4 masonry. 208 Figure 6.23: Large house of Type 5 masonry, from southwest. 209 Figure 6.24: Structure 17 (CO-100), from the northeast. Note tenon supports

for second floor 209 Figure 6.25: Large house of Type 5 masonry, from the south. 209 Figure 6.26: Wall fragment of Type 6 masonry, from the east. 210 Figure 6.27: Southwest corner of Structure 31 210 Figure 6.28: Southeast corner of Structure 31 211 Figure 6.29: House footprint area histogram 213 Figure 6.30: Boxplot of house area, grouped by masonry type 214 Figure 6.31: Boxplot of house area, grouped by site 215 Figure 6.32: Kallanka structure at YA-050, showing orientations of photos 220 Figure 6.33: Detail of facade fragment of kallanka (Structure 56) at CO-163 221 Figure 6.34: Corner remnant with Cuzco Inka style cutstone blocks in Yanque

(YA-041), facing east 222 Figure 6.35: Overview of the Casa Choquehuanca area, Yanque (YA-041) 224 Figure 6.36: Colonial house in Yanque (Casa Choquehuanca) 225 Figure 6.37: Large, three-storey chullpa with two doors (right side),

CO-098 (Fatinga) 230 Figure 6.38: Two-storey chullpa at CO-098 231 Figure 6.39: Adjoining two-storey chullpas at CO-098 231 Figure 6.40: Remnants of red pigment, chullpa at CO-098 232 Figure 6.41: Two-storey chullpa at CO-098 232 Figure 6.42: Small chullpa under rock overhang. Site CO-117 233 Figure 6.43: Cocoon style vegetal fiber mummy encasing 234 Figure 6.44: Primary feeder canals in the survey area on the north side

of the watershed 239 Figure 6.45: Primary feeder canals in the survey area on the south side

of the watershed 240 Figure 6.46: Panorama of abandoned feeder canals and Chilacotacocha reservoir 243 Figure 6.47: Large maqueta at site YA-162 248 Figure 6.48: Maqueta at YA-162 249 Figure 6.49: Large maqueta at YA-162 249

viii

Figure 6.50: Maqueta at YA-014 250 Figure 6.51: Small maqueta at site YA-032 (Chiquero) 250 Figure 6.52: Map of Pukara fortifications 253 Figure 6.53: Pukara fortification of site CO-165 254 Figure 6.54: Pukara fortification of site CO-168 255 Figure 6.55: Rectangular chullpa on the exterior of fortification wall, site CO-165 257 Figure 6.56: Pumachiri peak (CO-158) 259 Figure 6.57: Outer stone wall (denoted by arrows) of CO-158 from above 260 Figure 6.58: Inner wall encircling Pumachiri peak (CO-158) 262 Figure 6.59: Pukara fortification to the west of Tuti 263 Figure 6.60: Fortified settlement to the southeast of Tuti. Note ancient road

in foreground 264 Figure 6.61: Oblique airphoto of Uyu Uyu (YA-050) 267 Figure 6.62: Architectural Map of Uyu Uyu 268 Figure 6.63: Facade of structure 104 at Uyu Uyu, from the east 272 Figure 6.64: Airphoto of San Antonio/Chijra (CO-100) 276 Figure 6.65: Promontory of San Antonio (Sector N) from the southwest 279 Figure 6.66: Airphoto of Llanka (CO-127), including canals 282 Figure 6.67: Airphoto of Kitaplaza (CO-164) 284 Figure 6.68: Airphoto of Llactapampa and Tunsa, with major canals 285 Figure 6.69: Airphoto of Llactarana 289 Figure 6.70: Architectural map, habitational sector of Llactarana 290 Figure 6.71: Colonial/Republican era house in Yanque (YA-041) 292 Figure 6.72: Detail of masonry on gable end 292 Figure 6.73: Probable LIP/LH house at the southern edge of Coporaque (CO-161) 294 Figure 6.74: Same LIP/LH house, from southeast, showing doorway 294 Figure 6.75: Church of Coporaque 295 Figure 6.76: Detail of masonry, northwest corner of church of Coporaque 295 Figure 6.77: Circular house and adjoining corral at Jibillea (YA-093) 296 Figure 7.1: Examples of colonial ceramics 301 Figure 7.2: Colonial Period settlement pattern 303 Figure 7.3: View of kallanka and probable chapel at San Antonio (CO-100) 313 Figure 7.4: Map of the probable chapel at Uyu Uyu 316 Figure 7.5: View of probable chapel at San Antonio, from the northeast 317 Figure 7.6: Frontal view of probable chapel at San Antonio 318 Figure 7.7: Abutting interior join of vestibule wall in probable chapel at San Antonio 323 Figure 7.8: Facade of the chapel of San Sebastian, Coporaque 323 Figure 7.9: Perspective of chapel of San Sebastian after restoration 323 Figure 7.10: Interior of doorway, chapel of San Sebastian, Coporaque 324 Figure 7.11: Perspective of possible chapel in Yanque, from the east 325 Figure 7.12: Frontal view of possible chapel in Yanque 325 Figure 7.13: The reduccin of Yanque, showing community spatial divisions 332 Figure 7.14: The reduccin of Coporaque 333

ix

Figure 7.15: Abandoned irrigation features and terraces on Cerro Pallaclle 339 Figure 7.16: Abandoned irrigation features, Yurac Ccacca 341 Figure 7.17: Abandoned Inca canal and associated terracing 342 Figure 8.1: Schematic, ideal sociopolitical organization of Inkaic Province

of Collaguas 352 Figure 8.2: Schematic, ayllu outliers in herding villages subject to kurakas

in Yanque and Coporaque 363 Figure 8.3: Crop declarations by village, Yanquecollaguas, 1591/1604/1615-1617 371 Figure 8.4: Average landholding area per household 374 Figure 8.5: Landholding area per household 375 Figure 8.6: Comparison of average landholdings per household amongst

households with maize fields and households without maize fields 379 Figure 8.7: Proportion of households with livestock, by village 381 Figure 8.8: Livestock declarations by village, Yanquecollaguas

1591/1604/1615-1617 384 Figure 9.1:Dot density map of crop declarations in Coporaque, 1604/1615-1617 397 Figure 9.2: Landholdings by moiety, Coporaque 1604/1615-1616 399 Figure 9.3: Panorama of the Coporaque area 403 Figure 9.4: Landholding distribution, ayllu Cupi, Coporaque Hanansaya 408 Figure 9.5: Landholding distribution, ayllu Aipi/Cupi, Coporaque Hanansaya 409 Figure 9.6: Landholding distribution, ayllu Calloca, Coporaque Hanansaya 410 Figure 9.7: Landholding distribution, ayllu Checa Malco, Coporaque Hanansaya 414 Figure 9.8: Landholding distribution, ayllu Collana Malco, Coporaque Hanansaya 415 Figure 9.9: Landholding distribution, ayllu Icatunga Malco, Coporaque Hanansaya 416 Figure 9.10: Landholding distribution, ayllu Yumasca, Coporaque Hanansaya 417 Figure 9.11: Dot density map of left and right ayllus, Coporaque Hanansaya 420 Figure 9.12: Landholding distribution, ayllu of official state potters, Coporaque

Hanansaya 422 Figure 9.13: Landholding distribution, ayllu Collana, Coporaque Urinsaya 426 Figure 9.14: Landholding distribution, ayllu Pahana Collana Pataca,

Coporaque Urinsaya 427 Figure 9.15: Landholding distribution, ayllu Pahana Taypi Pataca,

Coporaque Urinsaya 428 Figure 9.16: Landholding distribution, ayllu Pahana Cayao Pataca,

Coporaque Urinsaya 429 Figure A.1: Preliminary Colca valley ceramic sequence 453 Figure A.2: Key to illustrations of ceramics 454 Figure A.3: Chiquero body sherds 456 Figure A.4: Rim diameters of Chiquero neckless ollas 457 Figure A.5: Chiquero neckless ollas 460 Figure A.6: Chiquero neckless ollas 461 Figure A.7: Chiquero neckless ollas 462 Figure A.8: Chiquero collared ollas 463 Figure A.9: Chiquero handles and lugs 464

x

Figure A.10: Chiquero jars (A-C) and ollas 465 Figure A.11: Rim diameters of Middle Horizon bowls 471 Figure A.12: Local Middle Horizon red-slipped rim bowls 472 Figure A.13: Local Middle Horizon red-slipped rim bowls 473 Figure A.14: Local Middle Horizon red-slipped rim bowls 474 Figure A.15: Local Middle Horizon bowls 475 Figure A.16: Local Middle Horizon decorated bowls 476 Figure A.17: Local Middle Horizon bowls (A-G) and drinking vessels (H-J) 477 Figure A.18: Rim diameters of Collagua I bowls 483 Figure A.19: Rim diameters of Collagua II bowls 484 Figure A.20: Collagua I cumbrous (A-C, E), and a collared (D) bowls 489 Figure A.21: Collagua I cumbrous bowls 490 Figure A.22: Collagua I vertical-walled (A-D) and cumbrous (E-G) bowls 491 Figure A.23: Collagua I cumbrous (A, F) and slightly open (B-D) bowls 492 Figure A.24: Collagua I vertical walled (A-C, F) and cumbrous (D, E) bowls 493 Figure A.25: Collagua I vertical walled (A-E) and flaring (F) bowls 494 Figure A.26: Collagua I flaring bowls 495 Figure A.27: Collagua II vertical-walled bowls 496 Figure A.28: Collagua II rounded open (A, B), and flaring (C-E) bowls 497 Figure A.29: Collagua II rounded open bowls 498 Figure A.30: Collagua II rounded open (A-C) and flaring (D-F) bowls 499 Figure A.31: Collagua II cntaros 500 Figure A.32: Comparison of Firing Quality, Collagua I-Collagua Inka 503 Figure A.33: Comparison of slip color, Collagua I-Collagua Inka 504 Figure A.34: Rim diameters of Collagua III bowls and plates 507 Figure A.35: Rim diameters of Collagua Inka and Inka plates 507 Figure A.36: Collagua III bowls 512 Figure A.37: Collagua III bowls 513 Figure A.38: Collagua III bowls 514 Figure A.39: Collagua III bowls 515 Figure A.40: Collagua III bowls 516 Figure A.41: Collagua Inka bichrome bowls and plates 517 Figure A.42: Collagua Inka bichrome plates 518 Figure A.43: Collagua Inka bichrome plates 519 Figure A.44: Collagua Inka bichrome plates 520 Figure A.45: Collagua Inka bichrome plates 521 Figure A.46: Collagua Inka bichrome plates 522 Figure A.47: Collagua Inka bichrome (A-H) and polychrome (I, J) plates 523 Figure A.48: Collagua Inka (A, E-G, I) and Inka (B-D, H) polychrome plates 524 Figure A.49: Collagua Inka bichrome (A, B) and polychrome (C) flat-bottom bowls 525 Figure A.50: Collagua Inka bichrome pitchers 526 Figure A.51: Collagua Inka bichrome pitchers 527 Figure A.52: Collagua Inka arbalo rims and lugs 528 Figure A.53: Collagua Inka bichrome arbalo handle and body sherds 529

xi

Figure A.54: Inka bichrome and polychrome arbalo body sherds 530 Figure A.55: Inka bichrome and polychrome arbalo body sherds 531 Figure A.56: Colonial Period bowls 535 Figure A.57: Probable early colonial rimsherds. 536 Figure A.58: Probable early colonial rimsherd from YA-054 (Llactarana) 536 Figure A.59: Probable early colonial rimsherd from YA-041 (Yanque) 537 Figure A.60: Probable early colonial rimsherd from YA-050 (Uyu Uyu) 537 Figure B.1: Projectile Points 539 Figure B.2: Projectile Points 540 Figure C.1: Archaic Period sites 542 Figure F.1: Overview map showing locations of Maps F.2-F.5 586 Figure F.2: All sites registered, northwest block 587 Figure F.3: All sites registered, northeast block 588 Figure F.4: All sites registered, west-central block 589 Figure F.5: All sites registered, south block 590

xii

List of Tables Table 2.1: Dry and wet periods derived from glacial and lake sediment cores 53 Table 3.1: Visitas to the Colca valley. 86 Table 4.1: Local sequence of occupational periods 106 Table 4.2: Modern toponyms in Coporaque that also appear in the visitas 123 Table 5.1: Formative Period sites (N=30) 132 Table 5.2: Middle Horizon sites (N=37) 154 Table 6.1: Late Intermediate Period Sites (N=53) 179 Table 6.2: Late Intermediate Period Site Counts by Class 180 Table 6.3: Late Horizon sites (N=72) 188 Table 6.4: House count by settlement 197 Table 6.5: Masonry typology 203 Table 6.6: Masonry type building counts/percentages, by site 203 Table 8.1: Villages and ayllus of Yanquecollaguas Hanansaya, 1591/1615-1617 357 Table 8.2: Villages and ayllus of Yanquecollaguas Urinsaya, 1591/1604 358 Table 8.3: Collagua colonists (mitmaqkuna) living in La Chimba of Arequipa, 1582 364 Table 8.4: Crop declarations by village, Yanquecollaguas, 1591/1604/1615-1617. 372 Table 8.5: Non-local household fields, Yanquecollaguas 1591/1604/1615-1617 379 Table 8.6: Landholdings and livestock of craft specialists, Yanquecollaguas

1591/1604/1615-1617 388 Table 9.1: Summary statistics, left/right predominance, Coporaque Hanansaya,

1615-1616 419 Table 9.2: Land tenure summary, left vs. right ayllus, Coporaque Hanansaya,

1615-1616 419 Table 9.3: Summary statistics, left/right predominance, Coporaque Urinsaya, 1604 425 Table 9.4: Summary of pre-reduccin ayllu residence patterns 432 Table A.1: Summary counts and percentages: all ceramics 452 Table A.2: Chiquero ceramics surface treatment scores (exterior) 455 Table A.3: Middle Horizon ceramic surface treatment scores by paste type 467 Table A.4: Collagua I and II bowl surface treatment scores (exterior surface) 481 Table A.5: Collagua III and Collagua Inka bowl and plate surface treatment scores 504 Table A.6: Comparison of average surface treatment scores, Collagua I to

Collagua Inka 505 Table B.1: Projectile point types and date ranges 538 Table C.1: Projectile point counts, grouped by time period 541 Table D.1: Survey artifact registry 543 Table E.1: Site/sector collections sorted by ceramic style 562 Table E.2: Site/sector collections sorted by time period 574 Table F.1: Site/sector registry 591

xiii

Acknowledgements

The archaeological field research for this project was carried out with the

authorization of Resolucin Directoral No. 615 from the National Institute of Culture,

Lima. The field research for this project was made possible through the generous

funding of a Dissertation Research Grant from the Wenner-Gren Foundation for

Anthropological Research (Grant No. 6431). Initial post-field analysis and writing was

funded by a Wenner-Gren Lita Osmundson Grant. I gratefully acknowledge the support

of a Dumbarton Oaks Junior Fellowship in Precolumbian Studies during the 2001-2002

academic year. An Advanced Dissertator Fellowship from the University of Wisconsin-

Madison provided support for the final stages of write-up during the 2002-2003 academic

year.

In completing this project, I am reminded of what a privilege it has been to be

able to conduct research that fascinates me and, even more so, to be able to exchange

ideas and work with so many talented, dynamic individuals. I warmly thank my

committee members, Frank Salomon, Jason Yaeger, Neil Whitehead, Sissel Schroeder,

and Karl Zimmerer, for all of their advice and always-constructive criticism. My work

benefited tremendously from the input of such a diverse group of great scholars. I am

particularly indebted to my advisor, Frank Salomon, for my scholarly formation. His

intellect, curiosity, and incredible depth of knowledge of the Andean cultural world are

truly inspirational. I am grateful for his unflagging support throughout all stages of my

graduate studies. The early stages of formulating this project developed out of numerous

xiv

stimulating conversations with Neil Whitehead, who urged me to undertake an

intersectional course of study. I could never have anticipated the long and windingbut

ever-more scenicpath this decision would send me down. A long, strange trip indeed.

Jason Yaeger, Co-Chair of the dissertation, was tireless and selfless in providing me with

detailed feedback, often on short notice, through the analytical and write-up phases of the

project, and was always available just to talk. Thanks, friend.

I am grateful to Steve Stern, Florencia Mallon, James Stoltman, Mark Kenoyer,

and Gary Feinman for their support and advice during the early stages of the project. I

owe a particular debt of gratitude to Bill Denevan, who pulled out his maps and airphotos

during an office visit and, with infectious enthusiasm, encouraged me to go to the Colca

to explore research options. Two professors from my years as an undergraduate at the

University of Iowa, Mercedes Nio-Murcia and Laura Graham, had a more lasting impact

on me than they know, and inspired me to pursue graduate study in anthropology.

One incredible individual and dear friend, Maria Benavides, profoundly affected

my ability to undertake this project by so generously providing me with her entire

collection of archival documents related to the colonial history of the Colca valley,

including photocopies and transcriptions of the visitas that are the centerpiece of the

ethnohistorical portion of the study. My warmest thanks to her for her generosity and

good faith in my ability to make a worthy contribution with this treasure of

documentation. I treasure her wisdom and friendship.

In Lima, Kristoff Makowski, who took an interest in my project from the first

time I met him, opened his home to me and assisted me through the National Institute of

xv

Culture permitting process. I look back fondly on our many invigorating conversations

and thank him for his collegiality and generosity. Special thanks to Luis Jaime Castillo

Butters for his support in seeing my proposal through the proper channels at the INC.

My heartfelt thanks go to Willy Ypez Alvarez and Erika Simborth Lozada, who

assisted in all aspects of the archaeological fieldwork and postfield analysis. We

developed a special bond that only working and living together in the field can create. As

experts in the archaeology of the Department of Arequipa, their contribution to the

project transcended the technicalI hope this dissertation does their hard work justice.

Many thanks to Ericka Guerra Santander, Co-Director of the Colca Valley Regional

Survey Project, for her contributions and work. I am grateful for all of the assistance of

Tamara Flores Ramos, who worked for hundreds of hours entering the visita data into the

database, photographed artifacts, and digitized ceramic diagrams. Thanks to Evelin

Lpez Sosa for her help in the photo documentation of the artifact collections. Bruce

Owen graciously provided me with a copy of a ceramic data matrix and coding scheme

that proved very useful, and answered my other questions in detail while I was working

in the lab.

The archaeological portion of this project wouldnt have been possible if not for

the goodwill and graces of the villagers of Yanque and Coporaque, who embraced us,

educated us, and often even fed us and slaked our thirst with chicha as we stumbled

ponderously through their fields. My debt to my padrinos in Yanque, Gerardo Huaracha

Huaracha and Doa Luisa Cutipa de Huaracha, is greater than I can ever repay.

Diuspagarasunqui. I warmly thank the Mayors of Yanque and Coporaque, Ramn

xvi

Cayllagua Cayllagua and William Bernal Huarca, for their support and collaboration with

my efforts. My gratitude to Justino Inka Montalvo, President of the Irrigators Committee

of Yanque Urinsaya, for allowing us to accompany the community during the annual

cleaning of the Misme canal. Sister Antonia Kayser of the parish of the Immaculate

Conception of Yanque always opened her doors to me and cared for me when I fell ill.

In Arequipa, I lived and worked in the research house of the Centro de

Investigaciones Arqueolgicas de Arequipa (CIARQ)a true home away from home in

sunny Sachaca. I thank Karen Wise and Augusto Cardona Rosas, Co-Directors of

CIARQ, for all manner of logistical and bureaucratic support.

My knowledge of the archaeology and ethnohistory of Arequipa owes largely to

the many stimulating conversations with Mximo Neira, pioneer of Colca valley

archaeology, and with Flix Palacios Ros, Guillermo Galdos Rodriguez, and Augusto

Cardona Rosas. I thank Luis Sardn Cnepa, Director of the National Institute of Culture

in Arequipa, as well as my friends on the staff of the Department of Archaeology of INC-

ArequipaPablo de la Vera Cruz Chvez, Lucy Linares Delgado, Marko Lpez Hurtado,

and Cecilia Quequezanafor helping me through the bureaucratic aspects of the

permitting and inspection process.

After returning from the field, I had the good fortune of spending an academic

year with a vibrant community of scholars at Dumbarton Oaks. I am especially grateful

to Jeffrey Quilter, Director of Precolumbian Studies, for all of his great advice and input

on the project. George Lau, Allan Maca, and Carolyn Tate, my colleagues while at

Dumbarton Oaks, all helped me formulate my ideas in the early stages of writing, and

xvii

contributed immeasurably to my thinking. Warm thanks to Loa Traxler and Jennifer

Younger for their detailed feedback and help.

My family has always been a bedrock of love and support for me, and had more

faith in me than I sometimes deserved. My parents, Arlyn and Grace Wernke, always

encouraged me to explore my interests and find my own path. My sister, Suzanne Bautz,

encouraged me every step of the way. Without their supportpersonal, emotional, and

financialI wouldnt be where I am today.

Finally, my deepest thanks and gratitude go to my wife, Tiffiny Tung, who, more

than any other person, helped me through this process and was there for me at every

critical juncture, even as she was finishing her dissertation at the same time. Her

formidable intellect and keen observations have fortified this dissertation. Thank you for

always believing in me.

xviii

Abstract

This study investigates relationships between political organization and land-use

from Formative through early colonial times in the southern Andean highlands. The

project combines a detailed synchronic view of agro-pastoral production and exchange

with a diachronic view of how the cumulative effects of distinct land-use practices under

autonomous and imperial rule transformed the built landscape and shaped later land-use

and political economic strategies. The archaeological portion of the project is based on a

90 km2 systematic survey in the Colca valley of Arequipa Department, Peru. The

ethnohistorical portion reconstructs the areas local- and regional-scale economic

relationships through spatial analysis of landholding and livestock declarations in a series

of colonial censuses (visitas).

The survey findings reveal a marked expansion of settlement and irrigated

agricultural infrastructure during the Late Intermediate Period (LIP; AD 1000-1450),

when a series of villages with distinctive Collagua domestic architecture were

established. During this period, inter-settlement political relations appear to have

oscillated between competition and cooperation. Defensible settlement locations and

fortifications signal conflict, while hydrological relationships among long primary canals

indicate that water apportionment was coordinated at a supra-settlement scale. Upon

imperial incorporation of the valley, Inka administration hierarchized previously

heterarchical political relations. The Inkas established an administrative center, and

major LIP settlements became secondary administrative sites, where rustic Inka imperial

architecture was prominently situated in association with local elite domestic structures.

xix

Analysis of Spanish visitas provides a complementary view of how local kindreds

(ayllus) had been reordered into a nested administrative hierarchy under Inka rule. The

formal ayllu structure mimicked Cuzco Inka norms unevenly, with startlingly close

matches occurring in the lower moiety. Reconstructed early colonial ayllu land tenure

patterns reveals dispersed household landholdings organized by ayllu affiliation.

Comparison of ayllu land tenure constellations with the terminal prehispanic settlement

pattern provides a basis for retrodicting Inka-era ayllu residence patterns. This analysis

shows that the highest-ranking ayllus resided at the administrative center and at

secondary administrative sites. At the regional scale, ayllus and their leaders articulated

economic flows between high altitude herding populations and specialized maize

production enclaves in low-lying valleys to the south.

PART I: INTRODUCTION AND BACKGROUND

1

Chapter 1 Community and Landscape in the Andes

Introduction

The complex interplay between political organization and land-use patterning in the

ancient Andes has long been a subject of intense scrutiny and debate. Researchers have

illustrated how diverse complementarity relationships between the altitudinally-dispersed

resources of the Andean landscape and the cultural schemata and social forms devised to

exploit it developed throughout Andean antiquity. In the two generations of research

following the paradigmatic formulation of anthropologist John Murra (1964; 1968; 1972),

scholars have traditionally approached these complementarity relationships as a means of

cultural-ecological adaptation, guided by a putatively ancient Andean ideal of community

self-sufficiency. At the same time, household and regional-scale studies have illuminated

connections between changing land-use patterns and changes in, respectively, domestic and

political economy (see Masuda, et al. 1985). The wider temporal frame of archaeological

studies has had a relativizing effect; structured relations between ecology and economy

that in an ethnographic frame appear stable and locked in homeostatic balance appear more

dynamic and historically contingent over archaeological time spans (Erickson 1988, 1999,

2000; Rice, et al. 1989; Stanish 1989a, 1992, 2003; Van Buren 1993, 1996).

In the Colca valley, a major Pacific drainage in the semiarid western flanks of the

southern Peruvian cordillera (Department of Arequipa), researchers have contributed to this

dynamic view through combined archaeological, historical, and ethnographic study. The

Colca valley is one of the most intensively terraced locales in the New World (Denevan

2

2001; Donkin 1979), making it an ideal location for the investigation of prehispanic Andean

political dynamics and land-use patterning. A palimpsest mosaic of approximately 11,000

ha of agricultural terraces covers virtually all geomorphic surfaces below about 4000 masl

in the valley. Previous researchers have identified a sequence of major transformations of

the valley landscape in which unirrigated sloping fields and terraces that were constructed

and used at least as early as the beginning of the Middle Horizon (AD 500-1000) were

subsequently abandoned and replaced by massive complexes of irrigated bench terrace

systems during the Middle Horizon and Late Intermediate Period (hereafter LIP; AD 1000-

1450), a process that probably intensified during the Inka occupation during the Late

Horizon (Brooks 1998; Denevan 1987; 2001:185-205; Malpass 1987; Shea 1987; Treacy

1989).

Explaining such long-term dynamics in production and sociopolitical organization

remains a central issue of debate in Andean archaeology, as well as in other world regions.

Working from a top-down political economy perspective, many archaeologists have

focused on control over labor deployment and its produce as the primary factor in the onset

of agricultural intensification and social inequality, contending that the construction and

maintenance of intensive agricultural infrastructure is predicated on the compulsion of non-

elites to produce above their subsistence needs by a stratum of political elites (Earle 1987;

Kolata 1986, 1993, 1996a; Stanish 1994). By contrast, others have argued that households

or other small-scale corporate collectivities can, and have, constructed large-scale, intensive

agricultural systems such as high altitude irrigated terrace and raised-field systems in the

3

Andean highlands (Erickson 1988, 1999, 2000; Gelles 1990, 2000; Graffam 1990, 1992;

Treacy 1989).

Locally, hypotheses have focused on terracing and irrigation in the Colca valley as

technological innovations for adapting to climatic perturbations (Brooks 1998; Denevan

1987; 2001:185-205; Treacy 1989). Although they identify relevant variables, these

hypotheses appear incomplete, since they do not account for the ways in which such factors

vary in relation to demographic, economic, or political factors. Unlike most other areas of

the Andes, landscape archaeology in the Colca valley has outpaced our knowledge of

settlement organization and patterning. As a consequence, we know comparatively little

about the origins, development, and organization of the two major late prehispanic ethnic

polities of the valley, the Collaguas and Cabanas, and how their historical trajectories relate

to the observed changes in the valleys built landscape. Document-based studies of

protohistoric Collagua political and economic organization have presented the Collagua as

a classic example of a large-scale ethnic seoro (chieftaincy) with outlier ethnic colonists

settled in neighboring, low-lying valleys to the south, forming a vertical archipelago akin to

the archetypal example of the Lupaqa polity of the neighboring Lake Titicaca Basin

(Benavides 1987b; Galdos Rodrguez 1984, 1987; Mlaga Medina 1977; Pease G. Y.

1977). Test excavations have yielded a preliminary ceramic sequence that identifies a local

LIP/Late Horizon ceramic style associated with the Collagua (Brooks 1998; de la Vera

Cruz Chvez 1987, 1988, 1989; Malpass and de la Vera Cruz Chvez 1986, 1990; Neira

Avendao 1961; Shea 1997), and reconnaissance projects have recorded several major

settlements from these late prehispanic periods with well-preserved, distinctive local

4

domestic architecture (Brooks 1998; de la Vera Cruz Chvez 1988; Guerra Santander and

Aquize Cceres 1996; Neira Avendao 1961; Oquiche Hernani 1991; Shea 1986a, b, 1987).

However, no systematic, full-coverage survey had been conducted in the valley prior to this

project, leaving several critical questions unanswered, such as: What are the culture-

historical antecedents of the Collaguas and Cabanas, and how were these pre-LIP local

groups organized? How did the Wari and Tiwanaku Middle Horizon imperial states

influence local political and economic organization and, by extension, the configuration of

anthropogenic landforms in the valley? How did local Collagua communities develop out

of the post-Middle Horizon regional balkanization during the LIP, and how were they

organized prior to Inka imperial incorporation?

Ethnohistorical reconstructions of the political organization of the Inkaic Collaguas

province suggest a penetrating imperial presence in which local kindreds (ayllussee

below) were reordered into a formal hierarchy according to Inka categories of rank

(Benavides 1989; Cock Carrasco 1976-77, 1978; Gelles 2000; Guillet 1992:18-19;

Prssinen 1992:362-371; Pease G. Y. 1977; Robinson 2003; Rostworowski de Diez

Canseco 1983:121-123; Zuidema 1964:115-118). Yet in the absence of systematic

settlement pattern data, archaeological indices for Inkaic influence or presence in the valley

remained poorly defined and debated (Brooks 1998; de la Vera Cruz Chvez 1987; Malpass

1987). Thus, critical questions have been left unanswered regarding how local and imperial

structures of power articulated under Inka rule: Did the Inka significantly reorganize the

local settlement system, or was Inka administration mediated through extant settlements

5

and their agents? How did changes in Late Horizon settlement patterning relate to the

(re)configuration of local agricultural landforms?

In this dissertation, I address these questions through a spatially-integrated

archaeological and ethnohistorical investigation. This study tracks the origins and

development of the communities in the core area of the Collagua polity and explores how

their later incorporation into the Inka and Spanish empires related to changes in community

organization and land-use patterning. The project is multi-scalar, both spatially and

temporally, and combines analysis of two primary data sets: first, archaeological data from

a 90 km2 full-coverage survey in the central section of the Colca valley that coincides with

an area described in colonial documentary sources as the seat of authority in the Inkaic and

colonial province of Collaguas; and second, demographic and cadastral data from a series

of detailed early colonial censuses (visitas) of an area that overlaps with the archaeological

survey coverage. I use the archaeological data to provide a long-term view of the origins

and development of the communities in this core area of the Collagua ethnic polity. My

analysis focuses on changes in settlement and agro-pastoral infrastructural patterning that

occurred between the period of autonomous Collagua rule during the LIP and the

subsequent Inka and early Spanish imperial occupations. In complement to this temporally-

deep local view, I use the visita data, from the years 1591, 1604, and 1615-1617, to

reconstruct both the regional political and economic organization of the Inkaic and early

colonial province of Collaguas, and the land tenure patterns of local communities within the

archaeological survey area. Within the survey area, I also compare the early colonial land

tenure patterns of local ayllus to the terminal prehispanic settlement pattern as a means of

6

retrodicting their Inka-era residence patterns, thereby affording a detailed, culturally-

informed view of how hybrid local/imperial sociopolitical institutions were mapped onto

the settlement system under Inka rule.

A Political-Ecological Approach to Community and Landscape

The Human/Environment Interface: Landscape as Household and Habitat

Understanding diachronic relationships between communities and the landscapes

they inhabit requires analytical frameworks that can account for both constrictive, limiting

processes and constructive, generative ones. Recent approaches are working in this

direction, as the reactive, one-way causality implicit in the concept of "adaptation" has been

challenged by models of "interpenetrating," "dialectic," and "recursive" relationships

between humans and their environment. Under various rubrics, such as political ecology

(e.g., Bryant 1992; Zimmerer 2000), historical ecology (e.g., Bale 1998; Crumley 1994;

Kirch and Hunt 1997; Patterson 1994), and landscape archaeology (e.g., Bender 1992;

Crumley 1999; Erickson 2000; Knapp and Ashmore 1999; McGlade 1995; Rossignol and

Wandsnider 1992; Treacy 1994; Wagstaff 1987), recent formulations are collectively

changing underlying assumptions regarding human/habitat relationships, as dynamic

disequilibrium has come into favor as a guiding principal over homeostatic balance (Botkin

1990). Simultaneously, a shift toward understanding landscapes as historically-contingent

and anthropogenic has highlighted the shortcomings of the nature:culture dyad still

prevalent in the literature, which reduces explanations of landscape dynamics to a

comparison of the causal forces of natural versus cultural systems of activity. The

7

inadequacy of this ecology:economy opposition, however expressed, becomes especially

apparent when considering that both terms derive from the Greek root okow (oikos)

household/habitata fact that points toward an alternative approach that considers

human synergy with the environment as the dynamic process at the interface between

nature and culture (Whitehead 1998).

The paradoxes that arise from this economic:ecological ambiguity are especially

apparent in the Andean region. Discussion of human/environment relationships in the

Andes has been dominated by the vertical complementarity model of Murra (1964; 1968;

1972), which posits that the altitudinally-compressed ecological zonation of the Andes

required societies of varying scale and complexity to adopt mechanisms of access to a

sufficiently diverse resource base. Rather than relying on trade with outside groups or

centralized markets, the uniquely Andean solution, according to Murra, was the direct

colonization of multiple ecological tiers by a single ethnic group, forming a vertical

archipelago of settlements linked through intra-ethnic reciprocal and redistributive

exchange (Murra 1972). Thus, the maximum control of ecological tiers came to be

viewed by many as a pan-Andean cultural ideal of balancing resources and demography,

differentially achieved through space and time. The development of asymmetrical systems

of redistribution was therefore treated as epiphenomenal to the adaptive process of securing

ecologically-diverse resources (see Van Buren 1993; Van Buren 1996).

However, both the ecological and political-economic aspects of the verticality

model have come under increasingly critical scrutiny. Ecological anthropologists and

geographers have noted that the construction of stable and secure production systems in the

8

Andes is predicated on a simplification of highly localized ecological variability by humans

(Brush 1977:9). Humans transform ecological gradients into ecological tiers (Erickson

2000; Zimmerer 1999), themselves composed of smaller anthropogenic production zones

(Mayer 1985) created through human interventions such as irrigation, augmentation of

soils, and alteration of microclimates. These observations require that a more dynamic

view be developed that is mindful of the ways in which the landscape itself is altered by

humans so as to lessen ecological strictures. As Mayer (1985:47) observes,

When we think of production zones as man-made things, rather than as adaptations to the natural environment, our attention is directed to how they are created, managed, and maintained. Then the importance of the political aspects of control by human beings over each other in relation to how they are to use a portion of their natural environment will again come to the fore. Recent researchers have indeed emphasized the political aspects of control

inherent in vertical complementarity. Archaeological investigations of the prototypical case

study of a vertical archipelago systemthat of the Lupaqas of the Titicaca Basinhave

found that the outlier Lupaqa settlements in the lowland valleys of Moquegua were not of

the great antiquity hypothesized (Stanish 1985, 1989a, b, 1992), and probably never

functioned to provision whole populations (Van Buren 1993, 1996, 1997). Building on

these findings, Van Buren has presented a general critique of the verticality model as overly

functionalist and essentialist, suggesting instead that vertical complementarity is better

viewed as a dimension of social poweri.e., as a mobilization of labor and material by

ethnic lords towards specific political ends (Van Buren 1993, 1996).

In this dissertation, I propose a bridging position that approaches the dynamics of

land-use and sociopolitical change as complex, multi-causal processes both constructed

9

culturally and constricted ecologically. I explore human/environment synergistic processes

through a theoretical framework that posits a recursive relationship between a particular,

Andean oikosthat of the late prehispanic and early colonial Colca valley and surrounding

regionand the ecological and economic practices of the members of local Collagua

communities. Here, rather than documenting the formal attributes of a vertical

complementarity system as a social formation or structure, the processual variability of

land-use practices is of primary interest. Within my formulation, ecological and economic

practices are structured by, while also constituting through their aggregate effects, the

household/habitat of oikos. This approach therefore conceptualizes complementarity as a

particular kind of agency, related recursively to the structures of landscape and community

organization.

Analysis at several temporal and spatial scales is required to understand the

historical and spatial relationships between built landscapes and the communities that alter

and adapt to them. Diachronically, the construction of built features, such as canals, terrace

complexes, anthropogenic soil regimes, and so on, is an aggregative process that alters the

physical and energetic parameters of production at varying rates and scales, depending on

historical circumstances. Synchronically, anthropogenic landscape features constitute the

congealed labor (Lansing 1991:12) or landesque capital (Blaikie and Brookfield

1987:9-10) that structure production. While political changeas in the case of the Inka

and Spanish conquests of the Andean regioncan occur over short (event- and

conjuncture- level) time spans, landscape-scale infrastructures are generally designed to

be durable and stable (Braudel 1972). Systems of production, qua systems, are organized

10

to produce reliable returns on given inputs of labor and material, and in this sense are

inherently conservative in nature. These contrasting temporal rhythms of political and

economic-ecological change therefore produce a dialectic from which new, hybrid social

formations emerge (Smith 1992a, b). Thus, built landscapes have tangible agentive force

that both constrain and present opportunities for economic and ecological praxis. In order

to understand these varied scales of spatial and temporal interaction, this dissertation

combines analysis of long-term changes in the built landscape of the Colca valley, with a

synchronic view of regional and local-scale patterns of production and exchange during

early colonial times.

Communities, Natural and Imagined: The Case of the Andean Ayllu

My approach to the archaeological and historical study of ancient communities

builds on interactionist frameworks which posit that communities are dynamic, socially-

constructed institutions that structure and are structured by supra-household interactions

(for a review of this and other approaches, see Yaeger and Canuto 2000). I conceive of

communities as matrices of social interaction that both create and emerge from a sense of

common interest and affiliationa sense of shared identity. This orientation contrasts with

functionalist and behavioralist approaches, which consider community as the natural,

fundamental unit of supra-household social and biological reproduction, necessarily

constituted by co-residence, proximity, and shared economic/ecological praxis (e.g.,

Murdock 1949; Redfield 1955, 1956). According to the behavioralist conception, the

quality of distinctiveness (Redfield 1955, as cited in Kolb and Snead 1997:611) that

11

defines community identity is epiphenomenal to the everyday interactions of living and

laboring in a bounded and defined space (for critique, see Isbell 2000; Wolf 1956). Some

archaeologists have recently advocated this framework as the most pragmatic and testable

approach to reconstructing and comparing prehistoric community organization (Kolb and

Snead 1997).

Clearly, proximity and co-residence affect the patterning and frequency of

interaction between social actors and groups, and community identity often is expressed in

the idiom of settlement, neighborhood, or territory. Patterns of daily interaction constitute a

structure of power and meaning that constrain the parameters of imagination and action

(Bourdieu 1977), although through critical awareness and reflexive monitoring, agents can

question, contest, and alter those structures (Giddens 1979). However, community identity

is not only determined by the propinquity for frequent interaction, but is also imagined

(Anderson 1991). Communities can be composed of individuals who do not frequently

interact or even know one another, but nonetheless share a deep sense of affiliation and

common interest (Anderson 1991). The sense of solidarity in communities emerges from

discourses and practices of affiliation (Yaeger 2000) that essentialize within-group

commonalities and interests (us) in contrast to others (them) (Barth 1969). Within my

conceptualization, therefore, ethnicity is one of many dimensions of community identity

an approach, as I discuss below, entirely befitting of the Andean context. In sum, as a

dynamic process of social identity, community need not be associated with any particular

socio-spatial scale or unit; communities are just as likely to cut across spatial boundaries as

to adhere to them (Goldstein 2000).

12

Ever since the early years following the European invasion (Polo de Ondegardo

1917 [1571]), students of Andean communities have recognized their supra-local,

archipelagic patterns of residence, production, and exchange (Murra 1964, 1968, 1972).

The concept of ayllu was central to the social, political, and economic articulation of

territorially-discontinuous communities in the Andes (Abercrombie 1986; Cock Carrasco

1981; Isbell 1997; Platt 1982; Salomon 1991; Spalding 1984). Ayllu, commonly translated

as clan, was emically-defined as a multi-scalar concept that could reference any segment

along a continuum of socially- or biologically-related collectivities, from the consanguines

of a patrilineage, to clan-like groupings of patrilineages, moieties, and even an entire ethnic

group. For example, Platt (1986) has illustrated how the modern-day Macha of Bolivia

conceive of ayllus in a nested fashion, from minimal ayllus of small patrilocal groups of

neighboring households, to minor and major ayllus made up of groups of related ayllus,

and finally a maximal ayllu encompassing the ethnic group as a whole. The scalar

plasticity of the concept has led to a plethora of definitions that emphasize different aspects

of ayllu membership and organization (for reviews of definitions, see Goldstein 2000:184-

186; Isbell 1997:101-135; Salomon 1991:21-23; Spalding 1984:28-29). Common to all,

however, are two attributes: ayllus are resource-holding corporate collectivities, and ayllu

membership is reckoned by reference to an actual or fictive focal ancestor.

As resource-holding collectivities, protohistoric ayllus mediated households access

to agricultural land and other immovable assets (Patterson and Gailey 1987; Rowe 1946b;

Salomon 1991:22). While ayllus held these resources as corporate entities, ayllu members

gained access to land and other resources in reciprocity for their labor in collective work

13

projects (e.g., canal cleaning, terrace construction and maintenance) and their participation

in rituals of affiliation, including ancestor-veneration (see below). However, ayllu

landholdings were not always spatially discrete, and in this sense they were not necessarily

conceived of as bounded units (Murra 1980:30-31). Ayllus of different scales of

inclusiveness organized agro-pastoral infrastructural systems and the labor to construct and

maintain them, from the daily interactions of cultivation, terrace maintenance, and the

distribution of irrigation water, to larger-scale mobilizations such as canal construction and

cleaning (Gelles 1993, 1995, 2000; Guillet 1978, 1981, 1992; Mitchell 1976; Sherbondy

1982; Treacy 1989). Like their land tenure patterns, ayllus were not residentially discrete;

not only could ayllus occupy several vertically- and horizontally-dispersed settlements, but

members of several ayllus could also share single settlements. For example, early colonial

native testimony from Huarochir province indicates that prehispanic settlements were

composed of multiple ayllus (Salomon 1991:23-24). Also, the protohistoric Andean

concept of settlement, or llacta, did not necessarily coincide with a nucleated settlement,

but appears instead to have encapsulated a huaca (a shrine, object, or landscape feature that

embodies the superhuman founder ancestor of an ayllu), its territory, and the people within

it (Salomon 1991:23-24). So even the concept of a settlement in the late prehispanic and

early colonial Andes was not territorially bounded in the sense of a discrete cluster of

houses.

As ancestor-focused kindreds, ayllu affiliation was reckoned by reference to a

pointillist landscape of ancestral huacas that were hierarchically-related in space and time.

As Goldstein (2000:185) notes,

14

...even those aspects of ayllu identity that explicitly refer to place refer not to spatial boundaries, but to the huacas that link an ayllu to its ancestors...This association of ancestor worship to group identity suggests that as a community form, the ayllu is more genealogical than territorial in natureit is bounded by history rather than borders.

Members of minimal ayllus traced their affiliation by reference to a focal, chartering

ancestor, usually embodied in the actual mummified corpse of that individual. The criteria

used in the actual reckoning of inclusion is a perennial topic of debate in Andean

anthropology, but in practice, the precise genealogical relationship appears to have been of

less importance in the reckoning of ayllu membership than a persons social conduct and

political standing as a genealogically-connected individual (Salomon 1991:22; Spalding

1984:28-29). Ayllu members reaffirmed and reified their community affiliations by

consulting and feting their ancestral mummies, who occupied cities of the deadclusters

of above-ground, multiple-interment mortuary monuments (houses of the dead, or

chullpas in modern archaeological parlance), often situated on prominent hilltops or under

cliffs near settlements (see Dillehay 1995; Isbell 1997). Such ancestral mummies were also

considered huacas (Salomon 1995). As such, they were also conceived of as the proximate

descendants of a hierarchy of superhuman huacaseach the guardian of fertility in its

domain, and each increasingly remote, both spatially and in terms of kinship reckoning

terminating at its apex in the origin-place (place of dawning, or pacarina) of an entire

ethnic group, usually a prominent mountaintop (Salomon 1991; Spalding 1984).

As political entities, I argue that the ayllu represents an example of heterarchical

organization, in which ...each element possesses the potential of being unranked (relative

to other elements) or ranked in a number of different ways... (Crumley 1979:144). A

15

concept originally developed in the field of artificial intelligence (see Crumley 1987),

heterarchical organization is based on lateral connections between elements in an

organization, each of which can take a more dominant role depending on the circumstances.

By contrast, hierarchical organization is based on the vertical integration of elements of

fixed rank. Examples of complex, orderly, heterarchical systems abound in nature,

including the neural network of the human brain, or the patchiness of plant and animal

communities such as in the case of the ecological tiers of the Andes. Hierarchy represents

only a specific kind or state of heterarchical organization and in this sense is a more

restricted concept (Crumley 1979:145). Crumley (1975; 1979; 1987; Ehrenreich, et al.

1995) has proposed that heterarchy provides an important alternative metaphor and model

for the comparative and diachronic study of social complexity, which has long been

conflated with hierarchy. Thus, rather than charting the presence, absence, or degree of

complexity of a social system by reference only to its degree and strength of vertical

integration, the concept of heterarchy provides a more three-dimensional perspective that

also accounts for flexible, horizontal integration between social groups that can be

variously ranked according to historical circumstances.

Testimony in early colonial texts indicates that the ethnic polities subsumed by Inka

administration were self-defined as bundles of rival ayllus of fluid prestige, rank, and

wealth (see, e.g., Rostworowski de Diez Canseco 1983). The political organization of

protohistoric ayllus, as well as inter-ayllu power relations, were governed according to both

ascribed and achieved criteria of rank. The huacas of clan-level ayllus generally were

conceived of as siblings borne of the origin huaca of the ethnic group (its place of

16

dawning, or pacarina). For example, as we know from the Huarochir testimony, the

origin huacas of clan-like ayllu sets often were ranked according to their birth order from

the apical huaca in the ethnic charter myth. According to this ideology, ...the firstborn or

leading member of a set (e.g., noble heads of a villages component ayllus) functions within

the set as first among equals, but outside the set as the totalizing representative of it

(Salomon 1991:20). In this way, native lords (kurakas), as ayllu representatives, could also

be ranked according to an ideal social structure derived from the relative birth-order

position of their respective huacas. However, crosscutting these ideal-typical criteria of

genealogical inequality, individuals and ayllus could also achieve higher rank through

supremacy in warfare. The thousands of Late Intermediate Period hilltop fortifications

(pukaras) throughout the central and south-central Andes strongly indicate that both intra-

and inter-ethnic conflict was endemic prior to Inka consolidation. For example, the

distribution of pukaras in the northwestern Lake Titicaca Basin indicates that the Qolla and

Lupaqatwo major ethnic polities with populations of nearly 100,000 peoplewere much

more politically decentralized (i.e., heterarchical) prior to Inka incorporation than their

leaders early colonial memorial accounts depicted (Stanish 1997a; 2003:209-220). Indeed,

their coherence as political entities may have been most salient primarily when they were

faced with a common external threat such as the Inka army (Stanish 2003:209-220). Such

macro-scale hierarchical political organization appears to have been rather exceptional and

ephemeral in the Andean highlands during the LIP.

Elite manipulation of ayllu ideologymost elaborated by the Inkainvolved the

representation of asymmetrical, hierarchical relationships within and between clan-like and

17

maximal- ayllus in the same terms as the lateral, symmetrical links between consanguines

of lineage-like micro-ayllus (Murra 1956, 1980). Following in the substantivist tradition

of Murra, archaeologists and ethnohistorians have demonstrated how Inka strategies of

expansion and consolidation were so successful in part because they manipulated key

cultural principles and practices familiar to their subjects, such as vertical complementarity

as a logistical system for articulating populations and resources (see Masuda, et al. 1985),

conspicuous public feasting as a primary forum for displaying state largesse (e.g., Morris

and Thompson 1985), and ancestor veneration as a primary idiom of political discourse

(e.g., Rostworowski de Diez Canseco 1983).

Clearly, ayllus constituted the primary building blocks of imperial administration,

and Inka statecraft relied heavily on the representation of state/subject relations as an

extension of ayllu relations. However, in effect, Inka policies also significantly altered

intra- and inter-ayllu relations, and fomented the formation of new imagined communities

of the state. The Inkas sought to build stable, hierarchical structures of governance out of

heterarchically-organized ayllus by amplifying and codifying latent or extant inequalities of

rank and its means of attainment, while preserving the corporate character of the ayllu

itself. In Cuzco, the royal ayllus (panacas) sealed-off achieved status as a criterion of rank,

and membership rank was narrowly defined according to bilateral reckoning (Julien 2000;

Zuidema 1977). In the provinces, imperial administration submerged local differences as

some ethnic groups and elites were promoted in status relative to others. As a result of

Inkaic consolidation and centralization, the authority and domains of pliant ethnic elites

expanded greatly under Inka rule. The process of consolidating competing ethnic polities

18

into vertically-integrated provincial units has been demonstrated in many cases, from the

northern provinces of Pichincha and Imbabura in modern Ecuador (Salomon 1986), to the

Cajamarca province of northern Peru (D. G. Julien 1993), the Wankas and Xauxas of the

Mantaro valley in the central Peruvian Andes (Costin and Earle 1989; D'Altroy 1987,

1992), and the Qolla and Lupaqa of the Titicaca Basin (Hyslop 1976; Julien 1983;

Lumbreras 1974a; Stanish 1997b). In this study, I illustrate in detail how the autonomous,

heterarchical communities of the Collagua ethnic polity of the Late Intermediate Period

were reordered under Inka imperial occupation according to a formal hierarchy structured

in the image of Inka ideals.

Alternative Hypotheses and Their Correlates

Given the inherently broad temporal and spatial scope of the study, the following

hypotheses are designed to take soundings across the broad historical arc of settlement and

agricultural production in the Colca valley, ranging from the earliest period of sedentary

human occupation associated with unirrigated agricultural production, the intensification of

settlement and production during the LIP and Late Horizon, to the subsequent contraction

of settlement, population, and production during early colonial times.

As I discussed above, most field research in the Colca valley has been focused on

understanding changes in the built landscape of the valley. The findings of previous

researchers can be divided into a sequence of four principal periods of landscape

transformation: 1) an early period, dating at least as early as the beginning of the Middle

Horizon and almost certainly earlier (Brooks 1998; Treacy 1989), in which Colca valley

19

agriculturalists cultivated in unirrigated sloping field and terrace systems, many clustered

around, and augmenting, natural drainages (quebradas), 2) a subsequent phase, dating most

likely to the Middle Horizon and Late Intermediate Period, during which local peoples

abandoned these unirrigated field systems and constructed extensive canals and massive

complexes of rock-faced bench terrace complexes covering virtually all slopes and surfaces

in the lower reaches of the valley (Brooks 1998; Malpass 1987; Treacy 1989), 3) a period

of probable expansion and intensification of irrigated bench terrace production under Inka

rule during the Late Horizon (AD 1450-1532) (Malpass 1987; Shea 1987; Treacy 1989),

and 4) a period of deintensification of production during colonial and early republican

times, including the abandonment of many bench terrace complexes and canals, especially

along the upper margins of the valleys agricultural core (Denevan 1987; 2001:185-205).

Early Agriculture and Settlement in the Colca Valley

Building on these findings, the first hypothesis is aimed at clarifying the nature of

productive and sociopolitical dynamics during the earliest agriculturalist occupation of the

valley, associated with an accretionary pattern of unirrigated sloping fields and terraces.

Given that agriculture is impossible without irrigation in the valley today, previous

researchers have focused on climatological and technological factors for explaining the

functioning of these early rainfed- and runoff-based field systems, suggesting that, without

irrigation canals, early agriculturalists adapted to wetter climatic conditions by situating the

fields around natural drainages and harvesting water by diverting runoff to the fields

(Brooks 1998; Treacy 1989).

20

These explanations, however, only account for factors of ecological constriction (the

absolute moisture requirements of crops) and economic construction (water harvesting

techniques). In accord with the general model reviewed above, these factors are expected

to both structure and be structured by political organization. Treacy (1989:122-126) has

observed that the accretionary pattern of these early fields suggests that they were not built

as integrated, centrally-administered field complexes, but instead formed conglomerations

of segmented terraces, each of which could have been built by individual households or

other small-scale collectivities. The lack of settlement evidence, however, has precluded

any thorough evaluation of the timing of their construction and use, or of the social

organization of local groups that built and used them. In fact, prior to this project, no pre-

Middle Horizon settlements had been reported in the Colca valley. This is a major void in

the local occupational sequence, since the shift to sedentary, agricultural lifeways in

surrounding localesbest documented in the Lake Titicaca Basin to the east (Aldenderfer

1989, 1998; Bandy 2001; Bauer and Stanish 2001:138-141; Hastorf 1999; Stanish 1997a;

2003:99-109; Stanish, et al. 1994; Steadman 1995)occurred between about 2000 and

1300 years before the Middle Horizon, during the Early Formative (ca. 1500-800 BC).

Before this project, no Formative Period ceramics had been identified in the Colca

valley. Reportedly high (but unspecified) proportions of Middle Horizon ceramics were

recovered from surface reconnaissance of these early fields, but it seems unlikely that these

low intensity, unirrigated field systems would have been built and used as late as the

Middle Horizon, a period in which intensive irrigated agricultural systems spread

throughout the Andean highlands (Moseley 1992:216-230). By contrast, Brooks

21

(1998:400-401, 405) has recently reported much earlier radiocarbon dates from buried soil

horizons in unirrigated terraces, leading her to hypothesize that terrace construction began

in the valley sometime prior to 2400 BC. However, these dates, derived from pooled

charcoal in bulk soil samples, are also clearly out of sync with the regional culture-

historical chronology, since they fall within the Late Archaic Period (4800-3000 BC)a

time in which groups throughout the region were engaged in hunting and gathering

lifeways.

The systematic survey data and artifact collections from this project provide a basis

for clarifying the social organization and chronological placement of early agriculture in the

valley. Given the agglutinated, low-intensity pattern evident in these field systems, I expect

that political and economic units at the time of their construction and use would have been

small and minimally hierarchical, with low supra-household demands for labor and surplus

production. Such political and economic organization could have been supported by the

low productive potential and acephalous, aggregative construction of the system. This

pattern is consistent with Early Formative social organization in the Titicaca Basin (Stanish

2003:99-109), and I expect that early agricultural settlement in the valley would have

developed roughly coevally with this and other surrounding locales. Trade links with the

Titicaca Basin throughout the Formative are abundantly evident in the distribution of

obsidian from the Chivay sourcelocated just east of the survey area of this project

throughout the Lake Titicaca Basin, where the great majority of obsidian at Formative sites

was procured from the Chivay source (Brooks, et al. 1997; Burger, Asaro, Salas, et al.

1998; Burger, et al. 2000). So prior to this project, it appeared quite probable that a

22

significant Formative occupational component had yet to be identified. I discuss corollary

metrics in further detail below, but in general terms, the settlement pattern correlate of this

Formative Period social organization would be characterized by a dispersed settlement

pattern of small settlements of approximately equal size.

The Middle Horizon: Expansionist States and Their Local Influence

The influence or presence of Middle Horizon expansionist statesthe Wari and

Tiwanakuin the Colca valley also remains poorly understood. Local Middle Horizon

ceramics are clearly derivative of Wari, not Tiwanaku, regional styles, and no Tiwanaku

ceramics have been recovered in the valley (Brooks 1998; de la Vera Cruz Chvez 1989;

Malpass and de la Vera Cruz Chvez 1986, 1990; Neira Avendao 1961, 1990). One large

Middle Horizon settlement (Achachiwa) has been identified in the lower reaches of the

valley (de la Vera Cruz Chvez 1987, 1988, 1989, 1996), although it appears to represent a

local settlement, and not a Wari administrative center (Brooks 1998:87; Schreiber

1992:104). However, economic links with Tiwanaku are evident in the continued

predominance of obsidian from the Chivay source throughout sites in the Lake Titicaca

Basin, including the imperial center of Tiwanaku itself (Brooks, et al. 1997; Burger, Asaro,

Salas, et al. 1998; Burger, et al. 2000). Thus, the Colca appears to have constituted a

boundary zone between the Wari and Tiwanaku spheres of influence; the former perhaps

more political or ideological in nature, and the latter more economic in nature or based on

more ancient trade networks (Burger, Asaro, Salas, et al. 1998).

23

If this is the case, I suggest that local Middle Horizon settlement patterning could

have changed in either of two general ways. First, if Wari political influence was

penetrating and included a direct imperial occupation, there should be evidence for

significant expansion and change of the local settlement pattern from the previous

Formative occupation, including the development of a site size hierarchy, and/or the

establishment of a state administrative center, as indicated by diagnostic imperial

architectural traits and high quantities of elaborate imperial ceramics. By contrast, if Wari

political influence was more indirect or ephemeral, there should be more continuity in the

local settlement pattern, and less evidence for marked inter-site hierarchical organization,

and only indirect influence in local ceramic styles.

The Late Intermediate Period: Community Organization and Agro-Pastoral Production

Previous researchers have hypothesized that the stimulus for the initial construction

of irrigation canals and massive complexes of bench terraces in the Colca valley came from

a climatological shift from wetter (AD 760-1040) to drier (AD 1250-1310) than average

conditions during the Late Intermediate Period (Brooks 1998; Denevan 2001:185-205;

Treacy 1989:133-138). Leaving aside the technical difficulties of correlating these periods

with terrace construction phases, even if perfectly aligned, such correlations would not

signify a causal relationship, but instead would provide information regarding the limiting

parameters within which any agricultural production system must operate. The theoretical

framework developed above suggests that models of linear causality from ecology to

24

economy (and from economy to polity) do not account for the recursive effects between the

two or their relationship to political organization.

This project therefore recasts these factors in a broader frame by analyzing how they

were negotiated within a changing social, political and economic context that made

agricultural intensification advantageous, or that otherwise allowed supra-household leaders

to mobilize the labor needed for the construction, coordination, and maintenance of the

large-scale infrastructure of irrigated, terraced agriculture in the valley. While the local

social, political, and economic context during the LIP was poorly understood prior to this

project, I expected that the growth of the Collagua polity followed the post-Middle Horizon

trajectory of regional development common throughout the central and south central

Andean highlands. Thus, I expect continued growth in the overall scale of local settlement

(i.e., demographic expansion) from the Middle Horizon to the LIP, as well as evidence for

increased competition through time. In short, I hypothesize that Collagua political

organization was heterarchical in nature during the LIP; that is, composed of internally-

differentiated, supra-settlement communities whose ranks relative to one another were fluid

and variable. Inter-community relations probably shifted between coordination and

conflict.

I evaluate these hypotheses using several lines of evidence. First, I expect the

development of internally-differentiated communities to correlate with the development of

a marked site size hierarchy. However, in accord with a non-centralized, heterarchical form

of political organization, I expect that no single settlement would dominate the settlement

pattern in terms of size or centrality. Rank size analysis, in which each site is compared to

25

the others in terms of its absolute and rank-order size (from the largest to smallest),

provides a means for characterizing the site hierarchy (Falconer and Savage 1995; Haggett

1965:100-107; Johnson 1987; Paynter 1983). I suggest that a heterarchically-organized

polity will be reflected by a convex rank-size distribution, in which the largest sites do

not greatly differ in size. I discuss the implications of my rank size analysis in further detail

in Chapter 6. Also, as political relations oscillated between coordination and conflict, I

expect that most, if not all large settlements would be situated either in defensible locations

or near fortifications, similar to the nucleated hilltop settlement pattern common to many

LIP contexts (e.g., D'Altroy 1987; Hyslop 1976; Stanish 1997a). In Chapter 6, I also

analyze hydrological relationships between the long feeder canals that carry meltwater from

the surrounding peaks to evaluate how water apportionment was coordinated at a supra-

settlement, watershed-scale. Differences in the size and elaboration of domestic

architecture provide a good archaeological index for inequality between households, and

domestic architecture at most LIP and Late Horizon settlements in the Colca valley remains

very well-preserved. Previous researchers have noted the distinctive features that

characterize the Collagua architectural style (Brooks 1998; Guerra Santander and Aquize

Cceres 1996; Neira Avendao 1961), and my analysis in Chapter 6 provides a view of the

variability of size and elaboration within that style, which I interpret as reflecting

differences in the wealth and status of households within and between sites.

26

The Late Horizon: From Heterarchy to Hierarchy under Inka Rule

Prior to this project, the nature of Inka influence or presence in the Colca valley was

debated among archaeologists. Portable artifact media suggested considerable Inka

influence: local Late Horizon Collagua ceramics show clear Inkaic stylistic and formal

attributes (de la Vera Cruz Chvez 1988, 1989; Malpass and de la Vera Cruz Chvez 1986,

1990; Shea 1997), but architectural and settlement indices were more ambiguous. In

general, evidence for direct Inka administration was stronger in the lower reaches of the

valley near the village of Cabanaconde, where structures of Inka style cutstone masonry

were reported, and a probable small administrative center (the site of Kallimarka) had been

identified (de la Vera Cruz Chvez 1987, 1988). In the central portion of the valley, where

the survey area of this project is located, no centrally-planned administrative center with

Inka public architecture had been documented in previous reconnaissance, leading some to

suggest that the Inka never established a direct imperial occupation there (Brooks 1998).

Others, noting the overall paucity of pertinent data, made no definitive statements but

suggested that Inka rule may have been administered through local settlements and elites

(Malpass 1986, 1987; Neira Avendao 1961, 1990; Shea 1986b, 1987, 1997).

However, ethnohistorical studies suggested that Inka administration had a much

more profound impact on the political and economic organization of the Collagua province

than these preliminary archaeological indices seemed to suggest. Although the Collagua

are not mentioned in most of the standard chronicles, a description of the province by the

Spanish magistrate (corregidor) Juan de Ulloa Mogolln from 1586 describes how the

ayllus of the Collagua were arranged according to a formal administrative hierarchy based

27

on dualistic, ternary, and decimal principles (Ulloa Mogolln 1965 [1586]). Forty years

ago, Zuidema (1964:115-118) noted the close parallels between the names of Collagua

ayllus as recorded in this and other 16th and 17th century sources and the names of the

ceque lines that mapped the ritual activities of the royal ayllus (panacas) of Cuzco over

space and time (cf. Bauer 1998:35-37). Based on Ulloas description and the names of

ayllus recorded in the colonial visitas of the Collaguas, subsequent research presented the

Collagua province as an example par excellence of Inka social engineering, whereby local

ayllus were reordered in a nested hierarchy of minimal-, minor-, major-, and maximal-

ayllus according to imperial categories of prestige and administrative scale (Benavides

1989; Cock Carrasco 1976-77; Prssinen 1992:362-366; Pease G. Y. 1977; Rostworowski

de Diez Canseco 1983:121-123).

In this project I evaluate the hypothesis that the Inkas established a centralized,

hierarchical structure of imperial administration out of the previously heterarchically-

organized local ayllus through integrated archaeological and ethnohistorical analysis.

Archaeologically, the correlates of centralized, coordinated imperial rule should be visible

not only in portable artifacts, but also in the form of significant changes in settlement

patterning and organization, either through the establishment of a large administrative

center with elite and public architecture, or the expansion and reorganization of an extant

settlement. Thus, I would expect a corresponding expansion in the site size hierarchy,

composed of four tiers of site sizes, ranging from small hamlets, to larger villages, towns,

and centers. In terms of rank-size distribution, an integrated administrative hierarchy is

often associated with a log-normal distribution, in which the largest site is approximately

28

twice as large as the second largest site, three times as large as the third largest site, and so

on (Falconer and Savage 1995; Haggett 1965:100-107; Johnson 1987; Paynter 1983). By

contrast, if the Collagua maintained relative autonomy during the Late Horizon, there

should be overall continuity from the LIP settlement pattern and no other evidence for

major state investments in the form of an administrative center or imperial architecture.

My interpretation of the settlement pattern data from my survey, however, relies not

only on the analogical reasoning of such archaeological correlates, but is also informed by

historically-and spatially-contiguous socio-structural homologies that I reconstruct through

analysis of early colonial visitas. These visitas, from the years 1591, 1604, and 1615-1617,

record some of the most detailed demographic and land tenure data of any colonial censuses

in the Andes, affording an opportunity to reconstruct patterns of land use and exchange at

regional and local scales.

Previous analyses of the Collagua visitas have illustrated how the ayllus of the

Collaguas, and the authority of their leaders, were distributed over villages in far-flung

production zones, forming a classic example of a vertical archipelago of settlements

linked to the core population in the central Colca valley by ayllu affiliation, ranging from

pastoralist enclaves in the high altitude grasslands of the upper valley, to maize-production

enclaves in the neighboring, low-lying valleys to the south (Benavides 1989; Galdos

Rodrguez 1984; Pease G. Y. 1977). My regional-scale analysis of visita landholding and

livestock declarations in Chapter 8 reconstructs how supra-local networks of ayllu

affiliation and authority synchronized varied complementarity practices between

agriculturalists and pastoralists under Inka and early colonial rule. As I illustrate, these

29

distinct complementarity practices were the result not only of adaptive processes to

ecological imperatives, but also of status and community affiliation.

At the local scale, my analysis of landholding declarations in the visitas of 1591,

1604, and 1615-1617 reveals how households access to agricultural fieldholdings was

mediated by their ayllu affiliation. In the censuses, each of the fields declared by

households were located using toponyms. As I discuss in Chapter 4, toponyms here and

elsewhere in the Andes tend to be extremely historically durable; thus, I have reconstructed

the locations of visita landholding declarations by reference to their modern toponymic

counterparts. Since the households of the villages of Coporaque and Y