Site Interaction and Political Geography in the Upper Usumacinta Region during the Late Classic: A GIS Approach Site Interaction and Political Geography in the Upper Usumacinta Region during the Late Classic: A GIS Approach Armando Anaya Hernandez BAR International Series 994 2001 British Archaeological Reports arc published by John and Erica Hedges Ltd and by Archacopress This volume has been published by: John and Erica Hedges Ltd British Archaeological Reports 1 Longworth Road Oxford 0X2 6RA England Tel/Fax +44(0) 1865 511560 E-mail: [email protected] www.barhedges.com Enquiries regarding the submission of manuscripts for future publication may be sent to the above address BAR S994 Site Interaction and Political Geography in the Upper Usumacinta Region during the Late Classic: A GIS Approach © Armando Anaya Hernández 2001. Volume Editor: Erica Hedges Printed in England by Biddles Ltd ISBN 1 84171 195 0 All BAR titles available from: Hadrian Books 122 Banbury Road Oxford 0X2 7BP England The current BAR catalogue with details of all titles in print, prices and meajis of payment, is available free from Hadrism Books All volumes are distributed by Hadrism Book Ltd ABSTRACT The reconstruction of the political organisation of the Maya lowlands during the Classic period has been the subject of continuous testing and debating of explanatory models. These have ranged from the strong centralised state of the Old Empire to the ephemeral peer polity/fragmented states models. Except for Puleston (1974) who argued on behalf of approaching a regional study from the periphery, everyone else has focused his or her studies from the centres. Hence most archaeologists have readily adopted models derived from Central Place Theory and Nearest Neighbour Analysis. However in order to apply these models archaeologists had to draw the line between primary centres and subsidiary centres. Before the advent of the “cpigraphic revolution”, site hierarchy was determined by quantitative methods, i.e. number of monuments present, number of courtyards, volumetric analysis, etc. With the discovery of the nature of the emblem glyph the pursuit of a site hierarchy continued, but following a different path. Based on their presence or absence the emblem glyphs where used to differentiate primary centres from subsidiary centres. But what about hierarchy between those sites that did have an emblem glyph. In this context first Barthel (1968) and later Marcus (1973) argued for the existence of four regional capitals in the Maya world. Mathews (1991) after a more detailed analysis of emblem glyphs concluded that these denoted the existence of completely autonomous polities. Hence the debate continued between those who favoured a strong centralised state, and those who favoured a scheme of basically small autonomous states. Recently Grubc and Martin (1995, 1998) have come up with a "middle ground” model. Through their analysis of a series of glyphs that express political subordination they have proposed the existence of two hegemonic powers, Tikal and Calakmul. In this scheme virtually the rest of the Maya world revolved in a succession of political alliances. Grube and Martin’s discovery seems to have resolved the contradictions that the two previous models were unable to resolve. Aliphat (1996) in his doctoral dissertation reached the conclusion that the physical landscape of the Upper Usumacinta played a pivotal role in shaping the settlement pattern. He also advanced the idea that access to, and control of the principal communication routes was a priority to the kings of the region. I followed this line of research in the prc.scnt thesis, arguing that it was precisely this determinant factor in regional supremacy. In this work 1 apply a conjunctive approach drawing from two lines of research to offer a reconstruction of the political organisation of the Upper Usumacinta region: Geographic Information Systems (GIS) and the hieroglyphic inscriptions. In the first instance I begun by creating a Digital Terrain Model for the region in order to model movement across the landscape. This enabled me to reconstruct territorial size and communication routes between polities based on cost- surface analysis. Later, I used an approach similar to that used by Tobler and Wineburg (1977) who applied the Gravity Model as a predictive model for the location of a series of sites in pre-Hittite Anatolia. The Gravity Model works on the premise that interaction between centres will be more intense if these are located closer together, and will decrease as distance increases. The inscriptions contained in the study contain mention of various presently unallocated sites. Of these Sak Tz’i’, Hix Witz, Man, and the “Knot-Site” are mentioned by more than one of the regional centres. I used these references and quantified the degree of site interaction in order as a measure of distance that would help me locate the unidentified sites. The Gravity Model identified the most likely areas for the location of Sak Tz’i’, Hix Witz and the “Knot-Site”, the distances obtained for Man set this site outside the study region. The Dempster-Shafer logic, which was used in the construction of is a probability model, was also applied to the data set. The results of Dempster-Shafer support those obtained with the Gravity Model. With these results and the site interaction information available for the region I proceeded to create a series of maps that offer a reconstruction of the political organisation of the region. Instead of presenting these maps based on fixed time periods, the critical points in the history of the region were identified to obtain those maps. These results offer a more dynamic view on how the political contestants of the Upper Usumacinta shifted their alliances from one to another of the hegemonic powers, and how these changes were reflected in the political geography of the region. 1 ACKNOWLEDGMENTS Many people and institutions were directly involved in the research and subsequent writing of this work, family, friends, and colleagues were a constant source of inspiration and knowledge. First of all my sincere appreciation to my dear friend and mentor Peter Mathews; in Peter I have found not only an inexhaustible source of knowledge; I have also found refuge in my many times of despair. Jane and Dave Kelley have also been critical in the development of my academic and personal life, both have opened their home and their hearts to me making the longing for home much more bearable. Scott Raymond went well beyond his duties as the Head of the Archaeology Department at the University of Calgary and has always been there for me, to him my deepest gratitude. If I were asked to single out which were the turning points in my professional career I would refer to two persons. One I have already mentioned Peter Mathews. The other is Nigel Waters from the Geography Department at Calgary. Nigel introduced me to an exciting dimension of spatial analysis. He patiently walked me through the otherwise cumbersome paths of spatial analytical methods and CIS that had always eluded me. Stanley Guenter has also made a decisive impact in the development of this work. Like a yapping puppy he would storm into my office sharing with me his new and exciting discoveries in the decipherment of a Maya inscription. This led to long discussion sessions, which gradually gave shape to my Chapter Five. Thanks are also due to Dr. Duane Marble, who kindly provided me with an exhaustive bibliography on movement over natural terrain, and his own paper on the same topic. These proved to be invaluable for the construction of my working model. In Mexico two persons also represent landmark moments in my professional life, Roberto Garcia Moll and Mario Aliphat Fernandez. Roberto drew me into the Maya area trusting me with his project at Bonampak. Mario honoured me with his friendship during the good times and especially during the bad times. It was then when he invited me to participate in the Proyecto Arqueoldgico El Cayo, and that in turn opened the doors to my postgraduate studies at Calgary. My recognition to my Mexican colleagues, Alejandro Tovali'n, a superb person and archaeologist, who unselfishly has shared much of his data from Bonampak with me and has provided his generous hospitality during my visits to his camp. By the same token Arnoldo Gonzalez, Daniel Juarez and Akira Kaneko have always given me proof of their generosity. On the institutional level my gratitude goes firstly to Mexican Council of Science and Technology CONACYT, which supported me through my studies at the University of Calgary. In Canada I also received the generous support of the Archaeology Department, the McMahon Stadium Society and the Dean of Graduate Studies Office. Finally in Australia I have received the overwhelming support of La Trobe University through the generous intervention of Professor Tim Murray. Without these institutions I could not have accomplished this work. 1 leave my family for the last because they fall in a completely different category. The level of their support is of course immeasurable. My lovely wife Yanira has been the driving force behind me throughout these years. She has suffered in silence the rigors that the life of a postgraduate student imposes. To my children Axel, Ixchel, Raul, and Max my love and my apologies for not being there all of the time. To my parents my endless love and admiration; throughout my life they have always encouraged and supported me in chasing my dreams. I thank my brothers Adolfo and Mario and my sister Leticia who through their example have shown me the value of family bonds; they have always been supportive and have made my existence more cheerful with their joyous view of life. II DEDICATION To Ihe three women of my life: my mother, my daughter, and my wife, without you I would be lost. lit TABLE OF CONTENTS ABSTRACT i ACKNOWLEDGEMENTS ii DEDICATION iii TABLE OF CONTENTS iv LIST OF TABLES v LIST OF FIGURES v LIST OF GRAPHS vi ONE SPACE THE FINAL FRONTIER 1. Introduction 1 2. Spatial Analysis and Archaeology 3 3. Central Place Theory 4 4. Nearest Neighbour Analysis 6 5. The Gravity Model 6 TWO SETTLEMENT PATTERNS IN THE MAYA LOWLANDS 1. Introduction 7 2. Spatial Models 8 3. The Epigraphic Approach 11 THREE LANDSCAPE ARCHAEOLOGY AND COMMUNICATION ROUTES 1. Introduction 17 2. Description of the Physical Environment 17 3. Traversing the Region, the Historic Accounts 20 FOUR GEOGRAPHIC INFORMATION SYSTEMS AND ARCHAEOLOGY PART I THE CONCEPTS 1. Introduction 28 2. Digital Terrain Models 28 3. Moving Over Natural Terrain 30 4. Archaeological Applications 33 5. Isotropic and Anisotropic Frictions 35 6. Predictive Modelling 36 PART II METHODOLOGY 1. Digital Terrain Model and Cost Surfaces 36 2. Modelling Territorial Extent 38 3. The Gravity Model 41 4. Predictive Modelling and Dempster-Shafer Theory 50 FIVE SITE INTERACTION AND POLITICAL GEOGRAPHY PART I: THE SOCIO-HISTORICAL LANDSCAPE 1. Introduction 56 2. Ancient Roads and Polities 56 iv PART II; A RECONSTRUCTION OF THE POLITICAL HISTORY OF THE REGION 1. The Early Period: 8.19.0.0.0 to 9.5.2.10.6 (AD 416-537) 59 2. The Early Palenque Wars: 9.8.5.13.8 to 9.9.11.12.3 (AD 599-624) 62 3. The SakTz’i’Years: 9.10.8.3.5 to 9.11.16.11.6 (AD 641 -669) 65 4. The Calakmul Alliance Wavers: 9.11.16.11.6 to 9.14.10.5.0 (AD669-772) 69 5. The Yaxchilan Expansion: 9.15.19.2.2 to 9.16.8.3.18 (AD 750-759) 75 6. The Reclaiming of the Western Frontier: 9.16.12.2.6 to 9.17.1.5.9 (AD 763-772) 77 7. The Final Years: 9.18.6.4.1910 10.1.14.9.17 (AD 796-864) 80 CHAPTER SIX; CONCLUSIONS 1. Site Interaction and Settlement Pattern 82 2. Spatial Models in Archaeology 84 APPENDIX 1: CARTOGRAPHIC MODELS 86 APPENDIX 2: SUMMARY STATISTICS 91 APPENDIX 3: SITE INTERACTION 93 BIBLIOGRAPHY 110 LIST OF TABLES TABLE 4-1 Predicted treadmill and relative ratios of energy costs 32 TABLE 4-2 Terrain coefficients 32 TABLE 4-3 Mean energy expenditure for slow walking with all three external loads 32 TABLE 4-4 Caloric requirements of overland bearers 34 TABLE 4-5 Estimated transport costs for prehispanic Mesoamerica 34 TABLE 4-6 Caloric costs per kilometre of walking 35 TABLE 4-7 Primary centres and their subsidiary sites 39 TABLE 4-8 Territorial extent of primary centres 41 TABLE 4-9 Numerical values for site interaction 42 TABLE 4-10 List of sites included in the initial data matrix 43 LIST OF FIGURES Figure 1-1 Christaller’s patterns of hierarchical distribution of central places 5 Figure 1-2 Point distribution patterns for Nearest Neighbour Analysis 6 Figure 2-1 Regional capitals of the Maya world, Copan Stela A/Seibal Stela 10 12 Figure 2-2 Components of the emblem glyph 13 Figure 2-3 Identification of autonomous polities based on presence of emblem glyph 13 Figure 2-4 Titles that imply political subordination 15 Figure 2-5 Expressions that imply political subordination 16 Figure 3-1 Map of the Western Maya Lowlands indicating the Upper Usumacinta 18 V Figure 3-2 Archaeological sites included in the region of study 18 Figure 3-3 Alonso Davila’s entrada to the region, 1530 22 Figure 3-4 Maudslay’s and Charnay’s routes, 1882 24 Figure 3-5 Maler’s entries into the Upper Usumaeinta 25 Figure 4-1 DEM of the Upper Usumaeinta region 37 Figure 4-2 Initial territorial extents estimated with ALLOCATE 38 Figure 4-3 Second stage: allocation of territory per site 40 Figure 4-4 Primary centres territorial extent re-adjusted 41 Figure 4-5 Gravity Model results for Sak Tz’i’ 44 Figure 4-6 Gravity Model results for Hix Witz 45 Figure 4-7 Gravity Model results for the “Knot Site” 46 Figure 4-8 Gravity Model results for Man 47 Figure 4-9 Gravity Model results for (a) Yaxchilan and (b) Piedras Negras 48 Figure 4-10 Proposed optimal territorial extents for the Upper Usumaeinta polities: (a) Initial extent with secondary sites overlaid, (b) Territorial extent After reclassification 49 Fugue 4-11 Belief results for the site hypothesis 52 Figure 4-12 Plausibility of the site hypothesis 53 Figure 4-13 Belief interval 54 Figure 4-14 Belief (a) and Plausibility (b) results for the non-site hypothesis 54 Figure 4-15 Dempster-Shafer & Gravity Model results 55 Figure 5-1 Potential communication routes in the Upper Usumaeinta region 59 Figure 5-2 Royal visits to Yaxchilan 60 Figure 5-3 Tab Balam I of Yaxchilan appearing as captive at Piedras Negras 61 Figure 5-4 Suggested territorial extents of the Upper Usumaeinta kingdoms at AD 416-537 62 Figure 5-5 Itzam Balam of Yaxchilan oversees accession of Chan Muwan I of Bonampak 63 Figure 5-6 Ruler 1 of Piedras Negras stands over his Palenque and Sak Tz’i’captives 64 Figure 5-7 Suggested territorial extents of the Upper Usumaeinta kingdoms at AD 599-624 65 Figure 5-8 Account of the Sak Tz’i’ wars on the Denver and Brussels Panels 66 Figure 5-9 Lacanha, Bonampak, and Yaxchilan lords paying homage to Piedras Negras Ruler 2 and his heir 67 Figure 5-10 Proposed escape route of Tikal’s king Nu-Bak-Chak, passing through Sak Tz’i’s territory at AD 659 68 Figure 5-11 Suggested territorial extents of the Upper Usumaeinta kingdoms at AD 641-669 69 Figure 5-12 The capture of Ah-Nik from Man by Itzam Balam of Yaxchilan 70 Figure 5-13 Tonina’s intervention in the Upper Usumaeinta 71 Figure 5-14 Sak Tz’i’captive displayed at Tonina 72 Figure 5-15 Sak Tz’i’s control over Lacanha and Bonampak 73 Figure 5-16 Suggested territorial extents for the Upper Usumaeinta kingdoms at AD 669-722 74 Figure 5-17 Yaxun Balam is presented with four captives 76 Figure 5-18 Suggested territorial extents at AD 750-759 76 Figure 5-19 Capture of SakTz’i’ lords by Chel Te of Yaxchilan (a) and Chan Muwan II of Bonampak (b) 78 Figure 5-20 Pomona captives at Piedras Negras 79 Figure 5-21 Suggested territorial extents at AD 763-772 80 Figure 5-22 The Randall Stela, Sak Tz’i’s latest known monument 81 LIST OF GRAPHS Graph 4-1 Energy expenditure values for walking in different types of terrain 33 Graph 4-2 Linear regression between expected energy expenditure values 33 VI