UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A mortuary analysis of the Dawenkou Cemetery Site, Shandong, China Kingscott, Anne Underhill 1983-04-19

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-UBC_1983_A8 K56_4.pdf [ 15.71MB ]
Metadata
JSON: 831-1.0058396.json
JSON-LD: 831-1.0058396-ld.json
RDF/XML (Pretty): 831-1.0058396-rdf.xml
RDF/JSON: 831-1.0058396-rdf.json
Turtle: 831-1.0058396-turtle.txt
N-Triples: 831-1.0058396-rdf-ntriples.txt
Original Record: 831-1.0058396-source.json
Full Text
831-1.0058396-fulltext.txt
Citation
831-1.0058396.ris

Full Text

A MORTUARY ANALYSIS OF THE DAWENKOU CEMETERY SITE, SHANDONG, CHINA by ANNE UNDERHILL KINGSCOTT B.A., Duke University, 1977 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Department of Anthropology and Sociology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1983 Anne Underhill Kingscott In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^/ftj/W^-L*^ CT/VJL ^rd^rtry^ The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date E-6 (3/81) ABSTRACT This study is concerned with the development of social ranking in Shandong province, China, and its environs, during the late Neolithic period. The Dawenkou cemetery has been considered by Chinese and western archaeologists as represent ing one of the earliest ranked societies in this region. There is some disagreement regarding the degree of status differentia tion reflected by the mortuary remains at Dawenkou. Opinions vary as to whether the site shows incipient ranking, or a more fully developed system, or whether there was a hereditary ruling class. The primary goal of this study is to provide a greater understanding of the nature and degree of status differentia tion represented at Dawenkou, by means of an in-depth mortuary analysis utilizing current archaeological methods. The methodology upon which this study is based is outlined in Chapter 2. Four analyses are included: an evaluation of the three relative chronological periods at the site (Chapter 3), an analysis to estimate the sex of unsexed burials on the basis of grave goods (Chapter 4), an exploratory assessment of social subgroup affiliation (Chapter 5), and the analysis of status differentiation (Chapter 6). I conclude that an increase through time in the degree of ranking is represented at the site. The Early period burials reflect elements of both an achieved and an ascribed (ranked) system of status differentiation. The Late period burials appear to reflect a highly developed ranked society. I propose that the Early period burials represent part of a social system in which members of a descent group were ranked. Also, the Late period burials represent members of a descent group that constitutes one status level in a regional status system. Mortuary analyses of other Dawenkou Culture sites roughly contemporaneous to the Early and Late periods at Dawenkou could reveal whether similar changes through time are apparent. It is likely that ranking first developed in the eastern seaboard region at an earlier date than previously considered. The secondary goal of this study is to make a methodo logical contribution to mortuary analysis. I argue that status may be symbolized by energy expenditure, grave goods, or both. Since the burials in a cemetery reflect more than one social system through time, an analysis of status should emphasize change through time. Finally, some of the techniques employed in this study should have utility for other mortuary analyses: the Simple Inspection Method to estimate sex (Chapter 4), and Ward's Method of cluster analysis and Torgerson's Metric Multidimensional Scaling for an analysis of status (Chapter 6), or for seriation (Chapter 3). iv TABLE OF CONTENTS Page ABSTRACT . ii TABLE OF CONTENTS iv LIST OF FIGURES vii LIST OF APPENDICES x ACKNOWLEDGEMENTS xii CHAPTER 1 Introduction 1 1.1. The Dawenkou Site and the Research Problem 1 1.2. Classification and Description of Dawenkou Culture 5 1.3. Cultural Evolution in the Late Neolithic Period 14 CHAPTER 2 Methodology 17 2.1. General Approaches in Mortuary Analysis2.2. The Processual Approach 19 2.3. The Symbolist Approach 36 2.4. The Approach Followed in this Mortuary Analysis 43 2.5. Method to Determine Status Levels . . 4 8 2.6. Consideration of Natural and Cultural Factors that may have Affected the Archaeological Record at Dawenkou . . 54 CHAPTER 3 Chronological Analysis 58 3.1. Introduction . 5 8 3.2. The Methods to Derive the Chronological Periods 61 3.3. Analysis 65 3.3.1. Introduction 63.3.2. Method  7 3.3.3. Data 74 3.3.4. Results 8 3.3.5. Multidimensional Scaling 79 3.3.6. Cluster Analyses . 84 3.4. Conclusion 8V Page CHAPTER 4 Analysis to Estimate Sex 95 4.1. The Problem 94.2. The Analysis 9 4.2.1. Introduction4.2.2. Method . . . 100 4.2.3. The Simple Inspection Method: Data 104 4.2.4. The Simple Inspection Method: Results . . . ' 107 4.2.5. Discriminant Analysis: Data . . . Ill 4.2.6. Discriminant Analysis: Results . 113 4.3. Conclusion 119 CHAPTER 5 The Nature of Social Subgroup Affiliation at Dawenkou 126 5.1. Method 125.2. Spatial Location of Graves 130 5.3. Correlation of Grave Orientation, Grave Form, and Body Disposition with Grave Location 131 5.4. Correlation of Ceramic Style with Spatial Location 14 3 5.5. Interpretation of the Potential Social Subgroup Affiliation -Related Variables 145 5.5.1. Introduction 145.5.2. Orientation5.5.3. Grave Form 149 5.5.4. Body Disposition . . . 151 5.6. Concluding Propositions . 153 5.6.1. Argument for a Descent Group at Dawenkou 15 3 5.6.2. Kinship or Residential .Groups at Dawenkou 156 5.6.3. The Mortuary Population at Dawenkou .157 CHAPTER 6 Analysis of Status Differentiation . . 160 6.1. Introduction 166.1.1. Procedure6.1.2. Test Implications 161 6.2. Analysis of Status, Early period . . 163 6.2.1. Data 166.2.1.1. The Multivariate Analyses ... 163 6.2.1.2. High Status Unique Items ... 169 6.2.1.3. Energy Expenditure 172 6.2.2. Results: The Multidimensional Scaling 173 vi Page CHAPTER 6 (cont'd) 6.2.3. The Three Types of Cluster Analysis 176 6.2.4. Interpretation of Status Distinc tions in the Early Period ..... 181 6.2.4.1. Distribution of the Unique Items 181 6.2.4.2. Argument for Four Status Levels 182 6.2.4.3. Differentiation by Energy Expenditure . 183 6.2.4.4. Body Disposition 188 6.2.4.5. Grave Location6.2.4.6. Test Implications for Achieved versus Ascribed Status 191 6.2.4.7. Multiple Burials 193 6.2.4.8. Conclusions . 195 6.3. Analysis of Status, Late Period ... 198 6.3.1. Data 196.3.1.1. The Multivariate Analyses ... 198 6.3.1.2. Energy Expenditure 201 6.3.1.3. High Status Unique Items .... 202 6.3.2. Results: The Multidimensional Scaling 203 6.3.3. The Three Types of Cluster Analysis 205 6.3.4. Interpretation of Status Distinc tions in the Late Period and Assessment of Change Through Time . 209 6.3.4.1. Distribution of the Unique Items 209 6.3.4.2. Argument for Three or Two Status Levels 210 6.3.4.3. Differentiation by Energy Expenditure . . . 216 6.3.4.4. Body Disposition and Grave Location . 217 6.3.4.5. Test Implications 219 6.3.4.6. Conclusions 222 6.4. Implications 227 CHAPTER 7 Conclusions and Suggestions for Future Research ... 229 7.1. Conclusions Regarding Dawenkou . . . 229 7.2. Future Research 233 7.3. Methodological Conclusions 238 BIBLIOGRAPHY ..... 241 APPENDICES 25GLOSSARY 318 vii LIST OF FIGURES Figure Page 1-1 The location of Dawenkou and other Neolithic sites 7 1-2 Rough chronological relationship of the Dawenkou Culture with other Neolithic cultures of China 8 1- 3 Rough chronological order of Neolithic sites in the eastern seaboard region of China 9 2- 1 Number of dimensional distinctions of the social persona according to form of subsistence 24 2-2 Average number of dimensional distinc tions according to subsistence cate gory 2 4 2- 3 Aspects of the social persona symbolized in mortuary treatment 26 3- 1 The 115 burials in the chronological analysis from Torgerson's Metric Multi dimensional Scaling, with clusters from Ward's Method 81 3-2 Dendrogram from Ward's Method of cluster analysis 86 3-3 List of burials in the newly assigned Early and Late periods derived from the chronological analysis, as well as undatable burials • . . 90 3- 4 Location of burials within the cemetery from each period as defined by the chronological analysis 91 4- 1 Distribution of sex-linked artifact types 102 4-2 Distance values for the unsexed burials from the Simple Inspection Method and assignment of sex, and the distance values for known sexed burials .... 109 viii Figure Page 4-3 Standardized Discriminant Function Coefficients for the 2 7 variables that distinguish between known male and female classes of burials .... 116 4-4 Discriminant scores for the 82 unsexed, single adult burials in the Discrimin ant Analysis 117 4-5 Known sexed and estimated sexed burials from the Simple Inspection Method and Discriminant Analysis ... 121 4- 6 Spatial location of the known sexed and estimated sexed burials in the cemetery 122 5- 1 Spatial groups of burials derived by visual inspection, Early period . . . 132 5-2 Spatial groups of burials derived by visual inspection, Late period .... 133 5-3 The range of grave orientations for Early, Late and undatable burials . . 135 5-4 Correlation of grave form, grave orientation, body disposition, sex and age with spatial location of grave, Early period ...... 136 5-5 Correlation of grave form, grave orientation, body disposition, sex and age with spatial location of grave, Late period 141 5- 6 Correlation of grave form, grave orientation, body disposition, sex and age with spatial location of grave, undatable burials 142 6- 1 Multidimensional scaling plot of Early period burials in the analysis of status, with Ward's Method clusters . . 175 6-2 Dendrogram of Early period burials from Ward's Method ..... 178 6-3 Distinguishing characteristics of the four status levels, Early period . . . 184 ix Figure Page 6-4 Age and sex composition of the status levels in the Early period 189 6-5 Location of the four status groups in the Early period 190 6-6 Multidimensional scaling plot of Late period burials in the analysis of status, with Ward's Method clusters . . . 204 6-7 Dendrogram of Late period burials from Ward's Method 207 6-8 Distinguishing characteristics of the three status levels, Late period .... 213 6-9 Age and sex composition of the three status levels in the Late period, and data regarding the one multiple burial in the Late period 218 6-10 Location of the three status groups in the cemetery, Late period 220 X LIST OF APPENDICES A3-1 The functional types, subtypes, and styles of pottery at the Dawenkou site 258 A3-2 The 15 pairs of intrusive burials 261 A3-3 Burials included in the chronological analysis 262 A3-4 The 83 ceramic forms in the chronological analysis . 263 A3-5 Distribution of ceramic categories among the burial sample 2 65 A3-6 Frequency of occurrence of ceramic categories after lumping 266 A3-7 Distribution of ceramic categories after lumping 267 A3-8 Composition of clusters from Ward's dendro gram in terms of ceramic styles 268 A4-1 List of sexed and unsexed burials : Early, Late, undatable 2 71 A4-2 The 129 artifact types in the Discriminant Analysis 272 A5-1 Orientation of graves, Early period . . . 274 A5-2 Orientation of graves, Late period . . . 275 A5-3 Distribution of ceramic styles, Early period 276 A5-4 Distribution of ceramic styles, Late period 281 A6-1 The 79 burials in the Early period analysis of status, with data on age and sex, grave form, grave size 285 A6-2 The 2 3 variables and their attributes in the Early period analysis of status ..... 287 A6-3 Distribution of serving stands, Early period 289 A6-4 Distribution of pottery vessels, Early period 290 A6-5 Distribution of stone tools, Early period . . 291 A6-6 Distribution of bone tools, Early period . . 292 A6-7 Distribution of pig skulls, Early period . . 293 xi Figure Page A6-8 Distribution of deer teeth, Early period . . 294 A6-9 Distribution of raw material pieces, Early period 295 A6-10 Grave area of Early period burials .... 296 A6-11 The mortuary attributes of the seven clusters from Ward's Method, Early period burials . .29 7 A6-12 Multiple burials from the Early period : age, sex, grave goods, energy expenditure . 302 A6-13 The 32 burials in the Late period analysis of status, with data on age and sex, grave form, grave size 303 A6-14 The 2 8 variables and their attributes in the Late period analysis of status 304 A6-15 Distribution of serving stands, Late period 306 A6-16 Distribution of tall stemmed cup, Late ' period 307 A6-17 Distribution of pottery vessels, Late period 308 A6-18 Distribution of stone tools, Late period . 309 A6-19 Distribution of bone tools, Late period . . 310 A6-20 Distribution of deer teeth, Late period . . 311 A6-21 Distribution of raw material pieces, Late period 312 A6-22 Grave area of Late period burials .... 313 A6-2 3 The mortuary attributes of the eight clusters from Ward's Method, Late period burials 314 A6-24 Data regarding the undatable burials : age, sex, grave goods, energy expenditure . . 317 xii ACKNOWLEDGEMENTS I am extremely grateful to several individuals for their valuable comments, encouragement and support throughout the duration of this study. I feel very fortunate to have had Dr. Richard Pearson, Dr. R.G. Matson, and Dr. David Pokotylo on my M.A. Committee. My advisor, Dr. Pearson, gave me much help and encouragement throughout the three years of my M.A. program. He generously shared his data, published and unpub lished papers, and ideas about the Dawenkou site and the Chinese Neolithic period. His knowledge and enthusiasm for East Asian archaeology have been an inspiration to me. Dr. Matson was very generous with his time regarding computer analyses and the interpretation of results. Dr. Pokotylo gave me detailed comments for every section of the study. I also thank the Department of Anthropology and Sociology for computer funds. Michael Blake, visiting professor at the Department of Anthropology and Sociology during 1983, provided valuable in sight into the methodology of mortuary analysis. Dr. Neil Guppy calmed my panic at the computer terminal on a few occa sions. Kian Kwok generously helped with translation problems. I am grateful to Zou Heng of Beida University, Beijing, and Gao Guangren of the Institute of Archaeology, Chinese Academy of Social Sciences, for their information. Clarence Shangraw of the Asian Art Museum, San Francisco, was extremely generous with his time and help. During his visit to this Department xiii in 1983, Dr. Lewis Binford gave me helpful comments on method ology. I could not have produced this thesis without the help of Moira Irvine, Pat Berringer, Julie Mandziuk and Rick Clements. I greatly appreciate the time Moira spent drawing several figures. Julie and Rick typed four chapters after working all day. Pat Berringer typed the remainder of the thesis and gave me valuable advice, and provided much support and encouragement. She is one of the most generous individuals I have ever met. Dana Lepofsky and Deanna Ludowicz were also very supportive. I especially thank Michael for his humor and loving support. My parents, as always, gave me much encouragement. Finally, I thank Dr. Michael Hammond for his support during my years at Duke University, and for stimulating my interest in the development of prehistoric, complex societies. - 1 -CHAPTER 1 INTRODUCTION 1.1. The Dawenkou Site and the Research Problem The Dawenkou site is a late Neolithic period site located in south-central Shandong province. The site was discovered in 1959 during the construction of a railway track near the town of Dawen on the Dawen River (Gao 1978:31). The first excava tion of the site took place in 1959 and the second in 1974. The 133 burials discovered in the first excavation are the subject of this mortuary analysis. A detailed description of these burials is provided in a site report published in 1974 by the Shandong Provincial Cultural Properties Commission and the Jinan City Museum (Dawenkou: Report of the Excavation of  a Neolithic Cemetery. Peking: Wen Wu Press). The burials from the 1959 excavation are regarded as critical for understanding the development of social ranking during the Neolithic period in the eastern seaboard region. Both western and Chinese archaeologists maintain that the Dawenkou cemetery represents a ranked society, but they disagree over the degree and nature of status differentiation represented at the site. Some researchers maintain that the site reflects incipient ranking and others, a highly developed system of ranking. The primary goal of this study is to provide a greater understanding of the site by means of an in-depth mortuary analysis utilizing current archaeological methods. My method-- 2 -ology and the outline of this study are presented in Chapter 2. The secondary goal of this study is to make a methodological contribution to mortuary analysis, in terms of overall approach and the specific techniques that have been employed to identify status distinctions in a cemetery. The authors of the Dawenkou site report classify the 13 3 burials into three periods: Early (75 burials), Middle (19 burials), and Late (25 burials). Fifteen burials are regarded as undatable. Samples for radiocarbon dates were not taken during the 1959 excavation. A recent article presents a revision of this relative chronology, stating there are two periods represented at the site: Early and Late. The Early period is said to represent approximately 600 years (from 3400 - 2800 B.C.) and the Late period, 400 years (2800 - 2400 B.C.) (Wu Ruzuo^ 1982: 268"). The method by which these dates were derived is not clear. The burials from the 1974 excavation represent an earlier time period than those excavated in 1959. Apparently, the report for the second excavation in 19 74 has not been completed and the exact number of burials found is not known (Gao, personal communciation, 1983). Two radiocarbon dates from material found during the 1974 excavations yielded the following recalibrated dates: 1) ZK 469(T10,4; lower H24) : 6155 + 140 (4205 B.C.) and 2) ZK 468(T12,2B; lower H3) : 6210 + 135 (4260 B.C.) (Wu Ren 1982:55). The Dawenkou site also has some Longshan period and Han Dynasty remains (Shandong Provincial Cultural Properties Commission and the Jinan City - 3 -Museum 1974: plate III). In The Archaeology of Ancient China, K. C. Chang (1979:160) states that Dawenkou and other Dawenkou Culture sites represent an incipient system of ranking. More recently, Chang (19 80:361, 1983a:513-514) states it is not clear if the Dawenkou Culture represents a level of cultural development which he calls "intervillage aggregates" (characterized by evidence for rich and poor groups, violence, crafts specialization, religious and village leaders such as chiefs, and intervillage leagues). He believes that Longshan sites of Shaanxi, Henan, and Shandong had reached this level of cultural complexity. Chinese archaeologists view socio-cultural evolution in terms of Marxist historical philosophy and disagree whether Dawenkou and other Dawenkou Culture sites reflect the begin ning of slave society (Chang 1983b:575). However, articles by the Wen Wu Correspondent (19 78), the Shandong Provincial Museum (1978), the Kao Gu Editorial Staff (1979), Zhang (1979), Gao (1978), Wei (1976), Luo and Zhang (1979), Wu (1973), the Nanjing Museum (1978) imply that ranking is fully developed at Dawenkou, especially by the Late period. An increase in the differentiation of burials in terms of grave goods from the Early period to the Late is noted by the Shandong Provincial Museum (1978) and implied by the Kao Gu Editorial Staff (1979). Luo and Zhang (19 79) note a concentration of high status burials from both periods in the northern sector of the site. Pearson's (1981) position on the degree of ranking at - 4 -Dawenkou is intermediate to that of Chang (19 79) and the Chinese archaeological literature cited above. Pearson (1981:1086) proposes that the presence of stone beads and stone and jade ornaments in both male and female graves indicates a system of ranking. Also, the wealth evident from some graves suggests the presence of a chiefly office. The lack of a complete segregation of wealthy graves may indicate the absence of a hereditary ruling class. Pearson's (19 81) comparative study of status differentia tion among several sites in the eastern seaboard region is the only study utilizing current archaeological methods of mortuary analysis to understand the nature and degree of status differ entiation at Dawenkou Culture sites. My study is an outgrowth of Pearson's (1981) study. I hope the application of current methods of mortuary analysis on the Dawenkou data can provide a better understanding of the nature and degree of status differ entiation at the site, as well as to resolve the controversy in the degree of ranking noted above. This study includes an evaluation of several trends through time in status differentiation that are concluded by Pearson (19 81). Pearson notes an increase in the mean number of ceramics in burials (1981:1081, Table 2, and 1082) as well as an increase in the percentage of burials with ceramics (1981: 10 85). Also, an increasing variation in the number of ceramics found in each burial through time is suggested (19 81:10 82,1083, Table 4). For the Dawenkou site there is also increasing - 5 -variation in quantities of tools and ornaments within burials (19 81:10 85). An increase in the quantities of ceramics in male burials over female burials through time is noted for some of the late Dawenkou Culture sites, but this increase is not apparent in the Dawenkou site burials due to the small number of sexed burials (1981:1083). A trend towards spatial segrega tion of high status burials by the Late period at Dawenkou is proposed (1981:1084) but is considered tentative until more assessment has been made (1981:1085). Pearson (1981:1086) concludes that by the Late Dawenkou Culture period at approx imately 2 000 B.C. (which includes the Late period at Dawenkou site), evidence of pronounced ranking is not present, nor is occupational specialization. He also proposes that the status of men in terms of power and wealth was increasing by the Late Dawenkou Cultural period and craft specialization was develop ing (ibid). 1.2. Classification and Description of Dawenkou Culture  Dawenkou is regarded in the Chinese archaeological litera ture as part of the Dawenkou Culture, which spans a period of approximately 2000 years (An 1979-80:38, 1981:258, 1982:58). Dawenkou Culture sites, the great majority of which are cemeteries, are located in the lower Yellow River valley, primarily in Shandong and Jiangsu (An 1979-80:38). Other Dawenkou Culture sites have been found in northern Anhui, eastern Henan and the Liaodong Peninsula (ibid). The location of Dawenkou and other Neolithic sites in the eastern seaboard - 6 -region is indicated in Fig. 1-1 (after Pearson 1981:1079). Only the sites in the northern areas of the map (northern Jiangsu and Shandong) are considered as part of the Dawenkou Culture. I am not aware of a published source that lists the total number of Dawenkou Culture sites that have been discovered, as well as the number of burials from each site and the geographic location of each site. Over 6,000 Neolithic sites have been discovered in China during the past thirty years (An 1979-80:35). The Dawenkou Culture is considered in the Chinese archaeo logical literature to exist from approximately 4000 B.C. to 2000B.C. Fig 1-2 (from An 1982:58) depicts the rough chrono logical relationship of Dawenkou Culture with other Neolithic cultures of China. Radiocarbon dates from individual Dawenkou Culture sites are discussed by Xia (1979), Wen Wu Correspondent (1978), Wu Ren (1982), Gao (1980), and Pearson (in press:7-8). The dating of the culture in general is discussed by An (1979-80). He explains (1979-80:38) that twelve radiocarbon dates have been taken from Dawenkou Culture sites. Two, ZK 90 and ZK 479, show a range of 4494 - 2690 B.C. (calibrated). Four dates (ZK 317, 319-0, 321, and 361-0) give a range of 2350 - 1905 B.C. (calibrated). An (1979-80:38) believes these late dates may be explained by difficulties in distinguishing Dawenkou Culture sites from Longshan Culture sites. The rough chronological order of sites in the eastern seaboard region (from Pearson 1981:1081) is indicated in Fig. 1-3. There is an ongoing controversy regarding the cultural - 7 -FIGURE 1-1. The location of Dawenkou and other Neolithic sites in the eastern seaboard region of China. (from PearBon 1981:1079) - 8 -FIGURE 1-2. Rough chronological relationship of the Dawenkou Culture with other Neolithic cultures of China. Chronology of the Neolithic Period in the Huanghe (Yellow) and Changjiang (Yangtze) River Valleys upper reaches of the Huanghe middle reaches of the Huanghe lower reaches of the Huanghe lower reaches of the Changjiang middle reaches of the Changjiang 1000 B.C. B A r S 8 Siba Culture Qijia Culture Shang Shang Shang Shang Longshan Culture Longshan Culture Majiayao Culture Liangzhu Culture Dawenkou Culture Majiabang Culture Longshan Culture ' 7$uj fating . —Culture | Daxi Culture Yangshao Culture lYangshao Culture 2000 B.C. 3000 B.C. I Dadiwan Culture Peiligang Culture I I Qlngliangang Culture Cishan Culture I U000 B.C. 5000 B.C. 6000B.C. (from An Zhimin 1982:58) - 9 -FIGURE 1-3. Rough chronological order of Neolithic sites in the eastern seaboard region (adapted from Pearson 1981: 1081). Dafanzhuang Xixiahou Dawenkou Dawenkou Dawenkou Dadunzi Songze Liulin Liulin Liulin Beiyinyangying Maj iabin Yuduncun Yuduncun Yuduncun Late Middle Early Huating layer Liulin layer Dawenkou Culture site Dawenkou Culture site Dawenkou Culture site Dawenkou Culture site excavation 1 Dawenkou excavation 2, upper layer Culture excavation 2, lower layer site Layer IB Layer 2 Layer 3 I - 10 -classification of the Neolithic sites from the eastern seaboard region. K.C. Chang (1979:136) maintains that the culture prior to Dawenkou, the Qingliangang (see Fig. 1-2), should be divided into a type including sites north of the Yangtze River and a type including sites south of the Yangtze. Chang (1979:138) also considers sites designated as Early Dawenkou Culture by Chinese archaeologists as part of the Qingliangang period. Chang (1979:144, 154-55) classifies Dawenkou Culture sites after approximately 3200 B.C. as part of the Longshanoid cultural complex, which includes Neolithic sites from the North Central Plain area. The Dawenkou site is considered as part of the Huating Culture, along with the Liulin and Huating sites (Chang 1979:160). A multivariate statistical test of the classification of Longshanoid cultures including Dawenkou by Lo (1977) resulted in a replication of the culture classification scheme in most cases. Huber's (1981:118, 1983:202) classification scheme is based partially upon ceramic development. She concludes that sites in northern Jiangsu and Shandong are part of the same cultural complex, but gives a different chronological sequence of sites from these provinces. The scheme utilized by Pearson (1981:1078, 1080) and Shangraw (1978:12) consists of the Qingliangang Culture (approximately 5000 - 3000 B.C.), includ ing sites in northern Zhejiang, Jiangsu, and southern Shandong such as the Liulin and Huating sites; and the Dawenkou Culture (approximately 3000 - 2000 B.C.). As Shangraw (1978:34) points out, terminological confusion - 11 -over the Dawenkou Culture sites has resulted in different terms being used by Chinese archaeologists from different provinces. Shandong archaeologists have labelled the sites in Shandong as Dawenkou Culture sites while those in Jinagsu have used the terms Qingliangang, Liulin and Huating periods. A symposium in 19 77 described by the Wen Wu Correspondent (1978) attempted to resolve these terminological problems. Archaeologists at this symposium had three different views regarding classification of sites north of the Yangtze. As explained by the Wen Wu Corres pondent (19 78), the first view, by the Nanjing Museum, advocated six periods consisting of certain strata from various sites. The Shandong Museum advocated eleven periods and Beijing Univer sity held the third view that there are eight periods. The participants of the symposium also grouped these periods to gether into larger stages. These stages appear to be derived on the basis of radiocarbon dates, stratigraphy, technological and stylistic characteristics of ceramics, stone tools and other aspects of material culture; and an assessment of general cultural complexity. The Shandong Museum advocated three stages while the proponents of the two stage view disagreed on the divisions of the first and second stages. Most Chinese investigators appear to advocate the three stage division of the Dawenkou Culture. According to Gao (personal communication 1983; 1980:49), the 1974 excavation at the Dawenkou site belongs to the first stage, the Early and Middle periods from 1959 belong to the second, and the Late - 12 -period burials belong to the Late stage. The Shandong Museum (19 78) describes three stages for the Dawenkou Culture: the Early stage consisting of burials from the 19 74 excavation at Dawenkou, as well as one layer of the Dadunzi site, the Middle stage as consisting of some graves from the 19 74 excavation, some graves from Yedian and Dadunzi and the Early and Middle periods from the Dawenkou site (1959 excavation), and the Late period burials from the 1959 excavation comprising the Late stage, along with burials from Xixiahou and Yedian. There was also disagreement at the 19 77 symposium regarding the relationship of cultures north and south of the Yangtze. Some participants regarded the northern sites as related but culturally distinct from the southern sites, while others regarded them as fundamentally the same. Because the term "Dawenkou Culture" has been used fairly consistently in the Chinese archaeological literature since this symposium, it appears that the majority of Chinese archaeologists now hold the view that the cultures north and south of the Yangtze developed from the earlier Qingliangang culture which was distributed in areas both north and south of the Yangtze. The major characteristics of sites in the eastern seaboard region include probable cultivation of rice and other plants suitable to the low lying, marshy environment of the region, as suggested by the discovery of the Hemudu site in Zhejiang (Pearson 19 83:124) — although no remains of rice have been found in sites north of the Yangtze (Chang 1979:136). There is - 13 -great consistency in burial practices, including: similar grave orientation and body disposition; similar grave good inclusions such as stone tool and ceramic forms, ornaments, pig skulls, deer teeth (Chang 19 79:160-162, Pearson, in press:17-24), animal shaped vessels (Pearson 1983:140); similar practices such as tooth extraction and skull deformation primarily for females (Pearson, in press:26), and keeping a clay ball in the mouth during one1s lifetime (Han and Pan 1980). Incised symbols on ceramic sherds at some sites may indicate a form of proto-writing (Chang 1979:161). A description of burial practices at specific Dawenkou Culture sites is included in different chapters of this study: Chapter 4 (artifact forms associated with the sexes), Chapter 5 (body disposition, grave form, orientation), and Chapters 6 and 7 (status-related grave goods). The environment of the region was probably warmer and moister than at present (Pearson, in press:25, 1983:134), and most sites are located on small rivers or lakes (Pearson, in press:2). Some rivers in Shandong may have been suitable for small to medium-sized irrigation projects (Pearson, in press:24). The only habitation remains of which I am aware at Dawenkou Culture sites are the remains of house foundations from the 1974 excavation at the Dawenkou site (Wen Wu Correspondent 1978) and a reference to a site in Shandong by Beijing University (19 83: 195). The Dawenkou site, as represented by the 1959 excavation, appears to be the only cemetery site containing a kiln. The - 14 -kiln may be associated with the Late period (Gao, personal communication, 19 83) because some Late period sherds were found in it. However, the fact that sherds from the Longshan and Shang periods were also found with the kiln (Shandong Provincial Cultural Properties Commission and the Jinan City Museum 19 74:114) may indicate the kiln is not dated to the Late period. Subsistence and settlement pattern data for the Dawenkou Culture are lacking (Pearson, in press:30). However, animal bones from Dawenkou Culture sites suggest a moist, forested environment: river deer, wild pig, wild bovid, species of turtle (Pearson, in press:25). Animal species mentioned in the Dawenkou site report include alligator, crane, different types of fish and birds, oysters, and domesticated pig and chicken (Shandong Provincial Cultural Properties Commission and the Jinan City Museum 1974:156-158). Some types of stone tools described in the site report reflect agricultural activities such as the sickle, stone spade and adze. Fishing is suggested by fish hooks, weaving and sewing by spindle whorls and bone needles, and hunting by bone arrowheads and knives. 1.3. Cultural Evolution in the Late Neolithic Period  As Fig. 1-2 illustrates, there are several cultural regions during the late Neolithic period. Chang (1979:155) maintains that by the latter part of the Neolithic period (approximately 3200 B.C.) in several of these regions, cultures became increas-- 15 -ingly similar to one another. These cultures comprise the Longshanoid cultural complex. The Longshanoid cultures evolved into the Longshan culture in several regions. Chang (1979:144) characterizes the Longshanoid phase as transitional but basic ally egalitarian and the Longshan phase as a "war-like and ranked society preparatory for the formation of civilization and the state". An (1979-80:45) disagrees with Chang's (1979) classifica tion of sites into the Longshanoid cultural complex on the grounds that each culture developed in a different manner. However, he states that the late Neolithic cultures from various regions are characterized by much interaction in terms of "exchange and influence on the one hand, and integration and unification on the other" (ibid). Chang (1981b:155) also em- J. phasizes the extent of regional interaction during the Long shanoid phase, saying that cultural systems in various regions evolved side by side with those in other regions. It is implied that ranking developed in more than one region during the Longshanoid phase. Investigators have noted a number of cultural continuities in burial practices from the Dawenkou Culture to the Shang period and have debated whether the Dawenkou Culture had an influence in the development of the Shang culture (Thorp 1980: 51, Chang 1980:345-46, 354, Chang 19 83:509-10). These cultural traits include the log tomb, second level platform, several forms of artifacts, and some types of raw material such as jade. The log tomb is also present in some Longshan sites (Cheng - 16 -1982:21). Historical texts also suggest that cultural evolution in the Dawenkou or Qingliangang culture area was related to the development of the state in China (Fried 1983: 488). However, Zou (personal communication 1983) maintains that Chinese archaeologists now believe the Dawenkou Culture did not directly influence the development of the Shang state. A new culture subsequent to the Longshan in Shandong has recently been discovered, to which the Dawenkou Culture is directly related (ibid). It is likely that new discoveries in Shandong and elsewhere will continue to modify our understand ing of socio-cultural evolution in the eastern seaboard region of China. - 17 -CHAPTER 2 METHODOLOGY 2.1. General Approaches in Mortuary Analysis During the past twenty years, mortuary sites have been analyzed primarily in order to provide an understanding of the social systems represented at these sites. The two major concerns of mortuary analyses have been: 1) status differ entiation among burials in a cemetery or in other types of disposal areas and 2) social relationships among burials in terms of kinship groups or residential groups. Some recent mortuary analyses have been concerned with understand ing the ideological component of culture represented by burial remains. Variability in grave good inclusions, grave form, human skeletal remains, and in spatial patterning of grave goods, grave form, or human remains are the major aspects of mortuary treatment that have been assessed in studies concerned with either social organization or ideology. Human skeletal remains analyzed by physical anthropologists have enhanced studies of status or social subgroups. Results of these studies have indicated a relationship between stature and social status (e.g., Haviland 1967, Buikstra 1976), degenerative joint disease and status (Tainter 19 80) and enamel defects and status (Cook 1981). Human skeletal remains have also provided information on biological distance - 18 -(e.g., Buikstra 1976) and paleodemography (Chapman and Randsborg 1981:19-20). The other aspects of material cul ture noted above are focused upon in this study. The primary concern of most mortuary studies has been status differentiation. An investigation of social subgroups is included in only some of these studies. Since the nature of status differentiation represented at Dawenkou is the research problem addressed in this study, general approaches to analyses of status differentiation are emphasized here. However, this chapter includes some discussion of general approaches to analyses of social subgroups because, as argued below, an understanding of the nature of status differentiation represented at a site is not complete without consideration of the nature of social subgroups represented. There are currently two general approaches to the study of status differentiation from mortuary remains in the archaeological literature: 1) the processual approach, which has the goal of reconstructing the social organization of the community represented by the burials in a mortuary site and 2) the symbolist approach, which has the goal of understand ing status distinctions in a society within the broader context of the symbolism expressed by material culture and the ideology of that society. Many of the articles following the symbolist approach have appeared in the. literature since the comprehen sive review of: archaeological mortuary analysis by Chapman and Randsborg (19 81). - 19 -The majority of mortuary studies in the archaeological literature follow the first approach (e.g. Tainter 1978, 1977a, 1977b, 1975a, 1976, 1973; Tainter and Cordy 1977, Shennan 1975, Wright 1978, Decker 1969, Braun 1977 and 1979, Van de Velde 1979, Rothschild 1979, King 1969, Goldstein 1981 and 1980, Hatch 1976, Peebles 1971, 1972, 1974; Peebles and Kus 1977, and Mainfort 19 77). The proponents of the second approach include Leach (1977), Hodder (1982a and b, 1980), Pader (1982, 1980), Shennan (1982), Pearson (1982), Shanks and Tilley (1982), and Blackmore et al. (1979). These authors maintain that the processual approach is totally inadequate for understanding status differentiation (Hodder 1980:161). I maintain that elements of both approaches are useful and that elements of both are problematic. The approach of the mortuary analysis presented here is based upon elements from each. 2.2. The Processual Approach The processual approach is based mainly upon the cross-cultural ethnographic tests of Saxe (1970) and Binford (1971), and to a lesser extent, from Tainter (1975b, 1973) and Goldstein (1976). These tests, described below, are important because they indicate that different social statuses of individuals and the general social organization in the society of the deceased are reflected by mortuary remains (Goldstein 1981:54). For the context of this mortuary analysis, "status" is defined as "relative rank in a hierarchy of prestige" - 20 -(Webster's Seventh New Collegiate Dictionary 1969, page 856). One can assume that differential treatment of an individual in death reflects his or her differential treatment in life (ibid). Although conclusions from these tests need to be confirmed from further testing with ethnographic and archaeo logical data (Chapman and Randsborg 1981:23; Bartel 1982:52), they are still useful for current mortuary analyses. More recent ethnographic studies have indicated that aspects of organization may not be directly or totally expressed by burial remains. However, the cross-cultural regularities derived from the tests of Saxe (1970), Binford (1971), Tainter (19 73, 1975b) and Goldstein (19 76) provide a means by which to assess variability in mortuary remains. No cross-cultural ethnographic tests regarding mortuary ritual have been pub lished by researchers who advocate the symbolist approach. The conclusions from Saxe (19 70), Binford (19 71), and Tainter (1973, 1975b) have been utilized by researchers following the processual.. approach to formulate test implications for achieved social status versus ascribed social status, as well as to identify various status levels in the society in which the deceased had lived. Saxe1s (19 70) test based upon three modern societies of varying social complexity (the Kapauku Papuans of New Guinea, the Ashanti of West Africa and the Bontoc Igorot of Luzon, Philippines) and Binford's (19 71) test based upon 4 0 non-state organized societies from the Human Relations Area Files both indicate that mortuary remains reflect the - 21 -"social persona" of the deceased, or the "composite of the social identities maintained in life and recognized as appropriate for consideration after death" (Binford 1971:17). During life, the various social identities of an individual are expressed in different situations. At the death of an individual, the living decide which identities of the deceased are the most important and should be expressed in mortuary ritual. According to Saxe (19 70:4-9), the determining factors in this decisions are the rights and duties of the living to the deceased. The aspects of the social persona that are symbolized as well as the form of the symbols themselves vary from culture to culture (Binford 1971:16-17). Saxe's (19 70) study involved the testing of eight hypotheses, most of which received partial support. The two hypotheses firmly supported are 1) The Components of a Given Disposal Domain Cooperate in a Partitioning of the Universe, the Resultant Combinations Representing Differ ent Social Personae (Saxe 1970:65). 2) In a Given Domain, the Principles Organizing the Set of Social Personae (Produced by the Cooperative Partitioning of the Universe of Disposal Components) are Congruent with Those Organizing Social Relations in the Society at Large (Saxe 1970:66). Saxe's (19 70) eighth hypothesis refers to social subgroup affiliation and is discussed in Chapter 5. Goldstein's (19 76) ethnographic test, a re-examination of Saxe's eighth hypo-- 22 -thesis, is also discussed in Chapter 5. Binford (.1971) tested and received support for three hypotheses. Binford*s (19 71) study is the most useful ethnographic study for mortuary analyses to date because it considers a number of aspects of the social persona and of mortuary treatment with material correlates. Binford con cludes from the testing of these hypotheses that ...the form and structure which characterize the mortuary practices of any society are conditioned by the form and complexity of the organizational characteristics of the society itself (Binford 1971:23). The first hypothesis tested by Binford (1971) is ...there should be a high degree of isomorphism between a) the complexity of the status struc ture in a socio-cultural system and b) the complexity of mortuary ceremonialism as regards differential treatment of persons occupying the different status positions (Binford 1971:18). The complexity of the mortuary ritual is measured by the number of dimensions of the social persona in a society's mortuary practices. These various dimensions are age, sex, social position or status, sub-group affiliation, cause of death, and location of death. Due to limitations of the data set, complexity of socio-cultural system is measured by form of subsistence (whether hunter-gatherers, shifting agriculturalists, settled agriculturalists, and pastoralists). - 23 -A greater number of distinctions of the social persona is found to be symbolized in the mortuary ritual of settled agriculturalists, generally accepted as having the most complex form of socio-cultural system than the other forms of subsistence (Binford 1971:18). Binford's (1971:20) results are shown in Figures 2-1 and 2-2. Hodder (19 80:168) criticizes the testing of this proposition on the basis of Binford's measure of societal complexity - which Binford (1971:18) admits is simplistic. Binford's (1971) second hypothesis is: We would predict that age and sex should serve more commonly as bases for mortuary distinc tion among hunter and gatherers; while among agriculturalists, social position, as varying independently of age and sex as well as sub group affiliation, should more commonly serve as the basis for differential mortuary treat ment (Binford 1971:20). Binford (1971:20) concludes from his tabulations (shown in Figure 2-1) that there is a marked difference in the number of cases from the total in which age and sex are symbolized in the mortuary ritual of agriculturalists versus hunters and gatherers, and in which social position is symbolized. This hypothesis is the basis of the test implications employed in processual mortuary studies for achieved versus ascribed social position. Binford's (19 71) third hypothesis is employed in processual mortuary studies to indicate status distinctions - 24 -FIGURE 2-1. Number of dimensional distinctions of the social persona symbolized in mortuary practices according to form of subsistence. dimensional hunters & shifting settled pastoralists distinctions gatherers agriculturalists agriculturalists conditions of 1 0 i death X o j_ location of 1 i 0 0 death X age 2 1 7 1 sex 12 4 10 3 social position 6 5 11 0 social affiliation 4 3 10 1 total cases 15 8 14 3 (from Binford 1971:20, Table 2) FIGURE 2-2. Average number of dimensional distinctions according to subsistence category. subsistence category average number of dimensional distinctions per category (1) hunters & gatherers 1.73 (2) shifting agriculturalists 1.75 (3) settled agriculturalists 3.14 (4) pastoralists 1.66 (from Binford 1971:20, Table 3) - 25 -or levels in the society of the deceased: ...the locus of mortuary ritual and the degree that the actual performance of the ritual will interfere with the normal activities of the community should vary directly with the number of duty status relationships obtaining between the deceased and other members of the community (Binford 1971:21). This hypothesis is supported and other correlations are found between the various characteristics of the social persona and different types of mortuary treatment. Figure 2-3 indicates the manner in which age and other aspects of the social persona may be distinguished in mortuary ritual. Social position is expressed by the greatest variety of aspects of mortuary treatment, and many different aspects may symbolize social position in one society (Binford 1971:22). Social position tends to be distinguished by the form and quantity of grave goods, particularly "status-specific" 'badges' of office and quantities of grave goods, and the location of interment (Binford 1971:23). Tainter (1978:121) disagrees with Binford (1971:23) that status is often reflected by grave goods. Tainter's ethnographic survey of mortuary practices revealed that grave goods reflect status in less than five percent of the cases (Tainter 1974:125). Tainter's (1973, 1975b) ethnographic test expands upon Binford's (1971) third hypothesis. Tainter (1973:6) proposes that the amount of community involvement and the degree of - 26 -FIGURE 2-3. Aspects of the social persona symbolized in mortuary treatment. condition location age sex social social of death of death position affiliation (1) preparation * ^> . . 5 "g (2) treatment (3) disposition 2 1 - - 2 2 2 13-2 1 1 - 1 - 3 1 3 - 9 3 - 7 - 8 15 16 5 - - 9 (4) form J5 > (5) orientation oo (6) location JJ (7) form only 3 5 (8) quantity only ** (9) form and quantity (from Binford 1971:22. Table 4) - 27 -activity disruption in mortuary ritual corresponds to the amount of energy expended in the mortuary ritual. The higher the social rank of an individual, the greater portion of the community and degree of activity disruption and the greater amount of energy expenditure (ibid). This hypothesis is initially tested from the archaeological and ethnohistoric data from the Kaloko cemetery, Hawaii (Tainter 1973) and in more detail from a sample of 103 ethnographic societies (Tainter 1975b). Energy expenditure is reflected by size and elaboration of grave, method of handling and disposal of the corpse, and the nature of grave goods (Tainter 1976:95). It is claimed that the hypothesis is clearly supported and that one may consider that different amounts of energy expenditure in mortuary ritual correspond to different levels or grades of ranking (ibid). The conclusions from Tainter's (1973, 1975b) ethnographic tests have been challenged by Braun (1981) and Kirch (1980). Braun (19 81:411) points out that Tainter did not test whether differences in energy expenditure can occur among individuals who do not differ in social status (although Tainter 19 81:419 insists that he did). Kirch (1980) tests Tainter's hypothesis linking energy expenditure and social rank with ethnohistoric and archeological data from a Tongan society mortuary site. Kirch (19 80:306) concludes that in Tongan society, energy expenditure reflects relative sociopolitical status more than social rank. Kirch (1980:304-5) finds that a high rank ing chief has a smaller grave monument than a chief of lower rank. Historical data indicate that a direct correlation - 28 -between societal rank and socio-political status was not present in Tongan society (Kirch 1980:305). Some individuals of high societal rank did not have much political power, while others with a great deal of political power were of lower societal rank (ibid). Recent ethnographic studies have also indicated that status differentiation can be inferred from mortuary remains. O'Shea (1981:49) concludes from his ethnohistorical and archaeological study of three Plains Indian mortuary prac tices that aspects of social ranking known from ethnohistoric data are clearly observable from the archaeological remains of Plains Indian mortuary practices. In the Plains cemeteries, ranking is symbolized in an obvious manner, by various symbols of wealth and increased energy expenditure (O'Shea 19 81:49-50). However, the total range of mortuary treatment is not observ able from the archaeological data. Contrary to Binford (19 71:22), O'Shea (19 81:49) concludes that distinctions of social subgroup affiliation are not clearly observable from archaeological remains of mortuary practices. Chapman and Randsborg (1981:8) mention a recent study which indicates that social status in rural Hungarian cemeteries has been symbolized by form and color of grave markers. Some authors (Ucko 1969, Leach 1977, Orme 1981) deny the potential of archaeological mortuary data for indicating information about status differentiation or any other aspect of social organization. Ucko (1969:265-266) states that grave goods may be offered but they may not be placed in - 29 -graves with the deceased. Also, the quantity and quality of grave goods may not reflect relative status of individuals (Ucko 1969:266-267), nor grave size (Ucko 1969:296-299). However, as Chapman and Randsborg (19 81:8-9) point out, Ucko (1969:270) actually confirms the conclusions of Binford's (1971) and Saxe's (1970) studies by saying: ...in the vast majority of cases known ethno-graphically, a culture or society is not charac terized by one type of burial only, but that, on the contrary, one society will undertake several different forms of burial, and that these forms will often be correlated with the status of the deceased (Ucko 1969:270). Ucko (1969:270) adds that location of graves particularly indicates status of individuals. Orme (1981:235) cites a few examples in which the mortuary ritual is more a reflection of social status of living relatives rather than that of the deceased. As Bartel (1982:47) points out, both Orme (1981) and Ucko (1969) select certain ethnographic cases which show that mortuary ritual does not reflect social status. Their studies would be more useful if a cross-cultural sample of mortuary practices had been included (ibid). The same criticism can be applied to Leach (1977:162). It is .clear from the ethnographic studies just described that social status and other aspects of the social persona are not expressed by the same aspect(s) of mortuary treatment in all social groups. For example, status may be expressed - 30 -by wealth in terms of grave goods, energy expenditure, or both. Also, different aspects of the social persona are considered important in different cultures. One study of modern mortuary practices indicates that cause of death is expressed more in mortuary ritual than social status, while another study concludes that social and religious affiliation tends to be expressed in times of political or economic stress (Chapman and Randsborg 19 81:8). Grave goods do not seem to express status in Danish Viking graves (Randsborg 19 81:112) and in a large group of burials from the Moche Valley representing 3500 years (Donnan et al. 1978:312). Huntington and Metcalf (1980:1) point out that there is a tremendous amount of variation in mortuary ritual throughout the world. However, the mortuary programs of individual societies display a large amount of redundancy. This conclusion has been made by Bartel (19 82:55) on the basis of his 19 73 study of 2 7 ethnographic societies from the Human Relations Area Files and Tainter (1978:114) on the basis of Saxe's (1970) formal analysis of mortuary treatment of the three societies mentioned previously. Thus, different social statuses, for example, symbolized in mortuary ritual should be identifiable by archaeological data due to repetition of certain mortuary treatment(s). For example, high status may be consistently symbolized in a cemetery by large quantities of artifacts and a particular form of artifact. - 31 -Some processual mortuary studies investigate status distinctions on the basis of only wealth in terms of grave goods (such as Rothschild 1979, Peebles 1972, 1974; Shennan 19 75) or only energy expenditure (e.g. Tainter 19 77a and b, 1975a, Tainter and Cordy 1977, Tainter 1973, Tainter 1976). These studies assume that status is expressed in the same manner in all socieites. Wealth and energy expenditure as well as other variables Binford identifies as reflecting status (body treatment, body preparation, body disposition, grave location) (1971:22) should be regarded as only potential status indicators at a mortuary site. This conclusion is also reached by Whittlesey (19 78:106). The problem with mortuary studies that follow the processual approach is that the derived status distinctions are interpreted as representing the major status levels that existed in one community at one point in time (M. Blake, personal communication, 1983). This manner of interpreta tion ignores the fact that most cemeteries contain individuals who were interred continuously throughout a certain length of time. The individuals from each chronological period within a cemetery do not represent one social system but any number of social systems. This point is recognized in the reviews by Chapman and Randsborg (19 81:15) and Braun (1981:409), the ethnohistorical and archaeological study by O'Shea (1981:40), and in the processual mortuary studies by Doran (1973:150-151), Jones - 32 -(1980:193), Chapman (1977:30), MacDonald (19 80:38-39), Shennan (1975:280), Van de Velde (1979:46), Gruber (1971:64-65), Hodson (1977:403, 1979:25), Greber (1979b:36), and Goldstein (1981:56-57). Goldstein (1981:56-57) points out that if the time factor is not considered, different perceived rank levels may be due to changes through time in mortuary treatment. O'Shea (19 81:45) notes a change through time in Pawnee mortuary treatment in the means by which a certain social distinction was symbolized and in the social distinc tion itself. There was a great change in all aspects of Arikara mortuary treatment, especially in the types of grave goods (O'Shea, 1981:48). O'Shea (1981:51) concludes that similar changes in mortuary treatment through time should be expected at any mortuary site. O'Shea (19 81:52) states that his study supports Binford's (1971:23) conclusion that mortuary treatment reflects the organization of the society in which the deceased had lived but that the possibility that the mortuary treatment changed through time must also be considered. The state of rapid change characteristic of the Plains societies is probably not typical of prehistoric societies, but some change in social organization should be expected (ibid). Braun (19 81:409) points out that the longer a cemetery is in use, the more likely there will be changes in social organization and in mortuary ritual. Although the mortuary studies cited above acknowledge the time factor, it is ignored by many researchers when interpreting the status levels represented at mortuary sites - 33 -within each chronological period (such as Shennan 1975, Tainter and Cordy 1977, Tainter 1975a, 1977a and b, 1973; Greber 1979a, Wright 1978, Peebles and Kus 1977, Rothschild 1979, Spencer 1982, Binford 1972, Peebles 1972, 1974, Buikstra 19 76). Some researchers that address the problem of the time factor make unwarranted assumptions about the population using the cemetery under study. Peebles and Kus (19 77:431) assume that the composition of a population using a cemetery (in terms of size, age and sex) does not change significantly through time, so that the age and sex popula tion of the site reflects that of the population through time. Shennan (1975:283) estimates the size of the living community from Branc and its age and sex distribution at an average point in time, assuming that the population size and age and sex structure did not change. Greber (1979a:43) assumes the individuals within the Seip mound are contemporaneous. I argue that such assumptions should not be made until an effort has been made to determine the chronological relationships of burials within the same broad chronological period. Only a few archaeological mortuary studies have made this attempt. These studies have attempted to determine if different spatial areas of a cemetery represent different periods of time within one chronological period (Goldstein 1981:66, 1980:122-123; Tainter 1976:102, MacDonald 1980:38-39, Shennan 1975:280, Doran 19 73:150-151, Jones 1980: 193, and Hodson 1979:25). A few mortuary studies do consider - 34 -change through time in the system of status differentiation in a cemetery with more than one chronological period (Braun 1977, Wright 1978, Mainfort 1977, Hatch 1977, Tainter 1977a and b). Pearson (19 81) discusses change through time in status differentiation within the Dawenkou site and within the Dawenkou Culture region as a whole. Even fewer mortuary studies have attempted to explain the cultural processes responsible for change within a site or cultural region (e.g., Wright 1978, Pearson 1981). The order in which the burials within a cemetery were interred is one aspect of the processes that affected the archaeological record at a cemetery. Consideration of these processes has been lacking in processual mortuary studies with the goal of understanding social ranking (Chapman and Randsborg 1981:11 and Tainter 1978:109). I maintain that these processes include natural and cultural factors that may have created site disturbance and three factors characteristic of the cultural group(s) using the cemetery: 1) the order of interment of burials as discussed above, 2) the nature of the social subgroup(s) using the cemetery, and 3) the nature of the mortuary ritual itself - whether the mater ial forms chosen to symbolize particular aspects of the social persona are preserved in the archaeological record and whether they are clearly recognizable. The mortuary studies cited previously that follow the processual approach attempt to reconstruct the status distinc tions in one particular society. It is implicitly assumed - 35 -that the people buried in the cemetery are from one community. It is known from ethnographical and archaeological data that there is great variability in the association between settle ments and cemeteries (Chapman and Randsborg 1981:15). A cemetery may represent one whole community, subsection(s) of one community, subsection(s) of more than one community, or more than one whole community. As O'Shea (1981:40) points out, consideration of the third culture-specific type of process mentioned earlier has especially been lacking in the literature. His study is a valuable contribution to the understanding of aspects of mortuary ritual that are observable in the archaeological record. O'Shea (1981:49-50) concludes that although social status tends to be clearly reflected in the mortuary remains from mortuary ritual, social subgroup affiliation tends to be symbolized by materials which do not preserve well or in a subtle manner. O'Shea (19 81:52) maintains there are regularities in the formation processes of mortuary sites, and although there tends to be a distortion between social organization as interpreted from archaeological mortuary data (especially social subgroup affiliation) and the actual social organization of a society, this distortion is predictable. Levine's (19 77) study of New Guinean mortuary rituals also indicates that aspects of the social persona are not directly expressed in mortuary ritual (Chapman and Randsborg 19 81:14). - 36 -2.3. The Symbolist Approach The mortuary studies following the symbolist approach also maintain that the goal of social reconstruction is problematic, as well as the lack of consideration of processes that affected the archaeological record of a cemetery. The symbolist approach is clearly expressed in Hodder (1982a, 1980). Hodder (1980) calls for a totally new approach in mortuary analysis than that employed in processual mortuary analyses. Hodder's (19 82a, 19 80) ethnoarchaeological work among the Nubia in Sudan demonstrates (like the work of O'Shea 19 81) that social organization may not be directly expressed by archaeological mortuary data. There tends to be "a distortion and structured disjunction" between mortuary patterning and patterning characteristic of the living society (ibid). The material culture of a society may either repre sent or misrepresent the social organization of the society, depending upon the ideology and symbolic codes of the society (Hodder 1982:210). Thus, the degree to which the material remains from mortuary ritual reflect the social organization of a society depends upon that society's ideology. The mortuary studies following the symbolist approach have made the informative conclusion that depending on the particular cultural attitudes towards death, status distinc tions or other aspects of the social persona may or may not be symbolized in the material remains from mortuary ritual. In some Nubian societies, many personal items of the deceased - 37 -are broken and placed on top of the burial mound instead of inside it (Hodder 1982a:163). The Lozi in Zambia do not symbolize high status by material goods (1982a:120). The fear of impurity created by death causes British gypsies to break objects and not place them in graves (Hodder 1980:167). Thus, the lack of differentiation of status or of any other aspect of the social persona by mortuary treatment does not indicate that this differentiation did not occur in life (Hodder 1980:166). Unlike O'Shea (1981:52), Hodder (1982a: 207-208) maintains that the nature in which the archaeological record reflects a transformation of social organization cannot be predicted, because different societies have different ideologies and symbolic codes which continually change with time. However, Hodder concedes that archaeological mortuary data do reflect some aspects of social organization: While burial behavior may distort and invert, it does not totally hide. There will always be some aspect of the societal organization which can be picked out in the gross cross-cultural reviews as being reflected in burial (Hodder 1980:168). Hodder's (19 80:168) solution to the problem of deter mining the degree to which the mortuary data from a site reflects the social organization of the social group(s) using the cemetery is to identify the ideology of the society with regard to death: - 38 -A new approach to burial must not expect simple correlations between social organization and burial. Rather, it must identify the way in which prevailing attitudes to death can be derived from different conceptions of the living practical world (ibid). Hodder (1982:208-9) points out that status differentiation in any society is more complex than has been realized in most mortuary studies. An understanding of social ranking requires consideration of attitudes towards domination, power and authority in a society (ibid). Hodder (1980:168) calls for new ethnoarchaeological studies which explain the relation ship between the ideology of societies and their mortuary rituals. A diachronic ethnoarchaeological study of the type advocated by Hodder (19 80:168) has been reported by Pearson (1982). Pearson's (1982) conclusions regarding the mortuary practices of Victorian and modern England are potentially applicable to prehistoric mortuary sites. Mortuary ritual may reflect ideal, not actual, relations of power within a society, and living social groups may manipulate the statuses of deceased individuals in order to elevate their own statuses or for other purposes (Pearson 1982:112). I maintain that until more individual or cross-cultural ethnoarchaeological studies of the kind advocated by Hodder (19 80:168) have been made, the only resort for the researcher is to make the most of the ethnographic and archaeological - 39 -data available. Also, although status relationships are most likely to be more complex in life than that apparent from mortuary treatment, it has been established that mortu ary remains do tend to reflect status relationships at least in part. The kind of ethnographic study advocated by Hodder may be as controversial as those presently employed in mortuary analyses because informants' accounts of mortuary practices tend to be unrepresentative of the total range of variability in the mortuary practices of a society (O'Shea 1981:43). Hodder (1982a, 1980) does not provide a method which can identify cultural attitudes towards death on the basis of archaeological data. Leach (1977:169) also maintains that cultural ideology regarding death should be assessed and does not provide a method. Other mortuary studies of the symbolist approach provide methods with which one can attempt to determine symbolic codes and cultural ideology regarding death with archaeological data. However, these methods are not based upon cross-cultural regular ities in the ideology of death. Also, some of these studies do not recognize that mortuary sites reflect a number of changing social systems through time. Pader (19 82, 19 80) maintains that the spatial patterning of artifacts within graves, particularly items of dress or bodily adornment, symbolizes social relations within a society. A summary of items of dress or bodily adornment as symbols in the mortuary ritual of ethnographic societies is included in Pader (1982). Pader (1982,1980) studies the - 40 -variability in the above type of patterning with regard to age, sex, skeletal position, and spatial location of graves within two Anglo Saxon cemeteries. Her study is synchronic. Pader maintains that a study of social ranking based upon the relative richness of burials alone is limited (Pader 1982:54, 1980:143). A study of ranking which ignores symbolism reflected by the spatial component (either within graves or between graves) ignores the cultural context of any derived status distinctions (Pader 1980:170). The cultural context is also indicated by variability in the other aspects of mortuary treatment at a site (ibid). Pader (1980:156) concludes that the male and female sexes within the two cemeteries are symbolized in different manners, so that a comparison of the quantity and quality of artifact types between the two sexes would be misleading. Like Pader (1982, 1980), Shanks and Tilley (1982:152) maintain that neglect of the symbolism reflected by mortuary treatment neglects the social context of derived status distinctions. They assess the treatment and arrangements of bones from skeletons in European Neolithic barrows. Com parison of the archaeological patterning with ethnographic data suggests that differential access to power by various social subgroups is being symbolized (Shanks and Tilley 1982:151). An aspect of mortuary treatment regarded in some mortuary analyses following either the symbolist or processual - 41 -approach as especially important for understanding social context is spatial location of grave. The spatial component is considered important by Pader (1982, 1980 - within graves as well as between graves) and particularly in the processual mortuary study described in Goldstein (1980, 1981 - between graves only). Chapman and Randsborg (1981:14) state that spatial patterning is another neglected topic in mortuary analysis. Goldstein (1980:9-10) maintains that spatial location of graves within a cemetery provides information on the nature of derived status distinctions and on how these distinctions operated within the society in which the deceased had lived. Like Pader (1982:170), Goldstein (1981: 56) asserts that a study of social ranking based exclusively upon artifacts is incomplete and disregards the social context in which the derived status distinctions should be considered: What does each group or status type mean? How do the groups relate to each other? What are the functions of each group, and what are the functional relationships between groups? While many of these questions may not be easily or reliably answered, current mortuary analysis does not even approach or attempt to ask these questions. Can we really say... that a culture in which we have determined seven social group ings is more complex than one in which we find six groupings? Some mortuary analyses following either the symbolist or processual approach examine mortuary sites in terms of - 42 -the regional social context as well as the local social context. These symbolic studies include Shennan (19 82), Hodder (1982b) and Blackmore et .al. (1979). Processual mortuary studies with this approach are Goldstein (1980, 1981), Peebles (1974), and Seeman (1979). The comparative study by Pearson (19 81) has been mentioned. The regional social context is yet another neglected aspect of mortuary analysis (Chapman and Randsborg 1981:23). Goldstein (1980, 1981) and Peebles (19 74) apply the same techniques that c identify different status levels to a number of sites in the specific cultural region under study. Seeman (19 79) provides an illuminating interpretation of the ritual and economic function of mortuary sites within the Hopewell culture region. Shennan (1982:160) suggests that different ideologies in different regions of the European Bronze Age resulted in the development of different types of ranking systems from region to region. Hodder (19 82b:175) maintains that changing design of ceramics and of other types of material culture reflect changing ideologies regarding regional expression of power and domination. In another study of regional ceramic design, Blackmore et al. (19 79:108) suggest that distinctively different designs from burial and settle ment sites symbolize competing ethnic groups. - 43 -2.4. The Approach Followed in this Mortuary Analysis This analysis of status differentiation utilizes test implications for achieved versus ascribed status derived from the cross-cultural regularities concluded by Saxe (1970), Binford (1971), and Tainter (1973, 1975b). The particular techniques utilized to identify the various status levels or distinctions represented among the Dawenkou burials are discussed in the following section. Variability in as many aspects of mortuary treatment that have been considered as potentially status-related is assessed. Variables that reflect wealth (such as quality and quantity of grave goods) and energy expenditure (grave form and size) are included. Other variables included are body disposition and spatial location of grave (see Figure 2-3). The Dawenkou site report does not provide a description of body treatment or body preparation. My exploratory study of status differentiation at Dawenkou conducted in 19 81 suggested that both wealth and energy expenditure symbolize status distinctions at Dawenkou. There is considerable variation in quantity of grave goods (from none to over 100 items), quality of grave goods (from utilitarian stone tools and ceramic vessel types to jade ornaments and items of carved ivory), and grave size in area (from less than 5.0m2 to 13.Om^), and grave form (from simple pits to log tombs). The status distinctions derived in the analysis of - 44 -status are interpreted with consideration of the fact that the Dawenkou cemetery represents a number of social systems through time. It is especially important to consider the time factor for Dawenkou because as mentioned in Chapter 1, the cemetery may have been used for 1000 years (Wu 1982). The number of derived status levels is not interpreted as being the exact number of status levels which existed in the society of the deceased. Instead, emphasis is placed upon distinct changes in the number and nature of status distinc tions in the cemetery through time. The exact number of status levels at one point in time is also not interpreted literally due to the probable distortion of status and of other aspects of the social persona as reflected in the cemetery. Some researchers have attempted to determine the degree of ranking represented in a cemetery at one point in time (Peebles and Kus 1977, Buikstra 1976, Tainter 1977a, 1977b, 1978; Tainter and Cordy 1977, Shephard 1979, Hatch 1976, Brown 1971). Buikstra (1976:35-37) describes the degree of ranking in Middle Woodland sites in terms of Fried's (1967) "ranked" and "stratified" societies. Peebles and Kus (1977), Hatch (19 76), and Brown (1971:102) attempt to determine whether a chiefdom in Service's (19 75) sense is represented at a site. Some researchers have criticized the character ization of sites in terms of evolutionary typologies (Hodder 1982a:208, Goldstein 1981:54, and Tainter 1978:115-117). In these studies it is argued that sites are pigeon-- 45 -holed into evolutionary types on the basis of a few traits and that such classification is description instead of explanation. Tainter (1977a, 1977b, 1978), Tainter and Cordy (1977), and Shephard (1979) give a calculation of the amount and degree of organization. These calculations have been criticized by Binford (personal communication, 1983), Braus (1981:408-412) and Goldstein (1981:55). The measurements of the amount and degree of organization assume that status represented in every cemetery is reflected by energy expenditure (see Tainter 1978:134). Also, by including in the calculations the number of individuals in every rank level (ibid), it is assumed that the cemetery directly reflects one entire social system. I argue it is more feasible with archaeological data to identify degree of organization in a relative instead of an absolute manner. Test implications for different degrees of ranking derived from ethnographic data have not been developed. Whether there is a change through time in the degree of status differentiation at Dawenkou is assessed in this study. A qualitative change is concluded if there is a change from achieved to ascribed status, and a quantitative change is concluded if there is a clear difference in the number of status distinctions through time. My exploratory assessment of status differentiation conducted in 19 81 suggested that there is a greater number of high status burials in the Late period (as defined by the authors of the site report) - 46 -in terms of grave goods and energy expenditure, located in more diverse spatial areas of the cemetery. In the final chapter of this study, there is a discussion of the nature of status differentiation represented at Dawenkou in comparison to other sites in the Dawenkou.Culture. Cultural processes such as those suggested by Pearson (1981) that may have been responsible for change in status differentiation through time are included. The analysis of status is conducted in Chapter 6. Three preliminary analyses are described in Chapters 3, 4 and 5. The first, in Chapter 3, is a chronological analysis of the burials. As mentioned in Chapter 1, there is a disagreement between the chronological periods given in the site report (Early, Middle, Late) and reported in the recent Chinese archaeological literature (Early, Late). The chronological relationships among burials must be understood before the analysis of status is undertaken; otherwise, any concluded variability in status distinctions may be a factor of the temporal dimension. The methods that were employed to derive the two chronological schemes are discussed and an effort is made to evaluate the two schemes. The second preliminary analysis, in Chapter 4, is an attempt to identify the sex of unsexed burials on the basis of the types of grave goods in known sexed burials. For an unknown reason, only 30 of the 12 7 single burials in the cemetery are sexed. Whether status differentiation cross cuts age and sex categories or whether it is on the basis - 47 -of age and sex categories is a key test implication that has been employed in processual mortuary analyses to indicate ascribed versus achieved status distinctions. An effort must be made to estimate the sex of the remaining burials in order to adequately assess whether status distinctions are ascribed (indicative of a ranking system) as Chinese and western archaeologists have proposed for Dawenkou. Another method by which female burials could be dis tinguished from male burials is identification of the burials with tooth extraction and skull deformation noted by Pearson (in press:26). Female burials could also be identified by determination of skeletons that had clay balls in their mouths during their lifetimes. Pearson (19 81:1084) notes that female skeletons from Dawenkou Culture sites tend to have the clay ball more than male skeletons. Unfortunately, information regarding tooth extraction, skull deformation, and the presence of clay balls for each burial is not provided in the site report. The third analysis, in Chapter 5, is the assessment of social subgroup affiliation represented in the cemetery. An attempt is made to understand the social context in which status distinctions were a part. Consideration of the number of portion(s) of communities represented at the cemetery is included. Some aspects of mortuary treatment at Dawenkou may reflect social affiliation as well as social position (see Figure 2-3): body disposition, grave form, spatial - 48 -location of grave. Variability in these aspects of mortuary treatment is assessed in both Chapters 5 and 6. The spatial location of graves within the cemetery is regarded as especially useful for understanding the nature of social subgroup affiliation. The spatial patterning of artifacts and skeletons within graves is not assessed here due to time constraints. Study of this patterning at Dawenkou is likely to be informative. There appear to be some regular ities in the placement of certain artifact types in relation to various body parts (Pearson 1981:1080). 2.5. Method to Determine Status Levels Despite the large number of mortuary analyses in the literature that follow the processual approach, a limited number of techniques have been employed to determine the status distinctions or levels represented in cemeteries. These methods have included: intuitive assessment based upon the quantity and quality of artifacts (Whalen 1983, Buikstra 1976, Larson 1971, Gruber 1971, Alekshin 1983, Jacobsen and Cullen 1981, Binford 1972, Greber 1979b, Milisauskas 1978), intuitive assessment and various significance tests (Rathje 1970, King 1969, Peebles 1971, Saxe 1971), the coefficient of variation, a measure of relative variability (Pearson 1981), formal analysis (Brown 1971, Decker 1969), ranking of burials or the assignment of wealth scores on the basis of quantity and quality of grave goods (Shennan 1975, Shephard - 49 -1979, Rathje 1973, Greber 1979a, Winters 1968, Blackmore et al. (19 79), and multivariate statistical techniques (monothetic - divisive or polythetic, in an R or Q mode). The monothetic - divisive techniques have included cluster analysis with the information statistic or with the sum of chi-square measure of association (Tainter 1975a, Goldstein 1980, 1981; Peebles 1972, 1974; Jones 1980, Mainfort 1977, Hatch 1976). Cluster analysis is the poly thetic method which has been employed to the greatest extent. The polythetic techniques have included Average Linkage cluster analysis (Shennan 1975, Tainter 1975a, Rothschild 19 79), Complete Linkage cluster analysis (Tainter 19 75a, Mainfort 1977, Hatch 1976), Single Linkage cluster analysis (Mainfort 1977, Hodson 1977), cluster analysis by Ward's Error Sum of Squares Method (Peebles 1974, 1972; Hatch 1976), cluster analysis by Program Mode (Peebles 19 72), factor analysis (Tainter 1975a, Bayard 1983), principal components analysis (Braun 1977, 1979; Van de Velde 1979, O'Shea 1981), and calculation of the amount and degree of organization by the information statistic (Tainter 1977a and b, 1978; Tainter and Cordy 19 77, Shephard 19 79). For a large data set such as Dawenkou, multivariate statistical techniques are more appropriate than intuitive assessment, formal analysis, or assignment of wealth scores. These techniques can group burials that co-vary in terms of the numerous variables considered as potentially status-related. Orton and Hodson (19 81) point out the statistical - 50 -difficulties in employing wealth score measures. The tech niques chosen for the analysis of status in Chapter 6 are three types of polythetic cluster analysis techniques (Average Linkage, Complete Linkage and Ward's Method) and multidimensional scaling. The above techniques are described in detail in Chapter 3. These techniques are employed in a Q mode to group burials of similar status. Most mortuary analyses in the literature are conducted in a Q mode, but some (e.g. Braun 19 79) have found it useful to employ an R mode analysis to identify co-varying sets of artifact types or grave forms. Factor analysis or principal components analysis is not suitable for the Dawenkou data set because the particular variables regarded as potentially status-related had to be coded on a presence/absence basis. The variables do not occur to a great enough extent among the burials to allow the use of frequencies. Polythetic cluster analysis techniques are chosen for the analysis of status because they have been considered useful for mortuary sites and for other subjects of archaeo logical research. In some mortuary analyses, it is asserted that monothetic - divisive techniques are better able to classify burials or artifact types than polythetic techniques (Tainter 19 75a, Goldstein 19 80, 19 81; Jones 19 80, Shephard 1979). The opinion in Tainter (1975a) is based upon a com parison of monothetic - divisive clustering techniques with - 51 -polythetic techniques on the same data set. Shephard (1979: 62) states that he compared several other types of clustering methods with the monothetic - divisive methods, but he does not name these types. Tainter's (1975a) comparison of techniques has been thoroughly criticized and considered inconclusive by Braun (1981:405). The opinion in Goldstein (1980:48, 1981:62-63) is based upon Tainter's (1975a) faulty comparison of techniques and Peebles' (1972, 19 74) use of monothetic - divisive methods for the Moundville site. However, Peebles (19 72, 19 74) maintains that both the mono thetic - divisive and polythetic types of cluster analysis yielded adequate results. Peebles (1972:10) states that the results from Ward's Method and Program Mode fulfilled a priori expectations, but some of the variability in the data set was not as apparent as in the results from the monothetic divisive analysis. Peebles (19 74:167-16 8) concludes that "each of the analyses captured slightly different aspects of the same sets of behavior" and that "each was valuable for the slight difference in their results". Comparisons of monothetic - divisive and polythetic clustering techniques have also been made by Mainfort (1977) and Hatch (1976). Mainfort (1977:78, 97) concludes that a polythetic clustering method (Single Linkage) yielded superior results, but as in Peebles (19 74), both types of methods emphasized different aspects of the varia bility in the data set. Hatch (19 76) employes a monothetic -- 52 -divisive method along with Ward's Method and Complete Linkage but does not comment on the effectiveness of one method over the other. The results from the studies above do not indicate that polythetic clustering techniques have no utility for mortuary analysis in comparison to monothetic - divisive techniques. Polythetic methods df classification are regarded by Braun (1981:405) as more appropriate than monothetic - divisive techniques for isolating redundantly symbolized status dis tinctions in a data set. Also, polythetic methods do not share the limitation of the monothetic - divisive methods, that define a given group on the basis of a unique set of features. Serious errors in classification result when an item that does not possess the one feature used to make a primary division is moved to a distantly related grouping, even if the item is related to the items of another grouping in every other feature (Sneath and Sokal 19 73:13). Three types of polythetic clustering techniques (Average Linkage, Complete Linkage and Ward's Method) are employed so that the patterning interpreted reflects the variability in the data set and is not a reflection of the mathematical characteristics of the particular techniques used (Matson and True 1974:72, Braun 1981:406). Patterning which is identified by more than one technique is regarded as characteristic of the data set. Interpretations from mortuary studies based upon only one type of multivariate - 53 -classification technique should be regarded as tentative until they are substantiated by another type of classification technique (e.g. Rothschild 1979, Goldstein 1980, 1981). Peebles and Kus (1977:43 8) and Shennan (19 75:283) employ more than one type of clustering technique but do not specify the types. To my knowledge, no mortuary analysis in the literature has utilized multidimensional scaling as a technique to identify status distinctions. Rothschild (1979:6 72) mentions she will use it in a future study. Multidimensional scaling has been regarded in the archaeological literature as capable of indicating important relations between items in a data set and as a comparable technique with cluster analysis (Johnson 1972). The results from the scaling and clustering techniques in Chapter 6 are compared in order to interpret the important status distinctions within the Dawenkou cemetery. The use of clustering and scaling to identify status distinc tions has potential utility for other mortuary sites (Binford, personal communication 19 83). In Chapter 6, it is demonstrated that the particular multivariate statistical techniques employed affect the status distinctions among the burials which result. Also, interpretation of clustering and scaling results involves a degree of subjectivity on the part of the researcher. This is recognized by many researchers and is discussed in Chapter 3. Therefore, there is a methodological reason as - 54 -well as a theoretical one for not interpreting the resultant status distinctions as the exact distinctions which existed in the society of the deceased. 2.6. Consideration of Natural and Cultural Factors that  may have Affected the Archaeological Record at  Dawenkou A natural factor that may have affected the archaeological record at Dawenkou is changing of the course of the Dawen River immediately adjacent to the cemetery, and subsequent flooding of a portion of the site. The river is shown in photo plate 1 of the site report. It is not clear whether any flooding has taken place since the interment of the burials during the Early and Late periods. It is possible that a portion of the burials originally laid during the Early and Late periods was not discovered by the excavators of the site in 1959, due to flooding or some other factor. Additional graves may have been discovered since 1959, too. A recent article on the Chinese Neolithic period by An (19 82) implies that more burials have been found at the Dawenkou cemetery. The English translation of An's article by the editors of China Reconstructs states: "So far some 1500 gravesites have been uncovered at Dawenkou" (An 1982: 62). Zou Heng (personal communication, 1983, interprets this unclear translation to mean that 1500 graves have been found at the Dawenkou site. I believe I am aware of the greater portion of the Chinese archaeological literature - 55 -available to the West on the Dawenkou cemetery. I have not found any information on additional burials discovered in the cemetery except for the burials excavated in 19 74 that belong to a time period earlier than those from the 1959 excavation mentioned in Chapter 1. Thus, I have not been able to determine whether any of the 1500 additional burials (provided Zou's interpretation is correct) are contemporary with those from the Early and Late periods of the 1959 excavation. If some or all of the 1500 burials at Dawenkou are contemporary with the 133 included in this mortuary analysis, the 13 3 burials may not be completely representative of the burial population as a whole. It is not clear whether disturbance of the 13 3 burials by later occupations occurred, either. It appears that the Longshan and Shang periods are represented only by a few pottery shards and vessels (shown in photo plates 110 and 111 of the site report). It does not seem that the use of the site during these two periods affected the 13 3 burials included in this mortuary analysis. However, intensive land use may have disturbed the archaeological record at Dawenkou. Five burials are noted in the site report as being disturbed. These burials are L3, E27, X39, L46, and L77. The time period(s) in which these burials were disturbed as well as whether both natural and cultural factors are responsible are not known. Only the extent of the distur bance in L77 is described as major. It is believed that - 56 -this burial originally contained a great number'.and'variety of grave goods (Shandong Provincial Cultural Properties Commission 19 74:14 7). The burials described as rich in grave goods in the site report are not said to be disturbed. A few burials are described as having incomplete skeletons; lacking a head or some other body part. It appears that the disturbance of these burials was limited to the skeletons and did not affect the grave goods in association. The specific excavation methods employed in 1959 at Dawenkou are not explained in the site report and cannot be assessed for their reliability. It is not known whether any problems in the reliability of reporting exist, either. The publication of the Dawenkou site monograph was interrupted during the recent cultural revolution in China (Goodrich 19 83: 16-17). Finally, the photographs of graves, skeletons and artifacts in the site report indicate that conditions of preservation at Dawenkou are good and that on the whole, the cemetery as excavated in 1959 is an accurate reflection of the material remains resulting from mortuary ritual during the Early and Late periods. Although the total effects of natural and cultural factors that may have affected the archaeological record at Dawenkou are not well understood, a mortuary analysis of the 133 burials excavated in 1959 is considered worthwhile. The 1959 burials are better reported than those excavated in 1974, and they have been considered by western and Chinese - 57 -archaeologists as critical for understanding the development of ranking in the eastern seaboard region. A mortuary analysis of the 133 burials utilizing current archaeological methods should be able to contribute to a more complete understanding of the nature of status differentiation at the site and in the Dawenkou Culture region as a whole. The complete description of each burial in the site report allows such an analysis to be made. The 133 burials described in the 1974 site report may be the most complete source of information on the Dawenkou site for many years. The problem of uncertainty over the representativeness of an excavated site is not new to East or Southeast Asia (e.g. MacDonald 1980:32 on the Bang site in Thailand and all other sites in Southeast Asia). As MacDonald (1980:32) points out, as long as the researcher realizes the results of his or her study are tentative and subject to new interpretation as more data becomes available, study of sites with the above problem is worthwhile. - 58 -CHAPTER 3 CHRONOLOGICAL ANALYSIS 3.1. Introduction In this chapter, an attempt is made to evaluate the two chronological . sequences that have been suggested for the Dawenkou burials. Without such an evaluation, it would be necessary to regard all of the burials as belonging to the same period. Thus, it would not be possible to deter mine whether the temporal dimension affects any of the status-related variability interpreted from the multidimen sional scaling and cluster analyses. If the burials could be assigned to two or three chonological periods, changes in status differentiation through time could be assessed. Before discussing the methods I believe were employed to derive the two chronological schemes, it is necessary to give a brief description of the ceramic classification system upon which they are based. The ceramic vessels in the 1974 site report are classified into 17 functional types, with a varying number of associated subtypes. The subtypes are subdivided into styles. The hierarchical relationship of the functional types, sub types, and styles can be seen in Figure A3-1. The majority of the functional types, subtypes and styles are described and illustrated in Chapter 5 of the site report. Represent-- 59 -ative photographs are provided elsewhere in the report. The descriptions, drawings and photographs suggest that the func tional types are distinguished by morphology and inferred function. Most of these terms are still in use. Chinese archaeologists have used the functional types to describe both ceramic and bronze vessel forms from a variety of time periods. The terms for the functional types of pottery vessels used today by Chinese archaeologists were apparently derived from catalogues of ancient bronze vessels compiled during the Song Dynasty (A.D. 960-1279) (Chang 1981a: 158-159), Song antiquarians used the terms for ancient bronze vessels given in classic historical texts or from inscriptions on the bronzes (Chang 1980:23). The use of the ancient terms has created confusion in the meaning of some vessel types (Chang 1981a:158,161) . The terms: from the classic texts are not always clear, and the terms from the inscriptions are of varying levels of inclusiveness (Chang 1980;23). Chinese archaeologists currently classify bronze vessels into general types according to their presumed functions and describe them by the traditional terms whenever possible (Chang 1980:24). It seems that ceramic vessels are classified in a similar manner. The subtypes from Dawenkou appear to be distinguished by morphology. The term for a subtype describes a prominent physical feature of a functional type. The subtypes are not - 60 -always mutually exclusive. For example, Figure A3-1 shows that the cai ("painted") subtype for many of the functional types is not distinguished by morphology. The vessels in this category could have any number of shapes. Another example from Figure A3-1 is the subtype wu bi ("no lugs") for the hu functional type ("storage vessel"). A vessel of this subtype could have any number of the physical features described for the other subtypes of storage vessel. Another potential problem is that the qi ta ("other") category contains a great variety of vessel forms. The style numbers indicate finer physical features of the subtypes. However, the difference between some of the styles is not obvious from the photographs. The functional types, subtypes and styles as described here appear to form a standard classification system used by Chinese archaeologists for other Dawenkou Culture sites, and perhaps for Neolithic sites in general (e.g., The Shandong Archaelogical Team 1964, The Nanjing Museum 1964, and the Shandong Provincial Museum 1972) . Four ceramic ware types are identified in the report, but these types are not included in the classification system described above. These types are: red (jia sha or "sand tempered" and ni or ni zhi, or "finely levigated clay") (C. Shangraw, personal communication, 1933), grey (sand tempered or of fine clay), black, and white (sand tempered or of fine clay). The significance of these ware types in terms of - 61 -status differentiation is discussed in Chapter 6. 3.2. The Methods to Derive the Chronological Periods The three periods defined by the authors of the Dawenkou site report (Early, Middle, Late) were derived from a judge mental assessment of changes in the ceramic vessel styles and subtypes. Chapter 7 of the site report. "The Method and Basis of the Periods", describes how this assessment was made. Ceramic forms are inferred as developing out of other forms. The changes in vessel forms from 15 pairs of burials that cut into one another (listed in Figure A3-2) form the basis of the relative chronology. The remaining burials in the cemetery were fitted into the chronological sequence by comparing the vessel forms in them to the sequence of forms in the above 15 pairs of burials. Except for the relation ship of the intrusive burials, stratigraphy was not employed as a basis for establishing the three periods. The site report does not provide any stratigraphic information for the 195 9 excavation of the cemetery. My interpretation of the method by which the three periods were derived has been confirmed and amplified by C. Shangraw (personal communication 1983). Gao's (1980) article includes a discussion of the three periods and their derivation. Chapter 7 of the site report indicates that the method to determine the sequence of developing ceramic forms is based upon implicit, intuitive criteria. The authors discuss - 62 -the evolution of the most important functional types, sub types and styles but do not explain how these forms were identified. It seems the most important forms are those with the most obvious physical features. The evolution of the various subtypes is described in the most detail. Certain morphological features are described as getting flatter, more bulging, etc. through time. When functional types or subtypes appear to die out is also noted. From the distribution of style numbers per period for the subtypes provided in Table 2 of the siteireport (page 130), I think that the perceived morphological changes in the sub types through time were marked by a different style number. From Table 2 in the site report, it is apparent that on the whole, the lowest style numbers for a subtype occur in the Early Period and progressively higher numbers occur in the Middle and Late periods. The style numbers also indicate developmental changes in the funtional types that are not subdivided into subtypes. I infer the authors used the vessels in the 15 pairs of burials which cut into each other to establish the general sequence in which the functional types and subtypes changed. For example, vessels of one subtype from a pair of intrusive burials were used to define "earlier" and "later" character istics, for..that subtype. Vessels of that subtype from other burials were judged as "early" of "late", or "intermediate" in time. The intuitive comparison of the vessels from the - 63 -remaining burials with those from the set of 15 burials resulted in a complete sequence of development for each functional type and subtype, marked by appropriate style numbers. The authors then examined the entire set of vessels for each burial, concluding that three evolutionary stages were represented. Chapter 7 of the site report mentions that a clear difference between the sets of vessels from the Early and Late burials is noticable. But vessel forms from the Middle period burials are considered transitional. The authors imply they are not certain about the chronological position of some Middle period burials. Two of the burials that cut into each other (#'s 31 and 62) are assigned to the Early period, yet they do not contain any pottery vessels. The method by which these burials were assigned to the Early period is not clear. The other (15) burials in the cemetery with no pottery vessels are considered undatable. Chapter 7 of the site report includes a summary of the distribution of ware colors and of some production tool types per period. Four ware colors are said to have chronological significance: red wares are predominant in the Early period, grey wares increase in the Middle period, black wares increase in the Late period, and white ware is only present in the Late period. This general trend through time is substantiated in many Neolithic sites from Shandong and northern Jiangsu (Shangraw, personal communication 1983, Shangraw 1978:12 and - 64 '-Gao 1980:61). Many of the stone and bone tool types in the report are given style numbers as well. I infer that these styles were also defined on the basis of implicit, intuitive criteria. The participants of the first symposium on the Dawenkou culture in 1977 also identified a distinction between the ceramic forms in the Early, Middle and Late period burials at Dawenkou (Shandong Provincial Museum 1978:59). However, the Early and Middle periods defined in the Dawenkou site':.report are grouped together as part of one main evolutionary stage of the Dawenkou Culture and the Late period burials into another main stage (ibid:61). The two-period chronological scheme (Early and Late, with the former Middle period as part of the Early period) advocated in the recent Chinese archaeological literature appears to have been based upon an intuitive comparison of the ceramic forms at the Dawenkou site with the forms from other Dawenkou Culture sites that have relative chronological sequences based upon stratigraphy, such as the Xixaihou site (Wu Ruzuo 1982:267-269). An independent check of the changes in ceramic morphology from the Dawenkou burials is necessary in order to assess whether the two-period or three-period scheme should be accepted for this mortuary analysis. - 65 -3.3. Analysis 3.3.1. Introduction Two types of multivariate quantitative methods are employed here in an effort to test the consistency of the relative dating method just described based upon change in ceramic morphology. Multidimensional scaling and cluster analysis in a Q-mode are used to group those burials that co-vary in terms of the presence of absence of ceramic styles, subtypes, and functional types they contain. The purpose of the analysis is to determine whether an ordering of burials from Early to Late can be generated. If the results from the analyses show burials of the same chronological period (Early, Middle and Late) grouped together, it can be concluded that the relative dating method based upon changes in vessel form used by the authors of the site report is consistent. The results could indicate that the more recent two-period chronological scheme should be supported. In this case, the results would show Early and Middle burials grouped together and Late burials as distinct. The: results may show two or three periods composed of different groups of burials than in the site report, or more than three periods. If the results from the analysis do not show burials grouped together by chronological period at all, neither of the two chronological schemes nor any scheme can be accepted and the burials must be regarded as belonging to the same period. - 66 -Since there are no radiocarbon dates or consistent stratigraphic data for the burials excavated in 1959, I will not be able to demonstrate whether changes in styles, sub types or functional types actually represent changes in time. I can only propose that two or more chronological periods (provided an ordering of burials results) are represented at the site. The placement of the 15 pairs of intrusive burials within the ordering of burials resulting from the multi dimensional scaling and cluster analyses can server.as~an independent check of the ordering. If the intrusive burials are grouped as expected (i.e., with the burials in each pair in different groups representing different time periods), the chronological periods interpreted from the multivariate analyses can be accepted with greater certainty. The results from the multivariate analyses cannot be used to confirm the duration of the time periods proposed by Wu (1982). If I propose that two chronological periods are represented at Dawenkou, I cannot be certain the earlier period represents 600 years and the later, 400 years, as in Wu (1982:268). The analysis will not indicate the rate of change of the various ceramic forms. Also, the possibility of abandonment of the cemetery between or within chrono logical periods cannot be precluded. Finally, it is beyond the scope of this analysis to assess the reliability of the terms for the functional types and subtypes in the site report. Even though there may be some problems with these terms, they - 67 -will be accepted for the purposes of this analyses and for the remainder of this mortuary study. An evaluation of these terms would require an independent classification of the ceramic forms from Dawenkou based upon firsthand observation. 3.3.2. Method The specific techniques of scaling and clustering employed in this analysis and in Chapter 6 are Torgerson's metric multidimensional scaling and three variants of cluster analysis: Complete-Linkage or Farthest Neighbor, Average Linkage (Unweighted-Pair Group Method) and Ward's Method (Error Sum of Squares). The multidimensional scaling program used here is from Matson (1975). This program is based upon Torgerson's (1958) algorithm, the equivalent of Gower's (1966) Principle Coordinates Analysis. The computer program for the cluster analyses is from Wood (1974). The scaling and clustering analyses are based upon a similarity matrix which describes the similarity between all pairs of items in the data set. Similarity (transformed into a pseudo-distance measure) is calculated by means of Jaccard's Coefficient of Association (described by Sneath and Sokal 1973:131, Sokal and Sneath 196 3:129,133; and Doran and Hodson 1975:141). This coefficient is useful for presence/absence data because it does not group items on the basis of shared absences (True and Matson 1970:1201). In the chronological analysis, the burials are grouped together only if they/share - 68 -certain pottery styles, subtypes, or functional types. The matrix for the present analysis is based upon the presence or absence of 83 ceramic forms in 115 burials. Binary data is necessary for this analysis because the ceramic forms do not occur with great frequency among the 115 burials. As mentioned in Chapter 2, more than one multivariate quantitative technique should be used in conjunction with one another to ensure that the ordering of items obtained is an accurate reflection of the relationship between the items, not an artifact of one particular method (Matson and True 1974:72). Various techniques of multidemensional scaling and of cluster analysis have been found compatible for comparative purposes (e.g., Johnson 1972, Matson and True 1974. Matson 1974, True and Matson 1970, Pokotylo 1982, Pokotylo 1981, Peacock 1976). Multidimensional scaling can reveal important relations between clusters (Sokal and Sneath 1973: 252), The use of more than one technique of cluster analysis is desirable because different techniques can produce diff erent results (Matson and True 1974). In the chronological analysis, the orderings of burials from the three clustering techniques are compared with one . another. A good fit between the groups of burials resulting from each technique would allow acceptance of the groups with certainty. A good fit of the groupings from the cluster analyses with the pattern produced by the multidimensional scaling would allow even greater confidence in the results. - 69 -Torgerson's metric multidimensional scaling (Torgerson 1958) has been used successfully by Peacock (1976), Pokotylo (1982), Matson and Lipe (1977), and Matson and True (1974), among others. The following description of the technique is based upon the explanations: provided in the four works just cited. Multidimensional scaling depicts the relationship of items in n-dimensional geometric space.. The distance between items in space reflects the degree of similarity between the items. The similarity or distance matrix calculated by the coefficient of association chosen forms the basis of a matrix of the products of the distances from the configuration's origin or centroid. This matrix is factored, resulting in dimensions of decreasing order of importance. The first dimension accounts for the greatest amount of variability in the data set, the second dimension accounts for the second greatest amount, and so on. A measure is given which indicates the amount of variability explained by each dimension. If one's data set does not meet the metric assumption required for Torgerson's method (Matson and True 1974:70), Kruskal's (1964) non-metric method may be used. Among the advantages of metric mulitdimensional scaling over non-metric scaling are that unique solutions are derived, the solutions are invariant under changes in dimensionality, and less computer time and money is required (Matson and Lipe 1977:4). Cluster analysis is an agglomerative, hierarchical - 70 -method which depicts the grouping of items in the form of a dendrogram. Items are successively grouped according to decreasing degrees of similarity (or increasing degrees of pseudo-distance) until all of the items form one large group. Matson and True (1974) point out that each of the three techniques has its own advantages and disadvantages. Complete Linkage cluster analysis has more conservative rules for the addition of items to extant clusters than Average Linkage, and it avoids the disadvantage of chaining common to Average Linkage. Chaining is an "elongate growth of single linkage clusters" (Sneath and Sokal 1973:223). Chaining obscures the relationship between items intermediate or connecting to the clusters with chaining (ibid). Average Linkage has been a standard method in the archaeological literature (Matson and True 1974:61), and the Unweighted Pair-Group method has been most commonly used in general (Sneath and Sokal 1973:230). Complete Linkage is a "robust" technique because slight changes in coefficients will not create significantly different results (Matson and True 1974:61). But items may join clusters with difficulty, resulting in small clusters that leave out many less related items (Sneath and Sokal 1973:222-223, 226). Matson and True (1974:61) cite Ward's method as yielding results which were closest to their intuitively derived expectations, in comparison to the results from Complete Linkage and Average Linkage. A detailed description of the three clustering techniques - 71 -is provided by Sneath and Sokal (1972:222-241) and Matson and True (1974). The major difference between the three tech niques lies in the criterion which groups an item or a cluster with extant clusters (Matson and True 1974:54). In Complete Linkage Cluster Analysis, the criterion is the distance be tween an item and the farthest member of an extant cluster (Matson and True 1974:57). In the Average Linkage technique, the distance between an item and the average distance for the extant cluster is the criterion (ibid). In Ward's Method, the criterion is not a single distance but a measure of total interpoint distances within a cluster. An item is grouped with a cluster if it optimizes the sum of squared interpoint distances within that cluster (Matson and True 1974: 58) . A critical point regarding the use of multivariate statistical methods is that the interpretation of the results must be made by the researcher. An interpretation is not provided by the analytical results (Matson and True 1974:72). For multidimensional scaling, the archaelogical meaning of the dimensions which result must be interpreted. Also, the researcher must decide which dimensions best describe the variability in the data set. For any variant of cluster analysis, one must judge which clusters have archaeological meaning. Intuitively-based knowledge of the data set as well as some independent data not included in the multivariate - 72 -analysis are required so that an objective interpretation can be made of the results (Matson and True 1974:72). At present, the use of multivariate analyses in archaeology requires some subjective judgement on the part of the researcher. As Matson and True pointed out in 19 74, There seems to be no way to eliminate these (judgemental) choices at the present time, and the results are always subject to re-interpretation as new data are recovered or improved methodologies are developed (Matson and True 1974:72). Seriation utilizing artifact forms such as ceramics by means of multidimensional scaling has been attempted by a number of researchers (including many with burial data) since the late 1960's (Orton 1982:85). Authors who utilize multi dimensional scaling for seriation purposes include Kendall (1971), Peacock (1976), Drennan (1976a, 1976b), Spencer (1982), and Matson and Lipe (1977), but Kendall's (1971) is the only study in which graves are the items being seriated. Torgerson's metric scaling is employed by Peacock (19 76) and by Matson and Lipe (1977). Both studies utilize cluster analysis in conjunction with the scaling. Researchers have found that when an ordering of items results, it often takes the form of a horseshoe when the items are plotted in two dimensions (Orton 19 82:85). However, a perfect horseshoe shape is rarely achieved in practice (Orton 1982:86). Horseshoe shapes resulted from Torgerson's metric scaling in Peacock (1976) and in Matson and Lipe (1977). - 73 -Researchers must also be aware that the dimensions responsible for the ordering of items may not represent time and that interpretation of the ordering must be justified (Orton 1982:88). Unless the ordering is checked with inde pendent chronological data such as stratigraphy, the ordering must be considered hypothetical (Marquardt 1978:287). As stated previously, the placement of the 15 pairs of intrusive burials within the ordering from the scaling and clustering will serve as an independent, but not conclusive, check of the ordering. Mortuary studies in which other techniques of seriation are employed are those of MacDonald (1980) , Doran (1971), and Hodson (1977). Marquardt (1978) provides a detailed descrip tion of various seriation techniques. Seriation studies such as this one are based on the proposition that stylistic similarity of ceramics is related to social interaction and that similar forms of ceramics are close in time (Doran 1971:424). Although ethnoarchaeological studies support the notion that ceramic types found in the archaeological record reflect time change, more studies are needed to provide a better understanding of the process responsible for the formation of the archaeological record and the effect upon the ceramic types represented at a site (Marquardt 19 78:199-2 30). Archaeologists have recognized the advantage of using artifact types from "closed finds" such as graves for - 74 -seriation (Kendall 1971:215). 3.3.3. Data A total of 115 burials of the 133 in the cemetery are included in the chronological analysis: 71 of the 74 from the Early period, all 19 from the Middle and all 25 from the Late. These burials are listed in Figure A3-3. Four of these burials are disturbed (E27, M4.6, L3, L77) , but the pottery styles or classes remaining should indicate the relative chronological placement of the burials. Another burial, L10, is only slightly disturbed, if at all. It is described as containing a form of vessel (shuang bi hu, style I, "jar with two lugs") that may have originated from-.the burial into which L10 intrudes, E26. Since the origin of the jar is not clear, it is considered here to be from L10. L10 contains a large number of vessels and the questionable placement of one should not change its position in the chronological ordering or its relative status position. Seventeen burials with no pottery vessels are excluded. Fifteen of these, are the burials considered as undatable by the authors of the site report. The other two are E31 and E62, the intrusive burials with no pottery vessels assigned to the Early period. Burial E108, assigned to the Early period, must be excluded as well. For reasons discussed below, the one type of vessel found in this burial cannot be included in the analysis. - 75 -A total of 83 mutually exclusive ceramic categories from the total of 176 known forms given in Figure A3-1 (excluding the unique animal shaped vessel and the uncertain forms - dou zuo and dou pan) are included in the analysis. A list of these 83 forms is provided in Figure A3-4. These forms include styles, subtypes not divided into styles, and functional types not divided into subtypes or styles. Unfortunately, it is not possible to include all 176 ceramic forms in the analysis. Study of the distribution of the 176 ceramic forms among the 116 burials (including E108 here) indicates that a . number of forms occur very infrequently. Figure A3-5 shows that 61 forms (34.7%) are present only once among the burials and 32 (18.2%) are present only twice. The forms begin to occur with greater frequency thereafter. From the graph in Figure A3-5, I judged that a ceramic form should be present at least three times among the burial sample in order for the multivariate analyses to be viable. The above decision required that the forms occurring only once or twice across the burial sample be either excluded from the analysis or grouped with other forms. If over fifty percent (52.9%) of the forms were excluded, the analysis could not adequately test the consistency of the relative dating method based upon morphological similarity. Therefore, an effort was made to group ceramic forms that are morphologically similar. Due to my uncertainty with the cai ("painted") subtypes - 76 -of vessels noted earlier, all 16 of these categories shown in Figure A3-1 are excluded from the lumping procedure and from the analysis in general. Since the only vessel in E108 is painted, this burial must be excluded from the analysis. Morphological similarity of the vessel forms was based upon the descriptions and drawings in Chapter 5 of the site report and the representative photographs. Chapter 5 describes certain parts of vessels such as rims, bodies, and bases as gradually changing form. Detailed study of the vessel forms tended to support my proposition that the forms closest in style number are those closest in style. For example, a vessel with a style of I appears more similar in form to a vessel (of the same subtype) with a style of II than a style of V. A decision was made to group a vessel form with a form having the nearest style number. For example, if a vessel with a style number of II had to be lumped, it was lumped with the vessels of style number I or III, depending on my judge ment of morphological similarity. As many original groups of vessels as possible were maintained. A subtype not divided into style numbers was not lumped with another subtype. If a vessel form did not in my judgement appear morphologically similar to the vessels in either adjacent style category, it was excluded. The qf ta or "other" subtypes and gai (pot lid) functional type have more styles excluded than other categories..: It.was necessary for a few entire categories to - 77 -be left out, such as "pen ("basin") and gui wu zu ("jug with no feet"). I acknowledge my limited understanding of the Chinese classification procedure. The site report only describes and pictures a representative group of vessels for each subtype and style. The range of variation within the style and sub type categories is not documented in the site report. Also, the written descriptions can be confusing to a western reader. For example, a translation of "bird-shaped beak" refers to a spout on a vessel and a "nose" refers to a lug. Although the incorrect lumping of styles may distort the resulting ordering somewhat by making the ordering less clear, it should, not result in a spurious ordering (R.G. Matson, personal communication 1983). There are some discrepancies for the figure given in various places in the site report for the total number of vessels in certain functional types, subtypes, or styles. The totals for some categories given in Chapter 5 of the site report and in Table 12, page 135, differ with my totals calculated from the descriptions of the burials. In each case, I accepted the totals from the descriptions of the burials After the lumping procedure was completed, another graph was made that depicts the frequency of occurrence of the ceramic categories in the burial sample. Figure A3-6 shows a more even distribution of the categories: 18.1% of - 78 -the categories occur three times among the 115 burials (excluding E108 now), 14.5% occur four times, 10.8% occur five times, and 16.9% occur six:times. Figure A3-7 shows that 13.9% of the burials have only one ceramic category present after lumping, 15.7% have two, 9.6% have three, and so on. A smaller percentage of burials having only one or two ceramic categories would have been more desirable. How ever, it was thought that the multivariate analyses would be viable. The 83 forms resulting from the lumping consist of 112 of the original 176 formal categories. Sixty-four forms could not be lumped and are excluded from the analysis. 3.3.4. Results The orderings of burials resulting from the multidimen sional scaling and three variants of cluster analysis show a clear distinction between the Early and Late burials, with some Middle burials grouped with the Early burials and others with the Late. This patterning is that described in the site report. Therefore, I conclude the changes in ceramic morph ology were assessed in a consistent manner to indicate time change. However, the orderings from the scaling and clustering support neither the three-period chronological scheme advocated in the site report nor the two-period scheme proposed in the recent Chinese 'archaeological literature, The orderings indicate a variation of the two-period chronological scheme. - 79 -They depict a two-period division of Early and Late, with some former Middle period burials placed in the Early period and some in the Late. This scheme differs from the two-period scheme in which the former Middle period is placed entirely within the Early period. The orderings produced from the scaling and clustering techniques are described below. Each of the three clustering techniques shows the clear division, between Early and Late period.burials. However, each shows a different ordering of burials within the Early period and within the Late. Some of the differences seem due to the inherent nature of each clustering technique.. Others may be due to the nature of the data set used in the analysis. The ordering of burials from Ward's Method is chosen for the comparison with the multi dimensional scaling results. The placement of the intrusive pairs of burials within the orderings from the scaling and Ward's Method provides independent support that the orderings represent chronological relationships among the burials. 3.3.5 Multidimensional Scaling Five dimensions were generated, accounting for 29.56% of the- variability within the data set. This low percentage may be due to the low number of ceramic categories within many of the burials, and thus the low number of ceramic categories shared between many burials. Figure A3-7 shows that 29.64% of the burials only had one to two ceramic - 80 -categories present. The first dimension accounts for the greatest amount of variability, 9.175%, with the second and third dimensions .accounting for only 5.597% and 5.748%, respectively.- In theory, the dimensions from Torgerson's metric multidimensional scaling method should account for decreasing amounts of variability. The eigenvalues for the first three dimensions are 4.782, 2.917, and 2.996. Slight deviations often result in association with small eigenvalues (R.G. Matson, personal communication 1983). One inequality out of 246,905 was violated in the program; a low proportion of 0.0000405, indicating that this data set is indeed metric. Figure 3-1 shows the ordering of the 115 burials along Dimensions 1 and 2. Dimension 1 is interpreted as representing time. The burials are divided into an Early sector on the left side of the axis and a Late sector on the right. The dividing point is at approximately 0.140 on the X axis. A horseshoe shape is not clearly discernable. The former Early period burials and some Middle period burials tend to be scattered while the former Late period burials are in a more restricted area, suggesting one half of a horseshoe. This difference in spatial positioning and the lack of a good horseshoe shape may be caused by differences in the number of ceramic categories per burial within each of the three periods. The Early period burials account for most of the burials with one to two ceramic categories shown in Figure A3-7-(12 of the 16 having one category and 16 of the 18 burials - 81 -FIGURE 3-1. The 115 burials in the chronological analysis from Torgerson's Metric Multidimensional Scaling, Dimensions 1 and 2, clusters from Ward's Method. E82A EI32# Ell , E34 E65« E13I EI29//' El 10 EI3 "3A DIM 2 E5I + AEI0?AE56 A .EI06 . M2I £23* EI03„ EI09 EI8 , E7 E32< E,llO E80 A" //A El 19 * EII5 OE73 Qa E59 A ,EII4 C- EI20 E99u E58 E43" E4I.E89 EI2° E8° E87° E810 E66 E76o E530 E48 A A EI02 E54* EII21 »E6 E29^ E45 * O EBI El 16 E9Ia E28 ^E33* E94 M49 'EI30 M42 E86 E88 E90 A Ml 18 E55 E6l"- £ AM36 A jL77 ,E27 LI L60 + +LI7,L25 £26^" „ ° L72 j D • 'LI05-EI0I M69 +3.0 CL6*~T^ DIM, E30" LI23*> °LI" L2U • „ DMI6* D • . L4 a a LI26 ,0 M96b L47 M9* -2.0 M93 M46" LI04 L3 LI00 / °U24* D LI25 • M75 o M22 M35 Clusters I 2 3 4 5 6 7 8 E Early Burioi M Middle Burioi L Late Burial intruding burial - 82 -having two). Thus, Early period burials are either very similar or very different, not allowing a continuous horse shoe shape. The plot of burials also shows a few burials located at the same point in geometric space. These burials are identical in composition of ceramic categories. The placement of the pairs of intruding burials supports the interpretation that Dimension 1 represents time. The Late period burials that intrude into Early burials (see Figure A3-2); L10, L15, and L24, are located some distance along Dimension 1 from the burials they cut into: E26, E33, and E30, respectively. Four of the intruding burials were excluded from the analysis because they do not contain pottery: E31, E62 and the undatable burials X70 and X133. E18 is not marked on the plot because its intrusive pair (E31) was not in the analysis. This is true for E71 as well (having the undatable burial 70 as .its pair). There is less spatial' separation of the intruding pairs of burials from the same former time period: E54 and E58, L123 and L124, E32 and E61, and E78 and E129. Three of the four former Middle period burials that cut into Early period burials are spatially separated from the latter burials: M9 and E23, M16 and E61, and M121 and E132. M44 and E43 are barely separated along Dimension 1, and M4 4 is closer to the majority of the Early period burials than E4 3. The locations of these two burials do not provide additional support for the ordering derived from the scaling, but they make sense - 83 -when M44 is considered an Early period burial. It is located well to the left of the 0.140 dividing line on the X axis. Because E30 and E61 have only one ceramic category present, their locations on the scaling plot are not as secure as that of the other intruding burials. Since all the other intruding burials contain more than two categories, their locations can be relied upon. It is possible that Dimension 2 represents time as well as Dimension 1. For three of the five pairs of Early period intrusive burials in Figure 3-1, the older burials (E43, E58, E12 9) are located in the upper left quadrant and the younger burials (M44, E54, E78) are located in the lower left quadrant. All the Early burials in the lower left quadrant may be younger than the Early burials in the upper left quadrant. However, the placement of the burials in the one Late period intrusive pair does not support the possibility that Dimension 2 represents time, with older burials in the upper section of Figure 3-1 and progressively younger burials towards the lower section of Figure 3-1. The older burial, L124, is located below the younger burial, L12 3. Due to the fact that the ordering of burials from Early to Late is also: exhibited in dendrograms of the cluster analyses,- as well as the lower percentage of variability accounted for by Dimensions 3 and 4, Dimensions 3 and 4 are not interpreted here. - 84 -3.3.6. Cluster Analyses The ordering from the Complete Linkage technique compares poorly with those from Average Linkage and Ward's. Small groups of burials join together only at high distance figures. This problem is common with Complete Linkage due to the conservative rules of joining items to extant clusters. While Early burials are in different clusters than Late burials, all of the Early clusters and Late clusters do not eventually join into one large cluster each. A macrocluster of Early burials and one of Late burials (with Middle burials inter spersed) is apparent in the dendrograms from both Average Linkage and Ward's Method. The dendrograms from Average Linkage and Ward's Method agree fairly well. It is apparent that this agreement reflects relationships among burials. The orderings of burials within the Early period from Average Linkage and Ward's Method compare poorly but the orderings of burials within the Late period compare quite well. This difference may be due to the greater number of ceramic styles, subtypes and functional types per burial within the Late period. The Early period clusters may not be as robust as the Late clusters because they contain burials that have few ceramic categories in common. Some chaining is evident in the Average Linkage dendrogram, but not an excessive amount. The dendro grams from both methods place the 19 Middle period burials in the same manner: eight burials in the Early period and 11 - 85 -in the Late. The ordering from Ward's Method is chosen to: compare with the multidimensional scaling results. Ward's Method has been found to yield results with the same data set that are more similar to intuitively derived expectations than those from Average Linkage and Complete Linkage (Matson and True 1974:61). The results from Ward's Method are also favored over those from Average Linkage and Complete Linkage by Peebles (1974:100). The dendrogram for Ward's Method is shown in Figure 3-2. The extremely clear break between the Early and Late burials is apparent. Figure 3-2 indicates that no Early and Late burials are in the same cluster. Also, all of the Early period burial clusters join together, and all of the Late join together. The Early and Late period macroclusters do not join until the 4.3578 distance level, indicating the clear dissimilarity of the Early and Late period burials in terms of ceramic categories. While a number of solutions could be compared with the multidimensional scaling plot, the Eight Group Solution at 2.0023 distance is chosen as a man^-ageble, interpretable one. The ceramic categories represented in each cluster are given in Figure A3-8. The placement of the intruding pairs of burials in the clusters (each burial in a pair of different clusters, at varying distance levels) is further support that the ordering from Ward's Method reflects chronology. These burials are marked in Figure 3-2. The pairs from the former Early and H G s w fD 3" i O NJ & • O hi o 3* o i—• c rt fD H > H O O O <JQ H-O cn a w m • ^< w H-cn Dj fD 3 Oi O i£« H OJ 3 Hi H O 3 (D H 00 - 87 -Late periods are widely separated: L10 and E26, L15 and E33, L24 and E30. Intruding burials of the same former period are not as widely separated. They also are in different clusters, but at lower distance values: E54 and E58, L123 and L124, E32 and E61, E78 and E129. Unlike the patterning in the scaling plot, M44 is separated from E43. The other Middle and Early burial pairs are separated, too: M9 and E2 3, M16 and E61, M121 and E132. The placement of the five disturbed burials is clear in both the Ward's dendrogram and in the scaling plot (E27 and M46 are with Early period burials and L3, L10 and L77 with Late period burials). The locations of the eight clusters from Ward's Method on the scaling plot (see Figure 3-1) depict very good agree ment for the Late period burials and a fairly good agreement with the Early ones. As noted earlier, the low ceramic content of the Early burials may be responsible for the more flexible group composition. The agreement between the multi dimensional scaling and cluster analysis results noted above, the extremely clear distinction between the former Early and Late burials by both methods, and the placement of the Late-Early intruding burials in the results of both methods allow confidence in accepting an Early-Late chronological division for the cemetery. During my initial interpretation of the clustering and scaling results, I thought there was a discrepancy in the placement of five of the 19 former Middle period burials - 88 -(M67, M9, M35, M118, M96). The Ward's dendrogram places M67, M9, M35 and M118 within the new Late period (in Cluster 8) and M96 within the new Early period (in Cluster 4). I thought the multidimensional scaling plot showed the four former burials with the Early burials because these four burials are located to the left of the 0.140 dividing point along Dimension 1. I did not realize that if the position of these four burials within the partial horseshoe in Figure 3-1 is considered, these four burials are indicated as Late by the multidimensional scaling results as well. The discrepancy with the placement of M96 remains because M96 is clearly near the Late burials within the horseshoe. Believing that there was a discrepancy in the placement of the five burials, I chose to accept the ordering of burials from the multidimensional scaling for the remainder of this mortuary analysis. I considered M67, M9, M35, and M118 as Early period burials and: M96 as Late. I judged the ordering from the scaling to be the better indicator of chronological relationships among burials because multidimensional scaling depicts relationships between items in terms of trends instead of discrete groups like cluster analysis. I realize now that there is no discrepancy in the chronological placement of M6 7, M9, M35, and M118 between the clustering and scaling results. I should have considered these four burials as belonging to the Late period. Also, I probably should have accepted the position of M96 from Ward's dendrogram instead of from the - 89 -scaling plot. Cluster analysis depicts relationships among items in terms of more variability within a data set than multidimensional scaling, which separates the major dimensions of variability. In Chapters 5 and 6 I have noted when my interpretation should be changed due to the original incorrect chronological placement of M67, M9, M35, and M96. 3.4. Conclusion The results from the chronological analysis allow two chronological periods to be accepted with confidence, an Early period and a Late period. The ordering of burials on the basis of changing ceramic style by the authors of the site report is clearly supported by the analysis. The results show that some former Middle period burials are grouped with Late burials and others with Early burials. The transitional nature of the ceramic styles in the Middle period burials is noted by the authors of the site report. Figure 3-3 lists the burials in the newly derived Early and Late period that are utilized for the remainder of this study, as well as the undatable burials. Figure 3-4 shows the locations of the Early, Late and undatable burials-within the cemetery. The figure indicates that all spatial locations within the cemetery were used in both periods. Three burials assigned to the Early period by the authors of the site report, 31, 62, (both intrusive) and 108 (with one painted vessel) could not be included in the chronological - 90 -FIGURE 3-3. List of burials in the newly assigned Early and Late periods,derived from chronological analysis, as well as undatable burials. Early: 85 total 6, 7, 8, 9, 11, 12, 13, 14, 18, 19, 20, 21, 23, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 38, 41, 42, 43, 44, 45, 48, 49, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 65, 66, 67, 69, 71, 73, 76, 78, 79, 80, 81, 82, 84, 86, 87, 88, 89, 90, 91, 94, 97, 99, 101, 102, 103, 106, 107, 108, 110, 111, 112, 114, 115, 116, 118, 119, 120, 129, 130, 131, 132. Former Middle period burials: 9, 21, 35, 36, 42, 44, 49, 67, 69, 97, 118. Late: 33 total 1, 2, 3, 4, 5, 10, 15, 16, 17, 22, 24, 25, 46, 47, 60, 64, 72, 75, 77, 93, 96, 98, 100, 104, 105, 117, 121, 122, 123, 124, 125, 126, 127. Former Middle period burials: 16, 22, 46, 75, 93, 96, 98, 121. Undatable: 15 total 37, 39, 40, 50, 57, 68, 70, 74, 83, 85, 92, 95, 113, 128, 133. - 91 -FIGURE 3-4. Location of burials within the Danwenkou cemetery from each period as defined by the chronological analysis. 64 63. n tatty JlC^-1 70 10 48 49 67 5C 39 41 57 36 "C3 ^ 42 <=J\ 53 45 .56 52 -,65 384 El • 33 79 • 80 C3 C3 81 82 40 I 128 130 131 ^3 46 e=a5 38^2. ^8 'a 2 9 ^ I5,=>=j33 / \ pottery kiln 21 3 680C3 73 66 17 27 r 35. 6C 'CZTQ 25 £J i4^2^50 Q 85 ^« 76 87^86^9 ,29 ^8 90 10?CD*°B_!09 127" »t=> 78^3^94^^ B5 S2 a Oa r-riii a7a 102 ri?J «E3 - - '23 , 124 S-JI05 e£TC» iota f->» io«r-i ,22<J cj125 113 a? <^ 100121/^133 C>I5 ^,32 JOII4 92 10 m I • • i • J CD EARLY PERIOD BURIAL HI LATE PERIOD BURIAL UNDATABLE BURIAL (adapted from The Shandong Provincial Cultural Properties Cornnission and The Jinan City Museum 1974:4) - 92 -analysis. These burials are accepted as belonging to the Early period for the remainder of the study. I conclude the vessel in burial 108 must have been examined by the method of stylistic comparison which the multivariate analyses show as consistent. It is likely the vessels in the burials (18 and 33) that cut into burials 31 and 62 were thought to resemble Early period styles. Since burials 31 and 62 are older than 18 and 33, burials 31 and 62 were thought to be Early. The fact that several burials within the accepted Early and Late periods are cutting into each other suggest each period is quite lengthy. It is not likely that mourners would deliberately disturb an ancestor's remains when digging a grave for a deceased person. It appears that no surface features mark the graves. However, intensive land use or use of the site during later cultural periods may have destroyed surface features. There may have been such a great time gap between the burials in question above that the locations of ancestor's graves were forgotten.. The fact that some people in the Late period were buried before the use of the pottery kiln in the cemetery and others were buried later (mentioned in Chapter 1) also suggests a lengthy Late period. The lengths of the periods proposed in the Chinese archaelogical lit erature (600 and 400 years for the Early and Late periods, respectively) do not seem unlikely. The only known chronological relationships of burials within each of the newly derived periods are indicated by the - 93 -pairs of intrusive burials within each period. Four pairs of intrusive burials in the new Early period (E32 and E61, M44 and M43, E54 and E58, E78 and E129) and one in the new Late period (L123! and L124) were included in the chronological analysis. These pairs are located in a variety of spatial areas within the cemetery, suggesting that all spatial areas were used continuously during each period. It does not appear that different spatial areas represent different periods of time as noted by Shennan (1975:280) for the European Neolithic cemetery of Branc and Chapman and Randsborg (1981:15) for a modern rural Hungarian cemetery. Due to the lack of infor mation on the chronological relationships of burials within each period, emphasis is placed on change through time from j the Early period to the Late for the remainder of this study. The distribution of ceramic ware colors within the newly derived Early period and Late period is comparable to the distribution for Dawenkou Culture sites mentioned previously. The distribution is additional support for accepting the newly assigned periods. My calculations indicate the dist ribution as follows: Early period: red ware is 54% of the total, grey 31%, black 10%, painted 5%; Late period: red ware is 12%, grey 41%, black 14%, white 31%, painted 1%. The distribution of ware color will be discussed in more detail in Chapter 6. The chronological analysis also indicates that some ceramic classes and forms are temporally sensitive. - 94 -The following forms are present only in the newly assigned Early period (see Figure A3-1): ding cai ("painted tripod"), dou guan shi pan ("serving stand with a guan-type dish"), dou da lou kong ("serving stang with large cut out holes"), dou cai ("painted serving stand"), guan cai ("painted jar"), he cai ("painted spouted vessel"), kui xing qi ("helmet... shaped vessel"), zuo ("stand"), and bo cai ("painted bowl"). These forms occur only in the Late period: bei dan ba ("cup with simple handle"), dou shuang ceng pan ("serving stand with double-layered dish"), hu kuan jian ("storage vessel with wide shoulders"), ping ("bottle"), gui kong zu ("hollow footed tripod pitcher"), and dou tong xing ("tubular shaped serving stand"). - 95 -CHAPTER 4 , ANALYSIS TO ESTIMATE SEX 4.1. The Problem Chapter 2 Outlines the methodological problem created by the low proportion of sexed to unsexed burials in the burial population. A key test implication for achieved versus ascribed status is whether status differentiation is based upon age and sex or if it cross-cuts age and sex categories. With only 19 out of 80 sexed single burials in the Early period, nine of 32 single sexed burials in the Late Period, and two out of 15 in the undatable group of burials, it would be difficult to test whether status differ entiation is based upon sex. A greater proportion of sexed burials would also enhance discussion of social subgroup affilitation represented in the cemetery. The burials confidently estimated as male or female are added to the sample of known males and females for the analyses in Chapters 5 and 6. In this analysis, two multivariate classification methods are employed: a simple inspection method and discriminant analysis. These methods are described in detail in the next section. Each method classifies the unsexed adult burials into a male or female class on the basis of the artifact types included in the known male and female - 96 -graves. An exploratory study of the distribution of artifact types within the known sexed graves suggested that some types of artifacts (especially production tools) occur exclusively or nearly so with males ,and others with females. This suggestion has been made for Dawenkou Culture sites by Pearson (1981:1084), the Shandong Prov incial Museum (1978), Luo and Zhang (1979), the Shandong Archaeological Team (1979) and Zhang (1979). Pearson (1981) includes a comprehensive summary of the distribution of sex-related artifact types at Dawenkou Culture sites. Binford's (1971) ethnographic study indicates that in mortuary ritual, sex is commonly expressed by artifact form (see Figure 2-3). These artifact forms may be clothing, forms that express the personality of an individual, or tools that reflect the male-female division of labour (Binford 1971:22). Only unsexed adult burials are included in this analysis. Since 29 of the 30 sexed burials are adults, the artifacts associated with these burials may be a reflection of relative (adult) age as well as sex. Multiple burials are excluded from this analysis because it is difficult to identify the artifacts associated with each individual. Sexed and unsexed single adult burials from the undatable group of burials are included in the analysis. Binford's (1971) study indicates that orientation o.f the grave may reflect the sex of an individual as well. The variability in grave orientation at Dawenkou is discussed in Chapter 5. The small number of aged skeletons at Dawenkou also limits the effectiveness of the analysis of status differ entiation. There are only four known subadults in the Early period and one in the Late. Six are in the undatable group of.burials. The other skeletons are described as adult, with no differentiation as to relative age. According to Binford's (1971) ethnographic study, age in mortuary ritual may be reflected by grave location, grave form and body disposition (see Fig.2-3). Variability in these three aspects of mortuary treatment is also discussed in Chapter 5. Whether relative age is also reflected by certain artifact forms at Dawenkou cannot be tested here due to the small sample of burials of known relative age and the paucity of grave goods in the majority of the subadult graves. This analysis assumes that the sexing of the Dawenkou skeletons was done accurately by the investigators. It is not clear why only 30 skeletons from single burials and eight skeletons from multiple burials were sexed. The investigators may have encountered fragmentation of critical skeletal parts or poorly developed suxual characteristics, problems which have affected other cases of sexing (Bender 1979:186). Since some of the sexed skeletons have only a few common artifacts present, it seems the sexing was done on the basis of skeletal characteristics and not on artifact types. It is not likelyithat in 1959 investigators utilized a multi variate sexing method, preferred today for its accuracy over the method of applying single criteria to each skeleton - 98 -(Weiss 1973:58). But for the purposes of this mortuary study, it is assumed the method employed is adequate. That subadults were not sexed is not unusual - other paleodemographic analyses have not been able to sex sub-adults due to inadequate development of sexual characteristics (Bender 1979:186). It is not known if the group of sexed skeletons from Dawenkou has the common problem of a bias towards males (Weiss 1973:58). A bias may not have occurred because approximately equal numbers of males and females were identified (14 males and 16 females from the single burials and four males and four females from the multiple burials). Since males comprise approximately fifty percent of most ethnographic adult populations, a sexed skeletal population should have a similar male-female ratio, unless cultural factors are responsible (ibid). Cultural factors that could upset this ratio include cases in which males died in other territories (Bender 1979:187). It is possible the known male-female ratio at Dawenkou may not be representative of the male-female ratio in the population as a whole. It is also not clear why the relative ages of the adult skeletons were not determined. The reason may be the inherent difficulty and uncertainty in the aging of skeletons (Weiss 1979:59). For the purposes of this mortuary study, it is assumed that the majority of subadults in the cemetery were reliably identified. Four burials in the cemetery do not contain a skeleton (one in the Early period and three in the Late). The authors of the site report suggest that these burials were - 99 -prepared for people who died in another territory, a practice common to modern fishing peoples (Shandong Provincial Culture Properties Commission and the Jinan City Museum 1974:7). Since these burials contain grave goods, they will be included in the analysis to estimate sex. 4.2. The Analysis 4.2.1. Introduction None of the extant mortuary studies in the literature has employed a simple inspection method or discriminant analysis to estimate sex. The two studies in which an attempt is made to estimate sex on the basis of artifact inclusions (Hodson 1977; Shephard 1979) do not make explicit use of the patterning of artifacts in known sexed burials. Hodson (1977) utilizes Single Linkage Cluster Analysis in an R mode, and Shephard (1979) utilized Single Linkage Cluster Analysis and principal coordin ates analysis in an R mode. Both studies interpret the "resultant groups of artifacts as either male-related or female-related. The interpretations seem made on the basis of modern western concepts of male and female artifacts. It is not clear whether the patterning of artifacts from known male and female burials contributes to the interpretations. Doran (1973) mentions that sex and age will be estimated in a future study for a sample of European Iron Age burials, but apparently the results have not yet been published. - 100 -Hodson (1979:25) mentions four studies in German that attempt to estimate sex on the basis of artifact inclusions but does not describe the methods employed. Since the exploratory study of the Dawenkou sexed burials suggest specific male and female related artifacts, a multi variate technique that can best make use of the known sexed group of burials should be useful. The simple inspection method and discriminant analysis are two such techniques. Of course, Single Linkage Cluster Analysis and principal coordinates analysis may work for other data sets. 4.2.2. Method The estimations of sex are based upon a comparison of results from the simple inspection method and the more rigorous method of discriminant analysis. Discriminant analysis is theoretically well suited for a classification problem such as the one here. However, the simple inspection method from Sneath and Sokal (1973:406) seems equally suitable. The simple inspection method is utilized to classify the unsexed single adult burials on the basis of a few artifact types that were noted in the exploratory study as exclusive to the known males or females. The simple inspection method could not identi fy all the male or female-lfnked artifact forms. Therefore, discriminant analysis is utilized to classify the unsexed burials on the basis of all the artifact forms present in the known sexed - 101 -burials. Discriminant analysis can search the artifact forms known as exclusive to either sex, the forms noted as shared between the sexes and the forms too numerous for study by visual inspection to determine the artifact forms that best discriminate between the known sexed burials. However, discriminant analysis is not always more effective than simple inspection methods in classifying items (Sneath and Sokal 1973:406). It has been argued (Thomas 1980:344) that a simple statistical method should be chosen over a more complicated one when it is known that both methods are equally capable of achieving the desired results. In cases such as this one, it is not clear that the simpler method will be as effective as the more complex method. I argue that a. comparison of results from both a simple and complex method should indicate the relative effectiveness of each method for the particular data set utilized. A comparison should also identify the relationships between items in the data set, not relation ships which are a factor of one method. In the simple inspection method, ideal types are derived from the variables that distinguish the known classes (in this case, the known male and female burials). Each ideal type consists of a set of sex-linked artifact types. The set of sex-linked artifacts, on the basis of the artifact types present in the known male and female burials (see Figure 4-1), is used to calculate the similarity of the unsexed burials to - 102 -FIGURE 4-1. Distribution of sex-linked artifact types Artifact whetstone llshi fish hook gdu tusk knife ya' dao stone knife shf dao disk, pendant bi, zhiii stone chisel shx zab bone chisel gu zafo arrowhead zu spoon, spatula bi head ornament, neck ornament toushi, j ingstii Known male burials Known female burials (14 total) (16 total) small round stone xiao shi bing - 103 -the ideal types. This calculation is by means of Jaccard's Coefficient, described in Chapter 3. The taxonomic distance of each unsexed burial from each ideal type is based on the sex-linked artifact types. A burial is estimated to be male if the value for it is closer to the "male" ideal type or female if it comes closer to the "female" ideal type. Discriminant analysis classifies items on the basis of a set of independent variables into one of two or more mutually exclusive categories or classes (Morrison 1969:442). In this case, the analysis classifies the unsexed burials into a male or female class on the basis of a discriminant function. The discriminant function is calculated from the pooled variances and covariances between the variables from each known class (Sneath and Sokal 1973:401). Here,, the discrim inant function is calculated from the quantities of all the artifact; types present in the known male and female classes of burials. The discriminant funtion is a statistical statemehtcof :-thervaf tables:.:found ;to:"distingUishr between vthe established groups in question (Hettinger 1979:456). The discriminating variables are weighted and linearly combined so that the classes are as statistically distinct ..as possible (Klecka 1975:435). The centroid for each class represents the average of the discriminant scores for the variables, or the average location of the variables in each class in the discriminant function (Klecka 1975:443). The discriminant scores for each case (here, the unsexed single adult burials) - 104 -reflect the probability for membership into one of the classes. The Direct Method of Discriminant Analysis from the SPSS Subprogram DISCRIMINANT (Klecka 1975) is utilized here. In the Direct Method, all independent variables are entered at once and the discriminant function (one in the case of two known classes) is derived from the set of variables. Unlike the Step-wise method, the discriminating power of each inde pendent variable is not considered. In this version of discriminant analysis, a "tolerance" test is mandatory, resulting in the rejection of perfectly correlated variables. This step may be statistically justified, but it seems unjustified in the present case. 4.2.3. The Simple Inspection Method: Data Thirty sexed burials (19 from the Early period, nine from the Late and two undatable burials) and 32 unsexed adult burials are included in the analysis by the simple inspection method. Figure A4-1 lists the sexed and unsexed burials for each period and for the undatable burials. The small sample size of known male and female burials in each of the two periods precluded a separate analysis for each period. In the interpretation section, consideration is taken of the two temporally sensitive artifact types included in the analysis. Eleven artifact types are included: nine that are exclusive to the 14 known males.(but not present in all male burials) and two that are exclusive to the 16 known females (but not present among all females). A few types are items of clothing but most are production tools. The distribution - 105 -of these 11 types is shown in Figure 4-1. Some of these artifact types have style numbers assigned to them. Style numbers are not included in the simple inspection method because the distribution of the various numbers according to sex is too difficult to ascertain by eye. Ceramic styles are excluded from this analysis for the same reason. Some artifact types have the same form but are made of a different material (stone and bone chisel, and tusk and stone knife). Since these types are purposely separated by the authors of the site report and they are exclusive to the known males, all are utilized in this analysis, Ceramic subtypes and functional types were not found to be exclusive to either sex except for a few forms that are rare in the burial population. Only 32 of the 85 unsexed single adult burials in the cemetery have one or more of these artifact types present. These burials include: one undatable burial, 19 Early period burials (one of which has no skeleton), and 12 Late period burials (of which two have no skeleton present), The artifact types do not occur with great enough frequency across the burial sample to allow the use of frequency data. Two of the disturbed burials, E27 and L3, are included in the analysis. All of the artifact types from the male subadult burial, L117, are included in the analysis because they occur with male adult burials. Thus, these types are not a function of age. I judged that the artifact types should be present at least four times among the burial population in order for the - 106 -analysis to be viable. The jiao zhui and bl are lumped together in order to meet this requirement, as well as the neck and head ornaments (jing shi and tou shi). A known female has both the head and neck ornament types. One type exclusive to males must be rejected (stone hammer, chui) as well as two types exclusive to females (elephant ivory comb, xiang ya shu and stone pendant, huarig) . At the time of the analysis, on the basis of the photographs and drawings in the site report I interpreted jiao zhui to indicate a horn pendant. I have learned since then that the term may refer to a net weight (Zou Heng, personal communication 1983). Therefore, the jiao zhui probably should not have been grouped with the bi (elephant ivory disk). However, since there are only two cases in which bl are present in the burial sample (both from the Early period), the results from the analysis should be reliable. The uneven distribution of male and female . artifacts limits the ability of the analysis to identify unsexed burials as female. Another limitation is that the absence of the nine "male" artifacts does not necessarily indicate a female burial because six of the 14 known male burials do not contain these artifact types, either. However, because the proportion of known females without these types is much higher that the known males without these artifact types (16 female burials out of 22 total with no "male" art^ ifact types/ versus six male burials out of 22 total), it is - 107 -more likely that a burial with none of these artifact types is female. Therefore the variable "no male artifact types" is included as a twelfth variable in the analysis. The last two cases included in the analysis are "ideal female" and "ideal male". The "ideal female" is coded as having the nine male artifact types absent and the three female artifact categories present (the two given in Figure 4-1 and the category "no male artifact types"). The "ideal male" is coded as having the nine male artifact types present and the three female catagories absent. Coding of the 32 unsexed burials showed a distribution of the nine male artifact types and two female ones which supported the proposition that at least some of the 11 arti fact types are sex-linked. The unsexed burials tend to have either some of the nine male artifact types present and none of the female artifact types, or vice versa. 4.2.4. The Simple Inspection Method: Results The similarity matrix generated from Jaccard's Coefficient is not difficult to interpret because the unsexed burials are clearly similar to either the "ideal male" type or to the "ideal female" type (with similarity expressed by a low dis tance value and dissimilarity expressed by a value of 1.000). None of the unsexed burials have values close to both the "ideal male" and the "ideal female". The distance values for the sexed and unsexed burials as well as the assignment - 108 -of "male" or "female" for the unsexed burials are given in Figure 4-2. A total of 26 unsexed burials appear to be male and six, female. The assignment of the six "female" burials is not as reliable as that of the 26 "male" burials. Four of the "females" (E49, E58, L5 and L60) only contain one female artifact, the small round stone (xiao shi bing). This artifact type is only found in one known female grave. Burials L3 and L47 are more reliably "female" since they contain an artifact type that is found: in three known female graves, the neck or head ornament (jing shi or tou shi). This artifact type is uncommon in the burial population and is restricted to the Late period. It may be a high status artifact type (this possibility is discussed in Chapter 6). Thus, there may be no artifact type to reliably identify low status females in the Late period or any females in the Early period. As discussed previously, the lack of the male related artifact types does not necessarily distinguish a female burial. The artifact types present in one of the unsexed burials suggests that the small round stones may not be female related after all. One of the estimated "females" (L5) contains an artifact type that may be male related, the stone hammer. The majority of the estimated "males" (16 out of 26 new male burials, or 61.5%) only contain one male related artifact type (all those burials with a value of 0,889 in Figure 4-2). The types present in these burials are bone FIGURE 4-2. Distance values for the unsexed burials from the Simple Inspection Method and assignment of sex, and the distance values for known sexed burials. Unsexed burials Ell E12 E19 E26 E27 E38 E49* E53 E54 Ideal male 0.889 0.889 0.778 0.444 0.889 0.889 1.000 0.889 0.778 Ideal female 1.000 1.000 1.000 1.000 1.000 1.000 0.333 1.000 1.000 E58* E61 E63 E66 E87 E103 E106 E110 E118 Ideal male 1.000 0.889 0.889 0.889 0.889 0.333 0.667 0.889 0.889 Ideal female 0.333 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 E119 L3* L4 L5* L17 L22 L24 L25 L47* Ideal male 0.889 1.000 0.333 1.000 0.333 0.889 0.778 0.667 1.000 Ideal female 1.000 0.0 1.000 0.333 1.000 1.000 1.000 1.000 0.333 L60* L75 L98 L126 X40 Ideal male 1.000 0.889 0.889 0.889 0.778 Ideal female 0.333 1.000 1.000 1.000 1.000 Note: assigned "females" marked*; assigned "males" unmarked. Burials with lower distance values are closer to the "ideal male" or "ideal female". L3 is identical to the "ideal female" because it has all of the "female" artifacts present. Cont'd. FIGURE 4-2 continued Known females E7 E28 E30 E55 E67 E82 E102 E115 E130 Ideal male 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 Ideal female 0.667 0.667 0.667 0.667 0.333 0.667 0.667 0.667 0.667 i'i' E131i L10 L72 L105 L121 X57 X85 Ideal male 1.000 1.000 1.000 1.000 1.000 1.000 1.000 Ideal female 0.667 0.667 0.333 0.333 0.667 0.667 0.667 The female burials with the lower distance figures have more of the "female" artifact types present and are more similar to the "ideal female". All of the known females are the maximum distance from the "ideal male". Known males E9 E34 E59 E73 E91 E99 E107 E109 E112 Ideal male 0.333 0.556 0.667 0.889 1.000 1.000 1.000 0.778 1.000 Ideal female 1.000 1.000 1.000 1.000 0.667 0.667 0.667 1.000 0.667 L15 L117 L122 L123 L125 Ideal male 0.889 0.778 1.000 1.000 0.222 Ideal female 1.000 1.000 0.667 0.667 1.000 The male burials with the lower distance figures have more of the "male" artifact types present and are more similar to the "ideal male". Eight of the known males are the maximum distance from the "ideal female". Six of the known males are more similar to the "ideal female" because they have none of the "male" artifact types. Males with a value of 0.889 only have one of the "male" artifacts present. - Ill -arrowhead, bone spoon or spatula, and whetstone. Twelve of these 16 burials are from the Early period and four are- from the Late. Only five of the 12 estimated "male" burials from the Early period have more than one male related artifact type present, while one-half of the estimated "male" burials from the Late period have more than one male related artifact present. Thus, the greater number of artifact types in the Late period burials makes the estimation of sex in these burials more reliable. None of the "male" identifications is based solely upon the presence of the uncertain netweight and stone disk category. 4.2.5 Discriminant Analysis: Data The classification procedure in the discriminant analysis program allows the inclusion of a large number of variables. In this analysis, 129 artifact types coded by frequencies and 111 burials (29 known sexed burials and 82 unsexed) are included. In order that the results be comparable to those from the simple inspection method, a separate analysis is not done for each chronological period and the 11 possibly sex-linked variables from the simple inspection analysis are included as they were in that analysis. The 129 artifact types are listed in Figure A4-2. Relationships among items in the data set (ie., the sex-linked artifact forms) should be identified whether .data are coded by presence/absence or by frequencies. - 112 -In addition to the 11 original artifact types, all other types of grave goods from the known male and female burials are included. The netweight and stone disk categories are separated here because I became uncertain about my initial interpretation at this point. Tool and ornament types that are known to be shared between the known male and female burials are included. They are coded by style number whenever possible in order to determine whether any styles are exclusive to either sex. The shared types that cannot be coded by style are included in case they are found more likely to be associated with either sex. The ceramic styles (in original, unlumped form) from the known male and female burials are included as well. The chronological analysis of Chapter 3 indicated that the variability in style numbers is at least partly explained by chronology. Discriminant analysis should be able to determine whether any of the styles are also sex-linked. The types and classes of ceramics or of other artifacts which are known to be temporally sensitive are considered in the interpretation of the results if any of these forms are found to be important discriminators of sex. All artifact types are included which are present in the known sexed burials and in at least one unsexed burial. Some of the artifact types shared between the sexes may be surprising to western archaeologists; fanglun (spindle whorl) , zhen (bone needle) , j_I (stone and bone hairpin) , and bihuan (stone and bone bracelet). If these were utilized in an R mode analysis similar to that by Shephard (1979) and - 113 -Hodson (1977), incorrectly interpreted male and female dimensions could result. One of the sexed burials included in the analysis by simple inspection is excluded here (E28, female with infant). In case any of the artifacts in the burial were intended for the infant, E28 will be treated as a multiple burial for this and all subsequent analyses. This problem was not recognized earlier. 82 of the 85 unsexed single adult burials (listed in Figure A4-1) contain at least one of the 12 9 artifact types from the known sexed burials. The ex cluded burials are E43, E71, and X128. Only those artifact types or styles from the subadult male burial (L117) that are found in other adult graves of known sex are included. Types or styles found exclusively in L117 that could reflect age are excluded. Some discrepancies in the quantities of tool and ornament types were noted from Chapters 4 and 6 of the site report (on tools and ornaments, respectively), the summary chart 5 page 131, and from the descriptions of the burials. As in the chronological analysis of Chapter 3, the quantities from the burial descriptions are employed here. 4.2.6. Discriminant Analysis: Results One discriminant function that separates the known male burials from the known female burials is derived, accounting for 100% of the variance. It is unfortunate that 102 of the - 114 -129 original variables entered in the analysis are not a part of the discriminant function. These variables were rejected because they failed the minimum tolerance test, not reaching the minimum tolerance value in the program of 0.00100. The tolerance level of a variable is "the propor tion of its within-groups variance not accounted for by other variables in the analysis", or 'unique variance1 (Hull and Nie 1981:293). Discriminant analysis searches for independent variables that discriminate between groups and rejects vari ables that are found to co-vary with others among the known sexed burials. Only five of the 12 artifact types employed in the analysis by simple inspection achieved the minimum tolerance level and were included in the discriminant function: stone chisel, bone arrowhead, bone spoon or spatula, tusk knife and the horn netweight. Both possible female artifacts were rejected (the head and neck ornaments and the small round stone). The difference in the centroid value for the known male burials (3.86 313) from the value for the female group of burials (-3.60558) indicates that:thejtwo .groups are distinct. The standardized discriminant function.coefficients identify the strongest discriminators as stone chisel, ring (II), awl (I), adze (medium-sized II), and tusk knife. Fairly strong discriminators are bone arrowhead, bone pointed tool (II), and horn netweight. Four of these variables (the chisel, tusk knife, bone arrowhead and horn netweight) are - 115 -known from the simple inspection analysis to be male discriminators. The standardized discriminant function coefficients are listed in Figure 4-3. Thirteen (15.9%) of the 82 unsexed burials are class ified as male and 69 (84%) as female. That is, 13 burials are more likely to be male than female on the basis of the male or female discriminating artifacts they contain. It appears that the assignment "female" to the majority of the unsexed burials is by default because these burials do not contain any male related artifacts. From my exploratory study of the distribution of artifacts in the known sexed graves, I suspect there are no strongly female-related discriminating artifacts. The discriminant scores for the 82 unsexed burials are shown in Figure 4-4. None of the disturbed burials are identified as male (L46, L3, E27, L77). Eight of the 13 males were also identified by the simple inspection method. Five of the burials (E79, E81, E101, E129, and L124) are identified as male on the basis of variables not included in the analysis by simple inspection. The classification of the unsexed burials by the dis criminant analysis cannot be considered totally reliable since it is not based on the total range of artifact types in the known sexed burials. Figure 4-4 shows that the discriminant scores for the estimated males and females fluctuate a great deal about the centroids for the known male and female groups of burials, indicating a poor classification - 116 -FIGURE 4-3. Standardized Discriminant Function Coefficients for the 27 variables that distinguish between the known male and female classes of burials. Listed in order of importance, according to absolute value. 1 adze (medium sized II) -13.61991 2 awl 9.86757 3 stone chisel -8.40071 4 tusk knife 6. 63768 5 ring (II) 6.10280 6 pendant or net weight 4.18537 7 painted tool (II) 4.04045 8 arrowhead -3.90091 9 spade (III) 2.42101 10 spade (IV) -2.21648 11 spear (I) -2.13114 12 sickle 2.08805 13 spoon or spatula -1.53598 14 lower jaw bone of pig 1.49133 15 ax (II) 1.43660 16 spade (V) 1.34083 17 bracelet (IV) -0.67213 18 hairpin (II) -0.49056 19 three legged vessel with bent body(III)-0.48811 20 hairtie -0.40644 21 spindle whorl -0.40418 22 hairpin (IV) 0.39779 23 pig skull -0.23252 24 hairpin (I) 0.22943 25 ring (I) -0.19595 26 deer teeth 0.19533 27 bracelet (I) -0.05743 - 117 -FIGURE 4-4. Discriminant Scores for the 82 unsexed, single adult burials in the Discriminant Analysis, Direct Method. ( * = male burials identified by the Simple Inspection Method also.) Assigned males: * L4 L24 L25 E26 L75 E79 E81 E101 E103 E110 L124 L126 E129 1.3178 4.0476 14.0075 55.3965 6.3347 4.1213 22.4919 0.3587 22.9345 46.0268 3.8262 6.0396 18.1390 * * * * * * Assigned females: L2 L3 L5 E6 E8 Ell E12 E14 E16 E17 E18 E19 E20 E21 L22 E2 3 E27 E29 E32 E33 E38 X40 E41 E42 -4.2893 -3.1088 -7.2403 -4.3630 -3.6990 14.4705 -8.3470 -2.8137 -4.8795 -1.7808 -5.1746 66.4834 -3.1088 -2.8137 -7.9781 -3.6253 -7.8306 -3.7728 -6.3550 -4.9532 16.7576 38.1530 -3.1088 -3.9204 Cont'd... - 118 -FIGURE 4-4 continued Assigned females: cont'd. E4 4 -3.8966 E45 -3.4777 L46 -3.1088 L47 -7.5354 E48 -3.1088 E49 -4.6581 E51 -3.1088 E52 -3.9941 E53 -9.7499 E54 -4.8057 E56 -12.6999 E58 -4.3630 L60 -3.1088 E61 -3.1088 E62 -2.9613 E63 -4.8795 L64 -3.4039 E65 -5.6172 E66 -6.9452 E76 -3.6253 L77 -2.9613 E78 -3.6253 E80 -2.8137 X83 -2.6662 E84 -3.4039 E86 -3.7728 E87 -116.2092 E88 -3.6253 E90 -3.7728 L93 -1.5595 L96 -3.8466 E97 -2.9613 L9 8 -1.3382 L100 -2.9613 L104 -10.9292 E106 -2.5186 E108 -3.2564 X113 -2.9613 E116 -4.3630 E118 -8.1994 E119 -4.2155 E120 -3.1088 L127 -11.5932 E132 -1.3382 X133 -3.1088 - 119 -of the unsexed burials. The majority of artifact types are not a part of the discriminant function because they were found to co-vary with other artifact types. The criterion in the discriminant analysis program that co-varying variables be rejected seems nonsensical in the present case. It seems likely that in mortuary ritual, mourners would purposely place sets of items reflecting the social persona of individuals in graves. Therefore, there is probably a tendancy for artifact types to co-vary with others for the purpose of symbolic communication, not for artifact types to be independent of one another. The particular data set used in this analysis may be of poor quality for discriminant analysis as well. Dis criminant analysis is best suited for data sets that can meet the assumption of a normal distribution (Sneath and Sokal 1973: 404). The 129 artifact types in the sexed burials do not occur with great frequency or with much variation in quantity across the burial sample. 4.3. Conclusion The results from the simple inspection method cannot be adequately compared with those from the discriminant analysis due to the rejection of the majority of variables considered potentially sex-linked. Only five of the 12 variables from the simple inspection method (with horn netweight separated from the stone disk artifact type) are included in the - 120 -discriminant function that distinguishes the male and female classes. The fact that the majority of male burials (eight of 13) identified by the discriminant analysis were identified by the simple inspection method is some support for the simpler classification method. It seems likely that if more of the original (12) artifact types had been included in the disc riminant analysis, the number of male burials identified by both methods would have increased. For the Dawenkou data set, the simple inspection method has proven more useful than discriminant analysis. But mortuary data sets exhibitingrmoreecohtinuous variation may find discriminant analysis more suitable and more capable of classifying unsexed burials. Figure 4-5 lists the burials for which sex has been estimated. Their spatial, locations in the cemetery are depicted in Figure 4-6. Figure 4-6 shows that males and females were buried in all locations of the cemetery during both time periods. Eighteen estimates considered as "reliable" are those with two lines of support: either they were identified by the simple inspection method and discriminant analysis, or they were identified by the simple inspection method and they are supported by a systematic comparison of sex-linked production tools from Dawenkou Culture sites in Pearson (1981), The eight males identified by both classification methods - 121 -FIGURE 4-5. Known sexed burials and estimated sexed burials, from the Simple Inspection Method and Discriminant Analysis. Early period (9) known males: 9, 34, 59, 73, 91, 99, 109, 112. (9) known females: 7, 30, 55, 67, 82, 102, 115, 130, 131. (11) reliably estimated males: 11, 12, 19, 26, 54, 61, 63, 66, 103, 106, 110. (0) reliably estimated females: none. (10) fair estimated males: 27, 38, 53, 79, 81, 87, 101, 118, 119, 129. (2) fair estimated females: 49, 58. (34) adult burials of unknown sex: 6, 8, 14, 18, 20, 21, 23, 29, 32, 33, 41, 43, 44, 45, 48, 51, 52, 56, 62, 65, 71, 76, 78, 80, 84, 86, 88, 90, 97, 108, 116, 120, 132. Late period (5) known males: 15, 122, 123, 125, 117. (4) known females: 10, 72, 105, 121. (6) reliably estimated males: 4, 17, 24, 25, 75, 126. (0) reliably estimated females: none. (3) fair estimated males: 22, 98, 124. (4) fair estimated females: 3, 5, 47, 60/ (10) adult burials of unknown sex: 2, 16, 46, 64, 77, 93, 96, 100, 104, 127. Undatable burials (0) known males: none. (2) known females: 57, 85. (1) reliably estimated-males: 40. (0) reliably estimated females: none. (0) fair estimated males or females: none. (4) adult burials of unknown sex: 83, 113, 128, 133. - 122 -FIGURE 4-6. Spatial location of known sexed and estimated sexed burials in the cemetery. I I EARLY PERIOD BURIAL E3 LATE PERIOD BURIAL £3 UNDATABLE BURIAL MALE A FEMALE - 123 -are marked in Figure 4-5. The following five burials are considered "reliable" for the latter reason: Ell, E12, E61, E6 3, and E66. These burials were identified on the basis of only one "male" artifact, and the artifact types in question have been identified as male related for other Dawenkou Culture sites. These types are the bone arrowhead and whet stone. Pearson (1981:1036) concludes that bone arrowheads (as well as bone and stone chisels, other male-related art ifacts identified by the simple inspection method) occur exclusively with male burials from four Middle Dawenkou period sites (Liulin, Dadunzi, Xixiahou, and Dawenkou). Whetstones occur approximately three times more on the average with males than with females from these sites (Pearson 1981:1085). Five other burials are also considered reliably "male": E19, E54, E106, L17, and X40. Although identified as male by only one classification method (simple inspection), these burials contain more than one of the probable male-related artifacts (see Figure 4-2). "Fair" estimated males include the five identified only by discriminant analysis (E79, E81, E101, E129, L124) and eight identified by the simple inspection method on the basis of one artifact type, with no other support from the arch aeological literature on Dawenkou Culture sites. These burials are: E27, E38, E53, E118, E119, L22 and L98. Each contains a bone spoon dr..spatula. Since a female burial at the Xixiahou site (roughly contemporaneous to Dawenkou, Pearson 1981:1081) contains a bone spoon or spatula, this - 124 -artifact type may not be male-related (The Shandong Arch aeological Team 1964:104). Unfortunately, none of the females identified by either classification method can be considered "reliable". As explained previously, the majority of females seem to have been identified on the basis of the lack of male artifacts rather than the presence of clearly female-related artifacts. Six of the known male burials also lack the male-related artifacts. The six "females" identified by both the simple inspection method and the discriminant analysis are considered "fair" estimates: L3, L47 (with the ornaments) and E49, E53, L5 and L60 (with the small round stone). Pearson's (1981) comparative study provides additional support for the artifact types regarded as male-linked in the simple inspection and discriminant analyses. Besides the arrowhead, chisel, and whetstone types, the stone and tusk knives are shown to be male-linked, although not exclusively (Pearson 1981:1085). The difference in the distribution of some artifact types between this analysis and that of Pearson (1981) (i.e., artifact types such as the whetstone that are exclusively male at Dawenkou but shared between the sexes in other Middle Dawenkou period sites) may be a reflection of chronological differences among sites. Pearson (1981:1084, from Luo and Zhang 1979) notes a general lack of exclusively female tools at Middle Dawenkou period sites. Females have fewer tools for agriculture, - 125 -hunting and maintenance but greater quantities of needles and spindle whorls (ibid). One possible reason for this lack of female-related artifacts may be that females were lower in status than males during the Middle Dawenkou Culture period. Pearson (1981:1086) suggests that the status of women may have been declining through time in the eastern seaboard region as agricultural systems became more intensive. This possibility is considered in Chapter 6. Although the simple inspection method and discriminant analysis classification procedures were somewhat problematic, a greater proportion of "sexed" burials in the burial populations for the Early, Late and undatable burials resulted. If all of the estimated burials are accepted, the number of "sexed" adult burials in the Early period is increased from 18 (known) to 41, with 34 of unknown sex remaining. The increase in the Late period is from nine (known) to 22, with 10 burials of unknown sex. The undatable sexed adult burials increase from two (known) to three, with five unsexed remaining. - 126 -CHAPTER 5 THE NATURE OF SOCIAL SUBGROUP AFFILIATION AT DAWENKOU 5.1. Method This chapter is an exploratory assessment of social sub group affiliation represented at the Dawenkou cemetery during the Early and Late periods. Four aspects of mortuary treat ment noted as potential indicators of social subgroup affilia tion (spatial location of grave, orientation of grave, grave form and body disposition, from Binford 1971:22 - see Figure 2-3) are assessed here. Figure 2-3 indicates that grave orientation and grave location are the most likely to reflect social subgroup affiliation. Binford regards social affilia tion as including a variety of membership groups such as clans, kindreds, and lineages (Binford 1971:22). Assessment of variability in ceramic style among the burials is also assessed here. The methods that have been used in the literature to assess social subgroup affiliation are limited. Some researchers have attempted to identify the specific type of social subgroup represented at a site: Longacre (1970: re type of descent and residence group), Binford (1972:411-412, type of descent and residence group), Saxe (1971:52, type of residence group), Van de Velde (1979:43, type of lineage), Decker (1969:78-79, type of descent group), and Shennan (1975:286, type of descent - 127 -group). Studies that attempt to identify specific type of social subgroup are criticized by Allen and Richardson (19 71) because the material correlates that reflect different types of social subgroups are not known. Tainter (1978:131) also notes the lack of cross-cultural ethnographic studies that indicate different types of social subgroups by ,different patterns of material culture. Some investigators (Goldstein 1981, 1980; King 1969, Tainter 1976, Chapman 1981, 1977) propose that corporate descent groups are reflected in a cemetery on the basis of the ethnographic tests by Saxe (1970) and Goldstein (1976). In other mortuary studies, variables from Binford's (19 71) study that may indicate social affiliation are assessed on an individual basis. These were written by Peebles (1974, 1971), Gruber (1971), Wright (.1978), King (1969), Whalen (1983), and Brown (1971) . Goldstein (1981, 1980) utilizes the spatial organization of the site as the organizing framework for an assessment of social subgroup affiliation. This approach is followed here. The method utilized here is to assess whether there are correlations in spatial location of grave, grave orientation, grave form and body disposition during each period at Dawenkou. First, spatial clusters of graves within the cemetery for each period are determined. Then, the variability in grave orienta tion, grave form and body disposition is assessed for each spatial cluster of burials. Social subgroup affiliation is expected to be reflected if the spatial clusters are distin-- 128 -guished by particular grave orientation(s), grave form(s) and body disposition(s). Goldstein (1980:3) and Petersen (19 81:185) point out that the differential location of graves in a cemetery implies differential treatment of individuals in life. O1Shea's (1981) study suggests that separate burial areas within a cemetery are suggestive of social subgroups but additional lines of evidence are required to make this suggestion more reliable (1981:50). Identification of contrasting sets of discretely occurring attributes is better evidence for the representation of social subgroups in a cemetery, although the cultural meaning of these sets may be ambiguous (O'Shea 1981:50-51). Covariation of age and sex with spatial areas is also assessed here. However, because many types of social subgroups are composed of an age and sex distribution similar to that in the total mortuary population (O'Shea 1981:50), patterning by age and sex alone is not taken to reflect social subgroups. An effort is also made to assess whether patterning in the potential social subgroup affiliation - related variables is a reflection of time. In the last section of this chapter, propositions regarding the nature of social subgroup affiliation at Dawenkou are offered. Spatial areas of graves within the cemetery are also assessed for composition of ceramic style. Pearson (1981: 10 86) suggests that correlations of artifact styles within the Dawenkou cemetery could represent lineages. I am not - 129 -aware of any mortuary studies since Longacre (19 70) that utilize ceramic style to assess social affiliation. The assumptions and methods in Longacre (19 70) are criticized by Allen and Richardson (19 71) . Since the relationship between ceramic style and social affiliation does not have support from cross-cultural ethnographic studies, a lack of correlation of ceramic style with spatial location of grave is not to be taken as a lack of representation of social affiliation in the cemetery. The definition of location of grave used in this assessment differs slightly from that in Binford's (1971) study. Binford's (1971:21) definition is "whether the facility was differentially placed in the life space of the community, or in spatially differentiated burial locations". Orientation of grave in degrees from north is assessed as well as four types of grave form: simple pit, pit with ercengtai ("second level platform", or ledge), log tomb, and log tomb with second level platform. Binford's (1971: 21) definition of body disposition is "distinctions made by differential disposition - placed in a grave, on a scaffold, disposed of in the river, etc.". In this assessment, a modified definition of body disposition is employed because the site report describes variability in terms of the disposition of skeletons within graves (whether supine or prone, for example). - 130 -5.2. Spatial Location of Graves The only methods that have been employed to derive spatial patterning within a cemetery of which I am aware are visual inspection to identify rows of burials and linear regression to check the rows (Goldstein 1981, 1980), nearest-neighbor analysis (Tainter 19 76, Peebles 1971, Shennan 19 75, King 1969, Whalen 1983) and visual inspection alone (Cole and Harding 1979, MacDonald 1980). The method that I consider most appropriate for the Dawenkou cemetery is visual inspec tion. Linear regression is not appropriate because the visually distinct groups of burials in the cemetery are more concentric in form than linear. Nearest-neighbor analysis has been considered problematic for mortuary sites with dense groups of burials (Goldstein 1981:58). However, there is a more fundamental reason for deriving spatial groups of burials by visual inspection. The site map shows some visually distinct clusters of burials from each period. Although some clusters are not totally clear and may be subject to different interpretation by different researchers, clusters apparent by visual inspection are those that are most likely to be culturally significant. As Goldstein (19 81:5 8) points out: It is reasonable to assume that the spatial principles used by a society will be fairly distinct and apparent. It is probable that rules for placement of individuals will have been followed consciously by the members of a society (especially in a cemetery, barrow or mound), and it is unlikely that these - 131 -people had computers or random-number tables to assist them in developing a visually incomprehensible pattern. Clusters of burials identified by visual inspection for the 85 Early period graves are shown in Figure 5-1. There are five roughly concentric groups of burials labelled numerically, five pairs of burials labelled by letters, and three isolated burials (E48, E14, and E36). The clusters of burials identified for the 33 Late period graves are shown in Figure 5-2. Four clusters of burials are labelled numerically. The locations of the 15 undatable burials can be seen in Figure 3-4. Two undatable burials, X12 8 and X92, are spatially isolated from other burials in the cemetery. 5.3. Correlation of Grave Orientation, Grave Form and Body  Disposition with Grave Location The majority of Early and Late period and undatable burials are oriented roughly east-west. An attempt is made to determine whether the minor differences in orientation among burials are culturally significant as suggested from the map of the cemetery. The two exceptions to the east-west trend are E45 (270°) and X12 8 (oriented north-south). Grave orientation in the Early period ranges from 43° -2 70° from north. There are no sharp breaks in the distri bution of values for the majority of burials. From the map of the cemetery, it appears there are three types of orienta tion: east-west, northeast, and southeast. - 132 -FIGURE 5-1. Spatial groups of burials derived by visua inspection, Early Period. CD EARLY PERIOD BURIAL EH LATE PERIOD BURIAL £3 UNDATABLE BURIAL - 133 -FIGURE 5-2. Spatial groups of burials derived by visua inspection, Late Period. I I EARLY PERIOD BURIAL EH LATE PERIOD BURIAL 23 UNDATABLE BURIAL - 134 -As Figure 5-3 indicates, northeast orientation is arbitrarily defined here as 43° - 80°, east-west as 83° -100°, and southeast as 102° - 122°. The skeletons lie with their skulls to the east (as shown in Figure 5-3) . The only exception to this rule is E45 with the orientation of 270°. The skeleton's head in E45 is towards the west. Figure A5-1 shows the distribution of orientation for the Early period graves. Figure A5-2 shows the distribution for the Late period burials. The majority of Early period burials are oriented east-west (52, or 61.2%), while 11 (12.9%) are oriented northeast and 22 (25.9%) southeast. The orienta tions of some burials look slightly different on the site map than one would expect from the orientation in degrees given with the descriptions of the burials. I assume that the values in degrees are the more accurate description of grave orientation. The distribution of the three arbitrarily defined types of orientation within the various spatial groups of Early burials is shown in Figure 5-4. The three types of orienta tion do not tend to be distributed exclusively in different spatial locations. I acknowledge that different definitions of an east-west, northeast and southeast orientation could result in a slightly better (or worse) correlation of orientation with spatial location. The lack of correlation noted here between orientation type and spatial location could indicate the orientation types are culturally insignifi cant or that orientation is not a reflection of social - 135 -FIGURE 5-3.- The range of grave orientations for the Early, Late, and Undatable burials. Orientation varies from 43°-270°. o° 180° ,1 Three types of orientation, arbitrarily defined (by eye): NE (43° - 80°) E-W (82° - 100°) SE (102° - 125°) Unique orientations: Burial E45 at 270° Burial X128 at 180° See text for explanation. FIGURE 5-4. Correlation of grave form, grave orientation, body disposition, age and sex with spatial location of grave (as in Fig. 5-1), Early period. (Intrusive pairs of burials indicated in Fig. A3-2). KEY overleaf. location grave form grave orientation body disposition age sex SP ER LT TE NE E-W SE 270° S LS RS P A C U RM M FM FF F MU 14 1 1 1 1 1 36 1 1 1 1 1 48 1 1 1 1 1 Pair A 1 1 1 1 2 2 1 1 Pair B 2 2 2 2 1 1 Pair C 1 1 2 1 1 2 1 1 Pair D 2 1 1 2 2 2 Cluster 1 (8 total: 1 multiple 1 no skel eton ) 4 3 1 8 8 8 1 4 2 1 1 1) M,F Cluster 2 (12 total) 11 1 3 8 1 8 111 12 11 1 Cont'd. FIGURE 5-4 continued location grave form grave orientation body disposition age sex SP ER LT TE NE E-W SE 270° S LS RS P A C U RM M FM FF F MU Cluster 3 (4 total) 3 1 4 3 1 4 2 1 1 Cluster 4 (14 total: 2 multiple) 12 2 16 7 16 15 1 12 1 1 1 1 1) F,C 2) ? ? Cluster 5 (34 total: 3 multiple) 26 4 4 2 17 15 34 3 1 34 4 17 4 7 5 5 DM, F,C 2) ? ? 3) M,F SP simple pit ER - second level platform LT log tomb TE - log tomb and second level platform NE = northeast orientation E-W = east-west orientation SE = southeast orientation S = supine LS left side RS = right side P prone A = adult > C = child i SU = subadult U — unsexed RM = reliably estimated male M = known male FM — fair estimated male FF — fair estimated female F = known female MU — multiple burial - 138 -affiliation in the Early period. The majority of the 85 Early period burials are simple pits (.68, or 80%), 10 (11.8%) are pits with second level platforms, four are log tombs (4.7%) and three (3.5%) are log tombs with second level platforms. Early period grave form types are shown in Figure 5-4. It is apparent that there is no correlation between grave form and grave location: each type of grave form is found in a variety of spatial locations. Four types of skeletal layout or body disposition occur among the Early period burials: supine, on the right side, on the left side, and prone. The majority of burials are supine (74, or 88.1%), four are on the right side (4.8%), five on the left (6.1%), and one is prone (1.2%). Grave E54 does not contain a skeleton. There is some variability in the placement of the hands and feet among all burials in the cemetery, but it seems largely due to type of skeletal layout. Only the more obvious variation in skeletal layout is considered here. There is also variability among all burials in the direction in which the skull faces (north, south or east, in one case). This variability is not assessed in this chapter because the direction of skull is not described for more than half of the burials. Also, the variability may be due to skeletal layout or post-depositional disturbance of skeletons. The ten Early graves in which skeletal position is other than supine are - 139 -shown in Figure 5-4. It appears- from Figure 5-4 that skeletal layout is not correlated with spatial location of grave. Thus, none of the three potential social affiliation-related variables, orientation, grave form, and body disposi tion is correlated with spatial location of grave in the Early period. Figure 5-4 also indicates that orientation, grave form, and body disposition are not correlated with each other, either. The only exceptions are E45 and E14. In E45, the unique orientation (2 70°) is correlated with the unique skeletal layout (prone). This distinguished mortuary treat ment could reflect social deviance (O'Shea 19 81:43) or a feared or unusual cause of death (Ucko 1969:271). E14 is the only burial with an orientation less than 70°,and it is in an isolated spatial location. Since visual inspection (see Figure 5-4) indicates there is clearly no covariation between any of the Early period potential social affiliation-related variables, I maintain that statistical tests of association are unnecessary. Chi square has been used (Peebles 19 71, 19 74) to test for an association between potential social affiliation-related variables. It is also not necessary to show with a Goodman-Kruskal statistic (MacDonald 1980:42-44) or a Kolmogrov-Smirnov test (Peebles 19 74:94) that orientation is not random. As Goldstein (19 81:58) points out, "One need not prove with statistics that which is apparent". - 140 -Figure 5-5 shows a lack of correlation between orientation, grave form and body disposition with the spatial groups of burials for the 33 Late period burials as well. Orientation is more homogeneous in the Late period, ranging from 80° to 114°. Using the arbitrarily defined types of orientation, one burial (3.0%) is oriented northeast, five southeast (15.2%) and 2 7 (81.8%) east-west. The southeastern oriented burials are not restricted to a certain spatial loca tion. Burials having a simple pit (25, 75.8%), a pit with a second level platform (1, 3.0%), log tomb (3, 9.1%), and log tomb with second level platform (4, 12.1%) are not restricted to particular spatial locations. Only three types of skeletal layout are found in the Late period: supine (2 8, 9 3.3%), right side (1, 3.3%) and left side (1, 3.3%). Burials L24, L60 and L127 do not contain a skeleton. Thus, skeletal layout is more homogeneous in the Late period as well as orientation. The burial with the skeleton on the left side, L117, is in a different spatial group than the burial on the right side, L25. Although both burials have log tombs with second level platforms, other burials with this type of grave form have the supine skeletal layout type. As in the Early period, there is no correlation between orientation, grave form, and body disposition. For the undatable burials, the association of orienta tion, grave form, and body disposition with spatial location cannot be properly assessed because their period and spatial groups are unknown. However, information in Figure 5-6 FIGURE 5-5. Correlation of grave form, grave orientation, body disposition, age and sex with spatial location of grave (as in Fig. 5-2), Late period. location grave form grave orientation body disposition age sex SP ER LT TE NE E-W SE S LS RE P A SU -> U RM M FM FF F MU Cluster 1 (3 total) 2 1 3 3 3 1 2 Cluster 2 (15 total: 1 multiple 2 no skel eton) 12 2 1 1 12 2 13 1 14 4 4 2 1 4 1 DM, F Cluster 3 (3 total: 1 no skel eton) 1 1 1 2 1 1 1 1 1 1 1 1 Cluster 4 10 1 1 10 2 12 12 4 1 3 2 2 KEY: Refer to Fig. 5-4 FIGURE 5-6. Correlation of grave form, grave orientation, body disposition, age and sex with spatial location of grave, Undatable burials. KEY F = female RM = reliably estimated male Unlabelled: simple pits, supine burials, adults and unsexed burials. Burials in rough order from northern section of cemetery to southern section: 70 E-W 37 E-W 39 E-W 40 E-W 50 E-W 57 E-W 83 E-W 68 E-W 74 E-W 85 SE 95 SE 113 SE 133 SE (90 (86C (84( (90C (88C (87C (92C (84C (91* (110 (110C (125C (124C unsexed multiple child child RM, left side child F child child F child flexed Isolated graves 92 SE (110°) , unsexed multiple 128 N-S (180 ) - 143 -regarding these burials does not contradict the conclusion that the potential social subgroup affiliation variables are not associated with spatial location. Both southeast and east-west oriented burials are located in more than one spatial location. However, the one north-south oriented grave (X12 8) is spatially isolated. Some spatial groups have more than one type of body disposition represented as well. All of the undatable burials have the same form of grave: the simple pit. 5.4. Correlation of Ceramic Style with Spatial Location For both periods, there is little correlation of ceramic style with spatial location of grave. Figure A5-3 shows that of 62 ceramic styles that appear in more than one burial in the Early period, only five (8.1%) are found exclusively in one spatial area (clusters 5, 4, 1). None of these ceramic styles is found in every burial in these clusters. Thus, ceramic style does not appear to be a reflection of social affiliation in the Early period. Figure A5-4 shows that of 66 ceramic styles that appear in more than one burial in the Late period, only 11 (16.7%) are exclusive to a particular spatial area. None of these styles is-, found in spatial cluster 2, and two in cluster 4. As in the Early period, none of these styles is found in all the burials within the particular spatial areas involved. Although the percentage of ceramic styles exclusive to a spatial area has doubled from the Early to the Late period, 16.7% is probably not a high enough figure to propose that ceramic - 144 -style is a reflection of social affiliation in the Late period, either. However, a trend towards the demarcation of spatial area by ceramic style may be developing. The Late period styles are associated with a variety of subtypes and functional styles. It is not clear whether a trend towards the development of symbolism of social affiliation by ceramic style is occur ring in the Late period and whether this trend pertains to certain spatial areas of the cemetery (e.g., Cluster 2) more than others. Due to time limitations, a study of the distri bution of production tool and ornament styles in the various spatial areas of the cemetery cannot be made here. Pearson (19 81:10 86) suggests that a study involving the styles of more than one type of artifact could reflect different lineages within the cemetery. An additional distributional study with tool types and ornaments could provide support for the possible trend indicated by ceramic style. Study of the spatial distri bution of artifact styles in other Dawenkou culture sites could indicate whether the trend is a regional one. A study of the distribution of ceramic design attributes in the Dawenkou site and in other Dawenkou Culture sites in Shandong could be useful (M. Blake, personal communication 1983). These design elements could symbolize social affilia tion within the region, as suggested in Wobst's (1977) information exchange model (Braun 1977:119). A distribu tional study of ceramic design attributes could be useful - 145 -for understanding Dawenkou Culture regional social organiza tion. A study of design elements would necessitate visual inspection of the ceramics from the various Dawenkou Culture sites. The design elements are not clear from published reports. The Dawenkou site report indicates that there may be some differences in painted design as well as incised and impressed design. 5.5. Interpretation of the Potential Social Subgroup  Affiliation - Related Variables 5.5.1. Introduction The lack of correlation between orientation of grave, grave form, body disposition, and ceramic style with spatial location indicates no additional support for the possibility that spatial groups of burials from the Early and Late periods reflect social subgroup affiliation. For reasons discussed below, it is also not likely that the variability in either orientation, grave form, or body disposition alone reflects social subgroup affiliation. I propose these variables reflect other aspects of mortuary ritual. 5.5.2. Orientation Figure 2-3 indicates that orientation of grave may reflect sex as well as social affiliation. The known sexed and aged burials from the Early and Late periods are shown in Figures 5-4 and 5-5. In the Early period, three (.30%) of the 10 known females from single graves have a northeast orientation, five (50%) have an east-west one, and two (20%) have a southeast one. One (1.1%) of the nine - 146 -known males from single graves has- a northeast orientation, four (4.4%) have an east-west one, and four (4.4%) have a southeast one. Inclusion of the reliably and fairly estimated males and two females from Figure 4-5 results in the follow ing distribution: northeast: four males (13.3%), three females (25%), east-west: 17 males (56.7%), seven females (58.3%), and southeast: nine males (30%), two females (16.7%). It is apparent that sex and orientation are not associated. In the Late period, none of the five known males are oriented northeast, four (80%) east-west, and one (20%) south east. None of the four known females is oriented northeast, three (75%) east-west, and one (25%) southeast. If the reliably and fairly estimated males and females are included, the distribution is: northeast: male (none), female (1, 12.5%), east-west: male (12, 85.7%), female (5, 62.5%), and southeast: male (2, 14.3%), female (2, 14.3%). Thus, sex and orientation are not associated for the Late period burials, either. The limited data on age of skeleton for both periods does not allow an assessment of whether differences in orientation are a function of age. All that is known is that two of the four children have graves oriented south east and two, east-west. The one subadult in the Late period is oriented east-west. - 147 -The pairs of intrusive burials within the Early and Late periods can be utilized to assess whether differences in orientation are a function of time. Figure 5-4 shows that within the Early period, three of the seven pairs of intrusive burials have the same orientation (east-west), and three have different types of orientation. In Figure 5-5, both burials in the intrusive Late period pair have the same orientation. If differences in orientation are a function of time, it would be expected that the older burials in each pair of intrusive burials would have a different orientation than the younger burials. A more likely explanation for differences in orientation in the Early and Late periods is season of interment. This proposition has been made for other mortuary sites by Gruber (1971:67) and Saxe (19 71:49-50). Gruber (1971:67) proposes that the slight variations in pattern of east-west orientation at the Mohr site in Pennsylvania are due to the position of the sun on the horizon at dawn. The mortuary ritual called for interment of graves in an east-west direction so that the heads of the deceased faced the rising sun in the east. The differences in orientation are due to the different positions of the rising sun at different seasons (ibid). When Gruber (.1971:6 7) calculated the position of the rising sun from season to season, he found that orientations of almost all graves in the cemetery fell within the solar arc. Some ethnographic groups have been known to place burials with respect to the rising sun (Gruber 1971:71). Saxe (1971:49-50) - 148 -concludes that some burials of the Wadi Haifa site in Sudan are oriented according to the position of the rising sun from season to season and others, according to the position of the sun at other times of day. Gruber's (.1971:73) calculations for the Mohr site indicate that summer interments range from 69° - 100° from north and winter interments range from 101° to 131°. It is possible that a calculation of the seasonal position of the sun at Dawenkou may result in a similar distribution. The orientations I interpret as northeast (43° - 80°) and east-west (82° - 100°) may reflect summer interments and those which are southeast (102° - 125°) may reflect winter inter ments. If the above interpretation is accepted, graves were interred in all locations of the cemetery regardless of the season. The east-west, northeast and southeast oriented graves are found in all spatial locations during the Early and Late periods (and for the undatable burials as well). The two exceptions are burials E45 (2 70°) and X12 8 (180°). If the season of interment is the explanation for variability in orientation, the greater consistency in the Late period orientations may be a factor of more burial in the summer. However, one would expect a fair number of individuals to die during the winter. Seasonal position of the rising sun may also explain the variability in grave orientation at other Dawenkou Culture sites. Pearson (in press:22-23) mentions three other Dawenkou Culture sites in Shandong with predominantly east-west - 149 -orientations: Yedian (Late Dawenkou Culture, 90° - 116°), Dafanzhuang (Late Dawenkou Culture, 70° - 9 0°) and Wangyin (fourth layer, Early Dawenkou Culture, 70° - 128°). The burials in the late site of Xixiahou have predominantly an east-west orientation (Shandong Archaeological Team 1964:58). The majority of graves from the Dadunzi site in Jiangsu also have an east-west orientation (Pearson in press:22). However, there is some variability for sites located in different regions and of varying time periods. Two sites earlier in time than Dawenkou (see Figure 1-3) have burials oriented north-south: Songze in Zhejiang and Liulin in Jiangsu. The early site of Beiyinyanying has a northeast pattern (all from Pearson in press: 22-23). Thus, there are differences in orientation pattern among the Dawenkou Culture sites which have not been explained. At some sites, burials may have been oriented according to distinct topographic features visible from the cemetery. Kao (19 83:14) proposes that burials at the Qianzhai site in Shandong (Late Dawenkou Culture) were oriented with respect to a mountain range located southeast from the site. 5.5.3. Grave Form Figure 2-3 indicates that grave form may reflect condi tion of death, age, and social position as well as social affiliation. Condition of death can be proposed for five burials. Burial E2 8 (female with infant) may represent a woman who died in childbirth. As mentioned in Chapter 4, the burials with no skeleton (E54, L24, L60 and L127) may - 150 -represent individuals who died away from the community. The sparse data on skeletal age indicates no apparent association between age and grave form. As Fig. 5-4 and Fig. 5-5 indicate, both subadults and adults have simple pits, log tombs, and second level platforms. All of the undatable burials have simple pits. Figure 5-4 indicates that three pairs of intrusive burials in the Early period have different grave forms: E54 and E58, E9 and E23, and E78 and E129. The one pair of intrusive burials in the Late period has the same grave form represented. Thus, it appears that grave form is not a reflection of time differences among burials. Figures 5-4 and 5-5 demonstrate that grave form is not associated with sex, either. Of the known females in the Early period, eight (80%) have simple pits and two (20%) have pits with second level platforms. Of the nine known males in the Early period, four (44.4%) have simple pits, three (33.3%) have pits with second.level platforms, and two (22.2%) have log tombs. Inclusion of the fairly and reliably estimated males and two females results in the distribution: pits: male (20 or 66.7%), female (12 or 75%); pits with second level platforms: male (six, 20%), female (three, 25%); log tombs with second level platforms: (two males, 6.7%); and log tombs: (two males, 6.7%). For the Late period, four of the five known males (80%) have simple pits and one has a log tomb with second level platform (20%). Inclusion of the fairly and reliably - 151 -estimated males indicates: 10 with pits (71.4%), one (7.1%) with a second level platform, two (14.3%) with a log tomb and second level platform, and one (7.1%) with a log tomb. Three (75%) of the known females have pits and one (25%) has a log tomb with second level platform. The inclusion of the four fairly estimated females gives: five (62.5%) with pits, one (12.5%) with a log tomb and second level platform and two (25%) with a log tomb. I am only aware of one other Dawenkou Culture site with the log tomb, a recently discovered site in Shanxi ( Zou, personal communication 19 83). Only one Dawenkou Culture site appears to have the second level plat form - the late site of Qianzhai in Shandong (Kao 19 83:14). 5.5.4. Body Disposition Body disposition may reflect condition of death, location of death, age, or social position as well as social affilia tion. Condition of death cannot be properly assessed since four of the five burials that may reflect condition of death do not contain a skeleton. A different location of death has been proposed for the burials with no skeleton. Body disposition cannot be assessed for association with age in either the Early or Late periods due to the limited sample of aged burials. However, both children and adults in the Early period have supine positions, and both adults and the one youth in the Late period are positioned on their sides. Child and adult undatable burials have supine positions. The cultural significance of the one flexed burial, in the cemetery CX133) and the one prone burial (E45) is not - 152 -apparent. The Wangyin site has a few prone burials with heads to the west, like E45. The Wangyin burials may reflect cause of death or a special social group (Shandong Archaeological Team 19 79).. Since each burial in the intrusive pairs of burials from the Early and Late periods has a supine position, the intru sive burials do not indicate that there is no association between time differences within each period and body disposi tion. Body disposition may be related to sex. For the Early period, none of the known females or fairly estimated females have a body position other than supine. The same holds true for the known and fairly estimated females in the Late period. Of the eight burials with a side body position in the Early period (either left or right), three (33.3%) are known males and two (20%) are fairly estimated males. Three of the burials with a side body position are of unknown sex. Of the two burials with a side position in the Late period, one (.20%) is a known male and one (16.7%) is a reliably estimated male. The one undatable burial with a side position is a reliably estimated male. Since the side position is not a characteristic of all males in the Early or Late period, some other aspect of mortuary ritual must be reflected as well. Finally, the reason for the greater homogeneity in skeletal layout in the Late period than in the Early period is not apparent. The supine position is most common to Qingliangang cultures (Pearson in press:18). At the Early Dawenkou - 153 -site of Wangyin, the majority of burials are supine and a few are on their sides such as at Dawenkou (Shandong Archaeological Team 19 79). The Wangyin site exhibits variability in body disposition not seen elsewhere: a large secondary burial pit (Pearson 1981:1985). The Nanjing Museum (1978) reports that burials from the northern Yangtze river area, including the area of the Dadunzi site are predominantly supine. In the southern area including the Shanghai region, burials tend to have a prone position (ibid). Wu (19 73) reports that the following sites have burials with mainly a supine position: Dadunzi (the Qingliangang stratum, Liulin and Huating strata), the Liulin stage of the Liulin site, and Xixiahou. Another Early Dawenkou Culture site with mainly the supine position is Beiyinyanying in Jiangsu (Chang 19 79: 164). Another site that exhibits variability in body disposi tion is Qianzhai, with over 50 ash-pits (Kao 1983:114). 5.6. Concluding Propositions 5.6.1 Argument for a Descent Group at Dawenkou It was seen that each spatial area determined by visual inspection for the Early and Late periods is not character ized by distinct grave orientation, grave form, body disposi tion, and ceramic style. The extent of variability in the above four aspects of mortuary treatment is exhibited in each spatial group. Therefore, the cemetery as a whole can be considered as consistent in the above aspects of mortuary treatment. This consistency in burial treatment from spatial - 154 -area to spatial area suggests that the whole cemetery reflects one type of social subgroup. I propose the burials from both periods represent a descent group. Due to the overall consis tency of mortuary ritual in terms of orientation, grave form, body disposition, and spatial location of grave, I propose that the same type of descent group is represented throughout the duration of the cemetery's use. Since test implications for a descent group versus another type of social subgroup have not been developed, the proposition is tentative. My argument is that a descent group is possible, given the importance of descent to agricultural ethnographic societies. Also descent systems were an important part of Shang socio-political organization. Historical data indicate the presence of ranked lineages (Chang 1980:78-79) and the importance of ancestor worship (Thorp 1979:152). The importance of descent in mortuary ritual may be another cultural trait that continues from the Dawenkou Culture to the Shang dynasty. Chang (1979: 110) claims that ancestor worship was established in China by the Longshan period. Pearson (19 81:10 86) proposes that lineages may be reflected at Dawenkou on the basis of the furnished graves with no skeleton. Also, an increasing importance of lineage by the Late Dawenkou period may be indicated by the spatial segretation of some burials composed of men, women and children (ibid). Ethnographic tests by Saxe (1970) and Goldstein (1976) indicate that cemeteries and other discrete disposal areas tend to reflect corporate descent groups (Goldstein 1980:7-8). - 155 -However, more ethnographic tests are necessary to substantiate this relationship (Chapman and Randsborg 1981:23). Also, the criteria to define an organized and formal disposal area have not been clarified (ibid). This problem is relevant to the Dawenkou cemetery. The possible use of the pottery kiln during the Late period may indicate that the cemetery was not used exclusively for the disposal of the dead during the Late period. Whalen (19 83:35) maintains that almost complete uniformity of burial position and orientation indicates a formal, well-organized disposal area. These criteria are met in the Dawenkou cemetery. Some ethnographic studies have indicated a relationship between descent, the ancestors, and mortuary ritual. Rights to land are justified by descent for particular ancestors among the Nuba (Hodder 19 80:164). Nuba mortuary ritual reflects matrilineal descent groups in an ideal manner: In practice in daily life, the matrilineal line is continually frustrated by male dominance and competing paternal rights. But in death, the matrilineal group is assembled 'pure', without the husband presence (Hodder 1980:165). The members of a matrilineal descent group buried together in a common cemetery lived in different settlements during their lifetimes (Hodder 1980:165). Goody (1962) describes the importance of descent in LoDagaa mortuary ritual (Chapman 1977: 22). The LoDagaa are also buried with members of their descent groups instead of with people from the settlements in which - 156 -they had lived (ibid). Goody (1962:412) points out that mortuary ritual is regarded by descent groups as important for the continuation of their descent groups (Chapman and Randsborg 1977:22-23). Corporate descent groups that regulated the inheritance of land and other resources could have developed during the Neolithic period in various world areas (Keesing 1975:17-18). Ethnographic data indicate that there is great variation in types of descent groups (Allen and Richardson 19 71:4 7-48, Pader 1982:64). In many processual mortuary studies, this known variation is not considered. Also, one should not assume a descent system continued in one locality through time (Allen and Richardson 1971:49). I argue that the general con sistency of grave orientation, grave form, body disposition and spatial location of grave from the Early period to the Late at Dawenkou does not reflect a change in type of descent group (or any other type of social subgroup) through time. 5.6.2 Kinship or Residential Groups at Dawenkou The spatial groups of burials for the Early and Late periods remain to be explained. Even though they are not characterized by distinct orientation, grave form, or body disposition, they could represent either a type of kingroiip or a type of residential group. In Nuba cemeteries, spatial clusters of burials reflect named community sections consisting of related individuals (Hodder 1980:165). Settled agricultural ists tend to have village subdivisions (MacDonald 1980:44). Ethnographic groups in Southeast Asia are known to have two - 157 -types of intra-village subdivisions: the family compound and the residential ward or neighborhood (MacDonald 1980:36). Family compounds are distinguished spatially in settlements (ibid). MacDonald (1980:36) proposes that spatial areas of burials at the Bang site in Thailand reflect one of these types of village subdivisions. Family plots in cemeteries are known from other ethno graphic data (King 1969:49, Chapman and Randsbord 1981:15). Spatial groups of graves are said to reflect family groups in the processual mortuary studies by King (1969:55), Petersen (1981:187) and Goldstein (1980:124). The multiple burials at Dawenkou could represent family groups. The earlier Dawenkou Culture sites of Liulin and Wangyin have distinct spatial areas of graves. Investigators think the burials from the second excavation of Liulin are located in five distinct areas, each of which reflects one or more families within a clan (Pearson 1981:1083). The four spatial groups at Wangyin are thought to represent a clan as well (Shandong Archaeological Team 19 79). The later Dawenkou Culture site of Dafanzhuang appears to have spatial groups of graves similar to those at Dawenkou (Archaeological Team of ^ingi, County 1975). An eastern and western sector of burials is reported for the later Dawenkou Culture site of Qianzhai (Kao 1983:14). 5.6.3. The Mortuary Population at Dawenkou If the cemetery was actually used for 600 years in the early period and 400 years in the Late, it seems clear that - 158 -85 burials (or 91 people, including the individuals from multiple burials) and 33 burials (or 34 people, including the one multiple burial), respectively represent only a portion of the living population for each period. It is likely that the cemetery reflects a portion of one or more communities. Thus, the mortuary population supports my proposition that a descent group is represented at Dawenkou. Chang (1979:108) proposes that late Neolithic cemetery sites in the North China Plain region are composed of individuals from more than one village. This proposition is plausible for Dawenkou and possibly other sites in the Dawenkou Culture region. The numbers of burials in Dawenkou Culture cemeteries vary from 11 at the later site of Xixizhou (Shandong Archaeological Team 1964) to 885 at the earlier site of Wangyin (the majority of which appear to be in the same stratum) (Shandong Archaeological Team 19 79). Other earlier sites have large numbers of burials: 344 from the second excavation of Dadunzi in Jiangsu (The Nanjing Museum 1981) and 197 from Liulin, also in Jiangsu (Zhang 1979). Other later sites contain relatively fewer burials: 15 at Xedian and 26 at Dafanzhuang, both in Shandong (Pearson in press:22), and 74 at Qianzhai (Kao 1983:14). It is not possible to • determine whether this variation in burial population is a function of time. The variation may be a function of inten sive land use in certain areas or partial excavation (Pearson, personal communication 1983). The large numbers of burials at Wangyin in comparison to other Dawenkou Culture sites may - 159 -be due to more complete excavation at that site. It appears that burial in the Dawenkou cemetery was restricted to a smaller number of people by the Late period. The mortuary population indicates a sharp decrease in the number of burials by the Late period (from 85 burials in the Early period to 33). It is not possible to ascertain whether this is also a difference in the number of children buried in both periods. Children may have been buried in a separate location by the Late period, since the only subadult in the Late period is the youth, L117. A separate burial location for some subadults is also conceivable for the Early period given the ..fact that subadults .are found in only, six .of the 85 graves (including multiple burials).. However, the six undatable child burials could be from the Early or Late period. The smaller mortuary population in the Late period could be a function of the proposed shorter Late period (400 years, versus 600 years for the Early period). However, I propose that a more plausible explanation for the smaller mortuary population in the Late period is a change in the system of status differentiation from the Early period to the Late. - 160 -CHAPTER 6 ANALYSIS OF STATUS DIFFERENTIATION 6.1. Introduction 6.1.1. Procedure The purpose of this analysis is to understand 1) the nature of status differentiation during the Early and Late periods, and 2) change through time in status differentiation at Dawenkou. The analysis of status for the Early period burials is described in section two and the analysis for Late period burials in section three. Change through time in status differentiation at the cemetery is discussed in section three, as well. Section four is comprised of conclusions and implications for the Neolithic period of the eastern seaboard region. Comparison of my results with the published articles such as Pearson (19 81) regarding the nature of status differentiation at Dawenkou is included. I conclude the nature of status differentiation in each period in terms of wealth (indicated by quantity and quality of grave goods) on the basis of the multidimensional scaling and cluster analyses results. Then, differentiation among burials in terms of energy expenditure (indicated by grave form and grave size) is compared with that by grave goods. Variability in body disposition and grave location is also assessed. Status differentiation in terms of grave goods, energy expenditure, - 161 -body disposition and grave location is evaluated by means of test implications for achieved versus ascribed status. The multiple burials from each period are not included in the multi variate analyses because it is difficult to determine the grave goods associated with particular individuals. The social statuses of the multiple burials are discussed separately. Finally, the intrusive pairs of burials within each period are employed in an attempt to assess change through time in status differentiation within each period. Status differentiation among the undatable burials is also discussed. 6.1.2. Test implications Since status is symbolized in different manners in different societies, any number of the following attributes of mortuary treatment for achieved or ascribed social status may be found to characterize Dawenkou in either the Early or Late period. A society in which social status is achieved (or an egal itarian society) is expected to exhibit differentiation of burials according to the age and sex of the deceased, on the basis of any number of the following characteristics: 1) certain forms of grave goods (for example, certain forms may be regularly associated with certain age or sex classes), 2) relative quantities of grave goods, 3) elaborateness of grave form, 4) relative grave size, 5) grave location (for example, children and infants buried in separate location than adults), and 6) certain body dispositions(s). - 162 -A society in which social status is ascribed (or a society with hereditary ranking) is expected to exhibit differentiation of burials which cross-cuts age and sex classes, on the basis of any number of the following characteristics: 1) certain forms of grave goods (including "badges" which symbolize particular high ranks, imported goods, goods made from imported raw materials, goods which require much skill in manufacture, goods made from rare raw materials symbolizing high rank and goods made from locally abundant materials and not requiring extraordinary skill in manufacture symbolizing low rank), 2) relative quantities of grave goods (greater quantities of all forms of artifacts for high status burials and small quantities for low status burials), 3) elaborateness of grave form (elabor ate for high status burials, simple for low), 4) relative grave size (large for high status burials, small for low status burials), 5) grave location (separate location for certain status levels, especially individuals of high status), and 6) certain body disposition for certain status levels. A ranking pyramid is also expected - or successively fewer burials the higher the social rank. The highest status category is expected to have only one or.a few burials. Children and adults of both sexes are expected to characterize each status level except the highest. Some differentiation on the basis of age and sex is expected in addition to the above differentiation which cross-cuts age and sex categories (the "subordinate dimension" from Peebles and Kus 1977:431). - 163 -Some deviations to the standard test implications described above that have been utilized in processual mortuary analyses have been suggested. They are considered here in sections two and three. Rothschild (1979:666) maintains that high ranking females may be symbolized in a different manner than high ranking males. Thus, not all mortuary attributes which symbol ize high status are expected to be shared between the sexes. This possibility is also noted by Chapman and Randsborg (1981: 9) and Pader (1982:59). Others comment that wealthy child burials do not necess arily indicate ascriptive ranking. Hodder (1980:163) and Braun (1979:68) maintain that wealthy child burials may reflect the achieved status of their parents instead of the ascribed status of the child. Bayard (19 83:18) states that wealthy child burials may reflect parental affection instead of ascribed status but their presence does allow the suggestion to be made that ascriptive ranking is present. 6.2. Analysis of Status, Early Period 6.2.1. Data 6.2.1.1. The Multivariate Analyses A total of 23 variables representing different aspects of mortuary treatment and 79 burials are included in the multi variate analyses. The 79 burials are listed in Fig. A6-1. One burial, E27, is disturbed. The 23 variables considered as - 164 -potentially status-related during the Early period are listed in Fig. A6-2. The frequency of occurrence of each attribute among the burial sample is shown in Fig. A6-2 as well. The variables included in the multivariate analyses for each period reflect grave goods in form, quantity, and raw material. Variability in energy expenditure in terms of grave form (either log tomb, log tomb with second level platform, pit with second level platform, or simple pit) and grave size is deliberately excluded from the multivariate analyses for both periods in order to compare status differentiation based upon grave goods and that based upon energy expenditure. A good correspondence would strengthen the interpretation of relative status since each approach is independent of the other and would indicate that both energy expenditure and grave goods symbolize status. Another check on the patterning from the multivariate analyses is my expectations derived from my exploratory study of the cemetery. As much variability in grave goods as possible that poten tially symbolizes status distinctions is included in the multi variate analyses for each period. Each variable was coded on a presence/absence basis as most variables did not occur very frequently. Some potential status related variables could not be included because they occur only once or twice among the burials. Three occurrences was chosen as the minimum number to make the analyses for each period viable because a minimum of - 165 -three occurrences among the 115 burials was viable in the chronological analysis of Chapter 3. I regard these excluded artifact types as probable high status items. Therefore, the distribution of these artifact types in the derived groupings representing equivalent status from the clustering and scaling serves as still another check on the analytical results. For each period, the variables are coded from the information in the "grave goods" column (sui zang qi wu), pages 136-155 of the site report In some cases, additional artifacts are listed under the "remarks" (bei zliu) column. Two ceramic ware types from the Early period may reflect high status: painted and black wares. The painted wares from Dawenkou culture sites are described as finely made by Thorp (1979:7) and as rare and little understood by Shangraw (1978: 31). The black wares at Dawenkou culture sites are considered finely made (with thin vessel walls and polished) by Shangraw (1978:35). The ceramic forms (functional types) thought to reflect high status at Dawenkou are the wine vessel (zun), serving stand (dou), and tall stemmed cup (bei gao). Bronze wine vessels are found in high status contexts in Shang tombs (Thorp 1980: 56 ). Since there are cultural continuities in mortuary practices from the Dawenkou Culture period to the Shang, it is possible the wine vessel symbolizes high status at Dawenkou as well. This possibility is also raised by Rawson (1980:29). The serving stand is regarded by Pearson (personal - 166 -communication, 19 83) as possibly having a ritualistic function and as symbolic of high status. The serving stand and the high stemmed cup appear finely made and as not having purely util itarian functions in the photographs provided in the site report. The high stemmed cups and serving stands are considered as having ritualistic functions by Shangraw (1978:34). Rawson (1980:29) describes the serving stands and high stemmed cups from several Dawenkou Culture sites as fragile and suggests these forms had a ceremonial function. Three ornamental forms considered status-related are included in this analysis: ring (zhThuan), bracelet (bihuan), and hairpin (ji_) . Bone hairpins have been found in high status contexts in Shang sites (Hay 1973:55). The functions of several artifact types from both periods of the cemetery are unclear. Three artifact forms included here may be ornamental or formerly part of an ornament: possible hair ties (shu fa qi', Gao 1978:31; Pearson 1981:1080), small round flat stone (bing), and small flat piece of bone (ban), (appearing worked in photograph 97 of the site report). The variable problable ornament parts includes forms which appear to be the same as those forms which are part of neck+. laces for the head or neck (tdu slfi or jlng shi) in known female burials: flat, thin piece of stone (pian), stone or bone tubular bead (guan), stone circular bead (zhu), small stone annular shaped piece, possibly an earring (huan). Also - 167 -each of these forms appears ornamental from plate 9 7 of the report. These forms are grouped together because each is present only once or twice across the burial sample. Other artifact forms in the analysis with a probable decorative as well as utilitarian function are elaborately carved bone or elephant ivory carved cylinder (diaotong), turtle shell (gui jia) , and spoon or spatula (b_i) . Only one Early period burial contains an elephant ivory cylinder. Elephant ivory is one of two raw materials in the Early period that may symbolize high status. On the basis of the discussion of the local environment in Chapter 1 it is possible that elephant were obtained locally. The photographs in the site report indicate that some cylinders from both periods are inlaid with stone, but the descriptions of the burials do not mention inlaid stone (except for one Late period burial). It is not clear from .photographs in the site report whether the turtle shells are worked. Chang (1979:280) describes them as polished and states they were used for containers at several Dawenkou Culture sites. Gao (1978:31) states the shells were strung together to form a pouch. Seven variables are coded in an additive, presence/absence manner because the relative quantities of these variables are thought to reflect status distinctions. Frequency distributions showing relative quantities of these variables were drawn, and the natural breaks evident from the graphs were regarded as reflecting culturally meaningful distinctions. These variables - 168 -are the serving stand, total number of ceramic vessels, stone tool total, bone tool total, pig skulls (zhu_tou), deer teeth (zhang ya), and pieces of raw material (liao, either bone, horn, or tusk). The associated frequency distributions are in Figs. A6-3 to A6-9. Since only one Early burial in the analysis has no ceramic vessels, it was not possible to have a separate category for absence of vessels. The functions of the pig skulls, deer teeth, raw material pieces, and oyster shell pieces are not clear. The pig skulls at Dawenkou Culture sites may reflect wealth (Zhang 19 79, Luo and Zhang 19 79, Lu 19 76). A contemporary society from Hainan Island in southern China uses pig jaws for this purpose (Luo and Zhang 19 79). Since pigs may have been important in sub sistence, pig skulls and other pig parts may reflect wealth in terms of food supply. Some or all of the deer teeth may have been part of cutting tools for agriculture (Gao 19 78:31, Chang 1979:160). The raw material pieces may have been for manufacture of various goods. The variable pig parts includes front teeth (zhu men ya), various pieces of bone (zhu gu kuai), and lower jaw bone (xia he gu). The oyster shell pieces (bang pian) are not shown in a photograph in the site report. It is not clear if they are worked. The ceramic categories in this analysis and in the Late period analysis are partially redundant because ware color, form and quantity are all included. The impact of this redundancy on the patterning of status distinctions in these analyses is judged to be minimal. - 169 -6.2.1.2. High Status Unique Items The artifact forms or materials excluded from the multi variate analyses that are considered probable high status items in the Early period are: elephant ivory disk (bl), elephant ivory comb (shu), animal shaped ceramic vessel (tab shou xing qi) , red ochre (chi. tie kuang shi, "iron ore", which I interpret as red ochre), and jade (yu). The elephant ivory disk is found in two single burials and in one multiple burial. None are found in Late period burials. The authors of the site report identify the object as a long hollow object with rectangular sides called a corig. However, on the basis of photo plate 9 4 in the site report, Shangraw (personal communication 19 83) identifies the object as the bi disk, maintaining the Chinese character in the site report is incorrect. The bl form in jade is known from historical data to have ritualistic significance during the Shang period (Rawson 1980:83). It is called the symbol of heaven in the Zhou Li, a text from the late Zhou dynasty (Rawson 1980:83, Sullivan 1977:48). The Zhou Li also states that the jade bl symbolized the lower ranks of viscount and baron (Sullivan 1977:48). The elephant ivory bl may be a badge of high rank in the Early period. Pearson (in press:29) suggests that some types of ornaments from Dawenkou resembling ornaments from the Shang period are badges and reflect inherited social position. Pearson (1981:1086) suggests that certain jade forms from Late Dawenkou Culture sites that do not resemble - 170 -utilitarian artifacts are badges of high status. He has since indicated that these forms may be of other materials as well (personal communication 19 83). The elephant ivory comb is found in one Early burial and in one Late. Only one animal shaped ceramic vessel is found in the cemetery. This vessel resembles a pig. The same burial (E9) contains the two pieces of red ochre. Ochre is not present in Late period burials. It is possible that the animal vessel and the ochre reflect ritual rather than status distinctions. Other animal shaped vessels have been found in late Dawenkou Culture sites, in high status contexts (Pearson 19 83:140). Only one Early burial contains jade while it occurs .in four Late period.graves. Burial E78 contains one jade huan (small annular object). This object is identified as jade from page 96 of the site report, not from the description for the burial. Jade was highly regarded in Chinese historical periods for its beauty and ritual significance (Rawson 1980:81). It was an important part of Shang mortuary ritual (Chang 19 80:156, Rawson 19 80:32). Rawson (1980:32) maintains that nephrite is the mineral which was worked in east coast Neolithic cultures. Jade is extremely difficult to work; abrasives must be used as even modern steel tools cannot cut it (ibid). The source for jade found in east coast Neolithic sites is presently unknown (ibid). However, at least one western investigator, Jeffrey Kao, Depart ment of Anthropology, Harvard University, is currently research ing the problem (Pearson, personal communication 19 83). - 171 -Wu (19 73) suggests some jade was available from Zhejiang and Anhui provinces. Lu (1976) maintains that jade was available from the Tai Shan region of Shandong. Chang (1980:156-157) suggests southwestern Henan and parts of Shaanxi were sources during the Shang period. Rawson (1980:32) suggests jade was imported over long distances from the north (Baikal region) or west (Central Asia) - the main source of jade in historical times. The amount of labor involved in procuring jade and working it strongly suggests it symbolized high status and its use was restricted for mortuary ritual or other types of ritual (ibid). Sullivan (1977:40) mentions that early historic texts state that jade was obtained from several locations in China. Also, a source for a variety of jade other than the highly valued nephrite has been found in Henan (ibid). The interpretation of the results from the multivariate analyses for both periods takes into account the known associa tions of particular artifact types with sex and age. As discuss ed in Chapter 4, the small flat stone (bing) is not definitely female-related. The elephant ivory disk (bi), although possibly incorrectly grouped with the horn zhui in Chapter 4, may be male-related. One of the two single burials in which it is found (E59) is a known male and the other (E26) is a reliably estimated male. None of the 2 3 variables included in the multi variate analyses are exclusive to the Early period child burials. In fact, none of the subadults in the cemetery (including Late period and undatable burials) contain artifact forms or raw - 172 -materials not found in adult burials. 6.2.1.3. Energy Expenditure Several researchers have pointed out the difficulty of identifying culturally significant differences in energy expend iture (Tainter 1978:128, Brown 1981:29, Goldstein 1981:55-56). The differences in grave form for the burials of both periods are self evident: pit, pit with second level platform, log tomb, log tomb with second level platform. It was previously concluded that grave form does not seem to reflect social sub group affiliation or age, in either the Early or Late period. The authors of the site report differentiate between three types of log tombs that reflect varying degrees of elaboration and energy expenditure. These types were not discussed previously. These types may reflect status distinctions. As Pearson (19 81: 10 80) describes, type one is a lid of undressed logs and represents the lowest amount of energy expenditure. Type two has a lid, four sides, and a bottom of loose logs. Type three, the most elaborate, is a chamber made of trimmed logs fitted together and placed in the middle of a large pit. The authors of the site report (1974:5-7) do not indicate the type represented in each burial with a log tomb. Only examples of each type are given. The example of a type one log tomb is burial E5 3 and type two, E94. Only one Early burial and six Late burials have log tombs of type three. The Early burial is a multiple burial, E13. The map of the site indicates great variability in grave - 173 -size for both periods. Grave size is described in the site report in terms of length, width, and depth in meters. Grave volume is not calculated due to my uncertainties over the depth measurements in the report. It is not clear whether depth was measured from the same point in every grave. Therefore, area of grave is employed as the measurement of grave size. The variability in grave form and grave size for the 79 Early burials is indicated in Fig. A6-1. The distribution of values for grave area is shown in Fig. A6-10. The sharp break in the distribution 2 of values after 2.50m is taken to reflect a culturally signif icant distinction in grave size. Seven graves (9.1%) have areas 2 2 over 2.50m (up to 3.82m ). These graves are considered large and the others, small. The areas for two graves, E4 3 and E51, could not be calculated due to lack of length and width measure ments. In a few cases from both periods, grave size as indicated by my calculations does not seem to agree with relative grave size as indicated on the site map. The calculations are consid ered the more accurate indicator of grave size. However, since the size of E43 and E51 appear clearly small on the site map in relation to the seven large graves identified above, these two graves are interpreted as small for the purposes of this analysis. 6.2.2. Results: The Multidimensional Scaling The multidimensional scaling analysis for the Early period burials resulted in ten dimensions, the first two accounting for 51.2% of the variability in the data set. No triangle inequalities in the program were violated,indicating this data set is metric. Dimension 1 accounted for 36.1% of the - 174 -variability and Dimension 2, 15.1%. The third and fourth dimensions accounted for 10.2% and 8.8% of the variability, respectively. When the first and second dimensions are plotted together, it becomes evident that Dimension 1 can be interpreted as status differentiation. The plot of the 79 burials is in Fig.6-1. Clusters from the cluster analysis are superimposed on Fig.6-1. I interpret the four burials on the extreme left of the plot (E26, E9, E54, E59) as highest in status, the burials in the mid section of the plot as intermediate, and the 17 burials in the extreme right section as lowest in status (E29, E36, E41, E43, E51, E71, E86, E90, E114, E120, E48, E27, E20, E45, E62, E88, E89). The four burials at the extreme left end have the greatest occurrence of the attributes I expected to reflect high status, the burials in the midsection have the next greatest amount, and the 17 burials at the extreme right end have the least. The burials at the extreme left and extreme right ends of the plot are spatially separated from other burials and appear to reflect distinct differences in status. The burials in the midsection of the plot along Dimension 1 seem to represent a status con tinuum because they do not seem to be clearly separated spatially from other burials. The burials located at the same point (for example, the ten burials at the extreme right end) have the same attributes in terms of grave goods. These 10 burials appear to reflect a redundantly symbolized status level. My tentative interpretation for Dimension 2 is also status - 175 -FIGURE 6-1. Multidimensional scaling plot of Early Period burials in the analysis of status, with clusters from Ward's Method. large in area (?) small in area (72) log tomb (6) second level platform (12) 43 disturbed burial (I) 43 controversial former Middle Burial (3) /o no skeleton (I ) 43 high status, unique objects (4) « child (4) Note: Clusters 1-7 from Words seven group solution ore circled. In the four group solution, grouped together are Clusters 18 2, Clusters 4 8 5, ond Clusters 6 8 7. Cluster 3 is distinguished by both group solutions. - 176 -differentiation. The burials in the lower midsection of the plot appear to be lower in status than those in the upper mid section of the plot (particularly E7, E78, E129, E67, E49, E58). The latter six burials contain more of the variables which represent ornamental or decorative types of grave goods than those burials in the lower midsection of the plot. Status differentiation may be represented as a diagonal through hyper-space in a roughly southeast to northwest direction. 6.2.3. The Three Types of Cluster Analysis The dendrogram judged to best represent the status relation ships among the burials is from Ward's Method. This judgement was made after studying the clusters of burials in each dendro gram from each of the three techniques at several levels of similarity to determine the best ordering, on the basis of 1) the occurrence of the mortuary attributes shared among the burials in each cluster and 2) my expectations of the status relationships among the burials from my former exploratory analysis of the cemetery. The dendrogram from Average Linkage compares fairly well with that from Ward's Method, while the dendrogram from Complete Linkage does not compare well with either that from Ward's or Average Linkage. As in the chrono logical analysis, the Complete Linkage dendrogram shows small clusters of burials at several distance levels and the Average Linkage dendrogram exhibits some chaining. The fairly good comparison of the clusters from Ward's Method with those from Average Linkage and the good comparison with the ordering of burials on the scaling plot also lends support to the acceptance of the Ward's dendrogram. Only the dendrogram from Ward's Method - 177 -is presented here. The dendrogram from Ward's Method is presented in Fig.6-2. The seven cluster solution (at the 1.0022 distance level), the four cluster solution (1.3262 distance level), and the two cluster solution (2.2437 distance level) are judged to indicate status relationships among the burials most clearly. The seven-four, and two cluster solutions are labelled in Fig.6-1. Each of the three solutions contains clusters composed of similar mortuary attributes and compares fairly well with the multi dimensional scaling solution. At first, I thought the seven cluster solution best depicts status relationships among the burials. The mortuary attributes in common for each cluster of this solution are given in Fig. A6-11. This solution seemed to fit my a priori expectations of status relationships based upon my exploratory study. However, further appraisal of the results indicated that the four group solution better depicts status relationship among the Early period burials. The mortuary attributes characteristic of the four group solution can be seen in Fig. A6-11 as well. The clusters of burials from the seven group solution appear at first to indicate equivalent status levels (see Fig. A6-11). Clusters 1 and 2 appear to represent low status clusters, with Cluster 2 of slightly higher status. The burials in these two clusters contain only a few utilitarian artifact types. Clusters 4, 5, 6, and 7 appear intermediate in status, with Clusters 5 and 6 slightly higher in status than Clusters 4 and 7. Clusters 4, 5, 6, and 7 contain greater quantities - 178 -FIGURE 6-2. Method. Dendogram of Early Period burials from Ward' distance Two Croup Solution - 179 -of artifacts, a greater variety of artifact types (including some ornaments and items of probable decorative function) and black and painted wares. Clusters 5 and 6. contain burials with slightly greater quantities of artifacts. Cluster 3 clearly represents the highest status level, with great quantities of artifacts, several ornamental and decorative artifact types and painted and black wares. I think the four group solution better depicts status rela tionships among the burials because it groups the clusters from the seven group solution that are not significantly different in terms of the potential status-related variables. Clusters 1 and 2 of the seven group solution are grouped together as well as the clusters judged as intermediate in status. Clusters 4 and 5 of the seven group solution are grouped together, and Clusters 6 and 7 of the seven group solution. The clearly high status group of burials is set apart in the four group solution as well. The two group solution is clearly evident in Fig. 6-2 as a "natural break" in the dendrogram. It distinguishes the distinctly low status group of burials (Clusters 1 and 2 of the seven group solution) from the other Early period burials. The clusters from the seven group solution of Ward's Method were drawn on the multidimensional scaling plot (Fig. 6-1) because I originally thought that the seven group solution best depicted status relationships among the burials. Information regarding the clusters from the four group solution is included in Fig. 6-1. Fig. 6-1 depicts a fairly good agreement between the scaling and clustering results (from the seven and four - 180 -group solutions). The high status cluster (3) is located on the left end of the plot, the clusters intermediate in status (4, 5, 6, 7) in the midsection, and the low status clusters (1,2) at the right end. Like the chronological analysis of Chapter 3, the results from Ward's Method and Torgerson"s Metric Multi dimensional Scaling are in sufficient agreement to allow an interpretation of status differentiation to be made based upon the results from both methods. Although both methods clearly depict the high, intermediate, and low status burials, there are some differences in the order ings from each method. The Ward's solution depicts burials 63, 102, and 12 as high in status along with 26, 9, 54, and 59. The scaling solution shows 63, 102, and 12 as spatially separated from the latter burials. As Fig. A6-11 shows, burials 63, 102, and 12 do not contain as many of the attributes I interpret as related to high status as the other four burials. There is also some disagreement over the placement of burials in the lowest status group. Ward's solution includes more burials in the low status group than the scaling plot: burials 80, 82, 97, 21, 91, 119. The latter six burials are spatially separated along Dimension 1 from the other 17 burials interpreted as low in status. Other disagreements in the orderings from both methods include burials 81, 6, and 101 (located along Dimension 1 with burials from Ward's Cluster 6 yet included with the burials from Cluster 7) and burial 30 (included in Cluster 7 yet located with burials from Cluster 2 along Dimension 1). - 181 -6.2.4. Interpretation of Status Distinctions in the Early Period 6.2.4.1. Distribution of the Unique Items The correspondence between status differentiation based upon grave goods included in the multivariate analyses with that based upon the distribution of the probable high status unique objects can be seen in Fig. 6-1. The distribution of the unique objects supports my interpretation that the order ing of burials from the clustering and scaling techniques reflects status differentiation. Five of the six high status objects are found in the high status group of burials (burials 26, 9, and 59 in Fig. 6-1). The two elephant ivory disks (bi) are in burials E59 and E26. E26 also contains the elephant ivory comb. E9 contains the red ochre and animal shaped pot. The jade huan is found in E78 (located in Cluster 5 on Fig.6-1). I originally interpreted the burials in this section of the scaling plot as being higher in status than those in the lower midsection of the plot. These burials also seemed higher in status than other burials of intermediate status from the seven group solution of Ward's Method. Due to the good corres pondence between the differentiation of burials from the clustering and scaling and from the probable high status unique objects, I conclude that status differentiation is at least partially symbolized by wealth in terms of grave goods during the Early period. - 182 -6.2.4.2. Argument for Four Status Levels The differences in the ordering of burials between the results from Ward's cluster analysis, the scaling results, and the distribution of the probable high status unique items preclude a definitive conclusion on the approximate number of status levels based upon grave goods represented among the Early period burials. I argue that four status levels are the most clearly represented. The difference in composition of mortuary attributes among the clusters of Ward's four group solution reflects status differentiation in a more obvious manner than in the clusters of the seven group solution. Because the composition of these four clusters in terms of burials agrees well overall with the multidimensional scaling results, these clusters are concluded as representing the major status distinctions of the Early period. I recognize that the placement of some burials is open to debate. According to the scaling results, the placement of three burials in the high' status cluster (63, 102, 12) on Fig. 6-1 may be incorrect. Likewise, the placement of a few other burials in the low status group (eg., 119, 91, 80, 82, 97, 21) may be incorrect. However, each method seemed to distinguish burials of high, high inter mediate, low intermediate and low status. Since the objective of this analysis is to identify the approximate, rather than the exact, number of. status- .distinctions . that existed in the society of the deceased, the differential placement of a few burials does not significantly change the interpretation - 183 -of status distinctions among the Early period burials. All the characteristics of the four status levels are indicated in Fig. 6-3. Additional unique items found in Early period burials but not considered as represent ing high status are in E103 (a bird claw bone, zhao, from photo plate 101), and E19 (a stone ball, qiu, photo plate 101). It is not clear whether this ball is of the type that have been found in the mouths of skeletons from other Dawenkou Culture sites (Han and Pan 1980). Burial E34 contains the only horn zhui from the Early period. 6.2.4.3. Differentiation by Energy Expenditure Differentiation of burials based upon the four types of grave form and the two types of grave size does not coincide to a great degree with the differentiation based upon grave goods. This lack of correspondence was expected on the basis of my exploratory study of status differentiation. As shown in Fig. 6-1, only three of the seven burials in the high status level have grave forms that represent more energy expenditure than the simple pit. Burials E9, E59, and E12 have pits with second level platforms. None of the high status burials contain the grave forms that represent the greatest amount of energy expenditure, the log tomb, or the log tomb with second level platforms. However, none of the low status burials (Clusters 1 and 2 in Fig. 6-1) have log tombs or second level platforms. Burials with the grave form representing the greatest amount of energy expenditure, log tomb with second level FIGURE 6-3. Characteristics of the Four Status Levels, Early period. high status high inter- low inter- low status mediate status mediate status (7 burials) (17 burials) (32 burials) (23 burials) ~ % of tot. + % of tot. + % of tot. + % of tot, black ware 6 23.1 15 57.7 5 19.2 0 0 painted ware 5 29.4 1 5.9 11 64.7 0 0 wine vessel 5 55.6 2 22.2 1 11.1 1 11.1 serving stand 0 0 0 0 0 14 41.2 20 58.8 serving stand 1-2 4 9.8 16 39.0 18 43.9 3 7.3 serving stand 3-9 3 75. 0 1 25. 0 0 0 0 0 tall stemmed cup 0 0 3 100.0 0 0 0 0 total pots 0-7 0 0 9 14.1 32 50.0 23 35.9 total pots 8-14 4 33. 3 8 66.7 0 0 0 0 total pots 19-30 3 100.0 0 0 0 0 0 0 carved cylinder 2 40.0 1 20.0 2 40.0 0 0 turtle shell 2 40.0 0 0 3 60.0 0 0 ring 2 25.0 1 12.5 5 62.5 0 0 bracelet 0 0 3 60.0 2 40.0 0 0 hairpin 4 50.0 4 50.0 0 0 0 0 spoon 4 30. 8 2 15.4 5 38.5 2 15.4 stone tool total 0 0 0 10 22.2 13 28.9 22 48.9 stone tool total 1 4 18.2 6 27.3 11 50.0 1 4.5 Cont'd. FIGURE 6-3 continued stone tool total 2-7 stone tool total 9-16 bone tool total 0 bone tool total 1-7 bone tool total 13-25 hair tie ornament part blng  ban pig skulls 0 pig skulls 1-3 pig skulls 4-5 other pig parts deer teeth 0 deer teeth 1-3 deer teeth 4-12 raw material pieces 0 raw material pieces 1-10 raw material pieces 16-40 oyster shell high status (7 burials) + % of tot. 2 22.2 1 33. 3 0 0 5 20.8 2 50.0 2 25.0 0 0 0 0 4 57.1 3 6.1 4 16.0 0 0 0 0 1 3.7 5 10.9 1 16.7 3 4.4 2 25.00 2 66. 7 3 50.0 high inter- low inter- low status mediate status mediate status (17 burials) (32 burials) (23 burials) + % of tot. + % of tot. + % of tot. 1 11.1 0 0 12 23.5 4 16.7 1 25. 0 4 50.0 4 80.0 3 100.0 2 28.6 14 28.6 1 4.0 2 40.0 1 11.1 2 7.4 13 28.3 2 33.3 17 25.0 0 0 0 0 0 0 6 66.7 2 66.7 17 33. 3 14 58.3 1 25.0 2 25.0 0 0 0 0 1 14. 3 9 18.4 20 80.0 3 60.0 5 55.6 11 40. 7 18 39.1 3 50.0 25 36.8 6 75.0 1 33. 3 3 50.0 0 0 0 0 22 4 3.1 1 •4.2 0 0 0 0 1 20.0 0 0 0 0 23 46.9 0 0 0 0 3 33. 3 13 48.1 10 21.7 0 0 23 33. 8 0 0 0 0 0 0 - 186 -platform, are E94, E53, and E81. E94 is in the high inter mediate status group (Cluster 4 in Fig. 6-1) and E53 and E81 are in the low intermediate status group (Clusters 6 and 7). E5 3 has a type 1 log tomb and E9 4 has type 2, which represents greater energy expenditure. The other burials with log tombs are E107 and E99 (in the high intermediate status group, Cluster 4 on Fig. 6-1) and E116 (low intermediate status, Cluster 7 on Fig. 6-1). The other burials with second level platforms are Ell, E109, and E3 8. The first two burials are in the low intermediate status group (Cluster 6) and the second, in the high intermediate (Cluster 4). Differentiation of burials based upon grave size also does not coincide to a great degree with that based upon grave goods. Three of the seven large graves are located in the high status group (E9, E59, and E12) . The other large graves are E78, E99, Ell, and E8. E78 is in the high intermediate status level (Cluster 5 in Fig. 6-1). However, this burial contains jade huan. E99 is also in the high Intermediate status group (Cluster 4),and the latter two burials are in the low inter mediate status level (Cluster 6). None of the large graves is located in the low status group. There is a better correlation between the differentiation based upon grave goods and that based upon both grave form and grave size. Six of the seven burials large in size have grave forms other than the simple pit: E9, E59, E12, E7 8, and Ell have second level platforms and E99 has a log tomb. - 187 -E8 is the only grave that is large in size and has the simple pit grave form. E9, E59, and E12 are in the high status group. E78 and E99 are in the high intermediate status group (Clusters 4 and 5). Ell is in the low intermediate status level (Cluster 6 on Fig. 6-1). It appears that energy expenditure in terms of grave form and grave area either does not symbolize status distinctions among the Early period burials or that it symbolizes a differ ent aspect of status than grave goods. The grave forms that reflect great energy expenditure and large grave size are not exclusive to either the highest status group or the second highest. On the basis of the comparison of status differentiation based upon grave goods and energy expenditure, I conclude that high status is characterized by exclusive access to only three forms of artifacts: the elephant ivory disk (bji) , the elephant ivory comb, and the animal shaped ceramic vessel. High status is distinguished by exclusive access to two types of raw material: elephant ivory (the comb and the one carved cylinder) and red ochre. The third characteristic exclusive to the high status burials is the highest quantity of pottery vessels. The high intermediate status level is characterized by exclus ive access to the high stemmed cup, the small round stone (bing) and the jade huan. The other mortuary attributes occur at least once in more than one status level. - 188 -6.2.4.4. Body Disposition Data regarding age, sex and body disposition are included in Fig. 6-4. Fig. 6-4 indicates that on the whole body disposition is not associated with status in the Early period. The high status burials all have the common supine position. Each of the intermediate groups contain burials with side positions. However, the only prone burial (E45) is in the low status group. The low status position of E45 does not contradict my proposition in Chapter 5 that E45 may represent a social deviant or a person who died of an unusual cause. 6.2.4.5. Grave Location The locations of the burials within each status group are shown in Fig. 6-5. Multiple burials (to be discussed shortly) are included in this figure. Five of the seven high status burials are located in the northern cluster of burials. E9 is located in the middle section of the cemetery and E102 in the southern end. The burials of high intermediate and low intermediate status are located in all areas of the cemetery. None of the low status burials are located in the northern spatial cluster, but they are located in several other areas. It appears that a trend towards the spatial segregation of high status burials is developing, as noted by Luo and Zhang (1979) and Pearson (1981:1086). - 189 -FIGURE 6-4. Age and sex composition of the status levels, Early Period burials. (Body dispositions other than supine are marked below, as well as children.) High Status Group: .7 total 63, 59, 54, 12, 26, 102, 9 known male: 59, 9 known female: 102 reliably estimated male: 63, 54, 12, 26 54 has no skeleton intrusive burials: 26, 9, 54 High Intermediate Status Group: 17 total 99, 34, 130, 118, 38, 32, 94, 132, 107, 76, 14, 67, 49, 129, 58, 78, 7 age: 1 child (94) known male: 99, 34, 10 7 known female: 130, 6 7, 7 fair estimated male: 129, 38, 118 fair estimated female: 49, 58 body disposition: 38 on right side, 34 on left side intrusive burials: 58, 32, 78, 129, 132 Low Intermediate Status Group: 32 total 103, 87, 79, 131, 110, 109, 73, 106, 19, 23, 11, 53, 18, 65, 8, 115, 61, 55, 33, 30, 81, 66, 56, 112, 52, 108, 44, 42, 116, 101, 84, 6 known male: 109, 73, 112 reliably estimated male: 103, 110, 106, 19, 11, 61, 66 fair estimated male: 81, 101, 79, 87, 53 known female: 131, 115, 55, 30 body disposition: 81, 52, 109 on left side 73, 84, 110 on right side intrusive burials: 23, 33, 30, 18, 44, 61 Low Status Group: 23 total 48, 120, 114, 90, 86, 71, 51, 43, 41, 36, 29, 97, 82, 80, 89, 88, 62, 45, 119, 27, 91, 21, 20 age: 3 children (114, 36, 89) known male: 91 known female: 82 fair estimated male: 119, 27 body disposition: 45 prone, 9 7 on left side intrusive burials: 62, 43, 71 - 190 -FIGURE 6-5. Location of the four status groups in the Early Period, in terms of the spatial areas derived in Chapter 5. Multiple burials are included. <8> 41 57 a5* 40 I I 46 t I 128 126 tZU EARLY PERIOD BURIAL EU LATE PERIOD BURIAL G3 UNDATABLE BURIAL £k HIGH STATUS © HIGH INTERMEDIATE STATUS si INTERMEDIATE STATUS 6" LOW STATUS - 191 -6.2.4.6. Test Implications for Achieved versus Ascribed Status Status differentiation among the Early period burials seems to reflect aspects of both an achieved and an ascribed system. Unfortunately, a conclusive test of whether status differentiation is based upon age and sex or whether it cross cuts age and sex categories cannot be made. Even after estimating the sex of several burials in Chapter 4, a chi-square test to assess whether the derived status distinctions are different from one another in terms of age and sex (Rothschild 1979) is hot possible due .to low sample size of sexed burials. Thirty-four of the 75 single adult burials in the Early period remain unsexed. However, some insight into whether status differentiation cross-cuts age and sex categories or not can be made by a visual inspection of the age and sex groups within each status group (see Fig. 6-4). Including estimated sexed burials, the high status group contains six males and one female. However, the four distinc tively high status burials are male (E26, 59, 54, 9). This fact may reflect an achieved system of status differentiation because females may not have access to the highest status positions. Burial 5 4 has no skeleton. On the other hand, it is possible that the ivory disk (bl) is a badge of rank that symbolizes only male high status. It is conceivable that the disk could be a marker of a particular achieved status. This possibility is not considered in the literature. The two burials with the high status raw material of elephant ivory - 192 -are E26, a reliably estimated male, and E59, a known male. The ochre and animal shaped pot in E9 may reflect ritualistic status rather than status based on wealth. However, this burial is grouped with the other high status burials in terms of grave goods. The high intermediate status level contains six males and one female. The fact that this female and the female in the high status group (E102) are of higher status than other females in the cemetery suggests that status differentiation in the Early period is ascribed. The burial in this status level with the jade huan is of unknown sex. The low intermediate status group and the low status group contain both males and females, suggestive of ascribed status. The sex composition of the low status group is especially unclear. The fact that three of the four children in the Early period are located in this cluster is suggestive of achieved status. The fourth child, E94, is in the high intermediate status group. This factor is suggestive of ascribed status, as well as the fact that the burial has a grave form that represents the greatest amount of energy expenditure, the log tomb with second level platform. Also, the log tomb is type two, which represents more energy expenditure than the other log tombs among the single Early burials (although the type of log tomb is not described in every case). The other child burials are in simple pits. The small number of child burials in the Early period - 193 -(4 out of 79) inhibits a comparison of mortuary treatment of children versus adults. The position of E94 along Dimension 1 in Fig. 6-1 is quite far to the right, away from many high status burials. Also, if energy expenditure in terms of grave form does symbolize ascribed status for E94, there should be a greater correlation of high status in terms of grave goods and high status in terms of grave form for the adult burials. It is also possible that the grave form of E94 reflects the achieved status of his or her parents. The number of indivudual burials in each status level (including multiple burials) basically indicates a ranking pyramid, expected by many processual mortuary studies to reflect ascribed ranking: eight burials in the high status group, 17 in the high intermediate group, 36 in the low inter mediate group, and 2 4 in the low status group. It is possible, though, that the presence of a ranking pyramid is not a reliable test implication for ascribed status in every case. Suttles (1960:297) states that high status individuals in Coastal Salish communities comprise the majority of the population, not the minority. 6.2.4.7. Multiple Burials It is difficult to compare the status of individuals in the six Early period multiple burials because the grave goods associated with each individual are not apparent. However, the status positions of the individuals in the multiple burials can be broadly assessed. My interpretation of status differentia-- 194 -tion among the multiple burials also indicates that elements of both achieved and ascribed social status are represented in the Early period. In Fig. A6-12, the multiple burials are described in terms of the mortuary attributes included in the analysis of the single burials. Grave 13 seems clearly a high status burial, equivalent to the burials in the high status group. In terms of grave goods, it has many of the traits exclusive to the high status group: the highest quantity of pottery, one elephant ivory carved cylinder and two elephant ivory disks. It also has 14 pig skulls, where the greatest quantity among the single burials is five. Another distinguishing feature is its large size. While the • 2 areas of the other multiple burials range from 2.04 to 2.90m , 2 the area of grave 13 is 6.46m . The largest area of the single 2 burials is 3.82m . Even though two bodies are in E13, the grave area is comparable to the large single graves. Grave 13 is also the only Early period grave with the most elaborate log tomb type, three. Significantly, grave 13 is located in the northern spatial cluster (see Fig. 6-5) with five of the high status burials. The body disposition of both individuals in the grave is supine. The relative statuses of the other multiple burials can be tentatively interpreted as follows: the individuals in E31 can be considered low in status;and E28, E69, Elll, and E35 as intermediate in status. The relative statuses of males and females in multiple burials from Dawenkou Culture sites have received much comment in the Chinese archaeological literature (Pearson 1981:1084-- 195 -1085). Pearson (1981:1084) explains that E35 (and LI) are often cited as indicating low status of females in comparison to males because the majority of grave goods within the grave are placed on the side of the male. The photos of E13 and Elll in the site report also show the majority of grave goods near the male. This disparity is more apparent in E13 and E35 than Elll. Significantly, the male in E13 is associated with the two elephant ivory disks and the elephant ivory carved cylinder. The disks are placed directly on the male's upper chest. How ever, the large quantities of ceramics are placed over the head of both the male and female. The pig skulls are placed on the side of the female. It is difficult to determine from the photographs whether the males and females in each grave were buried at separate times or at one time. The association of the high status items in E13 is support for the possibility that the Early period represents a system of achieved status differ entiation. 6.4.2.8. Conclusions It is not possible to ascertain conclusively whether status differentiation is achieved or ascribed, in terms of the test implications that have been utilized in processual mortuary studies. The Early period appears to have elements of both types of status systems. Rothschild (19 79:672) has similar results at an Archaic period and a Mississippian period site. Regardless of the descriptive term applied to the Early period system of status differentiation ("ascribed" or "achieved") the burials are differentiated into four broad - 196 -status levels on the basis of grave goods. Even though the placement of some burials into certain levels is debatable, the four levels are indicated by the results from both the multidimensional scaling and Ward's Method of cluster analysis. However, the lowest and highest status levels on the basis of grave goods are the most clearly distinguished. The individual burials in six of the seven pairs of intrusive burials are not in the same status level (see Figs. A3-2 and 6-4). Since each status level includes some intrusive burials within the Early period, it is possible that none of the four status levels is restricted to one part of the Early period. It is not clear why the differentiation of burials in terms of grave form and grave size differs from that based upon grave goods. Because some of the high or high intermediate status burials in terms of grave goods also have log tombs, second level platforms, or large graves and none of the low status burials have these traits, it seems that energy expend iture reflects status differentiation at least in part. Although elements of both an ascribed and achieved system of status differentiation are present in the Early, period, I propose that an ascribed system is more likely. The disparity in the quantity and quality of grave goods in the high status group in comparison to the low status group is suggestive of ascribed status. However, the high status group only has exclusive access to a few artifact forms or raw materials. - 197 -Pearson (1981:1086) proposes that the presence of stone orna ments in male and female graves is suggestive of ranking. I think the presence of hairpins, rings, and bracelets in male and female Early graves could reflect ranking. I think that the quantity and quality of grave goods in the high and high intermediate status groups by itself is suggestive of ascribed status. My analysis also depicts a partial segregation of wealthy graves. Pearson (ibid) states that lack of a complete segregation may indicate lack of a hereditary ruling class. Or, lack of a complete segregation may indicate a chiefly office (Pearson 1983:140). Although evolutionary typologies such as those by Fried (1967) and Service (1975) are controversial (see Chapter 2), they have utility for descriptive purposes. The Early period burials may reflect a "ranked" society in Fried's (196 7) sense. In Chapter 5, I argued that all the burials from both periods could reflect a descent group. Fried (1967:116) states that one of the major characteristics of a ranked society is organ ization by kinship in terms of descent. Members of a ranked society are related by descent usually in terms of a lineage or clan (Fried 1967:125-126). However, Fried does not discuss the nature of a ranked.society in terms of its mortuary ritual. One possibility is that the Early period burials represent a conical clan in Kirchhoff's (1955) sense. I argue that Neolithic communities could have kept a record of relations of individuals to particular ancestors of concern by location of graves in the same cemetery in which the ancestors of concern are buried. - 198 -The Early period burials could reflect a chiefdom in Service's (1967) sense. Service (1967:79) includes the con cept of conical clan after Kirchhoff (1955) in his definition of chiefdom. Peebles and Kus (1977:422) point out that the term "chiefdom" includes much variation in degree of socio-cultural complexity. Due to the lack of extreme differentia tion among the majority of burials, the Early period may reflect a simple chiefdom. Individuals of high ritualistic standing are expected in both the "ranked" (Fried 1967:137) and "chiefdom" (Peebles and Kus 1977:422) types of societies. As discussed previously, E9 may reflect ritualistic standing. However, I realize now that E9 is probably dated to the Late period instead of the Early (see Chapter 3). Even though this burial may have been incorrectly classified as Early, I do not think my inter pretation of the high status group of burials changes signific antly. If E9 is a Late period burial, high status during the Early period is not characterized by the animal shaped vessel and red ochrer but by the elephant ivory disk, comb and carved cylinder, and the greatest quantity of ceramic vessels. 6.3. Analysis of Status, Late Period 6.3.1. Data 6.3.1.1. The Multivariate Analyses Twenty-eight variables reflecting grave goods and 32 single burials are included in the multivariate analyses for the Late period. The 32 burials are listed in Fig. A6-13. Three of - 199 -these burials are disturbed: L3, L46, and L77. The 28 vari ables and their attributes are listed in Fig. A6-14. As in the analysis of status for the Early period, black-ware and painted ware are included. A third type of ware appears in the Late period that is likely to reflect high status, white ware. The white ware vessels from Dawenkou are extremely finely made and :. were fired, at. high.temperatures (Shangraw 1977:386). They are extremely brittle and were probably not used for utilitarian purposes (Shangraw, personal communication 1983). White ware is found in apparent high status Shang dynasty graves (Li Chi 1977:202). The color white may have symbolized the Shang state (Chang 19 80:57) or the soul during the Han Dynasty (Yu 1981:83). The frequency of occurrence of mortuary attributes is given in Fig. A6-14. Four forms of ceramic vessels are included in the analysis. The wine vessel, serving stand and tall stemmed cup, but not the bottle (ping), occur in Early burials as well. The appear ance of the bottles from the photographs of the site report, as well as the fact that a few burials contain great quantities, suggest they are high status items. The bottle is not coded in an additive manner because it does not occur in many burials. The serving stand and tall stemmed cup are coded in an addi tive manner. The frequency distributions for all variables coded in an additive manner are in Figs. A6-15 to A6-21. By the Late period there is a slight increase in the frequency of the serving stand in burials and a great increase in the - 200 -frequency and quantity of the tall stemmed cup. The totals of ceramic vessels increase a great deal by the Late period as well. The ornamental forms included in this analysis which are in the Early period analysis are the ring, bracelet, hairpin, hair-tie,, small round stone (bong), and probable ornament parts. There is an increase in the frequency of occurrence of the brace let by the Late period. No huan (the small flat annular piece of stone) occur in the Late period. Since the small flat piece of bone (ban) is in only one Late burial, it must be excluded from the multivariate analyses. Additional ornamental categories in the Late period analysis are the neck or head ornament and the horn zhui (as discussed in Chapter 4, either an ornament or a net weight). In Chapter 5, I concluded the head and neck ornaments are most likely female artifacts and the zhui, a male artifact. The artifact forms with a decorative function which were in the Early period analysis are turtle shell, bone or horn spoon, and the carved cylinder. The latter artifact type occurs with greater frequency by the Late period, in both elephant ivory and bone. Jade occurs more frequently in the Late period but still appears to be a restricted material. It occurs in four burials in five forms: hairpin, bracelet, ring, chisel, and spade. The two jade spades (chan) are unused and were apparently made for symbolic purposes (Gao 1978:33). This form may be a badge of rank. The remaining variables were also in the Early period analysis. Pig skulls do not occur as frequently and there is - 201 -little difference in the quantities among burials. The deer teeth (additively coded) are present in varying quantities. Oyster shell pieces and other pig parts are included. Other additively coded variables are stone tool total, bone tool total, and raw material pieces. As in the Early period status analysis, some of the included variables are slightly redundant. The three ware colors are slightly redundant with the four ceramic forms, and the two additively coded ceramic forms with the total amount of vessels. Upon retrospect, I believe inclusion of the category "jade" created more than a slight redundancy with the hairpin, bracelet, and ring categories. The burials with both jade "present" and either of the three ornamental categories above are: L10 (bracelet, ring), L25 (hairpin), and L117 (hairpin). The possibility that the status of these burials is misrepresented by the multivariate analysis is considered in the interpretation, section. Since the chisel was included in the stone tool totals, the status of the burial with both the chisel and jade present (L4) should not be misrepresented. The unused jade spades were not included in the stone tool totals. 6.3.1.2. Energy Expenditure The Late period burials that have a grave form other than simple pit are marked on Fig. A6-13. Six of the seven log tombs are type three, the type which reflects the greatest amount of energy expenditure. The authors of the site report do not state the type of the seventh log tomb, L10 4. It is implied this tomb is either type one or two. Only one of the - 202 -seven Early period log tombs was a type three. The distribution of grave size for the Late period single burials is in Fig. A6-22. While two distinctions in area of grave were noted for the Early period, three are apparent for the Late. Three burials are of large size, nine of medium size, and 19 small. The area of one grave, L12 7, could not be cal culated due to lack of information provided in the site report. Due to its small relative size on the site map, it is tenta tively called a small burial for the purposes of this analysis. The large size is much larger than the largest single burial 2 2 in the Early period (13-15.00m versus 3.82m in the Early period). Also the medium size range in the Late period encompasses the large size of the Early period burials (2.60-7.99m2). 6.3.1.3. High Status Unique Items The four artifact forms excluded from the multivariate analyses that are considered probable unique, high status items are: one elephant ivory comb (in L10), two semi-circular stone pendants (in L72, huarig) , 84 pieces of alligator hide (in L10), and turquoise in the form of a necklace (in L10) and inlaid in a bone carved cylinder (in L4, described on pages 101-102 of the report). In the western and Chinese archaeological litera ture regarding Dawenkou, the alligator hide in L10 is referred to as crocodile hide. However, the term Alligator sinensis (page 15 7 of the report) refers to alligator, not crocodile. This difference in genus and species is significant because the Chinese crocodile probably ranged only as far north as the - 203 -southern tip of China, while the Chinese alligator probably was present in the lower Yangtze River valley region (Campbell 1981: 37). The source area for turquoise is unknown. Inlaid turquoise items have been found in Shang Dynasty graves (Chang 19 80:345). Huang pendants of jade have been found at Shang and Zhou period sites (the Zu Hai dictionary 1978:315). The huang form may be a badge of rank during the Late period. 6.3.2. Results: The Multidimensional Scaling Nine dimensions resulted from the multidimensional scaling analysis. The first two accounted for 43.04% of the variability in the data set - with the first dimension accounting for 28.25% and the second, 14.79%. The third and fourth dimensions accounted for 12.60% and 8.84% of the variability, respectively. Fig. 6-6 depicts the relationship of the Late period burials when the first two dimensions are plotted together. Like the Early period scaling results, Dimension 1 on Fig. 6-6 can be interpreted as clearly reflecting status differentiation. Burials on the far left side (L10, L25, L117, L47) are inter preted as highest in status, those in the midsection (the majority) as intermediate in status, and those at the right end (L123, L93, L16, L122, L96, L15, L100, L46) relatively low in status. The inferred high, intermediate and low status burials contain the mortuary attributes from Fig. A6-14 that I expected from my exploratory study of status differentiation to reflect high, intermediate and low status, respectively. Fig. 6-6 does not depict any sharp spatial breaks between burials along Dimension 1. The high status burials (L25, L10, L117, L47) - 204 -FIGURE 6-6. Multidimensional scaling plot of Late Period burials in the analysis of status, with clusters from Ward*s Method. large in area (3) medium in area (9) small in area (20) log tomb (7) second level platform (5) 43 disturbed burial (3) 43 controversial former Middle Burial (I) 7 youth (I) la no skeleton (3) 43 high status, unique objects (3) Note: Clusters 1-8 from Ward s Method eight group solution are circled. In the three group solution Clusler I is the same; grouped together are Clusters 2-5 ond Clusters 6-8. In the two group solution grouped logether are Clusters 1-5 and Clusters 6-8. - 205 -are most clearly distinguished. It appears that status dis tinctions among most burials are gradual and not abrupt. Unlike the Early period results, a group of low status burials is not distinguished. The low percentage of variability explained by Dimension 1 may be due to the low number of attributes in common among the Late period burials, as in the Early period. The dimension of variability explained by Dimension 2 is not apparent. Assessment of the mortuary attributes of burials along Dimension 2 does not indicate that Dimension 2 reflects status distinctions. In the analysis of status differentiation for both the Early and Late periods, I do not attempt to interp ret dimensions three and four due to time limitations, the fact that Dimension 1 seems to clearly reflect status distinctions, and the fact that there is fairly good agreement in the status distinctions exhibited by the multidimensional scaling and cluster analysis. 6.3.3. The Three Types of Cluster Analysis Once again, the dendrogram from Ward's Method best reflects status distinctions among the burials. This judgement is based upon assessment of the composition of each burial in terms of the mortuary attributes within the clusters at varying distance levels and from my a priori expectations based upon my explor atory assessment of status differentiation. The dendrograms from the Complete Linkage and Average Linkage techniques exhibit some of the properties that were encountered in the dendrograms of the Early period burials. Again, the average - 206 -Linkage dendrogram compares better with the Ward's dendrogram. The Ward's dendrogram is presented in Fig. 6-7. The clusters from four distance levels in the Ward's dendrogram can be interpreted as exhibiting status relationships among burials: the eight group solution at the 0.6209 distance level, the three group solution at 0.9813, and the two group solution at 1.0018. The attributes characteristic of the eight clusters are shown in Fig. A6-23. At first, the eight clusters appeared distinct in terms of composition of mortuary attributes and seemed to meet my expectations from my exploratory study. However, the three and two group solutions suggest that some of the status distinctions indicated by the eight group solution are not significant. Cluster one contains four burials that are clearly high in status and distinct from other clusters. These burials contain great quantities of ceramics, tall stemmed cups, serving stands, and bone and stone tools. They also contain jade, white ware, and a variety of ornamental types. The three burials that may have been misrepresented are in this cluster: L10, L25, and L117. However, the fact that these burials are distinguished by several attributes as well as jade and jade ornaments indicates the status of these burials may not have been over-represented in the multivariate analyses. L10 and L117 are also distinguished by the presence of high status unique items. - 207 -FIGURE 6-7. Dendrogram of Late Period burials, Ward's Method. I i i i i i i i i I i o I -L2 -L5 -L64 -L77 -L127 -L104 -L105 Cluster 7 -L121 L15 L100 -L46 -L96 -L16 -L122 -L93 -L123 -L124 -L3 -L72 -L24 -L60 -L22 -L75 -L98 -L4 -L125 -L17 -L10 -L25 -L47 -L117 l l I I l I I I l l cor*-u"i ncs Ooor^ m *y cso rH iH rH i-t rH rH O O OO OO Cluster 8 Cluster 6 Cluster 5 Cluster 4 Cluster 3 Cluster 2 Cluster 1 distance - 208 -Cluster two is characterized by the largest quantity of bone tools, stone tools, and raw material. Cluster 4 is characterized by white ware, great quantities of serving stands, and a few ornament types. Cluster 5 is characterized by white ware, the medium quantity of ceramics, and the necklace and other ornaments. Although clusters 2, 4, and 5 are distinct in terms of mortuary attributes, it is difficult to judge whether one cluster represents higher status than another. Clusters 3, 7, and 8 appear distinct in terms of mortuary attributes but all appear third highest in status. Burials in these clusters contain a few ornament types and decorative items. Cluster 6 appears lowest in status, with the lowest quantities of ceramics and bone or stone tools, and few other artifacts, most of which are utilitarian. In the three group solution (see Fig. 6-7), the highest status cluster (1) of the eight group solution is distinguished. The clusters I had interpreted as intermediate and low in status from the eight group solution are grouped slightly differently: clusters 2, 3, 4, 5, and clusters 6, 7, 8, are grouped together. The two group solution is a clear natural break in the dendro gram in Fig. 6-7. Clusters 1, 2, 3, 4, 5, of the eight group solution are grouped together and clusters 6, 7, and 8. The clusters from Ward's eight group solution are plotted on Fig. 6-6. There is a fairly good agreement in the ordering of burials by both the scaling and clustering. The high status cluster (for the three group solutions as well) is located at - 209 -the left end of the plot. Cluster 2, one of the clusters interpreted as next highest in status, is located adjacent to cluster 1 on the plot. The low status cluster, 6, is located at the right end of the plot. The clusters also previously interpreted as next highest in status (4 and 5) are located in the same area along Dimension 1 as Cluster 2. The clusters previously interpreted to represent another low status level, 3, 7, and 8, are located between clusters 2, 4, and 5 and cluster 6 along Dimension 1, as expected. However, there is less agreement between the three and two group solutions of Ward's Method and the scaling plot in Fig.6-6. The high status cluster in the three group solution is distin guished on the scaling plot. However, the other two clusters do not match well with the scaling results. The clusters from the two group solution do not agree well with the scaling results either. 6.3.4. Interpretation of Status Distinctions In the Late period and Assessment of Change through Time 6.3.4.1. Distribution of the Unique Items The correspondence between status differentiation as reflected by grave goods in the multivariate analyses and the probable high status, unique objects is not as good as that among the Early period burials. Some of the objects are found in clusters other than the highest status grouping. The clearly identified highest status group (Cluster 1) in Fig. 6-6 contains - 210 -only one of the three burials with the probable high status unique objects. However, the burial in question, L10, contains three of the high status objects. This burial is shown on the scaling plot as the burial of highest status in Cluster 1. The status of L10 may be exaggerated due to its presence of jade, a jade bracelet, and a jade ring. The status of L25 and L117 may not be as exaggerated because they each contain two redundant categories instead of three (jade and the hairpin). L10 contains the elephant ivory comb, the alligator hide, and the turquoise necklace. L4, located in Cluster 2 (Fig. 6-6) contains the bone cylinder inlaid with turquoise. L72, with the two huang pendants, is in Cluster 5. Ward's Clusters 2 and 5 were inter preted as next highest in status. 6.3.4.2. Argument f>6r Three or Two Status Levels It is even more difficult to infer an approximate number of status distinctions based upon wealth in terms of grave goods for the Later period burials than among the Early. The differ ences in status distinctions as indicated by Ward's eight, three and two group solutions, the multidimensional scaling results, and the distribution of the probable high status unique objects suggest that many burials are not distinctively different in status in terms of grave goods from one another. I infer that either three or two status levels based upon wealth in terms of grave goods may be argued as being represented among the Late period burials. However, the composition of these status levels in terms of burials is open to some debate. - 211 -The burials in Ward's Cluster 1 (L10, L25, L117, L47) should be considered the highest status level because they are distinguished on the scaling plot by the eight and three group solutions from the Ward's dendrogram and by the probable high status objects in L10. I argue that Clusters 2, 4, 5, on Fig. 6-6 should be considered the next highest status level. I originally interpreted these clusters as distinct, but roughly equivalent in status on the basis of their mortuary attributes. Ward's three group solution depicts Clusters 2, 3, 4, and 5 together. The multidimensional scaling plot depicts Clusters 2, 4, and 5 as in the same area along Dimension 1 with Cluster 3 clearly separated from them. I conclude that the multidimensional scaling plot depicts the status relationships among the Late period burials more reliably than Ward's Method. Clusters 2 and 5 contain the other probable high status, unique objects. Clusters 2, 4, and 5 are located adjacent to Cluster 1 on the scaling plot. I interpret the burials in Clusters 8, 3, 7, and 6 in Fig. 6-6 as representing a low status level. Cluster 6 was originally judged as distinct in terms of mortuary attributes. However, many of the burials in Cluster 6 are located in the same area along Dimension 1 as burials in Cluster 7 and 3. Also, Clusters 6, 7, and 8 are grouped together in Ward's three and two group solutions. It is possible that Cluster 1 should be considered grouped with Clusters 2, 4, and 5 to form one status level and Clusters - 212 -8, 7, 3, and 6 to form another. Three of the four burials in Cluster 1 contain the redundant categories (L10, L25, L117). Also, the clearly depicted two group solution in the Ward's dendrogram groups Cluster 1 with 2, 4, and 5. The mortuary attributes that characterize the three status levels are depicted in Fig. 6-8. In the Early period, high status is characterized by exclusive access to only a few artifact forms and raw materials. The differences between the status levels are mainly quantitative, not qualitative. In the high status group, burial 10 is distinguished in terms of its quantity of bottles (38 versus 4 or 5 for the other burials in this status level) and quantity of ceramics (the highest in the cemetery: 9 3). The next highest quantity of ceramics is 71, in L126. Burials L10 and L117 each contain the unused jade chan spade, the possible badge of status. As mentioned previously, L10 also contains the elephant ivory comb, turquoise necklace and alligator hide. Burial L126 is distinguished among the burials of inter mediate status. The position of L126 in the scaling plot is very close to the burials in the highest status group. L126 contains the second highest quantity of bottles: 26. Burial L72 with the stone pendants (huang) and L4 with the inlaid turquoise are in this status group. The disturbed burials in the low status group are L77 and L46. L77 contains an item found only in one other burial, multiple burial LI, six end pieces of arrows. It is possible the arrows symbolize a specific status and that L77 is a high status individual as the authors - 213 -FIGURE 6-8. Characteristics of the Three Status Levels, Late period. intermediate high status status low status (4 burials) (8 burials) (20 burials) + % of tot. + % of tot. + % of tot. white ware 4 40.0 4 40. 0 2 20.0 black ware 4 26. 7 4 26. 7 7 46.7 painted ware 2 33. 3 3 50.0 1 16. 7 wine vessel 1 10.0 2 20.0 7 70.0 serving stand 0 1 8.3 2 16.7 9 75. 0 serving stand 1 0 0 4 28.6 10 71.4 serving stand 2-12 3 50.0 2 33.3 1 16.7 bottle 3 42.9 2 28.6 2 28.6 tall stemmed cup 0 0 0 1 7.1 13 92.9 tall stemmed cup 1-6 1 6.7 7 46.7 7 46. 7 tall stemmed cup 14-16 3 100.0 0 0 0 0 total pots 1-13 0 0 2 10.0 18 90.0 total pots 16-21 0 0 4 66. 7 2 33.3 jade 3 75. 0 1 25.0 0 0 bone cylinder 3 50.0 2 33.3 1 16. 7 ivory cylinder 2 40.0 2 40.0 1 20.0 turtle shell 1 33. 3 0 0 2 66.7 ring 2 40.0 2 40.0 1 20.0 bracelet 4 33.3 2 16.7 6 50.0 hairpin 4 50.0 2 25.0 2 25.0 spoon 1 20.0 1 20.0 3 60.0 stone tool total 0 1 7.1 3 21.4 10 71.4 stone tool total 1-3 2 15. 4 2 15.4 9 69.2 stone tool total 5-19 1 20.0 3 60.0 1 20.0 Cont'd.. . - 214 -FIGURE 6-8 continued intermediate high status status low status (4 burials) (8 burials) (20 burials) + % of tot. + .%. of tot. + % of tot bone tool total 0 3 17.6 3 17.6 11 64. 7 bone tool total 1-6 0 0 2 18.2 9 81. 8 bone tool total 10-27 1 25. 0 3 75.0 0 0 hair tie 2 20. 0 1 10. 0 7 70. 0 neck or head ornament 2 40.0 2 40.0 1 20.0 zhu"i 1 25.0 2 50.0 1 25.0 ornament part 1 33. 3 1 33. 3 1 33. 3 bing 0 0 2 66. 7 1 33. 3 pig skulls 3 42.9 2 28.6 2 28.6 other pig parts 2 22.2 4 44. 4 3 33.3 deer teeth 0 1 16.7 2 33. 3 3 50.0 deer teeth 1-2 1 5.0 5 25.0 14 70.0 deer teeth 3-5 2 33.3 1 16.7 3 50.0 raw material 0 4 15.4 2 7.7 20 76.9 raw material 1 0 0 3 100.0 0 0 raw material 11-•26 0 0 3 100.0 0 0 oyster shell 0 0 2 40.0 3 60.0 - 215 -of the site report suggest. Burials L125, L17, L4 and L72 also contain items not included in the multivariate analyses. Burial L17 contains three jaw bones of the genus Felis, L125 contains two deer leg bones , L4 contains some grains of sand (sha li.) , (from page 27 of the report), and L72 has the small flat piece of bone (ban). The high status level is not characterized by exclusive access to any forms of material except the unused jade spade and the elephant ivory comb. The only raw material exclusive to this level is alligator hide. The only form of artifact present in the greatest quantity is the tall stemmed cup. Even if the two status level interpretation is accepted (combining the high and intermediate status levels, above), there are still only a few artifact forms, raw materials, or quantities of artifacts exclusive to the high status level: the greatest quantity of ceramics, tall stemmed cup, and raw material pieces; the presence of jade, turquoise, alligator hide; the elephant ivory comb, the unused jade spade, and the stone pendant (huang). Many of the potential high status forms, materials, or great quantities of artifacts are found in burials of the low status cluster, such as white ware, black ware, painted ware, great quantities of the serving stand, hairpins, head or neck ornament, and the elephant ivory carved cylinder. The low status burials in the Early period were much more clearly set apart from other burials in terms of grave goods. - 216 -6.3.4.3. Differentiation by Energy Expenditure There is a change from the Early period to the Late in terms of correspondence in status differentiation based upon grave goods with that based upon energy expenditure. There is a better correspondence in the Late period than in the Early, as expected from my exploratory study of the cemetery and by Pearson (1981:1082). In the high status level, three of the four burials contain log tombs with second level platforms. The fourth burial (L47) has the log tomb but no platform. The multidimensional scaling solution depicts this burial as being slightly lower in status than the other three burials of this status level. In the intermediate status level, there are two more burials with log tombs. All the log tombs in the high and intermediate status levels are type three, the most elaborate. The last log tomb is in L104. This log tomb is probably type one or two. L10 4 is in the low status level. However, L10 4 also has a second level platform unlike burials L47, L60, and L126 which are located in higher status levels in terms of grave goods. The last burial with a second level platform is L9 8, in the low status level. Thus, two burials identified as low in status in terms of grave goods have grave forms other than the simple pit. None of the Early period burials identified as low status have log tombs or second level platforms. As in the Early period, the correlation between status in terms of grave size and status in terms of grave goods is not strong. The three large graves (L10, L60, L126) are in the high and intermediate levels. Three of the nine medium-sized graves - 217 -are located in the high status group and two in the intermediate group. Four are in the low status level. Thus, every status level has at least one burial with an area that represents more energy expenditure than the norm. The three burials that are large in size (L10, L60, and L126) also have high-energy grave forms. Grave L10 is distinct because it has the grave form that represents the greatest amount of energy expenditure (the type three log tomb with second level platform). Five burials of medium size have high-energy grave forms: L25, L117, and L104 have the log tomb and second level platform; L47 has the log tomb; and L98 has a second level platform. The high and low status levels are repre sented by . the above burials. In sum, although the correspond ence between status differentiation based upon grave goods and energy expenditure is better among the Late period burials than the Early, each could symbolize a different aspect of status differentiation. 6.3.4.4. Body Disposition and Grave Location The body dispositions other than the common supine position are indicated on Fig. 6-9, as well as age and sex data for each status group. The only two burials with skeletons lying on their sides are in the high status level. These burials are L117, the male youth, and L25, a reliably estimated male. The high status females in this status level (L10 and L47, a reliably estimated female) have the supine body position. It seems that by the Late period, the side body position was reserved for high - 218 -FIGURE 6-9. Age and sex composition of the three status level Late Period. (Body disposition other than supine is marked below, as well as the one subadult burial.) Characteristics of the one Late Period multiple burial are included. High Status Group: 4 total 25, 10, 117, 47 1 youth: 117 known male: 117 known female: 10 reliably estimated male: 25 fair estimated female: 4 7 body disposition: 25 on right side 117 on left side Intermediate Status Group: 8 total 126, 60, 24, 17, 125, 4, 3, 72 no skeleton: 60, 124 known male: 125 known female: 72 reliably estimated male: 17, 4, 126, 24 fair estimated female: 3 Low Status Group: 20 total 127, 77, 5, 2, 64, 98, 75, 22, 121, 105, 104, 12 16, 122, 123, 93, 15, 100, 96, 46 no skeleton: 12 7 known male: 12 3, 122, 15 known female: 121, 105 fair estimated male: 124, 75, 22, 98, 77 fair estimated female: 5 intrusive burials: 123, 124 Multiple burial LI: (1 male, 1 female) pit, medium-sized white ware, blackware, 1-13 pots, 1-6 tall stemmed cups 1 serving stand, turtle shell, hairpin, arrow remnants as in L77, 5-19 stone tools, 10-27 bone tools, 8 deer teeth, probable ornament parts (including one jade tubular bead), 1 bone ban, 5 raw material pieces Judged as intermediate in status. - 219 -ranking males. In the Early period, the burials with the side body position are in a variety of status levels. The spatial location of burials in each of the three status levels is shown in.Fig. 6-10. Unlike the spatial patterning in the Early period, the high status burials in the Late period (status level one) are located in all areas of the cemetery. The northern area only contains one burial of the high status level. If the two status level interpretation is accepted, the majority of high status burials are located in all areas of the cemetery as well. 6.3.4.5. Test Implications As in the analysis of status in the Early period, it is difficult to ascertain conclusively whether status differentia tion cross-cuts age and sex categories to indicate ascribed status. There is only one subadult in the Late period and a minority of burials for which sex is known or estimated. The age and sex composition of the burials in the three status levels is in Fig. 6-9. However, the clear high status of the male youth, L117, allows the proposition to be made that status differentiation in the Late period is ascribed. L117 is more clearly of high status than the child in the Early period, E9 4. L117 is in the high status group and contains the possible badge of status, the unused jade chan. The jade spade is also in the female adult burial L10. L117 is also distinguished from some adult burials in terms of grave form (the type three log tomb - 220 -FIGURE 6-10. Location of the three status groups in the cemetery, Late Period, in terms of the spatial areas discussed in Chapter 5, the one multiple burial included. CD EARLY PERIOD BURIAL EU LATE PERIOD BURIAL £3 UNDATABLE BURIAL A\ HIGH STATUS (72) INTERMEDIATE STATUS 64 LOW STATUS - 221 -and second level platform) and a medium grave size. It is possible that the status of L117 is a reflection of his parents' position. I do not think this possibility is likely given the distinctiveness of L117's status position among the other Late period burials. It is also clear that status differentiation crosscuts sex categories to a greater degree in the Late period than in the Early. Both male and female burials are in the high status level. L10, a known female, was described as consistently dis tinguished from the other burials in the high status level by both grave goods and energy expenditure. It may be the burial of highest status in the Late period. Male and female burials are in each of the other two status levels. The known female L72 has the stone pendant, another possible badge of status. Both male and female burials have grave forms and grave sizes that represent a greater than average amount of energy expend iture. Fig. 6-9 suggests that a ranking pyramid is represented among the Late period burials. The difference between the rank levels, however, is not great. The grave goods, grave form and grave size of the one multiple burial in the Late period, LI, is indicated in Fig.6-9. Due to the presence of jade in the burial, Ll can probably be characterized as similar in status to the burials in the inter mediate status level. As in the multiple burials from the Early period, it appears the status of the female is lower than that of the male because the majority of grave goods are placed by - 222 -the male. However, other females in the Late period are clearly-higher in status than some males. There is only one intrusive pair of burials within the Late period, L123 and L124. Both burials are in the low status group. Due to the limited information regarding time relationships within the Late period, it is not possible to judge whether the three status levels existed throughout the duration of the period. 6.3.4.6. Conclusions In the Late period as well as the Early, high status is characterized by exclusive access to only a few artifact forms, raw materials, or great quantities of objects. The elephant ivory comb symbolizes high status in both periods. However, other forms symbolic of high status had changed by the Late period, from the bi disk and animal shaped vessel to the unused jade chan (and the stone pendant, if the two status level inter pretation is accepted). The raw material types that had changed are elephant ivory and red ochre in the Early period to alli gator hide in the Late (and jade and turquoise if two status levels are accepted). Since jade was not found in the highest status level in the Early period, access to it may have become restricted by the Late period. Elephant ivory was exclusive to high status individuals in the Early period but not in the Late. Great quantities of ceramics characterized the high status Early and Late period burials, if the two status levels are accepted. Another change is that the tall, stemmed cup is exclusive to the high intermediate status group in the Early - 223 -period while in the Late it occurs in the burials of all status levels. However, by the Late period, the highest status level is characterized by the greatest quantity of tall stemmed cup. If E9 (with the animal shaped vessel and red ochre) is a Late period burial, I estimate it to be equivalent in status to the intermediate status level (see Fig. A6-11 for the grave goods in burial 9 and Fig. 6-8). If this estimation is accurate, then the animal shaped vessel and red ochre are not characteristic of high status in either the Early or Late periods. It appears that the nature of status differentiation at Dawenkou changed from a system with elements of both an achieved and an ascribed system in the Early period to a system that clearly represents ascribed ranking in the Late. Thus, an increase in cultural complexity from the Early period to the Late is indicated. This increase in cultural complexity is also suggested by the general increase in quantity and variety of grave goods or raw materials among the burials from the Early period to the Late. There is also an increase in the complexity of grave form by the Late period (the prevalence of the type three log tomb) and an increase in grave size. As in the Early period, few types of grave goods or raw materials are exclusive to the Late period high status burials. However, the Late period high status burials are characterized by high energy expenditure in terms of grave form or grave size more than the Early period high status burials. It may be that high status had become more redundantly expressed by the Late period, such that both grave goods and grave form or size were symbolic of - 224 -high status. The apparent increase in cultural complexity by the Late period is accompanied by few types of grave goods exclusively associated with high status and by a lack of extreme differenti ation between the high and low status burials. Low status burials in the Early period are distinctively different from the intermediate and high status levels. In contrast, the Late period low status burials are not as distinct from other burials. The decrease in the number of status levels represented in the cemetery from approximately four in the Early period to three or two in the Late is unexpected in light of the apparent increase in cultural complexity by the Late period. I argue that these apparent contradictions can be explained if all or most of the Late period burials are regarded as high in status. I propose that burial in the cemetery became restricted to relatively high status individuals by the Late period. The cemetery either became restricted to individuals of one high status level or of two to three high status levels. It is possible that some differences in status among the Late period burials are a factor of the "subordinate" dimension of Peebles and Kus (1977:431). Also the ranking system represented among the Late period burials may have been part of a regional ranking system and cemeteries in the Dawenkou Culture region (perhaps Shandong) may have reflected this regional system. The individuals in the Late period at Dawenkou may represent one status level in the regional ranking system. Goldstein (1980: 136) points out that individual Mississippian cemeteries appear - 225 -egalitarian because each as a whole represents one status level in a regional ranking system. I propose that the trend towards spatial segregation of high status burials noted for the Early period had developed by the Late period such that the entire cemetery was restricted to relatively high status individuals. The Late period burials may represent a "stratified" society in Fried's (1967) sense. According to Fried, status differentiation had increased from the level of ranked society such that there is unequal access to basic resources (1967:186). The stratified society is organized by a mechanism that trans cends kinship systems (ibid). In Chapter 5, I propose that the same descent group persisted in the cemetery from the Early period to the Late. It is possible that the whole descent group represented at the cemetery had become ranked in relation to other descent groups in the Dawenkou Culture region by the Late period. Other Dawenkou Culture cemeteries roughly contemporan eous with Dawenkou may be ranked higher or lower than Dawenkou. The Late period burials could also reflect a fairly complex chiefdom. Whether this chiefdom was as complex as that repre sented by Moundville and other Mississippian sites is unclear. I do not have an explanation for the few child burials in the Early and Late periods, especially in the Late period. A separate location for child burials is expected in achieved status systems. Pearson (1981:1084) notes that child burials at Dawenkou tend to contain fewer ceramics in comparison to adults than at the earlier Dawenkou Culture sites of Liulin. - 226 -It is possible that children from the Late period were not buried in the cemetery until they reached a certain age. There is a child in multiple burial 35, which probably belongs to the Late period (see Chapter 3). Including multiple burials, a change from 85 burials in the Early period to 33 in the Late in itself suggests a change in use of the cemetery, provided the length of time and popula tion base remained roughly the same. The changes through time in status differentiation just described would probably hold regardless of the period to which the 15 undatable burials belong. The relative status of each undatable burial in comparison to Early and Late period burials appears low. The age, sex, grave goods and energy expenditure information for the undatable burials is shown in Fig. A6-24. Seven of the eight burials lacking grave goods in the cemetery are undatable burials. Five of the burials lacking grave goods are children. Like the high status graves in both periods, the empty graves in terms of grave goods (including multiple burial E31) are not restricted to one spatial area of the cemetery. The undatable burials appear relatively low in status in terms of grave goods and energy expenditure. All of the burials are pits and the largest area 2 is 2.49m , smaller than most burials of both periods. There are no photographs of the two multiple burials to allow a comparison of the status of males versus females. - 227 -6.4. Implications Two of the trends through time in status differentiation noted by Pearson (1981) are supported in this analysis (see Chapter 1): an increase in the mean number of ceramics in burials by the Late period and an increase in the percentage of burials with ceramics. The multivariate analyses did not depict greater variation in quantities of ceramics, tools, and ornaments among burials in the Late period compared to the Early as expected by Pearson (1981:1082, 1085). On the basis of my exploratory study of status differentiation in the cemetery, I also had expected greater variation among the Late period burials. Pearson (19 81:10 86) suggests that the Late period does not reflect a highly ranked society and that craft special ization was developing. Also, the status of men in relation to women and children was increasing (ibid). I conclude that the Late period does reflect a highly ranked society. Also, both men, women and children had high status positions (eg., the female burial L10 and the youth, L117). The Wen Wu Correspondent (1978) states that Dawenkou represents the development of private ownership and social classes. I suggest the Late period represents a highly ranked society in which social classes may have developed. My analysis shows the increase in status differentiation among burials from the Early period to the Late noted by the Shandong Provincial Museum (1978) and the Kao Gu Editorial Staff. - 228 -If my proposition is correct that the Early period at Dawenkou represents a ranked society with a trend towards spatial separation of high status graves, and that by the Late period the cemetery may represent one or more high status levels in a regional ranking system, Dawenkou probably does not reflect an incipient system of ranking as Chang (19 79:161) maintains. It is likely that ascribed ranking developed in the Dawenkou Culture region at an earlier date than was previously thought. - 229 -CHAPTER . 7 CONCLUSIONS AND SUGGESTIONS FOR FUTURE RESEARCH 7.1. Conclusions regarding Dawenkou The primary goal of this mortuary analysis has been to understand the nature of status differentiation at Dawenkou in the Early and Late periods, as well as the nature of change through time in status differentiation. Dawenkou has been regarded in the western and Chinese archaeological literature as critical for understanding the development of ranking in the eastern seaboard region. This study, utilizing current arch aeological method in mortuary analysis, could help resolve the controversy in the archaeological literature regarding the nature of status differentiation at the site. The methodology by which to accomplish the primary goal was outlined in Chapter 2. I argued that since status differ entiation is expressed in a variety of manners and to different degrees in ethnographic societies, an investigation should regard all variables in an analysis of status as only potentially status-related. I found that quantity and quality of grave goods distinguished the Early and Late period burials at Dawenkou. However, energy expenditure in terms of grave size and grave form was not clearly status-related in either period. The. relationship between status and energy expenditure was more evident with the Late burials. - 230 -In Chapter 2 I stated that cemeteries must be regarded as reflecting more than one social system through time. Derived status distinctions should not be interpreted as the exact number of status levels that existed in the community of the deceased. Another theoretical reason for not interpreting the exact number of status levels is the partial distortion of social organization reflected in mortuary chains, noted by Hodder (1980) and others. A methodological reason for not interpreting the exact number of status levels is the subjectivity and difficulty in choosing optimum clustering and scaling solutions that reflect status distinctions. Mortuary analyses should attempt to assess change through time in status differentiation. In this study, change from the Early period to the Late is emphasized rather than change within each period. The intrusive pairs of burials within each period are the only source of chronological varia tion of burials within each period. I argued that a relative assessment of change through time in the degree of status differentiation at a site may be more feasible than discussing change in terms of evolutionary typologies. Although it was difficult to characterize the nature of status differentiation among the Early burials, I concluded an increase in the degree of status differentiation had occurred by the Late period. A system of ranking seemed present in both periods, but there appeared to be a qualitative change in the nature of the ranking system by the Late period. If evolutionary typologies are considered, I conclude there may be a change from a "ranked" to a "stratified" society in Fried's (1967) sense or from a simple to a complex chiefdom. - 231 -My understanding of status differentiation at Dawenkou would not have been as complete without the three preliminary analyses. In Chapter 3, the ordering of burials on the basis of changes in ceramic styles, subtypes, and functional types by the authors of the site report was closely replicated by the cluster and scaling analyses. Recent articles in the Chinese archaeological literature conclude that the Middle burials are closer in ceramic style to the Early burials. My analysis showed some Middle burials grouped with the Early and some with the Late. The resultant grouping of burials as Early and Late allowed me to accept two chronological periods for the duration of the study. Since the same typeuof ceramic classifi cation system appears to be utilized at some other Dawenkou Culture sites, chronological analyses of burials utilizing clustering and scaling techniques could be conducted when radio carbon or stratigraphic data are limited. Addition of the estimated sexed burials from Chapter 4 to the sample of known sexed burials enhanced the interpretation of status differentiation in Chapter 6 and the assessment of social subgroup affiliation in Chapter 5. Temporal and geograph ic variation in sexed-linked artifact types could be better understood if the simple inspection method and discriminant analysis were applied to other Dawenkou Culture sites. It would be useful to determine whether males are exclusively associated with a greater number of artifact types than females at other Dawenkou Culture sites. It is possible that this ratio of male to female artifacts reflects increasing status - 232 -of males compared to females in the eastern seaboard region as proposed by Pearson (1981:1086). The analysis of status differentiation in Chapter 6 indicated females with equal status to males. In Chapter 2, I argued that an analysis of status differ entiation is more complete with an assessment of social subgroup affiliation represented at a site. The assessment in Chapter 5, a correlation of spatial location of grave with grave orienta tion, grave form, body disposition and ceramic style indicated much consistency within the cemetery in terms of these variables. There were some changes from the Early period to the Late such as the greater consistency in orientation by the Late period. I proposed that the consistency during both periods in grave form, grave orientation and body disposition from spatial area to spatial area reflects one type of social group. I proposed that a descent group is likely, although test implica tions for a descent group have not been developed. I also proposed that the spatial areas in the Early and Late periods could reflect kin groups or residential groups. The analysis also indicated an increase in number of ceramic forms that are exclusive to a spatial area by the Late period. It would be useful to determine whether this trend through time is a regional one. The analysis of status provided some insight regarding the controversy over the degree of ranking represented at Dawenkou. I concluded the degree of ranking had increased from the Early period to the Late. The system of ranking may have changed - 233 -from one in which members within the descent group at the cemetery were ranked to one in which whole kin groups in the region were ranked. My results tentatively support Pearson's (1981:1086) proposition that lineages were becoming increasingly important by the Dawenkou Culture period. Also, the greater correlation between spatial area and ceramic style in the Late period (see Chapter 5) may indicate that symbolism of kin groups was increasing. 7.2. Future Research The process of understanding development of cultural complexity in the eastern seaboard region must involve assess ment of status differentiation at the regional level. An analysis of status at other Dawenkou Culture sites roughly contemporaneous with the Early and Late periods at Dawenkou in Shandong or Jiangsu in particular, could indicate whether the same trends noted here are apparent at other sites. The develop ment of a regional system of status differentiation in Shandong or Jiangsu should be reflected in other Dawenkou Culture sites. At sites contemporary with the Late period at Dawenkou, one might expect to find the same aspects of mortuary treatment that characterize high status at Dawenkou. The jade chan, a possible badge of status in the Late period, may symbolize high status throughout the region. Possible supralocal symbols (Peebles 1971: 69) include the chan and ceramic design attributes. One could determine whether the other attributes characteristic of high status during the Late period are associated with high status - 234 -items at other Late Dawenkou period sites: white ware, alligator hide, huang pendant, jade, turquoise, great quantities of ceramics (of the tall stemmed cup, in particular). Also one could determine whether burials from later sites show little differentiation in terms of grave goods — indicative of the cemetery being used for one status level. For sites roughly contemporary with the Early period of Dawenkou, one could assess whether the trend towards spatial segregation of high status graves is evident. The Liulin site shows such a trend. Pearson (1981:1083) concludes that the spatial areas at Liulin were a strong determinant of variability in terms of quantities of ceramics. A few mortuary attributes characteristic of high status in the Early period found in roughly contemporaneous Dawenkou Culture sites (such as the bi disk, elephant ivory objects, and great quantities of ceramics) may indicate the initial development of a regional system of status differentiation. Pearson (1981:1085) notes an increase in the degree of status differentiation through time in the eastern seaboard region. However, some of the artifact forms or materials I interpreted as characteristic of high status in either period at Dawenkou have been found in sites earlier than Dawenkou: white ware and painted ware . from the first excavation at Dadunzi in Jiangsu (The Nanjing Museum 1964); the stone huang pendant, jade ax, jade knife and painted ware from the second season at Dadunzi (The Nanjing Museum 19 81); and the jade spade (chan), white ware, turquoise and painted ware at Wangyin - 235 -(Shandong Archaeological Team 1979). Artifact forms or materials found at later Dawenkou Culture sites are: white ware, black ware, turquoise, ceramic bottle, full stemmed cup at the Dafanzhuang site (Archaeological Teamof Linyi County 1975); a jade huan, black ware, white ware at Yedian (Shandong Provincial Museum 1972); and white ware, painted ware, and the second level plat form at Xixiahou (The Shandong Archaeological Team 1964). The presence of these artifact forms and materials in early sites may indicate that access to these forms and materials became restricted only later in time. Status differentiation at these sites may be of a greater degree than currently considered. Instead, the variation in status differentiation among Dawenkou Culture sites may be a factor of geographical differences in resource distribution, for example, or in demography. Mortuary analyses of other cemeteries from varying time periods in the Dawenkou Culture region must be completed so that insight into the processes that may have influenced the development of cultural complexity from the Early to the Late period at Dawenkou is gained. Possible processes include population pressure (Fried 1967:196), agricultural intensifica tion (Pearson 1981:1086), warfare or trade. Warfare is poss ible, given that a burial at Dadunzi has a stone arrowhead embedded in a thigh bone (Yang 1982:58). However, as Pearson(ibid) has pointed out, there is no evidence for violent burial of low status individuals from sites in the eastern seaboard region as at some Longshan sites. Trade of ceramics, jade or elephant - 236 -ivory in either period is possible. If the kiln in Dawenkou cemetery was used during the Late period (see Chapter 1), it could have been used for production of trade wares. Shangraw (personal communication 19 83) suggests that the cemetery had a sacred function by the Late period. The kiln may have been used for production of white ware for purposes of mortuary ritual. The brittle nature of this ware suggests that it was used for ritualistic purposes. My results support the notion of the evolutionary relation ship of Dawenkou with the Shang Dynasty mentioned by Chang (1980:345-346, 19 83:509-510), Thorp (1980:51) and Fried (1983: 488). Many traits characteristic of high status in the Shang period appear to reflect high status at Dawenkou: the log tomb, second level platform, wine vessels (although ceramic, not bronze), jade, white ware and ornaments such as hairpins. I also think that the concern for lineage possibly reflected at Dawenkou is another cultural continuity with the Shang period. Ranked lineages were an important part of the Shang system of sociopolitical organization (Chang 1980:163). Communities consisted of members of common lineages (Chang 1980:161). Line age membership may have been symbolized in decorative styles of bronzes (Chang 1980:165). At their deaths, individuals may have been buried with members of their lineage. At a recent excavation at Anyang described by Chang (1980:369), eight spatial areas of graves were found, each with a distinct grave orienta tion, burial method and pottery assemblage. The investigators maintain that each of these spatial areas represents a lineage (ibid). They also found that there was much variation in grave - 237 -goods within each spatial area, indicating that ranking of individuals existed within lineages. For both periods at Dawenkou I found that the spatial areas within the cemetery contained more than one status level. However, I tentatively interpreted these spatial areas as representing kin groups such as families or residential groups. Chang (1980:235) notes that records of economic activities are lacking in the Shang state and that writing was more concerned with social identification (1980:247). Writing for this purpose may have begun during the Neolithic, at sites in several regional areas of China, includ ing Dawenkou Culture sites (1980:243). According to Chang (1980:161) lineage members symbolized relationship of the lineage to the ancestral line by the Cong ritual chamber. The cong is not present at Dawenkou but it has been found in Neolithic sites from the lower Yangtze River area (Rawson 1980:36). Thus, the concern for lineage may have been well developed during the Dawenkou Culture period. In conclusion, this mortuary analysis of Dawenkou suggests that the Dawenkou Culture region (Shandong province, in parti cular) should be regarded as a dynamic, independent region of cultural evolution during the Neolithic period. My reasoning is based upon the increase in cultural complexity noted for the Early period to the Late, the increase in cultural complexity through time noted by Pearson (19 81) for Neolithic 'sites in Shandong and other provinces of the eastern seaboard region, and the cultural continuities from the Dawenkou site to the Shang period. I maintain the results from the analysis of status - 238 -indicate that Dawenkou and other Dawenkou Culture sites do not represent incipient ranking. It is likely that that ranking was first developed in the eastern seaboard region at a date earlier than Chang (19 79:160) asserts. My results suggest that by the Late period, the Dawenkou site reflects the "Level of Village Aggregates" defined by Chang (1980:361) as characteristic of Longshan cultures in Shandong, Shaanxi, and Henan. It is likely the Late .period burials reflect internal differentiation into poor and rich groups, specialized handicrafts and a socio political organziation with intervillage leaders (as described by Chang 19 79:361). When the nature of regional social organ ization in the Dawenkou Culture region is better understood, attempts can be made to assess the nature of interaction between cultural regions expected by Chang (1981:155) and An (1979-80: 45) for late Neolithic Lungshanoid cultures. 7.3. Methodological Conclusions The secondary goal of this study was to make a methodolog ical contribution to mortuary analysis. My methodology was derived from both the processual and symbolist approaches to mortuary studies. In both the chronological and status analyses, I found that Ward's Method of cluster analysis yielded results that best met my a priori expectations based upon my intuitive study of the cemetery in comparison to Complete Linkage and Average Link age. The results from Ward's Method also compared reasonably well with those from Torgerson's Metric Multidimensional Scaling. In the Late period analysis of status, I judged the scaling - 239 -results to better depict the relationships in the data set than Ward's Method. I recommend the use of Ward's Method and multidimensional scaling in conjunction with each other for mortuary analyses of status differentiation. I conclude that efforts to identify the sex of unsexed burials by means of grave goods and the visual inspection method of Chapter 4 are worthwhile. Discriminant analysis may produce better results with data sets exhibiting more continuous varia tion. Social affiliation in Chapter 5 could only be assessed in an exploratory manner, due to the lack of test implications that would identify one type of social subgroup versus another. It became evident in the status analysis of Chapter 6 that mortuary analyses of status differentiation must have a region al approach, especially when ranking is expected. Also, the test implications which exist for achieved versus ascribed status are probably too simplistic. As Rothschild (1979:659) points out, investigators have assumed a polarity between egalitarian and ranked societies. The Early period at Dawenkou did not appear clearly ranked or egalitarian on the basis of the test implications in the literature. As mentioned in Chapter 2, systems of status differentiation are quite complex in most societies. It was not clear whether the lack of a correspondence between status differentiation based upon energy expenditure and grave goods meant that different aspects of status were being symbolized. It is possible that the correspondence between status differentiation based upon energy expenditure and upon - 240 -grave goods is better the more complex the society. Finally, the multivariate techniques proved extremely useful for the analysis of status differentiation. Thomas (19 78:236) claims that numerical taxonomy does not provide any insight in relationships among items in a data set that visual inspection cannot identify. The ordering of burials that resulted from the clustering and scaling techniques in Chapter 6 indicated a general lack of significant status distinctions among the burials, especially in the Late period. On the basis of my exploratory study, I had expected more extreme status distinctions among the Late period burials. I conclude with Matson (1980:342) that multivariate techniques can be more effective than visual inspection. - 241 -BIBLIOGRAPHY Alekshin, V.A. 19 83 Burial customs as an archaeological source. Current Anthropology 24(2):137-149. Allen, William and James B. Richardson III 1971 The reconstruction of kinship from archaeological data: the concept, the methods, and the feasibility. American Antiquity 36(l):41-53. AnZhimin 1979- The Neolithic archaeology of China: a brief survey 80 of the last thirty years. Translated by K.C. Chang, Early China 5:35-45. 1981 China's Neolithic period. Kao Gu 3:252-260 (in Chinese). 1982 China's Neolithic period. China Reconstructs XXXI (6):57-62. Archaeological Team of Linyi County 1975 Excavated burials from the Neolithic site of Dafanzhuang in Shandong. Kao Gu 1:13-22 (in Chinese). Bartel, Brad 1973 A multivariate analysis of European death ritual. Ethnologia Europaea 7:111-128. 1982 A historical review of ethnological and archaeo logical analyses of mortuary practices. Journal of  Anthropological Archaeology 1:32-58. Bayard, Donn 1983 Rank and wealth at Non Nok Tha: the mortuary  evidence. Paper presented at symposium KIE, XV, Pacific Science Congress, Dunedin, New Zealand. Beijing University Department of Archaeology 19 83 Cultural relics from Shandong province. Kao Gu 3:193-198 (in Chinese). Bender, Susan J. 19 79 Paleodemographic analysis of a Late Woodland site in southeastern Michigan. Mid-Continental Journal  of Archaeology 4(2):183-208. Bettinger, Robert L. 1979 Multivariate statistical analysis of a regional subsistence-settlement model for Owens Valley. American Antiquity 44 (3): 455-470. - 242 -Binford, Lewis R. 1971 Mortuary practices: their study and their potential. In Approaches to the social dimensions of mortuary practices, edited by James A. Brown. Memoirs of  the Society for American Archaeology 25:6-29. 1972 Galley Pond' Mound. In An archaeological perspective, edited by L.R. Binford, pp.390-420. Academic Press, New York. Blackmore, Charles, Mary Braithwaite and Ian Hodder 1979 Social and cultural patterning in the Late Iron Age in southern England. In Space, hierarchy and society, edited by Barry C. Burnham and John Kingsbury. BAR International Series 59:93-111. Braun, David P. 1977 Middle Woodland-early Late Woodland social change  in the prehistoric central midwestern U.S. Unpub. Ph.D. dissertation, Department of Anthropology, University of Michigan. 19 79 Illinois Hopewell burial practices and social organization: a reexamination of the Klunk-Gibson Mound group. In Hopewell archaeology: the Chilli- cothe Conference, edited by David S. Brose, N'omi Greber, pp. 66-79. The Kent State University Press, Kent, Ohio. 1981 A critique of some recent North American mortuary studies; American Antiquity 46 (2):398-416. Brown, James A. 19 71 The dimensions of status in the burials at Spiro. In Approaches to the social dimensions of mortuary, practices, edited by James A. Brown. Memoirs of  the Society for American Archaeology 25:92-112. 1981 The search for rank in prehistoric burials. In The archaeology of death, edited by Robert Chapman, Ian Kinnes, Kla s Randsborg, pp. 25-37. Cambridge University Press. Buikstra, Jane E. 19 76 Hopewell in the Lower Illinois Valley: a regional study of human biological variability and prehistoric mortuary behavior. Northwestern University Archaeological Program Scientific Papers No. 2. Campbell, George R. 19 81 Jaws, too. The story of Sanibel's alligators and  other crocodilians. Sanibel-Captiva Conservation Foundation, Sanibel, Florida. Chang, K.C. 19 79 The archaeology of ancient China. Third edition. Yale University Press, New Haven. 19 80 Shang civilization. Yale University Press, New Haven. 1981a Archaeology and Chinese historiography. World  Archaeology 13(2):156-169. 1981b In search of China's beginnings: new light on an old civilization. American Scientist (69):148-160. 19 83a Sandai archaeology and the formation of states in ancient China: processual aspects of the origins of Chinese civilization. In The origins of Chinese  civilization, edited by David N. Keightley, pp. 495-521. University of California Press, Berkeley. 19 83b Concluding remarks. In The origins of Chinese  civilization, edited by David N. Keightley, pp. 565-581. University of California Press, Berkeley. Chapman, R.W. 1977 Burial practices: an area of mutual interest. In Archaeology and anthropology: areas of mutual interest, edited by Matthew Spriggs. BAR Supple  mentary Series 19:19-33. 1981 Archaeological theory and communal burial in prehistoric Europe. In Patterns of the past:  studies in honour of David Clarke, edited by Ian Hodder, Glynn Isaac, Norman Hammond, pp. 387-411. Cambridge University Press. Chapman,Robert and Klaus Randsborg 19 81 Approaches to the archaeology of death. In The  archaeology of death, edited by Robert Chapman, Ian Kinnes, Klavs Randsborg, pp. 1-2 4. Cambridge University Press. Cheng, Te-k'un 19 82 Studies in Chinese archaeology. The Chinese University Press, Hong Kong. Coles, J. and A. Harding 19 79 The Bronze Age in Europe. Methuen, London. - 244 -Cook, Delia C. 1981 Mortality, age structure and status in the inter pretation of stress indicators in prehistoric skeletons: a dental example from the Lower Illinois Valley. In The archaeology of death, edited by R. Chapman, I. Kinnes, K. Randsborg, pp. 133-134. Cambridge University Press. Decker, Dean A. 1969 Early archaeology on Catalina Island: potential and problems. Archaeological Survey Annual Report 11:69-84. Department of Anthropology, University of California, Los Angeles. Donnan, Christopher B. and Carol J. MacKay 1978 Ancient burial patterns of the Mbche Valley, Peru. University of Texas Press, Austin. Doran, J.E. and F. R. Hodson 19 75 Mathematics and computers in archaeology. Edinburgh University Press. Doran, James 19 71 Computer analysis of data from the la Tene cemetery at Munsingen-Rain. In Mathematics in the archaeo  logical and historical sciences, edited by F.R. Hodson, D.G. Kendall and P. Tautu, pp. 422-431. Edinburgh University Press. 1973 Explanation in archaeology: computer experiment. In Explanation of culture change, edited by Colin Renfrew, pp. 149-153. Gerald Duckworth, Ltd., London. Drennan, Robert D. 19 76a A refinement of chronological seriation using nonmetric multidimensional scaling. American  Antiquity 41 (3)-.290-302 . 19 76b Fabrica San Jose and Middle Formative society in the valley of Oaxaca, edited by Kent Flannery. Memoirs of the Museum of Anthropology, University  of Michigan 8. Fried, Morton H. 1967 The evolution of political society. Random House, New York. 19 8 3 Tribe to state or state to tribe in ancient China? In The origins of Chinese civilization, edited by David N. Keightley, pp. 467-493. University of California Press, Berkeley. - 245 -Gao Guangren 1978 Dawenkou: Neolithic culture find. China Reconstructs XXVII(8):31-35. 1980 A statement proposed for the stages of the Dawenkou Culture. Wen Wu Ji Kan 1:47-65 (in Chinese). Goldstein, Lynne 1976 Spatial structure and social organization:. regional  manifestations of Mississippian society. Unpub. Ph.D. dissertation, Department of Anthropology, Northwestern University. 1980 Mississippian mortuary practices. A case study of two cemeteries in the Lower Illinois Valley. Northwestern University Archaeological Program  Scientific Papers No. 4. 1981 One-dimensional archaeology and multi-dimensional people: spatial.organisation and mortuary analysis. In The archaeology of death, edited by Robert Chapman, Ian Kinnes, Klavs Randsborg, pp. 53-69. Cambridge University Press. Goodrich, David W. 1983 Recent archaeological publications in China. China Exchange News 11(1):15-18. Goody, Jack 1962 Death, property and the ancestors. Tavistock, London. Gowe r, J.C. 1966 Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53:325-338. Greber, N'omi 1979a Variation in social structure of Ohio Hopewell peoples. Mid-Continental Journal of Archaeology 4(1):35-78. 19 79b A comparative study of site morphology and burial patterns of Edwin Harness Mound and Seip Mounds 1 and 2. In Hopewell archaeology, edited by David S. Brose and N'omi Greber, pp. 2 7-38. The Kent State University Press. Gruber, Jacob W. 1971 Patterning in death in a late prehistoric village in Pennsylvania. American Antiquity 36(l):64-76. - 246 -Han Gangxin and Ban Jifeng 19 80 Artificial deformation of the jawbones of Neolithic people at Dadunzi and Wangyin. Kao Gu 2:185-191 (in Chinese, translated by Kazue Pearson). Hatch, James Willis 19 76 Status in death: principles of ranking in Dallas Culture mortuary remains. Unpub. Ph.D. dissertation, Department of Anthropology, Pennsylvania State University. Haviland, William A. 1967 Stature at Tikal, Guatemala: implications for ancient Maya demography and social orgnaization. American Antiquity 32 (3):316-325. Hay, John 19 73 Ancient China. The Bodley Head, London. Hodder, Ian 19 80 Social structure and cemeteries: a critical appraisal. In Anglo-Saxon cemeteries, edited by P. Rahtz, T. Dickinson and L. Watts. BAR British  Series 82:161-169. 19 82a Symbols in action. Cambridge University Press. 19 82b Sequences of structural change in the Dutch Neolithic. In Symbolic and structural archaeology, edited by Ian Hodder, pp. 162-177. Cambridge University Press. Hodson, F.R. 1977 Quantifying Hallstatt: some initial results. American Antiquity 42 (3):394-412. 19 79 Inferring status from burials in Iron Age Europe: some recent attempts. In Space, hierarchy and  society, edited by B.C. Burnham and J. Kingsbury. BAR International Series 59:23-30. Hubers, Louisa G.F. 19 83 The relationship of the Painted Pottery and Longshan Cultures. In The origins of Chinese civilization, edited by David N. Keightley, pp. 177-216. University of California Press, Berkeley. 1981 The traditions of Chinese Neolithic pottery. Bulletin of the Museum of Far Eastern Antiquities 53:1-253. Hull, C. Hadlai and Norman H. Nie 1981 SPSS update 7-9. McGraw-Hill Book Company, San Francisco. - 247 _ Huntington, Richard and Peter Metcalf 19 80 Celebrations of death. The anthropology of mortuary ritual. Cambridge University Press. Jacobsen, T.W. and Tracey Cullen 19 81 A consideration of mortuary practices in Neolithic Greece: burials from Franchthi Cave. In Mortality  and immortality. The anthropology and archaeology  of death, edited by S.C. Humphreys and Helen King, pp. 79-101. Academic Press, San Francisco. Johnson, L.,.Jr. 19 72 Introduction to Imaginary models for archaeological scaling and clustering. In Models in archaeology, edited by D.L. Clarke, pp. 309-379. Methuen, London. Jones, Richard 1980 Computers and cemeteries: opportunities and limit ations. In Anglo-Saxon cemeteries, edited by Philip Rahtz, T. Dickinson, L. Watts. BAR British  Series 82:179-195. Kao, Jeffrey Y.T. 1983 Notes from an archaeological year in China. China  Exchange News 11(1):12-15. Kao Gu Editorial Staff 19 79 Discussion on the nature of the society of Dawenkou, and related problems. Kao Gu 1:33-36 (in Chinese, translated by Kazue Pearson). Keesing, Roger M. 1975 Kin groups and social structure. Holt, Rinehart and Winston, Inc.,New York. Kendall, David G. 1971 Seriation from abundance matricies. In Mathematics  in the archaeological and historical sciences, edited by F.R. Hodson, D.G. Kendall and P. Tautu, pp. 215-252. Edinburgh University Press. King, Linda 1969 The Medea Creek cemetery (LAn-243): an investigation of social organization from mortuary practices. Archaeological Survey Annual Report 11:23-68. Department of Anthropology, University of California, Los Angeles. Kirch, Patrick V. 19 80 Burial structure and societal ranking in Vavalu, Tonga. The Journal of the Polynesian Society 89(3): 291-308. - 248 -Kirchhoff, Paul 1955 The principles of clanship in human society. Davidson Anthropological Journal 1:1-10. Klecka, William R. 1975 Discriminant analysis. In Statistical package for  the social sciences. Second Edition. Edited by N. Nie, CH. Hull, J. Jenkins, K. Steinbrenner, D. Bert, pp. 437-467. McGraw-Hill, San Francisco. Kruskal, J.B. 1964 Multidimensional scaling by optimizing goodness to a nonmetric hypothesis. Psychometrika 29:1-27. Larson, Lewis, Jr. 19 71 Archaeological implications of social stratification at the Etowah Site, Georgia. In Approaches to the social dimensions of mortuary practices, edited by James A. Brown. Memoirs of the Society for American  Archaeology 25:58-67. Leach, Edmund 19 77 A view from the bridge. In Archaeology and anthro pology: areas of mutual interest, edited by Matthew Spriggs, pp. 161-176. BAR Supplementary Series 19. Levine, H.G. 1977 Intra-cultural variability and ethnographic descrip  tion: a decision-making analysis of funerary  behavior among the New Guinea Kafe. Unpub. Ph.D. dissertation, University of Pennsylvania. Li Chi 19 79 Anyang. University of Washington Press, Seattle. Lo Shyh-Charng 19 77 A review of Lungshanoid sites using cluster analysis  and multidimensional scaling. Unpub. M.A. thesis, Department of Anthropology and Sociology, University of British Columbia. Longacre, William 1970 Archaeology as anthropology: a case study. University of Arizona Anthropological Papers No. 17. Lu Zhong 19 76 The emergence of private ownership as indicated by the archaeological finds of the late primitive society from Jiangsu province. Kao Gu 3:165-167, 152,(in Chinese, translated by Kazue Pearson). Luo Gun and Zhang Yongshan 19 79 Change of clan system seen through the Dawenkou Culture. Zhongguo Shi Yanjiu 2:122-131 (in Chinese, translated by Kazue Pearson). - 249 -Ma Chengyuan 19 80 The splendor of ancient Chinese bronzes. In The great Bronze Age of China, edited by Wen Fong, pp. 1-19. The Metropolitan Museum of Art, New York. MacDonald, William K. 1980 The Bang site, Thailand. An alternative analysis. Asian Perspectives XXI(1):30-51. Mainfort, Robert C. 1977 The Fletcher site cemetery (20 BY 28) Bay County,  Michigan: a study in the Social dynamics Of the  contact period. Unpub. Ph.D. dissertation, Department of Anthropology, Michigan State University. Marquardt, William H. 1978 Advances in archaeological seriation. In Advances  in archaeological method and theory, Vol. 1, edited by Michael Schiffer, pp. 257-314. Academic Press, New York. Matson, R.G. 19 74 Clustering and scaling of Gulf of Georgia sites. Syesis 7:101-114. 19 75 Metric multidimensional scaling computer program. Department of Anthropology and Sociology, University of British Columbia. 1980 The proper place of multivariate techniques in archaeology. American Antiquity 45(2):340-344. Matson, R.G. and W.D. Lipe 19 77 Seriation of Pueblo ceramic assemblages from Cedar  Mesa, southeast Utah. Manuscript in the possession of the authors. Matson, R.G. and D.L. True 1974 Site relationships at Quebrada Tarapaca, Chile: a comparison of clustering and scaling techniques. American Antiquity 39 (1) :51-75. Milisauskas, S. 19 78 European prehistory. Academic Press, New York. Morrison, Donald G. 1969 Discriminant analysis. Journal of Marketing  Research 6:156-163. Nanjing Museum 196 4 Trial diggings at Dadunzi in Si Hu Zhen, Bei Xian, Jiangsu province. Kao Gu Xue Bao (2):9-56 (in Chinese). - 250 -Nanjing Museum (cont'd) 19 78 Inquiry into some problems concerning the Neolithic culture in the Yangtze River area. Wen Wu 4:46-57 (in Chinese, translated by Kazue Pearson). 19 81 Second season of excavations from the Dadunzi site. Kao Gu Xue Jikan (1):27-81 (in Chinese). Orme, Bryony 1981 Anthropology for archaeologists. Cornell University Press, Ithaca, N.Y. Orton, Clive 19 80 Mathematics in archaeology. Cambridge University Press. Orton, Clive and F.R. Hodson 1981 Rank and class: interpreting the evidence from prehistoric cemeteries. In Mortality and  immortality. The anthropology and archaeology  of death, edited by S.C. Humphreys and Helen King, pp. 10 3-115. Academic Press, San Francisco. O'Shea, John 19 81 Social configurations and the archaeological study of mortuary practices: a case study. In The  archaeology of death, edited by Robert Chapman, I. Kinnes, K. Randsborg, pp. 39-52. Cambridge University Press. Pader, Ellen-Jane 1980 Material symbolism and social relations in mortuary studies. In Anglo-Saxon cemeteries, edited by P. Rahtz, T. Dickinson, L. Wa11s. BAR British  Series 82:143-159. 19 82 Symbolism, social relations and the interpretation of mortuary remains. BAR International Series 130. Peacock, William 19 76 Cluster analysis and multidimensional scaling of cultural components at Glenrose. In The Glenrose Cannery s*ite, edited by R.G. Matson. Archaeologica 1  Survey of Canada Paper No. 52:214-2 30. Pearson, Michael P. 19 82 Mortuary practices, society and ideology: an ethno-archaeological study. In Symbolic and structural  archaeology, edited by Ian Hodder, pp.99-113. Cambridge University Press. - 251 -Pearson, Richard 1978 The Neolithic cultures of the lower Yangtze and coastal China. Paper submitted to the Scandinavian Institute for Asian Studies for inclusion in the publication of the symposium on The origin of agriculture and technology: west or east Asia. Moesgard, Denmark, in press. 1981 Social complexity in Chinese coastal Neolithic sites. Science 213:1078-1086. 1983 The Qingliangang Culture and the Chinese Neolithic, (written with the assistance of Shyh-Charng Lo). In The origins of Chinese civilization, edited by David N. Keightley, pp. 119-145. University of California Press, Berkeley. Peebles, Christopher 1971 Moundville and surrounding sites: some structural considerations of mortuary practices II. In • Approaches to the social dimensions of mortuary practices, edited by James A. Brown. Memoirs of  the Society for American Archaeology 25:68-91. 19 72 Monothetic-divisive analysis of the Moundville burials: an initial report. Newsletter of  Computer Archaeology VIII(2):1-13. 19 74 Moundville: the organization of a prehistoric community and culture. Unpub. Ph.D. dissertation, Department of Anthropology, University of California at Santa Barbara. Peebles, Christopher and Susan M. Kus 1977 Some archaeological correlates of ranked societies. American Antiquity 42 (3): 421-448. Peterson, Fredric F. 19 81 The excavation of a Bronze Age cemetery on Knighton Heath, Dorset. BAR British Series 98. Pokotylo, David 19 81 Numerical taxonomy and mollusc remains: an inter pretive study from the Ryukyu Islands. In Subsistence and settlement in OkinaWan prehistory—  Kume and Iriomote, Ryukyu Archaeological Research Team, pp.56-71. Laboratory of Archaeology, Depart ment of Anthropology and Sociology, University of British Columbia. 1982 A multivariate analysis of interassemblage varia bility in the Shield Archaic Tradition. In Approaches to Algonquian archaeology, edited by M.G. Hanna and B. Kooyman, pp. 213-231. University of Calgary Archaeological Association. - 252 -Randsborg, Klavs 19 81 Burial, succession and early state formation in Denmark. In The archaeology of death, edited by R. Chapman, I. Kinnes and K. Randsborg, pp.105-121. Cambridge University Press. Rathje, W. L. 1970 Socio-political implications of Lowland Maya burials: methodology and tentative hypotheses. World Archaeology 1 (3):359-374. 1973 Models for mobile Maya: a variety of constraints. In The explanation of culture change, edited by Colin Renfrew, pp.731-759. Duckworth, London. Rawson, Jessica 19 80 Ancient China. Harper and Row, New York. Rothschild, Nan A. 1979 Mortuary behavior and social organization at Indian Knoll and Dickson Mounds. American Antiquity 44 (4) :658-675. Saxe, Arthur A. 19 70 Social dimensions of mortuary practices. Unpub. Ph.D. dissertaion, Department of Anthropology, University of Michigan. 1971 Social dimensions of mortuary practices in a Meso-lithic population from Wadi Haifa, Sudan. In Approaches to the social dimensions of mortuary practices, edited by James A. Brown. Memoirs of  the Society for American Archaeology 25:39-57. Seeman, Mark F. 1979 Feasting with the dead: Ohio Hopewell charnel house ritual as a context for redistribution. In Hopewell Archaeology, edited by David Brose, N'omi Greber, pp. 39-46. Kent State University Press. Service, Elman R. 1975 Origins of the state and civilization. W.W.Norton and Co., Inc., New York. Shangdong Archaeological Team, Institute of Archaeology, CASS 19 79 Excavation of the Neolithic site at Wangyin in Yanzhou, Shandong province. Kao Gu 1:5-14 (in Chinese, translated by Kazue Pearson). Shandong Archaeological Team, IACC 1964 Excavations (first season) at Xixiahou in Ju Fu County, Shandong province. Kao Gu Xue Bao 2:57-106 (in Chinese). - 253 -Shandong Provincial Cultural Properties Commission, and the Jinan City Museum 19 74 Dawenkou, report of the excavation of a Neolithic cemetery. Wen Wu Press, Beijing (in Chinese). Shandong Provincial Museum 19 72 Preliminary report on the trial diggings at the Neolithic burial site at Yedian (Zou xian), southern Shandong. Wen Wu 2:25-30 (in Chinese). 1978 Notes on the Dawenkou Culture. Wen Wu 4:58-67 (in Chinese, translated by Kazue Pearson). Shangraw, Clarence F. 19 77 Early Chinese ceramics and kilns. Archaeology 30 (6):382-393. 1978 Origins of Chinese ceramics. China Institute in America, New York. 1983 Treasures from Shanghai. Archaeology 36(3):60-63,79. Shanks, Michael and Christopher Tilley 19 82 Ideology, symbolic power and riutal communication: a reinterpretation of Neolithic mortuary practices. In Symbolic and structural archaeology, edited by Ian Hodder, pp. 129-154. Cambridge University Press. Shennan, Stephen 1982 Ideology, change and the European Early Bronze Age. In Symbolic and structural archaeology, edited by Ian Hodder, pp. 155-161. Cambridge University Press. Shennan, Susan 1975 The social organization at Branc. Antiquity XLIX (193):279-288. Shepard, John 1979 The social identity of the individual in isolated barrows and barrow cemeteries in Anglo-Saxon England. In Space, hierarchy and society, edited by Barry C. Burnham and John Kingsbury. BAR  International Series 59:47-79. Sneath, Peter H.A. and Robert R.1 Sokal 19 73 Numerical taxonomy. The principles and practices of numerical classification. W.H. Freeman, San Francisco. Sokal, Robert and Peter Sneath 196 3 Principles of numerical taxonomy. W.H. Freeman, San Francisco. - 254 -Spencer, Charles S. 1982 The Cuicatlan Canada and Monte Alban. A study of  primary state formation. Academic Press, New York. Sullivan, Michael 1977 The arts of China. Revised Edition. University of California Press, Berkeley. Suttles, Wayne 1960 Affinal ties, subsistence, and prestige among the Coast Salish. American Anthropologist 62(2):296-305. Tainter, Joseph 19 73 The social correlates of mortuary patterning at Kaloko, North Kona, Hawaii. Archaeology and  Physical Anthropology in Oceania VIII(l):1-11. 19 75a Social inference and mortuary practices: an experi ment in numerical classification. World Archaeology 7(1): 1-15. 19 75b The archaeological study of social change: Woodland  systems in west-central Illinois. Unpub. Ph.D. dissertation, Northwestern University. 19 76 Spatial organisation and social patterning in the Kaloko cemetery, North Kona, Hawaii. Archaeology  and Physical Anthropology in Oceania XI(2):91-95. 1977a Woodland social change in west-central Illinois. Mid-Continental Journal of Archaeology 2(l):67-98. 19 77b Modelling change in prehistoric social systems. In For theory building in archaeology, edited by Lewis Binford, pp. 327-351. Academic Press, San Francisco. 19 78 Mortuary practices and the study of prehistoric social systems. In Advances in archaeological method  and theory, Vol. 1, edited by M. Schiffer, pp. 105-141. Academic Press, New York. 19 80 Behavior and status in a Middle Woodland mortuary population from the Illinois Valley. American  Antiquity 45 (2):308-313. 19 81 Reply to "A critique of some recent North American mortuary studies". American Antiquity 46 (2): 416-420. Tainter, Joseph and Ross H. Cordy 1977 An archaeological analysis of Social ranking and residence groups in prehistoric Hawaii. World Archaeology 9(1):95-112. - 255 -Thomas, David Hurst 1978 The awful truth about statistics in archaeology. American Antiquity 43 (2):231-244. 19 80 The gruesome truth about statistics in archaeology. American Antiquity 45 (2):344-345. Thorp, Robert 19 79 Mortuary art and architecture of Early Imperial  China. Unpub. Ph.D. dissertation, University of Kansas. 1980 Burial practices of Bronze Age China. In The great  Bronze Age of China, edited by Wen Fong, pp. 35-50. Alfred A. Knopf, New York. Torgerson, W.S. 195 8 Theory and methods of scaling. John Wiley and Sons, New York. True, D.L. and R.G. Matson 1970 Cluster analysis and multidimensional scaling of archaeological sites in northern Chile. Science 169:1201-1203. Ucko, Peter J. 1969 Ethnography and archaeological interpretation of funerary remains. World Archaeology 1(2):262-280. Van de Velde, Pieter 19 79 The social anthropology of a Neolithic cemetery in the Netherlands. Current Anthropology 20(1):37-58. Wei Jin 19 76 The earliest class differentiation in ancient China as indicated by the archaeological finds of the primitive society. Wen Wu 8:1-5 (in Chinese, translated by Kazue Pearson). Weiss, Kenneth M. 19 73 Demographic models for anthropology. Memoirs of the  Society for American Archaeology 27. Wen Wu Correspondent 1978 Summary of the symposium on Neolithic culture on the lower reaches of the Yangtze River. Wen Wu 3:35-39 (in Chinese, translated by Kazue Pearson). Whalen, Michael E. 1983 Reconstructing early formative village orgnaization in Oaxaca, Mexico. American Antiquity 48(1):17-43. Whittlesey, Stephanie 19 78 Status and death at Grasshopper Pueblo: experiments  toward an archaeological theory. Unpub. Ph.D. dissertation, Department of Anthropology, University of Arizona. - 256 -Winters, Howard 196 8 Value systems and trade cycles of the Late Archaic in the Midwest. In New perspectives in archaeology, edited by S.R. Binford and L.R. Binford, pp. 175-221. University of Chicago Press. Wobst, H.M. 19 77 Stylistic behavior and information exchange. University of Michigan Museum of Anthropology, Anthropological Paper 61;317-342. Wood, John J. 19 74 A computer program for hierarchical cluster analysis. Newsletter of Computer Archaeology IX(4):l-9. Wright, Gary A. 1978 Social differentiation in the Early Natufian. In Social archaeology: beyond subsistence and dating, edited by C.L. Redman, M.J. Berman, E.V. Curtin, W.T. Langhorne, Jr., N.M. Versaggi, J.C. Wanser, pp. 201-223. Academic Press, New York. Wu Ren 1982 The developing order of prehistoric cultures in Shandong. Wen Wu 10:44-56 (in Chinese). Wu Ruzuo 1982 On types and periodization of the Dawenkou Culture. Kao Gu Xue Bao 3:261-282 (in Chinese). Wu Shanjing 19 73 A brief discussion of the Qingliangang Culture. Wen Wu 6:45-61 (in Chinese, translated by Kazue Pearson). Xia Nai 1979 Carbon-14 dates and the Dawenkou Culture. In A collection of studies that discuss the Dawenkou  Culture, pp.1-12. Wen Wu Editorial Committee (in Chinese). 19 77 Carbon-14, authenticated chronology and the pre historic archaeology of China. Kao Gu 4:217-227 (in Chinese, translated by Kazue Pearson). Yang Hong 1982 China's ancient weapons. China Reconstructs XXXI (1) :58-62. Yu Yingshi 19 81 New evidence on the early Chinese conception of afterlife — a review article. The Journal of Asian Studies XLI(1):81-85. - 257 -Zhang Zhongbei 19 79 Analysis of the remains of the Liulin period of the Dawenkou Culture. Jilin Daxue Xuebao (l):87-95 (in Chinese, translated by Kazue Pearson). FIGURE A3-1. The functi Functional Types b§i (cup) bo (bowl) ding (tripod food vessel) dou (footed serving stand for food) gai (potlid) guan (jar) 1 types, subtypes and styles of pottery at the Dawendou Site. Subtypes Style Numbers cai (painted) - - -gi ta (other) I, II t6ng xing (tubular shaped) I - III gao bing (tall stem) I - VIII dan ba (simple handle)  - V cai (painted) I, II zhe fu (bent body with ridge) I - X qi ta (other) I - XI cai (painted) f II 1 yu'an f u (round body) I - V to guan shi pan (guan-like dish) I - VII shuang ceng pan (double layered dish) I, II xi bing (thin stem) I - V da ldu kong (large cut out holes) I jia (first), I yi (second), CO tong xmg (tubular shaped) cai (painted) II jia, II yi, III, IV I - IX cai (painted) I - III qi ta (other)  - VI yuan fu (round body) I - V zhe fu. (bent body with ridge) I - V cont1d. FIGURE A3-1 continued Functional Types guan (cont/'d) gui he hu (tripod pitcher) (spouted vessel, possibly for wine) (storage vessel, possibly for wine) Subtypes shen fu (deep body) wu zu (no legs) shi zu (solid legs) kong zu (hollow legs) ping di (flat, level base) Style Numbers cai (painted) san zu (3 legs) cai (painted) cai bei (painted with spout at back) shuang bi (2 lugs) wu bi (no lugs) kuan jian (wide shoulders) bdi (1 spout at back of vessel) kui xing qi - -(helmet shaped vessel) pen (basin for water) Ping  (bottle) shou xing qi. - -(animal shaped vessel) wan I, II, IV jia, I - IV I - III I - VI 1/ II I - VI If II I -V I - IX I - VII I - V I, II I, II I - III I, II <_n (bowl) cont'd. FIGURE A3-1 continued Functional Types yi (ladle) zun (wine vessel) zuo (stand, pedestal) Subtypes ping dl (flat or level base) quan zu (ring footed) Style Numbers I, II I, II I - IV Other: these types are given in the burial descriptions, but they are not discussed in Chapter 5 of the site report nor are they included in the photographs. dou zuo (in burials E105, E51) (Functional type - dou ?) ddu pan (in burial L9 8) (Functional type - dou ?) Note: The translations for the above terms are from The Chinese-English Dictionary (The Commercial Press Ltd., Hong Kong, 1979); translation terms compiled by Stanford University, Chang (1981a) and Ma (1980). - 261 -FIGURE A3-2. The 15 pairs of intrusive burials.; L = Late period; M = Middle period; E = Early period; X = Undatable, (dapo = 'cuts into1). Younger Older L10 E26 M9 E23 E33 E62 L15 E33 L24 E30 E18 E31 M44 E43 E54 E58 L123 L124 E32 E61 M16 E61 E78 E129 M121 E132 M121 X133 X70 E71 Note: There is a discrepancy in the burial descriptions for E32, E61, M16. The description for E61 says E36 cuts into it. This description seems to be in error. E32 is described as cutting into E61 and M16 as cutting into E61. Also, Chapter 2 of the site report and the burial descrip tions for X70 and E71 state that X70 cuts into E71. The V site map in the site report (page 4) shows X70 as cutting into Ell. I interpret the burial descriptions as being correct and the map as labelled incorrectly (with burials 11 and 71 switched). All the maps in this study show Ell in the southern part of the cemetery and Ell in the north ernmost, contrary to the site report, page 4. - 262 -FIGURE A3-3. Burials included in the chronological analysis. Early burials from the site report: 6, 7, 8, 11, 12, 13, 14, 18, 19, 20, 23, 26, 27, 28, 29, 30, 32, 33, 34, 38, 41, 43, 45, 48, 51, 52, 53, 54, 55, 56, 58, 59, 61, 63, 65, 66, 71, 73, 76, 78, 79, 80, 81, 82, 84, 86, 87, 88, 89, 90, 91, 94, 99, 101, 102, 103, 106, 107, 109, 110, 111, 112, 114, 115, 116, 119, 120, 129, 130, 131, 132. Middle period burials from the site report: 9, 16, 21, 22, 35, 36, 42, 44, 46, 49, 67, 69, 75, 93, 96, 97, 98, 118, 121. Late period burials from the site report: 1, 2, 3, 4, 5, 10, 15, 17, 24, 25, 47, 60, 64, 72, 77, 100, 104, 105, 117, 122, 123, 124, 125, 126, 127. - 263 -FIGURE A3-4. The 83 Ceramic Forms in the chronological analysis. 1 bei tdng xing I, II 2 bei tdng xing III 3 bei gao bing gao bing I, II, III 4 bei IV 5 bei gao bing V 6 bei gao bing VI 7 bei gao bing VII, VIII 8 bei dan ba III 9 bei dan ba IV 10 bei dan b"a V 11 • k° I 12 ding zhe fu I 13 ding zh'e fu II 14 ding ding zhe fu III 15 zhe fu IV, V 16 ding zhe fu VI 17 ding zhe fu VII 18 .1 ding zhe fu VIII, IX 19 ding zhe fu X 20 ding qi ta I 21 ding qi ta III 22 ding qT ta IV 23 ding qi ta VI, VII 24 ding yuan fu I 25 ding yuan fu II 26 ding yuan fu III, IV, V 27 dou guan shi pan I 28 ddu guan shi pa'n II, III 29 ddu guan shi pa'n IV, V 30 ddu shuang cehg pan II 31 ddu xi bing I 32 ddu xi bing II, III 33 ddu xi bing IV 34 ddu da ldu kong I j ia 35 ddu da ldu kong da ldu kong 1 £1 36 ddu II jia 37 ddu da ldu kong II yi, III 38 ddu da ldu kong IV 39 ddu tdng xing 40 gai I 41 gai II 42 guan qi ta I 43 guan qi ta yucin fu V 44 guan I, II 45 guan yuan fu III, IV, V cont1d... - 264 -FIGURE A3-4 continued 46 guan zhe fu I 47 guan zhe fu II, III 48 guan shen fu I, II 49 guan shen fu III jia 50 guan shen fil I yi 51 guan shen fu IV jia 52 giian shen fu V yi 5 3 guan shen fu V 54 g\li shi zu I, II 55 gai k5ng zu56 gai k5ng zu III 57 h£ ping dl I, II 58 he ping di II59 he' s5n zu60 hti shuang bi I 61 hd shuang bi II 62 ha shuang bi II63 hu wti bi" , II 64 hu wu bi" II65 hu wu bl IV, V 66 hu wd bf VI, VII 67 hu kuan~jfan II 68 hu kuan jfan II6 9 ha kuan jian IV, V 70 ha kuan jIan VI 71 hd bei I 72 hu bei I73 hu bei III 74 hd bei V 75 hd bei V 76 kui xing ql I, II 77 ping II, III 78 w&n I 79 yT~ ^ I, II 80 zun ping dl I, II 81 zun guan zu , II 82 zun guan zd II8 3 zun guan zu IV - 265 -FIGURE A3-5. .Distribution of ceramic categories among the burial sample. 70 -' i i i i i i i i i i i i i t u 01 - 34.7 60 H u o oo cu u co o 01 •§ C 50 4 40 30 A 20 H io ^ 18.2 11.4 7.4 7.4 4.0 3.4 1.2 1.1 1.1 i i i J_J • till I I l l l I i l l 1 0.6 o CM 0.6 -number of times a ceramic category is present among the burial sample - 266 -FIGURE A3-6. Frequency of occurrence of ceramic categories after lumping. 36 i i i i i i i i 2% § CO CO •rl a <U 4J 60 c O c Q) CO cu ex o 60 0) •U CD U s cd I-I <u o cn cu B <4-l O a) a I] 1.2% 20 2.4% 1.2% 15 ~] 1. 10 2% 4.8% 1.2% 3.6% 9.6% 4.8% 16.9% 10.8% 14.5% 18.1% 'i i i i i i i i i i i o o o r-i CN number of occurrences out of 83, the total number of ceramic categories after lumping - 267 -FIGURE A3-7. Distribution 'of ceramic categories after lumping. i i i i i i i i i i • 25 1 1 I 1-7% 20 CO cu •H u o 60 CU 4-1 rt CJ o •H u cu o u cu ! 15 1.7% 0.9% 10 3.5% 1.7% 5.2% 3.5% 6.1% 5.2% 9.6% 8.7% 11.3% 9.6% ~\ 1 1 1 r o rH 15.7% 13.9% i r o number of burials out of 115 - 268 -FIGURE A3-8. Composition of clusters from Ward's dendrogram (See Fig. 3-2) in terms of ceramic styles. Cluster 1 E90, E88, E86, M42. (4 total) ding zhe fu III (4 occurrences). Cluster 2 E109, E65, E23, E132, E34, E131, E110, E14. (8 total) dou da lou II jia (8), dou guan I (2), hu bei I (2), ding yuan II (3) , ding zlie fu III (3) , guan shen III jia (1). Cluster 3 E119, E89, E41, E112, E38, E116, E32, E33, E29, E103, E19, E7. (12 total) guan yuan I, II (11), guan yuan III-V (1), guan zhe I (3), guan shen III jia (1), guan shen IV jia (1), hu shuang I (1), hu shuang III (1), hu wu I, II (3), hu bei II (1), bo I (2), ding zhe I (4), ding zhe III (8), ding qi I (3), ding qi III (1), ding qi VI, VII (3) , ding yuan I (1) , doti da lou I yi (2) , dou da lou II jia (3), dou da lou II yi, III (1), wan I (2). Cluster 4 E71, E52, E101, M96, M69, E27, E80, E120, E43, E115, E61, E114, Ell. (13 total) dou guan I (1), dou da lou II yl, III (2), hu shuang I (1), hu shuang III (3), hu bei I (2), hu wu III (2), yi I, II (2), wan I (1), guan shen V (4), ding qi I (3), ding qi VI, VII (2), bei tong I, II (1). Cluster. :5 E106, E107, E82, E51, M21, E102, E59, E54, E26, E91, E79, E56, M36, E63, E58, E129, E13. (17 total) ding zhe I (1), ding zhe II (3), ding zhe III (5), Cont'd... - 269 -FIGURE A3-8 continued Cluster 5 cont'd. ding yuan I (10), ding qi III (1), ding qi IV (2), gui shi I, II (9), gai I (5), gai II (4), bei tong I, II (4), guan yuan I, II (4), guan yuan III-V (1), guan shen I, II (3), guan sh~en IV jia (3), guan shen IV yl (1), guan zhe I (1), kui xing qi I, II (2), wan I (3), hu wu I, II (2), hu wu III (1), hu shuang I (4) , hu shuang II (3) , hii shuang III (1) , zun guan I, II (6), zun guan III (1), zun ping I, II (1), dou da ldu I jia (3), dou da lou I yiL (2), dou da lou II jia (4), dou da lou II yi, III (1), dou da lou IV (4), ddu xi I (1), dou guan I (1), dou guan IV, V (1), he ping I, II (4). Cluster 6 E55, M44, M97, E73, M49, E28, E76, E45, E18, 1 E84, E48, E87, E12, E130, Elll, E99, E20, E66, E8, E81, E53, E94, E78, E6. (25 total) ding zhe I (7) , ding zhe II (3), ding zhe III (12), ding yuan II (5), ding qi III (1), ding gf IV (1), bo I (1), bei gap I-III (2), bei gao V (1), bei long I, II (3), kui xing qi I, II (1), gui shi I, II (3), gai I (2), gai II (1), he" ping I, II (1), hu shuang I (1) , hu shuang II (3), hu shuang III (1), hu bei I (4), hu bei II (5), hu wu I, II (14), zun guan I, II (3), guan yuan I, II (1), giian zhe I (1), guan shen I, II (2) , guan shen III jia (5), guan shen IV jia (1), guan shiBn IV yi (1) , guan shen V (1) , dou guan I (1) , dou guan II, 111(2), dou guan IV, V (1), dou da lou I jia (8), dou da lou I yi (1), ddu da lou II jia (2), dou da lou II yl, III (4). Cont'd... - 270 -FIGURE A3-8 continued Cluster 7 L60, L25, L126, L117, L10. (5 total) bei gao VI (4), bei gap VII, VIII (3), bei dan III (5), bei dan V (3), bei tong III (1), wan I (1), yl I, II (1), gai I (5), gai II (1), he san II (1), gui kong III (3), ding zhe X (5), guan zhe II, III (2), guan qi V (2), ping II, III (4), dou t8ng (1), ddu shuang II (3), ddu xi II, III (2), hu bei III (1), hti* bei IV (3), hu bei V (3) , hti shuang I (1) , hu wu VI, VII (2) , hu kuan IV, V (4), hu kuan VI (1). Cluster 8 L105, L77, L125, M121, M98, M67, L24, M93, M46, L124, L100, L122, L17, M118, L15, M75, M9, M35, M22, L5, L104, L127, M16, L72, L4, .u ' L47, L3, L123, L2, L64, LI. (31 total) be"! tong I, II (10) , bei tdng III (6) , bei gao I-III (3) , bei gao IV (8) , bei gao V (7) , bei gao VI (5) , bei dan III (3), bei dan IV (5), bei dan V (3), bo I (1) , wan I (2), yl I, II (1), gai I (4), ping II, III (2) , he san (2), he ping III (4), gui kong II (3), gui k5ng III (1) , ddu tong (2) , ddu xi I (2) , ddu xi II, III (6), dou xi IV (10), ddu guan I (1), ddu guan IV, V (1), zun guan I, II (4), zun guan III (6), zun qdan IV (3), zun ping I, II (3), hu wu I, II (2), hu wu III (1), hu wu IV, V (12), hu wu VI, VII (1), hu kuan II (5), hu kuan III (3), hu kuan IV, V (2), hu kuan VI (2), hu bei III (17), hu bei IV (3), ding  yuan III-V (3), ding zhe III (4), ding zhe IV, V (5), ding zhe VI (3), ding zhe VII (13), ding zhe VIII, IX (3) , ding zhe X (2) , ding qi" III (1) , ding qi IV (1) , / guan qi I (5), guan qi V (2), guan zhe I (3), guan zhe II, III (6), guan yuan III-V (8), guan shen III jia (4), guan shen III yl (3), guan shen IV yl (7), guan shSn V (7) . - 271 -FIGURE A4-1. List of sexed and unsexed burials: Early, Late, and Undatable. Early period: 85 total (10) Female: 7, 28, 30, 55, 67, 82, 102, 115, 130, 131 (9) Male: 9, 34, 59, 73, 91, 99, 107, 109, 112. (4) Children : 36, 89, 94, 114. (57) Unsexed single burials: 6, 8, 11, 12, 14, 18, 19, 20, 21, 23, 26, 27, 29, 32, 33, 38, 41, 42, 43, 44, 45 , 48, 49, 51, 52, 53, 54, 56, 58, 61, 62, 63, 65, 66, 71 , 76, 78, 79, 80, 81, 84, 86, 87, 88, 90, 97, 101, 103, 106, 108, 110 , 116, 118, 119, 120, 129, 132. (5) Multiple burials: 13 male, female 35 male, female, child 111 male, female 31 2 unsexed adults . 69 2 unsexed adults Late period: 33 total (4) Female: 10, 72, 105, 121. (4) Male: 15, 122, 123, 125. (1) Subadult: 117. (23) Unsexed single burials: 2, 3, 4, 5, 16, 17, 22, 24, 25, 46, 47, 60, 64, 75, 77, 93, 96, 98, 100, 104, 124, 126, 127. (1) Multiple burial: 1 : male, female Undatable burials: 15 total (2) Female: 57, 85. (6) Children: 37, 39, 50, 68, 74, 95. (5) Unsexed single burials: 40, 83, 113, 128, 133. (2) Multiple burials: 70 : two unsexed adults 92 : two unsexed adults - 272 -FIGURE A4-2. The 129 artifact types in the Discriminant Analysis, from the 2 9 known single, sexed burials. whetstone fish hook tusk knife stone knife stone chisel bone chisel arrowhead spoon, spatula head or neck ornament small round stone net weight or pendant elephant ivory disk hairpin (styles I, II, IV) ax (style II) spade (III, IV, V) ring (I, II) bracelet (I, IV) awl (I, III) adze (small I, II; medium II, III, IV; large II) painted tool (II, III) elephant ivory or bone carved cylinders (I, II, III) spearhead (I) hammer spindle whorl needle needlecase hairties sickle raw material pieces of tusk bone flat piece, slab deer teeth pig skull small flat piece of stone elephant ivory comb bone raw material pieces pig incisors pieces of pig bone lower jaw bone of pig three legged ceramic vessel with bent body (I-VII, X) three legged vessel, "other" (III, IV, VI, VII) three legged vessel with round body (I, II) serving stand, jar shaped (I, III, IV) Cont*d. - 273 -FIGURE A4-2 continued serving stand, thin stem (I-IV) serving stand with large cut out holes (I yi, II jia, III, IV) storage vessel with spout at back of vessel (II, III/ IV) storage vessel with no lugs (I, III, IV, VII) storage vessel with two lugs (I, II) storage vessel with wide shoulders (III, VI) jar with bent body (I, II) jar with round body (II, IV) jar with deep body (II, III jia, III yi, IV jia, IV yi, V) other shaped jar (I, V) simple handled cup (III, V) tall stemmed cup (IV, V, VI, VIII) tubular shaped cup (I, III) jug with hollow legs (II, III) tripod pitcher with solid legs (I) spouted vessel with flat base (I, III) spouted vessel with three legs (II) spouted vessel painted pot lid (I, II, VII, IX) basin (I) bowl (wan I) wine vessel with ring foot (II, III, IV) helmet shaped vessel (II) bowl (bo I) bottle (III) do,u zuo Note: the style numbers for the artifacts are those from the classification system in the site report; some types consist of a variety of materials. - 274 -FIGURE A5-1. Orientation of graves, Early Period i i i i i i i i i 270 JJ T3 O •H U <D J-I CO W CO CU > cfl ri 60 o CO 0) CU u 60 cu T3 O <u •H M O n o r t h e a s t t -o-w e s t w e s t n o r t h e a s t 121-125 116-120 111-115 106-110 101-105 96-100 91- 95 86- 90 81- 85 76- 80 71- 75 66- 70 61- 65 56- 60 ~ 51- 55 46- 50 41- 45 J] i IT number of burials - 275 -FIGURE A5-2. Orientation of graves, Late Period CO a) > CO u 60 CO 0) CU T3 U O 60 -H <U U T3 CU •H <U 4J C CO u cfl c 01 •H t-i o n o w r e t s h t "e a s t t _Q_ n o r t h w e s t iir-ii5 106-110 101-105 96-100 91- 95 86- 90 81- 85 76- 80 e a s t 1 o I o number of burials FIGURE A5-3. Distribution of ceramic styles among the spatial locations derived in Chapter 5 for the Early Period burials (see Fig. 5-4). Ceramic style: cr H M H H W M H w O o O n n fl cn 00 M CTl I—1 H-1 M h-1 Hi 0 OJ VO VD OJ cn fl fl fl fl A CTl 00 o > cn cn cn cn cn Oj pj rt rt rt rt rt H rt M M H M M CD CD CD CD CD CD CTl CO 00 M -J l-i Hi Hi Hi Hi W & -J o OJ OJ H-1 H1 r—1 to OJ tubular shaped cup I x x x tubular shaped cup V  x bowl (bo) I xxx bent body tripod I  x x x x 1 bent body tripod II x  x CTl bent body tripod III  xx xxxxx , other tripod I xxx other tripod II  x other tripod III x x other tripod IV  x other tripod VI  x other tripod VII  x painted tripod I x x round bodied tripod I  x x x x round bodied tripod II x xxxxx serving stand with  x x x guan dish I Cont'd... - 277 -Cluster 5 Cluster 4 Cluster 3 Ks* KA K/t K/l KA KA Kj* K/t * rS A KN A >A •NAKN TJ" 4-1 A o c_> Cluster 2 Cluster 1 •V S** M M rN t/N rS rS •V* ly* -V KN rN KN E66, E73 E130,E131 E81, E82 E:79/ E80 E49, E67 X X X X E48 E36 TJ cu fl •rH 4-> a o o Isolated burial E14 ro I IT) < D U H &H .fl 4-> •H H & H H TJ C .fl rrj to 4-> -H cn TJ tn fl fl JO -H *3 > M CU CD * tn! .fl 4-) •H & TJ C rrj 4-> CO TJ tn fl fi ,rd -rl^ > tn H OJ CO .fl 4-1 •H £ H H TJ fl g rrj CU 4-1 4-> CO CO tn C •H > U CU CO fl; •si X! X! — 4-> 4-> •H CO S M •H TJ m fl — (0 -P H 10 CO tn CU fl H TJ fl O o CU CO o .fl Xi 4-> -H •5 TJ fl — rtJ 4-> H CO CO tn CU fl rH •H O > Xi H CU CO Xi -P •H TJ C rd H 4-> H CO CO tn CU C 4-1 CO H •H m o .fl X! 4J •H £ TJ c (TJ 4-1 CO CO tn CU fl TJ C O O CU CO H o .fl 4-> •H TJ fl rrj 4-> H CO to tn CU fl <—I •rl > H CU CO H o X! X! 4-) •H & TJ C rrj > 4-1 H CO CO tn cu fl •rH > H CU CO o Xi TJ •H rH 4-1 o a> H TJ •H rH 4-> O 04 FIGURE A5-3 continued tr H C tn n o H- H CU fu I—1 rt CD M a, h-1 it-painted jar jar with round body I jar with round body II jar with round body III jar with bent body I jar with bent body II jar with deep body II jar with deep body III (first) jar with deep body IV (first) jar with deep body IV (second) jar with deep body V tripod pitcher with solid legs spouted vessel with flat base I *painted spouted vessel painted storage vessel II M M H M H O o o o n -0 CO h-1 r—1 t-1 CO CD M U) CTi c O CO w en cn ^ rt ft rt rt rt H W H M H (D n> ro (D CD CTi 00 CO H -J r< H r-i h o to 00 U) h-1 H" to Ul X X X X X X X X X X X X X X X X X X X X X X X X to 00 X X X X X X X X X X X X X X X X Cont'd. FIGURE A5-3 continued cr H C co M 1-1 0 u> H- M CTi CD fu fu H rt CD M Cb painted storage vessel with spout I storage vessel with two lugs I storage vessel with two lugs II storage vessel with two lugs III storage vessel with no lugs I storage vessel with no lugs II storage vessel with no lugs III storage vessel with no lugs IV storage vessel with spout I storage vessel with spout II w M a M w o o o o o 00 h-1 CTi h-1 h-1 H M h-1 co CO r-1 co CTi 0 d N o CO CO CO CO CO rt rt ft rt rt M M M H H CD CD CD CD CD CTi 00 00 r—' l-i r-i H o Ni CO CJO h-1 M to U) *>. Ul X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Cont'd... FIGURE A5-3 continued tr H M W W H C CD H M -J 00 I—1 n 0 H H- r—1 CO •* o pj pj M rr H M M H CD CTi oo CO M H Ch o to U) I—1 \-> *» storage vessel with spout III x helmet shaped vessel II basin I x bowl (wan) I ladle II *wine vessel with flat base I wine vessel with ring foot I wine vessel with ring foot II x wine vessel with ring foot III stand, pedestal x Note: Styles that occur in only one burial are excluded x = present * = style exclusive to one spatial area - 281 -FIGURE A5-4. The distribution of ceramic styles among the spatial locations derived in Chapter 5 for the Late Period burials (see Fig. 5-5). rH CN Ceramic style: Cluster Cluster Cluster Cluster *other cup II X tubular shaped cup I X X X tubular shaped cup II X X tubular shaped cup III X X X tall stemmed cup III X X tall stemmed cup IV X X X tall stemmed cup V X X tall stemmed cup VI X X X X *tall stemmed cup VII X tall stemmed cup VIII X X tripod with bent body IV X X tripod with bent body VI X X tripod with bent body VII X X X tripod with bent body VIII X X tripod with bent body X X X X tripod other IV X X tripod other V X X *tripod with round body IV X *serving stand with layered dish I serving stand with layered dish II X X X serving stand with thin stem I X X serving stand with thin stem II X X serving stand with thin stem III X X Cont1d... - 282 -FIGURE A5-4 continued serving stand with thin stem IV *serving stand with thin stem V tubular shaped serving sta potlid I potlid VII potlid VIII potlid IX other jar I other jar V jar with round body IV jar with bent body II jar with deep body III (first) jar with deep body III (second) jar with deep body IV (second) jar with deep body V *tripod pitcher with hollow legs I tripod pitcher with hollow legs II tripod pitcher with hollow legs III *spouted vessel with level base III spouted vessel with three legs II painted storage vessel VI •H CN ro H H u H CD CD CD CD -P 4-> 4-1 4-> CO (0 CO CO 3 3 3 •H rH rH rH CJ CJ CJ CJ X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Cont'd... FIGURE A5-4 continued storage vessel with two lugs IV storage vessel with no lugs IV *storage vessel with no lugs V storage vessel with no lugs VII storage vessel with wide shoulders II storage vessel with wide shoulders III storage vessel with wide shoulders IV storage vessel with wide shoulders V storage vessel with wide shoulders VI storage vessel with spout III storage vessel with spout IV storage vessel with spout V bottle III bowl (wan) I *ladle I *wine vessel with flat base II *wine vessel with ring foot II - 284 -FIGURE A5-4 continued wine vessel with ring foot III wine vessel with ring foot IV cup with simple handle III cup with simple handle IV cup with simple handle V x x CN ro u u H u OJ 'OJ OJ CD 4-> 4-> •P 4-> co CO CO CO rH rH rH rH u P u P X X X X X X X X X X X X Note: Styles that occur in only one burial are excluded. x = present * = style exclusive to one spatial area. FIGURE A6-1. The 79 burials in the Early Period, analysis of status: age and sex, grave form, grave size. 6 7 8 9 11 12 14 18 19 20 21 23 26 27 29 30 32 33 34 36 38 41 42 43 44 45 48 49 F LA M, ER, LA RM, ER, LA RM, ER, LA RM RM D, FM M C ER, FM KEY total (4) C — child (adults unlabelled) (9) F = known female (9) M = known male (1) NS - no skeleton (11) RM = reliably estimated male (10) FM = fair estimated male (2) FF = fair estimated female (1) D = disturbed (6) LT — log tomb (12) ER — second level platform (7) LA = large in size Simple pits and small graves undesignated. FF Cont1d. FIGURE A6-1 .continued 51 90 52 1 M 53 LT, ER, FM 94 C, LT, ER 54 NS, RM 7 55 F 99 M, LT, LA 56 101 FM 58 ER, FF 102 F 59 M, ER, LA 103 61 RM 106 RM 62 107 M, LT 63 RM 108 65 9 M, ER 66 RM 110 RM 67 F, ER 2 M I 71 114 C NJ 73 M 115 F » 76 6 LT 78 ER, LA 118 FM 1 79 FM 119 FM 80 120 81 LT, ER, FM 129 FM 82 F 130 F 84 131 F 86 2 87 FM 88 89 C FIGURE A6-2. The 2 3 variables and their attributes in the Early Period analysis of status. (Coded presence/absence or by additive coding - see text for explanation.) Variables Presence/Absence black ware 26 (32. 9%) / 53 (67. 1 & ^ o / painted ware 17(21. 5%) / 62 (78. 5%) wine vessel 9 (11. 4%) / 70 (88. 6%) tall stemmed cup 3 ( 3. 8%) / 76(96. 2%) bone or elephant ivory carved cylinder 5(6. 3%) / 74 (93. 7%) turtle shell 5 ( 3. 3%) / 74 (93. 7%) bone or stone ring 8 (10. 1%) / 71 (89. 9%) stone, bone or ceramic bracelet 5( 6. 3%) / 74 (93. 7%) i stone hairpin 8 (10. 1%) / 71 (89. 9%) bone or horn spoon, spatula 13 (16. 5%) / 66 (83. 5%) 00 -~J hair tie (pair or one half) 8 (10. 1%) / 71(89. 9%) probable ornament parts: 5 ( 6. 3%) / 74(93. 7%) 1 bone or stone tubular bead - guan thin, flat piece of stone - pi an stone annular object - huan stone circular bead - zhu small, round flat stone (bfng) small, flat piece of bone (ban) 3( 3. 8%) / 76 (96. 2%) 7( 8. 9%) / 72 (91. 1%) other pig parts 9 (11. 4%) / 70(88. 6%) oyster shell pieces 6( 7. 6%) / 73 (92. 4%) Cont1d... FIGURE A6-2 continued Additive Coding serving stand total number of ceramic vessels stone tool total bone tool total pig skulls deer teeth pieces of raw material: bone, horn, or tusk 0/1-2/3-9 : 34(43.0%)/41(51.9%)/4(5.1%) 0-7/8-14/19-30 : 6.4.(81. 0%)/12 (15. 2%)/3 (3. 8%) 0/1/2-7/9-16 : 45(57.0%j/22(27.8%)/9(11.4%)/3(3.8%) 0/1-7/13-25 : 51(64.6%)/24(30.4%)/4(5.1%) 0/1-3/4-5 : 49 (62.0%)/25 (31. 6'%)/5 (6 . 3%) 0/1-3/4-12 : 27(34.2%)/46(58.2%)/6(7.6%) 0/1-10/16-40 : 68(86.l%)/8(10.l%)/3(3.8%) - 289 -FIGURE A6-3. Distribution of serving stands, Early Period. 30 -' 20 H number of serving stands (dou) - 290 -FIGURE A6-4. Distribution of pottery vessels, Early Period, j i i i i i i i i i i i i i i i i i II '-D 25 5 20 M PM u W to 8 15 CO 01 > u 01 4-1 o 14-1 o u CU ! 10 I 1 i i i i l I I l i i l i I i I i i I i' a in o m o CN number of burials - 291 -FIGURE A6-5. Distribution of stone tools, Early Period. 20 10 H i i i i i stone tool totals, Early Period - 292 -FIGURE A6-6. Distribution of bone tools, Early Period. t i i i i i i i 3 0 crj 4J O 25 T3 O •H U CU 20 u CO W CO rH o o cu c o rQ o rJ <U ! 15 10 o CM I i number of burials - 293 -FIGURE A6-7. Distribution of pig skulls, Early Period. 12 -i i i i i i i 8 A cn rH CO •rt U 3 X> o 4 u a) •i 3 a i i i i i i i number of pig skulls - 294 -FIGURE A6-8. Distribution of deer teeth, Early Period. _i i I i i i i i i i I I _ !_ 20 -r 15 -CO rt 10 •H U 3 rO o u CU •§ 3 5 -I I I O number of deer teeth, Early Period - 295 -FIGURE A6-9. Distribution of raw material pieces, Early Period. T3 o •H u cu Hi 40 u Cu w 20 o cfl •H C H O rC c o rQ 15 to 3 rt •H U <D 4-1 e cti o CO cu o (U •H & M-l O rJ cu ! number of burials - 296 -FIGURE A6-10. Grave area of Early Period burials. » i i i i i i i i i 3.50 3.00 2.50 2.00 1.50 1.00 0.50 i o i c number of burials - 297 -FIGURE A6-11. The mortuary attributes characteristic of the clusters from the seven group solution of Ward's Method, Early Period burials. (Underlined addi-tively coded attributes represent the crreatest quantity of that artifact type.) Cluster 1 Burial 48 0-7 pots, probable ornament parts 120 0-7 pots 114 0-7 pots 90 0-7 pots 86 0-7 pots 71 0-7 pots 51 0-7 pots 43 0-7 pots 41 0-7 pots 36 0-7 pots 29 0-7 pots Cluster 2 Burial 97 1-2 serving stands,1-3 deer teeth, 0-7 pots 82 1-2 serving stands, 1-3 deer.teeth, 0-7 pots 80 1-2 serving stands, 1-3 deer teeth, 0-7 pots 89 0-7 pots, 1-3 deer teeth 88 0-7 pots, 1-3 deer teeth 62 0-7 pots, 1-3 deer teeth 45 0-7 pots, 1-3 deer teeth 119 spoon, 1-7 bone tools, 1-3 deer teeth, 0-7 pots 27 0-7 pots, spoon 91 0-7 pots, 1 stone tool, other pig parts, 1-3 deer teeth 21 wine vessel, 0-7 pots, other pig parts, 1-3 deer teeth 20 0-7 pots, other pig parts Cluster 1, j Cluster 2 is the same as Cluster 1 from the Four Group Solution • Cluster 3 Burial 63 blackware, painted ware, wine vessel, 1-2 serving stands, 8-14 pots, bone flat piece (ban) .4-12 deer teeth, carved cylinder, 1 stone tool, 1-7 bone tools 59 blackware, painted ware, wine vessel, 1-2 serving stands, 19-30 pots/ carved cylinder, 1-3 deer teeth, hairpin, spoon, 1 stone tool, 1-3 pig skulls,1-7 bone tools, hair tie Cont'd... - 298 -FIGURE A6-11 continued Cluster 3 (cont'd.) Burial 54 blackware,.painted ware, 19-30.pots, wine vessel, 3-9 serving stands, hairpin, spoon, 1 stone tool, 1-7 bone tools, flat piece of bone, 1-3 pig skulls, 1-10 raw material pieces, 1-3 deer teeth 12 painted ware, 1-2 serving stands, 8-14 pots, turtle shell, hairpin, 1 stone tool, 1-7 bone tools, small flat piece of bone, 1-3 pig skulls, 26 blackware, painted ware, hair tie, probable orna ment parts, 1-3 pig skulls, wine vessel, 1-2 serving stands, 8-14 pots, 1-3 deer teeth, carved cylinder, oyster shell, turtle shell, ring, hairpin, spoon, 2-7 stone tools, 16-40  raw material pieces, 13-25 bone tools 102 blackware, 3-9 serving stands, 8-14 pots, 2-7 stone tools, 1-7 bone tools, 1-3 deer teeth, oyster shell, 1-10 raw material pieces 9 blackware, wine vessel, 3-9 serving stands, 19-30 pots, ring, spoon, 9-16 stone tools, 13-25 bone tools, flat piece of bone, 1-3 deer teeth, oyster shell, 16-40 pieces of  raw material Cluster 3 is the same as Cluster 2 from the Four Group Solution. Cluster 4 Burial 99 probable ornaments parts, 1-3 deer teeth, black-ware, 1-2 serving stands, 0-7 pots, 1 stone tool, 1-7 bone tools, other pig parts 34 blackware, 1-2 serving stands, 0-7 pots, 1 stone tool, 1-7 bone tools, 1-3 deer teeth,(ban) 130 blackware, wine vessel, 1-2 serving stands, 8-14 pots, 2-7 stone tools, 1-3 deer teeth 118 blackware, wine vessel, 1-2 serving stands, 0-7 pots, spoon, 1 stone tool, 1-3 deer teeth 38 blackware, 1-2 serving stands, 8-14 pots, carved cylinder, spoon, 1 stone tool, 1-7 bone tools, 1-3 pig skulls, 1-3 deer teeth 32 blackware, 1-2 serving stands, 8-14 pots, 1 stone tool, 4-5 pig skulls, 1-3 deer teeth 9 4 blackware, painted ware, 1-2 serving stands, tall stemmed cup, 1-7 pots, 1-3 deer teeth 132 blackware, 1-2 serving stands, 0-7 pots, 4-12 deer teeth Cont'd... - 299 -FIGURE A 6-11 continued Cluster 4 (cont'd.) Burial 107 blackware, 1-2 serving stands, 1-7 bone tools, 1-3 deer teeth 76 blackware, 1-2 serving stands, 1-3 deer teeth 14 blackware, 1-2 serving stands, 0-7 pots, 1-3 deer teeth Cluster 5 Burial 67 1-2 serving stands, tall stemmed cup, 8-14 pots, bracelet, hairpin, hair tie, 13-25 bone  tools, small round flat stone (bing) 49 1-2 serving stands, tall stemmed cup, bracelet, 0- 7 pots, 1 stone tool, hair tie, probable ornament parts, small round flat stone (bing), (ban), 4-12 deer teeth 129 blackware, 3-9 serving stands, 8-14 pots, ring 58 blackware, 1-2 serving stands, 8-14 pots, hairpin, (bing), other pig parts, 4-5 pig skulls, 1- 3 deer teeth 78 blackware, 1-2 serving stands, 8-14 pots, bracelet, hairpin, probable ornament parts, 1-3 deer teeth 7 blackware, 1-2 serving stands, 8-14 pots, hairpin, hair tie, probable ornament parts, 1-3 deer teeth Cluster 4, Cluster 5 is the same as Cluster 3 from the Four Group Solution. Cluster 6 Burial 103 1-2 serving stands, 0-7 pots, spoon, 2-7 stone tools, 13-25 bone tools, oyster shell, 1-10 raw material pieces, 1-3 pig skulls, 4-12  deer teeth 87 0-7 pots, spoon, 2-7 stone tools, 1-7 bone tools, 1- 3 pig skulls, 1-3 deer teeth, 79 0-7 pots, 2-7 stone tools, 1-3 pig skulls, other pig parts, 4-12 deer teeth, oyster shell 131 blackware, 1-2 serving stands, 0-7 pots, ring, 2- 7 stone tools, 1-7 bone tools, 1-3 pig skulls, 1-3 deer teeth Cont'd... - 300 -FIGURE A6-11 continued Cluster 6 (cont'd.) Burial 110 1-2 serving stands, 0-7 pots, turtle shell, ring, 1 stone tool, 1-7 bone tools, 1-3 pig skulls, 1-3 deer teeth 109 1-2 serving stands, 0-7 pots, carved cylinder, ring, spoon, 1-7 bone tools, 1-3 deer teeth, 1-3 pig skulls 73 1-2 serving stands, 0-7 pots, 1-7 bone tools, 1-3 pig skulls, 1-3 deer teeth, bracelet 106 painted ware, 1-2 serving stands, 0-7 pots, turtle shell, ring, spoon, 9-16 stone tools, 1-7 bone tools, 16-40 raw material pieces, 4-5 pig skulls 19 painted ware, 1-2 serving stands, 0-7 pots, turtle shell, ring, 9-16 stone tools, 4-12 deer teeth, 1-7 bone tools, 1-3 pig skulls, other pig parts, 1-10 raw material pieces 23 painted ware, 1-2 serving stands, 0-7 pots, (ban), 1-3 deer teeth, 1-10 raw material pieces 11 painted ware, 1-2 serving stands, 0-7 pots, 1-7 bone tools, other pig parts, 1-3 deer teeth, 1-10 raw material pieces 5 3 painted ware, 1-2 serving stands, 0-7 pots, spoon, 1-3 pig skulls, 1-3 deer teeth/ 18 painted ware, 1-2 serving stands, 0-7 pots, 1-3 pig skulls, other pig parts, 1-3 deer teeth 65 painted ware, 1-2 serving stands, 0-7 pots, 2-7 stone tools, 1-3 pig skulls, 1-3 deer teeth 8 painted ware, 1-2 serving stands, 0-7 pots, 1 stone tool, 1-3 pig skulls, 1-3 deer teeth Cluster 7 Burial 115 blackware, 0-7 pots, 2-7 stone tools 61 blackware, 0-7 pots, 1 stone tool 55 painted ware, 0-7 pots, 1 stone tool 33 blackware, 0-7 pots, 1 stone, tool, hair tie, 1-3 deer teeth 30 0-7 pots, bracelet, 1-3 deer teeth, hair tie 81 0-7 pots, 1-2 serving stands, stone tool , 1-7 bone tools, 1-3 pig skulls, 1-3 deer teeth, 1-10 raw material pieces 66 0-7 pots, 1 stone tool, 0-7 bone tools, 4-5 pig skulls, 1-10 raw material pieces Cont'd... - 301 -FIGURE A6-11 continued Cluster 7 (cont'd.) Burial 56 painted ware, 0-7 pots, 1-7 bone tools, 4-5 pig skulls 112 0-7 pots, carved cylinder, 1 stone tool, 1-3 pig skulls, 1-3 deer teeth 52 0-7 pots, 1 stone tool, 1-3 pig skulls 108 painted ware, 0-7 pots, 1-3 pig skulls, 1-3 deer teeth 44 0-7 pots, 1-3 pig skulls, 1-3 deer teeth 42 0-7 pots, 1-7 bone tools, 1-3 pig skulls, 1-3 deer teeth 116 1-2 serving stands, 1 stone tool, 0-7 pots 101 1-2 serving stands, 0-7 pots, 1 stone tool, 1-7 bone tools, oyster shell 84 1-2 serving stands, 0-7 pots, 1-3 pig skulls 6 blackware, wine vessel, 1-2 serving stands, 1-7 bone tools, 1-3 pig skulls Cluster 6, Cluster 7 is the same as Cluster 4 of the Four Group Solution. - 302 -FIGURE A6-12. Multiple burials from the Early Period: age, sex, grave goods and energy expenditure. 13 (Male, Female) log tomb, type 3. large in size blackware, wine vessel, 3-9 serving stands, 19-30 ceramic vessels, elephant ivory carved cylinder, 14 pig skulls, hair pin, spoon, two elephant ivory disks (bi), 1 stone tool, 1-7 bone tools, 1-3 deer teeth - interpreted as high status 28 (Female with infant) pit with second level platform. small size 1-2 serving stands, 1-3 pig skulls, 1-3 deer teeth, 0-7 pots - interpreted as intermediate in status 31 ( ?, ? ) pit. only burial in Early perios totally empty of grave goods. small size - interpreted as low in status 35 (Male, Female, Child) pit. small size. blackware, 1-2 serving stands, 8-14 pots, 2-7 stone tools, 1-7 bone tools, hair tie, ring, 1-10 raw mate rial pieces, 1-3 pig skulls, oyster shell, 1-3 deer teeth. Additional: two pig hooves (ti) - interpreted as intermediate in status 69 ( ?, ? ) pit. small size 0-7 pots, 1 stone tool, 1-7 bone tools, 1 horn pendant or netweight (zhui), turtle shell - interpreted as intermediate in status 111 (Male, Female) pit. small in size 8-14 pots, wine vessel, 1-2 serving stands, tall stemmed cup, 2-7 stone tools, 1-7 bone tools, spoon, ring, 1-10 raw material pieces, 1-3 deer teeth. Additional: 4 lower jaw bones of deer (si lu) - interpreted as intermediate in status FIGURE A6-13. The 32 burials in the Late Period, analysis of status: age and sex, grave form, grave size. 2 3 D, FF, MD 4 RM 5 FF 10 F, LT, ER, LA 15 M 16 17 RM 22 FM 24 NS, RM 25 RM, LT , ER, MD 46 D, MD 47 FF, LT , MD 60 NS, LT , LA, FF 64 MD 72 F 75 RM 77 D 93 96 98 ER, MD , FM 100 104 LT, ER , MD 105 F 117 s, M, LT, ER, ; 121 F 122 M 123 M 124 FM 125 M, MD 126 RM, LT , LA 127 NS KEY total (1) S = subadult (4) F = known female (5) M = known male (3) NS = no skeleton (6) RM — reliably estimated male (3) FM = fair estimated male (4) FF = fair estimated female (3) D = disturbed (7) LT log tomb (5) ER — second level platform (3) LA — large in size (9) MD = medium-sized Simple pits and small graves undesignated. FIGURE A6-14. The 2 8 variables and their attributes in the Late Period analysis of status. (Coded by presence/absence or by additive coding.) Variables Presence/Absence white ware black ware painted ware wine vessel bottle jade bone carved cylinder elephant ivory carved cylinder turtle shell ring (bone,stone,jade) bracelet (bone,stone,jade) hairpin (bone,stone,jade) spoon hair ties neck ornament, head ornament horn (zhui)(function uncertain) probable ornament parts: (plan, guan, zhu) small round stone (bing) pig skulls other pig parts oyster shell 10(31. 3%) / 22 (68. 7% 15 (46. 9%) / 17 (53. 1% 6 (18. 8%) / 26(81. 2% 10(31. 3%) / 22(68. 7% 7 (21. 9%) / 25(78. 1% 4 (12. 5%) / 28(87. 5% 6 (18. 8%) / 26(81. 2% 5 (15. 6%) / 27 (84. 4% 3(9. 4%) / 29 (90. 6% 5 (15. 6%) / 27 (84. 4% 12(37. 5%) / 20 (62. 5% 8 (25. 0%) / 24 (75. 0% 5 (15. 6%) / 27 (84. 4% 10(31. 3%) / 22 (68. 7% 5 (15. 6%) / 27 (84. 4% 4 (12. 5%) / 28(87. 5% 3( 9. 4%) / 29 (90. 6% 3(9. 4%) / 29 (90. 6% 7 (21. 9%) / 25(78. 1% 9 (28. 1%) / 23(71. 9% 5 (15. 6%) / 27 (84. 4% Cont'd. FIGURE A6-14 continued Additive Coding serving stand tall stemmed cup total number of ceramic vessels stone tool total bone tool total deer teeth raw material pieces 0/1/2-12 : 12(37.5%)/14(43.7%)/6(8.8%) 0/1-6/14-16 : 14(43.7%)/15(46.9%)/3(9.4%) 1-13/16-21/38-93 : 20(62.5%)/6(18.8%)/6(18.8%) 0/1-3/5-19 : 14(4 3.8%)/13(40.6%)/5(15.6%) 0/1-6/10-27 : 17(53.1%)/11(34.4%)/4(12.5%) 0/1-2/3-5 : 6(18.8%)/20(62.5%)/6(18.8%) 0/1/11-26 : 26(81.2%)/3(9.4%)/3(9.4%) - 306 -FIGURE A6-15. Distribution of serving stands, Late Period.. 3 CO T3 C n) 4J CO 60 C •rl O > -H r-l U O) CU CO cw 0) O 4-1 cO U HJ •i c I I I I I I I I I I I I I I I. D 10 i \ 1 1 1 1 1 1 1 i i i r o LO o number of burials - 307 -FIGURE A6-16. Distribution of the tall stemmed cup, Late Period. •s 60 .•3' I I I I L 15 CO o< a -a t3 cu o I u at <u 4-> FN CO cu & 4J 60 CJ •rl- i-J 4-1 o CU •i-3 (3 10 i i—i—i—r~ O CO vO number of burials - 308 -FIGURE A6-17. Distribution of pottery vessels, Late Period. 93 V 71 o 0 u a cu 55 al 0 1 38 20 o 15 e c 10 i o number o f burials - 309 -FIGURE A6-18. Distribution of stone tools, Late Period. 10 i i i i i i i i i i i i i i i i 1 1 cd •H M <4-4 o u CD 1 a •5 0 -f I 1 • I I I I I I I I I o CN stone tool totals, Late Period - 310 -FIGURE A6-19. Distribution of bone tools, Late Period. o •H U CU PH cfl O O CU C O I I I I I I I I 30 25 20 n 15 u cu rO Cfl 4J O 10 1—i—i—r- T" O CN -J-number of burials - 311 -FIGURE A6-20. Distribution of deer teeth, Late Period. J i i i i i_ 10 -m o u Qi 42 o -I—I—I—I—I number of deer teeth, Late Period - 312 -FIGURE A6-21. Distribution of raw material pieces, Late Period. 01 G o l_ 26 CO 3 4-1 rH CO •H U cu CO e TJ !5 o cO •H U l-l 0) 4-1 P4 o 01 CO 4-1 QI CO O 01 •iH P. s o 4-1 /CO O •rl rH — 01 42 c 6 M 3 O (3 43 10 number of burials - 313 -FIGURE A6-22. Grave area of Late Period burials. i i i i i i i i i i i i i CO u cu 4J CU e cu u CTJ 3 cr CO c •H CO CU u cu > cfl u M 1 a r g e m e d i u m s m a 1 1 14.00-13.00-12.00-11.00-10.00-9.00-8.00-7.00-6.00-5.00-4.00-3.00-2.00-1.00-0.90-0.80-•14. •13. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. 0. 0. 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 89 i I o |-number of burials - 314 -FIGURE A6-2 3. The mortuary attributes characteristic of the clusters from the eight group solution, Ward's Method, Late period burials. (Underlined additively coded attributes represent the greatest quantity of that artifact type.) Cluster 1 Burial 117 white ware, black ware, painted ware, bottle, 14-16 tall stemmed cups, 38-93 pots, jade, bone carved cylinder, elephant ivory carved cylinder, bracelet, hairpin, TO-* 27 bone tools, hair tie, horn(zhui) 47 white ware, black ware, wine vessel, 2-12 serving stands, 14-16 tall stemmed cups, 38-93 pots, turtle shell, bracelet, hairpin, 1-3 stone tools, hair tie, head or neck ornament, pig skulls, 3-5 deer teeth 25 white ware, black ware, 2-12 serving stands, bottle, 14-16 tall stemmed cups, 3 8-93 pots, jade, bone carved cylinder, ring, bracelet, hairpin, spoon, 5-19 stone tools, 1-6 bone tools, pig skulls, other pig parts, 1-2 deer teeth 10 white ware, black ware, painted ware, 2-12 serv ing stands, bottle, 1-6 tall stemmed cups, 38-93 pots, jade, bone carved cylinder, elephant ivory carved cylinder, ring, bracelet, hairpin, 1-3 stone tools, head or neck orna ments, probable ornament parts, pig skulls, other pig parts, 3-5 deer teeth Cluster 1 is the same as Cluster 1 of the Three Group Solution. Cluster 2 Burial 17 10-27 bone tools, other pig parts, 1-2 deer teeth, 11-26 raw material pieces, black ware, 1 serv ing stand, 1-13 pots, elephant ivory carved cylinder, ring, bracelet, 5-19 stone tools 125 black ware, wine vessel, 1 serving stand, 1-6 tall stemmed cups, 16-21 pots, ring, spoon, 5-19 stone tools, 10-27 bone tools, hair tie, (zhui), pig skulls, 3-5 deer teeth, 11-26 raw  material pieces, oyster shell 4 painted ware, wine vessel, 1 serving stand, 1-6 tall stemmed cups, 1-13 pots, jade, bone carved cylinder, turtle shell, 5-19 stone  tools, 10-2 7 bone tools, 1-2 deer teeth, pig skulls, 11-26 raw material pieces, oyster shell Cont'd... FIGURE A6-2 3 continued Cluster 3 Burial 98 75 22 1-6 tall stemmed cups, 16-21 pots, spoon, 1-3 stone tools, 1-6 bone tools, 1-2 deer teeth, oyster shell black ware, wine vessel, 1 serving stand, 16-21 pots, spoon, 1-6 bone tools, 1-2 deer teeth, oyster shell 1 serving stand, 1-6 tall stemmed cups, 1-13 pots, ring, spoon, other pig parts, 1-2 deer teeth Cluster 4 Burial 60 126 white ware, black ware, 2-12 serving stands, bottle, 1-6 tall stemmed cups, 3 8-9 3 pots, probable ornament parts, (bing), other pig parts white ware, 2-12 serving stands, bottle, 1-6 tall stemmed cups, 38-9 3 pots, bone carved cylinder, 1-3 stone tools, 1-6 bone tools, other pig parts, 1-2 deer teeth, 1 raw material piece Cluster 5 Burial 72 white ware, wine vessel, 1-6 tall stemmed cups, 16-21 pots, hairpin, head or neck ornament, other pig parts, 1-2 deer teeth 3 white ware, black ware, painted ware, wine vessel, 1-6 tall stemmed cups, 16-21 pots, bracelet, head or neck ornament, (bing), pig skull, 1-2 deer teeth, 1 raw material piece Cluster 2, Cluster 3, Cluster 4, Cluster 5 is the same as Cluster 2 of the Three Group Solution. Cluster 6 Burial 124 wine vessel, 2-12 serving stands, 1-6 tall stemmed cups, 1-13 pots, bracelet, 1-3 stone tools, 1-2 deer teeth 123 painted ware, wine vessel, 1 serving stand, 1-13 pots, 1-3 stone tools, 1-2 deer teeth 93 1 serving stand, 1-13 pots, 1-3 stone tools, 1-6 bone tools, hairtie, 1-2 deer teeth Cont'd. - 316 -FIGURE A6-23 continued Cluster 6 (cont'd.) Burial 122 1 serving stand, 1-13 pots, 1-3 stone tools, other pig parts 16 black ware, 1 serving stand, 1—13 pots, 1-3 stone tools 96 1-13 pots, hair tie, 1-2 deer teeth 46 1-13 pots 100 1 serving stand, 1-13 pots, 1-2 deer teeth 15 1 serving stand, 1-13 pots, (zhui), 1-2 deer teeth Cluster 7 Burial 121 wine vessel, 1-13 pots, bracelet, 1-6 bone tools, hair tie, pig skulls, 3-5 deer teeth, ovster shell 105 1-6 tall stemmed cups, 1-13 pots, bracelet, hair tie, head or neck ornament, 1-2 deer teeth 104 wine vessel, 1-13 pots, bone carved cylinder, bracelet, hair tie, 1-2 deer teeth Cluster 8 Burial 127 white ware, black ware, 1-2 serving stands, bottle, 1-13 pots, 1-3 stone tools, 1-6 bone tools 77 black ware, bottle, 1-6 tall stemmed cups, 1-13 pots, elephant ivory carved cylinder, 1-3 stone tools, 1-6 bone tools, 1-2 deer teeth 6 4 white ware, black ware, 1-6 tall stemmed cups, 1-13 pots, turtle shell, 1-6 bone tools, hair tie, 3-5 deer teeth 5 black ware, 1 serving stand, 1-6 tall stemmed cups, 1-13 pots, bracelet, hairpin, 5-19 stone tools, hair tie, (bing), 1-2 deer teeth 2 black ware, 1-13 pots, bracelet, hairpin, 1-3 stone tools, probable ornament parts, 3-5 deer teeth Cluster 6, Cluster 7, Cluster 8 is the same as Cluster 3 of the Three Group Solution. - 317 -FIGURE A6-24. Data regarding the undatable burials: age, sex, grave goods, energy expenditure, body disposition. (All single pits, small in size.) 1 37 child. no grave goods. 2 39 child. no grave goods. disturbed. 3 40 reliably estimated male. skeleton on left side. 2 stone tools, 4 bone tools, 1 turtle shell, 1 pig skull, 2 pieces of raw material. 4 5 0 child. no grave goods. 5 5 7 known female. 1 stone tool, 1 deer tooth, 1 pig skull. 6 68 child. 1 bracelet, 1 deer tooth, 2 pig skulls. 7 70 multiple, unsexed burial. 1 stone tool. 8 74 child. no grave goods. 9 83 3 deer teeth. 10 85 known female. 1 stone tool, bone tool, oyster shell. 11 92 multiple, unsexed. no grave goods. 12 95 child. no grave goods. 13 113 1 piece of raw material, 1 deer tooth. 14 128 no grave goods. 15 133 flexed. 1 raw material piece, 3 deer bones. Note: unsexed adults not designated. - 318 -GLOSSARY ban: small flat piece (of bone), possibly part of JFHL an ornament bang plan: oyster shell, flat thin piece bei: cup ben: adze *-fr bi: spoon, spatula bi: round flat disk with hole in center biao: dart-like weapon, harpoon ^ bihuan: bracelet ^ bing: small flat circular piece of stone (function uncertain, possibly part of an ornament) bo: bowl chan: spade, shovel chikou: wide-mouth chi tie kuang shi: iron ore, red ochre? chui: hammer fk a & it %Y To cong: long hollow object with rectangular sides J-rjx dab: knife dapo: intrusive (eg. burial) diao tong: carved cylinder •n ding: tripod, food vessel ft dou: footed stand, serving stand for food 3. er: handle ercengtai: second level platform, or ledge on which =- £ grave goods were placed /7 aa a ix% eyu;lin: alligator hide T fanglun: spindle whorl fangwei: orientation - 319 hua wen: incised huan: small round piece (of stone), possibly an earring huang: semi-circular pendant jl: chicken XL fu: prostrate, prone fu: axe, hatchet gai: pot lid gou: fish hook gu: bone R gu jia: skeleton H ~yf ^ _ A"">r guan: tube, probably part of an ornament guan: jar gui: tripod pitcher ^jc1 a gui jia: turtle shell T he: crane he: spouted vessel, possibly for wine hu: storage vessel, possibly for wine ^ hua wen: motif ^ ji: hairpin on which hair is bound at back of head ^ jia sha: sand tempered, sand paste ^ ]ian wei: arrow remnant ]iao: horn, antler ift jiao qi ba: horn tool handle jing shi: neck ornament; types: shi zhu (stone bead), shu yao zing zhu (waist belt shaped beads?), shi pian (small flat pieces of stone), guan zhuang shi zhu (tubular shaped stone beads) - 320 -ju ch! wen: sawtooth design kui xing q1: helmet shaped vessel M_ / h% lian: sickle lian kou: constricted mouth liao: pieces of raw material lishi: whetstone liu: spout - - - nut 4> lu tui gu: deer thigh bone bK 7Jk- n lusongshi: turquoise ^ 0 raao: spearhead k mo bang: ground stone stick mu zang: grave, burial ni, ni zhi: smooth clay, finely levigated 'jb ^ niao zhao: bird claw pen: basin, for water /7t plan: flat thin piece (of stone) probably part of j^j an ornament ping: bottle qu: bent (flexed) qiu: ball sha li: grains of sand & if sheng wen: cord decoration ""^ shi: stone D s Q3lT£ D,° shou xing qi: animal effigy vessel ~FT~"/ A oo shu: comb jft shuf aqi: possible pair of hair ties it? 7^v 00 suo xing q i.: weaving shuttle shaped object ^ ~fi 3^ tao pian: ceramic sherd t^^| til % - 321 -tao qi: pottery, ceramic tao yao: pottery kiln 120 tdu shi: head ornament; types: chang fang shi pian ^ (flat rectangular stone pieces), guan zhuang shi zhu (tubular shpaed stone beads), huan xing (cir cular) , chang fang xing (rectangular), bu gui ze xing (irregular) wan: bowl 7B xiang ya: elephant tooth, ivory jpc. ^ xie fang ge wen: cross hatched design ^ ^ yan: rim, lip yang: supine i^^ yi: ladle 2) yu: jade yu gu bao: fish bone bundle or lump zang ju: burial apparatus, (log tomb) ^-p-zad: chisel, punch zhang: river deer ^% zhang ya gdu xing qi: deer tooth fish hook shaped ^h. ^ ^£/^'3S tool zhen: needle ft F zhen guan: needle case zhi htfan: ring (for' finger) zhu: circular bead, probably part of an ornament ^ETJC zhu menya: pig incisors j-j zhu gu kuai: various pieces of pig bone ^j£-zhu ti gu: pig hoof bones 3^' ^ zhu tdu: pig skull ^ zhu xia he gu: lower jaw bone of pig ^ ~f 6^ ,3, zhui: awl zhui: a hanging object, a plummet (possibly a net [Ms weight or pendant) ^ - 322 -zu: arrowhead zun: wine vessel zuo: stand, pedestal Sources: 1) The Chinese-English Dictionary, 1979, The Commercial Press, Hongkong 2) translation terms compiled by Stanford University 3) Chang (19 81a) and Ma (19 80) for some ceramic terms 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058396/manifest

Comment

Related Items