MSU LIBRARIES n. RETURNING MATERIALS: PIace in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. r' W} I “.382 W 3 07? I929 Aueminssg l rial 2000 j e9 ’9 t9 KIN GROUPS AND MORTUARY PRACTICES: ETHNOGRAPHIC IMPLICATIONS FOR ARCHAEOLOGY BY Robert George Kingsley A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Anthropology 1985 Copyright by ROBERT GEORGE KI NGS LEY 1985 ABSTRACT KIN GROUPS AND MORTUARY PRACTICES: ETHNOGRAPHIC IMPLICATIONS FOR ARCHAEOLOGY BY Robert George Kingsley This study seeks to address a long-neglected area of archaeological mortuary analysis, that of the symboling of "horizontal" societal groups (kin groups and sodalities) in mortuary context. A sample of 115 ethnographic societies forms the basis for an examination of such symboling. The results show that such groups, particularly clans and lin- eages, are frequently symboled in mortuary context, and a series of theoretical postulates and test implications is provided to allow the archaeologist to discern such groups from mortuary data. The implications for discovering kin groups is inte- grated into a multidimensional research program designed to elucidate socio-political structural form in extinct sys- tems. The patterning of kin groups, differential status, settlement structure, and other factors are used to explain formal variability. Three alternative models of structural form in ranked systems are offered as heuristic devices against which archaeological data may be evaluated. This construct is applied to a body of late prehistoric data .4 ~ - o ' - ,v ¢-' r . . ‘- ”\A1 I... - - 0 - a ‘ ‘ . I .4 '. w ' ' . i v a ' 7. a a: J. 5 ll. - - "o'l. .. ' . II . ‘ . - ' ‘3 'u ," [11 ““ f r I . a - J AL! A I \I o ) .1 a“ . i—\ .q C? C . 1" D L4 ,9» lk'. § rm,- I . x.— ‘V .. L " \L . ! .,.. -- i y“? . L‘ . w 5 ‘ I . "1'14 , 2': l ‘r l s I \ 5 ,.‘. .. 6. ,..‘ '3' . .. J 4,1: ,, _‘ l l V - . pd -~I‘ ' \r u-‘JLJv.- -’. ‘J ’3 l .r-natr. ,. q ' . . ' . '_ . fit .‘ ‘ .fl' 0 K" o ,a . a" v . ‘ .9 \ 0 ~. I -A .4 ‘ .. a. . u— w .L .a '- J -. ~_l .. I - . ' . .1 . - ~. . . . .‘- . -. . _ ‘.' - l 1 \ ~. u n .- . a. u - — -- .-. . .o. . 0 . , . .7 .. , . o , .. ‘ 1‘ - . ' , ' - ' ' A ‘ ‘ v. ). . a . _ ‘ a a a - ,. x s . < ~ . as . .-. . _. r x. - ‘ ‘ - -. v -o . .. . <. .. . . ~ ‘ r - ‘ . v . . . <. - - . » ., | ~ ‘ _ . I _ . \ . ~ ‘ n . ‘ , . , ‘ , .. ' ' s . . . v . . -. - - s . \ - . . . , . n a o . . . . ‘ k . - . . . P A u . .- a ~ . ‘ \ - . .' I \ ‘ v n .- . - . . . . . .- A . - . I n . . \ _ . ‘ . .' ‘ s < o- ‘. V . . . . , . . - . . . V . < a Robert George Kingsley from the central Mississippi River valley, and it is shown that variability between systems can be explained in terms of differences in kin group structures, the nature of diff- erential status, and settlement configuration between these systems. The significance of this approach lies in the fact that it allows the more precise estimation of socio-politi- cal structural form in an extinct system, and thus goes far beyond previous approaches that focus on differential status alone, and thereby result in simple classifications of sys- tems into gross evolutionary categories. An ethnographic test of "Hypothesis 8" is conducted, and the hypothesis is rejected as a proposed explanation for the use of formal disposal areas. "Hypothesis 8" stated that disposal areas will be used primarily by economic corp- orate groups; the present study shows that disposal areas are usually used by kin groups regardless of the nature of ECOIIOIIIIC corporate nes S . This dissertation is dedicated to the memory of DONALD G. KINGSLEY, Jr. ii ACKNOWLEDGMENTS This dissertation could not have been possible without the assistance and encouragement of many people. First and foremost, I would like to thank the members of my disserta- tion committee: Dr. William A. Lovis, Chair; Dr. Charles E. Cleland; Dr. Moreau S. Maxwell; and Dr. Lawrence H. Robbins. Special recognition must be accorded to Dr. Lovis who, as advisor and friend, provided considerable advice, sugges- tions, and constructive criticism. Thanks also goes to Ms. Nanette Carignan for typing the final manuscript. Encouragement and moral support was given by many individuals too numerous to name. However, particular thanks is offered to Dr. Margaret E. Holman, Mr. James A. Robertson, and Dr. Elizabeth B. Garland for support at par- ticularly crucial junctures. Finally, this study could not have been done without the continued support, encouragement, and above all patience of my wife, Lynn S. Kingsley. To her is offered my deepest gratitude. iii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . CHAPTER 1. CHAPTER 2. Data INTRODUCTION . . . . . . . . . ANALYSIS AND RESULTS . . . . . . Collection . . . . . . . . . Concepts and Definitions . . . . . . . Horizontal groups . . . . . . . . Socio-political categories . . . . . Theory . . . . . . . . . . . . Presentation of Results. . . . . . . . Introduction. . . . . . . . . . Disposal modes . . . . . . . Horizontal group symboling . . . . . Lineage and clan . . . . . . Descent groups, cognatic descent systems, and disposal areas . . . . . . Descent groups, disposal areas, and settlement patterns. . . —T_ . . Disposal modes and subsistence patterns. Disposal patterning and descent . . The problem ofihouse- related disposal . Summary . . . . . . . . . . . . . Discussion . . . . . . . . . . A multidimensional approach. . . . . Conclusion . . . . . . . . . . vi .viii 1 13 13 20 20 23 27 36 36 37 43 55 66 73 86 88 89 92 92 101 110 CHAPTER 3. AN ETHNOGRAPHIC TEST OF "HYPOTHESIS EIGHT"112 Introduction . . . . . . . . . . . 112 An Ethnographic Test. . . . . . . . . . 119 Discussion . . . . . . . . . . . . . 127 CHAPTER 4. TOWARD THE ARCHAEOLOGICAL STUDY OF SOCIO-POLITICAL STRUCTURAL FORM . . 132 .Introduction . . . . . . . 132 Archaeological Approaches to Formal Variability. 134 Socio-political Complexity in Ranked Societies . 142 Archaeological Implications . . . . . . . 163 iv The political role of UDGs Institutionalizatiofi_and the symboling "Complex "B " of status . "CompIEx II" Chiefdoms IT-Chiefdoms. Rafiking. a31c CHAPTER 5. Introduction A Mississippian Test Cas Early Mississippian Period, AN ARCHAEOLOGICAL APPLICATION 700-1000 . Middle Mi351ssippian Period, _ s-s- 52. Late Mississippian Period, 1000-1350 Summary Discussion 1350-1650 Conclusion CHAPTER 6. APPENDIX A. APPENDIX B. APPENDIX C. BIBLIOGRAPHY DISCUSSION AND SUMMARY. MASTER DATA LIST REFERENCES FOR SOCIETIES IN APPENDIX A AN ETHNOGRAPHIC TEST OF HYPOTHESIS EIGHT (CHAPTER 3): DATA SUMMARY 163 165 170 172 175 178 178 181 181 186 201 210 210 215 218 225 232 235 241 Table Table Table Table Table Table Table Table Table Table Table Table Table 1. 2a. 2b. 7a. 7b. 10. 11. LIST OF TABLES Disposal Modes, by Category. . . . . . 42 Frequency and Percent of Horizontal Group Symboling, by Category . . . 47 Frequency and Percent of Unilineal Descent Group Symboling, by Category . . . 47 Recoverability of Horizontal Group symboling . . . . . . . .47-A Horizontal Group Symboling, by Group, by Category . . . . . . . . . 50 Horizontal Group Symboling, by Occurrence, by Type, by Category . . . . . . 52 Horizontal Group Symboling, by Type, by Category . . . . . . . . 53 Lineage Symboling, by Strong vs. Weak, by Category . . . . . . . 58 Clan Symboling; Lineage and Clan Symboling, by Strong vs. Weak, by Category. . . . . . 58 Lineage and Clan Symboling, by Strong vs. weak, by Category, with Combinations of Symboling . . . . . . . . . 60 Correlation of Lineage or Clan Spatial Symboling with Artifactual and Ritual Symboling, with Combinations of Types. 63 Correlation of Disposal Areas with Unilineal Descent Groups. . . . . 67 Correlation of Disposal Modes with Non- Unilineal Descent Systems, by Category 74 vi Table Table Table Table 12. l3. 14. 15. Correlation of Lineage/Clan Disposal Areas with Lineage/Clan Settlement Configurations, by Category. . . . 76 Correlation of UDG Disposal Area Types with Attributes of Settlement Pattern 79 Correlation of Disposal Modes with Primary Subsistence Modes, by Category. . . 87 An Ethnographic Test of Hypothesis 8 . . 121 vii LIST OF FIGURES Figure 1. Early Mississippian Period Phases and Sites Discussed in Text. (Adapted from Morse and Morse 1983:Figure 10.1). . . . . . . . . . . . 183 Figure 2. Middle Mississippian Period Phases and Sites Discussed in Text. (Adapted from Morse and Morse 1983:Figure 11.1). . . . . . . . . . . . 188 Figure 3. Late Mississippian Period Phases and Sites Discussed in Text. (Adapted from Morse and Morse 1983:Figure 12.1). . . . . . . . . . . . 203 viii CHAPTER 1 INTRODUCTION A major emphasis in contemporary archaeology is the development of a body of theory to direct research and attempt explanations of phenomena observed in the archae- ological record. One area of inquiry that has received con- siderable attention in this regard is the analysis of social and political dimensions of mortuary practices. Since 1970, mortuary specialists have been reasonably secure in the fact that a well-developed body of theory exists that attempts to explain mortuary behavior and hence the structure observed in mortuary sites (e.g. Saxe 1970, Binford 1971, Brown 1971, Goldstein 1976, Peebles and Kus 1977, O'Shea 1981, 1984). From the vantage point of fourteen years of retrospec- tion On these accomplishments, one can now observe that many mortuary studies in fact supercede themselves, by illustrat- ing not what we think we do know about mortuary behavior, but rather what we do ngt_know. This observation is not wholly original; several papers appearing in the past few years have called attention to various deficiencies and problems in current and past work (e.g. Brown 1981, Gold- stein 1981, O'Shea 1984). The discussion that follows is an effort to continue this trend. Several related problem areas in contemporary mortuary theory and practice will be 1 2 isolated; these, in turn, form the substantive basis for the remainder of this work. It is of first importance to appreciate the goals of current theory and method for it is at this base level that a fundamental problem exists. Theoretical propositions and most data analyses to date have been directed toward the recognition of mortuary site patterning that might elucid- ate extinct socio—political status systems. Accordingly, theorists (e.g. Saxe 1970, Binford 1971) have incorporated Goodenough's (1965) concept of the social persona, which represents a composite of any and all social identities (i.e., roles, statuses, relationships with other personae) held by an individual during life (Saxe 1970:6-7). An in- dividual's social persona is not "fixed", but may vary as certain social identities are "selected" as appropriate for recognition in different social situations (ibid.). One such occasion is the event of death, in which "...a choice between incompatible social identities must be made. Those involving rights/duties counterparts with the great- est degree of influence, authority, and/or power by virtue of that set 9f relationships will be chosen" (Saxe l970:6; emphasis in original). In other words, the social persona symboled at death will, according to Saxe, consist of a composite of social identities directly relating to that individual's position in a hierarchy of status in the society (see also Binford 1971 for an identical argument). This theoretical pr0position has guided mortuary 3 analysis since Saxe offered it in 1970. It is important to note that this perspective directs the analyst toward the examination of 1) the positions of individuals (not groups) in society, and 2) what might be termed the "vertical structure" of status positions that these persons occupied. Most case studies, for example, have focused on the nature of vertical status in ranked societies (sengu Fried 1967). Given the theoretical perspective just reviewed, "this is hardly surprising since the character of the archaeological record tends to encourage the search for traces of social rank" (Brown 1981:25). Conversely, there have been few attempts to deal with what might be termed horizontal structures (gf. O'Shea 1981, 1984, and below), such as kin groups and sodalities. It could be suggested that the character of the archaeological record does not encourage the search for such structures. On the other hand, the strict perspective of examining individual vertical status does not encourage it either; it is usually difficult to discover something that one is not looking for. A phenomenon of mortuary analysis that appears to be the direct result of this focus on individual status sys- tems is the assumption - usually implicit - that any and all variability observed in a mortuary site is somehow rela- ted to the symboling of various social personae in the status hierarchy. In many studies (e.g. Tainter 1977a. Buikstra 1976) virtually every artifact and every nuance of interment are factored into the calculations of relative 4 status. Few analyses or theoretical papers to date have allowed the possibility that all observed variability may 39E in fact be status-representative (e.g. O'Shea 1981, 1984). The assumption that all variability is status- related occurs hand-in-hand with a largely quantitative analytical approach, a trend fostered by Saxe's seminal work. Saxe offered the proposition that higher status people will manifest a greater number of positive compo- nents (i.e. artifacts, attributes of interment) in the man- ner of their burial relative to lower status people (1970: 69-71). Thus, this idea has directed investigators toward the quantitative assessment of status, where more versus less attributes generally corresponds to more versus less status. This is not to say that Saxe's proposition is in- correct, quite the contrary: it is a valid and useful con- struct, but the point to be made here is simply that most studies have not considered the possibility that some vari- ability in the mortuary domain might not be status indica- tive, and may in fact symbol or signify something else. It should be mentioned in this context that attempts to qualitatively assess meaning or significance of mortuary variability are not altogether lacking. Brown (1971), for example, has coined the term "badges of office" to refer to rare, distinctive, non-utilitarian artifacts that appar- ently denote power and authority, such as have been found in late prehistoric contexts in eastern North America, e.g. headdresses, monolithic axes. Goldstein (1976, 1981) 5 has taken what might be considered a bold step in an era in which quantitative approaches are dominant, by advocating that the meaning of mortuary variability might be better sought through simple visual examination of the data, par- ticularly regarding spatial patterning. Braun (1979) has attempted to place controls on quantitative data analysis. His study began with a comparison of artifacts recovered from mound burials with those from habitation sites to better assess their assumed significance, i.e. to determine the extent to which they were common or truly "exotic". Those items found with burials but not in village middens' were taken to represent status markers. These studies are noteworthy because they attempt qualitative assessments or judgments about the variability observed, and what it might signify, rather than simply counting up items per burial and subjecting them to a battery of statistical tests. The theoretical perspective described above has tended to produce results that have been termed "pigeonhol- ing" (Goldstein 1981:54). That is, mortuary analysts have tended to measure their data against one or two evolution- ary stage schemes: Service's (1962) band, tribe, chiefdom, and state scheme, and/or Fried's (1967) egalitarian, ranked, stratified, and state model. That these models have been utilized is not a problem, but their manner of use is. The overwhelming majority of case studies have been "goodness-of-fit" tests, attempting to determine whe- ther the data best fits one or another of these categories. 6 What has until recently been ignored is the fact that vari- ability in scope, scale, and structural form exists within these categories - which has been amply documented ethnogra- phically - but which has not been considered archaeologi- cally. Put differently, emphasis has been placed on deter- mining the degree of structure, not the form of structure. This statement is not intended to be an overly harsh critique of previous studies. Indeed, any study that is able to determine that a prehistoric system belonged in, say,the ranked category would be an important contribution, whether this constituted pigeonholing or not. This state- ment is, however, intended to say that the attention of the mortuary analyst could and should be redirected toward the more explicit examination of form, as well as degree. At least one earlier writer on the subject recoqnized the problem. Saxe stated that his initial efforts were direc- ted toward examining degree of structure; "more specific predictions as to form we leave to future research..." (1970:118). The determination of structural form poses a challenging and difficult problem, and may not lend itself to quantitative analysis. Goldstein has commented on this, by pointing out that empirical measures of complexity that attempt to scale (rather than pigeonhole) societies, such as measures of relative entropy, may actually say more about an investigator's classification scheme than about the nature of societal complexity or organization (1981: 55). Her own study of "rural" Mississippian populations in 7 Illinois (1976, 1980) is a contribution toward the analysis of form; she was able to make statements about the nature of socio-political organization and group structure at the "lower end" of Mississippian society, and thus went beyond simply labeling the system as "ranked". A recent paper by Brown (1981) has attempted to address the problem. Brown focused on the study of ranked societies, and asserted that considerable variation exists within this general category. He makes distinctions between the concepts of rank, authority, and power (1981:26-28), and recognized that different forms of ranked society can and do exist, 315., those in which high social rank does not simul- taneously entail absolute (or any) power and/or authority, versus those that do (1981:26-27). This is an important point. Brown has outlined a construct in which variability within the category ranked society is accommodated, and hence allows a more precise estimation of structural form within this category. It is this writer's opinion that much of the reason why mortuary analysis has not progressed much past the pigeonholing stage is because when one asks questions as to structural form, one inevitably finds oneself in the realm of kinship, kin groups, and social organization. It goes without saying that the nature of kinship in a prehistoric society is difficult to determine; perhaps mindful of the criticism heaped upon earlier attempts to discern kinship and residence rules from material remains, it seems that 8 mortuary archaeologists have avoided this area. As stated at the outset, such horizontal structures - which in large measure constitute the structural form of society and deter- mine the manner in which it Operates - have been eclipsed by concern with the vertical dimension. O'Shea (1981, 1984) has tried to come to grips with this problem. He examined mortuary sites of three North American Plains societies in an effort to determine the extent to which vertical and horizontal dimensions were or were not symboled upon death. Through comparison with a series of expectations generated from logic and ethnohistor- ical data, O'Shea was able to show that vertical status was more consistently symboled in the mortuary domain than hori- zontal groups; he notes also that preservational factors greatly affected his ability to discern both vertical and horizontal structures, particularly the latter. He con- cludes, then, that the search for horizontal structures in a mortuary site might not be altogether fruitful, or at least not to the same degree as analysis of vertical status (1981:51-52). The chapter that follows generally concurs with O'Shea up to a point. He seems to be correct in asserting that vertical status is in fact symboled more often and more con- sistently than the horizontal dimension in most cases. At the same time, however, it is the position taken here that O'Shea's conclusions are too pessimistic. His study in- volved a direct comparison between ethnohistoric and 9 archaeological data for three societies; his results showed that horizontal groups - kin groups and non-kin sodalities - were either weakly symboled or not symboled at all in the mortuary domain. While these results are not encouraging, they should 293 be taken to indicate that any search for horizontal structure is doomed. O'Shea examined three Spe— cific cases within the same culture area. His unsatisfac- tory results should not deter mortuary archaeologists from attempting to discern general patterning of horizontal group symboling. Three societies hardly constitute an extensive sample upon which to base such a negative conclusion. O'Shea's work brings up a final problem in mortuary analysis. It is this writer's estimation that there has been too little use of the ethnographic record to generate hypotheses, expectations, and/or models of mortuary behavior. Uses of ethnographic data to date have tended to be either very general or very problem-specific. Binford (1971), for example, used a sample of 40 societies to prove his point that social phenomena are indeed symboled upon death. Simi- larly, Tainter (1977) used over 100 societies to demonstrate the general principle that higher status personae receive a greater amount of energy expenditure on their mortuary treat- ment than lower status personae. Goldstein (1976) examined 30 societies to test specific relationships between the nature of disposal areas and corporate group structure in society. O'Shea's study just mentioned sought correlations between ethnographic and archaeological data for three 10 specific cases. Finally, Saxe's (1970) work used three societies to test a previously generated series of hypothe- ses. It is often suggested that a more ambitious use of the ethnographic data base could be worthwhile (e.g. Chapman and Randsborg 1981, Hodder 1982); as it stands, it is a largely untapped resource. The present study, then, seeks to address the various problem domains outlined above. Specific goals can be sum- marized thus: 1) This study will undertake an examination of an extensive body of ethnographic data to examine the extent to which horizontal structures (i.e. kin groups and non-kin sodali- ties) are or are not symboled upon death. The material form of such symboling will be examined, along with the corres- ponding behavior correlates that produce these patterns. This phase of the study is intended to augment the mortuary specialists' ability to discern meaning from the archaeolo- gical record; it has been amply demonstrated that vertical status is inferable, and this study will posit expectations for infering horizontal structure as well. It will be shown that vertical status is not the only thing necessarily sym- boled in a mortuary site. 2) Yet another test of Saxe's (1970) and Goldstein's (1976) "Hypothesis #8" will be conducted. This hypothesis was first offered by Saxe to account for a presumed relationship between formal disposal areas and the nature of corporate group structure and resource abundance/scarcity in a system. 11 Goldstein tested and modified this construct. The present study tests it again; the results are very different from Saxe's and Goldstein's and suggest that Hypothesis #8 - in either form - is not a good explanation for the use of for— mal disposal areas. 3) This study will attempt an exploratory investigation into the archaeological elucidation of alternative socio- political structural forms. The dimensions of vertical status and horizontal group affiliation do not exist inde- pendently but are integrated into a functioning socio-poli- tical system. The independent correlation of a perceived vertical status pattern with a pattern of horizontal groups in a society would potentially yield important insights into the structural form of that system. Such an approach might help push mortuary analysis out of the pigeonholing stage and into a more rewarding line of inquiry. In short, this study will consist of an examination of the ethnographic literature in order to discover material- mortuary patterning pertaining to the horizontal social dimension and corresponding behavior correlates that can be discerned in archaeological context. This is undertaken for the purpose of deve10ping expectations of mortuary pattern- ing in an archaeological site that might in turn indicate such horizontal structures in an extinct society. Multidi- mensional lines of inquiry will be suggested for the testing of hypotheses regarding the presence of horizontal groups. This study is not intended to be a "cookbook". This 12 statement is, of course, a standard disclaimer; it is unfor- tunate that it cannot be a cookbook, with appropriate tests and measures for discovering lineages, clans, age-grades, etc., from prehistoric burials. The following analysis will demonstrate - as O'Shea has already suggested - that no such set formulae are forthcoming. While strong statements can be made regarding the discovery of horizontal structures in archaeological contexts, it would appear that, within the mortuary domain, the further away one moves from vertical status, the more ephemeral one's data becomes. In any case, this writer intends to explore the degree to which archae- ologists can make cogent statements about extinct society and polity from mortuary remains; it will be demonstrated that the situation is not as bleak as some have suggested. CHAPTER 2 ANALYSIS AND RESULTS Data Collection The data set used in this study consists of 115 eth- nographically documented societies drawn primarily from the Human Relations Area Files and a few additional sources. The HRAF used is housed on microfiche in the Main Library, Michigan State University, and is complete and up to date. The Master Data List is presented in Appendix A. Some comments regarding the collection of this body of data are in order. While data collection was planned to be a neat and orderly process, it actually was not, but prog- ressed in largely unplanned "phases". Initial examination of the HRAF - "phase I" - was conducted essentially as a feasibility study to determine two things: whether the per- tinent data on mortuary practices in general and regarding horizontal groups in particular were present, and if so, whether these data indicated that a lengthy study of hori- zontal group symboling in mortuary context would be a prof- itable undertaking. On the basis of a sample of 53 socie- ties, both questions were answered in the affirmative, and "phase II" was initiated which was simply the continued collection of data. 13 14 At different points in this process, various patterns and correlations presented themselves which had not been an- ticipated (at least not formally) at the outset. One such finding, which occurred about midway through the data collec- tion, was a positive correlation between certain disposal modes and unilineal descent group spatial arrangement (settlement pattern). Since settlement information had not been systematically sought up to this point, "phase III" consisted of a reexamination of all previous cases to in- clude settlement data. With this task accomplished, the collection process resumed with the addition of this new class of information. Similarly, many societies examined near the end of collection happened to be complex systems, Chiefdoms or states. This was completely fortuitous, since examination of the files was done on areal basis, and the last two areas happened to be Africa and Oceania. At this point, correlations between certain unilineal descent group disposal modes and form of socio-political structure were noted; so, "phase IV" necessitated the detailed reexamina- tion of socio-political structure for many but not all of the previously examined cases. Finally, the last phase of data collection consisted of various rechecks of problem cases or groups of cases, usually where information was vague or sketchy. This final phase led to the rejection of numerous cases. In addition to the HRAF, data was sought from other published sources, including previous mortuary studies and \ 15 ethnographies. The 115 society sample, then, is comprised of 103 cases culled from the HRAF and 12 gained from other sources. All societies in the files were examined with the ex- ception of "modern" industrialized state systems. The cate- gories used include the following: settlement patterns (361) , sodalities (575), kindreds (612), lineages (613), sibs (614), phratries (615), moieties (616), bilinear kin groups (617), clans (618), tribe and nation (619), community structure (621), funeral (764), and deviant mortuary practices (766). In the situation where complex societies were under consi- deration, and/or where information was sought to test Hypo- thesis #8, some additional categories were consulted: property system (421), real property (423), inheritance (428), castes (564), classes (565), territorial hierarchy (631), districts (634), and provinces (635). Murdock's "Ethnographic Atlas" (1967) was also consulted with regard to several data classes. The data collection process resulted in a voluminous body of information. To help expedite the undertaking, a standardized check-list form was developed early on whereby the relevant data on a society could be recorded in summary fashion. However, due to the "additive" quality of the task as described above, several sets of notes on each case were also taken. It was intended from the outset that the collection of the ethnographic data be as rigorous and selective as 16 possible; where data on a society were overly vague, contra- dictory, or absent, the fiche cards were cheerfully placed in the "reject" pile and another name was crossed off the (seemingly endless) list. It should be stressed that the multidimensional nature of the inquiry, wherein many and varied classes of information were considered, revealed the deficiencies and unevenness in coverage of many ethnogra- phies. It was not uncommon, for example, to find a detailed account of mortuary practices, but virtually no considera- tion of social anthrOpology; the reverse was almost equally typical. The necessity for "good" data on the varied topics resulted in an alarmingly high rejection rate of about 50 percent: 103 cases were accepted, 105 rejected. The rate for non-HRAF sources was about the same. In cases where the literature indicated that mortuary practices and/or social organization in a society had been somehow disrupted as a result of culture contact, these societies were usually omitted. Cultural changes of this sort were typically the result of either successful mission- izing or proselytizing efforts by representatives of one of the great religions, or by decree enforced by governmental authority. A common recurring pattern was a reported shift from aboriginal mortuary practices, whatever these may have been, to burial in a Christian cemetery, or a cemetery des- ignated and required by the government. Cultural changes were also noted to be the result of general interference and disruption through culture contact, particularly in North 17 America. If aboriginal mortuary practices were also des- cribed in sufficient detail, which was infrequent, then these data were used. The data were thus selected to repre- sent aboriginal behavior and not that which resulted from cultural contact or disruption. It should be stressed, how— ever, that every case where some kind of previous changes were indicated was not omitted, only those that were clearly the result of disruptive influences. While the writer is personally satisfied with this rigorous screening process and has confidence in the result- ing data set, it is appreciated that the data are only as good as the ethnographies they were taken from, and the eth- nographies are only as good as the observations made by the ethnographers that wrote them. Thus, one fundamental under- lying assumption of this study is a necessary acceptance that the ethnographic data are "accurate" and that the ethnogra- phers were sufficiently knowledgeable so as to not misrepre- sent that which they observed. Stated differently, the assumption will be made that the present data set is not somehow biased due to biases of the ethnographers. This variable constitutes an unknown quantity and will therefore be held constant, any potential skewing effects considered negligible. A perusal of the data list in Appendix A will reveal that the complete range of mortuary practices for most societies is not represented. Since this study was intended from the outset to examine the nature of horizontal group 18 symboling in mortuary context, and not symboling in general, only that information that was germane to the task was recor- ded. That is, data pertaining primarily to vertical status distinctions are not presented. For example, readers famil- iar with Ashanti mortuary practices (see Saxe 1970) will observe that their di5posa1 mode is recorded in Appendix A as simply "L cem", or lineage cemetery. No mention is made of status-related treatments, such as the Royal Mausoleum where- in are interred kings and various nobility. For the vast majority of Ashanti, matrilineal cemeteries are the normative mode of disposal and are for that reason considered to be the most important datum, since it is the lineage cemeteries that reflect horizontal group symboling. This is being brought up to emphasize that the data are not comprehensive, and that Appendix A therefore represents only those features (both positive and negative) relating to the horizontal dimension. The data have some shortcomings. First and foremost is the fact that the sample is not random, and was not drawn randomly; The HRAF as a whole is not a random sample. Partly for this reason, the following is not a statistical study, and numerical manipulation of the data will not go beyond simple frequency counts and percentages. Of greater significance is the fact that most of the data calculations presented below involve subsamples of the 115 society sample; rarely are all 115 societies included in a test. Thus, it was found that in most cases subsample sizes were sufficient- ly low so as to prevent meaningful statistical calculations. 19 For example, in the early stages of the analyses, chi-square tests were attempted to assess the presumed significance of certain associations. In every case, the resulting statis— tic was considered unreliable due to zeros or numerous other low frequencies in contingency tables, which were a result of small subsample sizes. The same thing occurred in biva- riate examinations. While other kinds of tests might be attempted, it is the position taken here that statistical testing would not greatly enhance the findings discussed below, and in any event, the data do not lend themselves to statistical reliability; indeed, preliminary chi-square tes- ting showed that as much time would have to be spent ex- plaining the mechanical reasons why various tests did not work as discussing those that did. There are two other factors that should be mentioned though their significance, or lack thereof, is unclear. The 115 case sample is uneven with regard to both area1.coverage and socio-political-economic category representation. Con- cerning the latter, the sample has been grouped into socio- political-economic categories, about which a discussion is provided below but generally corresponds to Service's (1962) band, tribal, ranked, and state. As Appendix A shows, the tribal category contains fully half of the cases, followed by ranked (n=28), band (n=20), and state, with only ten examples. It might be speculated that this distribution is indeed representative of the non-Western world in the first half of the twentieth century, though this is certainly 20 debatable. It is clear that, based on an examination of all accepted and rejected cases, tribal societies do dominate in the HRAF, but that the present sample is even slightly more skewed toward this category. Or, more correctly, the sample is slightly skewed against bands, ranked, and states. It is difficult to assess any potential negative effects the over- representation of tribal societies may have on the results of this study, excepting the obvious fact that data on the other categories is therefore less abundant. Areal coverage is also uneven. The breakdown by HRAF culture area is as follows: Africa (n=38), North America (n=30), Oceania (n=22), Asia (n=12), South America (n=11), and Middle East (n=2). Europe and USSR are not represented. Again, there is little to offer by way of comment on this distribution beyond simply pointing it out. Concepts and Definitions Some of the terminology used in this study should be explicitly defined at this point. Most of the terms and concepts are straightforward anthropological categories, while some require somewhat more precise parameters. Horizontal groups Unilineal Descent Group (hereafter abbreviated as "UDG"). A UDG is herein defined as any recognized societal group that reckons descent, either actual or stipulated, through either the male or female line, but not both. 21 Non-Unilineal, or Cognatic, Descent Group: Such groups are defined as any recognized societal group that reckons des- cent through neither line exclusively; bilateral descent is typical. Double Descent Groups: Double descent is defined as the recognition of descent and descent groups through both the male and female lines; thus egg typically can be said to be- long to two descent lines and to minimally two descent groups, one through the pater and one through the mater. Lineage: A lineage is herein defined as any societal group that reckons descent through either male or female line, and where such descent is demonstratable. Clan: A clan is any societal group that reckons descent through either the male or female line, but where such des- cent is stipulated and no longer demonstratable; clans can be characterized as groups of lineages related by stipulated descent. I The terms "sib" and "gens" will be herein incorporated under the category "clan", despite the fact that making a distinction between "sib" and "clan" could be of analytical value. Murdock (1949:47, 65ff) defines sib as a consangui- neal kin group that usually lacks residential unity, and the clan as a compromise (i.e., including affines) kin group that does have residential unity; the sib, then, is usually "dispersed" in space, the clan "localized". It will be shown below that the mortuary symboling of particular UDGs is related to some extent on group concentration/ 22 localization versus dispersal, and the sib/clan distinction might be meaningfully employed here. However, the term clan" has come to refer to either concentrated or dispersed kin groups of this order of magnitude, with emphasis placed on descent rather than on residence. This is the sense in which "clan" is used in this study, and the qualifiers "concentra- ted" and "dispersed" will be added when needed. The reader is free, of course, to substitute "sib" for "clan" whenever appropriate. Phratry: A phratry is defined as any societal group consis- ting of two or more clans related by stipulated descent, and where at least two other like groups also exist in the same society. Moiety: A moiety organization, or division, is considered in this study to be any division of society into two roughly equal halves; other, less inclusive UDGs are usually present but need not be; while moieties are herein lumped under the rubric UDG, membership by 239 into one or the other division need not be exclusively based on unilineal descent. Sodality: The term sodality will be defined as any recog- nized societal group where membership is not based on kinship or descent, and where such groups are relatively permanent (that is, an gg_hgg task group is not considered to be a sodality). This term encompasses other terms for such groups, e.g. "societies", "associations". This definition is more restricted than some, in which the term sodality also in- cludes UDGs; types of sodality noted in the present data set 23 include age grades, warrior societies, fraternal and sororal associations, secret societies, religious or medicine socie- ties, and others. Socio-political categories For purposes of analysis and comparison the sample societies have been divided into what are termed socio-poli- tical categories. This term has been coined in order to avoid the concept of evolutionary "stages", with all the theoretical and conceptual baggage it carries. This point requires emphasis. Throughout this study, reference will be made to these categories, but no claim is made that this is necessarily an evolutionary sequence. Rather, these categor- ies are held to be empirically recurring socio—political structures, or system states, within which variability does occur, but nonetheless, certain basic features exist which differ from other categories. These categories closely par- allel Service's (1962, 1978) model but with a few modifica- tions. A detailed discussion of all features of these cate- gories will not be undertaken at this point (see Service 1962, 1975, 1978); rather, in keeping with the present inter- est in horizontal groups, the differences in the nature of such groups in the categories is emphasized. gangs (n=20): Service defines band society as "...the least complex of societal levels, in the sense that it has no Special integrative mechanisms except those common to all human societies" (1978:4). In the parlance of the present 24 study, bands will be demarcated from tribes by the absence of pan-societal integrative mechanisms such as UDGs and so- dalities. The most inclusive kinship groupings are families or extended families and formal, defined lineages do not exist. The sample contains three exceptions to this rule, all of which are Australian Aborigine societies. These cas- es are paradoxical in that they constitute "classic" band- type demographics and adaptations, yet also incorporate society-wide kin-based social groupings. Considered in the broadest perspective, however, it seems desirable to retain them in the band category. The 20 band societies represent ethnographic studies of these societies undertaken in (more or less) contemporary times. Many anthropoloqists have speculated that these con- temporary societies, most of whom are living in marginal environments to which they have been forcibly exiled, are not necessarily representative of band-type structures of the more distant past (e.g. Service 1978:2-3; O'Shea and Zvelibil 1984:1-4). This is an important problem, particu- larly with regard to the development of band-related models and expectations for archaeoloqy from ethnographic data, but this study can do little in the way of providing a solution. Fortunately, band societies have the least impact of any on the results of this study, since the focus is on horizontal structures, and bands have been defined as lacking these. Nonetheless, where this analysis does offer statements about band societies, the reader is advised to bear the foregoing 25 in mind. Tribal (n=57): Of all the heuristic devices constructed to describe various socio-political forms, the concept of tribe has clearly been the most controversial. Fried has devoted many pages to the argument that tribes don't exist (e.g. 1967). Sahlins, in his definition of tribe, includes two general constructs, "segmentary" tribesanui"chiefdoms" (1968). The present study will continue to follow Service's formulation (1978:4-6): tribes can be defined as societies having pan-societal integrative structures, that cross-cut the entire tribe to its very boundaries. These integrative structures are primarily kin—based, usually UDGs, or can con- sist of non-kin sodalities. The point of importance for the present study is that tribal systems incorporate UDGs and sodalities while bands do not. Thus, this definition also parallels Sahlins' segmentary category. Another important distinction to be made is that tribes (and bands) are socially and politically egalitarian. Egali- tarianism is, of course, a relative thing; the term is here employed to denote the lack of ascribed, or inherited, posi— tions of status, power, or authority of any kind in a society. Most tribal societies do have status people or positions but - as herein defined - these statuses will be achieved, not ascribed. The condition of lack of status ascription consti- tutes the boundary between tribal and ranked systems. With this, the tribal category continues to follow Service, and departs from Sahlins. 26 Ranked, or Chiefdoms (n—28): If the tribal category can be said to be the most controversial, then ranked can't be far behind. The terms "ranked" or "social ranking" seem to mean different things to different people, especially archaeolo- gists. For many archaeologists, "social rank" freely trans- lates to "social status" or "social inequality" of any kind (see for example Renfrew 1981 and papers therein). Others have focused on the hierarchical or pyramidal character of ranked societies but have ignored the fundamental principle underlying the hierarchy (e.g. Mainfort 1979). This funda~ mental principal is that of status ascription, and the fact that the ranking pyramid is a kin-based phenomenon. Thus, a ranked society will be defined as one in which social or socio-political status persons or positions exist which are ascribed, or hereditary (Service 1978:6—8, Sahlins 1968:23— 27). The terms "rank", "social rank", or "rank level" will refer specifically to ranked-type societies. Persons in a tribal society may have high social status, but they do not have high (or any) social rank; persons in a ranked society may have high social status and a correspondingly high social rank. It might be added that "highness" and "lowness" of rank and status are usually isomorphic, though the reverse also occurs in some situations. It should also be added that in some ranked societies some measure of social status - but not rank 1 can be achieved as well. Perhaps “chiefdom” is a better term for this category: Chiefdoms are societies that incorporate and are structured 27 by the principles of ascriptive social ranking. Chiefdom and ranked society have come to be used interchangeably in the literature however, and in the present study they will be considered synonomous. Horizontal groups do not disappear or diminish in import with the advent of ranking. Rather, they continue to provide a pan-societal integrative function, generally simi- lar to their tribal counterparts. Yet there is more to it than that, since ranking is based on kinship. Horizontal UDGs form the socio-political "armature" upon which the ranked system is based and maintained. Thus, in many, if not most, ranked societies UDGs have a different - or addi- tional - character to them than those in the tribal category. State, 95 Primitive State (n-lO): This categoryijsstraight— forward and defined by Service: state societies are those having essentially non-kin based political institutions (1978:8-9). Succession to the office of paramount may still be based on ascription within a limited group, but the gov- ernmental institutions and personnel are no longer predomin- antly kin-based. Horizontal groups still exist but tend to be attenuated or truncated, and may no longer fulfill the same functions as those in tribes or Chiefdoms. Theory A study of this nature must of necessity be firmly based on the principle of uniformitarianism. Discussion on the principle of uniformitarianism has enjoyed a revival of 28 sorts in recent years, seemingly in concert with the advent of ethnoarchaeological studies (e.g. Gould 1978, Watson 1979, Salmon 1982). The most cogent presentation of uniformitarian principles is, in this writer's opinion, that provided by Binford (1981). Binford argues that a uniformitarian per- spective on archaeological subject matter is essential if we are to gain any understanding at all of the past: Insofar as our inferences regarding the past refer to the dynamics of the past, these inferences must be accomplished by appeals to principles or knowl- edge about dynamics and how static properties pre- served in the archaeological record may be derived from dynamics. Since the only access a researcher has to dynamics is through contemporary experience, all research directed toward the development of principles that serve to make possible inferences about the past must be conducted with documented dynamic situations generally in the present (1981:27). It is important to appreciate that the above is a statement about the scientific method, and not necessarily about cul- tural reality through time (ibid.). That is, while uniformi- tarian principles are necessary to the very fact of archaeo-— logical inquiry, archaeologists cannot safely assume a direct correspondence between any and all cultural processes operating in the present with those that may have Operated in the past. Yet some degree of correspondence between past and present must be assumed, lest the view be taken that the past was a totally unique (and hence unknowable) event. Binford states that the search for processual regularities between present and past is best conducted through "actual- istic" studies of living systems; only in the study of con- temporary, observable systems can relationships between '29 cause and effect, and their resulting material correlates, be elucidated with any measure of control (Binford 1981: 26,29). But how does one know whether a cultural phenomenon observed in the present also existed or occurred in the dis- tant past? In the context of the present study, this ques- tion becomes: Can the material patterns observed in the ethnographic data be legitimately used as explanations or identifications of similar patterning observed in the archaeological record? The problem is one of pattern recognition linked with the demonstration that the pattern is redun- dant and unambiguous, a diagnostic signature that discriminates one agent or set of agents from another (Binford 1981:26). Thus, the methodology for discovering potential uniformitar- ian patterns must involve the demonstration that the pattern observed in the contemporary world is "redundant and unambi- guous", and indeed accounts for all or nearly all observed examples of the phenomenon under study. Parenthetically, it is at this point where the use of uniformitarian methods and assumptions departs from the invocation of simple analogy. Pattern recognition alone is not enough however; two addi- tional questions must be asked of any observed correlate between dynamic behavior and static patterning: 1) is the observation indeed an incidence of cause and effect, or rather is the correlation merely coincidence, and 2) is the obserVed correlation also characteristic of the past (Binford 1981:27)? 30 In this context, Binford and others (e.g. Gould 1978) have speculated that certain classes of data may lend them- selves to uniformitarian assumptions better than others. Binford's suggestions for data classes relevant to uniformi- tarian assumptions include the study of spatial structure and artifactual patterning, and the study of animal species extant in the modern world that were also present in the past (1981:28). For present purposes, it must be determined whether or not mortuary phenomena can legitimately be con- sidered in the same light. All archaeological mortuary studies have been based on the assumption that spatial and temporal regularities in the manner of mortuary treatment exist. The temporal factor is a uniformitarian assumption, though this fact is usually left implicit in the particular studies. The following discussion will seek to demonstrate that uniformitarian assumptions regarding such presumed spatial-temporal regularities in mor- tuary practices are justified. As a point of departure O'Shea's recent theoretical statements will be considered. O'Shea (1984) has offered a coherent theory of mortuary analysis which is based on a consideration of previous theoretical statements and as such should form the theoreti— cal underpinnings of mortuary analysis for some time. Cer- tain of O'Shea's principles are particulary germane to the consideration of uniformitarianism. "Principle 1" states that "All societies employ some regular procedure or set of procedures for the disposal of the dead" (1984:33-34,38). 31 This observation is based on contemporary experience of course; no human society simply leaves peOple where they drop. It is emminently reasonable that this principle be projected backward, minimally to the advent of bioloqically modern humans, perhaps some 40,000 years. "Principles 3, 3a, and 3b" are stated as follows: Principle 3: Within a mortuary occurrence, each interment represents the systematic application of a series of prescrip- tive and prescriptive directives relevant to that individual. 3a: The nature of the society will pat- tern and circumscribe the practices for the disposal of the dead. 3b: The specific treatment accorded an individual in death will be consis- tent with that individual's social position in life (O'Shea 1984:38). These related principles postulate an isomorphism between the overall nature and structure of a society and the nature and structure of the disposal of the dead. Mortuary proce- dures are intended to deal with a dramatic and disruptive event, and are systematic and deliberate. They are perform- ed for the benefit of the living, and are the result of con- scious choice and purposeful action. As such, caprice or happenstance do not dictate mortuary practices. That this is a consistent and recurring phenomenon observed in thecxnr- temporary world suggests its temporal omnipresence. In short, O'Shea's theory presents principles and correlates between society and the manner of treating the dead that are, in Binford's terms, "redundant and unambig- uous", and therefore constitute a strong basis for uniformi- tarian assumptions. To this might be added two important,if 32 obvious, facts: 1) everybody dies, and 2) so far, when someone has died, there has always been someone left still living. Turning now to the present study, this analysis will seek to discover patterned regularities between horizontal groups and mortuary treatment. It has been established that the fact of death is a temporally uniform event, and the assumption has been made that observed mortuary regularities in the modern world are too. Now the assumption must be made that the horizontal groups observed in the present also had counterparts in the past. Stated differently, this study will assume that the social structural forms observed in the present sample can be projected backward. This as- sumption is logical and seems well-founded: all contemp- orary societies are based to a greater or lesser degree on recognition of socially significant groups based on kinship. In particular, the ubitquitousness and importance of UDGs in contemporary non-western society cannot be solely a contemp- orary phenomenon; the projection of similar such social groups into the prehistoric past is a sound uniformitarian assumption. It is concluded, therefore, that the various "redun- dant and unambiguous" patterns observed in the data are 1) actual cases of cause and effect and are not coincidental, and 2) that this causation is indeed relevant to the past. The claim is ngt_made that the 115 case sample is represen- tative of all possible horizontal group mortuary“variability 33 that ever existed, but that the patterning and variability observed in the sample do indeed have counterparts in the past. Based on the foregoing, then, this study constitutes an inductive pattern search whereby the ethnographic sample is examined for the purpose of discovering patterning and regu- larities between horizontal social groups, mortuary behavior, and material correlates. Such repetitive and recurring pat- terns will form the basis for a set of expectations and pos- tulates for the discovery of such groups and behavior from patterning observed in an archaeological mortuary context. Many of these findings are highly redundant and unambiguous, and their application will thus allow the archaeologist to make inferences of high probability. Other patterning is more equivocal, and therefore will allow inferences of some- what less conviction. The patterning and correlates to be presented below will be stated in terms of positive representation only. O'Shea has made the important point that much mortuary theory to date has not been adequately "translated" into archaeolo- gical expectations (1984:44-46). That is, "if a hypothe- sized relationship is derived from ethnographic testing, it must be demonstrated that the relationship will still pertain after transformation into the archaeological context" (1984:45). This theme is integral to Binford's theoretical perspective discussed above. For example, many mortuary studies have provided correlates between human behavior and 34 the archaeological record, but have also incorporated an assumed obverse/presence-converse/absence relationship that is not verifiable and that has ultimately proven false or otherwise untenable (e.g. Saxe 1970, Goldstein 1976; see chapter 3). That is, where the presence of an entity indi- cates the presence of a past behavior, the absence of that entity conversely indicates the absence of that behavior. It is this proposed converse/absence relationship that is often called into question, since the absence of the entity might indicate any number of alternative behaviors or factors. The necessity that proposed correlates be archaeologically rele- vant dictates "...a restatement of the correlates to reflect the directional constraints on inference imposed by archaeo- logical formation processes, allowing definitive statements only in instances of positive representation" (O'Shea 1984:45). The presentation of results will follow this use- ful advice. The fact that few correlates will work in the converse might be construed as a limitation of this study; on the other hand, recognition of this fact beforehand allows positive statements to be made that are unencumbered by bidirectional ambiguity. One final comment must be made before the presentation of findings. The results of this study involve expected archaeological patterning that will permit an archaeoloqist to make inferences about the nature of horizontal groups in a past society. However, these expectations should not be used in a vacuum, but rather as part of a multidimensional 35 research program designed to examine socio-political organi- zation in an extinct system. As such, these results regard- ing horizontal groups should be incorporated as one datum in an analysis involving different classes of data and converg- ing lines of inquiry. The present results may allow an archaeologist to make inferences of high probability, but analysis should not terminate at this point. If a strong inference can be made, such an inference should be used as a hypothesis for further testing with other data, such as data pertaining to the vertical status dimension, other non-mor- tuary data, and - most importantly - mortuary data sets from other related sites. Kin group mortuary and spatial pattern- ing is not a site specific thing, it is a system-wide thing, and independent confirmation for a particular inferred structure is best sought at another site. The strongest inferences to be made regarding horizontal group structures within a system will be those that derive from the study of as many related mortuary sites as possible; the strength of the inference will be proportional to the extent to which it can be demonstrated to be recurrent, repetitive, redundant, and unambiguous within the entire system. The preceding paragraph may sound like a thinly-veiled caveat, and in one respect, it is. As mentioned in the introduction, it has become abundantly clear that horizontal structures are not symboled in the mortuary domain with either the frequency or consistency that the vertical dimen- sion displays. Horizontal group symboling was described as 36 being more "ephemeral". Thus, the present results will probably have less applicability to mortuary data and use for the investigator than correlates pertaining to the ver- tical dimension. By the same token, however, if archaeolo- gists continue to assume that such structures are always invisible, then surely they will not be found. With that out of the way, the above demand that this work be used in conjunction with other lines of inquiry is not unique to this study, nor to mortuary analysis in par- ticular, but is characteristic of archaeoloqy in general, and increasingly so. Many mortuary analysts have correctly stressed the utility and necessity of a multidimensional approach (e.g. Goldstein 1976, Chapman and Randsborg 1981). The further archaeologists seek to push the theoretical and explanatory limits of the discipline, the greater the need for multiple avenues of testing and confirmation. If this statement can be taken as a general rule, then this study is no exception to it. Presentation of Results Introduction The presentation that follows has been structured so that the discussion will progress from more general observed patterning of horizontal groups to more specific correlates between particular UDGs, disposal patterns, and other cul- tural phenomena. Most results are presented in tabular form. Throughout this discussion, the reader is referred to the 37 Master List in Appendix A; all data to be presented here is included in the Master List in summary fashion. It should be explained that in most of the tables below there is included a notation labeled "cases omitted". This refers to various societies that were, for various rea- sons, omitted from a particular calculation. Most of these omissions are straightforward, e.g. where UDGs are under consideration, all cases lacking UDGs are logically excluded. A few other omissions were done for societies in which data on certain phenomena were inadequate, but where that lack did not warrant omitting the case altogether. Disposal modes As expected, a tremendous range of variation in modes of disposal of the dead was observed in the sample. It was nonetheless possible to categorize this variability into three gross disposal classes: 1) disposal associated with habitation structures, 2) use of some kind of recognized disposal area, where the disposal area is not associated with habitation structures, and 3) ”no pattern" disposal. These categories will be explained monentarily, but first the variability observed should be summarized: l) House-related: - house floor abandonment; individual - house floor cremation; individual - house floor burial; individual, family, UDG - house floor/near house burial; family, UDG - near house burial; family, UDG - house—related disposal for majority of popula- tion; high status or otherwise exceptional minority diSposed of differently 38 2) Disposal area: - cemetery; family - cemetery; single UDG - cemetery; multiple UDGs - cemetery; cognatic descent group - tomb; single UDG - crypt; single UDG - cremation, no burial; UDG, cognatic descent group - cremation, cemetery burial; UDG, cognatic descent group - cremation, no pattern burial; individual - scaffold area, no burial; multiple UDGs - disposal area for majority of population; high status or otherwise exceptional minority disposed of differently 3) No pattern: - abandonment, no burial; individual - "random" burial; individual - stream burial; individual - exhumation, bones kept; UDG - highly varied disposal pattern, e.g. subadults abandoned, unmarried males in cemetery, females in house floor, high status crema- ted, etc. — no pattern disposal for majority of population; high status or otherwise exceptional minor- ity disposed of differently In the house-related category, then, disposal - usually burial - takes place in or near the habitation of the deceased. In some cases, different locations relative to the house structure are prescribed for different peOple, e.g. status elders buried in the house floor, younger adults buried outside under the eaves, and subadults in the back. One exception to the house-related identification should be mentioned. The Tucano, a South American tribe, maintain a formal cemetery below a large, communal Great House, or Maloca, in which entire clans reside. Since this mode of disposal actually constitutes a disposal area unlike the more typical house-related patterns, it has been categorized as 39 such. The category disposal area contained the greatest amount of variability. Inclusion in this class was depen- dent upon the use of a recognized and defined (at least "loosely" defined) locus or loci for the disposal of the dead, that is or are exclusive and not associated with the specific habitations of the deceased (with the one excep- tion). Some form of interment in cemeteries is most common. All cases of cremation - with or without subsequent burial - involved formal cremation areas where all or most people were processed; no cases of "random" or no pattern cremation were noted, and the infrequent occurrences of in-house cre- mation, where the entire structure is burned down, were scored as house-related. A few communal tombs or crypts were noted, as was a single case of scaffold disposal. The no pattern category contains cases of disposal which indeed result in no pattern, whereby people are simply exposed to the elements, or buried on the spot where they died, wherever that may have been. A few cases, however, might be better characterized as "varied pattern"; as indi- cated in the list, a few societies utilized any number of disposal modes for different persons, for different reasons. This form of disposal has been lumped into the no pattern group because of this variability. This seems a logical thing to do since these examples are infrequent to begin with, and would appear ambiguous in archaeological context anyway. 40 Note that in the preceding list, each category con- tains a statement to the effect "house/diSposal area/no pattern for majority of population, something else for status or 'exceptional' minority". It was mentioned at the outset of this chapter that the present goal is to determine the nature of horizontal group symboling in a society, and this approach extends to cases where horizontal groups are not symboled (or are even present). That is, this study focuses on what might be termed "modal tendencies" in disposal of the dead, or how the majority of the population, the non- exceptional "everyday folk", are treated. Many cases of differential treatment of certain individuals was noted, but are not herein considered. For example, most societies in which shamans are important dispose of this individual in "atypical" ways, i.e., the deceased shaman is accorded some special rite or location of disposal. Similarly, many cases were noted in which the normal disposal routine involved, say, cemetery burial, but where persons who died from "abnor- mal" causes, such as drowning, being mauled by an animal, killed in battle, or struck by lightning, were treated dif- ferently or simply left on the spot. Persons considered "anti—social" in a society, such as murderers, thieves, adulterers, or general miscreants are occasionally denied a normal disposal. Finally, a few societies were noted in which certain individuals of some exceptional calibre, i.e., respected elders of advanced age, were permitted to select their location of burial, which was often in some secluded 41 spot. The point to be made here is that such behaviors are not considered - by this writer and usually by the people themselves — to be the modal or normative form of disposal for the majority of the population. Thus the focus on modal tendencies; in accordance with O'Shea's admonition that models and expectations about mortuary behavior be translat- able to the archaeological context, this procedure is justi- fied. Exceptional disposal of exceptional people may be discovered archaeologically, and may or may not be under- standable, but it does not reflect the modal mortuary prac- tices of the "non-exceptional" majority of the society. The three classes of diSposal modes were compared with the societies' categorization into bands, tribes, ranked, or states. Table 1 illustrates these associations. As shown in the graph, no pattern diSposal predominates in bands, with considerably fewer instances of disposal area or house-related. In terms of percentages of modes, tribal cases show the greatest variability, with about half diSpo- sal area and half house-related and no pattern. Disposal areas dominate in ranked, no pattern drops off to a single case. States show a downturn in disposal areas, an increase in house-related, and no no pattern disposal. The generally inverse relationship between disposal area and no pattern disposal modes is interesting. It is suggested that this pattern corroborates other investigators' statements about the structured utilization of space in more versus less complex systems. Many social researchers have 42 Table l. Disposal Modes, by Category House- Disposal No related Area Pattern n % n % n % n State: 6 60 4 40 0 0 10 Ranked: 8 29 19 68 l 4 28 Tribal: 16 28 27 47 14 25 57 Band: 5 25 3 15 12 60 22 N = 115 % .8 .7 .6 NP .5 .4 .3 H-R .2 DA .1 . O... 0 . . Band Tribal Ranked State Cases omitted: none 43 posited a linear relationship between overall societal com- plexity and the structured use of space in general. Speci- fically with regard to the mortuary domain, Goldstein (1976) posited that as overall socio-political complexity increases, so will the organized, structured use of mortuary space. The data in Table 1 nicely illustrate Goldstein's postulate. It is also interesting, especially in light of the foregoing statement, that in the state category, the inci~ dence of disposal areas declines and house-related increases. An explanation for this phenomenon can be offered which will be dealt with in detail later on. To anticipate, it will be suggested that the incidence of disposal areas correlates with the political importance or significance of UDGs in a society; such political importance of UDGs declines markedly in state systems as compared to most ranked societies, and the use of UDG disposal areas thus declines correspondingly. Horizontal group symboling At this juncture, the use of the terms "symbol" and "symboling" should be clarified. These terms will appear throughout this study, such as in the subhead above, as if to denote a conscious, deliberate symbolization of something on the part of the people involved. While such is indeed the case in the vast majority of the sample societies, this is not a necessary requirement for a particular pattern to be considered, say, "clan symboling". That is, some symboling in this study is in fact d3 facto symboling, whereby some 44 mode or form of mortuary treatment results in the signifying of a particular horizontal group that is observable archae- ologically, whether this is a conscious symbolization by the people or not. As stated, by far the majority of cases are reported to do so deliberately, to consciously do certain things in order to symbolize the deceased's status or hori- zontal group affiliation even after the fact of death. A minority of the cases evidently do not, but the manner of their mortuary treatment nonetheless results in patterning that is reflective of structure in the living society. The latter is considered "symboling", then, regardless of how the people actually feel about it. Symboling in the sample was recorded in three cate- gories: spatial, artifactual, and ritual. Spatial symbol- ing is fairly straightforward, and includes, for example, diSposal areas (usually cemeteries) for the exclusive use of a lineage or clan, or disposal areas with internal divisions used by different lineages or clans. No case of house- related or no pattern diSposal was considered spatial sym- boling; despite the fact that house-related disposal was occasionally reported to be intended by the people to symbol UDG membership, the vast majority of the cases did not, and UDGs were thus not isomorphic with house disposal. The artifactual category might be better termed "material, non-spatial". Many of the recorded instances of horizontal group symboling do involve the deliberate inclu- sion of group-defining artifacts with the deceased. Other 45 "material" representations were noted, however, including UDG totems or other insignia engraved or painted on grave markers or commemorative monuments, or in a few cases painted on the body or skeletal materials. A few cases of sodality-specific costumery were also noted, as was a single occurrence of lineage-specific grave shape. Ritual factors were recorded in this study princi- pally for heuristic and comparative reasons. None of these occurrences is recoverable archaeologically, which is how it was intended. That is, any and all mortuary behavior could be considered "ritual" or "ritualized", but the term is herein restricted to non-material ritual; the ritual act of burial in the lineage cemetery or painting the clan totem on the face of the deceased is not considered UDG ritual. Rather, examples of ritual include "paired group" structures, such as paired clans or moieties, where part of the duties of the "opposite" group are to organize and expedite the funer- ary event, or some part of it. A few cases of group-specific rituals were recorded in which each group (usually clans) in a society possessed their own unique procedures. This beha- vior was scored as ritual. Similarly, societies in which sodalities are present and where each or some sodalities are responsible for the disposal of their dead members were con- sidered "sodality ritual". Thus, with the exception of sod- alities, which are non—kin groups, to be considered UDG rit- ual involves more than the simple fact that the deceased‘s kin take part in the funerary event. This fact is universal, 46 and for a behavior to be considered UDG ritual, it must involve something beyond this. Tables 2a and 2b demonstrate the extent to which the 115 societies in the sample symbol horizontal groups. Note that these tables involve cases that symbol, not occurrences gf symboling; any society that symbols at least one group at least once is scored as "positive" regardless of whether they also symbol more than one group, or one group more than one way. Also, the tables include all forms of symboling - including ritual - and not just that which is archaeologi- cally recoverable. Percentages in parentheses are consid- ered unreliable due to small sub—sample sizes. Table 2a shows that of 93 cases having horizontal groups, 49 or 53% symbol at least one at least once. Ranked societies show the highest within category percent. Table 2b considers UDGs; non-kin sodalities are excluded. Here, 45 of 89 soci- eties with UDGs symboled at least one at least once, for 51%. Again, the incidence of symboling occurs most fre- quently in ranked societies. Table 3 examines the all im- portant factor of archaeological recoverability of horizon- tal group symboling. These data can be interpreted a number of ways. As summarized in the table, of 49 societies that symbol, 38 or 78% do so in a way that is archaeologically recoverable. This statistic is encouraging, but somewhat misleading, since in the ethnographic cases it is known beforehand whether or not a society symbols, while archae- ologically it is not. Thus, out of 93 cases that have 47 Table 2a: Frequency and Percent of Horizontal Group Symboling, by Category W/ N UDG,S Sym. % State: 10 8 2 (25) Ranked: 28 28 19 68 Tribal: 57 54 27 50 Band: 20 3 l (33) Total: 115 93 49 53 Cases omitted: none Table 2b: Frequency and Percent of Unilineal Descent Group Symboling, by Category W/ N UDG Sym. % State: 10 8 l (13) Ranked: 28 28 18 64 Tribal: 57 50 25 50 Band: 20 3 l (33) TOTAL: 115 89 45 51 Cases omitted: none 47-A Table 3. Recoverability of Horizontal Group Symboling N of Occur- cases Cases rences w/UDG,S Recov. % Spat. Art. Recov. % State: 8 1 (l3) 1 0 1 (l3) Ranked: 28 19 68 17 3 20 71 Tribal: 54 18 33 17 l 18 33 Band: 3 O 0 0 0 0 0 TOTAL: 93 38 41 35 4 39 42 SUMMARY: 38 cases recoverable = 78% of symboled cases 49 cases symboled* are recoverable 38 cases recoverable = 41% of cases w/UDG,S 93 cases w/UDG,S are recoverable 38 cases recoverable = 33% of cases in sample 115 cases total are recoverable 32 occurrences recov. = 42% of occurrences are 93 cases w/UDG,S recoverable 32 occurrences recov. = 34% of occurrences in 115 cases total sample are recoverable Cases omitted: none *Figure from Table 2a 48 horizontal groups, 38 symbol in some archaeologically recov— erable fashion, for a more sobering 41%. This figure is still somewhat unrealistic, since again, the archaeologist cannot know beforehand whether the society under investiga- tion definitely did or did not have horizontal groups (but see below). Therefore, 38 recoverable cases out of the 115 cases sample yields 33%. Table 3 also includes occurrences of recoverable symboling, i.e., where a society symbols more than one group or one group more than once. These figures are little different from the preceding: of 93 cases with horizontal groups, 39 occurrences are recoverable for 42%; 39 occurrences out of 115 cases yields 34%. Note that spa- tial symboling is most common and is always recoverable, while minimally only four instances of artifactual symboling were judged to be recoverable (3 non-perishable artifacts, 1 grave shape). What can be learned from these figures? From the standpoint of raw arithmetic, a bottom-line statistic of 34% recoverability rate might seem rather dismal. On the other hand, if it is recalled that most archaeologists have here- tofore not seriously considered the feasibility of discern- ing horizontal groups from mortuary remains, then 34% sounds like good news. This is an important point. This study is less concerned with simply how many societies in the sample actually symbol horizontal groups than with trying to dis- cover archaeologically interpretable patterning from those that do symbol. If a prehistoric society has chosen not to 49 distinguish its horizontal groups in mortuary context - or anything else for that matter - then the archaeologist can obviously do little. However,if the society does so symbol, then archaeologists ought to be able to recognize this, and the figure of 78% recoverability of symboled cases takes on new significance. Turning now to more specific examinations of symboling by particular group, Table 4 illustrates frequencies and percentages of individual group symboling by category, com- paring all societies with particular groups with the inci- dence of symboling. Some of the percentages are again unre- liable due to small sub-sample sizes. The table and accom- panying graph show some important patterning: the most fre- quently symboled group is the clan, which is the favorite of both tribal and ranked societies. Lineages are next most often symboled; phratries and moieties are rarely symboled, and then only by tribal systems. Sodality representation is also uncommon. Considering symboling by socio-political category, it can be observed that tribal systems are the only ones to symbol all five groups at least once, but clans and lineages are dominant. The latter intensifies in ranked societies, and phratries and moieties are not represented. State systems show a marked decrease in clan symboling, no lineages, phratries, or moieties, and an apparent slight increase of sodality symboling; note that phratries and moieties do not occur in the states in the sample (see Ap— pendix A). Also keep in mind that the state sub-sample 50 Table 4. Horizontal Group Symboling, by Group, by Category Linea e Clan Phratry Moiety Sodality N1N2% N1N2% N1N2% N1N2% N1N2% S 7 O 0 7 1 (l4) 0 0 0 0 O 0 4 1 (25) R: 22 7 32 23 12 52 l O O 4 O 0 12 2 16 T: 36 7 20 46 18 39 7 1 l4 l9 3 16 27 4 15 B. 2 0 0 l 1(100) 0 0 0 3 O 0 O 0 0 TOTAL: 87 l4 16 77 32 42 8 l 13 26 3 12 43 7 16 N1: N of cases with horizontal group N2: N of cases that symbol the group % .8 .7 .6 .5 .4 Clan .3 Lineage /, . .2 MOietY\r 0.0.0.7... Sodality #:000000000 .1 Phratry / 0 Band Tribal Ranked State Cases omitted: 1, 10, 90-91, 95, 99-115 51 sizes are small. Table 5 presents horizontal group symboling by speci- fic horizontal group, by occurrence, by type of symboling, and by socio-political category. To begin, note that spa- tial representations occur only in lineages and clans. Artifactual symboling is present in lineages, clans, and sodalities. Ritual is present in all. The graph demon- strates that clan-spatial is the dominant mode overall. Again, tribal systems display the most variability: clan- spatial and clan-ritual are nearly equally likely occurren- ces, lineage-spatial slightly less so, followed by fairly low occurrences of other forms. Ranked societies clearly favor clan—spatial, then lineage-spatial. All other forms are comparatively rare. The figures for state systems are again subject to small sample sizes, but show a weak prefer- ence for sodality ritual and clan—spatial, with no other forms occurring. Finally, Table 6 is presented to summarize some of the information presented in the preceding tables. This table shows the three forms of symboling by category, with the horizontal groups lumped toqether. In the graph, "ALL" refers to the percent of cases with horizontal groups that symbol at least one group once; the "SPAT", "ART", and "RIT" are the percents of that type of symboling by cate- gory. These data mirror that in Tables 4 and 5. Ranked societies show the highest occurrence of symboling, and the overwhelming majority of that symboling is spatial. Ritual 52 Table 5. Horizontal Group Symboling, by Occurrence, by Type, by Category Lineage Clan Phratry Moiety Sodality N S A R N S A R N R N R N A R S 7 0 0 0 7 l 0 O 0 0 O 0 4 0 l -n (14) (25)-% R: 22 6 l 0 23 ll 2 2 l 0 4 0 12 1 l -n 27 5 48 9 9 8 8 = T: 36 7 0 0 46 10 8 ll 7 l 19 3 27 2 3 =n 20 22 17 24 14 16 7 11 = B: 2 0 0 0 l 0 1 0 0 0 3 0 0 0 0 =n (100) =% TOTAL: 87 13 l O 77 22 11 13 8 l 26 3 43 3 5 -n 15 l 29 l4 17 13 12 7 12 = % .8 .7 l. clan rit .6 2. clan spat 3. lin spat .5 4. clan art 5. moi rit .4 6. phrat rit 7. sod rit .3 8. sod art ;‘\ 9. lin art \ .2 car. 5’: o]. 3;”00000::::..0 8’”..... . o . . 9. Band Tribal Ranked State Cases omitted: l, 10, 90-91, 95, 99-115 5 3 Table 6. Horizontal Group Symboling, by Type, by Category Spatial Artifactual Ritual N of N of n % n % n % cases sym. % State: 1 (l3) 0 O 1 (l3) 8 2 (25) Ranked: 17 61 4 14 3 ll 28 19 68 Tribal: 17 31 10 19 18 33 54 27 50 Band: 0 O l (100) 0 0 l 1 (100) % .8 .7 .6 .5 All '4 Rit .3 Spat . \\ .2 \. Art "n...” .1 \\»-—---— 0 . . Band Tribal Ranked State Cases omitted: l, 10, 90-91, 95, 99-115 54 and artifactual are infrequent. Tribal systems are again most variable, with spatial and ritual almost equally expec- table; artifactual is less common. The figures for states are somewhat inflated, but still reflect the data presented above. At this point, the reader might welcome some summary statements of the above data. Regarding specific horizontal groups, it has been shown that clans are the predominant groups that are symboled in mortuary context, followed by lineages. Sodalities, moieties, and phratries, roughly in that order, are relatively infrequent occurrences. Spatial symboling is the most common type for lineages and clans, and phratries, moieties, and sodalities are never represented spatially. Artifactual symboling is least common of all types; clan artifacts predominate, and phratries and moie- ties are never symboled artifactually. Phratries and moie- ties are only symboled in ritual, and this infrequently. Again, the clan is the most frequent locus of ritual; line- age ritual does not occur. From the standpoint of archaeo- logical recoverability, then, phratries and moieties are archaeologically invisible, and sodalities only slightly less so. Tribal systems display the greatest amount of varia— bility, both in terms of different groups symboled (all five) and in types of symboling (all three). Frequencies and percentages of group/type symboling are not dramatically different, as Table 5 shows, though clans are the generally 55 preferred group. Ranked societies show a marked preference for clan symboling, and this representation is overwhelm- ingly spatial. Lineage spatial is next most common. Phra- tries and moieties are not represented and are in fact rare in the ranked category. Ranked systems in general display less variability and greater consistency with regard to both groups symboled (clan and lineage) and type of symboling (spatial) than their tribal counterparts. It is suggested that this pattern reflects a more systematic utilization of space by ranked systems, as discussed earlier. Clans and sodalities are the only groups symboled by states. The state sample is small and the figures suspect, but never- theless show a marked preference for symboling no group; the clan-spatial and sodality—ritual symboling are represented by only one case apiece. Lineage and clan The selection or determination by a society of which of two or more horizontal groups are symboled in mortuary context is clearly not a random process. It was shown above that lineages and clans are the most common groups chosen for symboling, and Appendix A shows that in only two cases (Bemba, New Ireland) does a society with two or more groups (here, two apiece) symbol all of them. All five horizontal groups, then, do not appear to have an equal chance of being symboled in a mortuary site, even though all five may be present in the living society. 56 An explanation for this phenomenon is suggested by the ethnographic literature. Throughout the course of data col- lection, a frequent and recurring pattern emerged whereby if a society did indeed symbol a horizontal group, it was usually the group considered - by the people as well as the ethnologist - to be the most socially "significant" or "important" in the society. When two or more groups were symboled, then at least one was the "important" group. Stated another way, "stronger" horizontal groups were usu- ally selected over "weaker" groups. This writer will be the first to admit that assessment of "significance", "impor- tance", "strength", and "weakness" of one group over another is a relative and subjective thing, as are the terms them- selves. At the same time, however, examination of the lit- erature demonstrates that such assessments are a common preoccupation of most anthropological fieldworkers. The ethnographic works used in this study devoted a substantial portion to an anthropological study of a socie- ty's kinship and social structure; the lack of such an analysis in a study was usually grounds for its rejection from the sample. The vast majority of the writers provided a judgment - usually explicit, sometimes implicit - regard- ing which of two or more social groups appeared to be the most "important" or instrumental in structuring overall social relations and general day-to-day living in a society. These judgments were made on the basis of various criteria including how incest/exogamy was defined and how marriage 57 was arranged and regulated. Also, the degree of "corpor- ateness" of a social group was important, i.e., the extent to which a group functioned in an economic, political, and/ or religious manner. Further, a group's physical location in space was often a factor, such as concentrated lineages versus dispersed clans. In short, while the assessment of "strength" or "weakness" of one horizontal group over ano- ther may be subjective, it is nonetheless a real phenomenon, and most writers made such judgments as a matter of course. There is no valid reason to suspect or suggest that such assessments are somehow all incorrect. On the basis of the foregoing, assessments of "strong" versus "weak" groups in the societies were incorporated into the present study. The vast majority of these determina- tions were gleaned from the ethnographic works; in some cases where no judgment was made (or noted), no group was considered dominant. A few were made by this writer on the basis of the criteria listed above. If it is accepted that strong or weak horizontal groups can be accurately identi- fied in a society, then this analysis can examine the extent to which this relative strength and weakness correlates with mortuary symboling. In no case was a phratry, moiety, or sodality observed to be the single most important group in a society; in no case were these groups more instrumental in structuring social relations than lineages or clans. Tables 7a and 7b show the relationship between strong and weak lineages and 58 Table 7a. Lineage Symboling, by Strong vs. Weak, by Category Strong Weak Cases Cases Cases Cases n sym. not sym. n sym. not sym. State: 6 0 6 l 0 1 Ranked: 12 7 12 10 O 10 Tribal: 20 7 l3 l6 0 16 Band: 0 0 0 0 0 0 TOTAL: 38 14 24 27 0 27 %: 37 63 g 100 Cases omitted: l, 10, 88-115 Table 7b. Clan Symboling; Lineage and Clan Symboling, by Strong vs. weak, by Category Strong Weak Cases Cases Cases Cases n sym. not sym. n sym. not sym. State: 2 0 2 5 l 4 Ranked: 10 9 l 12 3 9 Tribal: 24 15 9 22 3 19 Band: 0 0 0 0 0 0 TOTAL: 36 24 12 39 7 32 %= 522 22 12 :33 L and C: 74 38 36 66 7 59 %: 51 49 ll 89 Cases omitted: l, 10, 88-115 59 clans, and the extent to which these are symboled in mortu- ary context. As summarized, 74 cases have strong lineages and/or clans, and of these 38 cases (51%) symbol and 36 don't (49%). Sixty-six cases have weak lineages and/or clans; only 7 symbol one or both (11%) while fully 59 or 89% do not. Note that in Table 7a strong lineages are symboled only 37% of the time, and weak lineages not at all. Table 7b shows that when present, strong clans are symboled fully 66% of the time while weak clans evidence only 18%. Mini- mally, then, there appears to be a strong correlation between the horizontal group that is "dominant" in a society and the group that ultimately gets symboled. Notice, too, that out of the three socio-political categories, ranked societies are more likely to consistently and frequently symbol the strong group, while tribal systems seem to be less so inclined. Table 8 is presented to add detail to the preceding findings. This table is rather complicated, but careful inspection will reveal conclusions similar to those above. Again, there is a very strong correlation between symboling and strong group. Further, symboling of the weak group where the other is strong never occurs, i.e., if one group is strong and the other weak, the strong group or neither is invariably selected. Interestingly, where both or nei- ther lineage or clan is strong/weak, the clan is by far most often symboled (9 clan to l lineage), and symboling occurs almost twice as often as not. Symboling of both groups by Table 8. Lineage and Clan Symboling, by Strong vs. Weak, by Type, by Category, with Combinations of Symboling 60 Lineage Clan L_§_E N S A S A R S R none oth. State L/L l l L,C 3 2 1,9. 2 2 L,C l 1 Lisa C/g l l Rafikéd L/L 6 2 l 2 L,C 6 2 4 L,C 3 1 l 1---l L,C 3 l----l 2 L'9 4 i---—1----1 C/g 6 5 l Tfibal L/L 3 2 l L,C l4 5 8 1.9 4 1:11;: L,C 4 1---i L,C 10 i _____ 2 6 cxg 14 3-5:}; 6 (underlined: strong; oth.: cases symboling group(s) other than lineage or clan; ing lineage/clan with Cases omitted: 10, 88-115 connecting lines denote cases symbol- combinations of types of symboling). 61 the same society occurs only once, and in this case both lineage and clan were considered strong. The combinations of lineages and clans further show that strong lineages, with or without the presence of clans, are symboled less often than are strong clans, with or without the presence of lineages. Finally, Table 8 demonstrates that the symboling of a phratry, moiety, or sodality without the simultaneous symboling of a lineage or clan is rare. The nature of the relationships between lineage and clan versus phratry, moiety, and sodality representation are straightforward and do not require a table. Where a society symbols the lineage, phratries, moieties, or sodali- ties are never symboled. If the clan is symboled, any of the three can occur: six (1 phratry, 3 moiety, 2 sodality) occur in tribal systems, 1 (sodality) in a ranked society. Where neither lineage or clan are symboled, sodalities can be, about evenly in tribes (n=2), ranked (n=1), and states (n=1). These data must be viewed with caution due to the relatively low occurrence (n=ll) of any phratry, moiety, or sodality representation, and the frequencies of these groups are highly skewed toward the tribal category. However, it might be suggested that phratry and moiety symboling, while always infrequent, is "linked" to clan symboling. That is, a society may feel that it is important to symbol the phra- try or moiety only if the clan is considered important too. Sodalities are evidently not nearly so constrained: three occur with clan symboling, four by themselves. It might be 62 mentioned in passing that three of the four cases of sodality symboling without lineage or clan are North Ameri- can Indian societies, perhaps suggesting the importance of these organizations in these societies relative to societies elsewhere. Finally, Table 9 illustrates relationships between the three types of symboling by lineage or clan, specifically the correlation of artifactual and ritual symboling with the predominant spatial mode. The table shows no overwhelming patterns, except that clan spatial symboling is more likely to also involve artifactual and ritual symboling than line- age spatial. The table shows that the artifactual and rit- ual symboling of a lineage or clan can occur about equally with or without concommitant spatial symboling. The presence of clan spatial patterning can be taken as a pre- dictor of artifactual or ritual only about half the time. To summarize this section, the data show that all hor— izontal groups do not have an equal probability of being symboled in mortuary context, and indeed, as discussed pre- viously, just under half of the cases in the sample that have such groups symbol none at all. For those that do symbol, the choice of group seems to be dictated, in the vast majority of the cases, by the degree to which a group is important in the living society, relative to other groups. Strong clans, then lineages, are by far the most common occurrences. Where the strong—symboled/weak-not symboled rule is broken or where neither lineage or clan is strong or 63 Table 9. Correlation of Lineage or Clan Spatial Symboling with Artifactual and Ritual Symboling, with Combinations of Types Lineage Clan L & C Spatial Art Rit Ar Rit A & R Lineage 1 1 Clan 2 3 2 ----- 2 7 Neither l l 2 64 weak, the clan is the most common choice, followed by infre- quent occurrences of phratry, moiety, or sodality. This correlation is important from the standpoint of archaeological expectations. A generalization can be offer- ed about this phenomenon, which is parallel to a generaliza- tion by Saxe regarding symboling in the vertical dimension. Saxe (1970:7lff) observed that upon the event of death, not all of an individual's social identities can or will be represented in the mortuary rite. Rather, The greater the social significance of the de- ceased the greater will be the tendency for the social persona represented at death to contain social identities congruent with that higher position at the expense of other (and less socially significant) identities the deceased may have had in life, and conversely (Saxe (1970:71; parens in original). That is, when the living make decisions about the treatment of the dead, factors "...which are structurally more signi- ficant should tend to take precedence over others" (Saxe 1970:72). In the jargon of the present study, then, the factors chosen for mortuary symboling of an individual will be those that are "stronger", "more important", or "more socially significant" than others. A parallel process is proposed in the decisions affecting the symboling of horizontal structures. It has been shown that when 3 horizontal group i§_symboled, it ig usually the one(§) that peOple consider dominant in the social structure gf_their society relative tg_and at the expense 9f other groups. This postulate regarding the horizontal dimension 65 differs from Saxe's vertical dimension hypothesis in one important way. Saxe's hypothesis is necessarily ego-Speci- fig, in that decisions must be made for every deceased indi- vidual upon the event of death. Of course, mortuary treat- ments are usually conventionalized or standardized for persons of roughly equal vertical status, but nonetheless, the necessity for ego-specific decisions can create distinc- tions or variability in the mortuary rite. The present postulate, on the other hand, is group-specific. The deci- sion to symbol upon death a person's horizontal group affiliation is based on only one factor: whether the group is symboled at all. Stated differently, when somebody dies, ego-specific symboling involves selection between various personal attributes, while group symboling is a phenomenon already decided upon, and was selection made between groups, not persons. Thus, vertical symboling can be fairly charac- terized as having a scalar, or continuous, quality, while horizontal is presence-absence, or discrete. When present, horizontal group symboling will be expressed ig the same manner for everybody in the group, or nearly so (gf. O'Shea 1981:46-47; 1984:49-50), whether spatial, artifactual, or ritual, or some combination of these. The point of this discussion is to illustrate that, while horizontal group symboling may be less frequent in mortuary context than the vertical, if present, it will be less variable than the vertical. A dead person's various social identites may be Open to selection, but his clan is simply symboled or it 66 isn't, and if it is, it will likely be symboled the same way any other clan member is symboled. The selection of various individual social identities based on vertical stat- us may create variability and even ambiguity in the archae- ological record, while the horizontal dimension is far less variable and is either symboled or it's not. Descent groups, cognatic descent systems, and disposal areas It has been shown above that lineages and clans are the most frequent horizontal groups in the sample that are symbolized in mortuary context, and that this symboling is predominantly Spatial, followed by ritual and artifactual. In terms of develOping expectations regarding the archaeolo- ical recognition of such groups, the spatial mode clearly is the most visible, the artifactual rather less so, and the ritual not at all. This is, then, a correlation of impor- tance to the archaeologist: that lineages and clans often dispose of their dead in specialized, more-or-less fOrmal disposal areas. These disposal areas would be recognizable; cemeteries are most common, followed by group crematories (with Or without subsequent burial), tombs, and crypts. At this point, the extent to which disposal areas are isomorphic with UDGs should be tested. Table 10 shows the breakdown of the 53 cases that use disposal areas by whom is disposed therein. Thirty-five or 66% of the disposal areas are used by lineages or clans. Nine (17%) represent disposal areas used by societies organized by non-unilineal 67 Table 10. Correlation of Disposal Areas with UDGs Disposal Disposal Disposal Area = Area # Area = UDG UDG non-UDG TOTAL State: 1 l 2 4 Ranked: 17 2 0 l9 Tribal: l7 6 4 27 Band: 0 0 3 3 TOTAL: 35 9 9 53 68 or cognatic descent; these cases are generally cemeteries or crematories used by the bilaterally related occupants of a single settlement. Finally, 9 or 17% of the cases are soci- eties that are organized on the basis of unilineal descent, but where the disposal areas include all members of a set- tlement, and where the disposal areas evidently do not dif- ferentiate between different UDGs. In sum, 66% of the dis- posal areas in the sample represent UDG-specific facilities, while 34% represent something else. Minimally, it can be Observed that, based on the pres- ent sample, one may expect a disposal area to represent a UDG about two-thirds of the time. This figure, by itself, is not encouraging. Fortunately, it is possible to go beyond it and offer additional statements regarding the reliability of the association between disposal areas and UDGS. First of all, it might be noted that in Table 11, 5 of the 9 non-unilineal descent cases are bands or states. Re- garding the former, the definition of band includes the assertion that UDGs do not occur. As discussed previously, the use of disposal areas by bands is in any case rare. Similarly, disposal areas are uncommon in state systems. To the extent that an investigator has some idea about the nature of socio-political-economic complexity in the society under study, he can then interpret the significance of a disposal area accordingly. That is, if an archaeologist has reason to believe, based on other data, that the society 69 under study was band—organized, then UDGs will not occur and a disposal area will thus not represent a UDG. If the so- ciety is believed to be state level, then the same caution applies. In short, other kinds of data must be brought to bear to assess the probabilty that a disposal area repre- sents a descent group or not. An indication of band or state level organization decreases that probability to zero and negligible, respectively. For ranked and tribes, the former category shows no cases Of cognatic descent. By definition, all ranked socie- ties will be organized on some variation of the principle of unilineal descent. Thus, if a study of the vertical dimen- sion of mortuary data suggests ranking, then UDGs must be present, and the association of a UDG with a disposal area is an inference of very high probability. Tribal systems are more equivocal and, as usual, display the greatest amount of variability. If previous analyses of mortuary and non-mortuary data suggest the absence of ranking but more complexity than band, then the presence of a disposal area can be taken as a "reasonably" strong inference that a UDG is interred therein. Actually, the presence of a dis- posal area in a tribal system should be the basis for a working hypothesis that UDGs may be present. The last point is important, and recalls a subject discussed earlier, that of the necessity for a multi-dimen- sional approach to horizontal groups in particular and to the analysis of prehistoric socio-political structure in 70 general. The use of multiple converging lines of inquiry is necessary for the testing of hypotheses regarding horizontal groups in a prehistoric society. The reader may have noticed by now that in the pre- ceding pages, only infrequent reference has been made to particular societies, the data instead being considered in aggregate. Non—conforming or contradictory cases have not been "explained away" by reference to unique or unusual structures or behaviors of particular societies (though frankly, such could be done legitimately as often as not). In the present context, however, certain non-conforming cases must be discussed, since some demonstrate archaeolo- gically distinguishable patterning, and others may be Otherwise atypical. Two cases were recorded in which UDGs were present, but disposal areas consisted of family members. The Mossi, a state system, reportedly buried their dead in family groups in agricultural fields near their houses. The Mossi have a strong lineage structure, and the family diSposal areas could be interpreted as lineage disposal areas. How- ever, since the luxury of second guessing was expressly forbidden, this extrapolation was not made. The point is, it would appear that a Mossi family cemetery would be archaeologically indistinguishable from a lineage cemetery, which in fact, it probably is. Such does not seem to be the case for the Nootka. Nootka are also reported to utilize family cemeteries, the 71 location of which is left up to the families. Often caves, islands, or other isolated spots are used; this mode of dis- posal could almost be considered no pattern. In any case, Nootka family cemeteries are distinctive and are unlikely to be misinterpreted as UDG disposal areas. The Pawnee, herein considered a ranked society, were divided into four societal segments, referred to in the literature as "bands". Each of the four "bands" was evi- dently spatially distinct, and each had its own communal cemetery. It can only be speculated whether such a cemetery would be distinguishable from a UDG cemetery, and logical arguments could be made either way. The remaining six cases appear to be clearly abberant: Omaha, Ojibwa, Fox, Mandan, HOpi, Zuni. All of these societies reportedly lived in multiple UDG settlements but evidently did not segregate the dead according to descent group membership. Rather, a general "village" disposal area was used, a scaffold area for the Mandan and cemeteries for the rest. It may be worth mentioning that the Omaha, Ojibwa, Fox, and Zuni symbol the clan through artifacts and ritual, the Zuni sodalities through artifacts and ritual, and the Hopi sodalities through artifacts. It is interesting that these six cases are all North American Indian societies; indeed eight of the nine non- conforming cases are North American. The writer has specu- lated that, in some cases at least, non-conformity mgy_be the result of cultural disruption prior to anthropolOgical 72 study. For the Zuni, for example, there are indications that village cemeteries may be a recent introduction replac- ing previous house floor burial (Cushing 1896:336, 365-366). However, no statements were discovered explaining exactly why this apparent shift occurred. Similarly, the Mandan are reported to have formerly practiced no pattern burial and later adopted scaffold disposal areas (Bowers l950:24,99). In lieu of clear evidence that the mortuary programs repor- ted for these societies are somehow reflective of cultural disruption through contact with Europeans, these cases could not be legitimately rejected at the outset, and will not be rejected now. However, the fact that eight out of the nine contradicting cases are North American may be more than coincidental. On the other hand, it may not, and these societies cannot be dismissed simply because they do not conform to the behavior of other people elsewhere in the world. At this point, two general principles will be offered that derive from the foregoing. First, it is posited that where UDGs are present ip g society, they will consistently segregate their dead, usually gp_the basis gf_one pf these groups. Out of 89 societies in the sample with UDGs, only 9 seem to contradict this rule, and there is some question about some of these. The dead are segregated by all three modes of disposal: house-related, disposal area, and no pattern. Only in the disposal area mode and possibly a few cases of house-related is UDG segregation a deliberate, 73 intentional behavior. Secondly, it is proposed that where present, a more- gp-less exclusive and formal dipposal area will contain the deceased members pf 3 UDG. As noted earlier, the accuracy Of this inference will vary with the nature of socio-politi- cal complexity in a society. Further, it has been stressed that while this is herein considered an inference of high probability, this study does ESE claim that any disposal area is g3 £3339 proof of the presence of a UDG. Rather, the presence of a disposal area can be considered the basis of a working hypothesis that requires further investigation. A brief consideration of cognatic descent systems is in order. Table 11 shows the 26 sample societies lacking unilineal descent broken down by disposal mode and category. There is little to learn here. The totals indicate that systems of this type tend to slightly favor no pattern dis- posal, though this fact is heavily weighted toward the band category. No dominant patterns other than that are appar- ent, and mode of disposal by these systems seems rather variable. It might be added in passing that four out of the seven tribal societies have non-kin sodalities, but these are not symboled in mortuary context. Descent groups, disposal areas, and settlement patterns It has been established that, with few exceptions, societies with UDGs will segregate their dead, usually on the basis of one of these groups, and that a formal and 74 Table 11. Correlation of Disposal Modes with Non-Unilineal Descent Systems, by Category House- Disposal NO related Area Pattern TOTAL State: 0 2 0 2 Ranked: 0 0 0 0 Tribal: l 4 2 7 Band: 5 3 9 17 TOTAL: 6 9 14 26 75 exclusive disposal area has a high probability of contain- ing the deceased members of a UDG. This section will seek to demonstrate how UDG disposal area spatial patterning is related to - and in some cases may be determined by - the spatial arrangement of UDGs over the landscape. Table 12 shows the spatial configuration of lineages and clans - whether single ("Si") or multiple ("Mu") - in a settlement, juxtaposed with the presence of lineage or clan disposal areas; non—conforming cases are also included. The single lineage, single clan, and single lineage and clan categories are essentially autocorrelations, showing that where a settlement consists of one discrete UDG, and where a disposal area is used, it will naturally be a single UDG disposal area. (Incidentally, it is in regard to some of these cases where the aforementioned g3 £2259 symboling enters in, i.e., it is not entirely clear from the litera- ture whether some of these cases maintain a disposal area for the purpose of representing a UDG, or whether a settlement-related disposal area is used simply out of expediency. In either case, the archaeologically visible result is the same). The two categories "multiple lineage, single clan" and "multiple lineage, multiple clan" show an interesting phenomenon. In the former, the symboling of the single clan occurs over the multiple lineage 8 to l; in the latter, the multiple clan and multiple lineage are evenly represented at 4 apiece. The combined totals are 12 clan to 5 lineage. The final categories, "multiple lineage, no 76 losooo so: moon I =I= .osasoass u :22. .oao mamasm =um. lucmsmHuumm ".ouomv msoflno pconHH mcmsfim 3oz Hmmnx as I mmcsom mxsm oxsm Donna: I :2 cameo mocosoama sandaawm HQSN coped: moonmd .3 xsua .mH Hdom mxuooz mflosqouH Cow Haocmumz Husband oz :2 won Hmmoz uflmaflaa flmcmaama OHOA osmose comoa copE¢ am a: omcm mm: masons Madame .mH Hmumom >HB onamm mpcmo am am UHSM onfihmow Monaco: woos maoou waaom Hm I omcoz mouoam condom: I am usxow mcHEHomcoo cmHO mommpmq cmHO mmowcfiq Guam 0 A Icoz Hogans ooxcmm mumum .uuom muommuoo an .mcoauousmfimsoo usmsmauumm CMHO\mmomcflA cues mound Homommfla GMHO\mmmocfiq mo coflumamuuov .NH magma 77 clan" and "no lineage, multiple clan", shows that these societies somehow spatially distinguish the different UDG dead through the use of disposal areas. Table 12 illustrates three related points. First, it is clear that while these societies (excepting the non— conformers) segregate their UDG dead, the smaller 9f the two groups i§_not necessarily the one that will pg_symboled. Rather, the reverse is more likely, that the clan will be symboled over the lineage. The second point may explain the first. In the 12 cases of clan over lineage symboling, 7 are strong clans, 3 are strong lineage and clan, and 2 are weak lineage and clan. Of the 5 lineages symboled, 4 are strong lineage and l is strong lineage and clan. Thus, it will be recalled that in cases where the clan is strong and the lineage weak, or both/neither are strong/weak, the clan is the favored group. In the "multiple lineage, single clan" category, the overwhelming dominance of clan disposal areas probably reflects the simple fact any local settle- ment is comprised of members of the same clan, and residen- tial or spatial contiguity promotes the effective function- ing of the group, and hence its relative importance; where single strong clans occupy a settlement, then it would seem to be "unnecessary" to symbol the lineage. In the "multiple lineage, multiple clan“ category, the symboling of the clan reflects the segregation of the more inclusive UDGs in the settlement gig 3 gig other like groups. The societies who chose to symbol the less inclusive group, the lineage, are 78 also g3 facto segregating their clan members as well. Finally, a third point derives from the foregoing: 1E ig gpparently_not possible Ep_distinguish 3 lineage from g clan 12.3 disposal area 93 the basis 9: spatial patterning alone. Minimally, the presence of a lineage or a clan can be infer- red from the presence of a disposal area; probability favors a clan, but the disposal area by itself evidently cannot indicate which. The data to be discussed in the remainder of this section is presented in Table 13. In order to add detail to the foregoing and to develop more explicit archaeological implications for the identification of UDGs, it is necessary to discuss the types of disposal areas involved and their internal structure; various individual cases will be brought in where necessary. Table 13 shows four categories of settlement/disposal area association, which will be explain- ed in turn: 1) Single UDG disposal area, one per settlement: This category is straightforward and contains 18 cases in which a single UDG settlement is associated with a single disposal area, in which all members are interred or processed. Note, too, that the table contains a class "concentrated vs. dis- persed". These terms describe the nature of the habitation sites. "Concentrated" settlements can be thought of as more or less compact, nucleated villages where habitation structures are located fairly close to one another. "Dis- persed" settlements are sometimes referred to (Murdock 1967) Table 13. 79 Correlation Of UDG Disposal Area Types with Attributes of Settlement Pattern Conc DA or UDG Div Type Disp? Case Sym Cat. sett? Si UDG cem conc Bemba L R - DA; 1 per Mbundu L R + Sett. Kurd L R - (n=18) Bella Coola C R — Thonga C R - crem/cem conc Tlingit C R - LOlO C R - Yokut L T + crypt conc Flores C R +? Ambon C R + cem disp Mapuche L T + Mongo C T +? Goajiro C T - Tiv L T - Tallensi C T - Mae Enga C T + Marshall Is. C R + crem/cem disp Monguor C R +? Si UDG cem conc Tucano C T - DA; 1 per Ganda C S - UDG tomb conc Tanala C R + (n=8) Easter Is. L R - crem/cem conc Khasi C R +? crem conc Toda C T - cem disp Truk L T + tomb disp Talamanca C T - Si UDG cem conc Siwans C R + DA; 1 per Dogon L T + UDG per W. Apache C T + Sett. Iroquois C T + (n=5) cem disp New Ireland C T + Mu UDG cem conc Ashanti L R - DA; 1 per (divided) Puka Puka L R + Sett; Fellahin L T - (n=4) di5p Yao L T + 80 as "hamlets" or "neighborhoods", where there is some degree Of dispersal or spacing of structures beyond that Of a con- centrated village, but where the aggregate whole is still recognized as "the settlement". In the concentrated settle- ments, the disposal areas are usually situated adjacent to the settlement, or rarely (e.g. Flores, Ambon) within its boundaries. For dispersed settlements, disposal areas are located somewhere within the area Occupied. 2) Single UDG disposal area, one per UDG per society: Societies in this category utilize one disposal facility for every UDG in the system. Deceased persons are brought from wherever they may have been living for processing or inter- ment in the facility. Types of facility and Spatial arrangements range from burial in a single clan cemetery beneath a clan Great House (Tucano), to the more common cemetery or tomb, either located within a fairly circum- scribed, concentrated UDG territory (e.g. Easter Island, Tanala), or otherwise located in some convenient, often "traditional" spot (e.g. Ganda). 3) Single UDG disposal area,one per each UDG per settle- ment: Five societies live in multiple UDG settlements, and maintain spatially separate disposal areas for each UDG. Disposal areas are located within the boundaries of a dis- persed hamlet-type settlement (e.g. New Ireland), adjacent to a concentrated settlement (e.g. Iroquois), or even within a concentrated settlement (e.g. Siwans). In one case (W. Apache), it was not entirely clear whether the disposal 81 areas were located nearby or somewhat removed from the habitation Site. 4) Multiple UDG disposal area with internal divisions, one per settlement: This final category includes four societies that live in multiple UDG settlements and maintain a dispo- sal area nearby. Deceased UDG members are interred in Spa- tially distinct and defined divisions within a cemetery. Based on the above, explicit implications can be derived for the recognition of UDGS. This UDG disposal area patterning must be contrasted with other patterning to determine the extent of possible obfuscation. Beginning with the first category, single UDG disposal area, one per site, archaeologically recognizable patterning would be straightforward: a concentrated settlement with an adjacent disposal facility, or several nearby hamlets with an asso- ciated facility. It is of paramount importance to point out that all non-conforming pp non-unilineal descent cases resemble the single concentrated settlement-Single disposal area pattern. The reader will recall that 9 cases lacked UDGS but maintained disposal areas, and 9 cases had UDGS but apparently did not segregate the dead. The literature on these 18 societies indicated that they utilized diSposal areas directly adjacent to the habitation settlement or, in one case (Java) somewhere within it. Thus, it can be con- cluded that the primary source of obfuscation is where dis— posal areas are located adjacent to habitation sites, since this pattern can represent single UDG disposal areas 82 (n=18), cognatic group disposal areas (n=9), or multiple UDG disposal areas (n=9) that lack internal differentiation. It is further noteworthy that all but one (Mossi) of the non-conforming/non-unilineal cases (even the three band societies) are reported to live in more or less discrete, concentrated settlements. Expectable patterning deriving from the second cate- gory, single UDG diSposal area, one per UDG, would be a lack of a disposal facility anywhere near most or all habitation sites, with the exception of the Tucano cemetery-under- Great House. Concentrated versus dispersed settlement has no bearing here. Disposal areas would be few, and presum- ably could be located anywhere. Further, since all UDG members are interred or processed through a Single facility in the same place, a large quantity of human remains would be expectable, particularly if a cemetery, tomb, or crypt is used. It might be added that this mortuary pattern was not noted for anything other than single UDGS, such as multiple UDGS maintaining the same facility. Also, there were no cases in which a single UDG maintained two or more facili— ties that were spatially removed from settlements; if more than one facility is used, they are invariably located some- where near the settlements of the users. The last two categories deal with multiple UDG settle- ments and demonstrate two highly recognizable Spatial pat- terns: distinct UDG disposal areas near or within a settle- ment, and Single disposal areas with internal divisions near 83 settlements. Such patterning would be unambiguous in archaeological context. Multiple cemeteries and divided cemeteries occur pply_where a society seeks to distinguish UDGs. No other cases were noted where multiple or divided disposal areas were intended to signify something else. It should be noted that the four divided cemeteries are divided into formal sub—areas; UDG affiliation is not represented by burial in rows or non-patterned family plots. It is also interesting to note that these four societies symbol the lineage rather than clan (one case, Puka Puka, reportedly lacks clans). At this point, other kinds of intra-disposal area patterning that does not symbol UDGS Should be summarized. Four cases reportedly bury in family plots within the ceme- tery (Kurd, Bella Coola, Java, Yao). It seems unlikely that such family plots would be confused with formal divisions, and indeed, the Yao do both. Two societies buried in rows: the Monguor buried cremated remains in family plots with generations in rows, and the Salish buried in rows for no apparent (or reported) reason. This discovery is disap- pointing from the perspective of some recent mortuary stud- ies that suggest that rows can be an important organization- al feature of cemeteries (e.g. Goldstein 1976, Milner 1984). Two societies (Mae Enga, Thonga) utilized cemeteries for male UDG members only; females and children are bured else- where. Three cases spatially distinguish by sex: Zuni in a cemetery, Toda by two adjacent crematories (no burial), 84 and Khasi by separate, adjacent cairns for cremated remains. Several societies utilized disposal areas for most people, but excluded some others Egg on the basis of status or deviancy. Many interred only adults in disposal areas, with subadults disposed elsewhere. The Mbundu practice no pattern burial for hunters. The Trukese will bury an indi- vidual somewhere other than the lineage cemetery after a deathbed request, which is apparently a rare occurrence. Some portion of the W. Apache tribe practiced no pattern disposal rather than the more common clan cemetery. In the Marshall Islands, clan cemeteries are used but also some no pattern and sea burial; an explanation for this was not found. Several cases were discovered where different body positions were used, but only one denotes UDGS. The Nyakyusa utilize different grave shapes to symbol lineage affiliation, and round or oval graves, being small, thus contain flexed burials, while oblong graves contain extend- ed burials. Finally, at the outset of this study considerable in- terest was placed on discovering cases where combinations of processing modes were utilized, such as cremation with in- humation, or primary with secondary burial, since such com- binations often appear archaeologically. The results are disappointing. The Yurok are reported to have cremated most people, but status individuals were buried. One case, Goajiro, combined primary burial with secondary urn burial 85 in a clan cemetery. Unfortunately, no explanation for this distinction was found. Further disappointing was the rela- tive rarity of any secondary burial. In archaeological context, secondary burials are often presumed to be indivi— duals who died away from the prescribed disposal area and who could not be brought there immediately for burial, e.g. during the winter months. In the present sample, the dead were evidently brought to the facility straight away. All secondary burial was apparently done deliberately and out of preference, or at least no other reasons were reported. One other correlation between settlement structure and disposal area patterning should be mentioned. Table 13 includes a class "Divided Settlement?" which denotes whe- ther a settlement is internally divided on the basis of some criterion. This information was taken from Murdock's (1967) categorized data, and is somewhat less than clear. Note that in Table 13 four societies that according to Murdock have "clan barrios" also have an accompanying ques— tion mark, meaning that the ethnographic data contradicted this, at least in this author's reading of it. However, possible discrepancies aside, taking the information at face value shows an interesting — if not overpowering - tendency: that single UDG settlements tend to be motinter- nally subdivided, and that multiple UDG settlements do. The latter Show the stronger association. The ambiguous factor is the criteria by which the settlement is divided, 86 which was not provided by Murdock and could not always be determined from the HRAF data. Many cases did, however, indicate division into UDG-based areas. In any case, if a habitation site can be determined to be internally divided into "wards" or "barrios" (1219-)! then the archaeologist can in turn hypothesize the presence of multiple UDGS. If multiple or internally divided disposal areas are discovered as well, then this becomes an inference of very high proba- bility. Considered separately, however, it is clear that multiple or divided disposal areas are a far better single indicator of multiple UDGS than are divided settlements. Disposal modes and subsistence patterns Appendix A shows subsistence practices for all socie- ties. These data were extrapolated from Murdock's (1967) tabulations. An examination of modes of disposal of the dead with dominant subsistence pattern (Table 14) reveals no surprises. The sample is heavily weighted toward primary agriculturalists, who demonstrate a marked preference for disposal areas, and then house-related disposal. Hunters and gatherers clearly favor no pattern disposal. The 9 primarily pastoral societies show no marked preference, and societies whose primary subsistence mode is fishing demon- strate a slight preference for disposal areas. These findings mirror those of Binford (1971), who showed that agricultural societies generally tend to mani- fest more structured and varied mortuary treatments than 87 Table 14. Correlation of Disposal Modes with Primary Subsistence Modes, by Category Primary Subsistence Modes Agri. Pastoral Fishing Hunting Gather. HR 6 State: DA 4 NP HR 8 Ranked: DA 16 3 NP 1 HR 11 4 l Tribal: DA 18 3 l 3 2 NP 6 l 5 2 HR 2 2 1 Band: DA 1 2 NP 1 l 3 7 HR 27 37% 4 3 1 TOTAL: DA 39 53% 3 6 3 2 NP _1 10% _2 _1 _§ _9_ 73 9 10 11 12 88 hunter-gatherer groups. However, this study concurs with O'Shea (1984:6-7) who notes that Binford‘s results are some- what misleading, in that the determinant of increased com- plexity is not necessarily agricultural subsistence versus hunting-gathering, but rather relative sedentism versus mobility. That is, greater or increased structure probably coincides with sedentism regardless of subsistence mode, and mobility does not promote such elaboration. The data in Table 14 can probably be taken as support for O'Shea's thesis, to the extent that agricultural subsistence and perhaps fishing correspond to sedentism. Further, 38 of the 49 cases that symbol horizontal groups are agriculturalists, demonstrating some degree of greater complexity in agricul- tural/sedentary societies relative to all others. DiSppsal patterning and descent An examination of disposal patterning and descent reckoning produced no useable results, and for this reason, the reader will not be encumbered with a table. The sample is highly skewed toward patrilineality (n=50), followed by matrilineal (n=24), bilateral (n=18), double descent (n=11), and undetermined (n=12). Because of the preponderance of unilineal descent systems, any results should be considered unreliable, but for the record, patrilineal and matrilineal systems favor disposal areas, and matrilineal societies tend to symbol some horizontal group more frequently than any other descent system. 89 The problem 95 house-related disposal It was mentioned near the outset of this chapter that observed instances of house-related disposal presented a conceptual problem, in that it was not clear the extent to which house-related disposal constituted horizontal group symboling. Despite the fact that the literature occasion- ally indicated that, for example, people are buried in the house floor in order to keep the lineage together even in death, house-related disposal is not herein regarded as symboling. The problem is twofold: 1) it is not certain, even in cases where there was some indication that the group "should be together", that this behavior was truly intended to symbol the horizontal group, and 2) unlike disposal areas, §g_fggpg spatial patterning of groups could not be interpol- ated. That is, even when the literature did indicate deli- berate symboling, in all but one case (Tucano; scored as disposal area) the members of the UDG lived in more than one structure, and the UDG was thus "symboled" in the ground under several of them. This pattern, then, is identical to house-related disposal where the UDG is not intentionally symboled (or even present), and is hence archaeologically ambiguous. During the course of data collection, however, certain patterning was noted that, while not being really redundant and wholly unambiguous, could be of value to the archaeolo- gist and Should be discussed at this time. Out of 35 socie- ties using house-related disposal, by far the majority are 90 indeed ambiguous, both with regard to horizontal group sym- boling and archaeological patterning. Several African socie- ties, however, displayed a pattern that can be called "com- plex house" burial, which is far more structured than most. Four cases were noted (Ngoni, Swazi, Zulu, Igbo) and three possible others (Azande, Rundi, Ila) in which kraal-type settlements are used, usually by a single UDG; these kraals, or compounds, constitute single UDG residence areas within a larger, often dispersed, multiple UDG settlement. Mortuary practices are of the house-related variety but are more com- plex than simply burying everybody in the floor. To gener- alize, the eldest male head of household is typically inter- red under the floor of his sleeping room or otherwise some- where in his house; non-elder males are buried somewhere else, say, within the compound near one of the walls. Fe- males will likewise have a spatially distinct area, perhaps just outside the compound wall, or on the opposite side of the compound from the males. Infants and children may be buried in some spatially distinct spot. The point to be made here is 1) that this kind of mortuary pattern is more complex and structured than more typical house-related dis- posal, and would be archaeologically interpretable, and 2) this kind of structured disposal program was not observed to be utilized by anything other than a distinct UDG (in the four cases mentioned, all lineages). Spatial locations and other elements of funerary treatment are determined on the basis of age, sex, and status, but within a UDG; the latter 91 provides the pool of peOple about whom age, sex, or status distinctions are made. Stated more generally, it appears that as complexity of house-related disposal increases, so does the probability that the mortuary programs are under- taken by a distinct UDG. It is tentatively suggested, then, that "complex house" mortuary programs, if discovered by the archaeologist, may indicate the presence of a single UDG. At the very least, such a pattern can be taken as a working hypothesis to be tested with additional data. The Second pattern worthy of mention involves the pre- viously discussed division of settlements into "wards" or "barrios", within many of which dwell members of separate UDGS. It was pointed out above that the probability of accuracy in inferring that a disposal area represents a UDG approaches 100% in ranked societies, since by definition all ranked societies have UDGS, and the probability that a dis- posal area represents something else is therefore negligible. A similar, if weaker, case can be made for house disposal. If clear evidence of ascriptive ranking in a system is dis- covered, then the presence of UDGS can be strongly inferred; if house-related disposal is discovered, and if clear, unequ- ivocal evidence of wards or barrios is present, then the in- ference may be permitted that these subdivisions may repre- sent UDG residential areas, and the dead are UDG members. The problem involved in an inference such as this is the fact that in some complex ranked systems, spatial subdivi— sions in habitation sites may not in fact be made along UDG 92 lines (see Chapter 4). That is, the spatial configuration of the domain is often determined by administrative or chiefly decree and need not conform to spatial arrangements of UDGS. In any event, the correspondence of house-related disposal and distinct settlement divisions may be taken to represent hypothesized UDGS, particularly in ranked systems. Again, however, this inference by itself is not terribly strong, but can serve as the basis for further testing. Summary Discussion With the presentation of findings completed, it remains to summarize and emphasize some of the main points in the discussion. First and foremost, it should be reit- erated that no claim is made that the patterning and corre- lates presented above are an exhaustive survey of the nature of horizontal group symboling in mortuary context. Rather, it is asserted that, through the invocation of uniformitar- ian principles, the patterning observed in the 115 society sample should have counterparts in the archaeological record. In this regard, it should be restressed that the frequencies and percentages of the 115 societies that symbol is less important than the fact that most of this symboling results in structure that is, for the most part, redundant and unam- biguous, and is archaeologically recoverable. Put another way, less concern is placed on, say, the fact that only about one-third of the cases symbol horizontal groups in the 93 archaeologically visible spatial mode than with the fact that this symboling is indeed recoverable and interpretable by the archaeologist. It can never be known what proportion of all prehistoric societies symboled horizontal groups or how they did it, but if some did, and in a manner similar to the ethnographic cases, then archaeologists are now in a position to recognize this. The results discussed above clearly demonstrate that societies do in fact symbol horizontal structures in mortu- ary context more often and more consistently than suggested in O'Shea's (1981, 1984) study of mortuary variability. While O'Shea is very correct when he posits that the verti- cal dimension seems to take precedence, horizontal dimension symboling isrxn:so rare or vague as he supposed. Further, it has been suggested that horizontal symboling can be ex- pected to be far less variable than vertical dimension sym- boling, to the extent that vertical dimension symboling involves the selection of various attributes of individuals, while the horizontal is represented the same way for every- one in a group. The present data set indicates that clans and lineages are the overwhelming favorites to be represen- ted in death, and that any and all horizontal groups in a society do not have an equal chance of being symboled. Thus, a theoretical postulate has been offered to account for this phenomenon: Postulate 1: When 3 horizontal group i§_symboled ig mor- tuary context, ip_will usually pg the one(§) 94 that people consider dominant ig their social structure relative Eg_and g; the expense 9f other such groups. In no cases were phratries, moieties, or sodalities found to be dominant in a society; the most important groups were always lineages or clans, and these stronger groups werenmme often and consistently symboled than any other groups. This fact is important in the context of O'Shea's expectations for horizontal dimension symboling. Based on a study of three North American Plains systems, O'Shea con- structed a logical model consisting of archaeological expec- tations for discerning horizontal groups (e.g. 1981:42, 1984:43-47). For example, Suppose a society differentiated individuals in death on the basis of moiety membership. Such a distinction if observed in an archaeological context would produce a pattern with a number of predictable characteristics, such as: 1. The mortuary occurrences would be divided into two groups of approximately equal size. 2. The demographic composition of the resulting two groups would be similar to that of the population as a whole. 3. A number of other classes of mortuary dis- tinction would cross-cut the group division ...(O'Shea 1984:46). It can be argued that O'Shea's expectations would more prob- ably indicate the presence of two local lineages or clans utilizing the same cemetery. This assertion is made on the basis that O'Shea's model is a logical construct indicating that if a society is divided into two halves (moieties), and if the disposal area or other symbolic indicators are like- wise bipartate, then these must be the moieties. While this 95 is certainly possible, it is unlikely, and the above results indicate that when multi-partate structures are present, probability suggests the representation of less inclusive UDGS, lineages or clans. Further, O'Shea's model does not take into account settlement arrangement. While a society may be divided into two moieties, these halves need not be represented in every settlement or disposal area. Again, probability favors the presence of two less inclusive UDGS. In short, this research has shown that there is no "auto- matic" isomorphism between horizontal group structure and mortuary representation, such as moiety organization divides into two, three phratries divide into three, eight clans divide into eight, etc. The configuration of any single disposal area will be dictated by 1) whether any group is symboled at all, 2) the relative strength or weakness of two or more groups present, and 3) the social configuration of the local settlement using the disposal facility. Given this, it is expectable that such a pattern as described by O'Shea is less likely to represent a moiety than two clans or lineages. In the presentation of results, heavy reliance has been placed on the spatial dimension of mortuary variability as an indicator of the presence of horizontal groups. With this concern with space, this study follows the lead of Goldstein (1976, 1980), who was perhaps the first to recog- nize the fact that mortuary variability includes something beyond the simple presence/absence of grave goods, or at 96 least first explicitly emphasized this fact. This emphasis has since been echoed by many (e.g. Chapman and Randsborg 1981, Brown 1981). The results Show a general linear cor- relation between overall socio-political-economic complexity in a system and the structured utilization of mortuary space, to the extent that the use of formal disposal areas indicates greater spatial structure than house-related or no pattern disposal. The use of space as a symbolic indicator of the horizontal dimension also increases in linear fashion with complexity at the expense of ritual and artifactual indicators, at least from band through ranked society. It can be concluded that, as one moves from band through tribal to ranked systems, 1) the use of disposal areas increases over house and no pattern, 2) the spatial symboling of hori- zontal groups increases over artifactual and ritual, and 3) the incidence of horizontal group symboling in general increases. Interestingly, this pattern seems to fall off dramatically at the theoretical apex of ranked society and in state systems, a phenomenon to be explored in a later chapter. In the context of spatial patterning, a second postu- late with three correlaries, has been offered. Postulate 2: Where unilineal descent groups are present ig g society, they will with few exceptions con- sistently segregate their dead, usually gg the basis pf membership ig_one pf these groups. 97 The qualifier "with few exceptions" is necessary since the data indicate that this fact is unfortunately not omnipre- sent. Almost, butrmnzquite. Similarly, "usually on the basis of membership in one of these groups" indicates that UDG dead may be segregated on the basis of no pattern dis- posal, which is not done on the basis of group affiliation. With this, the first corollary can be added: Corollary 23: The presence pf g_more p; less formal, exclu- sive disposal area indicates with high prob- ability the presence gf g discrete unilineal descent group. The phrase "with high probability" denotes just that, that this is indeed an inference of high probability but that exceptions have been observed to occur. The relative strength of that probability will be affected by the socio- political context of the disposal area: it has been argued that an identification of band or state system will decrease the probability, while tribal or ranked increases it. This finding should be encouraging to students of ranked society, since all ranked systems will be based on some variation of the principle of unilineal descent, and will thus ipso facto have UDGS. One whole category of non-conforming cases, cognatic descent systems, does not occur here, which greatly strengthens the probability that disposal areas will repre- sent UDGs. The fact that ranked systems have been shown to more consistently symbol UDGS and to do so spatially further pushes that probability to close to 100%. Stated 98 differently, in ranked systems there is an extremely low probability that formal, exclusive disposal areas could represent something else. The nature of spatial arrangement of UDGs affects the probability of a disposal area representing a UDG(s) in tribal, ranked, and state systems. Perhaps the strongest correlation derived in this study deals with multiple UDG settlements and the use of disposal areas. Thus: Corollary 29: Where disposal areas are divided into formal divisions, g£_where multiple disposal areas occur around/near g settlement, they will invariably contain members pf the respective multiple unilineal descent groups inhabiting the settlement. There are no exceptions to this rule, at least in the pres- ent sample. As well as from the empirical findings, the writer is intuitively confident of this generalization, since with but few exceptions it is difficult to imagine the partitioning of a disposal area on some other basis. One obvious exception would be on the basis of status, which for the most part should be archaeologically self-evident. Another possible basis suggested by Ucko (1969) is differen- tial wealth, which again can be expected to be discernable. The locus of most uncertainty lies with single UDG settle— ments and disposal areas, since all non-conforming cases resemble this pattern (but see below). One final formal corollary of UDG disposal areas 99 should be added. It has been shown that where lineages and clans are present in a society, and where the society sym- bols one or the other, the smaller one - the lineage - is not necessarily the one that is chosen. Clans are the across-the-board favorites, perhaps reflecting the greater importance, on the whole, of this more inclusive societal group over the other. From the archaeological perspective, a third corollary to the postulate that people consistently segregate UDG dead presents itself. Corollary 23: If unilineal descent groups are spatially symboled 1g disposal areas, probability favors the presence pf g clan(§), though lineages can pg similarly represented. Other kinds of data may or may not shed additional light on distinguishing a lineage from a clan, e.g. concommitant artifactual symboling would tentatively suggest clan. In any event, archaeologists will have to content themselves with inferring the presence of lineages g; clans, but appa- rently not precisely which. The results regarding the artifactual symboling of horizontal groups are somewhat disappointing, from the standpoint of both the relatively infrequent ethnographic occurrences of such symboling as well as the poor recover- ability factor. These conclusions may not be as discourag- ing as all that though, and at least it is important to bear in mind the fact that any and all artifactual inclusions with a burial may not be indicative of vertical status. 100 O'Shea's comments in this regard are of interest. O'Shea (1981:50-52) discusses a hypothetical case in which three kin groups utilized a cemetery and where each group disting- uished itself spatially and with an artifactual totem. He states that if only one of the totemic symbols is non- perishible, then "...it is no longer possible to establish unambiguously the existence of a tri-partate kin structure" (1981:50). The present findings argue to the contrary: the presence of three Spatially distinct disposal areas, with a kin group-specific artifact in one, can be taken as strong evidence for UDGS; in this case, the artifactual occurrences support the spatial inference. Further, this study does not concur with the assertion that "the pattern might be inter- preted erroneously as a two-level rank structure, with limi- ted access to the obvious socio-technic artefacts, and with ranked and non-ranked burial zones" (O'Shea 1981:50-51). This position is taken because elsewhere in the same article O'Shea states that, horizontal distinctions should be expressed through channels of 'neutral' value. Hence, 'unvaluable' tokens such as clothing, coiffure, symbolically distinctive artefacts, and elements of body posture and orientation, should be common indicators of horizontal differences (1981:49- 50). The present data support this observation, and the question of kin group artifact versus indicator of "rank" devolves upon the archaeologist's interpretation of the artifact. Two observations can be made: 1) that artifacts of "neu- tral" value should be distinguishable from indicators of 101 rank or status, and 2) each archaeological case must be evaluated independently, i.e., artifactual indicators of kin groups affiliation or rank/status can be expected to vary considerably (or theoretically even be inverted) from system to system. The point to be made here is simple: artifacts will remain symbolically ambiguous if the archaeologist con- tinues to consider all artifacts as indicators of vertical status. Minimally, it can be concluded that horizontal group-defining artifacts or other "material" phenomena do occur, and the archaeolOgist should consider an artifact from both the vertical and horizontal perspective. A multidimensional approach It has been stressed repeatedly in the foregoing pages that a multidimensional approach is the most fruitful line of inquiry for the discovery of horizontal groups in mortu- ary context, and indeed in mortuary studies in general. The problem is straightforward, that of multiple verification of an inference made on the basis of a perceived pattern. Cer- tain UDG disposal patterns are in themselves inferences of high probability, others are somewhat less so. In either case, independent sources of confirmation or verification must be sought. A sequential hypothesis testing procedure should be utilized in order to avoid potential tautology or self-fulfillment of the initial inference. What follows is a discussion of the kinds of multidimensional procedures that can be employed in the search for UDGS in archaeological 102 context. As pointed out, different configurations of disposal areas can indicate different things to the archaeologist. Discussion will begin with the least ambiguous configura- tions, the multiple disposal areas associated with a site and the single disposal area with internal divisions. In the present sample, these patterns indicated multiple UDGS utilizing the settlement and disposal area(s) without excep- tion. The nature of intra-disposal area divisions must be stressed: in the present sample, these are invariably for- mal subdivisions or sub—areas within the larger disposal area. While it is clear that internal disposal area patter- ning may be less than obvious to the eye, as Goldstein (1976) has aptly demonstrated, or may be incredibly complicated, as shown by Chapman and Randsborg (1981), it can only be sug- gested that such formal divisions can be discerned. If such divisions are not obvious, then clustering of burials might be sought; in this endeavor, statistical pattern recognition techniques might be usefully employed (e.g. Goldstein 1976, Robertson n.d.). It is important that subdivisions or clus- ters be unambiguous gig g gig other kinds of structure, such as small family plots or row burial that do not represent UDGS. It might be added that row burial could denote UDGS in spite of the fact that such was not the case in the pres- ent data. Thus, a second line of inquiry presents itself. Bio-anthropoligical study of the mortuary population should be undertaken to evaluate the expectation that each 103 hypothesized UDG indicated by internal divisions, multiple idisposal areas, or perhaps rows is comprised of a normal cross—section of the population as a whole. Possible sourc- es of Obfuscation must be kept in mind at this point, such as the disposal of certain individuals or sub-groups (e.g. subadults) elsewhere. Epigenetic and genetic distance studies (e.g. Buikstra 1976, Droessler 1981) should be prof- itable in assessing the degree of biological closeness with- in versus between the hypothesized descent groups. Intra-disposal area material patterning can be em- ployed as a further test of hypotheses. First and most obvious, UDG-defining artifacts, if present and interpreted as such, would provide clear confirmation. Failing that, other material nuances of interment can be considered, e.g. grave shape, burial posture or orientation. Further, non- random artifactual associations with proposed UDGS might be sought. That is, even if unequivocal "totems" are not found, the inclusion of contrasting artifact sets in UDG divisions might indicate burial accompaniments that, while not exactly imbued with some special significance, might reflect differ- ences in mortuary ritual between groups. The inclusion of certain utilitarian or otherwise mundane artifacts might be "customary" within one group but not in another. It is strongly suggested that many artifacts that have been here- tofore assumed to be status markers be reevaluated in this light. Similarly, stylistic studies of the various artifact classes can be brought to bear. Regarding artifacts in 104 general, caution should be exercised in consideration of artifactual associations with specific burials. For exam- ple, perhaps clan totems are only included with males, or distinctive utilitarian artifact sets are included only with females. The greatest conceptual problem is, of course, the assessment of whether such artifactual inclusions are indi- cators of horizontal group affiliation or of vertical status. This writer cannot offer any hard and fast rules for telling the difference, and each case must be considered on its own. Turning to one of the weaker correlations, that of divided disposal areas and divided settlements, a strong case can be made that if both are associated, then the pres- ence of multiple UDGS is an inference of very high probabil- ity. However, the converse does not necessarily obtain, i.e., if divided/multiple disposal areas are present but settlement partioning is not found, this does not diminish the probability that the disposal area patterning represents UDGS, though it does not support it either. The artifactual dimension enters in here as well. Intra-settlement studies of artifact type and style may isolate UDG wards. Cross- comparison of artifact style between habitation and disposal sites could prove useful in confirming the presence of multiple UDGS. The most profitable test of hypotheses is inter-site analysis. Should unambiguous patterning suggesting multiple UDGS be discovered at one location, independent confirmation is strongest when additional data sets are employed. The 105 reasoning is obvious: if multiple UDGS are hypothesized at one site, then this pattern should be redundant throughout the rest of the socio-cultural system. The same kinds of pattern reCOgnition procedures described above - spatial, artifactual, biological - should be employed at other sites of the same system. Further, if divided/multiple disposal areas are indeed discovered at two or more culturally rela— ted sites, then inter-UDG artifactual patterning could yield insights into the relationships of the various UDGS over space. Such a line of inquiry could provide interesting results beyond the more typical comparisons between whole sites and/or disposal areas. Comparing assemblages associa- ted with hypothesized UDGs might narrow the overall range of variability and facilitate pattern recognition. A second UDG disposal area pattern involves the proc- essing or interment of entire UDGS in a single facility. As discussed previously, this pattern also has no exceptions; . such as multiple descent groups using the same facility for all of their dead, or of a coqnatic descent society proces- sing all dead members in a single facility. ArchaeolOgi- cally, this mode of disposal would result in a lack of settlement related disposal areas or house-related disposal. Some kind of "central" facility would be used, one per UDG in the society. It is problematical whether these facili- ties would be totally isolated or be associated with one particular habitation site. If such a facility was used by a large group and/or over a considerable period of time, 106 then one would expect a large amount of human remains in tombs, crypts, or cemeteries. It is uncertain whether a mass crematory without subsequent burial could be identified as such. Additional lines of inquiry are similar to, though more limited than, those discussed above. Within cemeteries there should be be no formal divisions, though presumably rows or family plots could occur. If group specific arti- facts or other material factors are used, they should be homogeneous throughout the entire burial population. Biological analysis would necessitate mortuary popu- lations from more than one site. Again, such studies are herein considered very important for the testing of hypothe- ses regarding the presence of UDGS. Each disposal area Should contain a fair cross-section of the population as a whole, and should demonstrate greater biological closeness within versus between disposal areas. Settlement pattern studies can be directed toward the further evaluation of the pattern. As stated, disposal of the dead would not be expected to occur near or within set- tlements, or at least disposal of everybody; it is concei- vable that certain classes of individuals, such as subadults, could be disposed of in such a manner. The spatial rela- tionships between habitation sites and the disposal areas could indicate the spatial configuration of UDGS within the system. Analyses of material culture patterning and style can be integrated into the research to determine whether 107 these variables tend to cluster within suspected descent group areas. Single UDG disposal areas adjacent or otherwise nearby single UDG settlements present the greatest difficulty in interpretation. This pattern could be confused with cogna- tic descent systems who, in the present sample, often bury or cremate deceased settlement members in such nearby facili- ties, or with multiple UDGs who inter their members in a single facility without apparent spatial differentiation. Artifactual data can serve as a check on this. If a facil- ity was used by a cognatic group, then no group-defining artifacts would be expected; if a single UDG used the facility, then any group-defining artifacts would be homo- geneous throughout that population and contrast with any other disposal area. If the disposal area was indeed used by multiple UDGS, and if these groups employed group-speci- fic artifacts as do four of these cases in the sample (though not all are recoverable), then these should indicate to the archaeologist that the population may not be a single UDG. Biological study can be employed as discussed above. However, it is uncertain whether biological analysis could discriminate between a single UDG and a cognatic group, or could segregate members of multiple UDGS; each would call for very fine-grained analysis with excellent and abundant data. Inter-disposal area analysis could be profitable, and should seek to demonstrate higher within-disposal area 108 similarity and between-disposal area dissimilarity. Finally, intra-habitation site spatial and artifactual study could be brought to bear: if clear evidence of wards or barrios is discovered, then this could suggest something other than single UDGS. At the very least, it would suggest with some degree of confidence that this was not a COgnatic group. One more converging line of inquiry can be employed that has not been mentioned. This is the analysis of the vertical dimension of status. It must be emphasized that the probability that UDGS existed in a prehistoric society approaches 100% in ranked systems. Thus, if ascriptive ranking is clearly indicated, a disposal area of any config- uration has a correspondingly high probability of represent- ing a UDG. If a state or band is indicated, then the prob- ability is negligible to nil. In a tribal system the probability is also good, but it is here that the greatest amount of ambiguity can occur. Finally, a brief summary should be provided regarding various potential sources of obfuscation that the archae- ologist might encounter in a study of this nature. While most problem areas are obvious and none are unique to the kind of research advocated here, they do deserve comment. First and perhaps foremost is the nature of the data sets required for a successful analysis. In order to discern Spatial and/or artifactual patterning in a disposal area that can be demonstrated to be "redundant and unambiguous" requires quality data and largely intact sites, or at least 109 sites with a minimal amount of disturbance. Such disturb- ance - whether modern or aboriginal - may obscure the origi- nal patterning. At the same time, however, the lack of a virgin cemetery should not deter the archaeologist from the attempt. Several partial data sets could be profitably employed (see chapter 5). The multidimensional approach discussed above relies heavily on bio-anthropological analysis as an independent test of hypothesized UDG constitution. Obviously, poorly preserved human remains do not facilitate this operation. Another problem involves temporal control. In common with settlement system studies, any meaningful analysis of kin group structure necessitates the demonstration of con- temporanaity or near contemporanaity of the mortuary and habitation sites involved. Also, it is conceivable that the use of a disposal area over a prolonged period of time might complicate spatial patterning. Good chronological control can mitigate against such variability, and whether this can be accomplished will depend on the nature of the particular cases. The relative lack of fine-grained temporal control in archaeology is perhaps the single biggest problem faced by the discipline in general. Finally, it has probably become clear to the reader that this multidimensional research program as outlined above, requires research on a fairly large scale. For exam- ple, reference has been made to the testing of hypotheses with multiple data sets from many sites, which would require 110 a good deal of data. It is the present opinion that any large-scale, regionally focused, long term research program can recover the kinds of data necessary for the analysis of the horizontal dimension. This writer is not the first to advocate such long term regional research, and large-scale survey and excavation is a prerequisite to any study seeking to examine inter- and intra-Site Spatial structure. The point is, the kinds of data necessary for horizontal dimen- sion analysis can be recovered from existing programs if only the investigators seek it. Just as importantly, exis- ting data sets can be employed in pattern recognition, both as bases for hypothesis formation or as supplemental data for the testing of hypotheses generated from "new" data. Conclusion One final point will be made before moving on. In the present opinion, the elucidation of horizontal groups from archaeological data is as much a conceptual or perceptional problem as methodological. A large part of the problem seems to involve l) the previously discussed preoccupation with the vertical status dimension, and 2) the tendency for many archaeologists to conceive of their subject matter in terms of asocial adaptational systems. Concern with the vertical dimension directs attention toward the positions of individuals in a system. A systems approach directs atten- tion to the entire system, or to abstracted subsystems, i.e., economic, political, etc. It would appear that many lll archaeologists have lost sight of the fact that all human beings configure themselves into socially significant groups beyond the level of the individual and below the level of the system as a whole. In large measure, it is these social groups - not subsystems or the system as a whole - that com- prise the functioning units of society; human adaptation may be systemic in nature, but systems pgg gg_do not adapt, people do. The ethnographic research conducted in this study has demonstrated that social groups can leave consid- erable residue in the archaeological record: in a mortuary site, in a settlement pattern, and in general artifactual patterning. A habitation site or mortuary site was not used by a random assortment of peOple, nor is a settlement system simply an ordered assortment of various sites relative to the environment. A settlement system is also an ordered assortment of social groups across a landscape. Socially- defined groups of peOple occupying these sites are what adapt to their environment, not the sites or artifacts them- selves. In this writer's opinion, then, archaeologists should reconsider their data in terms of such socially Sig- nificant groups. This study by no means advocates the abandonment of the systems approach, only that a "social perspective", as it were, be integrated into it. The pers- pective advocated here may not yield all the answers, but it can surely promote the asking of some interesting questions. CHAPTER 3 AN ETHNOGRAPHIC TEST OF "HYPOTHESIS EIGHT" Introduction It is the purpose of this chapter to conduct an exam- ination and ethnographic test of an explanatory construct referred to as "Hypothesis 8". This hypothesis was offered by Saxe (1970), and attempts to relate the use of formal, exclusive disposal areas to a certain corporate group struc- ture in a society and to ecosystemic variables. Goldstein (1976, 1980) has further examined the construct and has offered modifications to it. The present study will reexa- mine the hypothesis in light of the ethnographic data base from the HRAF. Saxe's Hypothesis 8 is phrased as follows: To the degree that corporate group rights to use and/or control crucial but restricted resources are attained and/or legitimized by means of lineal descent from the dead (i.e., lineal ties to the ancestors), such groups will maintain formal dis- posal areas for the exclusive disposal of their dead, and conversely (Saxe 1970:119). This construct is essentially a mechanical model positing that when four dependent variables coincide, they will pro- duce the independent variable. That is, when D1) corporate groups are present in a society that D2) maintain lineal ties to the ancestors, for the purpose of D3) legitimizing 112 113 group rights to use/control D4) crucial but restricted resources, these groups will Il) maintain formal disposal areas for the exclusive disposal of their dead. The hypo- thesis is phrased as a law-like generalization; while it is expectable that Saxe, if queried on the subject, would allow that exceptions to the rule are inevitable, the hypothesis is nonetheless phrased as though it will obtain in every case where the dependent variables form a set. Of equal im- portance is the converse proposition that in societies where the dependent variables do not obtain, there will be no dis- posal areas. Saxe tested this model against a sample of three ethnographic societies, and found general confirmation (see also Saxe and Gall 1977). Goldstein's (1976) test of the hypothesis included 30 societies. She developed a modified and amended version of the hypothesis, thus: A. To the degree that corporate group rights to use and/or control crucial but restricted resource(s) are attained and/or legitimized by lineal descent from the dead (i.e., lineal ties to the ancestors), such groups will, by the popular religion and its ritualization, regularly reaffirm the lineal group and its rights. One means of ritualizing is by the maintenanEE_Of a permanent, specialized, bounded disposal area for the exclusive dispo- sal of their dead. I B. If a permanent, specialized bounded area for the exclusive disposal of a group's dead ex- ists, then it is likely that this represents a corporate group who has rights over the use and/or control of crucial but restricted resource(s). This corporate control is most likely attained and/or legitimized by means of lineal descent from the dead, either in terms of an actual lineage or in the form of a strong, established tradition of the critical 114 resource passing from parent to offspring. C. The more structured and formal the disposal area, the less number of alternate explana- tions of social organization apply, and con- versely (Goldstein 1976:61). The thrust of Goldstein's reformulation was to render Saxe's mechanical model into a statistical one, i.e., Gold- stein states that when the four dependent variables form a set, people will engage in deliberate, usually ritualized behaviors that are intended to reaffirm or reify a corporate group's rights to crucial but restricted resources. One:pos-w sible manifestation of this behavior involves symbolization in the mortuary domain through the use of formal, exclusive disposal areas wherein are interred deceased corporate group members. This practice would clearly reaffirm the group's rights, since the ancestors are buried in the land and, therefore, there can be no question as to control or owner- ship. Given the necessary conditions, then, there is a pos- tulated strong probability that a disposal area will result, though this is not inevitable (Goldstein 1976:58-62). Goldstein's version of Hypothesis 8 is an improvement over Saxe's original construct. Still, a very fundamental problem remains, yig., "item C" above: "The more structured and formal the disposal area, the less number of alternate explanations of social organization apply, and conversely" (Goldstein 1976:61). The problem is simple, that this state- ment, like Saxe's original hypothesis, implies quite clearly that the only reason a society will use formal disposal areas is in order to legitimize corporate group rights over crucial 115 but restricted resources. At this juncture, it is important to examine the approach taken by Saxe and Goldstein in their formulation and testing of Hypothesis 8, since the shortcomings of the hypothesis (see below) are a direct result of the approach. Also, their method differs from that used here. Saxe states that the hypothesis was stimulated in large measure by Meggitt's work with the Mae Enga, a highland New Guinea society. The Mae Enga are a society in which arable land is a crucial but restricted (by its scarcity) resource; Meggitt noted that the Mae Enga had strong corporate groups (local- ized clans) that utilized the principle of lineal descent from the ancestors to legitimize group rights to the land (Saxe 1970:119-121). Saxe further discovered that the Mae Enga maintained cemeteries for male members of the corporate groups. Thus, Saxe's Hypothesis 8 is essentially an exten— sion of the Mae Enga case, whereby he takes the Mae Enga situation one step further and rephrases it as a general construct, presumably applicable to any society (Saxe 1970: 121). He then went on to examine three additional societies as tests of the hypothesis; importantly, his perspective was to determine first whether the societies possessed the four dependent variables, and secondly whether or not they then utilized formal, exclusive disposal areas for members of corporate groups. Goldstein's test included 30 societies, and her approach was identical to Saxe's. In short, they both posed the question: "If a society has corporate groups 116 that attain and/or legitimize use/ownership/control of cru- cial but restricted resources, then how do they dispose of their dead"? Both concluded that peOple in this situation usually used formal and exclusive disposal areas, ostensibly in order to reinforce group control of the resource. Two problems present themselves. The first is straightforward, that Saxe developed a general behavioral model with expected material correlates that was based on only one case study; a more ambitious use of the ethnogra- phic record would have been desirable. The second problem is far more important however, and involves the analytical approach adOptedkanaxe and Goldstein. Their method is essentially unidirectional, whereby societies exhibiting the four dependent variables were sought, and then examined to determine the nature of mortuary practices and their presum- ed relationships with the dependent variables. That presum- ed relationship was stated to be one of cause and effect. What these investigators did ESE undertake was an examina- tion of whether other kinds 9i factors might also produce formal, exclusive disposal areas. In particular,oh mono moat gamenom: cmnu mood u oocflauooc: oocwsuoumoss ousmmoud monsommu n m Am.zv sobH usombo ouommoud oOHSOmoH u I Am.mv ammo: usououd musmmoum oOHSOmmH n + Am.Dv “can mono moat AHHEMM u m Am.DV mtoe cuouuom o:\oouoHoHIomson oEOm\3 mono omen u k Am.:v sac: moon mmflc oooncoc u z AI.DV ocmaouH comm down was u D 362 lm.sv masons A+.Dv Hoomcoz Am.Dv oHoq 1+.sc mcmzum 1+.sc cones AI.IDV cocoa: 1+.oc cfinmaaom Am.ac mouon Au.ov >«e AI.DV comoo AI.DV cameos Am.Dv fimcoaama AI.OV cocoa .III AI.DV “worm AI.DV mfloovouH Am.Dv Hosanna A+.«DV mmcm on: AI.:1 camamH 1+.sc ousflmow l-.sc crosses umummm 12.91 omcoz 1+.sv «Isa «Isa >HosH duo QHU can» maso haso \Emm moan Hmuomuw ousmmoum muommwum mum duo .m bond dmwo .v oz .m asosxca .m :uoomuom: .H m mflmonu0d>m no nude owsmmumocsum :4 .ma dance 122 AI.zv osom AI.mv mxuooz Am.zv mafia AI.zV oocSmm Au.zv Hnoq AI.zV nmflamm .mm AI.OV msoflwo Illl. “Lovmflam AI.SV smegma AI.SV maooo maaom Au.ov cuss AI.DV osxo» AI.DV ocmose 1+.ov mommHmH A-.zv somewas AI.DV mocmsmame Hamnmumz lu.ov passage A-.o1 msomma .m Am.sv xsue mumum\mmmflso dflsmuoCSO aflomnmcso H0\oco mono oz .m xmoz .n duo msofluo> .m AU.ucOOV ma OHQMB 123 generally it would appear that many societies maintain for- mal, exclusive disposal areas for their dead, but do not maintain corporate group rights to resources, do not endure "pressure" on their resources, and/or do not restrict their disposal areas to members of any particular economic corpor- ate group. By way of illustration, consider Table 15. On the left, Column 1 has been labeled "Perfect". While two cases appear here, only one, Puka Puka, actually conforms to all expectations of Hypothesis 8. The Mae Enga have been inclu- ded here though in reality they are "not quite" perfect, since not all members (i.e., females) of the corporate groups are interred in the disposal areas, nor are their disposal areas particularly formal or exclusive. Column 2, "Unknown Pressure Only", contains three cases in which all variables obtain with the exception of resource pressure, the nature of which could not be discerned from the literature. Thus, these three cases are potentially "perfect" also. Column 3, "No Pressure Only", shows four societies in which all varia- bles except restricted resources obtain; the literature on these cases indicated that resource pressure or scarcity is not a problem for these peOple. Note also that one case (Mbundu) does not dispose of all economic corporate group members in an exclusive facility. The next three columns contain societies in which dis- posal area populations do not reflect economic corporate group membership. These cases exhibit certain discrepancies 124 in the nature of ownership/control of resources and corpor— ate group disposal from the expectations of Hypothesis 8. Column 4, "DiSposal Area is Greater than Corporate Group", contains 11 cases in which members of more than one corpor- ate group are interred in the same disposal facility. The Tanala, for example, maintain patrilineal corporate group rights to ownership of land. Upon death, individuals are interred in clan tombs; members of several corporate patri- lineages are thus buried in a common - not exclusive - facility. The Iroquois are very Similar: matrilineages controlled rights of usufruct over land/resources, yet bur- ial occurred in clan cemeteries. It should also be mention- ed that in many cases resource pressure is not a problem, and one society (Thonga) does not include all corporate group members in the same facility. Column 5 is titled "Corporate Group plus Family/ Individual". This group of nine societies shows variability in the nature of land and resource ownership/control and tenure: some is owned by recognized economic corporate groups, some by families or individuals. On New Ireland, for example, clans own non-economic land, while families own their own agricultural plots. Inheritance of land is within the family, and there is no pressure on land or resources. Similarly, the Khasi have much clan-owned land, containing ancestor stones and cemeteries where clan members are buried. However, much productive land is owned by individuals and is open to buying and selling; land is not a restricted 125 resource. The six societies in Column 6 possess "Various Corpor- ate Groups and/or Ownership by Chiefs/State". These socie- ties show complex land tenure systems. In the Kurd case, various groups and individuals own land, including the Iraqi state, common peOple, and various Chieftains, who desire (or require) control of land as a source of personal wealth and hence personal power. Land is variously bought, sold, rent- ed. Many common people successfully usurp unused state land simply by settling on it for an extended period of time. A somewhat similar case occurs on the Marshall Islands, where certain chiefs own all the land. As often occurs in complex Chiefdoms, land equals power. Chiefs designate their own heirs, typically someone within their own rank. Usufructs are granted to commoners, and the usufruct is transmitted within families, though the owning chief reserves the right to revoke it at any time and for any reason. The users must pay tribute to the owning chief and to his administrators. The situation on Truk is a little different. Here, owner- ship is variable and can be Situational: lineages, extended families, and individuals can own various things, including land and resources. For example, a lineage may own a part- icular parcel of land, but an individual can own the trees or other resources on it. The final two columns, "Weak Ownership" and "No Owner- ship", are straightforward. In the former, most societies "own" loose, informal, usually "traditional" usufructary 126 rights to territories, but do not really claim to own the land, or to transmit/inherit it. In the latter column, rec- Ognized ownership of land or resources does not exist. As mentioned, these data can be differently interpret- ed. A "conservative" level of acceptance of Hypothesis 8 would allow that only one and potentially three cases actu- ally conform to the predictions of the hypothesis. A more "liberal" perspective on the data might indicate that about half of the cases support the construct. In Column 4 ("Dis- posal Area is Greater than Corporate Group"), for example, the disposal facilities contain two or more corporate no political importance. This fact should be reflected in the mortuary domain. First, it is expectable that the sym- boling of UDG affiliation should be rare, particularly 171 regarding artifactual representations. Secondly, UDG dis- persal should be common. In the present formulation, this is the critical factor distinguishing this form from all others. Such dispersal should be indicated in a number of ways: if disposal areas are used, they should be multiple or divided, reflecting the presence of multiple UDGs in any one settlement or area. If disposal areas are not used, then any number of alternative modes could be utilized. One would not expect single, discrete disposal areas either isolated or near settlements. Other spatial patterning can indicate a "Complex II" system. Major settlements divided into wards would suggest a highly structured utilization of space and considerable control over site planning by some individual or group; such wards could represent UDG residential areas or could be non- kin based. It is problematical whether the phenomenon of subdividing the domain into districts could be discerned archaeologically. However, even if actual districts (the boundaries of which could be wholly arbitrary) are not dis- cerned, evidence of this high level of systemic control could be. For example, the settlement pattern could indi- cate control of trade, transportation, and/or communication routes (9i. Steponaitis 1978), or may otherwise be structur- ed so as to enhance system regulation. Deviation from Peebles and Kus' ideal hierarchical pattern might be expec- ted in very complex forms, i.e., from the expectation that "minor centers" will cluster around "major centers", and 172 that smaller villages and hamlets will cluster around the former. Chiefly systemic control could result in settle- ments in "unusual" locations, such as settlements in areas of poor agricultural potential or otherwise in less than ideal physical and environmental conditions. Such a situa- tion would indicate the importance of political over subsis- tence factors in determining settlement structure. Pushing this argument to its logical extreme, it can be further suggested that atypical settlement locations could also in- volve atypical settlements, e.g. military outposts, garri- sons, or staging areas; isolated settlements in "foreign" territory representing colonization efforts. Regarding the structure of the chiefly domain in general, it can be posi- ted - for systems at the theoretical apex of ranking at least - that a considerable degree of intra-systemic homo- ‘geneity would be expected. That is, as chiefly control in- creases and UDG power declines, one would expect greater systemic integration and less regional variability to be manisfested; very complex chiefdoms tend to be defined and politically and geographically circumscribed into a known territory. At this level, the archaeologist would expect greater variability between two contemporaneous and adjacent chiefdoms than within either one. "Complex i: Chiefdoms The general category "Complex I" has been defined using a rather strict interpretation of Kirchoff's (op cit.) 173 conical clan concept. These systems differ from the former in that the territory of the chiefdom is divided into kin- based sub—territories, the occupants of which are predomin- antly concentrated UDGS, typically clans. Clans are ranked one to another; lineages within clans can be similarly ranked, and may also have their own lineage territories. This is a markedly different structural form from that labeled "Complex II". Here, UDGs enjoy a goodly measure of political power, which is a direct reflection of (probably a direct result of) the relative lack of institutionalized systemic control of the domain by the paramount or his group. "Relative" lack of control means relative to "Com- plex II" systems. To the extent that the paramount is clearly considered to hold the highest socio-political status position in the society, and the various other UDG Chieftains are compara- tively lower in status, then this fact could be reflected in vertical symboling. Peebles and Kus' two "dimensions" of status may not be so clear cut; rather, something approach- ing Brown's (1981) gradations of relative status would prob- ably obtain. Relative vertical status should have a spatial reflection, i.e., status interments should occur within the chieftains' settlements in each UDG territory. A generally hierarchical settlement pattern and site types would prob- ably obtain, reflecting in this case not so much "homeosta- tic" system control but the spatial representation of the social status hierarchy. Local subsistence autonomy would 174 also be expected, and each settlement or minimally each UDG territory would be subsistence self—sufficient., The extent to which monumental architecture, craft Specialization, and organized labor would be present is problematical, but should all increase as overall complexity increases. Like- wise problematical is evidence of chiefly control of the system; such should also increase with complexity but should not approach that seen in "Complex II" systems with- out a radical system transformation. As Sahlins (1968) remarked, conical clan chiefdoms can be discerned on the ground. UDG territories should be in":~ ferable from settlement patterning and from relationships between sites and the environment. If a hierarchical pat- tern obtains, then the settlements of the local Chieftains should be roughly comparable in size and configuration, with only one such in each territory. Minimally, some kind of Spacing between these sites should occur, though such spac- ing need not be regular; environmental variables must be considered. Smaller villages and hamlets should associate with these chiefly sites. In the mortuary domain, clear evidence of UDG concen- tration should be found. If disposal areas are used, then they should be single, with no internal divisions or dif- ferential artifactual patterning suggesting multiple UDGS. Logically, either settlement-specific disposal areas could be used, or disposal areas used by entire UDGs. If the latter obtains, the facility should occur within the UDG 175 territory. It could be added that in either case, the UDG chief and his immediate following can be expected to be treated differently, and perhaps disposed in a separate location, which most likely would be his own settlement. "Basic" ranking This category presents sizeable problems to which there may be no solutions. Variability in such minimal ranked systems can be marked, principally because 1) ascrib- ed status is not institutionalized, and 2) such systems have usually been observed to exist in a state of flux, i.e., where structurally equivalent chiefs compete for status, and where through time any chief's relative status can be expec- ted to rise or fall with shifts in the political winds. Further contributing to the lack of symbolic consistency, in the mortuary domain is the fact that, in the case of local- ized UDGs, variability could involve programming specific and unique to a particular UDG. It was suggested above that symboling of a chief's vertical status might depend upon the nature of his career, and his relative status at the time of his death. This observation follows Brown (1981) who states that such chiefs will receive funerary treatments reflecting their age, sex, and personal achievement. Further, it has been suggested that the symboling of a chief's heir after a premature death may be expected to reflect the chief's relative status at that time. 176 The basic problem is straightforward: if a petty chieftain's ascribed status is symboled on the basis of his age, sex, and personal achievements, then how can this treatment be distinguished from the symboling of simply achieved status? No solutions can be offered at this point. Further complicating matters is the fact that no necessary consistency, can be expected in the symboling of other mem- bers of the chiefly group; one clearly cannot anticipate the presence of two "dimensions" of status as are found in more complex systems. In the horizontal dimension, UDG localization might be demonstratable. An intra-UDG settlement hierarchy would not be expected, and if the condition of many chiefs but no paramount obtains, then no society-wide hierarchy would occur. Also absent would be any indication of chiefly sys- temic control. In the event that one Chieftain did become paramount over the entire society, and if this condition lasted over time, then archaeologically demonstratable patterning might emerge. Such a system would constitute a conical clan if the paramount and the lesser Chieftains were products of the same ranking group; if not, the system would represent another - albeit hypothetical - structural form of ranked society. To conclude, then, "Basic" ranking may not be consis- tently inferred from the archaeological record, given the current state of method and theory. Future research might 177 be directed toward distinguishing between mortuary status distinctions reflecting age, sex, and personal achievement that denote ascribed versus strictly achieved status. A detailed study of ethnological data on such societies could be profitable. CHAPTER 5 AN ARCHAEOLOGICAL APPLICATION Introduction It is the purpose of this chapter to apply the formal correlates of structural form in ranked societies to a body of archaeological data. With this, this study seeks to "bridge" theory and data; it should be recalled at this juncture, O'Shea's comments provided earlier, that much recent archaeological theory has not been adequately "trans- lated" into archaeological expectations, and is hence of dubious value. The following analysis will demonstrate that the theoretical postulates and formal correlates presented above can indeed be successfully used in the analysis of archaeological data. The data to be considered are various Mississippian Period manifestations in the Central Mississippi River Val- ley (see Figure 1). This data set has been chosen for a number of reasons. First is the fact that, until fairly recently, this region has been incompletely understood. That is, it has been known for some time that the Central Mississippi Valley, particularly in the northern areas, saw a substantial and impressive occupation by Mississippian 178 179 systems, but the nature and relationships between manifes- tations and phases has only recently been worked out. Phillips' (1970) ceramic studies have greatly contributed to improved spatial-temporal control, and a more recent syn- thesis of the archaeology of the region by Morse and Morse (1983) further places these data into a coherent cultural- historic framework. Significantly, the extant data - par- ticularly mortuary data - are now of sufficient quality and quantity to allow an exploratory application of the approach. Secondly, and of greater consequence, this data set will illustrate the utility of the method and demonstrate the kinds of insights that can be gained from the study of structural form; Mississippian in the Eastern United States constitutes a nearly ideal test case. To explain, climax Mississippian throughout the East has been recognized for some time as a cultural entity with a considerable degree of homogeneity or similarity over space. The term "Mississip- pian Pattern", to employ older terminology, was used in reference to this areal homogeneity, reflecting similarities in material culture, settlement types and patterns, subsis- tence economy, and esoteric iconography. Simultaneously, however, archaeologists have also recognized that a tremen- dous range of regional or local variability exists in Mississippian systems, despite the surficial homogeneity. It is this regional variability that has proven difficult to account for or explain. Regional phases have been worked 180 out, and it is often assumed that these phases represent independent chiefdoms, but beyond that, little has been sug- gested to explain within-phase structure; clearly it is easier to demonstrate regional variability than to explain it. The following discussion will attempt to show how this approach can be used to address and explain this variability through an analysis of structural form in Mississippian sys- tems. This examination of structure is based on the derived formal correlates of chiefdoms, which are in turn based pri- marily on the ability to infer horizontal group structure and configuration from mortuary patterning. Therefore, this chapter and application of method to data is intended to address the formal variability extant in one region of Mississippian development, and to offer hypotheses to ex- plain this variability in terms of differences in structural form. Also, and just as importantly, this chapter will illustrate the method employed in the study of structural form, because the approach is intended to be applicable gen- erally, and is not specific to any particular data set. It should be pointed out that the discussion that follows will be, to a degree, selective in the phases chosen for examination. All known or suggested phases in this area cannot be considered; such would be well beyond the present scope, and in any case, the data - particularly mortuary data - are insufficient for most known phases. Analysis will focus, then, on Mississippian systems for which 181 acceptable mortuary and settlement data exist. These phases tend to be the better known in the region. A Mississippian Test Case Early Mississippian Period, c.a. AD 700-1000 The earliest Mississippian manifestations in the Cen- tral Mississippi Valley are among the poorest known with regard to mortuary practices. Because of this, no firm statements regarding socio-political structural form can be offered for this period. A discussion of these data is in order, however, to provide background for the succeeding phases. By AD 700, socio-cultural systems that have been iden- tified as Mississippian were evolving in the Mississippi River Valley, particularly in the American Bottoms area in the vicinity of the Cahokia site, and further to the south in the Cairo Lowland region. The emergence of Mississippian is not yet completely understood, though it is clear that it was largely an in siiu evolution of resident Late WOodland societies into a more complex systemic form. As this proc- ess continued, various Mississippian groups seem to have moved into adjacent regions, intruding upon other less com- plex societies. The evolution of Mississippian systems probably cor- responds to the evolution of ascriptive ranking and chief- dom—type socio-political organization, or at least the 182 development of what has been referred to as complex ranking. It has been argued by some archaeologists that status ascription was present in earlier systems (e.g. Hopewell), but it is obvious that, if such does in fact prove to be the case, these earlier systems were of a very different struc- tural form than most Mississippian systems, and never devel- oped to comparable complexity. In the Cairo Lowlands, emergent Mississippian is re- ferred to as the Hoecake phase (Phillips 1970:902-903), which temporally bridges the transition from late Baytown Period manifestations into early Mississippian (Morse and Morse 1983:190). The Hoecake site (see Figure l) is the type site for the phase, containing anywhere from 31 to 54 mounds and covering some 80 ha (Morse and Morse 1983:215). The site is clearly multi-component and was occupied/util- ized over a considerable span of time, which greatly hampers interpretation of individual components. Morse and Morse interpret the early Mississippian component at Hoecake as a dispersed community made up of mounds and residential areas (ibid.). The relationship between mounds and residences is not known; there is, however, some evidence that the site layout was planned, minimally with regard to certain mound alignments, which foreshadows later planned sites in the Cairo Lowland. Very little of this large site has been excavated, so the extent and scope of community planning remains uncertain. Mortuary remains were excavated at Hoecake (J. 183 I'M." l ’4 "I...” 2 .‘flu a .‘ql 5‘4 \ ”an...“ .‘.f BIG LAKE PHASE 1 \ 'u“ N“ “nu tnu-I\\ *‘7 \ h .I u hal'un'. .- “I ‘lunnou u... "H "In... TENNESSEE MISSISSIPPI . .‘o. II'O" ’. Wu“... ‘2 Q. qp ’ pafli’l‘ ‘ 3.!“ "0., ’{ F0 . e4” .5: §° s fl E X \ SITES: 1 - Hoecake, 2 - Rich Woods, 3 - Kersey, 4 - Zebree Figure 1. Early Mississippian Period phases and sites discussed in text. (Adapted from Morse and Morse 1983:Figure 10.1). 184 Williams 1974) but they were few and difficult to interpret. At least one burial, with a c0pper artifact, suggests some degree of status differentiation. Burial occurred in mounds. Fourteen individuals were discovered in three log- lined tombs beneath one excavated mound; no artifacts were associated'(ibid,). Other sites that appear to be similar to Hoecake are known. One is the Rich Woods site, located 65 km to the west. This site reportedly contains 33 mounds (Morse and Morse 1983:215) in a dispersed pattern like Hoecake. Regarding Mississippian development in the Cairo Low— land, similarities in various artifact types suggest ties with Cahokia. The overall settlement structure of the Hoecake site (except for the mounds) resembles that of the Range site in the American Bottoms (Morse and Morse 1983: 192). While the precise relationships between the early Mississippian Hoecake phase and American Bottoms Fairmount phase are not yet known, it seems clear that the two re- gions were in some process of parallel development. Further to the south in the Central Mississippi Val- ley, the Big Lake phase has been defined by Morse and Morse (1983:217ff), and is largely based on their excavations at the Zebree site. The Big Lake phase is interpreted to be the earliest Mississippian system in this area, and is the probable result of a direct intrusion of Mississippian people into the Baytown Tradition (Morse and Morse 1983: 233). This phase appears to be rather different than the 185 Hoecake phase to the north, and there are no multi-mound sites comparable to Hoecake and Rich Woods. Zebree was a planned village covering about one ha, and included a cen- tral cypress post pit; apparent residential sub-areas were defined, but structures were not oriented into rows as is common in later periods. A ditch surrounded much of the site, but its function could not be positively ascertained, and no evidence of a palisade was found (Morse and Morse 1983:298-300). It is not clear if this feature was intend- ed for defense or was simply used for borrow. While Zebree seems to have been a fairly nucleated village, related dis- persed hamlets also occurred nearby. Few burials were found at the site, but those that were deserve mention, since minimally they may indicate differences in mortuary programming. Most interments were house-related, occurring in the general midden near residen- tial structures. One grave with multiple interments was found: four male and four female primary burials were dis- covered in this grave in one residential sub-area (Morse and Morse 1983:231-232). This area was found to contain the highest percentages of mammal bone on the site and most of the exotic artifacts, and Morse and Morse (ibig.) tenta- tively suggest that this may have been the ranking group in the village. In any case, some distinction in mortuary treatment is indicated. It is further interesting to note that 26 locations of isolated human bones were found in the village, which also suggests differential treatment of 186 certain dead. It has been suggested (ibid,) that Charnel- ling may have been conducted, particularly by the status group, though the primary interments in the grave do not support this. Charnelling does seem to be the case at a related site. Marshall (1965) describes 10 bundle burials within a circular structure at the Kersey site in Missouri. Morse and Morse (1983:232) note that the size of this structure is close to that of the Zebree grave. However, only at the Kersey site is Charnelling definitely indicated. While a more precise estimation of mortuary practices of the Big Lake phase must await additional data,cxuacan offer the gen- eral observation that in this early Mississippian phase, they seem to be rather varied. Thqueneral lack of data precludes any meaningful statements regarding structural form. Middle Mississippian Period, c.a. AD 1000-1350 By AD 1000, fully developed Mississippian systems were occupying the Central Mississippi Valley. The end of this period represents a climax of sorts for Mississippian in the Cairo Lowland; after AD 1350, this area seems to have been abandoned after a long and vigorous occupation (S. Williams 1977). Further to the south, abandonment did not occur, but rather dramatic system transformations took place. Impor- tantly for the present endeavor, the mature Mississippian systems of this period lend themselves to an examination of 187 structural variability in chiefdom societies. Beginning in the Cairo Lowland, this period has been referred to as the Cairo Lowland phase (Phillips 1970:925- 926). It has for long been recognized that the Cairo Low- land phase manifestations parallel the Stirling and Moore- head phase developments in the American Bottoms but as yet, the nature of any relationship between these regions is not known, though trade certainly occured. The Cairo Lowland phase (see Figure 2) is marked by an increase in population density in the region and an increase in the number of sites attributable to the phase; Morse and Morse (1983:237) describe this area as the population and cultural center of the Central Mississippi Valley. As will be shown, this system clearly demonstrates greater overall socio-political complexity than systems adjacent to the south. It is this writer's contention that the Cairo Low- land phase represents a distinct socio-political system, probably a single chiefdom, structurally equivalent to the Cahokia and Kincaid systems to the north and northeast. The Powers phase to the east would appear to be a politically autonomous system (Price 1978, Price and Griffin 1979). It might be mentioned that other investigators would further divide the Cairo Lowland phased Chapman (1980), for ex- ample, considers the major centers and associated sites in the area to be phases in their own right. The present study prefers to view these sites as elements in a single hier- archical socio-political system. 188 “mu-u I «M u "I CAN“) LOW LAND PHASE f. 111 .52?" 1 “I .... f}? .' MISSOURI ‘3' , - .2.» KENTUCKY v‘ ; ... s "- - ' T NN E Amnusms ; “3‘ E 5mm . .. w 3"." a" 3. 5 f‘ 3 ’3 u; 6" '9'. ' E .f‘ \ i 0' ' ..c" 3 . .3 33 3155f". .\ 3 \ . ‘ 711.99“ ‘s‘ "'"Nllu ,' lit‘ 3 .' v .- 2 3' ‘ : ~s ‘ r V“); :1 30“ us“; “9“ , 4N" CHERRY M)“ VALLEY f, 1 PHASE '5“ f I. \N . ° : §§ 5 t . I . I‘ I, .c' ‘ I p . J 3 Tanmmur ‘ 7 ’ 3 uwmamw: 'u‘ *0 0 x l 1 .10 a; "' s .- .... i a 2 - Towosahgy, 3 - Crosno, 4 - 6 - Matthews, 7 — Lilbourn, 8 11 - Old Town Ridge, 12 - l - McLeod Bluff, Sandy Woods, 5 - Sikeston, 9 - Zebree, 10 - Lawhorn, Langdom, webb, l3 - unnamed, l4 - Cherry Valley SITES: Figure 2. Middle Mississippian Period phases and sites discussed in text. (Adapted from Morse and Morse 1983:Figure 11.1). 189 The transition from the Hoecake phase to Cairo Lowland obviously involved many things. One important event seems to have been the abandonment of the Hoecake site and the establishment of the Towosaghy site nearby. The latter would evolve in time from a dispersed settlement (like Hoecake) into a planned, fortified town that can be justly called a civic-ceremonial center (Morse and Morse 1983:264). Towosaghy appears to be the largest and most complex site in an obviously hierarchical settlement system. Other large sites in the hierarchy include Mathews, Sikeston, Sandy Woods, Crosno, Lilbourn, and McLeod Bluff. These latter sites form a hierarchical settlement pattern generally clus- tering around Towosaghy; apparently all sites were fortified with a surrounding palisade. The land between these centers is dotted with small hamlets and farmsteads of a typical Mississippian dispersed settlement system. Evidently, the fortified centers offered a place of refuge for the dispers- ed population when needed. All of these centers display considerable evidence of pre-planning. Multiple mounds and plaza areas in regular relationships are characteristic. A controlled surface col- lection of Towosaghy revealed a non-random distribution of finely-made ceramics; these materials were found in a limi- ted area associated with structures surrounding a plaza, and may indicate the dwellings of high status peOple (Chap- man 1976:132). These major sites are all located within or adjacent to a meander belt zone and generally within areas 190 of environmental diversity (Morse and Morse 1983:265), which would facilitate a secure subsistence economy. Mortuary data from the Cairo Lowland phase can be best described as uneven. Good data sets exist from one site, but in general, mortuary practices are as yet poorly known. Fortunately, some of the known data are significant. Few burials were discovered at Towosaghy. Twelve burials were recovered representing extended, flexed, and bundled interments. Interestingly, of the twelve, eleven were adolescents and one an infant (Chapman e5 ai. 1977). None was noted to occur under a structure floor (Chapman 1976:143), and all seem to have occurred in the general midden (Chapman 35 ai. 1977). At the nearby Bryant site, a similar demographic phenomenon was noted: one mound con- tained l6 infant burials, and a cemetery area had 23 in- fants and seven adults. It seems obvious that the rela- tively limited excavations at Towosaghy and nearby sites have revealed only a small portion of the overall mortuary program. No formal cemeteries were discovered at Towo- saghy; a cemetery is known to exist immediately to the north of the site, but its relationship to Towosaghy is uncertain (ibid.). Chapman 33 ii. (ibid.) states that the presence of cemeteries at Towosaghy is expectable, and have heretofore been simply missed. The Lilbourn site has yielded the best mortuary data from a Cairo Lowland phase site. Here, four discrete cemetery areas were discovered, at least three of which are 191 clearly contemporaneous; the fourth may date somewhat earl- ier. The former three were located adjacent to and within the palisade walls, suggesting a planned arrangement (Chapman 1976:140). Only adults and subadults were inter- red in cemeteries; infants were buried under house floors (Chapman 3E ai. 1977). A few evidently high status burials were recovered at Lilbourn. Interestingly, these occurred not in a cemetery but on and under the floor of a house that had burned. Most notable was an adult male with a large, chipped stone mace lying on his chest (Chapman 1976:140-142). Three adult females were associated with him. These individuals were laid out on the burned floor of the structure. Prior to the burning, an infant had been buried below the floor. Some time subsequently, a grave was dug through the burned floor and an old adult female was buried; Chapman (ibid.) interprets this woman as a shaman, since her grave associations suggest a shaman's paraphenalia. The Lilbourn site shows differences in the mortuary programs for different individuals at the site. It would appear that most people are buried in a cemetery; infants are interred under house floors. Status personae are spa- tially segregated from the rest of the population, and are also buried under (or on) structure floors. The Matthews site is a smaller center than Lilbourn, but probably occupies a similar niche in the settlement system. Of importance is the fact that four separate 192 burial areas are known from this site (Chapman 1980:200), though the details concerning them are not clear. Also found were over 300 burials in a mound (ibid,). The rela- tionship between the mound and cemetery burials is not known. One additional site has yielded mortuary data of rea- sonable quality. At the Hearnes site (Klippel 1969), two mounds were excavated and over 70 burials recovered, most of which were secondary bundles. At the base of one mound, a circular post mold pattern was found, and Morse and Morse (1983:263) speculate that this may have been a charnel structure. Mortuary patterning at the Hearnes site does not resemble that at other Cairo Lowland phase locations. Importantly, Hearnes is probably earlier in time than the major centers (ibid.), and thus could reflect changing mor- tuary practices; the Hearnes situation more resembles Hoe- cake mound burial programs than the later forms. Turning to the region to the south of the Cairo Low- lands, a rather different system can be observed. The Middle Period Mississippian manifestations in this area are referred to as the Cherry Valley phase (Phillips 1970:929- 930), which succeeds the Big Lake phase. Settlements and the culture in general in the Cherry Valley phase do not seem to reflect the degree of development seen in the Cairo Lowland, though this could partly be due to the fact that Cherry Valley is less well known (9i. ibig.). A hierarchi- Cal settlement pattern is only suggested, and only for the 193 latter portions of the phase at that. The Morses' excava- tions at the Zebree site revealed a small hamlet-type site, in contrast to the earlier larger village. They suggest (1983:253) that Zebree may be part of a hierarchical system, as a hamlet associated with the Langdom mound and village, the Lawhorn small village, and the Old Town Ridge large village. The latter appears from aerial photographs to be a rectangular fortified town with houses organized into rows. Similar patterns may be represented elsewhere (Morse and Morse 1983:250). The situation at the beginning of the phase is dif- ferent. The Cherry Valley site, type site for the phase, is a mortuary site with no associated village. Cherry Val- ley was excavated by Perino (1967), and consisted of five mounds. Three of these were salvaged after considerable previous pothunting. Mound 2 was flanked on the north, west and south by other mounds, and is thus considered the central mound of the group; it measured about 17 feet by 68 feet. An impressive structure was found at the base: it was round and about.33 feet in diameter, with a 20 foot long entryway extending to the east (Perino 1967). Large and numerous post molds indicate that the structure, inclu- ding the entryway, had been roofed over. A rectangular fire basin was found in the center. This original structure had burned; the area was then covered with a layer of clean soil and the structure rebuilt, but without the long entry- way (ibid.). 194 Burials were found arranged around the wall posts, and most were bundles of otherwise fragmentary. Numerous other burials were found in the mound cap, indicating con- tinued use after the structure was finally capped. Mound 1 occurred immediately to the south of Mound 2. Its dimensions were estimated at 13 feet by 60 feet. Two distinct stages of construction are evident. Three large post molds oriented east-west were found at the base (appar- ently not part of a structure), as were a large number of bundle burials with few associated artifacts. Burials con- tinued to occur up into the secondary cap (ibidf). Mound 3, 13 feet by 65 feet, was adjacent and west of Mound 2. On the mound floor, three large post molds similar to those in Mound l were excavated, as was a small sub-floor burial pit that had been dug through an edge of a small cir- cular post pattern, evidently a small structure. In the pit were found five bundled and one extended child burials. Twenty-eight more bundles occurred on the ground surface. Like Mound 1, primary and secondary construction stages were noted, with burials occurring throughout. On the top of the primary mound a large basin interpreted as a possible crem- atorium was found (ibidf). All told, some 467 burials were recovered from the three excavated mounds. In addition, a local pothunter told Perino that he estimated that at least 100 more burials had been taken out of the upper levels of Mound 1 alone, prior to Perino's arrival at the site (ibid.). While the exact 195 number of individuals at Cherry Valley will never be known, there were obviously a lot of people buried here. The nature of vertical status at Cherry Valley is dif— ficult to assess from Perino's report, since every burial is not discussed, and bone preservation was extremely poor. The most common artifacts were ceramics, and exotic arti- facts of any kind were not abundant (ibid3). Morse and Morse (1983:242-243) interpret the large structure under Mound 2 as a charnel house, which seems a reasonable conclusion. Problematic are the three east-west post molds at bases of Mounds l and 3 (also, the charnel structure incorporated three similar posts in its construc- tion), as is the small circular post pattern under Mound 3. It is tempting to consider the charnel house in the central mound as the source for the hundreds of bundle burials in the adjacent mounds, and in the charnel house itself. If such is the case, however, then Mound 2 continued to be used for burial even after the charnel house was covered over. Cherry Valley represents the best known site of the phase. Other mound sites are known, however, but are poorly reported. One group of five, known as the Webb Mounds, was reported by Thomas (1894:200-203); preservation was poor, few burials were found, but pottery was abundant. The abun- dance of pottery and dearth of skeletal remains suggests total or nearly total decomposition, such as was sometimes observed at Cherry Valley. Morse and Morse (1983:246) men- tion an unnamed site containing five mounds, one of which 196 was salvaged. This mound covered a circular structure with central hearths and associated bundle burials. Burials also occurred in the mound cap. Charnelling is the inferred function of this structure (ibid.), and parallels with Cherry Valley are obvious. Habitation sites contemporaneous with the Cherry Val- ley site are apparently not known, and none was found any- where near the Cherry Valley mounds. It is somewhat para- doxical that this phase is known from mortuary sites only, or at least the earlier portion of it. As mentioned, the period after Cherry Valley is better known (though not in terms of mortuary data), and would seem to represent an evolution from whatever kind of system that is indicated by the Cherry Valley and related sites to a hierarchical pat- tern reminiscent of that of the Cairo Lowland (Morse and Morse 1983:247). Site sizes increased through time, and structures became oriented in rows; a consolidation of sorts was occurring, with the large civic-ceremonial center and associated dispersed hamlet pattern emerging (ibid.). Morse and Morse argue that this latter portion of the Cherry Valley phase may be an "...indication that chiefdoms in the Central Valley were evolving politically into more complex entities and that conflict between independent chiefdoms was intensifying" (1983:249). In any event, the isolated mor- tuary sites such as Cherry Valley seem to disappear, with burial evidently occurring near or within habitation sites. The information presented in the preceding pages can 197 be used to hypothesize the nature of socio—political struc- tural form in these societies. Mortuary data are not over- abundant, but will allow a preliminary estimation and the formulation of hypotheses. Beginning with the Cairo Lowland phase, it was noted that considerable structure exists at these sites, with con- siderable evidence of planning and structured utilization of space; ascribed status seems a certainty, though this has not yet been empirically demonstrated. Regarding Peebles and Kus' correlates, all would seem to obtain, with the assumption that two "dimensions" of vertical status probably existed. A hierarchical settlement pattern is clearly indi- cated, as is the potential for local subsistence autonomy. Organized labor, monumental architecture, and craft special- ization are present, and strong institutionalized control can be inferred from the presence of fortifications. This is, then, a bona fide ranked society. Other kinds of data have been specified above that might be sought to elaborate upon the simple ranked classi- fication. Of first importance is the fact that two Cairo Lowlands phase sites have been demonstrated to have discrete cemetery areas. At Lilbourn, these (three of four) are ex- clusive and contemporaneous; at Matthews, this cannot be demonstrated from the published report. According to ethno- graphic research, this pattern would strongly suggest the presence of multiple UDGs at these sites. In turn, multiple, dispersed UDGs can be correlated with a certain complex form 198 of ranked society, that which has been called "Complex II". In these complex chiefdoms, chiefly political control tends to override the political power of UDGS, and UDGS tend not to localize into discrete territories. If a hierarchical settlement system and military works are taken to represent such institutionalized control, as Peebles and Kus would have it, then there is a sound correlation between mortuary data and expected patterning observed in chiefdom societies in the contemporary world. In addition, Brown's postulate that status personae should be spatially segregated from the non-status masses is borne out. At Lilbourn, persons of apparent high status were not interred in cemeteries, but on house floors. It is reasonable to hypothesize that if the site could be totally excavated, additional status burials would be discovered in non-cemetery locations. This pattern should be repeated at similar, contemporaneous sites. The situation in the early Cherry Valley phase is obviously different. Here, at least three sites are known, all burial mound sites containing five mounds apiece. No village site was associated with Cherry Valley, and evident- ly not with the other two either (Morse and Morse 1983:266). Habitation sites have not been isolated or reported for this phase, which suggests that, wherever they are, they do not have associated mounds. If the interpretation of the Cherry Valley site is correct, that a central charnel facility was maintained that supplied burials for the surrounding mounds, 199 then this would constitute a situation similar to one of the ethnographically-derived programs, that of "single dis- posal area, one per UDG per society". It was pointed out that this pattern was observed to occur oniy_with single UDGs; no other societal grouping used group disposal areas spatially removed from habitation sites. The Morses have commented that The large number of burials at the Cherry Valley type site is unusual in that interment at the ceremonial center apparently was widely available. Other Cherry Valley phase mounds seem to have been more restrictive but still contain numerous bur- ials. Death is apparently this society's major excuse for ceremonial gathering (1983:246). The large number of burials at these sites supports the in- ference that entire UDGS - or at any rate a sizeable portion of the society - were interred at these facilities. Also, with this interpretation one need not assume some kind of unusual preoccupation with death, only that the society's dead are disposed in a few, rather than many, disposal areas. If this inference is correct, then the structural form indicated should approximate a "Complex I", or conical clan- type structure. Each of the three known mound sites should associate with a discrete UDG territory. There is no way to assess this possibility at present, except to note that the three sites are geographically separated by some distance (see Figure 2). It might be speculated that a conical clan-type system in the Cherry Valley phase is not unexpectable given the 200 nature of Mississippian occupation in this area. Mississip- pian seems to have been an intrusion into this area during the preceding Big Lake phase. It has been observed (e.g. Sahlins 1961) that territorial expansion often occurs along kin group lines, through descent group fission where a UDG or UDG segment buds-off and moves into new territory. While the extant data do not allow an assessment of this possibil- ity during the Big Lake phase, the configuration during Cherry Valley is consistent with this model. That is, the hypothesized conical clan-type structure could reflect colo- nization processes by Mississippian UDGS at the expense of less complex Baytown peoples. By Cherry Valley times, Miss- issippian peoples had consolidated their occupation into UDG territories and have begun processing their dead in central facilities. While these ideas are clearly speculative, they are nonetheless consistent with ethnographically known cul- tural dynamics and processes. As such, they could form the basis for further investigation. In a similar context, it has been noted that the late Cherry Valley phase (or perhaps after) witnesses a transfor- mation to a hierarchical settlement pattern more like that of the Cairo Lowland. Again, the data do not allow an eval— uation of this situation in terms of structural expectations, but obviously some kind of changes took place. As additional data - particularly mortuary data - accrue, it should be possible to determine whether the hypothesized conical clan— type system is retained, or whether the transformation 201 reflects UDG dispersal. .Should the former prove to be the case, then it can be observed that two adjacent Mississip- pian phases are organized into very different structural forms; if the latter obtains, then an evolutionary sequence from "Complex I" to "Complex II" structures is in evidence. The demonstration of either would form the basis for an in- teresting study of Mississippian adaptations, political or- ganization, and cultural dynamics. Late Mississippian Period, c.a. AD 1350-1650 At or slightly after AD 1350, a dramatic transforma- tion can be observed in the Mississippian systems of the Central Valley. Prior to this date several recognized pha- ses/systems were extant, each or most evidently representing an independent chiefdom. After AD.1350, independent chief- doms are also evident, but are of a markedly different nature. The entire Cairo Lowland region seems to have been abandoned by this date (S. Williams 1977). The area was not altogether devoid of people, but the Cairo Lowland phase chiefdom was no longer in Operation, and the major centers were no longer occupied. The same thing occurred in the Powers phase area. Morse and Morse (1983:282-283) have sug- gested that the Cairo Lowland pOpulation, or some portion of it, moved to the south and formed the Nodena phase. This is a reasonable assumption, and at any rate the Nodena and adjacent Parkin phases are both an apparent in-movement of 202 peOple from somewhere (see Figure 3). The reasons for this population shift are not certain, but various aSpects of the Nodena and Parkin systems are suggestive. Discussion will begin with the Nodena phase. Nodena sites occur in three geographical clusters of unequal size. Many sites are known; most contain mounds and fortifications and were obviously pre-planned. In contrast to previous phases (and indeed to most Mississippian systems in gener- al), dispersed hamlets or farmsteads appear to be rare, in- dicating a heretofore unheard of population nucleation into or very near the large towns (Morse and Morse 1983:280-284). Important sites include Pecan Point, which was probably the "capital" of the system (Morse and Morse 1983:285), Upper Nodena, Banks Village, Bradley, and others. Ecological factors seem to have played a role in the southward population shift that formed the Nodena phase, and hence partly structures its settlement system. The Morses explain that important differences exist in the configura- tion of productive soils within the Central Valley. All 10- cations could support a hierarchical, dispersed system, "But only the Meander Stream Surface (configuration) can accommodate the nucleated populations in fortified villages thought to be characteristic of the Late period Mississip- pian sites" (1983:283). Both Nodena and Parkin phases are situated within the Meander Stream Surface. Thus, whatever prompted the population shift and subsequent nucleated set- tlement pattern, it seems clear that these peOple selected 203 ‘- “‘1 ‘.‘ \ U \ 1 \, ... U 1!, Q‘ . n 0 .4" '7 "'u... S I hm 91' ".0: “’ mssoum ..w‘keurucxv ARKANSAS 4" “ Tennessee |\" g ,3 § 2 S s g » NODENA H; ‘ PHASE 1 I! I|“.” so I e‘” SITES: 1 - Campbell, 2 - Upper Nodena, 3 - Pecan Point, 4 - Banks Village, 5 - Parkin Late Mississippian Period phases and sites discussed in text. (Adapted from Morse and Morse 1983:Figure 12.1). Figure 3. 204 the appropriate environments in which to accomplish this. It should be added that this highly nucleated settlement pattern with a relative paucity of dispersed or "satellite" sites is not a typical Mississippian pattern. Mortuary data from the Nodena phase is abundant, but not all is well reported. One of the earliest known Nodena sites is Banks Village, which was excavated and reported by Perino (1966). This was a sizeable village, about 4 ha; no mounds were known from the site, but a plaza area was dis- covered. It is not known whether the site was palisaded. Perino excavated some 385 burials, which indicate two sep- arate burial programs. The vast majority of interments were of the house-related type. Precisely where one was buried relative to the house depended upon one's age and possibly sex: adults of both sexes were buried outside, between and generally parallel to the walls; children over about six years of age could be buried either outside with the adults or inside; children aged one to six were buried in the house, either under the central hearth, in a corner, or un- der a bench along a wall; and infants under one year were buried under or near the hearth. Age is clearly a discrim- inator of location. The apparent option for children over six to be buried in or out might suggest a sex distinction as well. The other form of programming involved group burial in pits. Four such pits were discovered, two in the central area of the village, and two in the northeastern sector. 205 All were about four feet deep and 15 feet in diameter. Bod- ies were deposited in levels, each separated by about six inches of soil (ibid.). These pits were evidently not with- in structures. The writer did not undertake a detailed examination of Perino's data for the purpose of determining the nature of vertical status. However, the importance of status ascrip- tion is probably reflected by the fact that "The finest and greatest number of artifacts were found with children be- tween two and ten years of age" (Perino l966:9). The Upper Nodena site covered about 6.2 ha; the site was palisaded and contained rectangular mounds and plazas, and one of the latter may be a chunky field (Morse and Morse 1983:287-288). Residential structures filled the rest of the area. A mortuary pattern very similar to Banks Village was discovered, in which numerous burials occurred near the structures, possibly in family plots (ibid.). One fascina- ting discovery at Upper Nodena was a mound containing 314 male and two female burials with few associated grave goods (ibig,). This writer will not hazard an explanation for this, and it clearly contrasts with the house-related pro- gram in the village area. House-related disposal programs were noted at Pecan Point by Thomas (1894:219-222), but these are not well des- cribed. Also noted at Pecan Point were burial in mounds and apparently in cemeteries as well. The House-related burials occurred in clusters, perhaps similar to Upper Nodena. It 206 can be speculated that at Pecan Point, different locations and programming probably reflect status differences. What may be an exceptiontx>the house-related norm occurs at the Campbell site, where Chapman and Anderson (1955) report two cemetery areas to the north and southeast of a central mound and plaza. One contained 18 burials, the other 15. A probable shaman was buried in the plaza. It is difficult to comment on this apparent deviation, ex- cept to note that the cemeteries do not contain numerous burials, at any rate not comparable to the hundreds found in house-related context elsewhere. Downstream from the Nodena phase along the St. Francis and Tyronza Rivers is the Parkin phase (see Figure 3). A geographical gap seems to exist between the two areas (Morse and Morse 1983:290). Parkin phase settlements are well- known as the "St. Francis—type" sites of Phillips, Ford, and Griffin (1951). These sites are typically rectangular and pre-planned and are almost invariably fortified. The area covered by most sites has been artificially built up by bringing in soil. The settlement pattern is interesting. The Parkin site, largest and type site of the phase, was situated in a strategic location, apparently controlling the mouths of the St. Francis and Tyronza Rivers. Moving up- stream, sites are evenly spaced: large sites occur about 8 km from Parkin and from each other; smaller sites occur at regular intervals between these, and almost all are palisa- ded (Morse and Morse 1983:292). Phillips remarked that 207 given this close spacing and regularity, “One would think the structures on the highest mounds could easily be seen from one settlement to another" (1970:932). Like Nodena, the settlement pattern appears to diverge from a more typi- cal Mississippian hierarchical arrangement. Further empha- sizing this assertion is the fact that absolutely no diSper- sed hamlets or farmsteads exist in the Parkin phase. The location of the Parkin site is also unusual with regard to subsistence agriculture. Parkin is a large site, covering about 7 ha, and evidently contained a large, nuc- leated population. A site catchment analysis by P. Morse (Morse and Morse 1983:293-294) showed that all sites of the phase except Parkin should have theoretically been able to produce enough food to feed their populations. Parkin, on the other hand, could raise only about half of the needed agricultural produce within a 1 km catchment. This phenom- enon suggests several things which will be considered momen- tarily. Mortuary practices of the Parkin phase are poorly known. According to Phillips, Ford and Griffin (1951:329) and Davis (1966), house—related disposal appears to be the dominant mode, including all ages and both sexes. Davis' excavations at Parkin recovered five burials, but preserva— tion was poor and relationships to structures could not be ascertained (1966:32-34). It is obvious that the Nodena and Parkin phases repre- sent complex societies; the presence of ranking seems 208 certain, though this has yet to be empirically demonstrated. It is interesting and important that all of Peebles and Kus' correlates of ranking do not obtain. Minimally, local sub- sistence autonomy is indicated for all but the Parkin site. Organized labor and monumental works are certainly in evi- dence. Chiefly control of the system is indicated by forti- fications and, at the Upper Nodena site at least, corncribs for the storage of foodstuffs. A truly hierarchical set- tlement pattern, however, is not indicated. It is suggested that these systems evidence a level of chiefly control beyond that indicated by Peebles and Kus' model. For example, the Parkin settlement pattern seems obviously designed for defense: sites are regularly spaced, and can even be seen one from another, greatly facilitating, one would think, intra-system communication. Dispersed, undefended hamlets are absent. Parkin, most certainly the chiefdom "capital", was located in an area where the inhabi- tants probably could not produce enough food to feed them- selves, indicating both non-subsistence related motivation in the selection of site location as well as some kind of economic subsidization of this site by the rest of the sys- tem (e.g. Steponaitis 1978). Parkin was probably located where it was in order to control access to and/or transpor- tation on the St. Francis and Tyronza Rivers. An identical phenomenon has been reported by Brain (1978) for Plaquemine/ Mississippian occupations in the lower Mississippi Valley. It may also be significant that Parkin's location is 209 furthest away (by river transport) from Nodena territory, with the rest of the settlements intervening. Finally, it was noted that an unoccupied area exists between the Parkin and Nodena phases. This is probably not accidental, but most likely represents a planned buffer zone between Nodena and Parkin and other phases to the south (Morse and Morse 1983). The Nodena settlement pattern, from what is known, does not seem to be nearly so regular, but is still obvious- ly defensive in nature. It is interesting to note that if Morse's three geographical site clusters are accurate and not the result of incomplete sampling, then the number of sites in each cluster decreases as one moves south, closer to foreign (e.g. Parkin and Walls phases) territory. This pattern may reflect Nodena expansion to the south. In this regard, Morse and Morse (1983:284) state that Nodena seems to be the dominant system in the Central Valley, and was ex- panding at the expense of the others. Further, they suggest that this fact was probably the motivation behind other phases' nucleation and defensive posture. According to the present construct, a "Complex II"- type system is indicated. Importantly, however, dramatic differences are evident between Nodena and Parkin and the "Complex II" Cairo Lowland phase. It would appear that in the later; phases, chiefly control has probably become nearly absolute. Not only sites, but entire settlement systems (Parkin) were pre-planned and politically 210 motivated. Virtually no political power or import of UDGs would be expected and the mortuary data bear this out. There is not a trace of UDG symboling in the spatial dimen- sion, and it is predictable that none would be evident in the artifactual. Doubtless the occupants of the major towns consisted of members of many UDGS, and the family or extend- ed family had probably become the important kin-based unit. The normative mode of disposal for commoners is house-relat- ed, in family or extended family groups. Spatial differen- tiation probably reflects differences in vertical status. In any event, systems of "Complex II"-type were extant in the Cairo Lowland, and at a later date further to the south. Their structural differences demonstrate a far greater level of systemic control in the latter than in the former. Summary Discussion With presentation of an assessment of structural form in the Mississippian of the Central Mississippi Valley con- cluded, it remains to review these findings and point out the utility as well as shortcomings of the approach. Sug- gestions for further research into structure in this region will also be provided. It has been posited that the climax Cairo Lowland phase manifestations are characteristic of "Complex II" ranked systems, where settlements are typically comprised of 211 multiple UDGs. At one and possibly two sites, discrete UDG cemeteries are present, where UDG affiliation was spatially symboled by interment in planned, exclusive disposal areas within the settlement. Further study could involve the search for artifactual patterning in the data that might support the spatial inference. More importantly, future excavation at other Cairo Lowland phase sites might be dir- ected toward the discovery of similar discrete cemeteries, since they should be present. It will be recalled that the best confirmation of a perceived UDG pattern at one site is the discovery of the same at other related sites. No statements can be made regarding the development of Cairo Lowland out of the Hoecake phase, primarily due to poor data from the latter. However, this situation may exemplify the greatest shortcoming of this approach: its inability to discriminate horizontal groups from multicom- ponent data sets. Hoecake is a complex, large site with multiple burial mounds. The present method involves the recognition of spatial or artifactual patterning in mortuary context that could signify horizontal groups. It is diffi- cult to be conclusive in this regard, but it is likely that such pattern recognition might be impossible at a site like Hoecake, even if it were completely excavated. Pattern rec- ognition would depend upon the precise temporal and spatial control of individual components; if a burial site was in fact utilized over several hundred years' time, and if cul- tural changes were taking place during that time (as seems 212 to be the case at Hoecake), then one might not be able to read any patterns at all. Stated differently, this approach seems best employed with data sets representing discrete mortuary events of fairly limited duration, and in relative- ly "stable" systems. Systems in transformation may not dis- play regularized mortuary behaviors, and may in fact, pro- duce behaviors that obfuscate mortuary patterning. This approach to socio-political structural form, then, does not in itself elucidate cultural processes or evolution (but see below). Mississippian in the central portions of the Central Valley has been posited to be the result of a population movement into the region; the source of this movement is not known, but north is a logical choice. The data from the Big Lake phase was not amenable to analysis, but the succeeding Cherry Valley phase has produced an interesting phenomenon. Based on what appear to be localized, isolated disposal areas, it has been posited that single UDGs are the likely users of these facilities. In turn, such a situation sug- gests localized, concentrated UDGs that, in a ranked society would resemble a conical clan system. This estimation was based on data from three mortuary sites, and the implica- tions for future research are obvious, that habitation sites must be located and studied, and major mortuary facilities must not occur here. In order to test this hypothesis, evidence suggesting UDG localization should be found in set- tlement patterning. Further, if a hierarchical pattern was 213 extant, then this should be observable. An analysis of the vertical status dimension would also be desirable. It was further suggested that the hypothesized conical clan system for Cherry Valley is consistent with known eth- nographic processes of system expansion. To further evalu- ate this observation, data from the Big Lake phase should be sought to determine whether UDGs are in fact localized at this earlier time, when they presumably would have recently moved in. Scant mortuary data from Zebree and Kersey indi- cate Charnelling at the latter, and possibly a house—related and status grave pattern at the former. Neither site con- firms or rejects the hypothesis, but does suggest that mor- tuary practices were different at this time. In short, if Mississippian in this area was the result of the movement of discrete UDGs, then this pattern should be consistent over time. Should Big Lake mortuary and settlement data indicate something else, then the hypothesized Cherry Valley conical clan might be in jeopardy, or at least the population move- ment aSpect of it. The late Cherry Valley manifestations are poorly known, but resemble more "typical“ Mississippian patterns, with fortified sites and hierarchical settlement pattern. A sys- tem transformation of some significance has evidently taken place and, based on expectations of structural form, a shift to multiple UDG settlement organization could be hypothe- sized. This hypothesis would be straightforward to test. It was stated earlier that one of the limitations of 214 this approach is that it cannot, by itself, elucidate dyna- mics and cultural evolution. It is, rather, an evaluation of structural form at discrete evolutionary moments. With this, it will be proposed that the method can in fact lead to the explication of processes and dynamics through the study of change in structural form through time. Cherry Valley is a case in point. A particular socio-political configuration for this phase has been proposed which must be further tested and evaluated in the field. If it stands, then hypotheses can be advanced regarding expected structure in the preceding Big Lake phase, gi§., that localized UDGs representing immigrant Mississippian groups are expectable. Thus, this form can be linked to known cultural dynamics of population movements in an hypothesis testing framework. Similarly, the configuration after Cherry Valley suggests dispersed UDGs and greater chiefly systemic control. If mortuary data confirm this, then one can observe a trans- formation from concentrated, single UDGs to dispersed, mul- tiple UDGs, which is consistent with known processes regard- ing the accumulation of chiefly power and the evolution of complex chiefdoms. It is equally important to point out that if any of the preceding hypotheses are'ngi borne out, then other processes or dynamics were evidently operating. The Nodena and Parkin phases represent the zenith of ranked society in the Central Mississippi Valley. Further study here should involve research into contemporaneous phases located to the south, in order to better assess their 215 relationships to Parkin and Nodena. For the latter, an evaluation of intra-site mortuary patterning would prove fruitful; a detailed study of vertical and perhaps horizon- tal structure, as reflected in spatial and artifactual vari— ability in the mortuary domain, would certainly lead to a better understanding of the ranked systems in operation here. It is noteworthy that these nucleated sites contain the remains of everybody who lived there (i.e., nobody seems to have been buried anywhere else), and thus, the full range of vertical and horizontal variability should be present. Conclusion This chapter has attempted to show how the recognition of potential horizontal groups in archaeological context can be integrated into a study of socio—political structural form in an extinct society. As stated in chapter 4, the three forms of ranked society developed here are considered as heuristic devices only, and are based on known configur- ations and dynamics observed in ranked systems. It must be stressed that the three forms have been developed to illus- trate how the horizontal and vertical dimensions of mortuary variability may be integrated and thus yield a more coherent picture of systemic complexity. The three forms probably cover most of the range of structural variability in ranked systems but certainly do not exhaust it; the development of other formal expectations or modifications of these should be done. Indeed, it might be desirable to divide "Complex 216 II" into relatively more versus less complex forms, based on the nature of system control and UDG symboling. The "Basic" category can stand further scrutiny and refinement as well. Stated more generally, the above approach is advoca- ted as an alternative to mortuary studies of the vertical dimension only, and as an alternative to the pigeonholing phenomenon. It has been shown quite clearly that this study of Mississippian phases goes far beyond simply label- ing them as ranked. The study has been multidimensional, with horizontal mortuary patterning being used as one datum in conjunction with other lines of inquiry. With this, this analysis has attempted to introduce a "sociological perspective" to mortuary, settlement, and other analyses. As was stated at the close of chapter 2, it is this wri- ter's belief that the recognition that spatial patterning of material remains reflects the socially significant groups of the users can lead to important insights into the structure of past social systems. Finally, it was noted that the attempted elucidation of structural forms is essentially a study of statics, or of static regularities in the archaeological record. In this respect, the forms are nothing more than refined, de- tailed pigeonholes. But unlike pigeonholing, this approach can lead an investigator to the discovery of dynamics and evolutionary processes. The present approach is a study of structure in discrete system states; this knowledge can form 217 the data base for processual studies examining continuity or change between states. It has been shown how certain struc- tural forms of ranking correlate with known cultural dyna- mics, such as system expansion, the competition for status, the institutionalization of power, and the decline of UDGs. Similar correlations of formal structure, material pattern- ing, and cultural dynamics can be undertaken. The applica- tion of this approach to the tribal domain might prove interesting, and could lead to a more precise understanding of the variability within these systems. CHAPTER 6 DISCUSSION AND SUMMARY The preceding study has attempted to expand the current state of mortuary method and theory to incorporate the horizontal dimension of mortuary variability. This has been done to provide an adjunct to studies of the vertical dimension which have heretofore dominated in mortuary analy- sis. The two perspectives are complimentary of course, but with a distinction: the vertical dimension focuses atten- tion toward the treatment of individuals, while the horizon- tal concentrates on the behavior of social groups. It is the latter that have been a traditional emphasis of social anthrOpology, but have proven difficult to deal with in archaeological context. This study has been based on a body of ethnographic mortuary and social anthropological data. In the present opinion, the ethnographic record has been overly neglected in mortuary studies, and the present analysis demonstrates that the use of ethnographic data can be highly profitable. With this, this study has incorporated uniformitarian prin— ciples and assumptions in order to render the results rele- vant to the past. It has been pointed out that regular, prescribed procedures for the disposal of the dead must have 218 219 existed in the past, and that horizontal groups must have also; based on an examination of cultures in the contempor- ary world, both are sound uniformitarian assumptions. It has been shown that horizontal groups are symboled more often and more consistently in mortuary context than has previously been believed or suspected. Of greater im- port is the fact that within the present sample at least, over 75% of this symboling is archaeologically visible and interpretable. Most visible patterning is spatial and in- volves the use of formal disposal areas of differing config- urations. From the various analyses conducted in chapter 2, two theoretical postulates with three corollaries have been derived that will allow the archaeologist to make inferences of high probability regarding the presence of horizontal groups in a mortuary context: Postulate 1: When a horizontal group is s mboled in mortu- ary context, ii will usually be the one(§) that pegple consider dominant in their social structure relative £2 and a; the expense gg other such groups. Postulate 2: Where unilineal descent groups are present in 3 society, they will with few exceptions con- sistently segregate their dead, usually 93 the basis gi_membershipin one gi these groups . Corollary 2a: Corollary 2b: Corollary 2c: 220 The presence of a more or less formal, ex- .— clusive disposal area indicates with high probability the presence of a discrete uni- lineal descent group. Where disposal areas are divided into formal divisions, or where multiple disposal areas occur around/near a settlement, they will invariably contain members of the respective multiple unilineal descent groups inhabiting the settlement. If a disposal area is present, probability favors the presence of a clan, but lineages can be similarly represented. As indicated, clans and lineages dominate in the spatial representation of horizontal groups; phratries, moieties, and sodalities are nearly invisible. The above results have been incorporated into a multi- dimensional research program designed to examine variability in socio-political structural form in archaeological con- text. A large part of this variability involves the spatial arrangement of UDGs within a system, and this research has shown how these patterns can be detected. This ability, coupled with data drawn from theoretical and synthetic anth- ropological literature, forms the basis for the development of three models of alternative socio-political structural form within one socio—political category, ranked society. These forms were called "Basic", "Complex I", and "Complex 221 II", and it has been stressed that they are heuristic devic- es designed to enable the archaeologist to assess variabil— ity in structure in ranked systems. It is clear that these constructs, particularly "Basic" and "Complex II", can be further refined and expanded. While the present focus has been on ranking, the procedure can be applied to any kind of system; the application of this approach to the tribal dom- ain might help explain some of the myriad variability extant within this category. As an illustration of this approach, a test case involving various Mississippian Period systems in the Cen- tral Mississippi Valley was selected. It was~shown how variability in mortuary practices, settlement patterning, and other cultural factors can be understood in terms of variability in formal structure. That is, it has been known or assumed that most or all cultural systems recognized as Mississippian were chiefdoms, but that they manifest consid- erable variability between systems; the present study has illustrated how such variability can be explained, and thus allows the archaeologist to make inferences that go far beyond the pigeonholing of these systems as "ranked". It was also discussed how this approach is a detailed study of static patterning in archaeological context: it is the attempt to elucidate structure in discrete system states. Importantly, this detailed study of static pattern- ing can lead to the study and discovery of cultural dynamics. The Mississippian example attempted to correlate variability 222 in structural form with cultural processes known from ethnographic ranked systems. Thus, while the present method itself is concerned with archaeological statics, the results can lead to the fruitful study of dynamics and processes. The results presented in chapter 2 regarding the asso— ciation of formal disposal areas with lineages and clans naturally led to the examination of a previous explanatory construct: Saxe's Hypothesis 8. This test of the hypothe- sis has shown that formal, exclusive disposal areas are no; the invariable result of corporate group control over cru- cial but restricted resources. These variables are far too restrictive and based to too great a degree on an economic premise to be of broad applicability. The primary problem with Hypothesis 8 is the attempted isolation of economic corporateness and specifiable economic corporate groups that actually own/control land or resources, since these condi- tions will not obtain in every case, and it is clear that societies will use disposal areas for other reasons. It has been argued that corporate behavior is but one aspect of socially—defined descent groups. Descent group corporate behavior of any form can vary considerably from society to society; there is no reason to expect that such behavior or group function should be uniform over time or Space. Thus, it is misdirected to utilize variable features, or behaviors, of descent groups as basic analytical units. Rather, the descent group itself should be the analytical unit, since these groups will be far less variable than the wide range 223 of corporate behaviors that they can potentially engage in (or not engage in). Finally, this test of Hypothesis 8 has demonstrated that one cannot infer intense competition for resources from the presence of formal disposal areas. To conclude, some brief comments will be offered about certain aspects of this study that, in addition to the sub- stantive results, are believed to be of value and should be employed in future mortuary studies. Following previous investigators, this study has demonstrated the utility of the spatial dimension of mortuary variability as an impor— tant indicator of structure. Future mortuary studies should continue this trend; the limitations of artifacts-only anal- yses have been aptly demonstrated. Similarly, this study has illustrated the value - indeed the necessity - of a mul- tidimensional approach to mortuary studies in particular and archaeological data in general. It is this writer's belief that for the development of an anthropological archaeology to continue will depend in large measure on the use of mul— tiple avenues of inquiry and the testing of hypotheses with multiple data sets. This study advocates that ethnographic data can be successfully used to generate expected material correlates of mortuary behavior. It is clear that the Human Relations Area Files contain a large quantity of excellent data on . mortuary practices: unlike many other aspects of culture, the fact of death must be dealt with by all societies, and is always an area of interest to anthropological 224 fieldworkers. It is heartily suggested that future mortuary studies employ a more ambitious use of these data. Finally, this presentation has advocated the adoption of a "social perspective" in archaeological studies. Empha- sis has been placed on the anthropological principle that all human societies configure themselves into socially sig— nificant groups based on the recognition of kinship rela- tionships between people. Further, it was demonstrated that these groups leave material traces in the archaeological record that have heretofore been only poorly elucidated. To be sure, the problem is partly based on the fact that such traces can be ephemeral or otherwise difficult to interpret; simultaneously, however, the current state of method and theory does not direct the archaeologist toward the discov- ery of these groups, nor provides the necessary analytical tools with which to do so. Hopefully, some of these short- comings have been rectified with the present study. Kin groups can be strongly inferred from archaeological remains, and the recognition of these groups should be integrated into a systems approach. It is in this context that the aforementioned factors - a concern with spatial phenomena, a multidimensional approach, and use of the ethnographic data base - can all be brought to bear on the problem. The present sutdy is intended to be an initial step in this direction. APPENDICES APPENDIX A MASTER DATA LIST Key to abbreviations LOC - location: 0— Oceania; AF - Africa; NA - North America; SA - South America; ME - Middle East; A - Asia. DISP MODE/UDG SYM - disposal modes/unilineal descent group symboling: R - ritual F - family A - artifactual L - lineage cem - cemetery (incl. tomb, C - clan crypt, scaffolds) P - phratry crem - crematory (no burial) M - moiety cr/cem - crematory, w/burial S - sodality H-R - house-related b - "band" NP - no pattern v - "village" sg - "segment" DES - descent: Pa - patrilineal; Ma - matrilineal; D - dou- ble; Bi - bilateral UDG/S - unilineal descent group and/or sodality present (see symbols above); underlined: "dominant" UDG(s); Slash(/): "very weak"; n-UD - non-unilineal descent UDG SET - unilineal descent group (lineage or clan) settle- ment pattern: Si - single UDG settlements; Mu - multiple UDG settlements; —9‘- "transi- tion"; seas — seasonal CONC/DISP - concentrated settlements (e.g. "villages", "towns") versus dispersed settlements (e.g. "hamlets") DIV'SET — divided settlement: settlement subdivided into recognized residential areas 225 226 APPENDIX A (cont'd) SUB — subsistence practices: A - agriculture; P- pastoral- ism; F — fishing; H - hunting; G - gathering; underlined: dominant subsistence mode 227 mam + a am . mo mm 566 o a Monaco: .Hm mmmw + 0 am I m a 566 o o mmuoam .om mom + o Anzalamv mwm mm Mum ma flocsm .ma amm + o as as ca ms 566 o o mpcmHmH Hamnmumz .mH em I a as as on ma mum o mamoxae .sa mmm< u o as as om mm msm m4 meanne .GH mg I U .92 .92 UM ME ETD A Md ..nvfimnmaw . mH cream + o 52 I w m: Emu o < Hmmnm .ea semen + u I a: mg a smo owe me spawns .mH mam + o I 9: mu o Mum o mosmm .NH mm« + a I a: a a mum o mmmmmmcoa .HH mm": omxzmm ummmm . o u I comm Hm smo > me anon .oa exam + a as :2 on ma mum we Humem .m mom a a as 52 cm mm mum m4 sass .m mmmm + o as as mom mam mum we acomz .s meow + a as as mom mm 566 m we ammo: .G mam + o as as mug mm mum «mum ma masses .m exam + u as u m a mum ma cum .4 comm + o I 52 mg no mum mm mosz .m seam u 0 am am on mm 966 o m< macaw .N m4 I U I I QDIC .mm ETD > O m>wb .H can: meaem mam 9mm amHo o a m\ca: mun sew on: com wechom\* mmac >Ho \ozoo sea was \mooz mmHo ip.ucoov « xHozmmm< 228 cam I a am am om mm mIm ms mumnmaa .ha swam I a Hm I w as ace 0 4m ousflmoc .oa on I 0 am I so am mIo “smuo o a mace .ma serum + 0 am I 0 am mIs o maaaxouo .aa mmsm I a Hm I m am as o camaMH .me swam + a Hm I m mm Emu u we omaos .ma swam + a I am H am 866 a «m maoaaas .Ha mam + no I am so am smo a s2 uaxo» .oe mess I o I am 4 mm as o axawamx .mm smug ammHme seam I o I I mss am «Is “mIm m4 amasxmsz .mm cum I o maamn msq as «Im a:60 n «z mmazmm .mm mom I o Ammmmv mos a Emu m as mxuooz .mm memo I o Ammmmv m mm as we uouamuuom .mm cmmam + a as as on ma mIm ms xaaaaam .am MM + O as .I 96 mm Soc 0 4 magafim . mm mam + c I as ma am 560 a 0 mean mxam .mm use I 0 am as msamq as sIo 1860\uo o «z pamaaae .Hm mam I o am as us as smo\uo o s oaoq .om use I 0 am I mm o mIm “smo 0 <2 aaooo maamm .mm exam + 0 am as ma mm mIs a immeacv assume .mm new + 0 am as mos mm 560 o 0 zones .am seam I 0 am as Ms ma m.sIu llpcsmo o as masons .mm new + 0 am as mm mm 566 u we Manama .mm exam I 0 am am om as mIo “Emu a we anamm .vm saw I 0 am am om mm smo a o pamHmH umumam .mm mas I no am am hm us mm 566 was as auas .mm Aa.uaoov « xHozmmas 1“. ‘ ' 6 5V J / I I) ’\ u— '-4 - j u. o o ' ‘- l h u ‘u \ ‘ ‘I .4 r . .. ,- l V I ‘H 4 -. ‘ i. . I' M . .. ~ ‘\ I u ,- ‘ ll , Q . I uro‘ :‘J momm 8% cem seam mmmw mwmhm 0mm mew mew swam mom worm he mmmm w mmflm 2 mmm mmm Ummm mom whom 0mm 0mm mm mwmm drum Km when macaw l++l Il++l + I-+4-+ l+ ll I+-+-+ I + I+I DUUQDUODUQQUUQUQQUUUQUOUQQUDQUU D2 52 52 DZ 52 52 SS 92 DZ 52 52 DZ 52 52 52 DZ 52 52 SS Hm Hm am am Hm Hm Hm Hm flm Hm Hm 52 msmma swq ms was msaom msaom 0M mom mom mos mm m2 m2 mm mm mm m2 mm 0 NE on MS Hm\mm mm ME ME Q MS MS mm mm mm mm mm mm mm mm mm mm mm mm m.¢IU «Emu ml: “Emu Emu Emu mz mz Emu ¢IU «mlm Mlm mum mlm mz m.4lo “Emu m.¢lo “Emu mz mIS.mIU «Emu mIm mIU “mz mz duo “Emu ADVEmO mum Emu m.4IU «Emu mz mlo «mlm mlm film mlm mlm Emu 1p.uaoov a chzmmm< QIJL)> Q >'> 0mg 0 A son mwosoouH some canmaamm Hmucmm ammo: om» momm cumuwm mHH oan onm>mz menano cameo sta osmHmHH 3oz ommoz mxomx moow> osmose mmcm om: 05A comoo Hmsmaama onouoo sosusossz MEOSM flcmasm cxsflo stz >fla .wh .55 .m5 .m5 .on .me .Nh .Hh .oe .mm .mm .hm .mm .mm .vm .mm .N@ .H@ .om .mm .mm .hm .mm .mm .om .mm .Nm .Hm .om .mv .mv 230 I+I++|l QQOUUUUUU UUUUUUUUQUUUUUQQU Anyone laaoac Aasmuamv 52 SE 92 52 DE DE DE DE DE 52 52 SS QDIG m.ooua m.coIa m.coIa QDIG QDIC m.ooIa m2 msm mswq mswm msoq amm msmoa mom mws msoa DQQNQ Hm flm m 0 am 02 ME mm ME ME m2 m2 02 m2 Ewo\uo > mIm «EGO mIs as “as «EwO “Emu lama «as sumo Ac.aaoov s xHozmams > b > > >=>U <2 <2 < 2 OOO<< <2 <2 <2 <2 <2 <2 mono onmmwu< sous» numEmHM 080m mmmnwmma mocmEmama oumu asaxam ononom wounds .3 secs Hana BOHO smosmz .voa .MOH .Noa .HoH .OOH .mm .mm .5m .0m 02fln MBMRGMHMM ommmmm Moscowuflno oumamomoz nuances .mHH .oaa .MHH .NHH .HHH .OHH .moa .mOH .hoa .moa .moa \omqmmbWNI‘" o o o o o 0 o 0 HI" HO 0 u 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. APPENDIX B REFERENCES FOR SOCIETIES IN APPENDIX A Java (Jay 1969; Nakamura 1982) Ganda (Roscoe 1911; Fallers 1960) Nupe (Nadel 1942; Forde 1955) Edo (Goldstein 1976) Yoruba (Ajisafe 1924; Forde 1951) Mossi (Mangin 1921; Delobson 1932) Ngoni (Barnes 1954; Read 1956) Zulu (Raum 1973; Service 1975) Swazi (Kuper 1963) Lozi (Gluckman 1941) Tongaeses (Beaglehole and Beaglehole 1941; Gifford 1929) Samoa (Mead 1930) Mbundu (Hambly 1934; McCulloch 1952) Khasi (Gurdon 1907) Ashanti (Rattray 1932; Saxe 1970) Azande (Evans-Pritchard 1971) Tikopia (Firth 1936, 1959) Marshall Islands (Wedgewood 1942; Mason 1954) Rundi (Meyer 1916; Roscoe 1924) Flores (Kennedy 1955) Monguor (Schram 1954, 1957) Kurd (Masters 1953; Barth 1954) Easter Island (Cooke 1899; Metraux 1940; Sahlins 1958) Bemba (Richards 1939, 1940) Tanala (Linton 1933) Thonga (Junod 1927) Ambon (Haar 1948; Kennedy 1955) Kachin (Leach 1954) Bella Coola (McIlwraith 1948) Lolo (Lei 1944) Tlingit (Jones 1914; Oberg 1937, Krause 1956) Puka Puka (Beaglehole and Beaglehole 1938) Siwans (Cline 1936) Shilluk (Westerman 1912; Dempsey 1955) Hottentot (Schapera 1930) Nootka (Sapir 1921; Drucker 1951) Pawnee (Wedel 1936; O'Shea 1981, 1984) Nyakyusa (Wilson 1951, 1957) Kapauku (Saxe 1970) 232 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 233 APPENDIX B (cont'd) Yokut (Gayton 1948; Kroeber 1953) Mapuche (Cooper 1946) Mongo (Maes 1924) Ifugao (Lambrecht 1932) Orokaiva (Williams 1930) Toda (Rivers 1906) Goajiro (Gutierrez de Pineda 1950) Lugbara (Middleton 1965) Tiv (Bohannon and Bohannon 1953) Nuer (Huffman 1931) Dinka (Deng 1972) Fulani (Goldstein 1976) Kwoma (Whiting 1941) Mundurucu (Murphy 1960) Dorobo (Huntingford 1954) Tallensi (Rattray 1932; Fortes 1945) Dogon (Palau Marti 1957) Luo (Hartmann 1928) Mae Enga (Saxe 1970; Goldstein 1976) Tucano (Goldman 1963) Vedda (Seligman and Seligman 1911) Kaoka (Hogbin 1964) Wogeo (Hogbin 1970) New Ireland (Powdermaker 1931) Tiwi (Hart and Pilling 1960) Omaha (O'Shea 1981, 1984) Ojibwa (Jenness 1935; Kinietz 1947) Navaho (Kluckhohn and Leighton 1946) Igbo (Leith-Ross 1939; Ottenberg 1971) Ila (Smith and Dale 1920) Herero (Vivelo 1977) Fang (Trilles 1912; Trezenem 1936) Yao (Stannus 1922) Masai (Hollis 1905) Santal (Culshaw 1949) Fellahin (Blackman 1927; Ayrout 1945) Truk (Gladwin and Sarason 1953) Iroquois (Morgan 1901; Fenton 1936) Fox (Tax 1955; Gearing 1960) Mandan (Deland 1908; Bowers 1950) Crow (Lowie 1935) Zuni (Cushing 1896; Bunzel 1932) Hopi (Voth 1912) W. Apache (Kaut 1957; Basso 1970) Bororo (Lowie 1946) Kikuyu (Middleton 1953) Garo (Playfair 1909) Talamanca (Stone 1962) Tapirape (Baldus 1944) 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 234 APPENDIX B (cont'd) Pomo (Loeb 1926) Klamath (Spier 1930) Yurok (Heizer and Mills 1952) Arapaho (Kroeber 1902; Higler 1952) Gros Ventre (Kroeber 1908) Comanche (Jackson 1972) Iban (Roth 1892; Freeman 1955) Murngin (Warner 1937) Aranda (Schulze 1891) Walbiri (Meggitt 1962) Andamanese (Radcliffe-Brown 1922) Ainu (Landor 1893) SE. Salish (Ray 1932) Washo (Lowie 1939) Ute (Lowie 1924) Copper Eskimo (Jenness 1922) Wichita (Goldstein 1976) Mescalero (Opler 1969) Chiricahua (Opler 1941) Papago (Underhill 1939) Karankawa (Newcombe 1961) Jivaro (Karsten 1935) Carib (Gillin 1936) Ona (Gusinde 1931) Calinago (Rouse 1948) Mbuti (Goldstein 1976) {Kung (Lebzelter 1934) 235 omucmu OmAm ma UCMA “mnmnamfi GMAO umsn uoc .mcomcm ou uso pH Amonmm AAA3 pan .ocma n30 mcmAu I + U U mocmw mmAAAEmm >9 owAAouucoo cam on #90 omAmoumm ma Gama pan .mmuwumm :30 mommmGAA I + A\m A UnmamH Hmummm Umuwumncflnflnumm “UcmA n30 mommmcflAIAHumm m + U\A U AmcmAAma confluwscfllAHumE up:MA £30 mommmqflAIAHumz m + A A Aucmnmm mmAAHEmm GA\3 .mocmuflnmncflufluumm «Amo Imam mumm HmnEmE comm unmumz .pcmA :30 mGMAU m + o o omcoz cMAU GA\3 mocmuflumscfl “pamA moumom G30 mCMAU + + U o mmnm om: mmAAAsmm cA\z .Auumm emu Ifiumncfl pamA «momso: .mcwonwm .mmwhwumewo chUsAOCA .mmmMAAA> afl\3 UGMA CBC mooanHA + + A A mxsm mxsm mozmeHmmmzH\mmozme mmmm mmHe oo onmBzoo\mHmmmmzzo momaommm\az¢A mmm muzm zoom A mmHQ ammo wm4zzbm Iflocfl mmafiuwaom can mommmGAAIAHumE Amfiflaflz pwuflumscfllwuumfi “000A 0030 Gwaolwuumz Gama pwnmaafl\c30 mmHAHfiwm ccmA “Auwscfl\czo mmAAAsmm mocmufluwncwlfluumm «Ummummmfio ma pan .msoum AMOOA w you ma Hume “=00A0 H0 >AAEmm poocmuxm m0 nocmun: 0 mA .000A mc3o Avon mocmuAuoscfl Ifluumm “000A uflumncfl\c3o mommmcflA H0 meAAfimm mmono :30 mmAAHEmm “mommafla map 0H\3 mmmum pan .Ump IAH0£0A um>m mAAmmu 00: .008A 0A0: mommmcflA mommafla IAHumE nosousu omuuAEmcmuu UQMA ou muosnmzmb omuflnmzcw 0:0 00:30 AAAEmm “000A mo mAmoumm moosmflucoo 0p SAMAU >mA mmAAAEmm omumAwH mo mommmcfla AmEAqAZ UwuHHchH IAHumm «ocma onuucoo mmmmch AMEHGHE AMUOA mmAAHEmm cH\3 cmufluwncfl Iflupmm “UGMA Aonucoo mommmcfiA AmEHGHE AMUOA 10.»:003 o xHozmmm¢ (‘0 (‘0 no (‘0 (‘0 (‘0 (‘0 A GAS A\m A\m nae CHE GAE 09D 92 GOD 32 Aqu Amom mmcone OHOA 00084 mmqum 0Amcma mfloswonH 00000002 500502 >AB 237 00:00Hu0::fl hAHE0w “0A00H0m Anm0 :30 m0AAAE0m «0:0A 0AE0:000|:0: :30 0:0AU 00:30 A050A>A0:H 00 00:30 :0A0 Eoum m:w>0& E0um>m “000AH0::A 0:0 00:30 >AA050A>A0:A\>AAE0M 0:0 us: .m0:0A :0AU 0:0»:0 00 H0su0m .30A EOA002 :msounu 00:00HH0::A “50:0 HA0m 0:0 as: :00 .mpOAm :00u0m :30 m0AAAE0m «m0Au0u0E00 :30 0:0AU 0A00H0m 0:0: 0:0:300:0A 0mH0A 0800 “00:00 IAH0::HIHH00Q .0:0A :30 A:0E ..0.flv 00HAHE0m m:HAA0m 0:0 m:H»:Q A090A>H0:A 00 m:H>oE E0um>m «0:0A :30 AA0 0A000A>H0:H .00AAAE0M .00000:AA A0EA:AS 0:0A m0 m:AAA0m 0:0 m:A>:Q £058 «00:00AH0::HIAHH0E >AA0:05 .0:0A :30 AAO ma0s0fl>fl0:fl .m0AAHE0w .m:0AU 0:0A m:AN0:m 00m: hAA0::EEOU 0:30 :0A0 «00:00 :Au0::A H0As0::>0IAHu0E “0A0on0m Aum0 :30 00AAAE0M .mx:0nu0>AH 0:0A0 0:0A onuu:00 :0a0 0:0A 00AH0::A\0:30 >AAE0m 0:0A ufln0£:A\:3o m0HAHE0m 10.0coov o xHozmmm0 +| + +l m\q :aE U\m U\m A :He was 1: GOD I: 0:0A0HH 302 Hosm:oz 0:03Am cflnmAAmm :omoo Amaze 00Afl000 0>0h can: 238 00:00H00::Hlfluu08 “0A A0H0:00\00: 00AAA80M .wuoufiuu0u 0>0: 0:0AU 0:000 >AM0A£0 :A\3 00:00AH0::A “0:0:08 I800 00 00000050: 0:0Hm “.000 .00000: .00:000n .0:0A "m:A:u>H0>0 :30 0m0a:00« 0:0A 0800 :30 mAA000: 00AAH80M «00:30 5A000Au0> .=00u0000 mo >SUH0H0H:= 00000AA0800«« msoum aAm0H30 :H\3 00:00 IAH0::A “0A 00000AA0 0:0 0:0A :30 0M0A:U«« 0:0A 00000 00 :oAu0mu:0: .m:Au:0H 00:8 “0:0A :30 AA0 0AQO0Q :08800 .0:A0um0A:0 .00000 “0:5:00 0:0 0A5000:30 x0Am80000 uosumdms 0:0 ua80:0uu 00AAA80M «00AAA80M 00:08800 ou 000:0m90: 0:000 .0:0A AA0 :30 0M0A:U«« A0:0Hu0:uA0 00 :00 “.000 .00An00>08 .000H0000H .0:0A :30 AA0 :00 0A000H>A0:A .00AAA80M .00000:AH "005:0: 0:0 QAS0H0:30 x0A0800«0 0:0A 0H80:0H0\:30 00m00:HA QA:0H0:30 A000A>A0:A 0800 00A0 “0A0mmsn :30 00AA Ifl80m «000: >AA0:58800 .0m000000m :30 0:0AU (\o 00 10.08000 0 xHazmmm< (\o (\o (\o U\m 00 *k. «0 «0 fix. 0* U\m was I: :a8 030004 :H00003 mxuooz Hue: 0:80m UHSM 00:0A0H AAmsmumz xsue H0002 0009 239 000:5000: :0 0:0A mo 0:00:0:30 oz«« 000:5000: :0 0:0A mo 0::0:0:30 ozi 000:5000: :0 0:0A mo 0::0:0:30 0200 000:5000: :0 0:0A mo 00:0:0:30 0200 050:0 00 050:0 80:0 00H:0> 0000:0:3o .800000 00x00 0: 0000:5000: .000:0 .0A00:00 00:30 000A0800 00:00::0::H :0 00:0:0:30 A00: 0: .050 00A00:00 0:0A “0000AAH> :00: 00:0A 00AA0:0:00 0000:0«0 00:00::0£:A :0 0::0:0:30 A00: 0: “00::O0A::00 :0>o 0005:0505 0A0: 00:000 >:o0:::00 no 005 A05:H0:00 mo x00A 0050:00 0>A00000 AAA3 005:0505 .00:00H:0::0 A00: 0: “00::00:::00 0::0:5: o0 0005:m I505 :30 0000AAH> :0A0 ::\3 000A080m A000A 00:00::0::H 00x00 0: .0000 I:0:30 A00: 0: “0A00:00 0:00 :0A0 .00:08 I000 0:00 :0>A: 0:00:00 .00A:00H::00 A0QA:B 000::0£:H :0 00:30 00: 0:0 050 .00A:00H::00 0>0£ 0:0AU 00.08000 0 xHQmem< +1 +l (\o +l “.4. 0* 00% 0* yak. 0m 0\0 GOD 52 GOD I: GOD I: GOD 52 GOD 52 :00:02 09:0: 80080:: 000::00 0800 00:300 manHflo 0A000 0HH0m 0:0059 00:080A0B 240 000:5000: :0 0:00 00 00:0:0:30 oz«« I 000:5000: :0 0:00 00 00:0:0:3o oz«« I 00.08000 0 x0020000 *0 *0 GOD I: was I: 5:00 £0000m .m.m BIBLIOGRAPHY BIBLIOGRAPHY Ajisafe, A.K. 1924 The laws and customs 9: the Yoruba People. Routledge and Sons, Ltd. London. Ayrout, Henry Habib 1945 The Fellahin. R. Schindler. Cairo. Baldus, Herbert 1944 Os Tapirape, Tribu Tupi no Brasil Central. Revisto do Arguivo Municipal, Vols. 96-105, 107-I24,_IZ7. Sao Paulo. (Translation by HRAF). Barnes, J.A. 1954 Politics in a changing society; a political his- tory of the Fort Jameson Ngoni. —Oxford Univers— ity PEEss. London. Barth, Fredrik 1954 Father's Brother's daughter marriage in Kurdi- stan. Southwestern journal of anthropology, Vol. 10, pp. 164-171. __ Basso, Kieth H. 1970 The Cibecue Apache. Holt, Rinehart, and Winston. New York. Beaglehole, Ernest and Pearl Beaglehole 1941 Pangai; village in Tonga. The Polynesian Society. Wellington. 1938 Ethnology of Puka Puka. Bernice 3. Bishop Museum, Bulletin 150. Honolulu. Befu, Harumi and Leonard Plotnicov 1968 Types of corporate unilineal descent groups. In M. Fried (ed.), Readings in anthrOpology, Volume £1. Crowell. New York. __ Binford, Lewis R. 1981' Bones: "ancient men and modern myths. Academic Press. New York 241 242 1971 Mortuary practices: their study and their poten- tial. In J.A. Brown (ed.), Approaches to the social dimensions of mortuary practices, Society for American Archaeology, Memoir N9. 25. Blackman, Winifred S. 1927 The Fellahinof_gpper Egypt. Harrap and Company, Ltd. London. Bohannon, Laura and Paul Bohannon 1953 The Tiv of central Nigeria. Ethnographic Survey of Africa, Western Africa, pt. 8. InternatiofiaI African institute. London. Bowers, Alfred W. 1950 Mandan social and ceremonial organization. Uni- versity of Chicago Press. Chicago. Brain, Jeffrey 1978 Late prehistoric settlement patternstIthe Yazoo Basin and Natchez Bluffs regions of the lower Mississippi Valley. In B. Smith (ed.), Mississ- ippian Settlement patterns. Academic Press. New York. Braun, David 1981 A critique of some recent North American mortuary studies. American Antiquity, 40(2):398-415. 1979 Illinois Hopewell burial practices and social organization: a reexamination of the Klunk— Gibson mound group. In D. Brose and N. Greber (eds.), Hopewell archEEology: the Chillicothe conference. MidcontInental journal of archae- ology, special paper No. 3. Brown, James A. 1981 The search for rank in prehistoric burials. In R. Chapman, K. Randsborg, and I. Kinnes (eds.), The archaeology of death. Cambridge University Press. Cambridge. 1971 The dimensions of status in the burials at Spiro. In J.A. Brown (ed.), Approaches to the social dimensions of mortuary practices.‘ Society for American Archaeology, Memoir, No. 25. Buikstra, Jane E. 1976 HOpewell in the lower Illinois valley: a regional approach to the study of human biological varia- bility and prehistoric behavior. ‘Northwestern 243 University Archaeological Program, Scientific Papers, No. 2. Evanston. Bunzel, Ruth L. 1932 Introduction to Zuni ceremonialism. Forty-sev- enth annual report of the Bureau of American Ethnology, pp. 4674544. Washington Chapman, Carl H. 1980 The archaeology 9f Missouri 1;. University of Missouri Press. Columbia. 1976 Internal settlement designs of two Mississippian Tradition ceremonial centers in southeast Miss- ouri. In C.E. Cleland (ed.), Cultural change and continuity. Academic Press. New York. Chapman, Carl H. and Leo 0. Anderson 1955 The Campbell site: a late Mississippian town site and cemetery in southeast Missouri. Missouri Archaeologist, 17. Chapman, Carl H. et a1. 1977 Investigation and comparison of two fortified Mississippian Tradition archaeological sites in southeastern Missouri. Missouri Archaeologist 38. Chapman, Robert and Klavs Randsborg 1981 Approaches to the archaeology of death. In R. Chapman, K. Randsborg and I. Kinnes (eds.), The archaeology 9: death. Cambridge University Press. Cambridge. Charles, Douglas K. and Jane E. Buikstra 1983 Archaic mortuary sites in the central Mississippi drainage: distribution, structure, and behav- ioral implications. In J. Phillips and J. Brown (eds.), Archaic hunters and gatherers in the American Midwest. Academic Press. New York. Cline, Walter 1936 Netee en the people 9: Siwah and E1 Garah,i_ the Libyan deser . Banta Publishing Company. Menasha. Cooke, George H. 1899 ‘Te Pito Te Henna, known as Rapa Nui; commonly called Easter Island, South Pacific ocean. 'Annual Leger; 0f the SmithSonian'InStitution for 1897. Washington. 244 C00per, John M. 1946 The Araucanians. Bureau of American Ethnology, Bulletin 143, vol. 2, pp. 687- 760. Washington. Culshaw, U.J. 1949' Tribal heritage: 3 study gf the Santals. Lutterworth Press. London. Cushing, Frank 1896 Outline of the Zuni creation myths.‘ Thirteenth Annual report of the Bureau of American Ethno- logy, pp. 321- 447. Washington. Davis, Hester 1966 An introduction to Parkin prehistory. ‘Arkansas Archaeologist, 7. DeLand, Charles E. 1908 The aboriginies of South Dakota, part II, the Mandan Indians. South Dakota Historical Collec- tions, vol. IV, pp. 273-748. Pierre. Delobsom, A.A. Dim 1932 L'Empire du Mogho-Naba, coutumes des Mossi de la haute-Volta. Institut de d' Ethnologie Juridiques, Vol. II. Les Editions Domat-Montchrest1en. 53815. (Translation by HRAF) Dempsey, James 1955 Mission 93 the Nile. Burns and Oates. London. Deng, Francis M. 1972 The Dinka gf the Sudan. Holt, Rinehart, and Winston. New York. Droessler, Judith 1981 Craniometry and biological distance: bio-cultural continuity and change at the Late Woodland-Miss- issippian interface. Center for American Archae- ology, Northwestern University. Evanston. Drucker, Philip 1951 The northern and central Nootkan tribes. Bureau g£_American Ethnology, Bulletin 114. Washington. Earle, Timothy 1978 Economic and social organization in a complex chiefdom: the Halelea District, Kauai, Hawaii. Anthr0pological Papers, Museum of Anthropology, University of Michigan, No. 63. “Ann Arbor. 245 Evans-Pritchard, E.E. 1971 The Azande: history and political institutions. Oxford University Press. London. 1940 The Nuer. Oxford University Press. London. Fallers, Margaret C. 1960 The eastern lacustrine Bantu (Ganda and Soga). ‘ Ethnographic Survey 9: Africa, east—central Africa, pt. 11, International African Institute. London. Fenton, William N. 1936 An outline of Seneca ceremonies at Cold Spring Longhouse. Yale University Publications 13 Anthropology, no. 9. New Haven. Firth, Raymond 1959 Social change in Tikopia. MacMillian and Com- pany. New York. 1936 We the Tikopia. Allen and Unwin Ltd. London. Forde, Daryll 1955 The Nupe. In D. Forde (ed.), Peoples of the Niger-Benue confluence. Ethnographic Survey gf Africa, western Africa, pt. 10, pp. 17-52. International African Institute. London. 1951 The Yoruba-speaking peoples of southwest Nigeria. Ethnggraphic Survey gf Africa, western Africa, pt. 4, International African Institute. London. Fortes, Meyer 1969 Kinship and the social order. Aldine. Chicago. 1945 The gynamics of clanship among the Tallensi. Oxford UniverEIty Press. London. Fortes, Meyer and E.E. Evans-Pritchard 1940 African political systems. International African Institute. London. Freeman, J. Derek 1955 Iban agriculture: a report on the shifting cul- tivation of hill rice by the Iban of Sarawak. ‘Colonial Research Studies, no. 18. London. Fried, Morton H. 1967‘ The evolution‘gf_politica1 society. Random House. New York. 246 1960 On the evolution of social stratification and the state. In 8. Diamond (ed. ), Culture and history: essays in —honor of Paul Radin. Columbia Univer— sity Press. New York. Friedman, Jonathan 1982 Catastrophe and continuity in social evolution. In C. Renfrew, M. Rowlands, and B. Seagraves TEds. ), Theory and Explanation in Archaeology: the Southampton conference. Academic Press. 1975 Tribes, states, and transformations.‘ In M. Bloch (ed. ), Marxist analyses and social antH_opology. Association of Social Anthropologists Studies No. 3. HalsEEad Press. New York. 1971 Marxism, structuralism, and vulgar materialism. Man, 9:444-469. Gayton, Anna H. . 1948 Yokuts and Western Mono ethnography. Univers1ty of California Press. Berkeley. Gearing, Fredrick O. 1960 Documentary history of the Fox project, 1948-1959. University of Chicago Press. Chicago. Gifford, Edward W. 1929 Tongan society. Bernice P. Bishop_Museum, Bulle- tin 61. Honolulu. Gillin, John 1936 The Barama River Caribs of British Guiana. Papers 9: the Peabody Museum gf American Archaeology and Ethnology, Vol. 14, no. 2. Cambridge. Gladwin, Thomas and Seymour B. Sarason 1953 Truk: man in paradise. Viking Fund Publications in Anthropology, No. 20. Wenner-Gren Foundation. Gluckman, Max 1941 Economy of the central Barotse Plain. Rhodes- Livingstone Papers 7, The Rhodes-Livingstone Institute. Livingstone. Goldman, Irving 1970 Ancient Polynesian society. University of Chicago Press. Chicago. ' 1963 The Cubeo: Indians of the northwest Amazon. Illinois Studies in Anth ropol ogy, No. 2. Univer- sity of Illinois Press. Urbana. 247 Goldstein, Lynne G. 1981 One dimensional archaeology and multidimensional people: Spatial organization and mortuary analy- sis. Ig R. Chapman, K. Randsborg, and I. Kinnes (eds.), The archaeology 9f death. Cambridge University Press. Cambridge. 1980 Mississippian mortuary practices:‘ g case study of two cemeteries in the lower Illinois valley. NSrthwestern University Archaeology Program. Evanston. 1976 Spatial structure and social organization: regional manifestations gthississippian society. Ph.D dissertation, Northwestern University. Goodenough, Ward H. 1965 Rethinking 'status' and 'role': toward a general model of the cultural organization of social re- lationships. In M. Banton (ed.), The relevance of models for social anthropology. Association of Social Anthropologists Monggraph N9. 1, pp. l-24. Gould, Richard 1978 Beyond analogy in ethnoarchaeology. £2 R. Gould (ed.), Explorations in ethnoarchaeology. Univer- sity of New Mexico Press. Albuquerque. Gurdon, Philip R.T. 1907 The Khasis. David Nutt. London. Gusinde, Martin 1931 Die Feuerland Indianer, Band 1. Anthropos. (Translation by HRAF). Gutierrez de Pineda, Virginia 1950 Organizacion social en la Guajira. Revista del Instituto Ethnologico Nacional, Vol 3, no. 2. Bogota. (Translation by HRAF). Haar, B. ter 1948 Adat law 13 Indonesia. Institute of Pacific Rela- tions. New York. Hambly, Wilfrid D. 1934 .The Ovimbundu of Angola. Field Museum of Natural History, Anthropological Series, 21(2):87—362. Chicago. Hart, C.W.M. and Arnold R. Pilling 1960 The Tiwi g: northern Australia. Henry Holt and Company. New York. 248 Hartmann, H. 1928 Some customs of the Luwo (or Nilotic Kavirondo) living in south Kavirondo. Anthropos, vol. 23, pp. 263-275. Heizer, Robert F. and John E. Mills 1952 The four ages 92 Tsurai: a documentary history gf the Indian village 92 Trinidad Bay. University of California Press. Berkeley. Higler, M. Inez 1952 Arapaho child life and its cultural background. Bureau gf American Ethnology, Bulletin 148. Washington. Hodder, Ian 1982 Symbols in action: ethnoarchaeological studies of material culture. Cambridge University Press. Cambridge. Hogbin, H. Ian 1970 The island gf menstruating men:' religion 13 Wogeo, New Guinea. Chandler Publishing Company. Scranton. 1964 A Guadalcanal society; the Kaoka speakers. Holt, Rinehart, and Winston. New York. Hollis, A.C. 1905 The Masai: their languages and folklore. Clar- endon Press. Oxford. Huffman, Ray 1931 Nuer customs and folklore. International Insti- tute of African Languages and culture. London. Huntingford, F.W.B. 1954 The political organization of the Dorobo. Anthropos, 49:123-148. Jackson, J. Brantley 1972 The Jared site: a Comanche burial at Fort Sill, Oklahoma. Plains Anthropologist 17:316-325. Jay,Robert 1969 Javanese villagers; social relations ig‘rural ‘Modjokuto. M.I.T. Press. Cambridge Jenness, Diamond 1935 The Ojibwa Indians of Parry Island; their social and religious life. 'Bulletin of the Canada 'Department gf_Mines, No. 78. Ottawa. 249 1922 The life of the Copper Eskimos. Report p: the .Canadian Arctic Expedition, 1913—1918, Vol. 12, pt. A. Acland. Ottawa. Jones, Livingston F. 1914 A study pf the Thlingets pf Alaska. Revell Com- pany. New York. Junod, Henri A. 1927 The life pf a South African tribe. MacMillian and Company. London. Karsten, Rafael 1935 The head-hunters of the western Amazonas: the life and culture of the Jibaro Indians of eastern Ecuador and Peru. Societas Scienizrum Fennica: Commentationes Humanarum Litterarium, Vol. VII, no. 1. Helsingfors. Kaut, Charles R. 1957 The Western Apache clan system: its origins and development. University g£_New Mexico Publica- tions 13 Anthropology 9. Albuquerque. Keesing, Roger 1975 Kinship and social structure. Holt, Rinehart, and Winston. New York. Kennedy, Raymond 1955 Field notes on Indonesia: Ambon and Ceram 1949- 50, Flores 1949-50. Human Relations Area Files. New Haven. Kingsley, Robert G. 1984 An analysis of the mortuary practices at the Nor— ton Mounds site. Wisconsin Archaeologist 65(2): 99-115. Kinietz, W. Vernon 1947 Chippewa village: the story of Katikitegon. Cranbrook Institute of Science, Bulletin No. 25. Bloomfield Hills. __ —— Kirchoff, Paul 1968 The principles of clanship in human society. In M. Fried (ed.), Readings ip Anthropology, Vol.—— II. Crowell. New York. Klippel, Walter 1969 The Hearnes site: a multicomponent occupation site and cemetery in the Cairo Lowland region of southeast Missouri. Missouri Archaeologist 31. 250 Kluckhohn, Clyde and Dorothea Cross Leighton 1946 The Navaho. Harvard University Press. Cambridge. Krause, Aurel 1956 The Tlingit Indians. University of Washington Press. Seattle. Kroeber, Alfred L. 1953 Handbook pf the Indians gf’California. Califor— n1a Book Company Ltd. Berkeley. 1908 Ethnology of the Gros Ventre. Anthropological Papers gfthe American Museum gngatural History, Vol. 1, pp. 141—281. New York. 1902 The Arapaho. American Museum'gf Natural History, Bulletin 18. New York. Kuper, Hidla 1963 The Swazi, a South African kingdom. Holt, Rine- hart, and Winston. New York. Lambrecht, Francis 1932 The Mayawyaw ritual, parts 1-5. Catholic Anthro- pglogical Conference, Publications, 4, no. 1-5. Washington. Landor, Arnold H. Savage 1893 Alone with the hairy Ainu. John Murray. London. Leach, Edmund R. 1954 Political systems pf highland Burma. Beacon Press. Boston. Lebzelter, Viktor 1934 Eingeborenenkulturen ip Sudwest und Sudafrika, Vol. 2. Karl W. Hiersemann. Liepzig. (Transla- tion by HRAF). Lei, Chin-liu 1944 Ancestor worship of the Lolo in Ch'eng-chiang, Yunnan. Pien-chenngung-lun (Frontier Affairs), Vol. III, no. 9, pp. 31—36. Chungking. Leith—Ross, Sylvia 1939 African women; a_study‘g£ the Igbo gf_Nigeria. Faber and Faber Ltd. London. Linton, Ralph 1933 The Tanala: a hill tribe of Madagascar. ‘Field "Museum‘gf Natural History, AnthropOlOgical Series, Vol. 22. Chicago. 251 Lloyd, Peter C. 1965 The political structure of African Kingdoms: an exploratory model. In M. Banton (ed.), Political systems and the distfibution pf power. Associa— tion of Social Anthropologists Monograph 2. Loeb, Edwin M. 1926 Pomo folkways. University of California Press. Berkeley. Lowie, Robert H. 1946 The Bororo. Bureau 9: American Ethnology, Bulle- tin 143, vol. 1, pp. 419-434. Washington. 1939 Ethnographic notes pp the Washo. University of California Press. Berkeley. 1935 The Crow Indians. Farrar and Rinehart. New York. 1924 Notes on Shoshonean ethnography.‘ American Museum pf Natural History, Anthropological Papers 20, pt. III. New York. Maes, J. 1924 Notes sur les populations des bassins du kasai, de la Lukenie, et di Lac Leopold II. Annales gp Musee gp Congo Belge, Nouvelle Serie, Miscella- nees, Vol. 1, Fasciculo 1. Brussels. (Transla- lation by HRAF). Mainfort, Robert 1979 Indian social dynamics in the period of European contact. Museum Anthropological Series, Michigan State University, Vol. 1, no. 4. East Lansing. Mangin, Eugene 1921 Les Mossi. Essai sur les pg 33 coutumes dg peuple Mossi ap Soudan occidental. Augustin Challamel. Paris. (Translation by HRAF). Marshall, Richard A. 1965 An archaeological investigation of Interstate Route 55 through New Madrid and Pemiscot Counties, Missouri. University of Missouri Highway Archae- ological Report 1. CoIEmbia. Mason, Leonard 1954 ReloCation pf the Bikini Marshallese: a study 1p ' rou migratibn. Ph.D. dissertation, Yale Uni- vers1ty. New Haven. 252 Masters, William M. 1953 Rowanduz: a Kurdish administrative and mercan- Eile center. Ph.D. dissertation, University of Michigan. Ann Arbor. McCulloch, Merran 1952 The Ovimbundu of Angola. Ethnographic Survey pf Africa, west-central Africa, pt. 2. Internation- al African Institute. London. McIlwraith, Thomas F. 1948 The Bella Coola Indians, Vol. 1. University of Toronto Press. Toronto Mead, Margaret 1930 Social organization of Manua. Bernice P. Bishop Museum, Bulletin 76. Honolulu. Metraux, Alfred 1940 Ethnology of Easter Island. Bernice P. Bishop Museum, Bulletin 160. Honolulu. Meyer, Hans 1916 Die Barundi: eine volkerkundliche studie aus Deutsch-Ostafrika. Ott Spamer. Liepzig. (Translation by HRAF). Middleton, John 1965 The Lugbara pf Uganda. Holt, Rinehart, and Win- ston. New York. 1953 The central tribes of the northeastern Bantu. Ethnoqraphic Survey pf Africa, east-central Africa, pt. 5. International African Institute. London. Milner, George R. 1984 Social and temporal implications of variation among American Bottoms Mississippian cemeteries. American Antiquity 49(3):468—488. Morgan, Lewis Henry 1901 League of the Ho-de-no-sau—nee, pp Iroguois. —-—————————.—_ Dod , Mead. New York. Morse, Dan F. and Phyllis Morse 1983 The archaeology pf the Central Mississippi Valley. Academic Press. New York. Murdock, George Peter 1967 Ethnographic Atlas. University of Pittsburgh Press. Pittsburgh. 253 1949 Social structure. The Free Press. New York. Murphy, Robert F. 1960 Headhunters' heritage. University of California Press. Berkeley. Nadel, S.F. v , 1942 A black Byzantium: the kingdom pf Nupe ip_Nig— eria. Oxford University Press. London. Nakamura, Mitsuo 1982 The cresent arises oVer the Banvan tree: a study gf_the Muhammadijah movement ip a central Javanese town. Ph.D. dissertation, Cornell Uni- versity, University Microfilms. Ann Arbor. Newcomb, William W. 1961 The Indians pf_Texas. University of Texas Press. Austin. Oberg, Kalervo 1937 The social economy pf the-Tlingit Indians. Ph.D. dissertation, University of Chicago. Chicago. Opler, Morris E. 1969 Apache Odyessey: ajou‘rney‘between two worlds. Holt, Rinehart, and Winston. New York. 1941 Ap_Apache life-way; the economic, social, and religious institutionsngthe Chiricahua Indians. University of Chicago Press. Chicago. O'Shea, John 1984 Mortuary variability: ag_archaeologica1 inves- tigation. Academic Press. New York. 1981 Social configurations and the study of mortuary practices: a case study. £2 R. Chapman, K. Randsborg, and I. Kinnes (eds.), The archaeology pf death. Cambridge University Press. Cambridge. O'Shea, John and Marek Zvelibil 1984 Olenostrovski mogilnik: reconstructing the so- cial and economic organization of prehistoric foragers in northern Russia. 'J0urna1 g: Anth- ropological Archaeology, 3:1-40. Ottenberg, Simon 1971 Leadership and authority in an African society; the Afikpo village-group. American Ethnological ‘Society, Monographs 52. University of Washington Press. Seattle. 254 Palau, Marti, Montserrat 1957 Les Dogon. Presses Universitaires.de France. Paris. (Translation by HRAF). Peebles, Christopher S. and Susan M. Kus 1977 Some archaeological correlates of ranked socie- ties. American Antiguity 42(3):421—448. Perino, Gregory 1967 The Cherry Valley mounds and Banks Mound 3. Central States' Archaeological Society, Memoir 1. 1966 The Banks Village site, Crittenden County, Arkan- sas. Missouri Archaeological Society, Memoir 4. Phillips, Philip 1970 Archaeological survey in the lower Yazoo basin, Mississippi, 1949-1955. Papers g§_the'Peabody Museum gf_Archaeology and EthnolOgy 60. Cam- bridge. Phillips, Philip, James A. Ford, and James B. Griffin 1951 Archaeological survey in the lower Mississippi valley, 1940-1947.’ Papers pf the Peabodpruseum pf Archaeology and Ethnology 25. Cambridge. Playfair, A. 1909 The Garos. David Nutt. London. Powdermaker, Hortense 1931 Mortuary rites in New Ireland. Oceania 2:26-43. Price, James E. 1978 The settlement pattern of the Powers phase. I3 B. Smith (ed.), Mississippian Settlement Pat- terns. Academic Press. New York. Price, James E. and James B. Griffin 1979 The Snodgrass site of the Powers phase of south- east Missouri. Anthropological Papers, Museum pf Anthropology: University pf Michigan, N9. 66. Ann Arbor. Radcliffe-Brown, A.R. 1922 The Andaman Islanders: a study in social change. Cambridge University Press. Cambfidge. Rappaport, Roy A. 1971a Ritual, sanctity, and cybernetics. ‘American ’Anthropologist 73:59-76. 1971b The sacred in human evolution. Annual Review pf Ecology and §y§tematics 2:23-44. 255 Rattray, Robert S. 1932 The tribes gf the Ashanti hinterland, Vol. 2. Clarendon Press. Oxf0rd. Raum, Otto Friedrich 1973 The social functions of avoidances and taboos among the Zulu. Hamburgische Museum'fur Volkerkunde, Monographien zur Volkerkunde 6. Berlin. Ray, Verne F. 1932 The Sanspoil and Nespelem: Salishan peoples of northeast Washington. University'gf Washington publications 12 Anthropology, 291. 5. Seattle. Read, Margaret 1956 The Ngoni pf Nyasaland. Oxford University Press. London. Renfrew, Colin 1973 Monuments and social organization in Neolithic Wessex. 12 C. Renfrew (ed.), The explanation pf culture change. Duckworth. London. Renfrew, Colin (ed.) 1981 Rankin , Resource, and Exchange. Cambridge Uni- ver51ty Press. Cambridge. Richards, Audrey I. 1940 The political system of the Bemba tribe - north- east Rhodesia. 12 M. Fortes and E.E. Evans-Prit- chard (eds.), African political systems. Inter- national African Institute. London. 1939 Land, Labour, and diet ip_northern Rhodesia: ap economic study pf the Bemba tribe. Oxford Uni- versity Press. London. Rivers, W.H.R. 1906 The Todas. MacMillian and Company. London. Robertson, James A. n.d. Social status and the fur trade. Unpublished man- uscript in possession of the author. Roscoe, John 1924 The Bagesu and other tribes‘gf'the Uganda Protec- torate. Cambridge University Press. Cambridge. 1911‘ The Baganda. MacMillian and Company. London. 256 Roth, H. Ling (ed.) 1892 The natives of Borneo. Journal pf the Anthropo- logical Institute gf Great Britian and Ireland, Vol. XXI, pp. 110-137. Rouse, Irving I 1948 The Carib. ‘Bureau of American Ethnology, Bulle- ~—— tin 143, vol. 4, pp. 547-565. Sahlins, Marshall D. 1972 Stone age economics. Aldine. Chicago. 1968 Tribesmen. Prentice—Hall. Englewood Cliffs. 1961 The segmentary lineage: an organization of predatory expansion. American Anthropologist, 63:322-345. 1958 Social stratification 12 Polynesia. University of Washihgton Press. Seattle. Salmon, Merrilee 1982 Philosophy and archaeology. Academic Press. New York. Salmon, Wesley C. 1984 Logic (third edition). Prentice-Hall. Englewood Cliffs. Sapir, Edward 1921 The life of a Nootka Indian. Queen'nguarterly, Vol. 28, pp. 232-243, 351-367. Kingston, Ontario. Saxe, Arthur A. 1970 Social dimensions pf mortuary practices. Ph.D. dissertation,.UfiiVersity of Michigan. University Microfilms. Ann Arbor. Saxe, Arthur A. and Patricia L. Gall 1977 Ecological determinants of mortuary practices: the Temuan of Malaysia. 1p W. Wood (ed.), 9215 tural-ecological perspectives‘gg Southeast Asia. Ohio UniVersity center for International Studiés. Athens. Schram, Louis M.J. 1957 The Monguor of the Kansu—Tibetan border, Part II. the political life. 'Transactions of the American Philosgphical Society, n.s., Vol. 47, pt. 1. Philadelphia. 257 1954 The Monguor of the Kansu-Tibetan frontier: their origin, history, and social organization. Transactions 9: the American PhilOsophical Society,‘n.s., V01. 44, pt. 1, pp. 1-138. Philadelphia. Schulze, Louis 1891 The aboriginies of the lower and middle Finke River... Transactions and Proceedings and Report of the Royal Society of South Australia, Vol. XIV, pp. 212-246. Ade—Iaide. Seligmann, C.G. and Brenda Z. Seligmann 1911 The Veddas. Cambridge University Press. Cambridge. Service, Elman R. 1978 Profiles ip_ethnology (third edition). Harper and Row. New York. 1975 Origins g: the state and civilization. Norton. New York. 1962 Primitive social organization. Random House. New York. Smith, Edwin W. and A.M. Dale 1920 The Ila-speaking pe0p1es gf_northern Rhodesia, Vol. 1. MacMillian and Company. London. Spier, Leslie 1930 Klamath ethnography. University of California Press. Berkeley. Stannus, H.S. 1922 The Wayao of Nyasaland. Harvard African Studies, N9. 3, pp. 229-372. African Department of the Peabody Museum. Cambridge. Steponaitis, Vincas 1978 Location theory and complex chiefdoms; a Missis- sippian example. In B. Smith (ed.), Mississip- pian Settlement‘Paiterns. Academic Press. New York. Stevenson, Robert F. 1968‘ Population'and political systems 13 trOpiCal Africa. Columbia University Press. New York. Stone, Doris 1962 The Talamancan tribes of Costa Rica. 'Papers of 'the Peabody Museum 92 Archaeology and Ethnology 43, no. 2. Camhridge. 258 Tainter, Joseph A. 1977a Modeling change in prehistoric social systems. In L. Binford (ed. ), For theory building in archaeology. Academic Press. New York. 1977b Woodland social change in west—central Illinois. Midcontinental journal 91 archaeology 2(1):67—98. Tax, Sol 1955 The Social organization of the Fox Indians. In F. Eggan (ed. ), Social anthropology of North American tribes. University of Chicago Press. Chicago. Thomas, Cyrus 1894 Report on the mound explorations of the Bureau of Ethnology. Twelfth Annual Report p: the Bureau 91 American Ethnology for 1890-1891. Washington. Trezenam, Edouard 1936 Notes ethnographiques sur les Tribus Fan du Moyen Ogoome. Journal de la Societe des Africanistes, yg1. 6, pp. 65-937— Paris. (Translation by HRAF) . Trilles, R.P.H. 1912 Le totemisme chez les Fan. Bibliotheque- Anthropos, Vol. 1. Aschendorffsche Verlagsbuchhandlung. Munster. (Translation by HRAF). Ucko, Peter 1969 Ethnography and archaeological interpretation of funerary remains. World Archaeology 1:262-280. Underhill, Ruth M. 1939 Social organization of the Papago Indians. Eg1e umbia University Contributions to Anthropology, 291. 30. New York. _— Vivelo, Frank R. 1977 The Herero of western Botswana. American Ethno- logical Society, Monograph 61. Voth, Henry R. 1912 Brief miscellaneous Hopi papers. Field Museum of Natural History, Publication 157, Anthropologicai Series 11, No. 2. Chicago. Warner, W. Lloyd 1937 A black civilization: a social study of an Aus- tralian tribe. Harper and Brothers. New York. 259 Watson, Patty Jo 1979 Archaeological ethnography in western Iran. Viking Fund Publications in Anthropology, No. 57. University of Arizona Press. Tuscon. Wedel, Waldo R. 1936 An introduction to Plains archaeology. Bureau 91 American Ethnologyp Bulletin 112. Washington. Wedgewood, Camilla H. 1942 Notes on the Marshall Islands. Oceania 13:1-23. Westerman, Diedrich 1912 The Shilluk people: their language and folklore. United Presbyterian Church. Philadelphia. Whiting, John 1941 Becoming a Kwoma; teaching and learning 13 a New Guinea tribe. Yale University Press. New Haven. Williams, Francis E. 1930 Orokaiva society. Oxford University Press. London. Williams, J. Raymond 1974 The Baytown phases in the Cairo Lowland of south-east Missouri. Missouri Archaeologist 36. Williams, Stephen 1977 Some ruminations gp_the current strategy pf archaeology in the Southeast. Paper presented at the Southeastern Archaeology Conference, Lafayette, Louisiana. Wilson, Monica 1957 Rituals and kinship among the Nyakyusa. Oxford University Press. London. 1951 Good company. Oxford University Press. London. IIIIIIIIIIIIIIIIII