Structure, Cultural Logic, and Transformational Dynamics in the Social Organization of Unstratified Societies: The Work of Raymond C. Kelly

Structure, Cultural Logic, and Transformational Dynamics in the Social Organization of Unstratified Societies: The Work of Raymond C. Kelly

Bruce M. Knauft, Department of Anthropology, Emory University.
Michael G. Peletz, Department of Sociology and Anthropology, Colgate University

Raymond C. Kelly’s principal contribution to anthropology has been to develop empirically based comparative frameworks and analytic models that illuminate and theorize the structural entailments, cultural logics, transformational dynamics, and evolutionary implications of social organization among unstratified societies. Whether relying on his own primary ethnographic data, synchronic or diachronic analysis of ethnographic, archival, or archeological material gathered by others, evolutionary argumentation, or a combination of these, his scholarship has been rigorous and methodical to the most exacting standards. As Marshall Sahlins (1977:vii) remarked in the forward to Kelly’s first book, “Here is anthropology in its highest form.”

Kelly’s publications have spanned the ethnography of Melanesia, Africa, and other world areas in tightly focused cross-cultural comparisons. In topical terms, his opus has ranged from kinship, gender, sexuality, inequality, and social structure, to witchcraft, cosmology, violence, warfare, and dialectical relations between cultural orientations and social and political organization. During the late 20th century, as cultural anthropologists often worked in an increasingly interpretative and impressionistic manner, Kelly has consistently and productively maintained a commitment to ethnographic, historical, and logical rigor. So too, while many cultural anthropologists redirected their attention to contemporary developments in state societies and transnational dynamics in contexts of modernity or globalization, Kelly has consistently demonstrated the broad intellectual significance of anthropology’s long-standing interest in unstratified societies. It is a fitting tribute to Kelly that in 2004 he was elected to membership in The National Academy of Sciences.

We first met Kelly during the 1974–1975 academic year. At that time Knauft was completing his undergraduate studies at Yale and had applied to a number of graduate programs, among them the University of Michigan, where Kelly was a young and untenured assistant professor. Knauft’s graduate application had included his library-based B.A. thesis on Melanesian cargo cults. When he visited Michigan prior to making his decision about graduate schools, Kelly had already read his undergraduate thesis, then met with him for an hour, and offered a range of incisive comments that helped him revise his thesis for publication in an academic journal. No professor at any of the other schools he was considering—or at Michigan—gave his work or his visit a fraction of similar attention. Knauft came to Michigan with Raymond Kelly as his graduate advisor. Looking back, he considers it one of the best professional decisions he ever made.

Peletz’s initial contacts with Kelly, though roughly contemporaneous with Knauft’s, occurred after he had begun his first year as a graduate student. Peletz had come to Michigan from the University of California, Berkeley, where he had developed interests in kinship, social structure, cultural theory, and Southeast Asia. During his second semester at Michigan he enrolled in Kelly’s course on the Social Organization of Tribal Societies (Anthropology 531), one of the core offerings taken by incoming graduate students and a scattering of brave undergraduate seniors. The course entailed careful reading and reassessment of data, arguments, and models presented in classic monographs on the Dobu and the Nuer, as well as critical reappraisal of the works of Claude Lévi-Strauss, Raymond Firth, Meyer Fortes, David Schneider, and others. It was unlike anything Peletz had ever experienced, particularly in terms of its conceptual and theoretical sophistication and rigor; it was in many ways the most intellectually demanding and rewarding course he ever took.

Also worthy of special mention is the course on research design and field methods (Anthropology 629) that Kelly overhauled and began teaching when he was first hired in 1971. He taught this course on a regular basis up until his retirement in 2004, thus providing generations of graduate students skills needed to craft successful grant applications and navigate some of the more daunting shoals of fieldwork. Both in Anthropology 629 and outside the classroom, Kelly offered students painstaking field advice concerning everything from patrol boxes to phonology, from field notes to fighting displays, from tape recorders to tapeworms, from cookware to cockroaches.

Kelly’s personal style and intellectual commitment pervaded his work with students as well as his scholarship. Over the years, he has been unflappably conscientious, systematic, and dedicated to the work at hand, effective in his comments, and wise in his counsel. In the same way that he has always taken care to clearly state, develop, and meet standards of empirical demonstration for specific ethnographic or analytic problems, he has both encouraged and expected the same in his students.

Amid this careful mentorship, one of Kelly’s greatest strengths is to take an argument on its own terms, be it his own or that of a colleague or student. Not one to preemptively elevate the assertions of a renowned colleague—or to preemptively diminish those of a student—Kelly’s trademark is to evaluate arguments through reasoned logic on the basis of the evidence presented. In this regard, he has always been careful to encourage students to develop and explore their own objectives, rather than to adopt, much less to parrot, his own. Early during Knauft’s primary fieldwork among the Gebusi of interior Papua New Guinea, he wrote to Kelly from the field and suggested that patterns of social organization he was finding might be explained by models developed by Kelly in his work with the Etoro who were just a few dozen miles distant. With great generosity, Kelly wrote back saying that Knauft should explore the information he had in terms of its own dynamics and in terms of his own research objectives and inclinations. He suggested that Knauft should not worry overmuch about similarities between the Gebusi and the Etoro—that their differences could be just as revealing. Sure enough, the inter-societal differences between Etoro and Gebusi have in subsequent years proved ethnographically salient for the comparative analytic and theoretical understanding of decentralized societies. Kelly has consistently stimulated and appreciated the systematic understanding of ethnographic differences and of controlled ethnographic comparison in the work of students and colleagues as well as in his own projects.

The qualities of mind that Kelly has brought to bear on issues discussed in the classroom, during office hours, in letters to and from the field, and of course in drafts of dissertation chapters and subsequent writing projects have inspired many of his students throughout their professional careers. To this is added his genuine willingness to engage student ideas and understandings—even when they were somewhat wide of the mark—and to take seriously their own research plans and aspirations even when they diverged from his own.

What of Kelly’s specific accomplishments and contributions? We limit ourselves to four main strands of Kelly’s research: (1) his early ethnographic work on the social structure of the Etoro of interior Papua New Guinea, which germinated in the 1960s and came to publication during the 1970s; (2) his later work on the Etoro, concerning the structure and genesis of gendered and intergenerational inequality, which also had its origins in Kelly’s initial fieldwork though it first took form in a short paper prepared for the 1974 American Anthropological Association (AAA) meetings and subsequently evolved into a 600-plus-page tome published in the early 1990s; (3) his reanalysis of Nuer warfare and expansion in Sudanic East Africa during the 19th century, which originated in a series of lectures Kelly prepared in the early 1980s; and (4) his extensive cross-cultural study of warless societies and the origin of war, which was begun in the 1990s and has continued since. The metacomment here is that Kelly’s commitments to anthropological empiricism, analysis, and theoretical advancement have never flagged; rather, they have intensified in scale and in scope during the past four decades. They have also resulted in the publication of towering books on a seven-to-eight-year cycle: 1977 (eight years after his primary fieldwork), 1985, 1993, and 2000.
From Chicago to Ann Arbor to New Guinea: the genesis of Etoro Social Structure

A career as an anthropologist was not at all what young Kelly had in mind in the early 1960s when he began his undergraduate studies at the University of Chicago, where he continued a venerable family tradition; his parents, uncles, and grandparents on both sides had all attended Chicago. The family plan was for Kelly to major in economics. However, he found the subject to be “a crashing bore.” He was predisposed in a general way to the explanation of human diversity by the social rationalism of his father, for whom human differences had reasonable purposes and functions, and by the diversity of childhood living environments that the young Kelly experienced. As his father took a series of upwardly mobile jobs, the Kelly family moved across the eastern and mid-western U.S., exposing Ray to students and lifestyles across a variety of different schools. At the University of Chicago, he took an introductory course in anthropology taught by Seth Leacock, a specialist in the social organization and religion of native South Americans. Although the course was engaging, Kelly’s initial experiences at the University were so mixed that he dropped out and headed west to live with friends, try his hand at various blue-collar jobs, and otherwise experience a version of “the good life”—a period Kelly remembers largely for its ample quantity of beer and tacos.

Roughly a year later, he made his way back to the University of Chicago, where he sampled courses in a wide variety of disciplines. He eventually enrolled in another of Leacock’s courses and others taught by members of the Department of Anthropology, including Sol Tax and Lewis Binford. The charismatic Binford was especially influential in shaping Kelly’s decision to pursue graduate work in anthropology—a decision about which his parents were dubious. Partly because Binford had earned his Ph.D. at the University of Michigan and was well acquainted with its program, its strengths in cultural ecology and cultural adaptation, and its faculty (which then included Marshall Sahlins, Elman Service, Roy Rappaport, and Eric Wolf), he both encouraged Kelly to apply to Michigan and, Kelly surmises, facilitated his acceptance into the program despite his “late bloomer” grade point average.

Kelly enrolled in the graduate program in Anthropology at the University of Michigan in 1965. Lacking a fellowship, he paid tuition and supported himself through a combination of loans and savings from summer jobs driving city buses in Chicago as well as trucks in the fleet of the Washtenaw Lumber Company, and a $1,000 graduation gift from his parents. Wanting to reduce his semesters of paid tuition, he practically doubled his course load so he could complete his M.A. requirements during this first year. Taking courses with Sahlins and Rappaport, who eventually served as co-chairs of his doctoral committee, and various other faculty members, he proceeded quickly to the “prelims” phase of the program. Before taking his preliminary exams, he wrote a pair of influential essays. One of these, on generalized exchange among the Dobu, he sent to Sahlins, who was in France at the time. The other, on demographic pressure and descent group structure in the New Guinea Highlands, was a course paper written for Sahlins and Mervyn Meggitt. Sahlins was so taken with Kelly’s incisive reanalysis of Reo Fortune’s data on Dobuan marriage and exchange that he shared the essay with Lévi-Strauss, who suggested that it be translated into French and submitted to the prestigious French journal L’Homme, where it was published in 1968 (Kelly 1968a). The other paper so impressed Meggitt that he brought it to the attention of the editor of the venerable journal Oceania, which published it as a lengthy article, also in 1968 (Kelly 1968b). The publication of these influential essays, one leaning toward French structuralism and the other considering demographic determinism in relation to British structural functionalism, helped establish Kelly’s reputation for his command of ethnography, analytical prowess, and theoretical insight, even before he went to the field.
Adventures in the field and academe

His fieldwork among the Etoro of interior Papua New Guinea was both personally rewarding and physically and logistically difficult. No effective translators could be found; the Etoro language had to be learned monolingually. It helped that he was accompanied by his wife, Mary, who had a B.A. in linguistics. The Etoro had barely been contacted by Australian patrols, and the way officials managed Kelly’s arrival in the area made some among the local population resentful. Kelly’s experiences strongly suggest both that he was poisoned and that his house was burned down by disgruntled Etoro who wanted him to leave. Upon the latter event, however, many Etoro touchingly supported him. Hobbling back to the patrol post, his shoes and boots having been burned in the fire, Kelly determinedly arranged to restock his most basic provisions and equipment. When he returned and resumed his research, the Etoro were amazed by his commitment to the community, and increasing numbers accepted and befriended him. He collected a large amount of data, first concerning ecology and then concerning social organization, but had trouble making sense of it. Indeed, he remembers “feeling when I got back that my fieldwork had been a flop, that I hadn’t understood things very well.”

After Kelly returned to Michigan, the Department of Anthropology found itself unsatisfied with candidates who had applied for their faculty position in social organization—and so turned to Kelly. He never wrote an application for the job, was not interviewed, and never gave a job talk! Indeed, as he was committed to writing up his thesis before starting to teach, he had to be pressured to take the job. He began teaching in 1971, six years after he initially enrolled as a graduate student. In writing his thesis, Kelly recalls the key point at which he realized that different widow remarriages had different structural implications. “I realized there were all these widows remarried in different ways. What was the logic of it? There were three forms of remarriage. It was methodologically like a twin study: Why were some of them remarried in one way as opposed to others? They were socially engineered in different ways. As I analyzed the data, things came together in my mind.”

Given Kelly’s teaching load, however, and his own commitment to teaching, he found it difficult over two years (1971–1973) to finish his thesis. Finally, he requested a semester of leave. His request being denied, Kelly announced he would quit so he could finish his doctorate. The department chair, James Griffin, finally relented and granted Kelly a leave. In 1974 he quickly finished and defended his thesis, which was accepted and published by the University of Michigan Press with virtually no changes.

Etoro Social Structure appeared in 1977. At the time, ethnographers of New Guinea typically drew upon one or another of three approaches that also informed work conducted by Anglo-American anthropologists in other world areas: structural-functionalism, cultural ecology, and symbolic anthropology. Kelly’s first impulse in the remote region of the Great Papuan Plateau that the Etoro called home was to pursue a demographic and ecological analysis. But he determined empirically after several months that the Etoro actually had an abundance of food, including protein, and that demographic and ecological necessities did not appear as primary constraints on or causes of Etoro cultural orientations or social organization. The Etoro lived in a mid-altitude habitat from which a diverse variety of floral and fauna resources could be obtained through a combination of hunting, foraging, rudimentary horticulture, and the rearing of semi-domesticated pigs. Moreover, influenza and measles epidemics had reduced the Etoro population such that the survivors tended to enjoy an environment relatively abundant in land and resources.

Turning instead to kinship and social structure, in which he had received early training from Sahlins and Meggitt, Kelly proceeded to collect information for what is the most complete and rigorous published analysis of social organization for a New Guinea society and quite possibly, though it is beyond our competence to say this with certainty, for the entire ethnographic record of pre-state societies. Developing a concept of structural contradiction that, as Kelly notes, “has roots in Sahlins’ writings,” and deploying it in the context of a “detailed analysis of social process” that “draws on Rappaport’s cybernetic approach,” albeit one “focused on social rather than ecological variables” (1977:xii–xiv), his exquisite combination of fine-grained documentation, analytic brilliance, and sweeping theoretical implications surpassed even classic monumental ethnographies, such as Mervyn Meggitt’s 1965 The Lineage System of the Mae-Enga of New Guinea, and Fortes’s tomes on clan dynamics and webs of kinship among the Tallensi (1945, 1949). Fortes himself acknowledged practically as much in his late-career 1978 essay inthe Annual Review of Anthropology, as discussed below.

Based on 15 months of fieldwork (1968–1969), Etoro Social Structure addresses a central question of kinship and social organization: Do culturally adduced principles such as agnation (e.g., patrilineal descent) always determine behavior, or are some social structures “loose?” That is, are they refractory to the cognized models of anthropologists and informants alike, and so admit idiosyncratic behavioral variance? Kelly illustrated, at least for the Etoro, that while models apotheosizing single ostensibly hegemonic principles such as descent were invariably doomed to failure, one could nonetheless construct alternative models with powerful potential to explain social practice “on the ground.” Arguing that Etoro social behavior is best understood through the interplay of mutually contradictory principles, Kelly demonstrated that two analytically distinct and potentially contradictory principles inform unit definition and unit interrelationship. For example, descent unit recruitment and alignment are both subsumed under the anthropological notion of “descent.” The vertical principle of (patrilineal) descent, which Kelly thus deconstructed, ideally governs recruitment to the basic units, the Etoro patrilineages; the horizontal principle of siblingship determines alignments among units. Structural distance is thus legitimately calculated across opposite-sex sibling bonds; for instance, a man’s sisters’ sons and their descendants may be incorporated into, or co-aligned with, his lineage and residential group.

Kelly encouraged an analytic formulation of siblingship relatedness and suggested that principles of siblingship—including the symbols, idioms, and metaphors keyed to them—were commonly invoked in societies whenever individuals, groups, or salient social categories were deemed to have equivalent, parallel, or essentially complementary rights, obligations, and experiences with respect to a specific territorial domain, political office, or other mediating element. This formulation has had broad significance in Melanesia and elsewhere in Oceania as well as in Southeast Asia, lowland South America, and beyond (see, for example, Kensinger 1985; Marshall 1981; Peletz 1988; Smith 1983).

More generally, the purportedly unitary domain of descent, and Etoro calculations of structural distance in particular, are shown to be ordered by two analytically separate principles, behavioral conformity to either of which may—and in many contexts clearly does—result in violations of the other. Manifestations of such violations, and of the structural contradictions that underlie them, are evident in, but by no means confined to, the internal organization and residential configurations of Etoro lineages. They also appear in processes of fission that demarcate lineage boundaries, in the prevalence of matrilateral siblingship in interlineage alignment, in the regulations and other circumstances surrounding the disposition of rights over widowed women, and in various features of kinship terminology and local mythology.

Kelly also demonstrated that Etoro social organization harbored a de facto moiety system that existed in behavior patterns “on the ground” even though it was not notionally recognized and did not appear to be consciously cognized by Etoro themselves. More broadly, Kelly argued provocatively that Etoro social structure was prescriptive in an “ex post facto” sense: all structurally significant behavioral variation was consistent with one or the other fundamental principle of the social structure.

Kelly sheds light on many contexts in Etoro society in which the entailments of principles of siblingship and descent are mutually incompatible, simultaneously emphasizing a more general point that is reminiscent of perspectives developed by the Manchester school of British social anthropology, particularly by Max Gluckman and the early Victor Turner (1957), as well as the more explicitly Marxist work of Maurice Bloch (1989), namely, that “the organization of contradictions is the essence of social structure.” His analytic model of structural contradiction applies as well to the classic disjunction between structure and behavior known as “the Nuer paradox” (Kelly 1977:3, 284–298). The scope and force of many of the contradictions Kelly examines turn on the conflicting imperatives of different categories of siblingship and descent, perhaps because among the Etoro and the Nuer, principles and idioms of brotherhood receive far greater cultural elaboration than do constructs of cross-sex siblingship and sisterhood. Some reviewers have suggested that by focusing on brotherhood as the dominant form of siblingship Kelly at times reduces siblingship in general to brotherhood specifically and thus glosses over critically important distinctions between the variants of siblingship that obtain among sisters and cross-sex siblings. This focus can lead to an insufficient appreciation of the various contexts in which the entailments of cross-sex siblingship and sisterhood may complement and reinforce, or, alternatively, compromise the moral imperatives of descent, as Smith (1983) and Peletz (1988) have shown for regions of Micronesia and Southeast Asia, respectively. However, caveats of the latter sort arise from studies conducted in different cultural regions. Kelly’s analysis appears cogent for the Etoro themselves; he documents, for example, that bonds between sisters are key for the construction of parallel marriage, by virtue of which men who marry females related to each other as sisters become “brothers” to each other, and this irrespective of the ways they may have been previously related through clan or lineage ties.

Kelly’s arguments that in some ethnographic contexts modalities of siblingship merit at least as much analytic consideration as principles of descent and alliance expanded the purview of a generation of scholars embroiled in increasingly fruitless debates concerning the pros and cons of theories based on descent versus those based on alliance. Partly for these reasons, they found themselves beset with what David Schneider (1965) famously characterized as “muddles in the models.” It warrants mention too that while Lévi-Strauss and his students, along with the Dutch structuralists who inspired them, are widely credited with developing alliance theory and with thus helping shatter the descent-based hegemony that dominated earlier studies of kinship and social structure in tribal societies, these alternatives appear much less radically opposed when viewed in relation to Kelly’s work. Indeed, Lévi-Strauss, like his descent-oriented forebears, focused largely on unilineally bounded groups and their external relations and reproduction through time. In short, descent-based units still constituted the fons et origo and ultimate loci of investigation. Though alliance through marriage was emphasized, little analytic provision was made for other structural principles, such as siblingship, which have important implications for the presence or absence of bounded groups and which inform a broad array of social relations and myriad domains of cultural order.

In his remarks on a panel devoted to sibling relations in lowland South America convened at the AAA meetings in 1983, Kelly provided additional insights on the value of taking sibling ties as one’s point of departure for observation and analysis:

Every analysis must begin somewhere, and the point of departure elected generally shapes the analysis that follows in very important ways. The pivotal position of siblingship as a major point of articulation between key symbols and core social relations provides both an entrée to two important levels of analysis and a strong interconnection between them (one that works against the analytic dismemberment of a coherent whole and the resultant concern with how the pieces fit together). [1985b:42]

These comments enjoin ethnologists to direct their analytic gaze on relations among the living rather than links between the living and the dead and/or the unborn; they likewise encourage a focus on bonds among individuals in the same relative generation. Such bonds constitute an important basis for daily social interactions in most societies. By urging greater attention to ways in which siblingship serves as a key symbol and central organizing principle in intragenerational relations, Kelly helped reorient and revitalize the study of kinship, which had been characterized by a largely unquestioned focus on intergenerational links since its very inception.

In the late 1970s when Meyer Fortes was invited by the editors of the Annual Review of Anthropology to summarize his life’s work and his views concerning the most productive directions for the future of anthropological research, he devoted the final three pages of his comments to Etoro Social Structure. Fortes characterizes the latter volume as a “remarkable recent study” that “covers every aspect of the social structure in almost microscopic detail,” doing so “with exemplary rigor of method and sophistication of theory” and “brilliant and conclusive development of ... [its] thesis” (1978:25–27). It is a great tribute to Kelly that Fortes singled out this newly published work by a freshly tenured associate professor in such approving terms, especially since Fortes could easily have highlighted his own accomplishments or those of one of his own many students. That Etoro Social Structure received praise from so many quarters (recall Sahlins’s accolade, quoted earlier) underscores this book as an enduring work of broad appeal, embodying the highest standards of scholarship.
Commitment to ethnographic and logical rigor

By emphasizing ethnographic rigor, logical deduction, and reasoned argument with powerful theoretical implications, Kelly instilled in his students an abiding respect for the relevance and explanatory power of empirical evidence. During an era in which anthropologists often produced evocative but impressionistic analyses, we were shown the value of systematically collecting and using empirical data and taking seriously the all-too-easy-to-sidestep larger patterns of what people actually do. We were encouraged to make sense of normative stories and evocative vignettes, to gather a systematic corpus of behavioral information that would reveal statistical patterns, and to discern these patterns in broadly conceived and theoretically relevant ways. The specific data set and issues to address were not predetermined. Genealogies, censuses and household surveys, patterns of residence and property transmission, case histories of various kinds, the content and structuring of narratives and other verbal renditions and reports could all provide key information.

In Knauft’s case, a desire to simultaneously pursue a symbolic analysis of cultural phenomena and a sociological analysis of social organizational led to the surprising discovery that Gebusi spirit mediumship and sorcery accusations together generated one of the highest rates of homicide documented in the cross-cultural record. For Peletz, who worked among “matrilineal” Malay Muslims in the state of Negeri Sembilan, Malaysia, a concern to investigate contemporary patterns of kinship and property relations, as well as their historical development during the period 1830–1980, yielded insights that had eluded previous researchers, the majority of whom engaged in purely synchronic analyses. Because such analyses tended to conflate distinctions between initial proprietorship and inheritance on the one hand, and long-term trends in the transmission of property rights on the other, they erroneously concluded that the system of matrilineal inheritance had completely “broken down” and given way to a system characterized by one or another degree of patriliny or bilaterality. Other such examples could be cited in the work of Kelly’s other students. The larger point is that during an era in which the empirics of social documentation were being eschewed by many anthropologists, Kelly’s commitment to this enterprise was all the more critically important, not at the expense of larger analysis or theory, but in the service of it.

In this as well as other respects, Kelly has been a doer rather than a preacher. Unlike many renowned academics, he has always disdained self-advertisement; he has been generous in acknowledging his intellectual debts and even-handed and temperate in responding to his critics. Even in personal interview, he attributes much of his early trajectory to luck and circumstance. More generally, “I’ve been known to say that the most important thing to remember about the past is that it’s over. So I never really think about it.” In the academic present, Kelly is not one to pontificate, much less to browbeat a student. He reads work carefully and determines what kinds of available information the student should marshal to adjudicate the propositions he or she would like to explore. By temperament, commitment, and dint of professional duty, Kelly has been as calm, unflappable, and consistent as he has been relentless.
Constructing Inequality: The Fabrication of a Hierarchy of Virtue among the Etoro

Kelly’s second monograph concerning the Etoro, Constructing Inequality, was published in 1993. While this came nearly a decade after his book on Nuer warfare, considered below, it is appropriate on empirical and thematic grounds to first discuss this later book and its implications. Constructing Inequality was conceived at a time when feminist ethnographers and theoreticians of various stripes had garnered well-deserved attention in anthropology and beyond. One consequence of their interventions was the revitalization and reconfiguration of kinship studies through a combination of feminist and Marxist perspectives. Many scholars working in this area sought to develop a coherent approach to the study of kinship and gender that focused on axes of gendered difference and associated inequalities. Both were held to be reflected in, buttressed by, legitimated—and sometimes contested through—systems of symbols and meanings.

In this intellectual climate, Kelly turned his attention back to the broader implications of his earlier work on the Etoro for understanding social and cultural inequality. More specifically, he analyzed cultural suppositions that attended Etoro notions of prestige, stigma, and virtue. Further, he articulated these with a detailed consideration of inequality as evident in Etoro gender relations and sexual practices, witchcraft, spirit mediums, the efforts and prerogatives of senior men, the gendered division of labor, the production and circulation of wealth, and the Etoro life cycle in general. The monumental result was Constructing Inequality: The Fabrication of a Hierarchy of Virtue among the Etoro.

As the title indicates, Kelly documented that relations of inequality, especially between Etoro men and women on the one hand, and older Etoro and younger ones on the other, were culturally constructed; they were not a product of sociodemographic or ecological conditions. Kelly suggested, moreover, that the existence of inequality, in particular, between the prestige that was culturally accorded to Etoro men as distinct from Etoro women, did not stem from physical domination or labor exploitation of women; men did not enjoy prestige on the basis of women’s work or through threat of violence but on the basis of their own physical effort and the prestige accorded to those efforts by the division of value ascribed to labor. As such, he suggested, the cultural configuration of virtue among the Etoro, including the privileging of male sexual substance, physical prowess, and the results of male endeavors, was earned and did not involve the exploitation of women. In a provocative argument, he further suggested that through exchange of male and female products predicated on the division of labor, Etoro women obtained items of greater material value from men than those they gave to them, including nutritional sustenance (e.g., 1993:498–499). In these terms, he suggested, Etoro women were actually advantaged, not disadvantaged, in relation to men. In an argument at turns orthogonal and contrary to feminist assertions, Kelly showed that Etoro cultural dimensions of gendered inequality were at least disjoined from economic ones, if not actually oppositional to them, despite the fact that the Etoro constituted a classic “brideservice society” (see below). This refutation of common feminist assumptions implicitly applied to a wide range of other unstratified societies.

Kelly also stressed that the empirics of cultural prestige or stigma in relation to economic effort or exploitation needed to be considered carefully in each case; they could not be reduced to a priori assumptions about male dominance, the exploitation of women, or presumed consistency between cultural prestige and the appropriation of labor. In the mix, Kelly supplied an enormously rich ethnography-in-the-round concerning many aspects of Etoro social, spiritual, economic, sexual, and gendered life.
In dialogue with feminism

To better convey both the specifics and broad implications of Kelly’s arguments, a brief excursus on the important scholarship of Jane Collier is in order, particularly since Constructing Inequality is offered as a systematic engagement with and critique of Collier’s work. Collier synthesizes certain Marxist and feminist perspectives in the course of presenting three ideal-typical models for analyzing the loci and entailments of gender and generational inequality among nineteenth-century Great Plains groups and other societies that lack stratification (bands, tribes, and ranked societies, but not chiefdoms). These were especially the “brideservice model,” developed in relation to Commanche data; the “equal bridewealth model,” constructed largely on the basis of Cheyenne data; and the “unequal bridewealth model,” heavily informed by Kiowa data. In each model, Collier systematically relates inequality to a particular form of validating marriage (e.g., performing brideservice, or paying equal or unequal bridewealth); she systematically links the latter with numerous other variables. These include production and circulation (the meaning of work, the content of statuses, the meaning of gifts), political processes (the causes of conflicts, the nature of leadership, folk models of social structure and human agency, practical action), and cultural representations (gender conceptions, rituals, etc.). Collier reformulates the materialist Marxism of Claude Meillasoux (1984)in light of contemporary feminist concerns while also committing herself to the more actor-oriented practice theory of Bourdieu. As such, she contributes to a reconfiguration of gender and kinship that bears centrally on issues of coercion, hegemony, and “misrecognition.”

Thus, Constructing Inequality becomes a key text in the reconstituted field of kinship and gender studies. It demonstrates with scrupulous attention to ethnographic detail and analytic logic that Collier’s model is undercut by misplaced causality; it focuses on marriage as the principal locus or generator of inequality in classless, especially brideservice, societies. Kelly’s objective is not merely to show how data from Etoro and other adjacent societies are out of keeping with Collier’s brideservice model, which cites Etoro and related groups as candidates for inclusion, but to develop an alternative model.

Defining social inequality as “social differentiation accompanied by differential moral evaluation” (Kelly 1993:4), Kelly emphasizes that stigma is the negative reciprocal of prestige; that analytic discussions of so-called prestige systems would benefit from greater terminological and conceptual precision; and that such systems are more accurately characterized as systems of prestige/stigma. Kelly’s goal is to describe and analyze all significant social inequalities among Etoro and related groups, including in particular inequalities that are not either “organized by marriage or derived from the means and relations of production” (1993:4). In realizing this objective, Kelly highlights how a shamanic elite plus senior men—along with initiation ceremonies, witchcraft beliefs, and associated cosmologies—are all implicated in the production of inequality. He also illustrates that social inequalities are entailed in marriage and in the division of labor, but that in the latter cases “the differential moral evaluation that engenders this is derived from the cosmological system rather than from economic processes per se” (Kelly 1993:9). This is to say that “both the means and relations of the production and allocation of prestige and moral superiority are...linked to the perpetuation of life across generations” (Kelly 1993:11); and, more broadly, that

[T]he gender asymmetry that Collier points to...turns on the differential age of marriage for males and females; marriage corresponds to the age of transition to an adult level of production for women but not for men...The orientating proposition that marriage organizes inequality collapses analytically pertinent distinctions and lumps together a number of independently variable causal factors as a single variable assessed in terms of presence or absence (i.e., married versus unmarried). [Kelly 1993:437]

Despite their divergent emphases, Kelly and Collier both call for a radical recontextualization of kinship and gender. And they both demonstrate that systems of kinship and gender are “about” difference and inequality, and, as such, are most usefully analyzed as components of more encompassing systems of distinction and hierarchy that are variably grounded in cosmology and political economy. These studies reinforce the theme that feminist and Marxist concerns alike lead back to kinship (or social relations). Indeed, Kelly and Collier both share with Pierre Bourdieu (1977), Jack Goody (1990), and numerous Marxists and feminists the focus on kinship as an array of key social relations that engage production and reproduction and that remain squarely embedded in practice.
In dialogue with kinship studies

Another important feature of Kelly’s 1993 volume seems to have been more widely appreciated in British, continental European, and Australian anthropology than in American circles, due perhaps to the lesser influence outside the U.S. of David Schneider’s work, in particular his insistence that kinship, like culture generally, is most productively approached “purely as a symbolic system, the catch phrase is, ‘in its own terms’” (Geertz 1973:17 cited in Schneider 1980:125). This feature of Kelly’s work concerns his analytic decision to conceptualize kinship as a domain of social relations (albeit social relations of a certain type, as noted below) rather than “purely as a symbolic system” à la Schneider. An advantage to this approach is that it offers a way out of the conceptual and analytic cul-de-sac caused by accepting the comparative and theoretical implications of Schneider’s oft-cited insistence that on purely cultural grounds, “there is no such thing as kinship,” which as many scholars have noted, is tantamount to completely “abandon[ing] the use of kinship as an analytic category” (Holy 1996:165; cf. Peletz 1995). In Kelly’s words:

Kinship relations are social relations predicated upon cultural conceptions that specify the processes by which an individual comes into being and develops into a complete (i.e., mature) social person. These processes encompass the acquisition and transformation of both spiritual and corporeal components of being. Sexual reproduction and the formulation of paternal and maternal contributions are an important component of, but are not coextensive with, the relevant processes. This is due to the ethnographic fact that a full complement of spiritual components is never derived exclusively from the parents...Foods may also constitute essential ingredients in the spiritual or corporeal completion of personhood... [And] maturation frequently entails...replacing, adding, and/or supplanting spiritual and corporeal components of personhood. There is no analytic utility in artificially restricting the category of kinship relations to relations predicated on some but not all of the constitutive processes of personhood because these processes are culturally formulated as components of an integrated system and the social relations they predicate are all of the same logical type, i.e., relations of shared substance or shared spirit. [1993:521–522]

One of many European scholars who find this conceptualization useful is Ladislav Holy. He concludes his 1996 book on the history and current state of kinship studies with a discussion of Kelly’s work, and a consideration of this extended quote in particular. Holy echoes Kelly’s position that kinship should be approached analytically as a domain of social relations keyed to “culturally specific notion[s] of relatedness deriving from shared bodily and/or spiritual substance and its transmission” (Holy 1996:171).
An objectivist anthropology

An additional, largely unrelated set of issues worth mentioning here bears on the rather formal, “objectivist” writing style characteristic of Constructing Inequality and of Kelly’s oeuvre generally. This style is consistent with and underscores Kelly’s disinclination, both in his writings and in the classroom, to recount emotional and other personal experiences except as they are directly relevant to the data and the objective analysis at hand; he rigidly eschews the a priori privilege of speaking from, or having to justify oneself on the basis of, one or another subject position. He retains an unwavering commitment to what others would call Weberian judgments of fact as opposed to judgments of value. He has been similarly committed to data-driven arguments, model building, and the relation between empirics and formal theoretical analysis. Taken together, these qualities might lead some scholars, especially younger ones, to the erroneous conclusion that Kelly’s scholarship is dissonant with broader concerns in contemporary anthropological analysis and theory.

Kelly’s work does not promote the view expressed in various streams of feminist, post-colonial, and subaltern scholarship that anthropology has a moral imperative to advocate on behalf of the powerless and that “anthropology that doesn’t break your heart just isn’t worth doing anymore” (Behar 1996:177). However, from the very beginning of his professional career Kelly has evinced a sharply focused commitment, both in his writing and in the classroom, to systematically describe and analyze axes of difference and inequality along with myriad other issues of vital importance to feminist anthropologists and others concerned with diverse modalities of inequality. His early work and subsequent writings on non-heteronormative sexualities among the Etoro and other Melanesian groups have had an enduring impact on those involved in pioneering the study of alternative sexualities in anthropology and fields such as sexuality studies, gay and lesbian studies, and queer theory. Kelly’s remarks about the differences in normative homosexuality between the Etoro and the neighboring Kaluli and Onabasulu people (1977:16) have been very widely cited in anthropology textbooks and in introductory courses. Gayle Rubin began her graduate studies in anthropology at the University of Michigan in the early 1970s, during which time she took courses on Melanesia with Kelly. She went on to produce a series of articles that remain among the most important texts written by any anthropologist on the subject of sexuality (Rubin 1975, 1984). She recently noted that

Ray Kelly had an incalculable impact on my development as an anthropologist, and he particularly helped shape my intellectual sensibilities about gender and sexuality. In 1973 I took his course on peoples and cultures of the Pacific. The course was mostly on New Guinea, and provided an extended tour of the astonishing variability of human sexual arrangements. I was already wrestling with issues of sexuality and gender, so I paid especially close attention to those aspects of the course contents. These were incredibly rich, and provided me with a fund of exemplary material from which I have drawn ever since. A great deal of that material could have been sensationalized, or, as was more typical at the time, avoided altogether. Ray’s descriptions of multiple intercourse, ritual semen collection, and homosexual initiation were completely matter of fact and utterly disarming. I learned the dizzying details of marriage customs of the Banaro, the Marind Anim, and of course the Etoro, as well as the multitudinous forms of institutionalized sodomy so extensively documented for New Guinea. Much of what I learned from Ray ended up in two papers, “The Traffic in Women” [1975] and “Thinking Sex” [1984]. Ray was also inspirational in his implacable, unflinching willingness to track down the implications of an analysis to their logical termination. His example helped consolidate my conviction that sexuality is, to use Polanyi’s terminology, “an instituted process.” In those days it went against the grain to argue that sexual practice is a product of different cultural logics and institutional structures. Ray was a formative influence, and I owe him an immeasurable debt. [Gayle Rubin, personal communication, August 2005]

As a scholar strongly committed to facts, Kelly has thrown into question a range of theoretical pretensions, intellectual hegemonies, and a priori biases by taking seriously the particulars of what people actually do, as well as what they say they do. In this regard, Constructing Inequality has become a benchmark in the cultural and sociological understanding of social and gendered inequality in pre-state societies. So far as we are aware, scholars working in feminist, Marxist, and other traditions have been unable to refute Kelly’s empirical evidence, though some critics have raised questions about the conceptual underpinnings and assumptions of his work. These include how cultural and economic phenomena are categorically configured and analytically distinguished, and the nominal or notional framework through which these are then connected logically and causally in his analysis. Even granting some validity to this criticism, as an overall ethnographic contribution Kelly’s 1993 magnum opus on the Etoro is unsurpassed. Constructing Inequality appropriately received the University of Michigan book award in 1994.
The Nuer Conquest: The Structure and Development of an Expansionist System

As noted earlier, upon completion of Etoro Social Structure, and before turning to Constructing Inequality, Kelly pursued a book-length project devoted to the Nuer conquest. Some of the basic analytic objectives of this project were forecast in an important article on the goals, explanatory power, and limitations of functional arguments that Kelly co-authored with Roy Rappaport in 1975, published in the first issue of this journal. The war-associated expansion at the center of the book’s analysis entailed the four-fold territorial increase of the Nuer at the expense of the Dinka and the Anuak in the southern Sudan from about 1818 to 1890. In examining the reasons for this expansion, Kelly re-tilled intellectual ground made ethnographically famous not only by E. E. Evans-Pritchard’s work concerning the Nuer (1940, 1951) and Godfrey Lienhardt’s concerning the Dinka (1961), but also by a subsequent swirl of analyses and controversies surrounding the cultural ecology of Nuer expansion. This included Sahlins’s paper on the “segmentary lineage as an organization of predatory expansion” (1961).

Against this background, Kelly designed a program of archival research and library-based study that took him to the Sudan and Great Britain in the spring of 1979. In university libraries and government archives in Khartoum, and at the Public Records Office, the British Museum, and other repositories of archival material in Great Britain, Kelly sifted through all the pertinent information scattered in published and unpublished sources, and discovered a striking pattern: the pronounced expansion of the Nuer at the expense of the Dinka, perhaps the most prominent instance of “tribal imperialism” in the ethnographic or archeological record, could not be explained by population pressure, protein scarcity, caloric need, or other commonly adduced variables. The case was especially poignant since the Dinka and the Nuer apparently developed from common cultural as well as demographic stock; their distinct ethnic identities developed only shortly before and during the period of Nuer expansion itself.

What made, and still makes, the Nuer case intellectually intriguing is that virtually all of the work conducted within then current ecological and cybernetic paradigms focused on self-regulating mechanisms of homeostatic systems, a point well illustrated in Rappaport’s masterful publications on the Tsembaga Maring (1967, 1979), which Kelly justly characterizes as containing “the most sophisticated and comprehensive formulation of concepts that are broadly applied within the field of ecological anthropology” (1985:237). Yet here was a case where central features of the system worked against homeostasis insofar as they “engendered a predisposition to unremitting territorial expansion” (1985:226). Nuer data call into question Rappaport’s fundamental dictum that “self-regulation is intrinsic to adaptation and therefore applicable, in principle, to every ethnographic case” (1985:236). Such data bear all the more significance in light of the fact that Nuer expansion is a “striking example of evolutionary success” (1985:1).

Kelly meticulously analyses the ecological, demographic, economic, and military features of Nuer expansion, including the role of savannah pastoralism and cattle demographics as well as the salient contrasts between the societies in the structure of kinship and marriage relations and social organization as a whole. At the risk of oversimplification, Kelly demonstrates that non-socio-cultural factors could not have been prime movers of Nuer conquest. Among other things, their military successes provided them so much extra land and cattle that expansion driven primarily by needs for these resources would have been curtailed if not reversed by the Dinka themselves, who were much more numerous and lived in more densely populated settlements than the Nuer.

The salient factors adduced by Kelly are practices and symbolics associated with the payment of bridewealth, particularly the Nuer’s strong and unwavering insistence that marriage payments from the groom’s kin to the bride’s include ideally at least 32–40 head of cattle (for qualifications, see 1985:230 et passim). This bridewealth is much larger than that expected by Dinka. Among the Nuer as well as the Dinka, bridewealth is a key symbol in Ortner’s (1973) sense. However, among the Nuer, but not the Dinka, bridewealth was subject to considerable inflation that “often involv[ed] intrinsically unobtainable cultural objectives” (1985:240). Thus, its significance ranged far beyond the “symbolic.” Indeed, bridewealth simultaneously provided a critical impetus for the organization of economic production and political mobilization for purposes of feuding and warfare. These and other factors “prevented [both] the devaluation of bridewealth” as a core symbol and “its displacement from a key position in the symbolic system” (1985:235).

Kelly shows how the combined cultural elaboration and social and material entailments of contrasts in bridewealth between the Nuer and the Dinka engaged a wide range of positive feedback features, including, as noted earlier, the use of segmentary lineage politics to mobilize larger numbers of Nuer warriors in raids and pitched battle against the Dinka. The result was a process that Gregory Bateson (1958) referred to as complementary schismogenesis: the Nuer and the Dinka became increasingly distinct in terms of ethnic markers and the structure and organization of their socio-political systems. In the process, the larger Dinka population was progressively displaced by the expanding Nuer. Of broader significance still, “[T]he Nuer case suggests that models of adaptive structure and regulatory process drawn from cybernetics and evolutionary biology may fail to bring out the distinctive pattern of relationships between relationships that obtain within socio-cultural systems” (1985:240–241).

Not surprisingly, Kelly’s provocative challenges to received wisdom attracted a number of materialist responses and rebuttals. Though we have not followed all of these debates in detail, we believe there is general agreement that Kelly’s arguments have held up well under intense critical scrutiny and debate. As in his earlier Etoro work, the implication was clear: cultural beliefs and practices are structured in ways that fundamentally shape patterns of social relations and more encompassing dynamics of social organization, reproduction, and transformation. These developmental dynamics cannot be reduced to environmental pressures, though they meticulously engage them. Among cultural anthropologists, Kelly has been particularly distinctive in taking seriously the empirics of ecological and sociological argument and showing on the basis of their own standards of evidence how social reproduction and transformation cannot be explained without considering how cultural factors inform social organization. In 1986, The Nuer Conquest won the University of Michigan Press Book Award.
Warless Societies and the Origins of War

During the 1990s, Kelly greatly expanded the scope of his scholarship in comparative purview as well as in historic and evolutionary depth. Building on and furthering a venerable tradition in Michigan Anthropology, he drew on both ethnographic and archeological evidence to explain the structure, organization, and transformational dynamics of simple societies and their evolution through time. Kelly began this work by scrutinizing the comparative ethnographic record of foraging societies as reflected in the data base of the Atlas of World Cultures (1981), conceived and configured by George Peter Murdock by means of codes that he and a colleague first published in the journal Ethnology in 1969 (Murdock and White 1969). Qualitatively oriented cultural anthropologists have often criticized reliance on the AWC and analogous compendia such as the Human Relations Area Files, which Murdock created in 1949, because of the ways that Murdock and his team of data coders pigeon-holed the rich ethnographic information concerning societies across the world into a small number of categories. This reduced the complexity of human practices and beliefs to conceptual comparability by means of classifications that were often seen as rigid, outmoded, or otherwise problematic.

For such reasons, Kelly’s use of the AWC was highly selective and cautious. He discerned key analytic distinctions for decentralized foragers by using AWC codings, and then went back to the original ethnographic sources to carefully examine the relevant comparative features across the full range of conceptually delimited but nonetheless revealing cases. Proceeding in this fashion, he was able to draw upon a worldwide sample, pare it down effectively for his analytic purposes, and use existing and “independent” ethnological assessments made by Murdock and others in the study of conflict, warfare, and social organization in decentralized societies. By this means, he surpassed the comparative achievements of his predecessors in the use of AWC and similar data bases for understanding the developmental organization and formative evolution of societies.
War and segmentation

Kelly’s analysis reveals how lethal violence among foragers is culturally cognized and socially organized. He makes an important distinction between foraging societies, where the organization of kinship and descent is “unsegmented” and warfare generally absent, and “segmented” societies where warfare tends to be present. Unsegmented societies are “characterized by the minimal complement of social groups,” for example, the family and the local community but nothing more (2000:45). By contrast, segmented societies are organized into enduring named groups, such as descent groups or other types of “segments” defined as “units that are equivalent in structure and function” (2000:45). “Segmental organization,” for its part, is thus “the combination of these like units into progressively more inclusive groups within a segmentary hierarchy” (2000:45). A salient feature of segmented societies, lacking in unsegmented societies, is the presence of a Radcliffe-Brownian logic of “social substitutability;” group members are considered both by themselves and others to be “substitutable” for one another across a broad range of social roles and activities. The structural equivalence of group members in these societies is variously reflected in patterns of kinship terminology and marriage, and in dynamics of collective social action including, prominently, sociopolitical features associated with armed conflict carried out by a group in its capacity as a group, i.e., warfare.

War in Kelly’s analysis is defined as collectively organized, premeditated, purposeful, and lethal armed conflict or aggression between societies or factions thereof. Of considerable relevance to all observers of human society, warfare is almost always seen as morally legitimate by its perpetrators. Unlike the aims of capital punishment and of lethal feuding, the goal of warfare is not to kill presumably guilty parties or their relatives. Instead, the targets of warfare are simply members of the larger social group (e.g., a faction or society) who may not and typically do not have personal involvement or direct culpability in the precipitating conflict (see Kelly 2000:3–10). On Kelly’s analysis, warfare is qualitatively different from both capital punishment and feuding that end in death; it is not universal in the sense of being present in all societies cross-culturally or throughout history or evolution. Put differently, just as warfare is not grounded in “human nature” and is not an inevitable feature of human society, human beings are not inherently violent or warlike, though under certain socio-cultural conditions they certainly can be. Of more immediate significance in terms of Kelly’s model is that deadly conflict as warfare is characteristic of segmented societies but not of their unsegmented counterparts.

Conflict, strife, and violence do occur among many foragers that have unsegmented social organization; these groups are not necessarily “peaceful.” And women, children, and domesticated animals are sometimes subject to physical coercion and aggression in these societies. Hence the carefully chosen words in the title of the book: Warless—not “Peaceful”—Societies. But in unsegmented societies, the occurrence of violence is typically met with processes of dispute management that reduce the escalation or collectivization of violence and that preclude rather than encourage warfare. These processes include negotiation, mediation, arbitration, capital punishment, and rituals of peacemaking.

As in his previous work, Kelly is clear that the key distinction he is making, in this case, between unsegmented and segmented foraging societies, is not a result of, and cannot be explained in any simple or direct way, by ecological features such as population density, degree of migratory mobility, food storage, or overall resource availability. Rather, the relevant contrast stems from—is literally entailed in—the structure and cultural logic of social organization. What is in many ways the larger and more compelling issue, at least or especially for students and the college-educated public, is how human violence and armed conflict have developed in organizational terms and been shaped over thousands of years.

In this respect, Kelly offers a new understanding of the evolutionary origins of human warfare, one that enables him to eschew hoary and typically irresolvable debates over whether humans are “by nature” warlike or peaceable. He demonstrates that war is a distinct organizational form of human aggression with a specific origin and an equally specific etiology associated, in turn, with specific structural and cultural features of social organization, and that “the earliest conclusive archaeological evidence of warfare dates from 12,000 to 14,000 B.P.” (2000:148), before the Neolithic and the development of agriculture. Before this, human societies experienced interpersonal conflict and violence but apparently had no knowledge or experience of warfare, or of the segmental social organization that was its most important precondition.

Kelly’s central thesis and corollary arguments receive strong support not only from AWC data, but from accounts published by analysts who had neither discerned nor made use of a distinction between segmented and unsegmented societies. Kelly’s broadly comparative project also entailed sifting through extensive archeological data and painstaking reconsideration of particular ethnographic cases for which fine-grained information was available and could be reexamined in detail. His brilliant re-analysis of historical and ethnographic material bearing on patterns of disputing and conflict management in the transitional case of the Andaman Islanders that had been collected by V. M. Portman, E. H. Man, A. R. Radcliffe-Brown, and others merits special mention here, as does his incisive utilization of Gebusi data gathered by Knauft more than a century after the early Andaman accounts were written (e.g., Knauft 1985; cf. Knauft 2002, 2005). Kelly’s respect for the specifics of ethnographic data resonates here with his letters to Knauft in the field in the early 1980s: each case should be taken on its own ethnographic terms rather than reduced to the patterns of another society or to an analytic or theoretical a priori. In a word, the ethnographer should show “from the inside,” based on the material itself, how ethnographic differences and similarities are both analytically important and theoretically significant.

Even more than his earlier books, Kelly’s project on the origins of warfare has attracted attention from a number of different fields, including archeology, physical anthropology, paleontology, primatology, evolutionary biology, sociobiology, sociology, political science, military history, and peace and conflict studies. In light of the caliber of his earlier work, it is not surprising that Warless Societies and the Origins of War has yet to encounter serious challenges on empirical grounds at least as far as we are aware. Scholarship of such scale and ambition can nonetheless give rise to objections concerning the way terms are defined and the manner in which conceptual orientations at the outset impact the cast and generality of the conclusions. For instance, it remains to be determined if interpersonal human violence, including psychological coercion, not meeting Kelly’s criteria of “war” was socially or systemically inconsequential in human evolution. As Kelly himself would emphasize, this is an empirical question.

The more general point is that aside from Kelly, there are very few anthropologists who can legitimately claim to have made important analytic and theoretical contributions to the archeology, history, and ethnography of unstratified human societies as well as the culture, social organization, and evolutionary implications of the specific delimited category of societies that we now realize, thanks to Kelly’s analysis, to be key to understanding the early history if not the origins of human warfare. That Kelly (2005) has recently turned his attention to analyzing the dynamics of lethal violence among non-human primates, especially chimpanzees and bonobos, so as to better understand the long-term evolutionary implications of hunting and killing among higher primates and from Paleolithic times onwards, further supports and amplifies this point. Kelly’s current work aims to compare and contrast these patterns with those that pertain in societies in which marriage payments are present versus absent. Kelly has identified in the archeological record types of material assemblages that are very likely associated with bridewealth, if not diagnostic of it. He is in the course of developing a sequenced model of when, where, and in what context this evidence appears prehistorically. This promising line of research and analysis has the potential to achieve what scholars have long desired: a powerful theory and nuanced analysis linking the evolution of human conflict with the evolution of social organization by means of artifactual evidence from the archaeological record.
Considerations and conclusions

In the years before Sir Edmund Leach died in 1989, he sometimes complained that word processors and personal computers made it all too easy for anthropologists to spin out insufficiently reasoned and poorly supported arguments—if not endless pages of drivel (Martha Macintyre, personal communication, December 2005). Leach would be reassured to learn that Kelly has written all of his books and articles, and most of his correspondence, in longhand, by pencil, on legal pads. When required to use a keyboard, as when composing emails, he depresses each key with the eraser end of a pencil. Whether or not one agrees with Leach’s position that contemporary technologies of writing compromise the quality of our intellectual output, the image of Kelly armed with a pencil, or hunting and pecking at the keyboard, is perhaps worth savoring as a reminder of the careful deliberation that goes into everything he writes.

Kelly’s overall work has focused on, and in this specific sense, privileged, the empirics of social phenomena, the construction of comparative frameworks, and the development of analytic models that elucidate and theorize the irreducible importance of structure and cultural logic in systems of social organization and the dynamics of social transformation. Kelly has applied this focus to an expansive range of cases and topics; most notably, to the structure and organization of Etoro social relations and the gendered system of prestige and stigma to which they are related; to the Nuer’s remarkable expansion at the expense of the Dinka; and to the origin and early development of warfare as revealed through the ethnographic and archeological record. Rather than relying on largely symbolic interpretations, or one or another version of cultural or subjective relativism, Kelly uses fine-grained conceptual distinctions to classify and compare phenomena and to tease out structures and logics through which cultural influences and social entailments are both intertwined and understandable. For the most part, the structuring of culture for Kelly is less the “thing” to be explained than the explicandum, less the object of investigation than the means of study, the entity that predicates social objects and behavioral phenomena that are otherwise hard to explain.

Kelly’s scholarship builds on the strengths of empiricism and classic Western social science generally and remains fully within the epistemological paradigms of these traditions. He has developed an innovative framework for the analysis of social process and transformational dynamics that is informed by cybernetic theory, structuralism, structural-functionalism, and symbolic anthropology. Some might argue that the variant of anthropology Kelly has pioneered also flirts with certain dangers of structural-functionalism. When deployed without great care, the epistemologies and methods at issue risk reifying terminological and conceptual distinctions as independent variables and then turning around to classify the material to be explained in ways that parallel and conceptually mirror the causes adduced to explain them. The specter of tautology arises when the logic of these distinctions infuses both the cultural cause and the behavioral result at the same time that these are separated logically as cause and effect. To put this crudely in terms of Kelly’s major publications, structural contradictions explain Etoro social organization completely, but ultimately in an ex post facto fashion; Nuer customs of bridewealth directly precipitate momentous changes, but are themselves vouchsafed from causation except secondarily through the result of their own effects; the cosmologically grounded but more or less autonomous system of prestige and stigma that helps constitute and differentiate Etoro men and women is conceptually divorced from a nexus of labor, exchange, and economic benefits, and then shown to be empirically disarticulated from rather than connected to it; war in simple human societies is defined in ways that evoke the definition of social segmentation, and is then found to be absent in societies in which segmentation does not exist.

Neither of us believes that potential criticisms along these lines abrogate the value of Kelly’s scholarship; his attention to conceptual as well as empirical detail affords his analysis a power not found in the hands of less competent craftspeople. Further, such criticism does not do justice to either the analytic sophistication or the empirical documentation of Kelly’s corpus. In practice, moreover, the cultural drivers that Kelly adduces in each case are not mandated by a theoretical master narrative. In this respect, his work contrasts with structural-functionalism as well as with the other major approaches he draws upon, just as it provides a strong counterweight to the explanation of cultural phenomena by non-cultural means or even by a priori theories of culture itself. Kelly thus achieves “large theory” in its most empirically responsive form while avoiding the excesses of “grand theories” that depend more on the force of a master theoretical narrative than on the empirical complexities of the phenomena at issue.

The larger point—and the broader value—as Kelly himself would emphasize, cannot be reduced to the sweep of summary review or categorical assessment. Significance must be read and appreciated in the specifics of the work itself. This, in a larger sense, and beyond even the enduring content of his particular ethnographic, analytic, and theoretical contributions, is Kelly’s lasting contribution to anthropology. Shakespeare wrote, “The play’s the thing.” For Ray Kelly, “The work’s the thing.” An unwavering meritocrat in the most positive sense of the term, he respects excellence in the work of others, demands it in himself, and has consistently encouraged and rewarded it in his students. Unlike so many in academia, Kelly has never been one to assume the specifics of ethnographic content, to favor politics or personal background over substance, to worry about fads of fashions, to curry favor with who might advance his career, or to take the short way around. The work is the thing, to be evaluated on the basis of clear evidence.

The keenest mind, the sharpest analysis, the most meticulous empiricism, the broadest implications: these are not just the substance of Raymond C. Kelly’s scholarship but a larger model of what anthropological excellence can be.
Acknowledgements

We would like to thank Michael Lambek for comments on an earlier version of this essay, and Ray Kelly for biographical and other factual information.