Archaeologists have long been interested in the process of state formation. However, attention has recently shifted towards investigation of lower levels of political complexity, particularly chiefdoms. The publication of a recent School of American Research seminar (Earle 1991b) and other research on chiefdoms (summarized by Earle 1987) has led to a fairly radical reinterpretation of our understanding of such “intermediate-level” societies. The functionalist notion that chiefs managed the distribution of resources has been replaced by a realization that chiefs were rather more selfish individuals out to extract a surplus from their followers, who in turn could curb the more despotic tendencies of their leaders by threatening to shift their allegiances elsewhere. This description, albeit superficial, reflects the central role given in recent work to discussion of how chiefs acquired and retained power, the essential components of which were control of the economy, war, and ideology (Earle 1991a: 9).

Chiefdoms are usually viewed in evolutionary models of state formation as the immediate precursors of states. These models have been the subject of a spirited attack by Yoffee (1993). He, interalia, points out the dangers inherent in subverting a taxonomy of recent societies into an evolutionary scheme whereby contemporary chiefdoms, for example, are viewed as representative of the historical precursors of states, allowing archaeologists to flesh out flimsy data with borrowed ethnographic detail. Yoffee (1993) has also cogently argued that states do not normally evolve from chiefdoms. Instead, states may arise from the competition among different nodes of power (economic, political, and ideological) within a society. In fact, successful chiefdoms would appear to be inimical to the development of states.

Introduction

The concept of civilization was for a long time defined in terms of the presence or absence of what was considered to be key elements, such as writing. It is now generally accepted that civilization can exist without the presence of writing, which Gordon Childe had taken for the one ingredient that usually marked the turning point (Childe 1950, quoted in Connah 1987: 7). Many scholars have made the point that these definitions were based on a limited knowledge of other parts of the world, especially Africa.

Like civilization, the concept of urbanization has for a long time been pervaded by Western representations. Common to most of the early theories about the nature of a town or city is the belief that the existence of an agricultural population, a mercantile class, and writing were a prerequisite. Also considered as necessary was the idea that the settlement had to perform the functions of a religious and an administrative center, capable of mobilizing a massive labor force for monumental architecture and the accumulation of wealth (Hassan 1993; R.McIntosh, this volume; O'Connor 1993).

Long-distance trade and religion have indeed influenced the growth of towns in some parts of the world. The rise of the mercantile towns of East Africa, for example, was influenced by Portuguese and Arab trade during the sixteenth and se venteenth centuries. Similarly, the Islamic towns of North Africa, with their sultanates, mosques, Islamic schools, and wall enclosures, testify to strong Islamic influence in this part of the continent prior to European colonization.

Introduction

A major current issue throughout southern Africa is the manner by which the false convergence of tribal/ethnic allegiance and ideology initiated by Europeans during the colonial era can be unraveled. This requires a new formulation of the history of the subcontinent – one which will rely on archaeological materials to provide a firm historical basis for assessing the trajectory of precolonial social and economic formations and their transformations in the colonial period. Yet, until recently, anthropologists and historians have tended to view the entire region before the latter part of the nineteenth century as having been populated by isolated cultural and linguistic entities with few connections among them. Khoisan-speaking people in the Kalahari, for instance, have been represented as foragers unaffected by Bantuspeaking agropastoralists until the twentieth century. All local cultures, however, have been historically constructed through ongoing dialogues with others – dialogues characterized as much by social flux as by stability. As Wilmsen and Vossen (1990: 11) point out, “the resultant constructed ethnicities rarely conformed to a people's prior self-identification.”

Current “debates” over the “authenticity” of some Kalahari peoples (Solway and Lee 1990; Wilmsen and Denbow 1990) thus exemplify more general trends in anthropological thought that question the historical utility of ethnographic concepts such as closed cultural traditions, that take on an apartheid-like appearance in the southern African context.

The organic analogy

High in the Canadian Rockies of Yoho National Park lie outcrops of early Cambrian shale that contain the fossils of creatures as exotically named – Ayesheaia, Wiwaxia, Anomalocaris, Hallucigenia – as they are themselves fantastic and, indeed, out of this world. They have, as Stephen J.Gould (1989) proclaims in the book from which our title is derived, changed our view of life, for, despite the small numbers of species represented, they are more disparate – more varied at the level of phylum and class – than the “entire spectrum of invertebrate life in today's oceans” (p.25), a spectrum now characterized by profligate diversification of a restricted number of classes.

This disparate, but not diverse, early Cambrian fauna appears to have arisen explosively in what must be the type specimen of punctuated evolution. There is, for the majority of these creatures, no lengthy preceding record of gradual change. Rather, some 600 million years ago, new patterns of life suddenly radiate through the Cambrian oceans. Elsewhere, Gould argues that this steep rise in organic diversity in the first 10 to 20 million years of the Cambrian, “was merely the predictable outcome of a process inexorably set in motion by an earlier Pre-Cambrian event … that initiated the evolution of complex life.” The event, or rather complex of events, was the development of eukaryotic cells and of sexual reproduction. Once these were in place “[t]he pattern of the Cambrian explosion seems to follow a general law of growth” (Gould 1979: 127), the sigmoid curve that classically characterizes the career of a colony of bacteria on a Petri dish (Figure 8.1a).

Archaeologists of my generation have experienced dramatic shifts in the ideas that orient our field – as students in the early 1970s we cut our teeth on the New Archaeology, and many embraced the neo-evolutionary models that informed a processual understanding of human development. The musty culture history of our instructors' elders seemed uninteresting, and our imaginations were fired by questions of origins and transitions: the origins of food production and the transition to sedentary village life; the origins of complex societies and the growth of cities. We were concerned with modeling, with cross-cultural comparison, and ultimately with processual questions of evolutionary change. During the 1980s we weathered another sea-change in theoretical currents with the upwelling of critical stances loosely grouped under the rubric of post-processual archaeology. Although this belies the diverse perspectives of archaeologists so-labeled – some marxist, some feminist, some post-structuralist (Kohl 1993) – they share a disdain for the evolutionary models that played a prominent role in processual archaeology. They criticize evolutionary models as universalizing, winnowing diversity by forcing an array of societies into rigid types; as hierarchical, ranking societies along dimensions of complexity and heterogeneity; as reductionist, implying a correlation between particular economic and social forms; and, as distancing simple societies in time by treating them as relics of earlier stages (Andah 1995; Shennan 1993; Thomas 1989; Upham 1990a, 1990b;Yoffee 1993). Others have examined the relationship of evolutionary ideas with imperialism, ethnocentrism, and inequalities of power (Bowler 1992; Fabian 1983; Gamble 1992a, 1992b; Rowlands 1989b, 1994; Schmidt and Patterson 1995; Stocking 1987; Thomas 1994; Trigger 1989). Thus progressive evolutionary schemes map societies in ways that are ideologically charged.

The ethnography of the Alur segmentary state

The Alur taught me the practice of the segmentary state, as well as their language. I had to make it theory in my language, the most stereotypical and intractable form of distancing the Other. The segmentary state is one in which the spheres of ritual suzerainty and political sovereignty do not coincide. The former extends widely towards a flexible, changing periphery. The latter is confined to the central, core domain. A number of such partially overlapping entities with political cores at the center of wider ritually based zones may be related to one another pyramidally at several levels. The Other may be partially assuaged by exploring the origin of the Alur segmentary state. Alur society emerged as a distinct entity in the sixteenth and seventeenth centuries, when Lwo migrants from the north crossed the Nile and moved west (Crazzolara 1950/51). The segmentary state of Atyak became the largest among the Alur (Southall 1956: 349), because its early leaders and their following moved into a part of the country where they were able to develop the most favorable combination of agriculture and pastoralism, while also remaining beyond the disturbing influence of Bunyoro (Southall 1956: 9, 16). As they moved up into the highlands, they were able to incorporate (Southall 1970) small groups of earlier settlers (Okebo, Lendu, Madi, Abira, etc.) who belonged to quite different ethnic groups and spoke mutually unintelligible languages (Southall 1956: 16–24). I have not found it possible to relate the concept of the internal frontier to this situation in any meaningful way.

My account represents the most plausible hypothesis based on Alur traditions and elders' statements.

Introduction

Between 1600 and 1800, in eastern Africa's Great Lakes region, the states of Bunyoro-Kitara and Buganda formed and began to export their bureaucratic, militaristic, and religious hegemony to neighboring societies. They did so with little or no connection to intercontinental or maritime trading systems. They did so in wet, dry, highland, and lakeshore environments. They did so using metals, cattle, bananas, grains, fishing, and hunting, in varying combinations, as their technological and agricultural bases. But they also did so with less visible resources, with funds of meaning and concepts of power and with sometimes abstract units of social organization. This essay will argue that the history of these visible and invisible roots of these two states between the Great Lakes reveals that any sort of evolutionist model for state formation must take these factors into account.

Exploring these issues begins with the evidence from Great Lakes societies which diagnoses their own analytical categories for social life. If scholars choose to settle on chiefdoms as important units of study they must understand their importance to those who lived in them. Such a condition highlights the necessity of interdisciplinary scholarship. Archaeologists, comparative linguists, and comparative ethnographers generate information on the varieties of chiefdoms and, more importantly, these scholars may suggest what other sorts of social and material relations – family structure and gender – may be implicated in the development of chiefdoms. For historical linguists, such data emerge from the comparative study of retention and innovation in the vocabulary (material and lexical) of power, settlements, social relationships, and gendered identities (see Appendix). But only careful survey and excavation by archaeologists will confirm these sequences of innovation and reproduced continuity.

In the Western popular imagination, sub-Saharan Africa has long been the domain of small-scale, dispersed, mobile, swidden agriculturalists. Yet within the Inland Niger Delta in the Sudanic zone of West Africa, some of the earliest documented agricultural communities display quite the opposite pattern: rapid population growth over the course of centuries-long occupation of high density settlement clusters. Commonly, a cluster comprises a large, central settlement mound of up to 10 meters in height and 20–80 hectares in area, surrounded by intermediate and smaller mounds at distances of 200 meters or less. In the vicinity of Jenné-jeno, where multiple clusters are present and there is evidence from survey and excavation for the extent of occupational contemporaneity among the various mounds, the total occupied mound surface exceeded 100 hectares within a millennium of the initial pioneering settlement c.250 BC. From a comparative perspective, the lengthy occupation sequences and the pattern and scale of population nucleation documented in the Inland Delta are relevant to a number of long-standing theoretical discussions in archaeology, among them the relation of population growth and the evolution of agricultural systems, and the relation of population size and political hierarchy. In this case study, both the lack of evidence for agricultural intensification and the distinctive pattern of nucleated settlement clusters are considered essential elements for understanding the particular trajectory that developing complexity assumed in the region.

Colonization and settlement clustering in the Inland Niger Delta

The Inland Niger Delta (IND), as its name implies, is an area of false deltaic hydrology located in the heart of the modern state of Mali over 1,500 km inland from any sea coast.

Over the past decade, sub-Saharan Africa has virtually disappeared from the screen of archaeologists engaged in broadly comparative, theoretical discussions on the emergence of complex society. Prior to the 1980s, the subcontinent was represented with some regularity at important archaeological conferences and discussions on these issues (e.g., Cohen and Service 1978; Friedman and Rowlands 1978; Moore 1974) even while the actual archaeology of sub-Saharan complex societies remained nascent. Since then, the visibility of Africa in comparative theoretical discussions has declined considerably, despite the surge of interest in societies organizationally intermediate between small-scale, non-stratified and locally autonomous groups and the internally differentiated state (e.g., Arnold 1996; Drennan and Uribe 1987; Earle 1987, 1991c; Gregg 1991; Price and Feinman 1995; Upham 1990) and despite the abundance and diversity of such societies throughout the subcontinent at the time of European colonial expansion. Sub-Saharan regions are represented briefly, if at all, in some widely cited works (Earle 1987, 1991c; Ehrenreich et al. 1995; Haas 1982; Price and Feinman 1995; Renfrew and Cherry 1986; Trigger 1993 is a notable exception). Ironically, the archaeology of complex societies in Africa has grown remarkably during this same period (see, e.g., Shaw et al. 1993).

The primary objective of this volume is to reintroduce an African perspective into archaeological theorizing about complex societies. This is a daunting task because the subcontinent is vast (over three times the size of the United States) and in historic times has exhibited an astonishing diversity of sociopolitical formations. Thus, any attempt at general coverage will necessarily suffer from incomplete and unsatisfactory geographic representation, and leave a host of relevant topics and potential insights unexplored.

This book project had its origins in my own frustration at the paucity of African models, African-inspired theories, and African case studies in the archaeological literature on the development of complexity. When Antonio Gilman asked me to organize an archaeology session at the 1992 annual meeting of the American Anthropological Association in San Francisco, I used the opportunity to showcase what that continent has to offer to ongoing debates about the origin and nature of complex societies. If archaeology was to live up to its claims to be a comparative discipline, it needed to take Africa, especially sub-Saharan Africa, seriously and systematically into account.

The papers at the invited session entitled “Intermediate-level Societies in Africa: Archaeological and Anthropological Insights” ranged widely over the continent and included presentations by anthropologists and historians as well as archaeologists. A number of the papers in this volume (those by Southall, S.McIntosh, Schoenbrun, Robertshaw, and Denbow) were originally presented at this session. Igor Kopytoff was also scheduled to participate, but circumstances prevented it. Fekri Hassan participated in the session, but in view of the strongly sub-Saharan focus of the rest of the papers, it was decided to strengthen that focus for publication and exclude ancient Egypt – arguably the best-known African complex society. Subsequently, papers presented by Ann Stahl and Nic David at the Complex Society Group's meeting in San Bernardino in October 1995 were added, and additional papers solicited. Not all of the additionally solicited papers actually materialized, leaving the coverage of some themes and areas envisioned for inclusion rather threadbare.

The problem of archaeological invisibility

This chapter examines certain aspects of the history of thought about the origins of pre-industrial cities in an attempt to better understand why African urban expressions in the archaeological record remained unrecognized and unexamined for so long. In some cases, such as that of the Inland Niger Delta (also discussed in this volume by S.K.McIntosh), these include massive tells that are, on the face of it, difficult to overlook. How could such distinguished observers of the ground as Raymond Mauny (1961: 101–2) or Théodore Monod (1955), for example, walk over the massive site of Jenné-jeno, yet dismiss it with barely a mention? And these are only the most distinguished of the cohorts of prehistorians and natural scientists who, during perhaps a century, have yawned their way over the high and extensive tell remains of many early Middle Niger towns. I contend that they lacked the intellectual toolkit to recognize and process the evidence under their feet. They failed to recognize the town they trod upon because they did not find the expected signs of pre-industrial urbanism, namely, encircling wall and citadel (or palace or temple) reflecting coercive political organization, elite tombs or residences or other accoutrements as monuments to economic social stratification, and monumentality of architecture as monument to state ideology (R.McIntosh 1991:203).

I argue here that our deeply rooted view of the non-Western city as despotic, depraved, and a place of bondage comes out of the tradition of Bible exegesis called Yahwism (or Jahwism).

Introduction

With its many shared linguistic, symbolic, and ideological characteristics, the Bantu-speaking population of central, eastern, and southern Africa offers an exceptional opportunity for comparative studies. How, with a common background, they managed to expand and adapt to the various local conditions they encountered in this gigantic area over a period of probably more than three millennia is fascinating. Although two-thirds of Fortes and Evans-Pritchard's seminal book African Political Systems (1940) was devoted to Bantu political organization, the potential contribution of this part of the world to the recent debate on chiefdoms, intermediate level societies, and the rise of early states has been almost completely overlooked in favor of Oceania, America, and, to a lesser extent, Europe (Earle 1987).

Bantu Africa offers not only side-by-side examples of major kingdoms, such as the Kuba, and autonomous villages with collective leadership, such as the Lele, but also hundreds of societies exhibiting a wide range of intermediate political systems. Considering only the Bantu living in the rainforest, Vansina (1989) notes

There is an evolutionary bias in the anthropological view of political development, a bias that makes it seem obvious that political forms should have moved from the simple to the complex. Thus, “chiefdoms” necessarily arose out of acephalous structures such as “bands,” usually in response to similar economic conjunctures, such as a rise in trade. And the chiefdom, unless mired in evolutionary stagnation, necessarily moved in the direction of the “state.” This view of political complexification is perhaps not unrealistic in the very long term, but it becomes misleading when we extend it to the formation of actual middle-range polities in a particular ethnographic area. If the grand evolutionary scheme had indeed been working itself out uniformly over the centuries, there should have been very few small-scale polities in the world by, say, the nineteenth century, and they should all have been of very ancient vintage. Yet, when we look, for example, at Africa in recent times, we find it full of small scale polities whose formation usually dates back but a few centuries and often less.

I have argued elsewhere (Kopytoff 1987) that most of the African polities we know did not evolve out of simpler forms. To the contrary, they sprang out of more complex polities, having grown out of settlements of immigrants from chiefdoms and kingdoms, immigrants who had moved into the “internal frontiers” that lay at the fringes of fully formed polities. To these settlers, the frontier represented an institutional vacuum – an area that was out of the reach of established polities, or was entirely empty, or was under the uncertain sway of weak local hegemonies.