Cannibals and Kings

by Marvin Harris

  The larger and denser the population, the larger the redistributive network and the more powerful the re-distributor war chief. Under certain circumstances, the exercise of power by the redistributor and his closest followers on the one side and by the ordinary food producers on the other became so unbalanced that for all intents and purposes the redistributor chiefs constituted the principal coercive force in social life. When this happened, contributions to the central store ceased to be voluntary contributions. They became taxes. Farmlands and natural resources ceased to be elements of rightful access. They became dispensations. And redistributors ceased to be chiefs. They became kings.

  To illustrate these momentous transformations in the context of a small preindustrial state, I shall call upon John Beattie’s description of the Bunyoro. Ruled over by a hereditary ruler called the mukama, the Bunyoro numbered about 100,000, occupied an area of 5,000 square miles of that portion of the central lake area of East Africa which is now known as Uganda, and earned their living primarily by raising millet and bananas. The Bunyoro were organized into a feudal but nonetheless authentic state society. Their mukama was a king, not a mere redistributor chief. The privilege of using all lands and natural resources was a dispensation granted by the mukama to a dozen or so chiefs, who then passed on the dispensation to the commoners. In return for this dispensation, quantities of food, handicrafts, and labor services were funneled up through the power hierarchy into the mukama’s headquarters. The mukama in turn directed the use of these goods and services on behalf of state enterprises. Superficially, the mukama appears to be just another “great provider” redistributor chief. In Beattie’s words:

  The king was seen both as the supreme receiver of goods and services, and as the supreme giver.… The great chiefs, who themselves received tribute from their dependents, were required to hand over to the Mukama a part of the produce of their estates in the form of crops, cattle, beer or women.… But everyone must give to the king, not only the chiefs.… The Mukama’s role as giver was, accordingly, no less stressed. Many of his special names emphasize his magnanimity and he was traditionally expected to give extensively in the form both of feasts and of gifts to individuals.

  But a comparison of the mukama with the Trobriand or Cherokee supreme chief reveals that power relationships had become inverted. The Trobriand and Cherokee chiefs were dependent on the generosity of the food producers; the Bunyoro food producers were dependent on the generosity of the king. The mukama alone could grant or withhold permission for blood vengeance, and failure to contribute to the mukama’s income could result in the loss of one’s lands, banishment, or corporal punishment. Despite his lavish feast-giving and reputation as a “great provider,” the mukama used much of his income to bolster his monopoly over the forces of coercion. With his control over the central grain stores he maintained a permanent palace guard and heaped rewards on warriors who displayed bravery in combat and loyalty to his person. The mukama also spent a considerable portion of the state treasury on what we would today call “image-building” and public relations. He surrounded himself with numerous officials, priests, magicians, and such regalia keepers as the custodians of spears, of royal graves, of the royal drums, of royal thrones, and of royal crowns, as well as “putters-on” of the royal crowns, cooks, bath attendants, herdsmen, potters, bark-cloth makers, and musicians. Many of the officials had several assistants. Other advisers, diviners, and retainers hung around the court in the hope of being appointed to a chieftainship. Also present were the mukama’s extensive harem, his many children, and the polygynous ménages of his brothers and of other royal personages. To keep his power intact, the mukama and portions of his court made frequent trips throughout Bunyoro land, staying at local palaces maintained at the expense of the chiefs and commoners.

  As Beattie points out, many features of Bunyoro kingship were also present in post-Roman feudal Europe. Like the mukama, William the Conqueror and his entourage traveled constantly about twelfth-century England, checking up on his “chiefs” and living off their hospitality. The English kings at that time still displayed evidence of their origins as “great providers” at the head of redistributive networks. William the Conqueror, for example, held three great feasts yearly at which he wore his crown and entertained great numbers of lords and subjects. As we shall see, however, the further evolution of state systems gradually led to the removal of all obligations on the part of rulers to act as “great providers” for their subjects.

  Under what circumstances would the conversion of a redistributive chieftainship to a feudal state be likely to occur? To intensification, population growth, warfare, storageable grains, and hereditary redistributors, add one more factor: impaction. Suppose, as Robert Carneiro has suggested, a population being served by redistnbutors has been expanding inside a region that is circumscribed, or closed off, by environmental barriers. These barriers need not be uncrossable oceans or un-climbable mountains; rather, they might merely consist of ecological transition zones where people who had broken away from overcrowded villages would find that they would have to take a severe cut in their standard of living or change their whole way of life in order to survive. With impaction, two types of groups might find that the benefits of a permanently subordinate status exceeded the costs of trying to maintain their independence. First, villages consisting of kinspeople forced to enter the transition zones would be tempted to accept a dependent relationship in exchange for continued participation in the redistributions sponsored by their parent settlements. And second, enemy villages defeated in battle might find it less costly to pay taxes and tribute than to flee into these zones.

  Very little direct physical coercion would be needed to keep the emergent peasantry in line. Kinship would be used to justify the legitimacy of differential access to resources on the part of junior and senior lineages or of wife-giving, wife-taking alliance groups (those who gave wives would expect tribute and labor services in return). Access to the stored grains might be made contingent upon rendering craft or military services. Or the “big men” of the more powerful group could simply begin taxation by redistributing less than they took in. External warfare would increase and defeated villages would be regularly assimilated into the tax and tribute network. A growing corps of military, religious, and craft specialists would be fed out of the central grain stores, amplifying the image of the rulers as beneficent “great providers.” And the social distance between the police-military-priestly-managerial elite and the emergent class of food-producing peasant drudges would widen still further as the scope of the integrated food production facilities increased, as trade networks expanded, as population grew, and as production was intensified still further through more taxation, labor conscription, and tribute.

  How well does the theory of environmental circumscription and impaction accord with the evidence? The six most likely regions of pristine state development certainly do possess markedly circumscribed zones of production. As Malcolm Webb has pointed out, all of these regions contain fertile cores surrounded by zones of sharply reduced agricultural potential. They are, in fact, river valleys or lake systems surrounded by desert or at least very dry zones. The dependence of ancient Egypt, Mesopotamia, and India on the flood plains of the Nile, Tigris-Euphrates, and Indus is well-known. In ancient China conditions of climate, soil, and topography limited intensive forms of agriculture beyond the river margins of the Yellow River Basin. Central highland Mexico south to Tehuantepec is also dry and in addition “suffers from severe rain shadow effects in the highland basins and stream valleys that were the aboriginal population centers.” And finally, the Peruvian coast is notable for the stark contrast between the lush vegetation bordering the short coastal rivers that flow down from the Andes and the desert conditions that prevail everywhere else. All of these regions present special difficulties to villages that might have sought to escape from the growing concentration of power in the hands of overly aggressive redistributor war chiefs.

 
; Furthermore, there is no doubt that all of these regions were the scene of rapid population growth prior to the emergence of the state. I mentioned earlier that the population of the Middle East increased fortyfold between 8000 and 4000 B.C. Karl Butzer estimates that the population of Egypt doubled between 4000 and 3000 B.C. William Sanders estimates that population tripled or quadrupled in the highland zones of early state formation in Mexico, and similar estimates also apply to Peru, China, and the Indus Valley. “For all areas one receives the impression of an increase not only in the total number of sites but also in the density of distribution, size, and elaboration of sites.”

  Malcolm Webb has also reviewed the evidence for warfare. Egypt’s legendary history begins with a tale of conquest, and specialized instruments of war and fortifications appear early in the archaeological record. In Mesopotamia weaponry and representations of slaves and battles are present in early predynastic times. Fortifications and documentary evidence indicate that Shang China, at the time of the emergence of the first Yellow River states, was an extremely militaristic society. Recent discoveries in the heartland of the earliest Indus River states have confirmed the existence of strongly fortified neolithic villages that were destroyed by conquest. In the New World “both coastal Peru and Mesoamerica show a long history of warfare”; archaeological “indications of fighting are present no later than the start of the first millennium B.C.”

  The kind of warfare that was conducive to the evolution of the state obviously must have involved long-distance external combat by large coalitions of villages rather than internal warfare of the Yanomamo variety. Matrilocality being a recurrent method of transcending the limited capacity of patrilineal village groups to form multi-village military alliances, it seems likely that societies on the threshold of statehood would frequently adopt matrilineal forms of social organization. According to Robert Briffault, there is a considerable body of literary evidence to support the view that ancient state societies did possess matrilineal institutions shortly before and shortly after their achievement of statehood. The great Egyptologist Flinders Petrie, for example, was of the opinion that the administrative divisions, or nomes, of early dynastic Egypt had once been matrilineal clans and that postmarital residence in earliest times was matrilocal. Strabo, the Greek historian, recorded that the ancient peoples of Crete worshiped predominantly female divinities, granted women a prominent role in public life, and practiced matrilocality. Plutarch says that in Sparta marriage was matrilocal and that “women ruled over men.” The great classicist Gilbert Murray was convinced that in Greece during Homeric times “sons went off to foreign villages to serve and marry women in possession of the land there.” Herodotus said of the Lycians at the eastern end of the Mediterranean, “They have one singular custom in which they differ from every other nation in the world: naming themselves by their mothers, not their fathers.” And of the early Germans, Tacitus wrote that “the sons of a sister have the same position as regards their uncle as with their father” and “some even consider the former as the stronger tie.”

  To a modern-day anthropologist this emphasis upon the tie between mother’s brother and sister’s son strongly suggests the existence of an earlier matrilineal organization. Moreover, Tacitus’ description of the relatively high status of women in ancient Germany is supported by discoveries of females dressed as warriors buried side by side with males dressed the same way. Livy reports that the curiae, or the earliest administrative divisions, were named after the Sabine women whom Romulus’ followers were supposed to have raped. Finally, Briffault points out that Roman kinship nomenclature preserved a distinction between the father’s brother and the mother’s brother. The former was called patruus; the latter, avunculus. The Latin word for ancestor was avus. Hence, as would be the case in a matrilineal system, the mother’s brother was designated by a term denoting common ancestry with the sister’s son. (The fact that the English word “uncle” survives from the word for “mother’s brother” suggests the former importance of mother’s brother-sister’s son relationships.)

  Female figurines and statues found among many prestate cultures of Europe and Southwest Asia provide another line of evidence suggestive of matrilineal organizations. On Malta, for example, the Temple of Tarxien, built before 2000 B.C., contained a six-foot stone statue of a rotund seated woman. The theme of “fat ladies” is echoed in several smaller versions found in Maltese temples, all of which are associated with human burials, altars, and the bones of sacrificed animals—suggestive of a cult of female ancestors.

  While most of this evidence pertains largely to the formation of secondary states in Europe, it is sufficiently consistent to warrant the inference that the pristine states had earlier passed through a similar matrilineal phase. But if there was such a phase, whether for pristine or secondary states, it must have been short-lived. What we glimpse through the writings of the classical Greek and Roman historians are the lingering traces of systems that had already reverted back to patrilineal descent. Very few ancient or modern state societies have matrilineal descent or practice matrilocality (which is why Herodotus described the Lycians as differing from “every other nation in the world”). With the rise of the state, women again lost status. From Rome to China they were legally defined as the wards of their fathers, husbands, or brothers. The reason for this, I believe, is that matrilocality was no longer functionally necessary for recruiting and training the armed forces. States wage war by means of military specialists whose solidarity and effectiveness depend on hierarchical ranks and strict discipline, not on common post-marital residence. The rise of the state, therefore, saw the old male supremacy complex reassert itself in full force. I do not think it is an accident that the pre-state Siuai, Trobrianders, and Cherokee engage in external warfare and have matrilineal institutions, while the Bunyoro state, which engages in even more external warfare, has patrilineal institutions and a strong male supremacy complex.

  Once pristine states have formed in a given region, secondary states begin to develop under a variety of special conditions. Some secondary states form as a matter of defense against the predatory inroads made by their more advanced neighbors; others develop as a result of attempts to capture control over strategic trade routes and the increased volume of goods in transit which usually accompanies the growth of states in any region. Still others form as part of an attempt by nomadic peoples living on the margin of a state to plunder its wealth. States found in relatively low-density, un-impacted regions must always be examined with these possibilities in mind before concluding that intensification and reproductive pressures did not cause the evolution of the region’s pristine states. For example, low-density pastoralist peoples—Turks, Mongols, Huns, Manchus, and Arabs—have repeatedly developed states but only by preying upon the preexisting Chinese, Hindu, Roman, and Byzantine empires. In West Africa secondary states developed as a result of Moslem and European attempts to control the slave, gold, and ivory trades, while in southern Africa the Zulu developed a state in the nineteenth century to meet the military threat posed by Dutch colonists invading their homeland.

  What I find most remarkable about the evolution of pristine states is that it occurred as the result of an unconscious process: The participants in this enormous transformation seem not to have known what they were creating. By imperceptible shifts in the redistributive balance from one generation to the next, the human species bound itself over into a form of social life in which the many debased themselves on behalf of the exaltation of the few. To paraphrase Malcolm Webb, at the beginning of the lengthy process no one could foresee the end result “Tribal equalitarianism would gradually vanish even as it was being appended, without awareness of the nature of the change, and the final achievement of absolute control would at that point seem merely a minor alteration of established custom. The consolidation of governmental power would have taken place as a series of natural, beneficial, and only slightly (if at all) extra-legal responses to current conditions, with ea
ch new acquisition of state-power representing only a small departure from contemporary practice.” By the time the remnants of the old council finally sank into impotence before the rising power of the king, no one would remember the time when the king had been only a glorified mumi whose exalted status rested on the charity of his friends and relatives.

  I urge those who feel that my explanation of the evolution of culture is too deterministic and too mechanical to consider the possibility that at this very moment we are again passing by slow degrees through a series of “natural, beneficial, and only slightly … extra-legal” changes which will transform social life in ways that few alive today would consciously wish to inflict upon future generations. Clearly, the remedy for that situation cannot lie in the denial of a deterministic component in social processes; rather, it must lie in bringing that component into the arena of popular comprehension.