The story is told about Herbert Spencer that when T.H. Huxley discovered that in his youth Spencer had written a tragedy, he remarked that he knew what the plot must have been: “a beautiful theory killed by a nasty, ugly, little fact.” It is a story which comes immediately to mind on reading the back jacket of this book by Eli Sagan, who teaches at Berkeley and has previously published books on cannibalism and on violence in ancient Greece. There we find the following astonishing comment by Professor Robert N. Bellah on Sagan’s thesis about the origins of the state:

As social theory, Sagan’s analysis of the necessity for the concentration of power in the hands of the tyrant in order to make the transition from kinship structures to nonkinship structures is one of the most profound insights that I know of in the whole range of social science.

Profound it may or may not be; but it is false. States—that is, permanent forms of central political organization independent of kinship—do often originate through, as Sagan has it, a “tyrannical monarch” with “the power to overthrow the kinship system.” But this is by no means always so. There are exceptions, and they disprove the rule.

The most obvious are those where the transition to statehood comes about not through a surrender of sovereign power into the hands of a single person but through its being placed in the hands of magistrates whose scope for abusing power is deliberately circumscribed in advance. Magistrates can, to be sure, turn into tyrants. Aristotle, who knew at least something about what had gone on during the period when the city-states of Greece came into being out of the anarchy that followed the collapse of Mycenaean civilization, is quite explicit about the risk that an elected “overseer” might use his term of tenure to usurp monarchical power, as happened, for example, in Miletus. And awareness of that risk by contemporaries themselves is strikingly confirmed by the survival of an inscription of the seventh century BC from Dreros, in Crete, in which a chief magistrate who has held annual office is specifically forbidden to hold it again for another ten years. But even if the restrictions imposed did not prevent the emergence of tyrants later on, when one or another war leader, administrator, lawgiver, or noble was able to secure a sufficient following to put down any opponents or rivals, it cannot be argued, as Sagan does, that the change from rule by the heads of lineages, clan chiefs, or village patriarchs to statehood in its proper sense comes about always and only through the successful assumption of quasi-divine omnipotence by charismatic heroes (or villains).

Even in early Mesopotamia, where kingship was intimately and explicitly associated with divinity and conquerors like Sargon of Akkad behaved as tyrannically as any of the rulers whom Sagan has in mind, the initial transition to statehood appears to have come about by way of local councils—reminiscent of those documented for some African societies—which in Mesopotamia vested authority in temporary war leaders who only later became kings. The origins of the Sumerian city-states cannot be reconstructed without some element of conjecture. But many specialists find convincing the hypothesis of Thorkild Jacobsen1 that the “king” (lugal) was at first a son of a rich landowner who was therefore able to draw on his father’s retainers and was appointed for a limited period by a bicameral assembly of “elders” and “men.” The central institution of the city-state was at this stage the temple of its tutelary deity, and the ideological fusion of god and state was only later expressed through the attribution of supernatural status to the holders of political power. Again, in other words, it was not tyrants who created states so much as states that created tyrants.

The origin of the state is a fascinating topic, and it has attracted a remarkable amount of scholarly attention in recent years from specialists in every relevant geographical area and from every relevant academic discipline. There is not, or at least not yet, a theory of the origins of the state that commands general agreement. But there is virtual unanimity on two points: first, no explanation by a single cause, such as trade, or war, or population growth, or religion, will stand up against the range of examples available in the historical and ethnographic record; second, no progress toward a general explanation can be made by extrapolation from a few examples selected to illustrate, rather than test, a chosen theory. Sagan, however, is not merely dismissive of the ecological and demographic influences which are broadly agreed to be of major, even if not decisive, importance in the formation of states—for example, the way a shift from stock rearing to arable farming was accompanied by changes in patterns of settlement and population size in archaic Greece. He also relies heavily on extrapolation from Buganda, Tahiti, and Hawaii as described by European explorers or missionaries who encountered them in their pristine condition.


He is, admittedly, quite right to point out that the ease with which an economic surplus can be extracted from the soil is not a sufficient condition of the evolution of statehood: Busoga had an even more “banana-beneficent” climate than neighboring Buganda, but never developed a centralized monarchical state. But this is not an argument for dismissing “materialist” explanation out of hand in favor of an equally one-sided overemphasis on psychology. The temperamental predispositions of those who become tyrants and their henchmen, on the one side, and followers or victims, on the other, are only one among several aspects of a complex process which cannot be fully explained without a detailed analysis of how psychological, ecological, demographic, ideological, and military influences act and react upon one another.

This said, a particular interest undoubtedly attaches to the kind of despotism and the institutions and practices surrounding it that Sagan regards as the archetype of statehood. The peculiar hold that even the most atrocious tyrants often succeed in establishing over a seemingly proud and vigorous people is documented for a wide range of times and places, and Sagan is fully entitled to maintain that the rituals and mythology of Buganda, Hawaii, and Tahiti can tell us something about the nature of that hold which has application beyond the particular descriptions on which he draws. But once again, he succumbs to the temptation to generalize both too widely and too deeply. Submission to the capricious and often sadistic exercise of arbitrary power is neither as willingly made nor as seldom resented as he seeks to imply. Puzzling as it may be that Mutesa of Buganda (1856–1884) and despots like him should get away with so much, and for so long, with the apparent acquiescence of their subjects, submission to a tyrant is often reluctant, and often, when not reluctant, straightforwardly explicable by obvious self-interest.

According to Herodotus, Deioces, who founded the kingdom of the Medes (in what is now northwestern Iran), began his career as an arbitrator so successful that he was enlisted by all the villages in the region. Realizing the strength of his position, he deliberately withdrew his services with the result that the Medes voted in general assembly to establish a monarchy which they invited him to fill, allowing him at the same time both a palace and a bodyguard of his choice. From this, it was a short and predictable step to a fortified capital, a network of spies, restriction of access to the royal person, and prohibition against laughing or spitting in the royal presence. The accuracy of Herodotus’ account is, of course, impossible to verify. But there are well-attested examples where a palace and a bodyguard enable a ruler to establish a degree of domination over his subjects which they may deeply resent but which they are by that time unable to reverse.

Often, indeed, it is only with the help of an outside power that a state is formed. Firearms, as Sagan acknowledges, can be crucial: without them, and the readiness of Europeans to supply them, neither Hawaii nor, among others, Madagascar would ever have been unified under a single king. Nor, for that matter, would monarchies have been established as they were among the German tribes if it had not suited the Romans to support local strong men with whom they could then do business: Tacitus makes it clear that a king like Maroboduus of the Suebi, who briefly wielded autocratic power complete with bodyguard and palace, was thoroughly detested by his subjects and toppled by his enemies as soon as they could. The notion that tyrants only establish themselves because of the “consent of the governed,” as Sagan has it, is true only if “consent” is taken to be proved by the mere occurrence of the fact.

In any case, even where the submission is voluntary, it may have much less to do with psychoanalytic considerations of the kind that Sagan draws from Anna Freud and Margaret Mahler than it pleases him to suppose. Is it really necessary to invoke “dependence on the father-king” and the wish of the members of primitive societies, like children, to “discover what we all long for: omnipotent support without the threat of symbiotic regression” to account for the wish of aggressive young men to place themselves at the disposal of the warlike head of a lineage or region other than their own in the hopes of land, booty, and military prestige? The sort of aspiring Anglo-Saxon king described in Bede’s Ecclesiastical History, such as Oswine of Northumbria, drew his support from a combination of noble birth and a reputation for personal prowess. In territories recently brought under cultivation by bands of settles who either drove off or enslaved the indigenous inhabitants, kin and nonkin groups came voluntarily together under leaders who could provide both internal political cohesion symbolized (if not enforced) by law codes and military organization directed against other similar embryonic states. It is true that the earliest of these kings claimed legitimacy not on grounds of lineage but as descendants of gods, and this perhaps gives some support to Sagan’s claim that the psychological appeal of a transition from chiefdom to kingship is bound up with a need to “identify with a monarch who seems capable of omnipotence.” But myths of divine descent are neither a necessary nor a sufficient condition of the successful establishment of a monarchical state. They may embellish and strengthen the kingly role, but they do not create it.


None of this diminishes the importance of the part that leaders with exceptional gifts of personality and temperament may play in breaking free of traditional ties of kinship and custom: Shaka, whom Sagan rightly singles out for his achievement in taking Zulu society from chieftainship to empire in a single generation in the early nineteenth century, is one of them. But when Sagan lumps together Nyungu-ya-Mawe of what is now western Tanzania, Kamehameha of Hawaii, Pomare II of Tahiti, Finow I of Vavau in the Tonga Islands, Alexander, Caesar, Ivan the Terrible, and William the Conqueror as “all of a piece: annihilators of people and builders of states” (and into the bargain “sadists and pirates and megalomaniacs”), he once again puts the cart before the horse. What William the Conqueror conquered was already a strong, rich, centralized monarchical state to whose throne he held himself to be the legitimate claimant. For Sagan to cite him as one of the “individuals who created the centralized state” in the “brave new world of politics that replaced the kinship system” is to betray an embarrassing ignorance of what had happened before 1066, by which time ties of kinship had already been superseded by ties of lordship and an English king like Athelstan could complain about the existence of men of “so much kindred” as to be unpunishable precisely because they were a hangover from an earlier age. In the creation of states, it is often more remarkable that the first ruler has no particularly outstanding gifts than that he has. Among the Pabir of northeastern Nigeria, for example, as studied by Ronald Cohen,2 statehood arose from the need for defense against Borno raiders and the village chief who became king was in Cohen’s words “not appreciably different from others”: lineage heads evolved from council elders into titled governmental officials without any sign of the psychological melodramatics which Sagan regards as the norm.

At this point, some readers may want to protest that I am criticizing Sagan for failing in something which he is not in fact trying to achieve. Judged by the standards of late-twentieth-century (as opposed to late-nineteenth) comparative anthropology, his speculations are, it may be admitted, amateurish and superficial. Having started by promising a “sociological study attempting to answer questions about the nature of society and cultural development on the deepest theoretical level,” he ends, after nearly four hundred pages of largely anecdotal illustration of a supposed link between the emancipation of children from their parents and the transition of societies from kinship to statehood, with the disarming confession that the question why some but not all “primitive societies” went down this road “cannot as yet be answered.” But is he, perhaps, concerned with insight rather than with generalization? Should he not be read as literature rather than science? When he wonders “why any person leaves the cozy, intimately hostile, familiar world of kinship for the cold, competitive, unfamiliar world of nonkinship politics,” is he not wondering less about causes and effects than about modes of feeling and thought?

To some extent, at least, this is surely so. When he asserts, “If the worst jingoistic expressions of patriotism frighten us, it is in part because we recognize how much our Nuer past still rules our lives,” Sagan cannot possibly mean that there is any historical connection between the forms of social organization developed among the pastoralists of East Africa who were studied by Evans-Pritchard and those developed among the European conquerors of the indigenous “Indians” of the North American continent; he is just reminding us that we are all still savages under the skin. The trouble, however, is that if judged as imaginative literature rather than social science, his book invites an immediate comparison from which it is bound to suffer badly. He does not include in his bibliography, and perhaps has never read, Canetti’s Crowds and Power. But Canetti has things to say about tyranny, obedience, sacrifice, the role of the executioner, the paranoia of rulers, and killing as the “lowest form of survival” which are directly germane to Sagan’s theme and far surpass anything in At the Dawn of Tyranny for both subtlety and memorability. Canetti is, admittedly, a writer of genius whose reputation rests on his aphorisms, among much else, and not even Professor Bellah is likely to be putting Sagan forward as a candidate for a Nobel Prize. But to anyone thinking of reading Sagan for imaginative insights about power and despotism the advice has to be: read Canetti instead. And to anyone thinking of reading Sagan for a well-tested explanatory theory of how states originate, the advice has to be: don’t.

This Issue

December 18, 1986