This book, or something very like it, was accepted for publication about six years ago, and its final, long-delayed appearance is welcome. It is a study of despotic rule by terror and violence, a form of government all too familiar both from the pages of history and the newspapers of our own time. It is a topic which, in recent times at any rate, has received surprisingly little attention from the academic analysts of political thought. On a world scale there have been many famous writers—Kautilya, Han Fei Tzu, Machiavelli, von Clausewitz, to name but a few—who have taken it for granted that a ruler is one who aims at total domination and that he is subject to no moral constraints of any kind. Even Max Weber, the prophet of rational legality, made a point of expanding Trotsky’s formula, “Every state is founded on force (Gewalt)” into “Every state claims for itself the legitimate monopoly of physical violence (Gewaltsamkeit).”

But our softer Western writers, with their utilitarian notions of social contract and of a general will motivated to achieve the greatest happiness of the greatest number, have tried to sell the idea that, in all “normal” societies, government is by consent. In contemporary newspaper language the word “terrorism” nearly always refers to illegitimate activity. It is assumed that the legal government maintains its authority by cultivating a consensus of respect for established law and order; the terrorist is part of a subversionary organization (such as the Stern gang in mandated Palestine) which tries to destroy that authority by systematic resort to violence. This type of terrorism is quite explicitly not within the field of our author’s present enquiry. It is Weber’s legitimate Gewaltsamkeit that he is after: “The central problem of the book is to find out why some men who already have authority choose to rule by violence and fear.” (p. vii)

However, instead of attempting a close study of despotism in the contemporary world, Walter proposes an extended examination of historical case histories. He envisages a subsequent volume which will concern itself with governmental terrorism in more complex societies, but the present book is designed to illustrate a general theory, and it does this by drawing on European accounts of pre-colonial tyrannies located south of the Sahara. This has the ironic consequence that the role of South Africa’s arch-villain is played not by Dr. Verwoerd but by the early nineteenth-century Zulu King Shaka.

Walter’s choice of materials does not stem from any subliminal bias against “savages” but from a desire to achieve detachment and simplicity. His point about detachment is that Western social scientists tend to be so thoroughly committed to a belief in the moral superiority of voluntary forms of cooperation that they almost take it for granted that terrorism could only be a pathological solution of last resort:

They come up with judgments that resemble the following improvisations: “The crisis in France was so acute that Robespierre had no choice but to initiate the Reign of Terror”; or “The only way for Stalin to modernize Russia and to open the Siberian wilderness was through terror and forced labor.” To the contrary, the argument in this book, based in part on the histories of simple societies, explains the policy of terror as a response to crises of integration and as one social choice among alternatives. The social choice is neither an inescapable event nor a solitary decision. [p. 282]

The case for simplicity is more shaky. Walter claims that:

In that setting [of simple societies] the problem can be stated in plain terms. The variables are relatively easy to grasp, and one does not have to disentangle factors introduced by changing economic forces, elaborate stratification, and other complexities of modern social structure. [p. viii]

There is a touch of Rousseau about this and at least a whiff of Hobbes’s murderous war of all against all. Although Walter is not himself an anthropologist he is impressively well-read and up to date in the literature of that genre. But his interpretation of the evidence is often very different from that of a professional. For example, when writing about West African secret societies he takes his cue from a remark by a heavily biased medical missionary—“the final secret of the Poro was frightfulness” (p. 88)—and then discusses the much more restrained and informative comments of the anthropologists from this orientation. And so it is with the main Zulu case also. Although he has gone back to the original sources for the details, he ends up, it seems to me, by getting a good deal of it wrong.

The view that pre-literate societies are everywhere simple and uncomplicated is, of course, very widely held, but complexity is a relative matter. The surface of the moon is simple or complex according to whether your range of interest is that of a poet or that of an astronaut, and in some contexts the simplicity of simple societies is no more than a convenient fiction for those who do not want to look too closely at the facts. Practicing field anthropologists are all too well aware that the more they learn the more complicated it will all become.


Walter seems to suppose that there is a Weberian ideal type of institution which can be labeled terroristic despotism and that despotism in primitive societies necessarily contains the stark essentials of this institution in more clearly distinguishable form than do the despotisms encountered in more sophisticated contexts. Therefore, if we study Shaka’s Zulu despotism in detail we shall be able to see what is typical of the general case. Unfortunately, although the Zulu case was most certainly a terroristic despotism, it had a number of characteristics which were quite peculiar, so that, by generalizing from this one instance, we can be led on to false conclusions.

Part of the trouble is that Walter shows too great a trust of his original observers, most of whom had only a very impressionistic understanding of the African societies which they were describing. Their contacts were with the leaders of the communities concerned and, for the most part, they worked through interpreters. Even those few who spoke the native languages never moved far from the royal court. Consequently they were particularly ill-informed about precisely those areas of African social life on which the modern social anthropologist has made himself especially expert, namely kinship and the land-rights of the commonality. In these circumstances, the apparent fit between Walter’s model and Walter’s history is simply a by-product of lack of information. The model derives in a quite straightforward way from Weber:

In terroristic states, the agents of violence are structurally detached, often living apart and usually organized as independent social units—armies, corps of executioners, alien mercenaries, special police, etc. In the Nazi state, the SS and the Gestapo were organized as specialists in terror, even dissociated from other staffs of violence such as the army and the police. [p. 342]

and from this Walter infers that:

Therefore, the kind of cooperation in which people associate and work together without being friends or trusting one another in the first place would endure the terror best. Since they are already cooperating without friendship and trust…. Curiously, then, a society in which people are already isolated and atomized, divided by suspicions and mutually destructive rivalry would support a system of terror better than a society without much chronic antagonism. [pp. 342-3]

Now it is apparent from various passages in his book (e.g., pp. 5, 282, 341) that Walter considers that the Soviet Union, Communist China, and Cuba, are all countries in which despotic terrorism flourishes, or has recently flourished, so we must infer that these are societies in which, from the start, the population has been “divided by suspicions and mutually destructive rivalry.” But all the evidence goes the other way. It would be difficult to devise a convincing cross-cultural scale of comparative friendliness and kinship solidarity, but I should myself have supposed that, however such a scale were constructed, the Russians, the Chinese, and the Cubans would be bound to come very near the top. So there seems to be something wrong somewhere.

Walter claims that his African evidence supports his thesis. He admits that the nineteenth-century British description of the Zulu King as a “bloodthirsty, barbarian despot with absolute control of a celibate manslaying machine” “bore scant resemblance to the complex realities of the Zulu state” (p. 234) in the form it had reached by 1879 (about which we know a great deal), but he implies that it might have been quite an appropriate description of the first two Zulu despots who ruled from 1818-1840 (about whom we know relatively little). He claims that these despots “took pains to control sexual behavior, marriage and the family and to reduce the social power of kinship. The later Zulu kings, working in conditions of limited monarchy, cooperated with the pressures to restore the traditional rights to marry. The family eventually returned to the province of custom, and kinship institutions recovered authority” (p. 332). Evidently we are expected to believe that in Zulu history there was a critical period of some twenty years during which the family had ceased to form a part of the “province of custom” but that shortly afterwards the situation was exactly reversed. If Walter’s arguments were correct it is surely very surprising that throughout the whole of the period under discussion the Zulu adhered to a rule of strict clan exogamy, practiced polygamy, and had an absolute prohibition against divorce!

The point is this. In the pre-despotic phase of the Zulu polity, the Zulu political system was much like that of many other African peoples; kinship groupings determined by common descent were co-terminous with political groupings. During the era of despotism this was not the case; the kings recruited their regiments (“armies, corps of executioners, alien mercenaries, special police etc.”) from right across the board without regard to kinship affiliation, and it was these organizations which attracted the attention of contemporary European observers.


During their period of military service the men were not allowed to marry. Their future wives were maintained in “seraglios” in the vicinity of the King’s court, and the sexual favors of the latter were under the strict control of the King and the King’s mother. The age of marriage for both sexes was delayed, but in the case of the males this delay was probably no greater than would normally be the case in any context where men commonly had more than one wife. The “seraglios” of unmarried girls were there to provide the unmarried warriors with some reassurance that the wealthier polygynists were not scooping the whole pool of available women. Indeed, in spite of Walter’s assertions and the misrepresentations of the contemporary chroniclers, there is really no evidence to suggest that in Shaka’s day, in Zulu society as a whole the institution of marriage was under attack.

One peculiar feature of the case which has evoked considerable comment was that, although both Shaka and his successor Dingane indulged themselves with concubines, neither took wives or allowed his concubines to produce heirs. Walter is justified in arguing that this seems to have been an act of policy rather than a consequence of psychological abnormality (as has been suggested by Max Gluckman); but he then seems to imply that this is the typical behavior of terroristic despots everywhere and thereby misses the point, which is spelled out quite explicitly in his source material. In the context of Zulu society a warrior was always a bachelor; therefore, the warrior king had to show himself as also a bachelor. It is as simple as that. Walter’s inference from the rules of sexual constraint imposed on the bachelor warriors is that kinship had ceased to matter, either socially or politically; but the evidence, as an anthropologist would read it, points in exactly the opposite direction.

Moreover if, like Walter, we are ultimately concerned with a hypothetical general type of despotism rather than with the special variety exemplified by the Zulu, we ought to be impressed by the fact that a consistent feature of all Asiatic cases (as well as many of the African versions) was the enormous size of the harems maintained by the ruling monarchs. Unlike the consorts of the Zulu Kings and of the Turkish Sultans, these ladies were mostly wives, not slave concubines; they maintained their kinship connections with their parents and their siblings so that the network of marriage claims and obligations centering on the person of the king ramified out in all directions all over the kingdom. The great importance of kinship through marriage in such contexts makes me highly skeptical about the possibility of delineating the characteristics of any general type. The Zulu and Asiatic varieties of terroristic despotism differ quite radically from one another and also from all the varieties of totalitarian regime which we have encountered in the present century. But Walter entirely ignores all such distinctions.

All the same, his book deserves close attention. The conventional assumption that despotic terrorism is a wholly aberrant form of government is statistically false; it is a common form of government. This poses for the political theorist a serious question (to which he usually offers no answer): How can we account for the fact that such regimes which, from any rational viewpoint seem totally unjust, are not only tolerated but supported with passionate enthusiasm? It is to Walter’s credit that he has faced up to this issue.

One simple-minded explanation is that, in the short run, terroristic despotisms are often very successful. Like others of his kind Shaka engaged in predatory warfare against all his neighbors as a matter of calculated policy. That he invariably won his battles was a consequence partly of the superior discipline of his troops but in a larger measure of superior morale. The defeated enemy were systematically massacred—men, women, and children—and knowledge of this prospective fate took the heart out of the opposition before battle was joined at all. This is a pattern which is familiar from other periods of history. In 1222 in the course of Genghis Khan’s expansionist phase, the great Turkish city of Herat (in Afghanistan) raised the flag of rebellion. In reprisal the Mongols massacred the entire population, allegedly over one and a half million people. The demoralization of the populations further west was so great that two years later the Mongols were able to conquer the whole of modern Iran and Iraq with a small army of about 3000 men.1

But victory in battle doesn’t explain the loyalty of political advisers under threat of instant execution. Consider, for example, the following astonishing eyewitness account of Shaka’s courtly routine:

Cattle and war formed the whole subject of his conversations; and during his sitting, while in the act of taking a pinch of snuff, or when engaged in the deepest conversation, he would by a movement of his finger, perceivable only by his attendants, point out one of the gathering sitting around him, upon which, to the surprise of strangers, the man would be carried off and killed. This was a daily occurrence. On one occasion I witnessed 60 boys under 12 years of age dispatched before he had breakfasted. No sooner is the signal given, and the object pointed out, than those sitting around him scramble to kill him, although they have good reason to expect the next moment the same fate themselves, but such apprehensions are far from their thoughts; the will of the King being uppermost.

I have seen instances where they have had opportunities of speaking while being carried off, but which they always employed in enthusiastically praising the heroic deeds of their King. [pp. 134-5]

This pattern too has its parallels from less “primitive” contexts. The following comes from a reliable account of the mid-nineteenth-century Court of the King of Burma:

Sudden deaths were not at all uncommon in the late king’s reign. An official displeased him in some way, and Mindohn Min said emphatically, “I don’t want to see that man any more.” The poor wretch left the royal presence to be seized by lictors outside and killed more or less rapidly. A day or two afterwards his majesty would ask where so-and-so was. “Alas! sire,” was the answer, “he died of chagrin shortly after the lord of the earth and ocean cast eyes of displeasure on him.” Then the Convener of the Fifth Great Synod quoted a pious saw from the Lawkaneedee, and turned his mind to other matters. He made it a special boast that never in all his reign had he ordered an execution. Yet many people died of “official colic” during the time he was on the throne, and Colonel Sladen arrived sixty seconds too late with a respite for one of the pious monarch’s own sons. [Shway Yoe, The Burman; His Life and Notions (1896), p. 486]

Why on earth should any population of sane human beings submit to such treatment? To those of us who have been brought up to believe that any political structure is necessarily a complex mesh of checks and balances it all seems very odd.

Walter discusses the issues involved with considerable critical detachment. He recognizes that the performances put on for the benefit of European observers contained an element of bravado and dramatization; the monarch was not so arbitrarily omnipotent as he appeared, the views of his counselors carried some weight. Walter’s final explanation depends on an extreme form of the organic analogy. The Zulu, we are led to believe, had, somewhere along the line, taken a crash course in Hobbes’s Leviathan, thus:

The underlying fiction—one is tempted to say, the single principle that makes the system intelligible—was that the state was like a single body: the body of the despot, moved by his feelings, fluctuating according to his emotional vicissitudes, regulated by his will. This principle, acted out by the Zulu but not formulated in so many words, was not merely a descriptive analogy or an organismic metaphor, but a mode of orientation and a mainspring, giving the state its structure and its destiny. Moreover, it had the force of an implicit Grundnorm, a touchstone of legitimacy. Therefore, it is not enough to say that the state was “like” the ruler’s body, but rather that according to the rules that may be inferred from the phenomena of collective life, the Zulu felt that the state ought to be the body of the despot—who was imagined as the great destroyer-provider—responding to his emotions and controlled by his will. This notion of the ruler and of the nature of his power was a collective idea, a social pattern of expectations to which the despot was bound and to which he was returned when he strayed from its demands. Thus violence and caprice were not simply his idiosyncratic deviations from legitimate rule, but essential elements in the living “constitution.” [pp. 256-7]

I do not know enough about the Zulu materials to know whether the evidence can be made to fit, but if this is intended as a general explanation then it won’t work. The Burmese King, for example, was a charismatic leader in his own right, not a personification of a mystical entity “the State.” Nor could the argument possibly apply to the case of Mohammed Tughluq, “the Wonder of the Age” (King of Delhi 1325-1351). His empire, which was entirely his personal creation, at one time embraced the whole of India but had dissolved again before his death. His India was not a unitary national state. Yet although Mohammed was the most gifted and accomplished prince of his age whose court became the intellectual center of the Moslem world, his terroristic exhibitionism was every bit as drastic as that of Shaka. It included the following:

his nephew…was flayed alive. His skinned body…was paraded through Delhi…His skin was stuffed and exhibited publicly. Rumor said that the nephew’s flesh was cooked with rice, and that Mohammed gave some of this to the dead man’s wives and some to the elephants. The stuffed skins lived long in memory, and centuries after his death one of Mohammed’s titles among Hindu story tellers was “lord of the Skins of Kings.” [A. Watson, The War of the Goldsmith’s Daughter (1964), pp. 51-52]

The trouble with general theories, such as that advanced by Walter, is that they need to take account of such a diverse variety of particular cases.

Actually, so far as I can see, the concept of a State is a redundancy in Walter’s model. The essence of the matter is the charisma of the leader as such and the religious devotion of his disciples. The leader need not personify anything at all except himself—he is God. Walter contrasts this idea of personal devotion (which he considers to be characteristic of a despot’s court) with the Western liberal idea that the state functions as a rational system of cooperating agents and agencies. The former, he suggests, might be expressed in some such formula as the following:

We, the people of the Despot, including the greater and lesser officers chosen by him and the formerly sovereign chiefs who have submitted to his will, rejoicing in the terrible majesty and radiant grandeur of the Master who has thrown his shield over us, in order to form a more perfect union, to conquer all peoples who do not submit, and to gather the wealth of the land which is his by right and by power, do pledge our lives, our honor, our children, and our efforts to magnify his glory, to immolate our wills in his service, to render ourselves, all people, and all things to dust at his command. [p. 258]

Fair enough. But this is still description rather than explanation, and Walter has the ambition to show us how it came about. Here he is much less successful. A political despotism is only one possible form of political integration; a rational bureaucracy is another, a decentralized federalism is a third, and so on. Since nineteenth-century Africa in fact yielded a variety of such forms Walter sets out to demonstrate that:

the alternative solutions to problems of integration were not inexorable products of historic events or in any sense inevitable. They were deliberate social choices. [p. 292]

But, as always in politics, the history of cause and effect proves to be sticky going, the more so since, at this point, Walter finds himself embroiled in a somewhat unreal academic debate which has sporadically engaged the interest of a number of distinguished anthropologists over the past fifteen years. In 1954 Gluckman committed himself to the view (with special reference to the Zulu and their neighbors, the Swazi) “that a periodic civil war was necessary to preserve national unity: sections fought for the kingship and not for independence from it.” Schapera writing two years later took exactly the opposite view, stressing the notable frequency of secession among all the Bantu-speaking groups of southern Africa.2 Played at third hand by Walter who worries over the integrational function of terrorism, the Gluckman/Schapera argument can be made to suit all comers depending on whether their sympathies are with those who stay behind and become integrated with the unitary state of the all-powerful conqueror or those who run away and establish a fragmented existence alongside other groups of refugees somewhere else.

Thus Walter’s thesis that what he calls the segmentary state was an alternative to terroristic despotism resulting from “a deliberate social choice” is macabre. It was simply the immediate consequence of devastation. Walter discusses a number of other “alternatives” very briefly—the Swazi, the Nupe, the Ankole, the Bemba, and so on—but Africa is a very large chunk of the map, and since these varieties of state organization do not form part of a single pattern their brief consideration in this context does not throw any light on why a tyranny such as that of Shaka should have aroused enthusiasm in its own day or why it should be viewed in nostalgic retrospect as a golden age.

So we end up pretty much where we began. Walter has posed an important question too long neglected by political theorists, but I cannot agree that he has offered us any kind of an answer. Right at the beginning (p. viii) he claims that “from the dynamics of the political communities examined in this book, we may abstract typical features that will help us to understand the process of terror in different environments,” but this is not how it works out. The “typical features” which are supposedly abstracted from Zulu despotism do not, in fact, arise out of the dynamics of African politics; they are importations from our Western experience imposed on ethnographic data which Walter does not adequately understand. Certainly it is an interesting book, but it does not fulfill the promise that it will have “the advantages of simplicity without the vacancy of an abstract model.”

This Issue

January 1, 1970