In response to:

Are Tyrants Necessary? from the December 18, 1986 issue

To the Editors:

So mired is he in predominantly irrelevant historical data that W.G. Runciman, in his review of my book At the Dawn of Tyranny, [NYR, December 18, 1986], manages never to address the theoretical problems raised by the book. In some primitive societies we have much evidence of tyranny expressed against women and against children (gender and generational tyranny) but we get no oppression of adult men by other adult men, no political tyranny in the narrow sense. Runciman contends that there are many different roads to the beginnings of state society, but he does not explain why it is that every state society, certainly every state before the existence of democratic states, practices some form of political oppression. He does not address the question of why every early state is a monarchy, why every early state creates social stratification and divides the world into empowered and unempowered classes, when there had been no classes in kinship-system societies. He does not discuss the fascinating question of why human sacrifice—which hardly exists in primitive society, nor in Archaic civilizations—is rampant in complex (early state) societies and is intimately involved with kingship.

My explanation of these phenomena is that the separation from the kinship system creates an almost intolerable burden of psychic anxiety which can be contained only by investing political power in an omnipotent masculine monarch. Also, that human sacrifice is a mechanism of defense erected against that same anxiety. Once a fundamental transformation of the kinship system has been accomplished—as in Archaic civilization—both human sacrifice and this exaggerated conception of tyrannical monarchy are unnecessary.

Runciman passionately disagrees with the whole approach, and he has his own view as to how we are to understand these awesome human problems. Read Elias Canetti, who is a genius and a Nobel Prize winner, he tells us. Runciman would have done a much greater service if he had spent more time giving some hint as to what Canetti’s explanations of tyranny, the paranoia of rulers, and sacrifice are, instead of the fanciful historical tales of Aristotle and Herodotus. Certainly, there is no more important intellectual task than the explanation of these oppressive human phenomena. All we are told is that killing is “the lowest form of survival.” One would have liked to hear more about how that makes Auschwitz and Hiroshima comprehensible.

And Runciman’s historical analysis is inadequate to the task because it fails to make the crucial theoretical distinction between the origins of (A) The state and (B) A state. He does not differentiate between (A) The pristine state and (B) A successor state. He does not really address, therefore, the issue of the appearance of a state society in the circumstance where there was no state before; i.e., the origins of the state out of a kinship-system society. And this is precisely the central theoretical problem addressed in the book.

The difficulty, unfortunately, is that Runciman has practically no knowledge of the vast literature on the origins of the state that has accumulated over the last twenty-five years. He manifests no acquaintance with the work of Fried, Service, Sahlins, Claessen, Skalnik, etc. Lacking this, he cites four historical examples which are intended to demolish my theoretical conclusions. Three of the four are irrelevant to the argument: 1. The Greek city-states, and 2. The first Mesopotamian states, did not develop out of kinship-system societies. They were not pristine states. 3. Herodotus’ fanciful folk-tale about the Medes is not even history, as Runciman himself acknowledges.

  1. The one circumstance Runciman does discuss that is about the origin of a pristine state, is that among the Pabir elaborated by Ronald Cohen. This is a complicated issue but it is worth discussing because it reveals the rather superficial quality of Runciman’s analysis. Cohen’s description of the rise of the Pabir state is not clear. At one point he writes that the pressure from the Borno state forced the Pabir kingdom “to evolve towards more centralized control and more hierarchy,” which would imply that the Pabir state was already established before the Borno pressure. But in another part of the discussion, he indicates that the state arose in response to this pressure, which statement Runciman picks up on. In order to clarify the situation, I spoke to Cohen on the telephone on December 12. Cohen explained to me that the kingdom evolved in a two-step process. Firstly, in the attempt to fight off the conquest pressure of the Borno state, village headmen were elevated to the position of chiefs. This was a peaceful, non-violent process and is the circumstance that Runciman gleefully pounces on to prove the absurdity of my notion that violence and tyranny are necessary for the origins of the state. The second step in the process, however, as Cohen tells it, is that the chiefs warred with each other and that the one chief who ultimately proved triumphant was proclaimed king. After the monarchy was established, Cohen’s text asserts, “clear-cut lines of stratification” appeared. The heads of prestige lineages, he told me, turned into nobles (i.e., an aristocracy was created). When I asked Cohen if he knew of any early state that was not a monarchy, he responded that he could not think of one, and that even if he could recall one or two, that was not relevant, because “statistically the statement is accurate.”

Runciman cannot distinguish between the origins of chiefdoms and the origin of kingdoms. He pays no attention to the first chapter in the section of my book entitled “The State as a Work of Art,” wherein I list seven stages in the process of political development from kinship society to a full-blown state: headman—transition—chieftainship—transition—simple kingdom—transition—complex kingdom. Chieftainships are not states, I contend, because kinship forms of social cohesion are still primary in that stage of social evolution. The early centralized state requires a monarchy. “Historical information indicates that every complex centralized monarchy was built on a foundation of military conquest. Small kingdoms were erected as a result of peaceful expansion, amalgamation, the consent of the governed, but for the centralized state this seems never to have been the case” (p. 316). Runciman’s obligation, it would seem, is to disprove that statement with reliable historical data taken from circumstances where the pristine state grows out of the kinship system.

I return to the questions of why every postkinship society practices political oppression, why human sacrifice seems to be a necessary symbolic form to break and transform the kinship system. I am criticized for “my one-sided overemphasis on psychology.” Yet, when praising Canetti’s work, Runciman emphasizes his psychological insight into “tyranny, obedience, sacrifice, the role of the executioner.” I suspect, therefore, that what the reviewer finds intolerable is the psychoanalytic basis of my theoretical approach. Runciman would imply, though he does not say, that having read Canetti, no one has need any longer to address the work of Freud, Menninger, Fromm, Lifton or Dinnerstein. Canetti, the reviewer insists, is “all ye need to know” about the significance of human destructiveness. Runciman’s view would seem to be that a psychoanalytic analysis of society and culture has no intellectual validity. We can be sure that his will not remain the last word spoken in that argument.

Eli Sagan

Englewood, New Jersey

W.G Runciman replies:
  1. “Mired…in predominantly irrelevant historical data” means, I am afraid, that I know a number of awkward facts in the historical record which cannot be squared with Sagan’s would-be general theory about the origin of states.

  2. I do not have to explain why state societies practice “some form of political oppression” or why every early state “divides the world into empowered and unempowered classes” since both these propositions are true by definition.

  3. I do not have an explanation to offer for human sacrifice, but nor do I see how an explanation, psychoanalytic or otherwise, could rescue Sagan’s theory of the origin of states.

  4. Pristine states are indeed very often monarchies, but not always. Those that are not, however few, cannot be explained in Sagan’s terms and those that are can be better explained in terms other than his.

  5. I brought Canetti’s Crowds and Power into my review as an example of the value of imaginative insights as opposed to social-scientific hypotheses about the psychology of power. To the extent that Sagan is (unlike Canetti) doing social science, he fails to substantiate his case. To the extent that he is (unlike Fried et al.) offering imaginative insights, Canetti does it enormously much better.

  6. Sagan’s assertion that I have “practically no knowledge of the vast literature on the origins of the state that has accumulated over the last twenty-five years” is as impertinent as it is inaccurate. The authors whom Sagan lists are precisely those I would recommend to anyone who wished to see the topic discussed according to standards of evidence and argument which Sagan fails to meet.

  7. I do not see how cases where one of several rival chiefs becomes a king by successfully monopolizing the means of coercion support the claim that kingship arises as a response to the “burden of psychic anxiety” imposed by “separation from the kinship system.”

  8. Likewise, the undisputed connection between conquest and “complex” monarchy can be accounted for (and my review gave some examples) without invoking “psychic anxiety”; and in any case, it often occurs at a later stage than the initial “separation from the kinship system.”

  9. I do not deny that psychoanalytic theory may be able to contribute to our understanding of society and culture. But this needs to be demonstrated more effectively than Sagan has managed to do.

This Issue

April 23, 1987