Beyond the democratic state: anti-authoritarian interventions in democratic theory

separate 
war from the state
…The Peace Chief was permanent, and though no autocratic 
ruler, he was often a very important personage. But he had nothing to do with 
war. He did not even appoint the war chiefs or concern himself with the conduct 
of war parties. Any man who could attract a following led a war party when and 
where he would…[His authority, however,] “…lasted only till the return of the 
war party.
Whereas Tilly argues that war-making was central to both the development of states and their 
expansion, Benedict contends that it was precisely by separating war-making from the Chief (the 
individual that was best positioned to become a nascent state), that war did not follow a 
centripetal logic. Clastres provides a similar account of both the Yanomami warrior Fousiwe 
and the Apache warrior Geronimo, both of whom tried to parlay success in an initially popular 
battle into more battles and more power. They both failed, with the former dying in a battle 
alone and the latter leading only a handful of men (Clastres 1987, 209 – 12). The key point here 
is that, in sharp contrast to centripetal logic advocated by Hobbes and explained by Tilly, war-
making functions – violence in its most organized and extreme form – were separated from the 
state. As a result: 
Primitive warfare…was neither centrally organized nor led by permanent 
chieftains; it was relatively infrequent; it was not war of conquest nor was it 
bloody war aimed at killing as many of the enemy as possible. Most civilized 
war, in contrast, is institutionalized, organized by permanent chieftains, and aims 
at conquest of territory and/or acquisition of slaves and/or booty” (Fromm 1973, 
171 – 72).
The separation of war-making from the Chief, and the radical decentralization of the capacity for 
violence that it implies, did not lead to “the impulse to conquer people or territory, to subjugate 
human beings, or to destroy the basis for their livelihood” (
ibid
. 176). In short, war did not tend 
toward centralization. In what sense, though, did it follow a centrifugal logic oriented toward 
dispersion?

---

98 
While Fromm finds a significant number of peaceful tribes along with more aggressive 
groups, Clastres (1994, 142 and 157) argues that, in the region he studies (Amazonia), “no 
primitive society escapes violence” and that “the possibility of war is inscribed in the being of 
primitive society.” And, it is in these violence-prone societies that the centrifugal logic becomes 
most clear. The argument is that Indian violence is distinct – in terms of both its different 
practices and different ends – from the unifying logic of Hobbesian violence. While there 
is
war 
for the Indian, it is not destructive war. 
Now all wars, as we know, leave a victor and a vanquished. What, in this case, 
would be the principal result of a war of all against all? It would institute 
precisely the political relationship that primitive society works constantly to 
prevent: the war of all against all would lead to the establishment of domination 
and power that the victor could forcibly exercise over the vanquished (
ibid
. 158).
Rather than a war of all against all, we see a network of alliances, but alliances that nearly 
constantly shift. Thus, Hobbes was correct that small groups are only capable of fragile and 
tenuous alliances.
12
He was wrong, however, that the result would be destructive. Rather, Indian 
societies need allies to ensure that they will not be the perpetual losers in conflict, but these 
alliances are constantly broken by betrayals – betrayals, however, that serve a specific function: 
they prevent the rise of any lasting alliances which would be capable of dominating other tribes.
Rather, these fragile alliances protect a 
balance of power
between tribes. No one tribe is so weak 
so that, without allies, it can be dominated and, yet, no tribe is so strong, that with its allies, it can 
effectively dominate. 
What ends could such a logic of violence serve? The answer that Clastres presents 
suggests that the logic of Indian violence and the logic of the Leviathan’s violence are 
12
“Nor is it the joining together of a small number of men that gives them this security; because in small numbers, 
small additions on the one side or the other make the advantage of strength so great as is sufficient to carry the 
victory; and therefore gives encouragement to an invasion” (Hobbes XVII 3).

---

99 
diametrically opposed – the former seeks to disperse, while the latter seeks to unify. Clastres 
(
ibid
. 164 – 65)
 
asks: 
What is the function of primitive war? To assure the permanence of the 
dispersion, the parceling, the atomization of groups. Primitive war is the work of 
a 
centrifugal logic
, a logic of separation, which is expressed from time to time in 
armed conflict. War serves to maintain each community’s political 
independence…Now, what is the legal power that embraces all differences in 
order to suppress them, that exists precisely to abolish the logic of the multiple 
and to substitute it for the opposite logic of unification? What is the other name 
of this One that primitive society by definition refuses? It is the State. 
While the Leviathan’s logic of violence is to unify violent capacity into a single entity, the 
Indian’s logic of violence seeks the opposite – to decentralize violent capacity so that no such 
Leviathan can emerge. “[I]t is not war that is the effect of segmentation, it is segmentation that 
is the effect of war. It is not only the end, but the goal…Primitive society wants 
dispersion…[and] primitive war is the means to a political end” (
ibid
. 153). Now the 
Leviathan’s violence provides certain benefits; perhaps most importantly, the likelihood that 
most people most of the time will not engage in violent conflicts because the Leviathan’s 
“terror” will hold potential perpetrators in “awe.” That said, as we have clearly seen throughout 
history, when statist violence erupts it occurs on an unimaginable scale. In contrast, violence for 
the Indian may be somewhat more frequent, though it will tend to be less destructive and less 
prone toward centralization – such is the cost of its logic of dispersion. While I think it is an 
important question to ask which logic of violence we prefer, that is not my aim here. Again, the 
argument in this section is not so much a defense of primitive violence as an explanation of how 
a centrifugal logic functions. The important point is simply that there is an alternative to the 
Leviathan. This alternative is a logic that seeks to disperse, rather than unify; a logic that accepts 
the possibility of frequent conflict from many directions, rather than the possibility of 

---

100 
domination from one; a logic that actively prevents, rather than actively establishes, the 
Leviathan, the Mortal God.