In their theorisation of the State, Deleuze-Guattari set off from the ancient autocracies (TP, 448/560), citing favourably (if only ultimately to criticise) the work of the anthropologist Pierre Clastres (1934–1977). Clastres tried to understand primitive societies as cabals, which oppose the State through warfare. War injects a constant dispersion into the collective, interrupting the concentration of power. As a consequence, according to Clastres, the Hobbesian proposition that ‘the State opposes war’ must be reversed: ‘war is against the State, and makes it impossible’ (TP, 357/442; original emphasis).

Now Deleuze-Guattari, by abstracting even further from these cabals, extract the concept of the ‘war machine’. War machines work on a logic different from that of the State, as ‘rhizomatic’ collections of formless and disparate forces. In opposition is the ‘tree’ diagram of the State defined as an ‘apparatus of capture’, which exists in the interstices of primitive societies capturing their wealth, as empire. The State therefore exists in a state of tension with the primitive societies, and not as their extrapolation. If anything, the State bursts into being in the interstices of the primitive societies all at once. In other words, the State has always-already existed (TP, 359–60/444–5).

The problem with Clastres’ account, however, is that in his zeal to highlight the anti-State nature of the primitive societies, he ends up conceiving them as self-sufficient entities with no relation to the society of the State (TP, 359/444). It is true that primitive societies have at their disposal a mechanism to ward off the State. Nevertheless, therein also exists an undeniable vector tending towards the State (TP, 431/537). What is more, primitive societies have always interacted with each other via the State.

Deleuze-Guattari point out that anthropologists (including Clastres) do not pay sufficient attention to the results of archaeology.