over its relations with that remainder. The question is not to find out whether what is
retained is natural or artificial (borders) because in any event there is deterritorialization.
But in this case deterritorialization is a result of the territory itself being taken as an
object, as a material to stratify, to make resonate. Thus the central power of the State is
hierarchical and constitutes a civil-service sector; the centre is not in the middle (au
milieu) but on top because the only way it can recombine what it isolates is through
subordination. Of course, there is a multiplicity of States no less than of towns, but it is
not the same type of multiplicity: there are as many States as there are vertical cross-
sections in dimension of depth, each separated off from the others, whereas the town is
inseparable from the horizontal network of towns. Each State is a global (not local)
integration, a redundancy of resonance (not of frequency), an operation of the
stratification of the territory (not of the polarization of the milieu).
It is possible to reconstruct how primitive societies warded off both thresholds, while
at the same time anticipating them. Lévi-Strauss has shown that the same villages are
susceptible to two presentations, one segmentary and egalitarian, the other encompassing
and hierarchized. These are like two potentials, one anticipating a central point common
to two horizontal segments, the other anticipating a central point external to a straight
line.^3 Primitive societies do not lack formations of power; they even have many of them.
But what prevents the potential central points from crystallizing, from taking on
consistency, are precisely those mechanisms that keep the formations of power both from
resonating together in a higher point and from becoming polarized at a common point:
the circles are not concentric, and the two segments have need of a third segment through
which to communicate.^4 This is the sense in which primitive societies have not crossed
either the town-threshold or the State-threshold.
If we now turn our attention to the two thresholds of consistency, it is clear that they
imply a deterritorialization in relation to the primitive territorial code. It is futile to ask
which came first, the city or the State, the urban or state revolution, because the two are
in reciprocal presupposition. Both the melodic line of the towns and the harmonic cross-
sections of the States are necessary to effect the striation of space. The only question that
arises is the possibility that there may be an inverse relation at the heart of this
reciprocity. For although the archaic imperial State necessarily included towns of
considerable size, they remained all the more strictly subordinated to the State the more it
extended its monopoly over foreign trade. On the other hand, the town tended to break
free when the State’s overcoding itself provoked decoded flows. A decoding was coupled
with the deterritorialization and amplified it: the necessary recoding was then achieved
through a certain autonomy of the towns or else directly through corporative and
commercial towns freed from the State-form. Thus towns arose that no longer had a
connection to their own land because they assured the trade between empires or, better,
because they themselves constituted a free commercial network with other towns. There
is therefore an adventure proper to towns in the zones where the most intense decoding
occurs: for example, the ancient Aegean world or the Western world of the Middle Ages
and the Renaissance. Could it not be said that capitalism is the fruit of the towns and
arises when an urban recoding tends to replace State overcoding? This, however, was not
the case. The towns did not create capitalism. The banking and commercial towns, being
unproductive and indifferent to the backcountry, did not perform a recoding without also
inhibiting the general conjunction of decoded flows. If it is true that they anticipated
Rethinking Architecture 298