models or by reference to other models already available to them. ‘Seeing’ markets
inaction reinforces market behaviour in turn validating markets as a model and
reinforcing them as institutions.
As I have just remarked, additional theories offer a fuller picture. They also cast
doubt on the notion that the spontaneous emergence of markets means that they
need not really exist to have real effects. From the time of David Hume and Adam
Ferguson, observers have grasped the institutional effects of agents’ self-interested
choices, even if agents do not.^49 Aswith markets, so with anarchy. Following John
Ruggie’s review ofTheory of International Politics, few scholars would venture to say
that the Western state system emerged spontaneously, whole cloth out of whole
cloth.^50 There is a mountain of recent scholarship explicating the system’s institu-
tional moorings, as if any student of international law, at least since Emmerich de
Vattel, could have doubted the extent and significance of the system’s institutional
features. And there is a sizeable literature on the social construction of sovereignty.^51
On Waltz’s account, anarchy is a model of the international system. A theo-
retically enhanced account would construe anarchy as a model and, at the same time,
an institution that system observers and states’ agents have made together. States’
agents are themselves system observers; models mediate whatever all observers (and
not just states’ agents) think they ‘see’; agents acting on what they see give rise to
institutions; observers take these institutions into account in forming models and
acting on their conclusions (whether reporting on them as observers or doing some-
thing about them as agents). In the process, models (structural models, constitutive
models) and institutions ‘structured’ by models (anarchy, sovereignty), lend each
other credibility.^52
Waltz seems to have been deeply, even hopelessly confused in explicating the
analogy between microeconomic theory and his structural theory. Yet his con-
cluding remarks redeem him.^53 ‘To say that “the structure selects” means simply that
those who conform to accepted and successful practices more often rise to the top
and are likelier to stay there’. To say that structure selects is to summarize the model’s
terms with a crisp metaphor made familiar by well-known evolutionary models.
What Waltz called ‘selection according to behavior’ is behaviour that accords with
the model but responds, in the first instance, to the institutional possibilities that
agents ‘see’ for themselves. Insofar as relevant models tell them what to see, their
behaviour may alsorespond to the constraints stipulated in those models.
Friedman’s ghost
Waltz’s careless tendency to move from the formal causes stipulated in his theoretical
model to institutional effects has troubled his critics less than has his careful
segregation of structural models and institutional reality. ‘[H]ow can something
which does not really exist ... “shape and shove” anything’?^54 If one believes that
structures really do exist even if they cannot be directly observed, then one never
needs to ask this question, much less engage in conceptual contortions to answer it.
And this is exactly what philosophical realists believe.
Structure? What structure? 97