3
The Use of Formal Methods
In IR theory there have traditionally been three general approaches to the use of formal
game-theoretic analysis. One is as a heuristic and prescriptive tool, another is as a predictive
tool, and the third is as the language of rational choice theory. These are general categories, not
mutually exclusive groups, and historically there has been significant overlap among the groups.
But this overlap is decreasing over time, and the work within the various groups is becoming
increasingly mutually exclusive. This pattern, of the development of separate research
communities for the various uses of formal modeling in IR, is having the effect of squeezing the
heuristic/prescriptive use out of the general IR literature, for reasons that will be discussed
below.
The use of formal logic based on assumptions of actor rationality as a heuristic tool first
came to international relations theory through the work of such scholars as Thomas Schelling,
and was further popularized in the field by scholars such as Glenn Snyder and Paul Diesing, and
Robert Jervis.
2
This approach uses models to illustrate general patterns of behavior in
international politics, and to prescribe behavior that would best allow decision-makers in the
field to achieve their goals.
These models are not intended to be accurate descriptions of what is going on in
international politics, in two ways. First, the models do not assume that actors behave rationally.
They are contingent – when actors behave rationally, the predicted outcome will occur, and when
2
E.g. Schelling 1960; Snyder 1961; Schelling 1966; Snyder and Diesing 1977; Jervis 1978.
Convention