29
heuristic formal models, this limit does undermine the precision with which the researcher can
determine the extent to which actor behavior approaches the ideal type of rational behavior.
This limit suggests that formal models that assume actor rationality can and should be
used in concert with other kinds of heuristic models. These other kinds include formal models
that begin with assumptions other than pure rationality, such as prospect theory.
62
They also
include various kinds of computer modeling in which agents act responsively rather than
strategically, such as agent-based modeling and complex adaptive systems modeling.
63
Yet other
kinds of heuristic modeling that can be productively used in concert with formal models include
those that do not use mathematical or computer skills, such as sociological models that often
appear in constructivist theorizing.
64
In all of these kinds of modeling, the basic logic of
heuristic models remains the same; to create an ideal-type model based on a clearly specified set
of assumptions about actor motivations and the interactions of these motivations. The measure
of usefulness of the model in each case remains the extent to which the model illuminates
international politics, rather than the extent to which it predicts or is compatible with
international politics.
The use of formal models in combination with other kinds of heuristic models in the
study of IR can in fact serve as a useful bridge between approaches to the study of IR based on
different understandings of the sources of human behavior, particularly when used for
explanatory purposes. Formal models assume that behavior is governed by cost-benefit
calculations; most computer modeling assumes that behavior is habitual, and that learning is
62
E.g. Levy 1997.
63
E.g. Axelrod 1984 and Cederman 1997 respectively.
64
E.g. Klotz 1995; Jackson and Krebs 2004. Interestingly, complex adaptive systems modelers
in IR are sometimes self-referentially constructivist, and see computer modeling as a useful tool
in developing sociological models of international politics, e.g. Cederman 1997 and Hoffmann
forthcoming.