(Page 3 of 33) - Reagan and Gorbachev: Altercasting at the End of the Cold War authored by Malici, Akan.

much cited “agent-structure debate” is still not comprehended sufficiently and remains
undertheorized (Buzan, Jones, and Little, 1993; Carlsnaes, 1992; Dessler, 1989; Gould, 1998;
Hollis and Smith, 1991, 1992, 1994). The reason can be traced to a lack of rigorous methods that
would enable the scientific study of ideational variables (Ruggie, 1999). Historically,
constructivists rely on interpretations as a methodological tool. Interpretations are by definition
subjective, however, which makes an effort to derive firm knowledge elusive. As one scholar
acknowledged, “Not all interpretations are equally valid,” leading to the problem of “how to
justify a claim about unobservables” (Wendt, 1999: 85).

My goal in this paper is to make the unobservables observable by engaging in analysis at the
interface of constructivism and cognitivism. The task of tracing the impact of ideational variables
on strategic interactions of states is a daunting one involving the solution to some major
methodological problems. Both cognitivists and constructivists have been criticized for
advancing underspecified models lacking rigorous conceptualization and operationalization
(Achen and Snidal, 1989; Morrow, 1997). In this paper I address these criticisms by using
methods capable of systematically and quantitatively identifying ideational variables. I also use a
model that conceptualizes ideational variables (beliefs) and preferences specified within a
rigorous framework provided by sequential game theory (Brams, 1994; Maoz and Mor, 2002).

Constructivists caution against endogenizing preferences within a game-theoretic framework
arguing that “game theory was not designed for this task and so its relevant conceptual repertoire
is relatively underdeveloped” (Wendt, 1999: 327). Yet, I argue that exactly this task ought to be
the goal – to develop the conceptual repertoire of game theory and thereby contribute to the
scientific study of ideational variables. Indeed, I argue that endogenizing preferences within a
game-theoretic model is especially promising because the latter carries positive statements about
what ought to be expected if decisions are indeed influenced by beliefs.

It is important to recognize at the outset the fundamental difference between what I argue and
traditional game theoretic approaches to problems of international conflict resolution. In
traditional game theoretic approaches interaction is regulative. Actors know what they want and
act accordingly. If they fail to achieve their goals, they engage in structural learning and adjust
their actions. I argue along with constructivists that “more is going on.” Actors are not only
trying to “get what they want,” they are also trying to sustain or change the conceptions of self
and other that generate those wants in the first place (Wendt, 1999: 316). I argue, in other words,
that interactions can also have constitutive effects via a cognitive process of experiential
learning. Here actors do not only alter actions towards each other. They also alter the beliefs they
hold about each other, which redefines the menu of rational foreign policy actions. Which form
of learning occurs – structural or experiential – constructivists argue, “is an empirical question
that needs to be investigated, not assumed away a priori” (Wendt, 1999: 334).

In this paper I will investigate this question in relation to the greatest change in recent history –
the end of the cold war. I argue that no two leaders exemplify the claim that “Anarchy is what

For previous interesting attempts to endogenize preferences within a general rationalist or game-theoretic

framework see Elster (1982), Cohen and Axelrod (1984), Raub (1990), Becker (1996) and Clark (1998).