No abstract

This essay proposes dialogical analysis as a method of modeling political interactions. The method combines the formal theory of dialogical disputation, a family of theories drawn from linguistic pragmatics, and formal proof procedures. By analyzing models of their explicit and implicit contents in context, the method identifies the argumentative thrust of negotiation dialogues and shows systematically how the parties signal intent and commitment to one another. The paper illustrates the method by applying it to superpower interactions in the 1980s INF negotiations. The analysis indicates that American force deployments did not motivate the Soviet retreat from their early insistence on compensation for European missiles. The change in the Soviet position is better attributed to their strategic reconceptualization of the Cold War insecurity dilemma.I shall also call the whole, consisting of language and the actions into which it is woven, the "language-game."--Ludwig Wittgenstein (1958:5)Game theory provides a powerful set of formal tools for explaining political interactions. Yet the formality of game models sometimes limits their empirical application. Contextual factors generally held exogenous to game models may prove decisive for outcomes. These factors include actors' beliefs about the nature of the interaction, their beliefs about other actors' beliefs, and the means by which actors convey and infer intentions to and from one another. Scholars concerned with the role of norms, beliefs, and identities in social interaction (e.g., constructivists, cognitivists, negotiation theorists, diplomatic historians) could benefit from a model of social interaction that captures these contextual factors.

Constructivists often refer to the end of the Cold War to illustrate their contention that social rules are not immutable. Agents can change the rules by performing actions that undermine them. In this article, we describe the Cold War as a set of social rules sustained by superpower speech acts. We show that, by altering their behavior, the superpowers undermined the felicity of these rules. In so doing, they progressively dismantled the rules of the Cold War. Our model captures the competing arguments in the ongoing debate about whether the rationalist buildup argument or the constructivist new thinking argument better explains the end of the Cold War. Within the model, we identify the rules that, when made infelicitous by the superpowers, resolves tensions in the Cold War rule system in ways consistent with each argument. We conclude by showing how these competing arguments are reflected in contemporary debates concerning the nature of the global security rules emerging in the post-cold-war world.

Scholars who apply artificial mtelligence to political questions seek, most generally, to expand the scope and relevance of political model analysis. By incorporating the effects of variable human notions, traditions, and meanings, they seek to humanize political models. Most early applications of artificial intelligence in political science research address substantive issues pertaining to political decision making. Most of these works apply production-system technology to construct choice models in for eign-policy decision contexts. In recent years, political applications have begun to diver sify. Today, lively research efforts flourish in widely varied application areas, such as computational text analysis, logic programming, computer learning, and conflict sim ulation. The works reviewed here constitute the early steps of a nascent program of study. Much remains to be accomplished. Nevertheless, the efforts conducted thus far suggest many potentially fruitful research avenues. Keywords: artificial intelligence, production systems, logic programming, computational hermeneutics, belief models, political simulation, political science.

No abstract