Adversarial Problem Solving: Modeling an Opponent Using Explanatory Coherence

In adversarial problem solving (APS), one must anticipate, understand and counteract the actions of an opponent. Military strategy, business, and game playing all requfre an agent to construct a model of an opponent that includes the opponent’s model of the agent. The cognitive mechanisms required for such modeling include deduction, analogy, Inductive generalization, and the formation and evaluation of explanatory hypotheses. Explonatory coherence theory captures part of what is involved in APS, particularly In cases involving deception. He who can modvy his tactics in relation to his opponent, and thereby succeed in winning, may be called a heaven-born captain. (Sun Tzu, 1983, p. 29) Many problem-solving tasks involve other people. Often, accomplishment of a task requires coordination with others, an enterprise that might be called cooperative problem solving. Unfortunately, however, we also face problems that require us t6 take into account the actions of opponents; this is adversarial problem solving (APS). Both kinds of social problem solving have been relatively neglected by cognitive science researchers who typically investigate people’s performance on nonsocial problems. This article investigates the cognitive processes required for APS in which one must anticipate and understand the actions of an opponent. A review of several domains of APS-military strategy, business, and game playingThis research is supported by contract MDA903-89-K-0179 From the Basic Research OFFice of the Army Research Institute For the Behavioral and SociaI Sciences. Thanks to Bruce Burns, Colin Camerer, Nicholas Findler, Keith Holyoak, Greg Nelson, and Alex ViskovatoFF For useful discussions and comments, and to David GochField For help with the simulation in Section 4.3.