• Choose a category
• All
• Classic-Fiction

By Root 1580 0

over as soon as he used the word “theory.” But they nodded approvingly when he spoke instead of the need for “generic knowledge.” It is not difficult to understand why they responded favorably to this phrase. Policy specialists recognize that many generic problems arise in the conduct of foreign policy—for example, the task of deterrence emerges repeatedly over time with different adversaries and in different contexts. Therefore, policy experts readily understand and agree that generic knowledge about the uses and limitations of every particular strategy or policy instrument can be helpful when one considers possible uses of that strategy in a new situation.526

Generic knowledge is not sufficient to determine what action to take, but it is useful to policy specialists who must first diagnose a new situation to see whether or not favorable conditions exist or can be created for employing a particular strategy. Good generic knowledge enables a practitioner to increase the chances of making the right decision about whether and how to employ a particular strategy. Generic knowledge is most useful when it identifies conditions, processes, and causal mechanisms that link the use of each strategy to variance in its outcomes.

CORRECT IMAGE OF THE ADVERSARY

The policymaker needs a correct image of the adversary whose behavior the strategy is designed to influence. Policy specialists and academic scholars agree on this fundamental point: in conducting foreign policy one must try to see events—and even assess one’s own behavior—from the perspective of the adversary. Only by doing so can the practitioner diagnose a developing situation accurately and select appropriate ways to communicate with and influence the other actor. Faulty images of an adversary often lead to major errors in policy, avoidable catastrophes, and missed opportunities.

Scholars and policymakers often assume that adversaries are rational, unitary actors. Both components of this assumption seriously oversimplify the task of understanding and influencing other actors. More discriminating “actor-specific” behavioral models are needed that recognize that an adversary is not a unitary actor, but often includes a number of individuals who may differ in important ways in their analysis of challenges and opportunities to be considered in deciding policy. Similarly, the particular rationality of an opponent may reflect values, beliefs, perceptions, and judgments of acceptable risk that differ from those of the side that is attempting to influence its behavior. Simple assumptions that one is dealing with a rational or unitary actor may be particularly dangerous when one is trying to deal with non-state actors, such as warlords, terrorists, or rivals in civil wars.527

We have identified three types of knowledge practitioners need for dealing with generic problems: general conceptual models, generic knowledge, and correct images of adversaries. We turn now to a discussion of the forms of knowledge most useful for policymaking.

What Forms of Knowledge Do Practitioners Need?

Much scholarly theory and knowledge is cast in the form of probabilistic generalizations of a broad character. These are not without value for policymaking, but leave the policymaker with the difficult task of deciding whether the probabilistic relationship in question applies to the particular case at hand. Political scientists should therefore make a move from theory and knowledge cast in probabilistic terms to conditional generalizations of more limited scope. For example, additional research is needed to transform the general probabilistic proposition that arms races are likely to lead to war into more specific contingent generalizations that will identify the conditions under which arms races are likely to lead to war.528 Conditional generalizations are more useful when they identify variables over which policymakers can exercise some leverage. Conditional generalizations may also be couched in probabilistic terms, but are more specific and more limited in scope than general probabilistic