The events of September 11 can be taken as a good place to explore the interaction of variables in explaining events. We know that many variables are at play in most complex situations. Taking account of them all is impossible, and even investigating several is messy. One way to tackle this problem is to pay closer attention to the interaction of variables--not to look at all of them but to explore carefully where they intersect and to detect the mechanisms of their relationships.

Many of the quarrels in our field arise out of a desire to assert aggressively the predominance of a particular, favored variable. Wearying of this sort of combat, some analysts have become more interested in examining the connections among variables, the ways in which one parameter acquires strength or signals causality because of its interaction with another. The events of September 11 show the values of this approach. It will surely involve some loss of parsimony, but it may lead to important insights and research programs. To pursue it, we need models of causality that work with interactions.