I have constructed a data set composed of all wars since 1648 involving at least one or more great or rising powers. In this chapter, I describe the data set and justify its appropriateness to my propositions. I open with a discussion of key terms and coding rules. I then discuss my findings, which, I contend, offer considerable support for my propositions and call into question some of the foundational assumptions and claims of opposing theories.

Definitions

In the introduction, I noted that peace and war are commonly treated as dichotomous categories, although in practice they represent two ends of a continuum. In between these poles, we encounter various states of cooperation and violence. The term “Cold War” was coined to represent one such in-between state: a tense, armed peace with periodic military conflicts between superpower client states or between one superpower and another's ally (e.g. the Chinese–American component of the Korean War). Distinguishing war from peace is further complicated by the fact that they are legal categories, giving states the option of fighting wars without declaring them, as the Soviet Union and Japan did in Mongolia in 1939 and the US and China did in Korea in 1950. For purposes of my data set, I count as a war any interstate conflict that produced over 1,000 deaths independently of whether either of the protagonists considered themselves to be at war.