It sometimes seems as if explaining the outbreak of war in 1914 is a holy grail for international relations specialists in war etiology. The First World War, of course, was not a minor event in the annals of international history and that helps to explain some of its allure. Its reputation as the war no one wanted also makes it something of a magnet for scholarly entrepreneurs. Explaining the inexplicable is always a worthy challenge. Moreover, the developments that transpired prior to the outbreak of war are sufficiently complicated that almost every model ever created in international relations seems to fit. Yet underlying the whole explanatory edifice is the early and continuing search for blame, its evasion, and its former implications for postwar reparations and war guilt. Which country was most responsible for bringing about the onset of the First World War? In addition, a disproportional number of the central research foci in international relations – security dilemmas, spiral dynamics, offensive–defensive arguments, crisis dynamics, alliances, arms races – stem to varying extents from interpretations of the onset of the First World War. If we get the outbreak of this “wrong” or have overlooked significant factors, we may be heading in the wrong direction in our search for general explanations of war causes.

Over the last decade considerable progress has been made in deciphering the causes of international war and conflict. Theoretical advances have been matched by empirical evidence uncovering for instance the democratic peace phenomena. Yet one potentially important research question has gone nearly untouched. How are crises between the same adversaries related to each other, if, indeed, they are? This question is important in its own right: does previous conflict predict future conflict? But it also addresses the prevailing tendency to examine conflict events as if they were entirely independent confrontations. If previous conflict predicts future conflict, the independence assumption will be difficult to sustain. Third, the serial crisis question also addresses the very conceptual foundations of rivalry analysis. If previous conflict does not predict systematically future conflict, an important premise for examining rivalries is eliminated. The problem is that we do not know whether and how past and future conflict is related. We assume that they are related but whether they are remains an interesting empirical question and, as noted, has implications for other analytical assumptions.

Most quantitative research on international conflict has treated confrontations between the same set of states as independent events. For example, it is conventional to treat the 1961 US-Soviet Berlin Crisis as a separate and distinct case from the 1962 Cuban Missile Crisis. It would be equally conventional to treat the sequences of Egyptian-Israeli or Indo-Pakistani crises as if the latest ones had nothing to do with earlier ones.

According to Bremer (1992), contiguous dyads lacking alliance linkages and composed of less developed autocracies are more prone to war than dyads lacking these features. Less important but still significant are two other characteristics – the absence of preponderance and the presence of one or more major powers – which also contribute to dyadic warproneness. In the past decade, however, a number of scholars increasingly have turned to a different set of “dangerous dyads” known as rivalries. Instead of assuming that all dyads have an equal probability of friction, why not look more closely at the very small number of dyads that create far more than their share of the world's interstate conflict?

In this chapter, we first view rivalries through the alternative lens of Bremer's “dangerous dyads.” Do the attributes found to identify warproneness in general also inform us about which states are most likely to become rivals? This is a reasonable question to raise given some presumed overlap between rivalry initiation and war-proneness. But we anticipate that there will be a number of reasons to expect that Bremer's profile of most dangerous states, while certainly highlighting important variables, will need considerable modification for use from a rivalry perspective, especially one that does not equate militarized dispute-proneness with rivalry relationships.

We have tried to suggest in the preceding chapter that the analysis of rivalry in world politics possesses some considerable potential for revolutionizing the study of conflict. Rather than assume that all actors are equally likely to engage in conflictual relations, a focus on rivalries permits analysts to focus on the relatively small handful of actors who, demonstrably, are the ones most likely to generate conflict vastly disproportionate to their numbers. For instance, strategic rivals, a conceptualization that will be developed further in this chapter, have opposed each other in 58 (77.3 percent) of 75 wars since 1816. If we restrict our attention to the twentieth century, strategic rivals opposed one another in 41 (87.2 percent) of 47 wars. A focus on the post-1945 era yields an opposing rival ratio of 21 (91.3 percent) of 23 wars. Moreover, their conflicts are not independent across time – another frequent and major assumption in conflict studies. They are part of a historical process in which a pair of states create and sustain a relationship of atypical hostility for some period of time. What they do to each other in the present is conditioned by what they have done to each other in the past. What they do in the present is also conditioned by calculations about future ramifications of current choices.

A variety of factors have constrained progress on deciphering escalation to war. One is the sheer complexity of the undertaking. International relations are hardly simple and uncomplicated phenomena. Laws of interstate interaction are not readily established and, in any event, will probably prove to be the wrong goal to pursue in unraveling highly complex behavior. A second factor that derives from the complexity is that we tend to focus our explanatory efforts on favored, relatively monovariate, causal elements. One group pursues alliances, another dwells on arms races, and still others specialize in the analysis of rivalry or crises. There is much to be gained from concentrating on specialized topics. But there is also a price to be paid if we remain concentrated too long. Conflict escalation, without doubt, involves multiple causation.

To develop satisfactory explanations of processes subject to multiple stimuli, we need to introduce multiple explanatory factors simultaneously, and possibly interactively, as opposed to separately. However, a third factor that has held us back is that our explanatory apparatuses tend to be linked to the preference for monovariate explorations. We have tended to study individual hypotheses about alliances, or arms races, or crises. Do alliances tend to be followed by war? Are arms races genuinely influential in escalation processes or merely symptomatic of underlying tensions? What types of crises are most likely to lead to war? These are all relevant questions but we need to test more complex arguments that make linkages among alliances, arms races, and crises.

The beginnings of new millennia suggest, accurately or otherwise, openings and opportunities for new forms of behavior. We enter the twenty-first century with several recent and dramatic legacies in interstate conflict. The Soviet-American Cold War ended. Russia and China managed to de-escalate their feud at almost the same time and, of course, not coincidentally. Southern Africa fairly quickly changed from a highly conflictual and potentially explosive region to one characterized by unusually pacific interaction. Several components of the conflicts linking Israel and its immediate neighbors have been defused at least temporarily and work continues intermittently on reducing the remaining conflict issues. In South America, the ABC powers seem more interested currently in economic integration than in maintaining their traditional rivalries. Argentina and Britain seem most unlikely to fight again over the Falklands and even Ecuador and Peru have devised a formula for de-escalating their long-running conflict.

At the same time, though, the Indo-Pakistani feud refuses to go away. The United States, China, and Russia often seem on the verge of resuming one or more of their former rivalries. China and Taiwan continue to exchange missiles and declarations of independence. North and South Korea persist in maintaining their hostile divided status. The boundaries of Iraq remain potentially elastic, with a number of implications for its rivals. Turkey and Greece remain at loggerheads, although their hostility may also be waning.

Predicated on the finding that most wars stem from a relatively small number of rivalries, some analysts are shifting attention away from general probabilities of state conflict behavior escalating to war and toward modeling why rivalries sometimes (de)escalate in hostility. In this chapter, we reexamine and extend one such effort that is predicated on the assumption that territorial disputes provide the fundamental motor for escalation to war among rivals. Vasquez's (1996) model differentiates two different paths to war for contiguous and noncontiguous rivals. While the model is attractive in many respects, it is doubtful that the major power subsystem, Vasquez's empirical focus in the 1996 study, is the most likely place to find empirical corroboration for an argument stressing territorial issues. Major powers do quarrel over territory but they also have a marked propensity for competing over positional issues. Accordingly, we develop a second theory that allows for variable interests in two types of issues – spatial and positional. Many, but not all, of the same hypotheses that Vasquez derives from his single-issue theoretical focus can also be deduced from the two-issue theory, but now both paths receive equal theoretical attention. Moreover, three more testable hypotheses pertaining to the prevalence of contiguous/noncontiguous rivalries, dyadic versus multilateral wars, and war-joining behavior can also be derived. We can also test the previously assumed linkage between contiguity and territorially focused rivalries. These several new hypotheses and the added explanatory power facilitate choosing between the two war escalation theories.

Underlying the interest in the role of rivalry processes as antecedents to interstate conflict is the simple idea that conflict within the constraints of rivalry works differently than conflict outside of rivalry. Rivalry is by no means necessary for conflict to occur but conflicts associated with rivalry processes erupt with a great deal more historical and psychological baggage than is likely in the absence of rivalry. Rivals have a history of conflict, often over the same issues. Vengeance for past defeats and worries about the probability of future defeats, therefore, intrude into the decision-making processes. Compared to nonrivals, rivals have more reason, whether accurately or not, to mistrust the intentions of their adversaries. They have had time to develop images of their adversaries as threatening opponents with persistent aims to thwart their own objectives. If rivals offer concessions, why should such offers be viewed as anything but attempts at deception? Concessions and movement toward some middle ground, accordingly, are more difficult to attain. The ultimate outcome is that rivals are less likely to settle their disputes than nonrivals. Rivals give themselves fewer opportunities to exit a conflict trajectory, once they are on such a path, than do two opponents coming into conflict for the first time, and, therefore, are also likely to generate more than their fair share of conflict and violence.

The passage above is an excerpt from a contemporary account of a Sumerian border conflict between Lagash and Umma that persisted roughly between 2500 and 2350 BCE. After the war described in the excerpt, the boundaries were redrawn by the winners only to be repeatedly contested by Umma, at least according to the Lagash version of events. Undoubtedly, this Lagash-Umma conflict was not the first territorial squabble between states but it is probably the first one on which we have some documentation. Since that time, states have multiplied, as have their borders, and so have their consequent disagreements about where those boundaries should be placed.

Some 4,350 years after the Lagash-Umma conflict, we have learned a number of things about the role of contested territory in the escalation of interstate conflict. For instance, there is little controversy that contested territory plays a central role in the escalation of force and the onset of war.