Identifying Rivals and Rivalries in World Politics
Author(s): William R. Thompson
Source: International Studies Quarterly, Vol. 45, No. 4 (Dec., 2001), pp. 557-586
Published by: Wiley on behalf of International Studies Association
Stable URL: http://www.jstor.org/stable/3096060
Accessed: 20-11-2015 13:22 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

International Studies Association and Wiley are collaborating with JSTOR to digitize, preserve and extend access to
International Studies Quarterly.

http://www.jstor.org

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

International Studies Quarterly(2001) 45, 557-586.

Identifying Rivals and Rivalries in World Politics
WILLIAM

R. THOMPSON

Indiana University
Instead of assuming that all actors are equally likely to clash, and that they do so independently of previous clashes, rivalry analysis can focus on the small number of feuding dyads that cause much of the trouble in the international system. But the value added of this approach will hinge in part on how rivalries are identified. Rivalry dyads are usually identified by satisfying thresholds in the frequency of militarized disputes occurring within some prespecified interval of time. But this approach implies a number of analytical problems including the possibility that rivalry analyses are simply being restricted to a device for distinguishing between states that engage in frequent and infrequent conflict. An alternative approach defines rivalry as a perceptual categorizing process in which actors identify which states are sufficiently threatening competitors to qualify as enemies. A systematic approach to identifying these strategic rivalries is elaborated. The outcome, 174 rivalries in existence between 1816 and 1999 are named and compared to the rivalry identification lists produced by three dispute density approaches. The point of the comparison is not necessarily to assert the superiority of one approach over others as it is to highlight the very real costs and benefits associated with different operational assumptions.
The question must also be raised whether all approaches are equally focused on what we customarily mean by rivalries. Moreover, in the absence of a consensus on basic concepts and measures, rivalry findings will be anything but additive even if the subfield continues to be monopolized by largely divergent dispute density approaches.

The analysis of rivalry in world politics possesses some considerable potential for revolutionizing the study of conflict. Rather than assume all actors are equally likely to engage in conflictual relations, a focus on rivalries permits analysts to focus in turn 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 article, 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 an historical process in which a pair of states create
Author'snote: The strategic rivalry data were collected with support from a National Science Foundation grant.
The present article has benefited from the criticisms of three reviewers, including Paul Diehl who finally has been allowed to review a rivalry paper.
? 2001 International Studies Association.
Published by Blackwell Publishers, 350 Main Street, Malden, MA 02148, USA, and 108 Cowley Road, Oxford OX4
1JF, UK.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

558

Identifying Rivals and Rivalries in WorldPolitics

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. Rivalries thus represent a distinctive class of conflict in the sense that rivals deal with each other in a psychologically charged context of path-dependent hostility in ways that are not necessarily observed in conflicts that occur in more neutral contexts.
We cannot yet say that we know a great deal about how conflict in rivalry operates differently from conflict in nonrivalry contexts (Gartzke and Simon,
1999). We have not really been sensitive to the significance of rivalry relationships for all that long a time. Much remains to be learned. However, before we are likely to make significant headway in reducing our collective ignorance about rivalry relationships, the problem of what rivalries are and how best to measure them must be confronted. It is no doubt expecting too much that we could develop a quick consensus on this matter. At the very least, though, we need to come to terms with the choices being made in undertaking the study of rivalry.
One of the most fundamental issues relates to how we know a rivalry when we see one. The basic tension analyzed here is between an interpretive emphasis on perceptions about threatening competitors who are categorized as enemies (strategic rivalries) and an empirical emphasis on satisfying a minimal number of militarized disputes within some time limit (enduring and interstate rivalries).
Must a relationship become sufficiently militarized before we recognize it as a rivalry? A related question is, what do we do with this recognition to translate it into a systematic data set for empirical analysis?The interpretive approach requires a labor-intensive investigation of historical sources. The empirical approach requires manipulating an existing data set according to various rules. Unless we can come to some early understanding about these questions, the findings of rivalry analysis will simply not be additive in any sense.
While the study of rivalry has been characterized by a large number of relatively casual references to the phenomenon in the historical literature on international relations, there is also a burgeoning empirical literature that, in most cases, has developed a convention of relying on data on militarized interstate disputes (MIDs; see Jones, Bremer, and Singer, 1996) to identify rivalry relationships.l Essentially, analysts require X number of disputes within Y number of years to tell them that a rivalry exists. They then employ this information as a filter for various studies of conflict onset, escalation, and termination. Even though the approach seems quite straightforward, there are in fact a host of problems associated with this practice. One of the problems is whether the dispute-density approach measures rivalry relationships per se or simply greaterthan-average-disputatiousness? Moreover, the last two decades have seen a number of formulae put forward for capturing the right dispute-density. How do we assess rivalry findings if they are predicated on a variety of different operational thresholds? Another problem is whether relying on information on the occurrence of disputes distorts our understanding of when rivalries begin and end?
Does a reliance on dispute activity discriminate against places and times where and when militarized dispute activity is less visible?
There are definite limits on how well we can answer these questions at this time. But they need to be addressed early on rather than later. Fortunately, it is also possible to address them in the context of an alternative way to identify rivalries. Rather than relying on data sets already in existence that were put together for other purposes, it is feasible to cull information from historical
1 In addition, case studies sensitive to rivalry processes are beginning to appear. See, e.g., Lieberman, 1995;
Stein, 1996; Mares, 1996/97; a number of chapters in Diehl, 1998; and Thompson, 1999; Hensel, 2001; Rasler, 2001; and Thompson, 2001.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

559

sources about when and with whom decision-makers thought they were in rivalry relationships. This approach emphasizing perceptions rather than disputes is not without its own problems. It is labor intensive. It requires a great deal of interpretation that renders replication difficult. But the question remains whether it is a more suitable approach to the substantive questions associated with rivalry analyses than the dispute-density approach.
Without knowing which type of approach is more accurate in capturing the
"true" rivalry pool, the best that can be done is to look for the apparent biases exhibited by the alternative approaches. Accordingly, the remainder of this analysis is devoted to a more detailed examination of the problems linked to alternative approaches to rivalry identification. The examination is conducted within the concrete context of four identifications of rivalry: three versions utilizing a dispute-density approach (Diehl and Goertz's [2000] 63 enduring rivalries,Bennett's
[1996, 1997a] 34 interstate rivalries, and Bennett's [1997b, 1998] 63 rivalries) in contrast to a new data set on 174 strategic rivalries predicated on systematizing historical perceptions about competitors, threats, and enemies. The sequence of discussion is to first discuss the definition and operationalization of strategic rivalry. A second section is devoted to comparing the three sets of rivalry identifications in terms of conceptualization, identification agreement, spatial and temporal coverage, and other types of characteristics. This is not a tournament in which one approach will be determined the victor. Each approach starts with a certain conceptualization and then proceeds to measure that conceptualization in distinctive ways. The ultimate question, therefore, cannot be which operational path is right or wrong. Rather, the fundamental question is what price or payoff for the analysis of rivalry is likely to be associated with pursuing one path versus another.

Strategic Rivalries
Strategic rivalries are very much about conflict. Thus, one needs to begin with some elementary assumptions about conflict. Inherently, conflicts are about relative scarcity and overlapping interests and goals. We cannot have as much as we would like of objects with value because there are usually not enough of them to go around. Someone's gain means somebody else's loss. We cannot attain all of our goals because to do so would interfere with somebody else's maximal goal attainment. Hence, conflicts are about real incompatibilities in attaining material and nonmaterial goals. They do not exist unless they are perceived and perceptual pathologies may make conflicts worse than they might otherwise have been. But they still tend to be based on some inability to occupy the same space, share the same position, or accept the superiority of another's belief system.
Disputes about territory, influence and status, and ideology, therefore, are at the core of conflicts of interest at all levels of analysis, but especially between states.
Conflicts of interest vary in intensity. Conflicts can be mild or extreme. Nor is behavior consumed by conflict. Actors also cooperate, and they do so in various amounts. One way to visualize the array of behavior is to imagine a conflictcooperation continuum. At one end are extreme cases of intense conflict; at the other, extreme cases of intense cooperation. In between are various mixes of conflict and cooperation of the relatively milder sorts. The relationships between most pairs of states can be located around the center of this continuum. That is, their relationships are normal and encompass some combination of conflict and cooperation. Some pairs of states have especially cooperative relationships (often called "special relationships"), either because they share certain affinities of culture, race, and language, or because they share important goals, or because one of the states in the dyad has no choice but to be highly cooperative.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

560

Identifying Rivals and Rivalries in World Politics

In the intense conflict zone of the continuum, pairs of states regard each other as significant threats to goal attainment. However, there are essentially two types of dyadic situations at this end of the continuum. Dyads encompass either roughly comparable states or circumstances in which one of the states is much more powerful than the other. When the dyad encompasses states with roughly equal capabilities, the conflicts of interest are likely to persist because it is less likely that one part of the dyad will be able to impose its will on the other actor successfully. When the dyad encompasses states with highly unequal capabilities, the conflicts of interest are less likely to persist, all other things being equal, because the more powerful actor can contemplate coercing the other actor to accept its superior position. If those same factors remain equal, the weaker party is likely to have incentive to yield on the question(s) at hand.
Of course, other things are not always equal. Stronger states do not always win their contests with weaker states. Hence, it cannot be assumed that conflicts of interest will not persist in cases of dyads with unequal capability. They may not be the norm, but it is possible for conflicts to emerge in these circumstances and, given the appropriate conditions, to persist. It is also possible for decisionmakers in weak states to delude themselves temporarily into believing they have more capability to act in international politics than it turns out they really have.
Decision-makers in strong states are also capable of exaggerating the menace posed by weaker neighbors.
Strategic rivalries might be thought of as the reverse image of the cooperative special relationships. All dyads located toward the intense conflict end of the continuum are not strategic rivalries. A very weak state confronted with an intense threat from a very strong state is unlikely to see the very strong state as a rival. Nor is the strong, threatening state likely to see the very weak state as a rival. Capability asymmetry does not preclude rivalry but it does make it less probable. Nor are rivals defined solely by intense conflicts of interest. Rivals must be selected. Three selection criteria appear to be most important. The actors in question must regard each other as (a) competitors, (b) the source of actual or latent threats that pose some possibility of becoming militarized, and
(c) enemies.
Most states are not viewed as competitors-that is, capable of "playing" in the same league. Relatively weak states are usually capable of interacting competitively only with states in their immediate neighborhood, thereby winnowing the playing field dramatically. Stronger actors may move into the neighborhood in threatening ways but without necessarily being perceived, or without perceiving themselves, as genuine competitors. If an opponent is too strong to be opposed unilaterally, assistance may be sought from a rival of the opponent. Other opponents may be regarded more as nuisances or, more neutrally, as policy problems than as full-fledged competitors or rivals.
For instance, Scandinavia was once a theater dominated by the strategic rivalry between Sweden and Denmark. As new and more powerful states, Prussia and
Russia in particular, entered the Baltic subsystem, the central rivalry was gradually supplanted and wound down as the traditional Baltic rivals found themselves outclassed by the new power of their neighbors. At the same time, Sweden and
Denmark ultimately came to a territorial arrangement that brought them less into conflict than had been the case in the past. Thus, several processes worked to de-escalate the Danish-Swedish rivalry without simply transforming the traditional rivalry into new ones.2 Sweden attempted to be a rival to Russia for a time but was forced to concede that it was no longer sufficiently competitive. Denmark and Prussia never really became rivals despite the contentious Schleswig2 On the Danish-Swedish rivalry see Lisk, 1967; Burton, 1986; and Fitzmaurice, 1992.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

561

Holstein dispute and the two nineteenth-century wars over the issue. Denmark was too weak and Prussia was more concerned about its Austrian and French rivals. Relatively strong states are apt to perceive more competition than weak states and engage in wider fields of interaction, but only some limited portion of this wider field is likely to generate strategic threats. Even the strongest states find it highly taxing in resources and energy to cope with several rivals simultaneously.
As a consequence, decision-makers, of both major and minor powers, are apt to downgrade old rivals once new ones begin to emerge. Taking on a new adversary often means putting some of one's old conflicts on the back burner. Both the supply and demand for rivals thus work toward actors being highly selective in whom they choose to threaten and from whom they choose to perceive strategic threats. As already noted, most states are unable to project threats very far in the first place. That fact of life also helps narrow the selection pool immensely.
The outstanding example of rivalry downgrading occurred prior to World
War I. Faced with an emerging German threat, British decision-makers negotiated significant reductions in the level of hostility associated with their main rivals of the nineteenth century: France, Russia, and the United States. Two of these de-escalations proved to be permanent. Only the Anglo-Russian strategic rivalry resumed when decision-makers found it more convenient to act on their conflicts of interest. The other intriguing dimension of this British example is that a case can be made that the source of Britain's greatest threat emanated from the United States, not Germany. It would not have been totally implausible if British decision-makers had decided to ally with Germany and to oppose their mutual, traditional rivals, France and Russia. But they did not; nor was the
United States placed at the top of the external list. That place was reserved for
Germany. British decision-makers selected Germany to be its principal rival, just as the Germans selected Britain as one of their primary rivals (see, e.g., Kennedy,
1976, 1980).
Similarly, Israeli decision-makers have done much the same thing by drawing concentric circles around their state boundaries (see, e.g., Brecher, 1972). Subject to some qualifications, Israel's rivals have been located in the most immediate geographic circle. Those located farther away are, or at least were, once less worrisome. Within the inner circle, further rank ordering took place, with Egypt and Syria regarded as more dangerous than Jordan. Much the same process was at work in southern Africa prior to the end of apartheid. States, such as Tanzania, that were not proximate to South Africa's borders were much less likely to be targets of South African attacks. Angola, Mozambique, Zimbabwe, and Zambia were a different story.
Precisely in that context, the most important criterion for identifying rivalries is their nonanonymity.3 Actors categorize other actors in their environments.
Some are friends, others are enemies. Threatening enemies who are also adjudged to be competitors in some sense, as opposed to irritants or simply problems, are branded as rivals. This categorization is very much a social-psychological process.
Actors interpret the intentions of others based on earlier behavior and forecasts about the future behavior of these other actors. The interpretation of these intentions leads to expectations about the likelihood of conflicts escalating to physical attacks. Strategic rivals anticipate some positive probability of an attack from their competitors over issues in contention. One side's expectations influence their own subsequent behavior toward their adversary and the process
3 This element is especially stressed in Kuenne, 1989. See, as well, works by McGinnis and Williams (1989),
McGinnis (1990), Vasquez (1993), Thompson (1995), Levy and Ali (1998), Levy (1999), and Rapkin (1999) for other definitions of rivalry that could be said to overlap on this issue. The enemy criterion follows the thread suggested some time ago by Finlay, Holsti, and Fagan (1967).

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

562

Identifying Rivals and Rivalries in WorldPolitics

continues from there. Both sides expect hostile behavior from the other side and proceed to deal with the adversary with that expectation in mind. One round of hostility then reinforces the expectation of future hostility (and rivalry) and leads to some likelihood of a further exchange of hostile behavior in cyclical fashion.
Whether or not the level of hostility spirals increasingly upward, the rivalry relationship, with time and repeatedly reinforced expectations, develops a variety of psychological baggage from which it is difficult to break free. The expectations become more rigid, less sensitive to changes in adversary behavior, and less in need of continued reinforcement.
This is not a mystical process in which somehow the rivalry takes over like a runaway train. The cognitive biases constructed to justify and maintain rivalries have their domestic political process counterparts. Rivalries develop their own domestic constituencies and those constituencies lobby for maintaining the rivalry.
Leaders may find that their room for external maneuver is circumscribed severely by the influence of these domestic constituencies. For that matter, leaders openly opposed to maintaining a prominent rivalry are less likely to be selected for major decision-making posts in the first place.
The combination of expectations of threat, cognitive rigidities, and domestic political processes make strategic rivalry a potent factor in world politics. They create and sustain dyadic relationships of structured hostility, with or without a
Once in place, they develop great deal of continuous, external reinforcement. substantial barriers to cooperation and conflict de-escalation.
Some level of conflict and distrust becomes the norm. Dealing with one's rivals entails juggling very real conflicts of interest within a charged context especially prone to various decision-making pathologies (in-group solidarity, out-group hostility, mistrust, misperception, and self-fulfilling prophecies). As a consequence, rivalry relationships should be particularly conducive to at least intermittent and serial conflict escalation. Not all interstate conflicts are embedded in their own history but those of rivals definitely are. Conflict de-escalation should thus also be much less likely within rivalry contexts than outside of them. To fundamentally alter this state of affairs becomes a matter of somehow overcoming expectational inertianever an easy process in the political or any other type of arena. It is not impossible to do so. Yet observers are often caught by surprise, for good reasons, when it is achieved.

Operationalizing
StrategicRivalry
This perceptual perspective on rivalry can be translated into operational terms by examining the appropriate evidence about whom actors themselves describe as their rivals at any given time. Foreign policy-makers not only talk and write of rivals, they also bias their activities by explicitly about their identification concentrating considerable energy on coping with their selected adversaries. Not history literasurprisingly, then, we have an extensive foreign policy/diplomatic ture well stocked with clues as to which, and when, states are strategic rivals.
Culling the information constitutes a labor-intensive task, to be sure, but it is possible to extract such information, systematize it, and generate a schedule of rivalries for all states in the international system as far back in time as one has the resources and inclination to do so.4
We no longer think twice about coding information on the existence and dates of onset and termination of wars, crises, deterrence attempts, alliances, or
4 Data on major-power rivalries going back to 1494 were also collected as part of this National Science
Foundation-funded project but they will be discussed in a separate article. One application is found in Colaresi
(forthcoming). Other uses of the major-power rivalry data but for shorter periods can be found in Rasler and
Thompson, 2000, 2001.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

563

trade.5 Collecting information on strategic rivalries is not really all that different an enterprise. No phenomenon is so clear-cut that counting it does not require some level of interpretation. What is a war? If the definition hinges on battle deaths, how does one assess the number of troops actually killed? If a crisis must pose a severe threat to the existence of a state, how do we tell what decisionmakers engaged in responding to the challenge are really thinking? If we want to know whether a deterrence attempt was successful, how do we go about determining whether an aggressor was really deterred from doing something that had been intended? What should we do with long-standing informal alignments that seem to be more meaningful than some formal alliances? Whose trade estimates should we trust: the importing state, the exporting state, or the vertically integrated, multinational corporation that evades labeling its production somewhere else as "trade"?
The point remains that measurement choices rarely boil down to interpreting the raw information versus allowing the facts to speak for themselves. Some interpretation of the raw information is inevitable. In the case of identifying strategic rivalry relationships, some more interpretation of the raw data is required than is normally the case with wars, alliances, or trade. The reason for this is that one is attempting to codify decision-maker perceptions without ever expecting to have direct access to these perceptions. In looking for proxies, minimal thresholds of violence or verbal threats as in the case of wars or crises have limited utility. These indicators may tell us something about the level of hostility at any given point in time but they are unlikely to tell us how long the rivalry has been in existence. Wars and disputes may come and go but rivalries can persist for generations. Strategic rivalries are not usually formally announced, as in the case of alliances, although official justifications for defense spending can approximate these formalities. Rivalries are sometimes declared to be over and sometimes the declarations can be taken at face value-but only sometimes.
The bottom line is that collecting information on strategic rivalries is not completely different from collecting systematic information on other topics of interest in world politics.6 The phenomenon being measured must be delineated as carefully and accurately as possible. Data collection rules and sources must be made as explicit as possible. But as long as the rivalry definition demands that we focus on decision-maker perceptions and categorizations of other states, the need for more interpretation than usual should be anticipated. The following coding rules were employed to generate data on strategic rivalries for the 18161999 period:
1. Strategic rivals must be independent states, as determined by Gleditsch and
Ward's (1999) inventory of independent states.7
2. Beginning and ending dates are keyed as much as possible to the timing of evidence about the onset of explicit threat, competitor, and enemy perceptions on the part of decision-makers. Historical analyses, for instance, often specify that decision-makers were unconcerned about a competitor prior to some year just as they also provide reasonably specific information about the timing of rapprochements and whether they were meaningful ones or simply tactical maneu5

"Thinking twice" means only that we are not intimidated by the task, not that we can do it well.
Collecting data on rivalries is very much like collecting information on military coups (Thompson, 1973) or ships-of-the-line (Modelski and Thompson, 1988).
7
Basically, the prime value of the Gleditsch-Ward approach is that it incorporates a number of non-European states earlier than do the conventions that have hitherto prevailed. This is important if one finds that a state engaged in an external rivalry but is not considered to exist by prevailing Correlates of War conventions. Those who wish to employ a more restrictive system membership need only remove the rivalry cases that do not match.
6

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

564

Identifying Rivals and Rivalries in WorldPolitics

vers. For instance, one might have thought there was a strong likelihood that some form of Spanish-U.S. rivalry over Cuba preceded the 1898 war. Yet one is hard pressed to find any evidence of much U.S. official concern about Spanish activities as a threat after the American Civil War. While Spanish decision-makers may have felt threatened by the presence and growing strength of the United
States, U.S. decision-makers often can be characterized as simply preferring the
Cuban problem to go away. Alternatively, they also worried that Spanish colonies would be taken over by some other European power (see, for instance, Langley,
1976; Combs, 1986). The two wars between Spain and Morocco prior to World
War I also do not seem to have been preceded by a rivalry (Burke, 1976; Parsons,
1976).
More often the identification problem is one of assessing a variety of different dates advanced as beginning and ending candidates. However, the candidates are not put forward in the relevant sources on the basis of a rivalry definition involving threat, competitor, and enemy criteria (or dispute-densities, for that matter). In actuality, it is often unclear what any given historian or decisionThe mere utterance of the terms maker means by the terms "rival"and "rivalry." by appropriate sources, therefore, does not suffice as sufficient evidence of the existence of a rivalry. The operational question is one of deciding whether all three rivalry criteria have been met. Of the three criteria, perceived threat and enemy categorization are the most straightforward to identify. The competitor status identification can be murkier and tends to hinge on how the threat is perceived. If the threat is too great to be met by the threatened acting alone or in conjunction with other states of similar capability, or if the threat is too insignificant to worry much about, the source of threat is not usually viewed as a competitor. For instance, Denmark's decision-makers probably felt threatened by the Soviet Union during the Cold War but there would not be much that
Denmark could do alone or in alliance with half a dozen other states of similar capability to meet the threat. Denmark was not in the same league as the Soviet
Union and neither Danish nor Soviet decision-makers were likely to think otherwise. Britain, on the other hand, had a long-lasting rivalry with Russia and the
Soviet Union that persisted after the end of World War II despite Britain's diminished capacity after 1945. As long as Britain tried to maintain its greatpower status and as long as the Soviet Union and others treated Britain as a competitor, Britain was able to maintain some semblance of its traditional competitor status until the Suez Crisis in 1956. After 1956, Britain continued to regard the Soviet Union as a threatening enemy but no longer could be viewed as a competitor, as evidenced by Britain's gradual retreat from great powerhood and the winding down of its once-global security strategy.
Another illustration of the way in which these terms require interpretation is offered by the Franco-German rivalry. One might have thought that 1945 would have sufficiently altered Germany's competitor status to end the rivalry but it did not. French decision-makers persisted in treating Germany in terms of its potential to regain competitive status until the French strategy toward Germany underwent a radical shift in the early 1950s. Why that happened is too complicated a story to try to explain quickly (see, among others, Milward, 1984:126-167; Heisbourg, 1998; Sturmer, 1998). Suffice it to say that the French acted as if the
Franco-German rivalry was still alive for nearly a decade after the German defeat in World War II. The initial French strategy, predicated on ensuring that a strong
Germany did not reemerge, evolved reluctantly into constraining the implications of the German reemergence through regional integration. One of the implications of this change in strategy was a reduction in the emphasis on the perceived threat of a nascent Germany. Thus, Germany began to regain its competitive status vis-a-vis France, but with much less of the threatening enemy

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

565

image of the previous eight decades. Accordingly, the termination date for this rivalry is interpreted as 1955, the year of French acceptance of the official emergence of a West German state.
The changes of status experienced by major powers offer good examples of the need for interpretation. But the problem is not restricted to major powers.
Cambodia/Kampuchea had never been competitive with Vietnam and most of the time, especially in the early nineteenth century, served as a buffer between
Vietnam and Thailand. Yet Cambodian decision-makers, including both Lon Nol in the early 1970s and Pol Pot in the mid-1970s, apparently came to believe they could compete with Vietnam, despite a capability ratio of roughly 10:1 in the
Vietnamese favor when the Vietnamese invaded Kampuchea (Porter, 1990; Alagappa, 1993). Objectively, the evidence indicates that Cambodia and Vietnam should not have regarded each other as competitors but Cambodian decisionmakers chose to ignore the objective evidence and act in a contrary fashion. The
Vietnamese obviously did not choose to ignore or excuse this presumption.
Something similar seems to have happened to Paraguay in the second half of the
1860s when it was crushed by Argentina and Brazil (Lynch, 1985; Perry, 1986).
Objective capability ratios do not always govern the way decision-makers behave.
The only recourse is to treat each potential case on a case-by-case basis in an attempt to assess decision-maker perceptions at the time. More often than not, though, and the Cuban-U.S. case is clearly another exception, threat perceptions and competitor/enemy status are closely correlated, and tend to rise and fall in tandem. As a general rule, the competitor criterion restricts rivalries to their own class within the major-minor power distinction. Major (minor) power rivalries are most likely to involve two major (minor) powers. Definitely, there are exceptions to this rule. Major-minor power rivalries emerge when minor powers become something more than nuisances in the eyes of major power decision-makers.
Capability asymmetry may still be quite pronounced but that does not mean that the major power is in a position to, or is inclined toward, the use of its capability advantage. Minor power dyads can also be characterized by high asymmetry in capability and one might think that rivalry in such cases is unlikely. For instance,
India and Nepal, China and Kazakhstan, or Israel and Lebanon suggest unlikely dyadic circumstances for the emergence of rivalry.Yet the India-Pakistan, ChinaTaiwan, China-Vietnam, and Israel-Jordan dyads are also characterized by unequal capabilities that have not prevented the emergence of rivalry perceptions. Ultimately, it depends on the decision-makers and their perceptions of sources of threat and who their enemies are.
3. No minimal duration is stipulated in advance. While one can certainly contend quite plausibly that longer enduring rivalries are likely to possess more psychological baggage than shorter ones, there may be a variety of reasons why some rivalries are nipped in the bud, so to speak. For instance, one state might eliminate its rival in fairly short order. We would not wish to suppress this information by definition. Presumably, assessments of the effects of rivalry duration will proceed more efficaciously if we allow the rivals themselves the opportunity to establish the minimum and maximum duration of hostility.
4. Various constituencies within states may have different views about who their state's main rivals are or should be. Unless they control the government, constituency views are not considered the same as those of the principal decisionmakers. If the principal decision-makers disagree about the identity of rivals, the operational problem then becomes one of assessing where foreign policy-making is most concentrated and/or whether the disagreement effectively paralyzes the rivalry identification dimension of foreign policy-making. More likely in such

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

566

Identifying Rivals and Rivalries in WorldPolitics

cases, the identity of the leading rival fluctuates with the political fortunes of domestic competitors (e.g., Caps and Hats in eighteenth-century Sweden or
Tories and Liberals in nineteenth-century Britain).
5. If two states were not considered rivals prior to the outbreak of war, they do not become rivals during the war unless their rivalry extends beyond the period of war combat. This rule is designed to avoid complications in assessing the linkages between rivalry and intensive forms of conflict. If every two states that opposed one another in a war became rivals by definition, we would be hard pressed to distinguish between genuine pre-war rivals and states that were never rivals yet nevertheless found themselves on opposite sides of a battlefield. We would also find it difficult to trace the linkage between rivalry and warfare.
6. One needs to be especially skeptical about dating rivalry terminations. Some rivalries experience short-lived and highly publicized rapprochements that turn out to be less meaningful than one might have thought from reading the relevant press accounts at the time. In other cases, decision-makers become too distracted by other pressing events such as a civil war or other external adversaries to pay much attention to sustaining an external rivalry. Some rivalries enter long periods of hibernation only to erupt suddenly as if nothing had changed. All of these situations may share the outward appearance of rivalry termination. What needs to be manifested is evidence of some explicit kind of a significant de-escalation in threat perceptions and hostility. In the absence of such information, it is preferable to consider a rivalry as ongoing until demonstrated explicitly otherwise. Nevertheless, one must also be alert to genuine de-escalations of hostility that resume at some future point. In such cases, the interrupted periods of threatening competition by enemies are treated as separate rivalries. For example, Greece and Turkey's first rivalry ended in 1930. The primary motivation for the de-escalation may well have been tactical-to meet mutual threats from third parties-but it is clear that the two long-time rivals suspended their dyadic hostility for a number of years. A second rivalry reemerged in 1955 initially over the status of Cyprus and remains ongoing.
Another example is provided by the relationships among several northwestern
African states. Morocco became independent in 1956 with aspirations toward creating a Greater Morocco-not unlike similar aspirations observed at times in other parts of the world (for example, Bulgaria, Greece, Somalia, Syria, Serbia,
China). A newly independent or less constrained state initiates a foreign policy agenda that entails expanding its territorial boundaries to encompass land controlled or thought to have been controlled in an earlier era. In the Moroccan case, Spain controlled small enclaves within Morocco and considerable territory to the south. The border between Algeria (not independent until 1962) and
Morocco to the southeast was poorly defined. Mauritania (independent in 1960) also lay within the claimed southern scope of Greater Morocco. In all three cases
Morocco threatened to retake territory by force if necessary. Irregular actions against Spanish enclaves began as early as 1956, with the pressure on Spain shifting south toward Ifni and the Western Sahara in the 1960s. Pressures on
Spain to withdraw from its Saharan territory built up in the 1970s, both from
Morocco and other sources, and ultimately led to a Spanish evacuation in 1976.
Morocco had renounced its claim on Mauritania in 1969 and gained occasional
Mauritanian collusion in controlling the former Spanish Sahara. However, SpanishMoroccan conflict over the northern Spanish enclaves (Ceuta, Melilla) continued intermittently, with some possibility of militarized clashes remaining tangible and aggravated by fishing rights disputes off the Atlantic coast of Morocco. Only in the early 1990s did Spanish decision-makers seem to become less apprehensive of a military attack by Morocco.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

567

Algerian and Moroccan troops began clashing over the disputed Tindouf region as early as 1962. While a resolution of the border dispute was eventually reached in the early 1970s, Algerian and Moroccan forces had also clashed over
Algerian support for resistance against Moroccan expansion into the Western
Sahara region. Once Spain withdrew, the main local opposition to Moroccan expansion became the indigenous Polisario movement, bolstered by unofficial
Algerian financial and military support. Moroccan-Algerian military clashes appear to have continued intermittently in the Western Sahara without either side choosing to admit it. Diplomatic relations between Algeria and Morocco have blown hot and cold but there is as yet no indication that Algeria is prepared to concede to Moroccan expansion and a stronger Moroccan position in northwest Africa.
Three strategic rivalries have emerged from these relationships. The AlgerianMoroccan one began with Algerian independence in 1962 and has yet to end.
The Mauritanian-Moroccan rivalry lasted only from 1960 to 1969. The SpanishMoroccan rivalry began in 1956 and appears to have terminated by 1991. It could resume because the enclave-fishing rights problems persist but there is no indication that decision-makers on either side are prepared to press their grievances. As long as that remains the case on both sides, the level of threat perleast as far as one can tell looking in from ception is reduced substantially-at the outside.
7. The most valuable sources for information pertinent to identifying strategic rivalry are political histories of individual state's foreign policy activities.8 Authors are not likely to identify rivalries precisely in ways a coder might desire because the concept of rivalry is not uniform in meaning. Nor do most historians consider it part of their job description to prepare their analyses in ways that political scientists can transform their interpretations into systematic data. Yet for many rivalries, the problem is not an absence of information but too much information and information that is in disagreement.
In the end analysis, the data collector must make a best judgment based on the information available and the explicit definitional criteria that are pertinent.
8. Reliance on students to collect data may be inevitable in large-N circumstances. In cases requiring interpretation and judgment across a smaller number of cases, however, student input should be restricted as much as is feasible. In this particular case, all of the decisions on how to code the strategic rivalry data were made by the author based on a direct reading of all of the sources employed for each case, as well as a number of other sources used to reject potential cases.
Whether other analysts might have reached exactly the same conclusions about the identity of rivalry relationships must await subsequent studies by individuals prepared and equipped to take on the labor-intensive examination of nearly two centuries of conflict throughout the planet, or perhaps to concentrate on specific sections of the planet. It should be assumed that errors of interpretation have been made and, hopefully, they will be revealed in time by the closer scrutiny of other analysts. Just how much error should be anticipated and/or tolerated is not clear. Ultimately, error assessments are both absolute and relative. One question is how much error is associated with the 174 identifications of rivalry. While some termination dates are clearly debatable, publication of these identifications assumes that most of the specifications will survive closer scrutiny.
The most likely source of error lies in omissions of rivalries about which we know very little and that are not well covered by historians orjournalists. Late twentieth8 The list of references utilized exceeds 50 pages. Most rivalries are quite capable of generating a dozen or more pertinent sources. In addition, Keesing's Contemporary Archives was examined for the 1990-1999 period in order to compensate for any paucity of discussion in published sources for the last decade.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

568

Identifying Rivals and Rivalries in WorldPolitics

Central Amercentury Central Africa is one good example. Nineteenth-century ica is another.
The relative error question is how well the 174 identifications fare in comparison to identifications made by other approaches to the rivalry question. This is a question to which we will return in the next section by comparing the 174 identifications with those of two other rivalry lists that appear to constitute the principal alternatives at this time. Yet since the principal alternatives are intended to measure distinctly different phenomena, there are major constraints on how far we can take comparisons of the three data sets' relative accuracy. If no one can claim to know what the full dimensions of the rivalry circle or pool are, it is rather awkward to assess relative accuracies. But, at the same time, it is also extremely awkward to simply leave the question of accuracy entirely open-ended.
At the very least, it should be made clear that there are implicit and explicit costs involved in choosing among the available measurements of rivalry between states.
Four Approaches to Measuring Rivalry
Table 1 lists the 174 strategic rivalries that emerge from an identification process predicated on a rivalry definition that combines competitor status, threat perception, and enemy status and focuses on the extraction of information about decision-maker perceptions from historical analyses. Along with the 174 strategic rivalries, information on the identification of 63 enduring rivalries (Diehl and
Goertz, 2000), 34 interstate rivalries (Bennett, 1996, 1997a), and 63 rivalries
(Bennett, 1997b, 1998) is also provided in Table 1.9 There is certainly more than one way to look at the contrasts suggested by the four columns of alternative
One way is to simply say that each type of rivalry conceptualizaidentification. tion must be looking at something quite different given the extensive disagreements characterizing the comparison of any two columns (about which more will be said below). If that is the case, users should simply adopt the identifications that come most closely to their own conception of rivalry. The problem with this approach is that there is much more agreement among these three approaches in defining rivalry than may be apparent. Where they really part company is in measuring their concepts. Evaluating the relative utility of conflicting approaches to measurement is a different process than comparing conceptual definitions.
Each approach has advantages and disadvantages that need to be made as explicit as possible. The ultimate questions are whether the advantages outweigh the disadvantages and whether such an outcome is equally true of all three approaches.
Diehl and Goertz (2000:19-25) begin their conceptual definition by stating that rivalries consist of two states in competition that possess the expectation of future conflict. This beginning point overlaps well with the notion of threatening enemy competitors associated with strategic rivalries. The expectation of future conflict is an important dimension in rivalries and can be conceptualized in various ways, including the synonymous concept of threat perception. At this point, then, the only real conceptual difference between strategic and enduring rivalries is the absence of the enemy identification criterion found in the strategic definition.
9 Studies employing dispute-density approaches to constructing rivalry variables other than the ones to be examined more closely here have employed or endorsed different mixes of dispute and duration thresholds, as well as different versions of the MIDs data set. See, for instance, Gochman and Maoz, 1984; Diehl, 1985a, 1985b, 1985c,
1994; Diehl and Kingston, 1987; Goertz and Diehl, 1993, 1995, 2000a, 2000b; Geller, 1993, 1998; Huth and Russett,
1993; Huth, 1996; Vasquez, 1996; Maoz and Mor, 1996, 1998; Wayman, 1996, 2000; Gibler, 1997; and Cioffi-Revilla,
1998. A related conceptualization is the idea of "protracted conflicts" found in ICB crisis studies (Brecher, 1984,
1993; Brecher and Wilkenfeld, 1997). Huth, Bennett, and Gelpi (1992) developed an alternative form of the labor-intensive approach to acquiring rivalry data but appear to have abandoned the further analysis of their rivalry data. This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

569

However, a genuine parting of the conceptual ways occurs when Diehl and
Goertz choose to introduce two additional criteria: the severity of competition and time. They restrict the competitions in which they are interested to militarized ones. For them, rivalry equals militarized competition because the recourse to military tools of foreign policy demonstrates the severity of the conflict. They go one step further empirically and require that competitors engage in at least six militarized disputes. Moreover, the disputes must also take place within a minimal interval of twenty years. The rationale is that the frequency of militarized competition establishes the expectation of further conflict and also creates another important dimension of rivalry, a history of past conflict. Brief encounters preclude much in the way of history establishment. Nor is there sufficient time to create expectations of future conflict.
The emphases on the history and future dimensions of rivalry are extremely well taken. Participants in rivalries are prisoners of the past and future. They select adversaries on the basis of past encounters, convert their interpretations of the past encounters into current and future expectations about the behavior of the adversary, and worry as well about how current decisions may benefit or penalize adversaries in the future. The problem lies in the six militarized disputes and twenty-year threshold. The obvious advantage is that such a threshold can be applied to an existing data set on militarized disputes to create a list of enduring rivalries. Some variation can also be created by developing multiple thresholds. Diehl and Goertz (2000) also generate lists of what they call "isolated" and "proto" rivalries which have less dispute-density over time than enduring rivalries.10 This procedure generates 1,166 rivalries and allows analysts to compare increasing levels or at least densities of dispute militarization.
The basic conceptual problem is that the Diehl and Goertz approach assumes that a fairly substantial amount of militarized disputation must occur in order to create rivalry histories and futures. While it may be true that more explicit conflict generates stronger expectations of future conflict and threat perception, the Diehl and Goertz approach rules out a full test of this proposition. We can only compare among different dispute frequencies at a number higher than 1.
We cannot compare how nonmilitarized rivalries might be different from those that become militarized for nonmilitarized rivalries do not even exist by definition. Yet it is less than clear that militarized disputes of any frequency are necessary to the creation of conflict expectations." The theoretical question is whether a sense of rivalry can precede actually coming to blows or the explicit (as opposed to implicit) threat thereof. The answer would seem to lie in the affirmative as long as actors are allowed to anticipate trouble. The Diehl and Goertz approach effectively eliminates this possibility in favor of requiring actors to find themselves embroiled in a sequence of conflict before the recognition of rivalry occurs. It follows from this observation that the Diehl and Goertz rivalry identifications are likely to be slow in specifying beginning points. If one does not equate
10 "Isolated rivalries" have only one or two disputes. "Proto rivalries" fall in between the criteria for isolated and enduring rivalries.
I In contrast, slightly more than half (94 or 54 percent) of all strategic rivalries have yet to experience a war.
All but 25 (14.4 percent) have experienced one or more militarized disputes but most have not had many of them.
About three fourths (72 percent) of the 174 rivalries have engaged in ten or fewer years in which militarized disputes were ongoing between them. In this respect, the strategic rivalry approach endorses Goertz and Diehl's
(1993:155) argument that rivalry analysts should seek to avoid precluding "a priori any class of protracted hostile interaction from consideration as a rivalry."As Goertz and Diehl observe in the same article, an emphasis on high conflict thresholds can cause problems for studying the origins, continuation, and endings of rivalries. In their own words, "[E]nduring rivalries definitions that use dispute data will [have problems detecting] truncation [starting a rivalry too late because an operational threshold is slow in being breached],censoring [not knowing when a rivalry actually ends because operational information is either missing-that is, not yet collected-or a fixed, post-conflict period has not yet been completed], and peaceful interludes [brief interruptions in intense conflict]" (163).

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

570

Identifying Rivals and Rivalries in WorldPolitics
TABLI1. Three Identifications of Rivalries in World Politics

Rivalries

Strategic

Afghanistan-Iran I
Afghanistan-Iran II
Afghanistan-Pakistan
Albania-Greece
Algeria-Morocco
Angola-South Africa
Angola-Zaire
Argentina-Brazil
Argentina-Britain
Argentina-Chile
Argentina-Paraguay
Armenia-Azerbaijan
Austria-France
Austria-Italy
Austria-Ottoman Empire
Austria-Prussia
Austria-Russia II
Austria-Serbia
Bahrain-Qatar
Belgium-Germany
Belize-Guatemala
Bolivia-Chile
Bolivia-Paraguay
Bolivia-Peru
Bosnia-Croatia
Bosnia-Serbia
Brazil-Britain
Brazil-Paraguay
Britain-Burma
Britain-China
Britain-France II
Britain-Germany I
Britain-Germany II
Britain-Iraq
Britain-Italy
Britain-Japan
Britain-Russia
Britain-Ottoman.Empire/Turkey
Britain-United States
Bulgaria-Greece
Bulgaria-Rumania
Bulgaria-Ottoman Empire/Turkey
Bulgaria-Yugoslavia
Burkino Faso-Mali
Burma-Thailand
Burundi-Rwanda
Cambodia-Thailand
Cambodia-S. Vietnam
Cambodia-N. Vietnam
Cameroons-Nigeria
Chad-Libya
Chad-Sudan
Chile-Peru
Chile-United States

1816-1937
19961947-1979
1913-1987
19621975-1988
1975-1997
1817-1985
19651843-1991

Enduring

1949-1989

InterstateI

Interstate II

1949-

19741984-

1962-1984

1862-1870
19911816-1918
1848-1918
1816-1918
1816-1870
1816-1918
1903-1920
1986-

1873-1984

1897-1984

1843-1919

1873-1909
1952-1984

1926-1930

1938-1954

1914-1940
1981-1993
18361887-1938
1825-1932
19921992-

1857-1904
1886-1938

1849-1965

1838-1863
1862-1870
1816-1826
1839-1900
1816-1904
1896-1918
1934-1945

1887-1921

1899-1955

1816-1904
1878-1953
1878-1945
1878-1950
1878-1954
1960-1986
1816-1826
1962-1966

1876-1923
1939-1985
1895-1934
1837-1861
1914-1952

1833-1907

1876-1907

1816-1903

1905-1926
1858-1903
1940-1954

1940-1956

1913-1952

1953-1987
1956-1975
1976-1983
19751966-1994
1964-1969
1832-1929
1884-1891

1919-1955
1984-

19581934-1943
1932-1945
1816-1956

1927-1938

1953-

1975-

1871-1929 continued This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

571

WILLIAM R. THOMPSON
TABLE 1. Continued

Rivalries

Strategic

China-France
China-Germany
China-India
China-Japan
China-Russia I
China-Russia II
China-S. Korea
China-Taiwan
China-United States
China-Vietnam
Colombia-Ecuador
Colombia-Nicaragua
Colombia-Peru
Colombia-Venezuela
Congo-Brazzaville-Zaire
Costa Rica-Nicaragua I
Costa Rica-Nicaragua II
Costa Rica-Panama
Croatia-Serbia
Cuba-United States
Cyprus-Turkey
Czechoslovakia-Germany
Czechoslovakia-Hungary
Czechoslovakia-Poland
Dominican Rep.-Haiti
Ecuador-Peru
Ecuador-United States
Egypt-Ethiopia
Egypt-Iran I
Egypt-Iran II
Egypt-Iraq
Egypt-Israel
Egypt-Jordan
Egypt-Libya
Egypt-Ottoman Empire
Egypt-Saudi Arabia
Egypt-Sudan
Egypt-Syria
El Salvador-Guatemala
El Salvador-Honduras
Eq. Guinea-Gabon
Eritrea-Ethiopia
Eritrea-Sudan
Ethiopia-Italy
Ethiopia-Somalia
Ethiopia-Sudan
France-Germany II
France-Italy
France-Russia II
France-Turkey
France-United States II
France-Vietnam
W. Germany-E. Germany
Germany-Italy
Germany-Poland

Enduring

1844-1900
1897-1900
19481873-1945
1816-1949
1958-1989

InterstateI

1870-1900
1950-1987
1873-1958
1862-1986

1898-1929
19501874-1951
1857-

1950-1987
19491949-1978
19731831-1919
1979-1992
1824-1935
1831-

1949-1972

InterstateIIr

19711894-1951
18981976-

19491949-1972

1969-1972

1899-1934
1963-1987

1840-1858
1948-1992
1921-1944
199119591933-1939
1919-1939
1919-1939
1845-1893
1830-1998
1868-1882
1955-1971
1979194519481946-1970
1973-1992
1827-1841
1957-1970
19911961-1990
1840-1930
1840-1992
1972-1979
199819931869-1943
1960-1988
19651816-1955

1987-

1959-1990
1965-1988

19791988-

1891-1955
1952-1981

1891-

19111972-

1948-1989

1948-1979

1968-1979

1960-

198019871866-1955

1923-1943
1960-1985
1967-1988
1830-1887
1911-1945

1850-1955

1881-1940
1816-1894
1897-1938

1920-1939

1914-1945

1939-1956

1830-1871
1858-1885
1949-1973
1918-1939
continued

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

Identifying Rivals and Rivalries in World Politics

572

TABLE Continued
1.
Rivalries

Strategic

Germany-Russia II
Germany-United States I
Germany-United States II
Ghana-Ivory Coast
Ghana-Nigeria
Ghana-Togo
Greece-Ottoman Empire/Turkey I
Greece-Turkey II
Greece-Serbia
Guatemala-Honduras
Guatemala-Mexico
Guatemala-Nicaragua
Guinea-Bissau-Senegal
Guyana-Venezuela
Haiti-United States
Honduras-Nicaragua I
Honduras-Nicaragua II
Hungary-Rumania
Hungary-Yugoslavia
India-Pakistan
Indonesia-Malaysia
Indonesia-Netherlands
Iran-Iraq I
Iran-Iraq II
Iran-Israel
Iran-Ottoman Empire/Turkey
Iran-Russia
Iran-Saudi Arabia
Iraq-Israel
Iraq-Kuwait
Iraq-Saudi Arabia I
Iraq-Saudi Arabia II
Iraq-Syria
Israel-Jordan
Israel-Lebanon
Israel-Syria
Italy-Russia
Israel-Saudi Arabia
Italy-Turkey
Italy-Yugoslavia
Japan-Russia
Japan-S. Korea
Japan-United States
Jordan-Saudi Arabia
Jordan-Syria
Kazakhstan-Uzbekistan
Kenya-Somalia
Kenya-Sudan
Kenya-Uganda
N. Korea-S. Korea
N. Korea-United States
Laos-Thailand
Libya-Sudan
Lithuania-Poland
Malawi-Tanzania
Malawi-Zambia

1890-1945
1889-1918
1939-1945
1960-1970
1960-1966
1960-1995
1827-1930
19551879-1954
1840-1930
1840-1882
1840-1907
1989-1993
1966-

Enduring

InterstateI
1908-1970

1866-1925
1958-1989

1829-1923
1958-

1907-1929

1948-1973

1948-

19481936-1943

1948-1986

1948-

1900-1945
1946-1958
194619911963-1981
1989-1994
1986-1995
1948-

1878-1923
1978-

1891-1915
1929-1962

1895-1962
1980-1987
1918-1947
1918-1955
19471962-1966
1951-1962
1932-1939
195819791816-1932
1816-1828
1979194819611932-1957
196819461948-1994

1884-1943
1918-1954
1873-1945

InterstateI

1947-1991

1947-

1967-

1953-

1953-

1973-

1908-1987

1933-

1967-1991
1961-

19911990-

1957-1981
1880-1924
1923-1956
1895-1984
1953-1982

1880-1923
1853-

19811908-1928
1953-1956
19171977-

1971-

1949-1991

1965-1989
19491950-1985
1960-1988

196819851968-

1949-

1989197019751980-

1974-1985
1919-1939
1964-1994
1964-1986
continued

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

573

WILLIAM R. THOMPSON
TABLE 1. Continued

Strategic

Rivalries
Mauritania-Morocco
Mauritania-Senegal
Mexico-United States
Morocco-Spain
Mozambique-Rhodesia
Mozambique-South Africa
Norway-Russia
Oman-S. Yemen
Ottoman Empire/Turkey-Russia
Ottoman Empire/TurkeySerbia/Yugoslavia
Peru-United States
Poland-Russia
Rhodesia-Zambia
Russia-United States
Russia-Yugoslavia
Saudi Arabia-Yemen I
Saudi Arabia-Yemen II
Spain-United States
Sudan-Uganda I
Sudan-Uganda II
Tanzania-Uganda
Thailand-Vietnam I
Thailand-Vietnam II
South Africa-Zambia
South Africa-Zimbabwe
N. Vietnam-S. Vietnam
Yemen-S. Yemen

Enduring

InterstateI

InterstateII

1960-1969
1989-1995
1821-1848
1956-1991
1975-1979
1976-1991

1836-1893
1957-1980

1836-1923

1859-1927
1979-

1956-1987
1972-1982
1816-1920

1876-1921

19781816-1923

1898-1923

1878-1957
1955-1992
1918-1939
1965-1979
1945-1989
1948-1955
1932-1934
19901816-1819
1963-1972
19941971-1979
1816-1884
1954-1988
1965-1991
1980-1992
1954-1975
1967-1990

1992-

1946-1986

1946-

1966-

1962-1984
1850-1875

1850-1898

1873-1898

1961-1989

1980-

Note: Roman numerals indicate that a dyad has engaged in more than one period of rivalry. In some major power cases, the earlier manifestation of the rivalry preceded the 1816 starting point for this data set. Similarly, all rivalries designated as beginning in 1816 actually began before the 1816 starting point.

frequent militarized disputes with rivalry, and this is the critical assumption, the
Diehl and Goertz identifications are also apt to be too quick in specifying termination points.'2 It seems also probable that some of the identifications will not focus on rivalries per se but, instead, identify dyads that merely have a sequence of militarized disputes. Similarly, any rivalries that lack a sequence of militarized disputes would be ignored entirely. Finally, one should expect some bias in a militarized dispute-based identification toward stronger actors that are most capable of foreign policy militarization and, as well, a bias toward areas in which these actors are most active.
This last expectation also suggests that the Diehl and Goertz listing is likely to
"over sample" situations in which strong actors apply coercion to weaker actors repeatedly. There is debate in the rivalry literature over whether capability asymmetry is absolutely necessary to rivalry development and maintenance. Vasquez
(1993), for instance, argues that it is necessary. Others, including Diehl and
Goertz, suggest that it should remain an open empirical question. The position taken here (and employed in the development of the strategic rivalry data set) is that, other things being equal, symmetrical capabilities should be expected to make rivalry more likely and more enduring, but that it is not a necessary requirement. For instance, a weaker member of a rivalry dyad may possess a
12

According to Diehl and Goertz (2000:46), an enduring rivalry ends ten years after the last dispute.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

574

Identifying Rivals and Rivalries in WorldPolitics

roughly equal capability position in a local arena in which the stronger member of the dyad is projecting some portion of its capability over considerable distance. At the same time, rivalries with asymmetrical capabilities are not likely to be all that common because both sides of such dyads are less likely to accord competitor status to the other side than they are in dyads with symmetrical dyads.
That does not mean it cannot happen, but only that it is not the norm. More specifically, we should expect major (minor) powers to form rivalry relationships with other major (minor) powers and major-minor combinations should be more rare than major-major or minor-minor rivalry dyads.
Bennett (1996, 1997a) defines interstate rivalries as dyadic situations in which states disagree over issues for an extended period of time to the extent that they engage in relatively frequent diplomatic or military challenges. The issues that are contested must be the same or related to preclude capturing situations in which states simply are disputatious. The outbreak of multiple disputes, continuing disagreement, and the threat of the use of force reflect long-term hostility, the seriousness of the policy disagreements, and the likelihood that states will consider each other as sources of primary threat. Bennett's empirical threshold for the "interstate I" rivalry data, in addition to the issue continuity, is five militarized disputes over at least twenty-five years. Rivalries end when the parties cease threatening the use of force and either compromise over the issues in contention or surrender their earlier claims. These terminations are recognized when a formal agreement is signed or claims are renounced publicly.
In Bennett, 1997b and 1998, a second rivalry identification procedure is advanced. Starting with an older Goertz and Diehl (1995) identification of fortyfive rivalries based on an earlier version of the MIDs data set, an "interstate II" rivalry is any dyad that satisfies a six MIDs criterion within a twenty-year interval, as long as there is no more than a fifteen-year gap between disputes.13 In this approach, rivalries begin only after the dispute-density criteria have been fully established; they end when the issue in contention is settled and no more milithe actual ending date tarized disputes occur in the ensuing ten years-although is then backdated to the formal agreement to terminate the rivalry.
Bennett makes a telling observation when he notes that while continuing to resolve issue conflicts, the militarized disputes indicates an unwillingness absence of militarized disputes does not necessarily tell us whether the disagreements have been resolved. For this reason, he requires a formal agreement or renunciation to demarcate a rivalry termination in addition to the dispute termination. But if the absence of militarized disputes cannot be equated with the absence of serious disagreement, then why should we assume that the presence of multiple militarized disputes is necessary for the existence of a rivalry? Yes, multiple disputes suggest the presence of conflict quite explicitly. But, as argued above, conflict, the expectation of conflict, and the perception of serious levels of threat can exist without the prerequisite of five or six militarized disputes.
Bennett's (1996, 1997a) approach, therefore, ends up duplicating Diehl and
Goertz's focus restricted to explicitly militarized competitions.
If we had an earlier established convention that rivalry requires militarization, the assumption would be more plausible. But we have no such convention. Nor do we know that a sense of rivalry demands militarization. It would seem preferable, then, to leave the role of militarization an open question, not unlike the role of capability symmetry. We could then ask what kind of rivalries become
13
Actually, there appear to be two versions of interstate II. In Bennett, 1997b, the starting dates of the rivalry identifications are based on the first dispute that begins the dispute-density qualification sequence. In Bennett,
1998, the starting date of the rivalry identifications are based on the first year after the dispute-density qualifying sequence has been established. In both articles, it should also be noted that Bennett has dropped the "interstate" modifier and simply refers to the identifications as rivalries.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

575

militarized as part of inquiries into conflict escalation dynamics. The operational approach taken by Bennett and Diehl and Goertz precludes this question by delimiting rivalries to situations that have already escalated considerably.
If they said they were interested for whatever reasons in dyadic situations involving serial militarized dispute behavior, that would be one thing. It becomes a different matter when the term "rivalry"is equated with, and restricted to serial militarized dispute behavior. The rich potential of rivalry analysis does not deserve to be handicapped in this fashion. Alternatively, the rich potential of rivalry analysis is unlikely to be fully realized if we choose to restrict our analytical attention to some small proportion of the rivalry pool from the very outset. The analytical problem will only be complicated further if some of the dyads so identified satisfy serial dispute-density criteria without also delineating accurately the rivalries in the pool. At the very least, we risk losing possibly important observations on the pre-militarization phase of rivalries. The risk is minimal if all rivalries begin with a militarized bang. It is much greater if only some do so.
Thus, in general, we should expect Bennett's rivalry identifications to possess many of the same disadvantages as Diehl and Goertz's list. Beginning and end points may not possess much face validity if they are geared to the occurrence of militarized dispute behavior. If they must complete six disputes in twenty years before they even begin, their life cycle will look vastly different than if the first dispute had been used as a starting point, or if one begins in some premilitarized phase. Bennett's modification of end point requirements, insisting on a formal treaty or renunciation of claims, may be a step in the right direction but it is not enough to delineate when participant perceptions of rivalry actually end.
That also is another empirical question in rivalry analysis that we have yet to answer. Thus, some "non-rivalries" will meet the empirical criteria and some genuine rivalries will be overlooked, and/or ended too early. Given the emphasis on militarized disputes, the bias toward higher capability actors should also be manifested in the Bennett rivalry lists.
Three factors interfere with a full comparison of the four data sets. One is that it is not possible to discuss each and every case in dispute. There are too many cases and too little space to address the disagreements.14 Given the conceptual disagreements, there is also no real way to resolve identification disagreements.
A third and lesser problem is that the first interstate rivalry list covers the
1816-1988 period, the second one encompasses 1816-1992, as do enduring rivalries, and the strategic rivalry list encompasses 1816-1999. Yet these 18161992 dispute-density lists must end by 1982 to count as having terminated. A number of rivalries have terminated toward the end of the twentieth century but we cannot always be sure how the interstate and enduring rivalry identification systems might have treated them. Nevertheless, there are a number of observations that can be made about agreement, disagreement, and various biases in the four lists.
Not surprisingly, the level of agreement is low across all four data sets. Since the strategic rivalry list has so many more rivalries than the other two lists, a low general level of agreement is inevitable. Less inevitable is the substantial level of disagreement found to characterize the three lists based on dispute-density measures. Forty-five enduring rivalries (72.5 percent of 62) are strategic rivalries while all but one of the first set of interstate rivalries I is a strategic rivalry.'5 Only
27 (43.5 percent of 62) enduring rivalries are interstate rivalries in the first iteration. Put another way, the enduring and interstate rivalry I lists agree on
14 Bennett (1997a:392) reports some fairly slight differences in outcome using enduring and interstate rivalry

data.
15

The Cambodia-Thailand dyad is the exception.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

576

Identifying Rivals and Rivalries in WorldPolitics

twenty-seven cases and disagree on forty-two. The two interstate lists (I and II) agree on twenty-eight cases and disagree on thirty-eight. The best agreement is manifested by the enduring and interstate II lists (agreeing on 57 and disagreeing on 10), but then the interstate II list was based on an earlier version of the enduring list.16 None of the lists shows much agreement about specific dates.
For instance, the enduring and interstate I lists agree only on three cases and are a year apart on a fourth case. The interstate rivalry II periodization is well designed to minimize dating overlaps.
Perhaps the level of disagreement should not be surprising given the various conceptual emphases. However, one of the asserted advantages of the disputedensity approach is its presumed objectivity. Somewhat more agreement than was found, one might think, should characterize three lists with overlapping operational emphases. The problem is compounded by the fact that the enduring rivalry list arrayed in Table 1 is the most recent version. An earlier version that was used in a number of published articles featured forty-five enduring rivalries. In moving from the earlier list to the most recent one, six rivalries were dropped and twenty-three added. Presumably, these rather extensive modifications were due to revisions of the MIDs data set, a dispute inventory that has expanded its N size several times since it was first introduced in the early 1980s.17
Further revision of the MIDs data set is probable so it is quite possible that we lists based to whatever may see further changes in the rivalry identification extent on dispute-density indicators.
If one adds the many earlier studies using different dispute thresholds for rivalry variables, three preliminary implications are clear. One, it is difficult to argue that a reliance on dispute-density avoids interpretation. There is after all some ambiguity about the appropriate density cutoff points that can never be removed because the number of disputes and number of years required for a full-fledged rivalry are fairly arbitrary. That is one reason so many density variations have been put forward. Hence, the interpretive element in disputedensity approaches is focused on thresholds as opposed to more direct evidence for rivalries. While it may be more convenient to both access and argue about the indirect evidence, it is not yet clear that any consensus has emerged concerning precisely what dispute-density is a necessary criterion for identifying a rivalry.
Even if a consensus had emerged early on, there still would have been multiple dispute-density lists thanks to the revisions in the MIDs data set. Either way, the outcome is that we have to be very careful in interpreting the analyses done on, or involving, rivalry data in the past two decades. It is not always clear what differences the various rivalry identifications might have made in the findings that have been produced. Given the low level of agreement in the most recent ones, which would only be compounded by citing the earlier identifications, we must assume that some of the findings would not have emerged if different rivalry identifications had been introduced. That is another empirical question that remains to be resolved. So while a dispute-density approach may constitute a more objective and replicable practice, the employment of such approaches has not had a salutary effect on the rivalry subfield so far.
One cannot assume that the findings of any two empirical rivalry analyses are complementary unless they were done by the same author(s) and actually emThese two conditions have yet to be ployed the same rivalry identifications. satisfied jointly very often.
16
However, Bennett (1997b) does express some misgivings about whether some of the rivalries his approach identifies should be viewed as rivalries.
17
Over the years, MIDs analyses have been based on inventories of disputes ranging from 800 to around 2,000 cases. This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

577

listed in Table 1, other
If we return to a close focus on the identifications observations can be advanced. The enduring rivalry data set identifies no rivalry before 1830 and lists only four as active after 1992. No new rivalry emerges after
1967. But, as captured in Table 2, the enduring list does respond to the increase in new states after World War II. The interstate I set starts with two rivalries and remains relatively flat or constant in number after World War II and throughout the Cold War era. No new rivalry emerges after 1968, but in fact most of the
"latest" rivalries in the set entered in the 1940s. As a consequence the interstate
I set registers the most modest post-1945 increase of the three sets of identifications while demonstrating an aggregated number of rivalries quite similar to the enduring list prior to 1945. The interstate II set converges on the number recorded by the other two dispute-density series around the turn of the century and then initially declines as the international system expands after 1945 before ramping upward from the 1960s on. It is quite clear that the three disputedensity series disagree about whether rivalry propensities are increasing, decreasing, or remaining about the same.
In contrast, the strategic list begins in 1816 with eighteen rivalries carried over from the pre-Waterloo era, rises gradually through the first three quarters of the unlike the other two series, before falling off more nineteenth century-not than the other two due to the effects of World War II. As many as precipitously twenty-one rivalries are listed as terminated between 1939 and 1945. The number of ongoing rivalries then almost triples in the post-World War era before declining in the second half of the 1980s and 1990s. However, the strategic list suggests that almost as many rivalries have persisted into the twenty-first century as the enduring list ever recorded in operation at one time. The number of strategic rivalries thought to be operating in 1999 is about three times as many in number as the interstate I list has ongoing in 1988 and about ten times the number of enduring rivalries listed as still functioning in 1992. The number of interstate II rivalries is converging on the number of strategic rivalries toward the end of the twentieth century, but, in part, only because the two series are characterized by opposing trends in that time period. Thus, in general, there are some discernible similarities in profile across all four series, but each one has some distinctive characteristics as to when and how much the aggregate number fluctuates. One of the more striking features of the enduring list is that we must presume that the following rivalries have ended: Algeria-Morocco
(1984), China-India
(1987), Cuba-United States (1990), Ecuador-Peru (1955), Greece-Turkey (1989),
India-Pakistan (1991), Iraq-Israel (1991), and Israel-Syria (1986). Other rivalries have terminated in this list but the dating of the eight terminations in particular involved in them. The would come as some surprise to the decision-makers
Ecuador-Peru rivalry appears to have terminated in 1998 but the others seem to be like Mark Twain alive and well at this writing despite rumors to the contrary.
Ironically, one of the rivalries declared ended by the enduring list has increased its probability of producing a nuclear war primarily since the rivalry was said to be over. The acute dangers associated with the India-Pakistan rivalry offer a dramatic lesson in the problems linked to over-relying on data on overt, militarized dispute activity--although this particular rivalry has continued to exhibit militarized disputes as well.
Of course, one can attribute some unknown portion of this problem to a censoring problem. The MIDs data set is currently being updated but at the time of this writing is available only through 1992. In the absence of complete data, one cannot know when or whether some rivalry identifications based on disputedensity measurement principles that were ongoing fairly recently are genuinely terminated. With more MIDs data, some of these rivalries might be seen in a different light. Note, however, that this liability does not appear to encourage

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

Identifying Rivals and Rivalries in WorldPolitics

578

TABLE 2. Four Rivalry Series

Enduring InterstateI InterstateII Strategic
Rivalries Rivalries
Year Rivalries Rivalries
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870

0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
2
2
2
2
2
3
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
6

2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
4
4
4
5
5
5
5
5
5
5
6
6
6
6
6
6
6
8
8
8
9
9
9
10
11
11
11
11
11
11
11
11
11
11
10
10
10
10

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
2
3
3
3
3
3
3
3
3
3
3
3
3

18
19
19
18
18
19
19
19
20
21
19
21
20
20
22
24
25
25
25
25
26
26
26
27
33
32
32
33
34
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
37
37
37
37
37
37
38
39
36

Enduring InterstateI InterstateII Strategic
Rivalries Rivalries
Year Rivalries Rivalries
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925

6
6
8
8
8
9
9
9
9
10
10
10
10
10
10
10
11
10
10
10
11
11
11
10
12
12
12
13
13
13
13
13
13
13
12
13
13
14
14
13
14
14
14
14
17
17
17
17
17
17
17
15
16
15
15

11
11
12
13
13
13
13
13
14
14
14
14
14
14
14
15
15
15
15
15
16
16
17
17
17
17
17
17
18
18
18
18
18
17
16
16
16
17
17
16
16
16
16
16
16
16
16
16
16
15
15
15
15
13
11

3
3
4
4
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
7
8
8
8
8
8
9
12
11
11
11
10
11
11
11
11
11
11
12
12
12
12
12
11
12
14
12
14
14
14
14
14
12

35
35
37
37
37
37
37
42
43
43
43
43
43
43
43
42
43
43
44
45
44
44
43
42
43
44
45
45
45
43
43
43
44
42
42
42
41
41
41
41
41
41
42
42
42
42
42
41
43
41
42
42
42
42
42 continued This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

579

WILLIAM R. THOMPSON
TABLE 2.

Continued

Enduring InterstateI InterstateII Strategic
Rivalries Rivalries
Year Rivalries Rivalries
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962

14
14
14
14
13
13
12
13
13
12
12
12
12
11
12
11
11
11
10
10
9
10
13
17
20
20
22
24
24
25
24
25
28
27
30
31
35

11
11
11
11

13
13
12
12

10
10
10
10
10

11
10
10
11
11
11
11
11

9
9
9
9
8
8
8
8
8
8
8
9
10
13
17
18
18
17
19
19
19
18
17
17
18
19
19
19

12
12
13
13
13
13
13
13
13
13
13
13
13
13
12
13
13
11
9

6
6
6
6
7
7

42
42
42
41
38
38
40
41
42
41
42
41
40
34
33
33
33
29
28
22
26
27
35
37
36
37
37
36
35
34
35
34
35
36
42
42
45

much hesitation in assigning endpoints
(1993:164) once observed:

Enduring InterstateI InterstateII Strategic
Rivalries Rivalries
Year Rivalries Rivalries
1963
1964
1965
1966
1967
1968
1969
1970
1071
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998

35
35
38
40
39
39
39
39
39
39
37
36
37
37
37
37
37
37
36
35
34
34
34
26
22
15
13
8
7
4

19
19
19
19
19
19
19
19
18
18
17
17
17
17
17
17
17
16
16
16
16
16
15
15
15
15

6
6
6
7
8
11

12
13
15
16
16
17
18
19

20
22
24
26
27
27
27
29
29
29
31
32
33
34
35
35

1999

47
50
54
53
54
55
53
50
50
51
52
53
55
57
58
57
56
58
58
57
56
56
54
53
51
48
49
48
45
44
41
38
39
39
39
39
39

to rivalry durations. As Goertz and Diehl

Another, often unstated, basis for judging any definition of enduring rivalries is that it match our intuition about what cases qualify as enduring rivalries and exclude those from historical knowledge that we think deserve to be excluded.

In respect to capturing termination dates accurately, dispute-density identifications, especially those based strictly on an absence of militarized disputes and some post-conflict waiting period, leave something to be desired.
Table 3 compares the four lists in terms of the types of actors involved in each identified rivalry dyad. It is not possible to say with any great authority what the distribution across the three dyadic types should be, although it was hypoth-

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

580

Identifying Rivals and Rivalries in WorldPolitics
TABLE 3. Rivalry Distributions by Types of Dyads

Rivalry Types

Major-Major

Strategic Rivalries
Enduring Rivalries
Interstate I Rivalries
Interstate II Rivalries

20
(11.5%)
9
(14.3%)
8
(23.5%)
6
(9.5%)

Major-Minor

Minor-Minor

18
(10.3%)

127
(74.0%)

21
(33.3%)
7
(20.6%)
22
(34.9%)

33
(52.4%)
19
(55.9%)
35
(55.6%)

esized earlier that the distribution should look something like a dumbbell, with major-majors and minor-minors more prevalent than major-minors. It is possible, though, to look for the types of biases that are exhibited in Table 1. In all three lists minor-minor rivalries are the largest category As predicted, however, the three dispute-density lists have quite a few cases involving major powers, and almost as many as the number of cases involving minor powers only. Since there have been only a handful of major powers and quite a few minor powers, such distributions should be disturbing. Either major power cases are overrepresented or minor power dyads are extraordinarily unlikely to generate rivalries. On the other hand, the problem may simply be that major powers are more likely to engage in militarized disputes than are minor powers.
For instance, if there have been something on the order of 170 minor powers in the past 200 years, that suggests there have been roughly 14,365 minor power dyads in the same time period. The 32 minor power dyads reported in the enduring rivalry list would then suggest that only 1 of every 500 minor power dyads might be expected to generate a rivalry. The 19 minor power dyads in the interstate I list suggest the ratio of 1:3 for every 1,000 minor power dyads. The interstate II list suggests the ratio is 2:4 per thousand. In contrast, the strategic rivalry list would predict the probability of a minor power rivalry at about 9 in every thousand. All three estimates are strikingly low. Minor power rivalries are not very probable by any measure, but there is still a rather wide range between
1.3 and 8.8 per thousand.
There is also disagreement about the frequency of major-minor rivalries. A third of the enduring rivalries constitute major-minors. About a fifth of the interstate I and slightly more than a third of interstate II rivalries combine strong and weak powers, while the same category accounts for only 10 percent of the strategic rivalries. If we have reasons to anticipate that major-minor rivalries are plausible but not all that common, the data set with the fewest such cases, proportionately speaking, should have greater comparative appeal.
Another type of bias to look for concerns the starting dates of rivalries. We are interested in rivalries either as a control variable or as a subject in its own right.
Either way, we need to capture the full life cycle of each rivalry as accurately as possible. If one stipulates that rivalries must begin with some sort of coercive bang, linking the start to militarized dispute activity is one way to proceed even though we have seen that there is not a great deal of agreement over which dispute we should begin with. If, on the other hand, we have no reason to assume that rivalries must begin with a bang or a bang density, then we need to try capturing when decision-makers began thinking and acting as if a rivalry existed. Without consensus on this starting point, it is difficult to say whose rivalry starting dates are right or wrong. But we can assess the potential for temporal distortion associated with each approach. Assuming we are better off

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

581

WILLIAM R. THOMPSON

erring on the liberal side than the conservative side on such an issue, let us separate the rivalry identifications in Table 1 that have more than one possible periodization advanced from those that only have one candidate. Then we need to establish the earliest date advanced as a base line and compute how far off each of the other candidate starting dates are in relation to the earliest one.
Such a test is not perfect but it does provide one more indicator of bias.
Table 4 summarizes the outcome in terms of two numbers. The first number is the number of years a given rivalry identification missed vis-a-vis another identification of the same rivalry that began earlier. But this absolute number should be qualified by the number of times an identification did not provide the earliest starting date. Otherwise, a list with the fewest overlapping identifications might appear to be the least biased in this respect. The second number is thus the absolute deviation from the earliest start date divided by the number of times another identification commenced at an earlier date.
If earlier starting rivalries, other things being equal, are advantageous, the least bias is associated with the strategic rivalry list which usually advances the earliest date, in part because it is not tied to dispute-densities. Only eight times does one of the other lists suggest an earlier start date. On average, 11.6 years are "lost" with this approach to identification. Not surprisingly, the most years lost is found in the interstate II list, at an average of almost thirty-six years per rivalry. The next most biased on the starting date dimension is the enduring list at 25.5 years per rivalry. The interstate I list falls in between the strategic and enduring lists at 18.6 years lost on average per rivalry.
Of course, putting forward the earliest starting date cannot be equated with possessing the most accuracy. But since we cannot know for sure which starting date is most accurate without privileging one approach over the others, it seems a reasonable test. Based on this test, all four lists possess some propensity for error on starting dates but the one with the least likely amount of error (compared to the other three) is the strategic list. The list with the most likely amount of starting date error is the interstate II list. We might conduct the same test with ending dates, giving the benefit of the doubt in this case to the latest date advanced, but there is simply too much ambiguity about which list actually advances the latest ending dates after 1982 to take us very far. Presumably, we would have to ignore all of the cases that are listed as ongoing. Even without doing any specific analysis of this question, however, the shortest rivalry durations have to be associated with the interstate II list, and it has already been noted that the enduring list tends to end a number of rivalries prematurely. The likelihood is that ending date biases mirror starting date biases.
Table 5 examines geographical distributions. The regional categories used in this table are fairly crude. It is possible to be more discriminating and to distinguish, for instance, among the three subregions in Europe (western, northcentral eastern, and southeastern), the three subregions of the Middle East
(Mashriq, Maghrib, Gulf), the four subregions in Sub-Saharan Africa (west, east, central, and southern), or even the continental and maritime distinctions in

TABLE 4. Starting Date Biases
Strategic
Rivalries

Absolute Number of Years "Missed"
Average Number of Years "Missed"

Enduring
Rivalries

Interstate I
Rivalries

Interstate II
Rivalries

93
11.6

697
24.9

334
18.6

2182
35.8

Note: The number of years estimate is based on accepting the earliest beginning rivalry as a baseline in contested cases and calculating the deviation of the other starting dates from the baseline.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

582

Identifying Rivals and Rivalries in WorldPolitics
TABLE 5. The Geographical Distribution of Rivalries

Strategic
Rivalries

Enduring
Rivalries

InterstateI
Rivalries

InterstateII
Rivalries

Number

Regions
American
European
Sub-Saharan Africa
Middle East/North Africa
Asia
Other
Total

%

Number

%

Number

%

Number

%

32
37
31
34
28
12
174

18.4
21.3
17.8
19.5
16.1
6.9

11
12
5
12
14
9
63

17.5
19.0
7.9
19.0
22.2
14.3

9
6
1
4
9
5
34

26.5
17.6
2.9
11.8
26.5
14.7

12
12
4
16
14
5
63

19.0
19.0
6.3
25.4
22.2
7.9

southeast Asia. But the relatively small numbers associated with the two disputedensity lists would result in a large number of empty cells if a more refined regional breakdown was imposed on the data.
The geographical distribution for the 174 strategic rivalries are quite evenly dispersed among the five areas.18 Each broadly defined area has generated twenty-eight to thirty-seven rivalries. The enduring rivalry list has a slight Asian bias but roughly the same numbers in the Asian, American, European, and
Middle Eastern zones. Only sub-Saharan Africa appears to be slighted with much less representation than the other macroregions. Less macroscopically but not demonstrated in Table 4, no or very few enduring rivalries are associated with
Central America (1), the northern rim of South America (0), north/central eastern Europe (0), western and southern Africa (0), maritime southeast Asia (0), or central Eurasia (0). The interstate I rivalry list places more than half of its rivalries in the Americas and Asia. Europe is in third place, with comparatively few rivalries assigned to the Middle East and Africa. Yet the interstate I list is especially weak in the same places that are poorly represented in the enduring rivalry list (Central America, the northern rim of South America, north/central eastern Europe, western and southern Africa, maritime southeast Asia, and central Eurasia). The interstate I list is also quite weakly represented in east Africa
(1), southwest Asia (1), and continental southeast Asia (1). The interstate II list shows more geographical balance than interstate I, but it, like the enduring list, discriminates against sub-Saharan Africa.
Each list, then, has a different geographical slant. Strategic rivalries have been found everywhere. Enduring and interstate II rivalries are particularly thin in sub-Saharan Africa, while the interstate rivalry I list detects little rivalry activity in the Middle East and Africa. Hence, all three of the dispute-density lists are noticeably weak in scattered parts of the globe located within the broader macin these roregions. Presumably, the areas that are discernibly underrepresented lists are the other side of the major power bias also found to be linked to dispute-density approaches. More specifically, what that means is that the disputedensity approaches overlook some important rivalry complexes, such as the many intra-Arab feuds, the southern African ones over apartheid, more obscure ones in East Africa, and new ones in southeastern Europe and central Eurasia.
Conclusion
There are no free lunches in choosing among alternative identifications of rivalries between states. Each list has advantages and disadvantages. The dispute18As much as is possible, the rivalry dyads are located in the areas in which they are primarily concerned. Dyads that cannot be restricted easily to one region are assigned to the "other" category.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

583

density lists reduce the need for subjective interpretation, even if they do not of dispense with it altogether. Their liabilities include the overrepresentation of hostile interrivalries involving major powers and the underrepresentation state activity in various parts of the world. They explicitly exclude cases that do not involve fairly high levels of militarized competition. Their dates of onset and termination, which, after all, have some significance for studies attempting to explain the timing of onsets and terminations, are rendered awkward by reliance on formal indicators that may or may not accurately capture the beginning and ending of the phenomena at hand. Since none of the dispute-density approaches that are very congruent with other dispute-density yield rivalry identifications error-in there must be considerable room for identification identifications, terms of both including the appropriate cases and excluding inappropriate cases.
There have also been a number of different dispute-density thresholds, all with different rivalry identifications, applied in the last two decades which suggests that all findings linked to these approaches must be viewed as highly tentative until some consensus should ultimately emerge.
An alternative approach is now available but it relies on an intensive interpreof rivalry that emphasizes tation of historical evidence and a conceptualization rather than militarized conflict. As such, it avoids artificially cenperceptions, soring and truncating the rivalry data, in terms of specifying onset and termination dates, in terms of excluding less militarized conflicts, and in terms of slighting some parts of the world. But the nature of its construction makes the rivalry identifications clearly less easily replicable. Acquiring systematic information on apparent decision-maker perceptions is not quite the same thing as recording the number of times two states have clashed. A substantial amount of interpretation seems inevitable if one seeks data on past, present, and future expectations in world politics for a large number of states and for a respectable length of time.
Given a very small country and temporal N, one might be able to reduce substantially the amount of historical interpretation involved. Ultimately, one might even be able to extend these intensive case studies throughout the planet.
But we are not there yet. In the interim, we are forced to choose among various types of "quick" and dirty short-cuts to the empirical categorizations that we seek.
Nonetheless, choosing among the alternatives also should reduce, in part, to what we think rivalry relationships are most about. Are they about a process of categorizing some competitors as threatening enemies with variable outcomes in the level of explicit conflict, as the strategic rivalry approach contends? Or, should the concept of rivalry be restricted for all practicable purposes to dyads that engage in a large number of militarized disputes? Most conceptual definitions of rivalry, outside of the dispute-density group, do not insist explicitly on a high level of disputatiousness. However, the nature of dispute-density measurements preclude a focus on anything but highly conflictual dyads-whether they regard one another as rivals or not. In the final analysis, the significance of rivalry analyses for the study of international conflict may simply be too important to leave them hostage to the existence of data collected earlier and for other purposes. At the same time, there is no reason why there must be only one definition of what interstate rivalry is about. Analysts who prefer the high conflict emphasis are likely to be more comfortable with dispute-density approaches.
Analysts who are uncomfortable with equating rivalry with intense conflict should be uncomfortable with dispute-density approaches. As long as we keep in mind what the different conceptualizations and measurement approaches entail and imply, we should be able to maximize the digestion and utilization of what we learn from analyses of "rivalry," even as we continue to disagree about how best to approach its identification. For some questions, it may not make all that much difference what approach is adopted. For others, it is likely to make considerable

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

584

Identifying Rivals and Rivalries in World Politics

difference. One of the things we need to do now is to determine fall into which category.

which questions

References
M.
ALAGAPPA, (1993) "International Security in Southeast Asia: Growing Salience of Regional and
Domestic Dynamics." In Securityof Third WorldCountries,edited by J. Singh and T. Bernauer, pp. 109-149. Aldershot, England: Dartmouth.
D.
BENNETT, S. (1996) Security, Bargaining, and the End of Interstate Rivalry. International Studies
40:157-183.
Quarterly
BENNETT,D. S. (1997a) Democracy, Regime Change, and Rivalry Termination. International Interactions 22:369-397.
D.
BENNETT, S. (1997b) Measuring Rivalry Termination. Journal of Conflict Resolution41:227-254.
D.
BENNETT, S. (1998) Integrating and Testing Models of Rivalry. American
Journal of Political Science
42:1200-1232.
M.
BRECHER, (1972) The ForeignPolicy Systemof Israel: Setting, Images, Process.New Haven, CT: Yale
University Press.
M.
BRECHER, (1984) International Crises, Protracted Conflicts. InternationalInteractions11:237-298.
M.
BRECHER, (1993) Crisesin WorldPolitics: Theoryand Reality.New York: Pergamon Press.
BRECHER, ANDJ.W. WILKENFELD
M.,
(1997) A Studyof Crisis.Ann Arbor: University of Michigan Press. in PrecolonialProtestand Resistance,1860-1912. Chicago:
E.
BURKE, (1976) Preludeto Protectorate Morocco:
University of Chicago Press.
BURTON,H. A. (1986) Scandinavia in the RevolutionaryEra, 1760-1815. Minneapolis: University of
Minnesota Press.
C.
CIOFFI-REVILLA, (1998) "The Political Uncertainty of Interstate Rivalries: A Punctuated Equilibrium Model." In Diehl, ed., 64-97.
M.
COLARESI, P. (N.D.) Shocks to the System: Great Power Rivalry and the Leadership Long Cycle.
Journal of Conflict Resolution.Forthcoming.
COMBS, A. (1986) The History of AmericanForeignPolicy,vol. 1. New York: Alfred A. Knopf.
J.
26:331DIEHL,P. F. (1985a) Arms Races to War: Testing Some Empirical Linkages. SociologyQuarterly
349.
DIEHL,P. F. (1985b) Armaments Without War:An Analysis of Some Underlying Effects. Journal of Peace
Resolution22:249-259.
DIEHL,P. F. (1985c) Contiguity and Military Escalation in Major Power Rivalries, 1816-1980. Journal of Politics 47:1203-1211.
DIEHL,P. F. (1993) Enduring Rivalries: Theoretical Constructs and Empirical Patterns. International
37:147-171.
Studies Quarterly
DIEHL,P. F. (1994) Substitutes or Complements? The Effects of Alliances in Military Spending in
Major Power Rivalries. InternationalInteractions19:59-76.
DIEHL,P. F., ED. (1998) The Dynamics of Enduring Rivalries. Urbana: University of Illinois Press.
DIEHL,P. F., AND G. GOERTZ
(2000) War and Peace in InternationalRivalry. Ann Arbor: University of
Michigan Press.
(1987) Messenger or Message? Military Buildups and an Initiation of
DIEHL,P. F., ANDJ. KINGSTON
Conflict. Journal of Politics 49:801-813.
ANDR. R. FAGAN
(1967) Enemiesin Politics. Chicago: Rand McNally.
FINLAY,
D.J., 0. R. HOLSTI,
FITZMAURICE,(1992) The Baltic: A Regional Future?New York: St. Martin's Press.
J.
E.,
GARTZKE, AND M. SIMON(1999) Hot Hand: A Critical Analysis of Enduring Rivalries. Journal of
Politics 63:777-798.
37:173-194.
StudiesQuarterly
D.
GELLER, S. (1993) Power Differentials and War in Rival Dyads. International
D.
GELLER, S. (1998) "The Stability of the Military Balance of War Among Great Power Rivals." In
Diehl, ed., 165-190.
GIBLER,D. (1997) Control the Issues, Control the Conflict: The Effects of Alliances That Settle
Territorial Issues on Interstate Rivalries. International Interactions22:341-368.
K.
GLEDITSCH, S., ANDM. D. WARD(1999) Interstate System Membership: A Revised List of Independent States Since 1816. InternationalInteractions25:393-413.
GOCHMAN, ANDZ. MAOZ(1984) Militarized Interstate Disputes, 1816-1976: Procedures, Patterns,
C.,
and Insights. Journal of ConflictResolution28:585-616.
G.,
GOERTZ, AND P. F. DIEHL(1993) The Empirical Importance of Enduring Rivalries. International
Interactions18:151-163.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

WILLIAM R. THOMPSON

585

G.,
GOERTZ, ANDP. F. DIEHL(1995) The Initiation and Termination of Enduring Rivalries: The Impact of Political Shocks. American
Journal of Political Science39:30-52.
GOERTZ, ANDP. F. DIEHL(2000a) "(Enduring) Rivalries." In Handbookof War Studies II, edited by
G.,
M. I. Midlarsky, pp. 222-267. Ann Arbor: University of Michigan Press.
GOERTZ, ANDP. F. DIEHL(2000b) "Rivalries:The Conflict Process." In WhatDo WeKnowAbout War?
G.,
edited byJ. A. Vasquez, pp. 197-217. Lanham, MD: Rowman and Littlefield.
F.
HEISBOURG, (1998) "French and German Approaches to Organizing Europe's Future Security and
Defense: A French Perspective." In Europe'sFranco-German
Engine, edited by D. P. Calleo and E.
R. Stael, pp. 47-70. Washington, DC: Brookings Institute Press.
P.
HENSEL, R. (2001) "Evolution in Domestic Politics and the Development of Rivalry: The BoliviaPolitics,edited by W. R. Thompson, pp. 176Paraguay Case." In Evolutionary
Interpretations World of 217. New York: Routledge.
HUTH, P. K. (1996) Enduring Rivalries and Territorial Disputes, 1950-1990. ConflictManagementand
Peace Science15:7-41.
AND C. GELPI(1992) System Uncertainty, Risk Propensity and InterHUTH, P. K., D. S. BENNETT, national Conflict Among the Great Powers. Journal of ConflictResolution36:478-517.
HUTH, P. K., ANDB. M. RUSSETT
(1993) General Deterrence Between Enduring Rivals: Testing Three
Competing Models. AmericanPolitical ScienceReview 87:61-73.
AND
(1996) Militarized Interstate Disputes, 1816-1992: RatioJONES,D., S. A. BREMER, J. D. SINGER nale, Coding Rules and Empirical Patterns. ConflictManagementand Peace Science15:163-213.
KENNEDY, P. M. (1976) The Rise and Fall of British Naval Mastery.New York: Charles Scribner's.
P.
KENNEDY, M. (1980) TheRise of theAnglo-German
Antagonism, 1860-1914. London: George Allen and
Unwin.
R.
KUENNE, (1989) Conflict Management in Mature Rivalry.Journal of ConflictResolution33:554-566.
LANGLEY, L. D. (1976) Strugglefor the American Mediterranean:United States-EuropeanRivalry in the
1776-1904. Athens: University of Georgia Press.
Gulf-Caribbean,
LEVY, S. (1999) "Economic Competition, Domestic Politics, and Systematic Change: The Rise and
J.
Decline of the Anglo-Dutch Rivalry, 1609-1688." In Thompson, ed., 172-200.
LEVY, S., AND S. A. ALI (1998) "From Commercial Competition to Strategic Rivalry to War: The
J.
Evolution of the Anglo-Dutch Rivalry, 1609-1652." In Diehl, ed., 29-63.
E.
LIEBERMAN, (1995) What Makes Deterrence Work? Lessons for the Egyptian-Israeli Enduring
Rivalry. SecurityStudies4:851-910. in LISK, (1967) The Strugglefor Supremacy the Baltic, 1600-1725. New York: Funk and Wagnalls.
J.
LYNCH, (1985) "The River Plate Republics from Independence to the Paraguayan War." In The
J.
CambridgeHistory of Latin America:From Independenceto c. 1870, vol. 3, edited by L. Bethell, pp. 615-676. Cambridge: Cambridge University Press.
MAOZ,Z., ANDB. MOR(1996) Enduring Rivalries: The Early Years. InternationalPolitical ScienceReview
17:141-160.
MAOZ,Z., AND B. MOR (1998) "Learning, Preference Change and the Evolution of Enduring Rivalries." In Diehl, ed., 129-164.
MARES,D. R. (1996/97) Deterrence Bargaining in the Ecuador-Peru Enduring Rivalry: Designing
Strategies Around Military Weakness. SecurityStudies 6:91-123.
34:111-135.
McGINNIS,M. (1990) A Rational Model of Regional Rivalry. InternationalStudies Quarterly
McGINNIS,M., ANDJ. WILLIAMS
(1989) Change and Stability in Superpower Rivalry. AmericanPolitical
ScienceReview 83:1101-1123.
A.
MILWARD, S. (1984) The Reconstructionof WesternEurope, 1945-51. Berkeley and Los Angeles:
University of California Press.
MODELSKI,G., ANDW. R. THOMPSON
(1988) SeaPowerin GlobalPolitics, 1494-1993. London: Macmillan.
F.
PARSONS, V. (1976) The Origins of the MoroccoQuestion, 1880-1900. London: Duckworth.
W.
PERRY, (1986) "Argentina: Regional Power Overstretched." In EmergingPowers:Defenseand Security in the Third World,edited by R. W. Jones and S. A. Hildreth, pp. 343-370. New York: Praeger.
PORTER, G. (1990) "Cambodia's Foreign Policy." In The Political Economyof ForeignPolicy in Southeast
Asia, edited by D. Wurfel and B. Burton, pp. 247-272. New York: St. Martin's Press.
D.
RAPKIN, P. (1999) "The Emergence and Intensification of U.S.-Japan Rivalry in the Early Twentieth
Century." In Thompson, ed., 337-370.
K.
RASLER, (2001) "Political Shocks and the De-escalation of Protracted Conflicts: The IsraeliPalestine Case." In EvolutionaryInterpretations World
Politics,edited by W. R. Thompson, pp. 240of
260. New York: Routledge.
RASLER,K., ANDW. R. THOMPSON
(2000) Explaining Escalation to War: Contiguity, Space and Position in the Major Power Subsystem. International Studies Quarterly
44:503-530.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions

586

Identifying Rivals and Rivalries in WorldPolitics

RASLER, K., ANDW. R. THOMPSON
(2001) Rivalries and the Democratic Peace in the Major Power

Subsystem. Forthcoming in Journal of Peace Research38.
STEIN,J. G. (1996) Deterrence and Learning in an Enduring Rivalry: Egypt and Israel, 1948-73.
SecurityStudies 6:104-152.
STURMER, M. (1998) "Deux Reves dan un Seul Lit: Franco-German Security Cooperation." In Europe's
Franco-GermanEngine, edited by D. P. Calleo and E. R. Stael, pp. 71-83. Washington, DC:
Brookings Institute Press.
W.
THOMPSON, R. (1973) The Grievancesof Military Coup-Makers.
Beverly Hills, CA: Sage.
W.
THOMPSON, R. (1995) Principal Rivalries. Journal of ConflictResolution39:195-223.
W.
THOMPSON, R., ED. (1999) GreatPowerRivalries. Columbia: University of South Carolina Press.
W.
THOMPSON, R. (2001) "Expectancy Theory, Strategic Rivalries, and the Sino-Soviet Rivalry." In
EvolutionaryInterpretations WorldPolitics, edited by W. R. Thompson, pp. 218-239. New York: of Routledge.
VASQUEZ,J. A. (1993) The WarPuzzle. Cambridge: Cambridge University Press.
VASQUEZ, A. (1996) Distinguishing Rivals That Go to War from Those That Do Not: A Quantitative
J.
40:531-558.
Comparative Case Study of the Two Paths to War. International Studies Quarterly
WAYMAN, E W. (1996) "Power Shifts and the Onset of War."In Parity and War,edited byJ. Kugler and
D. Lemke, pp. 145-162. Ann Arbor: University of Michigan Press.
WAYMAN, F. W. (2000) "Rivalries:Recurrent Disputes and Explaining War."In WhatDo WeKnow About
War?edited byJ. A. Vasquez, pp. 219-239. Lanham, MD: Rowman and Littlefield.

This content downloaded from 110.93.234.9 on Fri, 20 Nov 2015 13:22:55 UTC
All use subject to JSTOR Terms and Conditions