Foundations of Rivalry Research
Summary and Keywords
Rivalry is pervasive in many areas of life. Though rivalry is not isolated to international politics, interstate rivalries are particularly important given their conflict propensities and destructiveness. Tremendous progress has been made in determining the causes of rivalry initiation, maintenance, escalation, and termination. What we know empirically about rivalry, however, hinges on how the concept of rivalry is defined and which dyads are identified as rivals for which periods of time. Though the what, who, and when questions of rivalry may seem straightforward and the answers to such questions in some cases obvious, defining and identifying rivals has not been without scholarly controversy.
There are several approaches to conceptualizing and operationalizing rivalry. Each approach has strengths and weaknesses. Dispute density approaches, for example, which identify rivals as states that engage in repeated instances of militarized conflict over time, have higher levels of measurement reliability than validity. The strategic rivalry approach, on the other hand, which identifies rivals as states that view one another as threatening competitors and enemies, has a higher level of measurement validity than reliability. Rather than advocating one approach over another, the intent of this article is to lay bare the strengths and weaknesses of different ways of identifying cases of rivalry.
Existing rivalry research provides a foundation from which to further develop rivalry approaches. Given that the concept of rivalry has only recently been applied outside of the dyadic interstate context, additive research is still needed on rivalry in intrastate, triadic and multistate, and other settings. Due to the existence of several mature dyadic interstate rivalry approaches, on the other hand, developing additional distinct approaches for the dyadic interstate context is less imperative than integrating existing approaches. There are several ways this can potentially be done, such as by combining elements of multiple perspectives in ways that minimize weaknesses, through conceptual mapping, or by developing an ordinal measure of rivalry.
Rivalry is relevant to many facets of life. Intense competitors within almost any realm are often said to be rivals, whether it is businesses vying for market share, political adversaries seeking power, sports teams contending for championships, or even siblings seeking to outdo one another. Whether or not we consciously acknowledge it, rivalry is a part of our daily existence.
Though rivalry is not isolated to international politics, interstate rivalries are particularly important given their conflictual and destructive nature. Rivals, such as India and Pakistan, Israel and Syria, and North Korea and South Korea, account for a disproportionate share of international conflict. At the system level, approximately three-fourths of all wars occur among rivals (Colaresi, Rasler, & Thompson, 2007, pp. 88–89). Such conflicts cause widespread human suffering that includes (but is not limited to) fatalities, debilitating injuries, damage to infrastructure, and the displacement of individuals or even entire communities.
The importance of interstate rivalries spans beyond their conflict propensities. Engagement in rivalry, even in the absence of war, may result in the curtailment of civil liberties, as witnessed during the “red scares” of the Cold War. The funneling of revenue toward national defense to meet rival threats reduces the amount of resources available to support economic, social, environmental, and other needs. Involvement in interstate rivalry can even affect international sporting events as witnessed by the Cold War Olympic Games boycotts of the 1980s.
The study of interstate rivalry began in earnest a few decades ago. Since then, tremendous progress has been made in determining the causes of rivalry initiation, maintenance, escalation, and termination (see Diehl & Goertz, 2012; Goertz & Diehl, 2000). What we know empirically about rivalry, however, hinges on how the concept of rivalry is defined and which dyads are identified as rivals for which periods of time. Proper conceptualization and operationalization is fundamental to building cumulative knowledge about interstate rivalry.
This article focuses on the what, who, and when questions of rivalry. What is rivalry? Which dyads should be identified as rivals? When does rivalry begin and end? Such questions must be addressed before investigating the causes and consequences of rivalry. Though each question may seem straightforward and the answers to such questions, in some cases, obvious, defining and identifying rivals has not been without scholarly controversy.
There are two predominant ways of conceptualizing and operationalizing rivalry—the dispute density (Diehl & Goertz, 2000; Klein, Goertz, & Diehl, 2006) and strategic rivalry approaches (Colaresi et al., 2007; Thompson, 1995, 2001; Thompson & Dreyer, 2012). Other approaches, such as the issue rivalry (Mitchell & Thies, 2011) and crisis density (Hewitt, 2005) approaches alter rivalry criteria but tend to draw from the dispute density and/or strategic rivalry perspective(s). Choosing one approach over another often involves making trade-offs. Dispute density approaches, for example, have higher levels of measurement reliability than validity while the strategic rivalry approach has a higher level of measurement validity than reliability. The intent of this review is not to advocate one approach over another but to lay bare the strengths and weaknesses of different ways of identifying cases of rivalry.
This chapter covers not only dyadic interstate rivalry approaches but also recent attempts to extend the concept into intrastate and non-dyadic (triadic or multistate) contexts. Keeping the strengths and weaknesses of different approaches in mind, there are several paths forward for future research. While additive research is still needed in areas that the rivalry concept has only recently been applied (such as the intrastate and complex contexts), integrative research is needed on dyadic interstate rivalry that ties existing approaches together.
What, Who, and When?
Identifying cases of rivalry first requires specifying the conceptual dimensions of rivalry. Rivalry conceptualizations vary from those that identify only a few bare essential dimensions of the concept (e.g., Mitchell & Thies, 2011) to those that provide relatively detailed accounts of rivalry characteristics (e.g., Vasquez, 2009). Broadly, rivals can be defined as actors who engage in contention over time while more narrowly rivals can be restricted to cases in which there is also physical violence, psychological hostility, enemy identification, parity, etc. (see Dreyer, 2014). Of the many conceptualizations, this review focuses on those that have been used as a basis for identifying cases of rivalry, enabling the systematic large-N study of rival relations.
Properly coding cases of interstate rivalry involves operationalizing rivalry in a way that minimizes (or ideally eliminates) false positives and false negatives. In the rivalry context, false positives are non-rivals coded as rivals while false negatives are rivals coded as non-rivals. If cases are improperly coded, some irrelevant cases will be included in rivalry analyses while some relevant cases will be excluded. While it may be unreasonable to expect any approach to eliminate the coding of false positives and false negatives entirely, it is important to consider systematic biases that may exist in the coding of cases when choosing between approaches in order to conduct empirical analyses.
Identifying cases of rivalry involves not only specifying which dyads qualify as rivals but also pinpointing periods of time for which states were or have been rivals. The dates at which rivalries begin or end are not always clear. Diehl and Goertz (2000, p. 46) compare identifying when a rivalry has ended to determining when one has been cured of cancer. After some time has passed without symptoms one might assume that a rivalry has ended or that cancer has been defeated, but either may reappear after a period of dormancy. For cases in which rivalry begins and ends gradually, it may be particularly difficult to pinpoint initiation and termination dates.
Miscoding a few years at the beginning or end of a rivalry may seem relatively insignificant. Yet testing hypotheses on things such as rivalry initiation and termination without properly identifying when rivalries begin and end could lead researchers to draw erroneous conclusions. As Bennett (1997, p. 228) points out in relation to rivalry termination, for example, if rivalry end dates are incorrect, conclusions drawn about causes of rivalry termination based on covariation between variables may be incorrect. Determining proper rivalry initiation and termination years is essential to accurately testing hypotheses related to rivalry dynamics.
Though the what, who, and when questions are to some extent disentangled in the following discussion for analytic purposes, it should be noted that such questions are closely interrelated. How rivalry is conceptualized forms the basis from which rivalry is operationalized, which in turn provides the criteria used to identify rivals and periods of time for which states are rivals. The rivalry approaches surveyed in this article represent cohesive ways of thinking about and coding cases of rivalry. In some instances, one approach may be chosen over another for empirical testing due to an approach being especially suited for answering a particular research question. In general, however, rivalry approaches should be weighed based on overall comparative assessments of balances of strengths and weaknesses across approaches.
Rivalry approaches have proliferated over time. There are several approaches to conceptualizing and operationalizing dyadic interstate rivalry. Some approaches, furthermore, have variants (e.g., there are multiple dispute density measures). This section assesses major dyadic, complex, and intrastate rivalry approaches, starting with the two leading ways of conceptualizing and operationalizing interstate rivalry—the dispute density and strategic rivalry approaches.
Dispute Density Approaches
In identifying cases of rivalry, dispute density approaches emphasize repeated engagement in militarized conflict over an extended period of time. Diehl and Goertz (2000), in a seminal study of interstate rivalry, conceptualize rivalry along three dimensions—spatial consistency (meaning that rival relations are generally dyadic), time or duration, and militarized competitiveness or conflict. Operationally, enduring rivals are identified as two states that engage in six or more disputes (as coded by the Correlates of War project (Jones, Bremer, & Singer, 1996)) over a period of at least 20 years (disputes must also occur within 10–15 years of each other (the exact cutoff depending on how many disputes have occurred previously) to be considered part of the same rivalry). Their measure is one of several dispute density measures of interstate rivalry. Gochman and Maoz (1984), Bennett (1996, 1997), Maoz and Mor (2002), and others have also used dispute density criteria to identify rivals.
Dispute density measures have a high level of intercoder reliability. Lists of rivals produced by scholars using the same dispute density criteria and Correlates of War dataset should be identical. Once dispute density criteria have been established, such approaches ostensibly eliminate the need for subjective determinations in identifying cases of rivalry. Coding rivals using dispute density criteria is neither time nor labor intensive. Those wishing to identify interstate rivals do not need to have extensive knowledge of world politics. Once dispute density criteria have been set, coding rivals becomes a matter of searching Correlates of War data. Dispute density approaches allow for the easy identification of not only past and present rivals but also the continued identification of rivals into the future as long as Correlates of War data on militarized interstates disputes continues to be updated.
There are, however, several drawbacks to dispute density approaches, most of which stem from the reliance on repeated engagement in militarized conflict to identify cases of rivalry. Conceptually, militarization is not a necessary dimension of rivalry. Rivalry can be defined, in other words, without reference to militarized conflict. Broadly, rivals are actors who engage in competition over an issue or set of issues in which the past significantly affects present relations and there is an expectation of future contention (Dreyer, 2014). This is not to say that rivalry cannot be defined more narrowly to focus attention on the most conflictual dyads in the international system. Doing so, however, involves making a number of trade-offs.
Including militarization in definitions of rivalry makes it such that rivalry cannot be used to explain conflict. Explaining militarized disputes with a definition of rivalry that includes militarized disputes is circular (Thompson & Dreyer, 2012, p. 12). Dispute density measures have many uses. Such measures can be used to test hypotheses related to nuclear proliferation (Jo & Gartzke, 2007; Kroenig, 2009; Singh & Way, 2004), terrorism (Conrad, 2011; Findley, Piazza, & Young, 2012), arms races (Rider, Findley, & Diehl, 2011), etc. In seeking to explain militarized conflict, however, scholars must look elsewhere than the rivalry concept if rivalry is operationalized using dispute density criteria.
Relying on dispute density criteria to identify rivals may result in the coding of false negatives. States are not coded as rivals by dispute density approaches if they do not meet a minimum dispute density threshold even if they engage in long-standing hostile interactions. For example, despite continued contention, Iran and Israel are not considered to be rivals according to dispute density approaches due to a lack of repeated engagement in militarized conflict. In some cases, antagonistic states that seem to be rivals simply may not yet have had enough time to accumulate a sufficient number of militarized conflicts to qualify as being coded as dispute density rivals.
States that engage in long-standing non-militarized competitions at low levels of hostility, such as those centered on economic issues, are similarly excluded by dispute density approaches (in some cases perhaps rightfully so). The United States’ economic rivalry with Japan in the late 20th century (Hensel, 1999, p. 177; Rapkin, 1999, p. 337), for example, is not included on dispute density lists of rivals. Most significantly, the United States’ rivalry with China, which may be the central rivalry in world politics for some time to come (Thompson & Dreyer, 2012, p. 8), will not be included on dispute density lists of interstate rivals if the United States and China do not engage in a sufficient number of militarized disputes.
While dispute density measures may not properly identify some cases of rivalry as such, there may also be false positives. Dispute density approaches identify dyads as rivals if dispute density and other criteria are fulfilled even if the states involved do not view one another as rivals. Dyads such as Canada and the United States, for example, are identified as rivals by some dispute density approaches (e.g., Klein, Goertz, & Diehl, 2006) even though the leaders of such states would not commonly think of one another as such.
Other questionable cases of rivalry include those with high levels of capability asymmetry. Power asymmetry does not preclude states from becoming rivals. Though competition generally requires some level of parity, states with asymmetric capabilities can become rivals due to a number of factors, such as a stronger state’s preoccupation with other rivals, a weaker state’s overestimation of its own capabilities, a weaker state’s continued dissatisfaction with the status quo, etc. (see Dreyer, 2014, pp. 510–511). Nonetheless, some dyads are so unequal and uncompetitive, such as India-Nepal, United States-Ecuador, and United States-Haiti that including such dyads on lists of rivalries may be stretching the concept of rivalry too far (see Thompson, 1995, pp. 197–199).
Relying on dispute density criteria to identify rivals also results in bias toward major powers and the regions that they are most active in (Colaresi et al., 2007, p. 52). Given relatively limited capabilities, minor powers tend not to engage in as much militarized conflict as major powers. Furthermore, minor power disputes may be underrepresented in Correlates of War data as a result of less extensive reporting on minor powers in the sources used to code the data. Due to lesser engagement in and perhaps underidentification of militarized disputes, few dyads from sub-Saharan Africa, Central America, Central Eurasia, and other areas populated predominately by minor powers are included on lists of dispute density rivals (see Thompson, 2001, p. 582).
Identifying rivals with dispute density criteria is generally assumed to be less subjective than identifying rivals through historical analysis. When surveying qualitative historical evidence, different scholars may render different judgments as to which dyads qualify as being rivals. Though dispute density approaches may be less subjective than approaches rooted in historical analysis, there are subjective judgments that must be made (though they may be less obvious) in identifying dispute density rivals.
Correlates of War datasets are among the most comprehensive and rigorously coded datasets available on international conflict. It should nonetheless be noted that coding for militarized disputes (which dispute density rivals are identified by) involves searching qualitative historical sources, collecting and systematizing information, and drawing conclusions. Subjective judgments must at times be made when coding the data, particularly when information is conflicting or incomplete. In some cases, there may be different interpretations of the available historical evidence. For example, applying Correlates of War coding criteria, Reiter, Stam, and Horowitz (2016) found that they disagreed with the coding of at least one key variable in more than 30% of the wars coded by Correlates of War investigators.
Further judgments must be made when operationalizing dispute density measures. Dispute density scholars have used a variety of cut-points to determine which dyads are rivals. For example, according to four different operationalizations, for states to qualify as rivals, conflict must occur for a minimum of 0, 11, or 25 years, there must at least 2, 3, 5, or 7 disputes, and the disputes must occur within an interval of 0, 10, or 15 years (see Diehl & Goertz, 2000, p. 36). Which of these cut-points should be adopted? Diehl and Goertz (2000) look for empirical breakpoints to differentiate between types of rivalries. This is better than arbitrarily setting thresholds though it is relatively atheoretical.
Dispute density approaches code rivalry initiation as occurring upon the first of a series of disputes while rivalry termination is coded as occurring after a substantial number of years have passed (such as 10 or 15) following the end of a series of disputes. Using such criteria does not allow for the precise identification of rivalry initiation and termination dates, as Goertz and Diehl (2000, p. 236) acknowledge. More troubling, there may be systematic biases in the coding of dispute density rivalry initiation and termination years.
By identifying the beginning of interstate rivalry as the first of a series of militarized conflicts, dispute density approaches may miss identifying early stages of rival relationships (Colaresi et al., 2007, p. 52). Though the beginning of some rivalries coincide with the initiation of militarized conflict (for example, India’s rivalry with Pakistan), others develop gradually (such as China’s rivalry with the Soviet Union) (Thompson, 1999, p. 14). Rather than militarized conflict marking the beginning of rivalry, militarized conflict may not occur until after rival relations have later escalated. In linking rivalry onset with dispute initiation, there may be systematic bias in which some rivalries are coded as beginning later than when rival relations first develop.
There may also be bias in the coding of rivalry termination years. On the one hand, scholars using dispute density criteria may be too quick to identify termination years (Colaresi et al., 2007, p. 52), coding rivalries as having ended when they are merely in lulls. As long as the issues that drive a rivalry remain unresolved, the potential for conflict tends to remain. On the other hand, scholars using dispute density criteria may be late in identifying rivalry termination, coding rivalries as having ended long after the issues that drive a rivalry have been resolved and relations have been normalized. As Goertz and Diehl (2000, p. 236) note, if rivalry termination is identified as occurring 10 years after the last militarized dispute, the United States-Soviet Union Cold War rivalry would be erroneously identified as not having ended until the beginning of the 21st century.
Klein and colleagues’ (2006) measure of interstate rivalry (an updated version of the Diehl and Goertz (2000) measure) is likely the most widely used dispute density measure of rivalry. The new rivalry measure improves on the previous measure by dropping the isolated rivalry category (“rivals” with only one or two disputes) and including a “linked conflict” dimension to rivalry operationalization. Dropping the isolated rivalry category creates a clearer distinction between isolated instances of conflict versus rivalry while adding a linked conflict dimension to rivalry operationalization better accounts for the interrelated nature of disputes within rivalries.
In order to determine which disputes were related over time, Klein and colleagues produced historical narratives identifying the issues involved in each dispute. Though this improves the validity of the measure, it reduces reliability. Incorporating qualitative historical analysis increases the measure’s subjectivity and complicates identifying rivals in the future by requiring the development of narratives rather than simply being able to rely on Correlates of War.
In general, dispute density measures have high levels of reliability given the use of Correlates of War data to code cases of rivalry. Due to the inclusion of militarization as a dimension of rivalry, however, such measures arguably have lower levels of validity. Likely due in part to the scientific importance of transparency and replicability, many scholars use dispute density measures of rivalry in large-N empirical studies.
The Strategic Rivalry Approach
An alternative to dispute density approaches is the strategic rivalry approach according to which rivals are threatening competitors and enemies (Colaresi et al., 2007; Thompson, 1995, 2001; Thompson & Dreyer, 2012). Rivals are identified by the strategic rivalry approach through qualitative historical analysis. With no dispute density criteria, states are coded as rivals if they view one another as such regardless of whether or not there is repeated engagement in militarized conflict.
The strategic rivalry approach arguably better captures the essence of rivalry. Rivals, whether in politics, sports, business, or some other setting, do not necessarily engage in physical violence. Rivalry is a state of mind, rooted in perceptions of oneself and some other that may result in behavioral manifestations such as engagement in conflict. Identifying rivals according to whether or not conflict occurs may be linking rivalry identification to the presence of a symptom of rivalry rather than rooting rivalry identification in the essence of the concept itself.
By delinking rivalry identification from engagement in militarized disputes, the strategic rivalry approach allows for using rivalry to explain conflict. Caution in doing so is nonetheless warranted. There is likely significant endogeneity between rival perceptions and conflict. While rival perceptions may result in militarized conflict, militarized conflict may solidify rival perceptions. Explaining conflict with strategic rivalry is nonetheless less circular given that rivalry is not defined by repeated engagement in militarized disputes.
Delinking rivalry identification from involvement in militarized conflict may also result in the coding of fewer false negatives and false positives. States are identified as rivals if they view one another as threatening competitors and enemies even if they do not engage in a set minimum number of militarized disputes (e.g., Iran and Israel). States that engage in repeated conflict are not identified as rivals if they do not view one another as such (e.g., Canada and the United States). Asymmetric states are only considered rivals if a weaker state can legitimately compete with and threaten a stronger adversary (Rasler, Thompson, & Ganguly, 2013, p. 4).
Despite such advantages, there are limitations to the strategic rivalry approach, most of which stem from a reliance on qualitative historical analysis to identify rivals. Coding rivals by surveying the historical record is more subjective than searching Correlates of War data to identify cases of rivalry. Strategic rivals cannot be identified simply by coding instances in which states are referred to as “rivals” (Thompson, 2001, p. 564). Historians generally use the term “rivalry” colloquially. Identifying strategic rivals involves rigorously applying strategic rivalry coding rules to historical information.
Though there are clear guidelines for identifying strategic rivals (see Thompson, 2001, pp. 563–568), the historical record is at times incomplete, unclear, or even contradictory. In some instances, there may be a dearth of information. Coverage of interstate relations in Central America during the 19th century, for example, is limited (Thompson, 2001, p. 567). In other instances, there may be adequate historical coverage but conflicting information or accounts. By whom, when, and how a rivalry began is often a matter of controversy.
Identifying strategic rivals requires differentiating between competitors and rivals. Competitors and rivals both contend over issues. Yet while all rivals are competitors, not all competitors are strategic rivals. Distinguishing between competitors and strategic rivals entails assessing whether competitors sufficiently threaten one another to qualify as rivals (Rasler et al., 2013, pp. 3–4). Assessing threat perceptions, however, is not always easy.
Perceptions cannot be directly observed but must be inferred based on recorded statements and behaviors (see Thompson, 2001, p. 563). Accurately assessing leaders’ perceptions can be difficult given that leaders at times have incentives to conceal information (in hopes of gaining advantages or not exposing weaknesses), exaggerate threats (to rally the public or justify certain policies such as high levels of defense spending), or even bluff (hoping an adversary will back down). False negatives could be coded as a result of the concealment of information while false positives could be coded as a result of threat exaggeration and bluffing.
It may be difficult at times to pinpoint when strategic rivalries begin and end with accuracy. Strategic rivalry initiation and termination years are coded to coincide as closely as possible with leaders’ perceptions of others as competitors, threats, and enemies (Thompson, 2001, pp. 563–565). In some cases, such as when states are “born feuding” or when there is total defeat in war (see Wayman, 2000), it may be obvious when rivalries begin or end. In other cases, however, such as when rivalries ebb and flow without pronounced cut-points, it may be more difficult to determine rivalry initiation and termination dates. In 13 case studies of strategic rivals, scholars found rivalry initiation and termination years to be unclear in all but two cases (Thompson, 1999, pp. 14–15).
Disagreements or differences in emphasis among political leaders over which states constitute threats can further complicate identifying rivalry cut-points. States viewed as rivals may fluctuate depending on which political leaders or parties are in power (Thompson, 2001, pp. 565–566). Leaders may not always view their predecessors’ rivals as their own. Coding rivalries as beginning and ending with changes of government could lead to the identification of series of brief rivalries, which seems contrary to the notion that rivalries tend to be enduring. Yet coding rivalries as persisting across different administrations with different foreign policy priorities could lead to states being identified as rivals during periods in which leaders do not view one another as rivals.
Coding for strategic rivals is time and labor intensive. It requires the analysis of hundreds of years of historical data on interstate relations around the globe. Given that it is necessary to not only collect but also interpret historical information, Thompson (2001, p. 567) warns against using students to code strategic rivals. Scholars wishing to independently identify strategic rivals are consequently left with the monumental task of analyzing centuries of dyadic relations.
The independent identification of strategic rivals would serve as a reliability check and allow for refinement of the strategic rivalry measure. Yet the professional dividends of completing such a task at this point are seemingly low, given that doing so would likely be viewed by the academic community as replicating an existing study rather than producing novel research. As a consequence, it is more likely that scholars will develop yet another rivalry approach, leading to the further fragmentation of rivalry research.
The strategic rivalry approach has a lower level of measurement reliability than dispute density measures given the use of qualitative analysis to code cases of rivalry. Yet by excluding dispute density criteria, the strategic rivalry approach is arguably a more valid measure. In choosing between dispute density and strategic rivalry approaches, scholars must consider whether to prioritize measurement validity or reliability.
Rivalry approaches have proliferated beyond dispute density and strategic rivalry measures. Scholars have developed other dyadic interstate rivalry approaches (Bennett, 1997; Hewitt, 2005; Mitchell & Thies, 2011) and have also begun to explore rivalry in intrastate (DeRouen & Bercovitch, 2008) and non-dyadic contexts (Valeriano & Powers, 2016). Such approaches tend to draw from dispute density approaches and/or the strategic rivalry approach and have similar, though also some unique, strengths and weaknesses.
Bennett’s (1997) interstate rivalry approach incorporates issue settlement into measuring rivalry termination. Relying on a list of rivals identified with dispute density criteria (Goertz & Diehl, 1995), Bennett adjusts rivalry termination years so that they coincide with issue resolution. Given that leaders’ public statements are not always sincere and treaties are not always honored, rivalries are only coded as having ended upon formal agreements if 10 years subsequently pass without militarized conflict.
The issue settlement approach improves on dispute density measures by incorporating issues into rivalry termination operationalization, which is more congruent with how rivalry termination should be conceptualized and allows for the more precise identification of rivalry termination years. Rivalries are driven by goal incompatibilities. Once the issues that drive a rivalry are resolved, relations tend to improve. Coding rivalries as ending upon issue settlement is less arbitrary than coding rivalry termination as occurring after a set number of years following a series of disputes. In determining when the Soviet Union-United States rivalry ended, for example, the issue settlement approach correctly identifies the rivalry as terminating at the end of the Cold War rather than around the turn of the century.
Given that the issue settlement approach relies on dispute density criteria for identifying rivals, the approach suffers from many of the same limitations of other dispute density approaches. States may be rivals in the absence of militarized conflict. Dyads in which political leaders do not view one another as threats or enemies are coded as rivals. Even in regard to coding rivalry termination, the issue settlement approach does not entirely divorce itself from dispute density criteria by requiring 10 years to pass without militarized conflict following issue resolution for rivalries to be coded as having ended.
Rivalry termination likely often coincides with issue settlement. In some cases, however, rivalries may end sooner or later than when the issues that drive a rivalry are formally resolved. Despite lack of issue settlement, if relations nonetheless improve over time, lower threat perceptions may result in states no longer viewing one another as rivals. For example, although China and Vietnam still have conflicting claims over parts of the South China Sea, rivalry arguably ended in 1991 following the normalization of relations (Womack, 2011). Contentious issues may at times be tacitly settled or put aside as new threats emerge. Britain, for example, downgraded rivalries with France, Russia, and the United States to confront the growing German threat prior to World War I (Rasler et al., 2013, pp. 210, 245). Rivalries may persist despite issue settlement if rival perceptions linger once the issues that drive a rivalry are resolved. In the event of a peace settlement between the Israelis and Palestinians, for example, it is unlikely that psychological hostility and feelings of animosity would immediately evaporate given ingrained enemy images.
The issue rivalry approach (Mitchell & Thies, 2011), which defines rivalry along two conceptual dimensions—issues and militarization, further incorporates issues into rivalry operationalization. Issue rivals are operationalized (using Issue Correlates of War data (Hensel, 2001; Hensel, Mitchell, Sowers, & Thyne, 2008)) as pairs of states that compete over multiple issues simultaneously while militarized rivals are operationalized as pairs of states that engage in multiple militarized disputes (drawing from dispute density approaches) over a single issue. Pairs of states that compete over multiple issues and engage in multiple militarized disputes are coded as being both issue rivals and militarized rivals.
There are several strengths of an issue-based approach to conceptualizing and operationalizing interstate rivalry. A focus on issues is well-warranted given that rivalries are fundamentally rooted in issue competition. Issue rivalry (though not militarized rivalry) can be used to explain conflict given that issue rivals are not defined by engagement in militarized disputes. Coding rivalry initiation as occurring upon the onset of issue competition allows for the identification of early phases of rivalry for cases in which escalation to violence occurs later (or not at all). Identifying rivalry termination as occurring upon the settlement of issue claims (whether by force or negotiation) is consistent with the issue settlement approach according to which issue resolution removes the basis for rivalry.
Defining rivalry broadly (along only one or two dimensions), however, increases the likelihood that there will be false positives (states that are not rivals coded as such). By not including a temporal dimension to rivalry operationalization, dyads are considered rivals even if competition lasts for a year or less (e.g., Honduras-United Kingdom). Excluding threat and enemy perceptions from rivalry operationalization results in states being coded as rivals if they meet issue or militarized rivalry criteria even if they do not view one another as rivals (e.g., Canada and the United States). Dyads with pronounced power asymmetries in which it is unlikely that a weaker power could credibly threaten a stronger adversary are also identified as rivals (e.g., Dominican Republic-United States, Ecuador-United States, and Haiti-United States).
Given the fundamental importance of issue competition to rivalry, explicitly including issues in rivalry formulations is a positive step forward. Available data on contentious issues in world politics, however, is currently limited. Issue Correlates of War investigators have collected data on territorial, river, and maritime issue claims. States also contend over positional, economic, ideological, and other issues. Developing a comprehensive database of contentious issues for all dyads over centuries of interstate relations is a monumental task that may never be completed. The Issue Correlates of War datasets nonetheless enable the testing of hypotheses related to geopolitical rivals.
The crisis density approach, which identifies enduring rivals as states that have engaged in at least three crises (as identified by the International Crisis Behavior project [Brecher & Wilkenfeld, 1997]) over at least 20 years (crisis must also occur within 15–20 years (depending on how many have previously occurred) to be considered part of the same rivalry), draws from both the strategic rivalry and dispute density approaches (Hewitt, 2005). Data for the International Crisis Behavior project is coded, similar to the strategic rivalry approach, by assessing leaders’ perceptions. Crisis density rivals are coded, similar to dispute density approaches, as states that engage in multiple contentious events over time. In building off of the strategic rivalry and dispute density approaches, the crisis density approach bridges the gap between the two camps to some extent.
By focusing on states that engage in repeated crises rather than repeated militarized disputes, the crisis density approach corrects for some of the false negatives and false positives identified by dispute density approaches. Minor power rivalries are less likely to be underidentified by the crisis density approach. While low-level militarized disputes among minor powers may go underreported (and consequently be underrepresented in Correlates of War data), crises tend to be covered regardless of where they occur (and should consequently be appropriately represented in International Crisis Behavior data). States with mildly conflictual relations are less likely to be overidentified as rivals by the crisis density approach. While low-level militarized disputes may at times be relatively insignificant, crises are high stake events in which there is often a significant likelihood of escalation.
Despite such improvements, identifying rivals with crisis density criteria may still result in the coding of false positives and false negatives. Regarding false negatives, states with long-standing contentious relations that do not meet crisis density thresholds are not coded as rivals. Of the “consensus” rivalries identified by six major rivalry approaches (see Colaresi et al., 2007, p. 57), there are several that are not identified as either proto- or enduring (i.e., moderate or entrenched) rivals using crisis density criteria (e.g., Argentina-Chile, China-Russia, and China-United States) and several that are identified as proto- but not enduring rivals (e.g., China-Japan, Japan-Russia, and North Korea-South Korea). Regarding false positives, highly asymmetric dyads in which a weaker state may not be able to compete with a stronger foe are coded as rivals (e.g., Israel-Lebanon, North Korea-United States, and Syria-United States).
In identifying the beginning of rivalry as occurring upon the first of a series of crises and the end of rivalry as occurring at the end of the last of a series of crises (a substantial number of years must pass after a crisis without recurrence for a crisis to be identified as the last in a series), rivalry initiation and termination years may at times be misidentified. Early phases of rival relations may be missed by coding rivalry initiation as occurring only once a disagreement escalates to the point of a crisis. China’s rivalry with Vietnam, for example, is coded as beginning the same year the Sino-Vietnamese War began even though China and Vietnam viewed one another as rivals and issues began to accumulate for several years prior to the outbreak of war (Dreyer, 2010). Termination years may not be precise as a result of not taking into account issue settlement and other criteria. For example, the Cuba-United States and North Korea-South Korea rivalries are coded as having long ended by the crisis density approach despite subsequent contention.
Though interstate rivalry is generally defined in dyadic terms, rivalries may be triadic, multilateral, and/or interconnected in complex ways. Dyadic rivalries may become linked as demonstrated prior to World War I (Thompson, 2003). The China-Soviet Union-United States Cold War triad is perhaps the most well-known case of triadic rivalry (Goldstein & Freeman, 1991). More broadly, the Eastern bloc versus the Western bloc during the Cold War may qualify as a case of multilateral rivalry (Diehl & Goertz, 2000, pp. 19–20).
The complex rivalry approach represents an initial attempt at identifying triadic interstate rivalries (Valeriano & Powers, 2016). Complex rivals are conceptualized as three or more states with relations linked by common issues, alignments, or conflict joining dynamics. Drawing from dispute density approaches, complex enduring rivals are operationalized as triads with at least four militarized disputes over a minimum of 10 years (disputes must also occur within 10–15 years of each other to be considered part of the same rivalry). Drawing from the strategic rivalry approach, complex strategic rivals are operationalized as competitive triads linked by threat and enemy perceptions.
There is an extremely wide range of cases that could be coded as complex rivals. Valeriano and Powers (2016) identify forty-seven rivalry triads, none of which are located in Africa and only three of which are located in Latin America. In contrast, in an examination of rivalry linkages (rivalries connected to other rivalries through security interests or by geography), Diehl and Goertz (2000, pp. 252–254) only found one dyadic enduring rivalry (out of 63) that was not linked to at least one other rivalry at some point in time. The average rivalry, they found, is linked to17 other rivalries. Some rivalries are extremely interconnected, such as the Soviet Union-United States Cold War rivalry, which was linked to 38 other enduring rivalries. There is a high upper bound on the number of cases that could potentially be identified as complex rivals.
Caution is nonetheless warranted in grouping rivalry dyads together. There may be significant variation in the duration, processes, and importance of rivalry dyads within rivalry networks (Diehl & Goertz, 2000, p. 20). Bilateral rivalry initiations and terminations within complex rivalries may not coincide. For example, although triadic rivalry between Colombia, Ecuador, and Peru is coded as having ended in 1935 by the complex rivalry approach, Ecuador and Peru continued to engage in territorial conflict until 1998. Bilateral levels of cooperation or conflict within complex rivalries may vary significantly. For example, the Iraq-United Kingdom-United States triad featured two states aligned (the United Kingdom and United States) against a common enemy (Iraq) while the Turkey-Russia-United Kingdom triad consisted of three rivalry dyads (each state aligned against each of the other two states in the triad). Certain dyads within complex rivalries may be more important than others. India’s rivalry with Pakistan, for example, has been the dominant rivalry in South Asia’s security complex (Buzan, 1983, pp. 113–114; Diehl & Goertz, 2000, p. 20).
Whether shifting the unit of analysis from dyads to something more complex would result in an empirical payoff remains to be seen. As it currently stands, there are limitations to analyzing complex rivalries empirically. International relations data is generally collected at the individual, state, dyadic, or system level. There are not many readily available variables with which to systematically analyze complex rivalries. Dyadic rivalries often differ significantly from one another, which complicates further aggregating data. Most interstate rivalry analyses will likely continue to be conducted at the dyadic level, though network analysis (Hafner-Burton, Kahler, & Montgomery, 2009) may provide a future means through which to analyze complex rivalries.
Beyond interstate relations, the concept of rivalry has been applied to the intrastate context. The enduring internal rivalry approach defines intrastate rivals along two dimensions—time and militarization (DeRouen & Bercovitch, 2008). Enduring internal rivalries are operationalized as protracted conflicts between governments and insurgencies in which there are at least 25 deaths over a minimum of 10 years. Applying the concept of rivalry to the intrastate context is well-warranted given obvious examples of long-standing conflict between states and non-state groups such as Israel versus the Palestinians.
The enduring internal rivalry approach has similar limitations as dispute density interstate rivalry approaches. Intrastate rivalry cannot be used to explain rivalry duration or conflict if rivalry is defined by duration and conflict (Thompson, 2015, p. 10). There may be false negatives and false positives as a result of not incorporating non-dispute density criteria into rivalry operationalization. Even though the enduring rivalry approach does not rely exclusively on conflict data for identifying rivalry initiation and termination, like all approaches, there are difficulties involved in pinpointing enduring internal rivalry initiation and termination years (such as determining whether rebel groups belong to the same insurgency) (see DeRouen & Bercovitch, 2008, pp. 59–60).
The rivalry concept could potentially be applied to relations between states and extra-state non-state actors as well (e.g., the United States and the al Qaeda or Islamic State terrorist groups). Unlike interstate rivalries, rivalries involving state versus non-state actors tend to be asymmetric given that states can raise revenue through taxation while non-state actors typically cannot. In applying the rivalry concept outside of the interstate context, care must be taken in adapting the concept into what may be much different settings.
Given that the concept of rivalry has only recently been applied outside of the dyadic interstate context, additive research is still needed on rivalry in intrastate, triadic and multistate, and other settings. Due to the existence of several mature dyadic interstate rivalry approaches, on the other hand, developing additional distinct approaches for the dyadic interstate context is less imperative than integrating existing approaches. There are several ways this potentially can be done, such as by combining elements of multiple perspectives in ways that minimize weaknesses, through conceptual mapping, or by developing an ordinal measure of rivalry.
Scholars interested in intrastate, complex, and other forms of rivalry should continue to borrow from dyadic interstate rivalry research. Experts on intrastate conflict, for example, could work toward developing a strategic intrastate rivalry approach. Scholars interested in triadic and multistate rivalries could develop an issue-based complex rivalry approach in which states engaged in competition over the same issue or set of issues (as coded by the Issue Correlates of War project, for example) for an extended period of time are identified as rivals. The existing literature on dyadic interstate rivalry provides fertile ground from which to further develop rivalry approaches in other contexts.
In the dyadic context, rivalry approaches could be integrated by combining elements of different perspectives in ways that maximize the strengths and minimize the weaknesses of existing approaches. This would admittedly not be easy, particularly given that efforts to increase measurement reliability may reduce validity while efforts to increase measurement validity may reduce reliability. It is nonetheless possible to imagine a way forward. For example, rivals could be defined as states that engage in issue conflict (drawing from the issue rivalry approach) over time (drawing from dispute density approaches) and view one another as threats and enemies (drawing from the strategic rivalry approach). Relying on a rigorously coded and publicly available dataset such as the Issue Correlates of War to identify a potential pool of rivals would reduce concerns over measurement reliability, while adding perceptual and other criteria would reduce concerns over measurement validity.
Rivalry approaches can also be integrated through conceptual mapping. For example, rivalry can be conceptualized along a min-max continuum in which the minimal end represents rivalry in its most basic form (in which many cases will qualify as being rivals) while the maximal end represents rivalry in its most extreme form (in which fewer cases will qualify as being rivals) (Dreyer, 2014). Existing approaches can be arranged along a min-max continuum and integrated approaches can be developed starting at the minimal end of the continuum and adding conceptual dimensions.
Interstate rivalry is generally treated as a dichotomous concept (i.e., states are either rivals or they are not). This is useful for certain purposes, such as for using rivalry as a case selection device and for testing hypotheses related to rivalry initiation and termination. Yet treating rivalry as a dichotomous concept is a simplification. Rather than states uniformly being rivals or non-rivals, there are degrees of rivalry.
An ordinal measure of rivalry would better capture rivalry nuances and allow for the investigation of rivalry dynamics over time as rivalries wax and wane. There is significant variation in rivalry interactions over time (see Colaresi, 2005). Some periods of rivalry are “hot” while others are “cold.” Though rivalries tend to be conflictual, rival relations at times significantly deescalate. Rivals even at times cooperate with one another. An ordinal measure of rivalry would facilitate investigating rivalry dynamics beyond initiation and termination.
Although rivalry initiation and termination are generally thought of as events, it may be more appropriate to think of initiation and termination as processes. As discussed in this article, accurately identifying rivalry initiation and termination dates is fraught with difficulties. This is partly because rivalries do not always begin or end abruptly. Even when rivalries seemingly do begin and/or end with a bang, transformations in rival relations often take time to develop. Kupchan (2010), for example, identifies four phases that states go through as they progress from enemies to friends. An ordinal measure of rivalry would perhaps obviate the need to identify rivalry start and end dates and could be used to investigate how rivalrous relations gradually change over time.
If scholars nonetheless want to use rivalry as a dichotomous variable for case selection or to test hypotheses concerning rivalry initiation and termination, an ordinal measure of rivalry could easily be transformed into a dichotomous measure by establishing a measurement cut-point, as is done with other measures. For example, though the Polity measure of democracy is ordinal, states with polity scores of six or greater can be coded as being democratic (Marshall, Gurr, & Jaggers, 2016). Establishing an ordinal measure of rivalry would allow scholars to use the rivalry concept in ordinal or dichotomous fashion as needed.
Developing an ordinal measure of rivalry would not be easy. The starting point, perhaps, would be at the minimal end of a min-max rivalry continuum. Dimensions could subsequently be added to a minimum conceptualization moving toward more fully developed forms of rivalry. Though developing an ordinal measure would be difficult, doing so would open the door to further investigating rivalry complexities.
There are several ways of conceptualizing and operationalizing rivalry. This article provides a comparative assessment of predominant dyadic, complex, and intrastate rivalry approaches. Each approach provides answers to the what, who, and when questions of rivalry. How rivalry is conceptualized and operationalized provides the foundation that empirical knowledge of rivalry processes is built upon.
Each perspective has strengths and weaknesses. A perfect rivalry approach has not yet been, and probably never will be, developed. Choosing between rivalry approaches consequently entails assessing strengths and weaknesses across perspectives. This article provides a basis from which scholars can make judgments on which approach is best suited for their analytic purposes.
There are two primary ways of identifying rivals—using quantitative data or historical narratives. Using quantitative data increases measurement reliability though it may be at the expense of validity, while relying on historical narratives reduces measurement reliability though it may increase validity. Identifying rivals with quantitative data measuring conflict propensities can lead to the identification of false positives and false negatives and the miscoding of some rivalry initiation and termination years. There may be fewer false positives and false negatives identified using historical analysis though there are still difficulties involved in properly identifying rivalry initiation and termination years that may lead to miscodings. Unfortunately, when it comes to the straightforward what, who, and when questions of rivalry there are no easy answers.
Scholars will nonetheless hopefully continue to grapple with pinning down the elusive concept of rivalry. There now exists a wealth of rivalry research from which scholars can draw from. Such research can be used as a basis from which to integrate knowledge on dyadic interstate rivalry and apply the rivalry concept to areas outside of the dyadic interstate context. Advancing rivalry research requires being aware of and seeking to minimize weaknesses of existing rivalry conceptualizations and operationalizations when further developing rivalry approaches.
Understanding rivalry is important not only due to the effects of rivalry but also due to the constitutive nature of rivalry. We define ourselves in part in relation to our rivals. Adopting rivalry roles affects how we view ourselves and how we behave toward those inside and outside of groups with which we identify. Rivalry can perhaps only truly be transcended by coming to grips not only with our perceived adversaries but also ourselves and the impulses that lead us to define others as rivals and act accordingly.
Bennett, D. S. (1996). Security, bargaining and the end of interstate rivalry. International Studies Quarterly, 40(2), 157–183.Find this resource:
Bennett, D. S. (1997). Measuring rivalry termination, 1816–1992. Journal of Conflict Resolution, 41(2), 227–254.Find this resource:
Brecher, M, & Wilkenfeld, J. (1997). A study of crisis. Ann Arbor: University of Michigan Press.Find this resource:
Buzan, B. (1983). People, states, and fear. Boulder, CO: Lynne Rienner.Find this resource:
Colaresi, M. P. (2005). Scare tactics: The politics of international rivalry. Syracuse, NY: Syracuse University Press.Find this resource:
Colaresi, M. P., Rasler, K., & Thompson, W. R. (2007). Strategic rivalries in world politics. Cambridge, U.K.: Cambridge University Press.Find this resource:
Conrad, J. (2011). Interstate rivalry and terrorism: An unprobed link. Journal of Conflict Resolution, 55(4), 529–555.Find this resource:
DeRouen, K. R., Jr., & Bercovitch, J. (2008). Enduring internal rivalries: A new framework for the study of civil war. Journal of Peace Research, 45(1), 55–76.Find this resource:
Diehl, P. F., & Goertz, G. (2000). War and peace in international rivalry. Ann Arbor: University of Michigan Press.Find this resource:
Diehl, P. F., & Goertz, G. (2012). The rivalry process: How rivalries are sustained and terminated. In J. A. Vasques (Ed.), What do we know about war? (2d ed., pp. 83–109). Lanham, MD: Rowman & Littlefield.Find this resource:
Dreyer, D. R. (2010). One issue leads to another: Issue spirals and the Sino-Vietnamese War. Foreign Policy Analysis, 6(4), 297–315.Find this resource:
Dreyer, D. R. (2014). Unifying conceptualizations of interstate rivalry: A min-max approach. Cooperation and Conflict, 49(4), 501–518.Find this resource:
Findley, M. G., Piazza, J. A., & Young, J. K. (2012). Games rivals play: Terrorism in international rivalries. Journal of Politics, 74(1), 235–248.Find this resource:
Hensel, P. R. (1999). An evolutionary approach to the study of international rivalry. Conflict Management and Peace Science, 17(2), 175–206.Find this resource:
Hewitt, J. J. (2005). A crisis-density formulation for identifying rivals. Journal of Peace Research, 42(2), 183–200.Find this resource:
Gochman, C., & Maoz, Z. (1984). Militarized interstate disputes, 1816–1976: Procedures, patterns, and insights. Journal of Conflict Resolution, 28(4), 585–616.Find this resource:
Goertz, G., & Diehl, P. F. (1995). The initiation and termination of international rivalries: The impact of political shocks. American Journal of Political Science, 39(1), 30–52.Find this resource:
Goertz, G., & Diehl, P. F. (2000). (Enduring) rivalries. In M. I. Midlarsky (Ed.), Handbook of war studies II (pp. 222–267). Ann Arbor: University of Michigan Press.Find this resource:
Goldstein, J., & Freeman, J. (1991). U.S.-Soviet-Chinese relations: Routine, reciprocity, or rational expectations? American Political Science Review, 85(1), 17–35.Find this resource:
Hafner-Burton, E., Kahler, M., & Montgomery, A. H. (2009). Network analysis for international relations. International Organization, 63(3), 559–592.Find this resource:
Hensel, P. R. (2001). Contentious issues and world politics: The management of territorial claims in the Americas, 1816–1992. International Studies Quarterly, 45(1), 81–109.Find this resource:
Hensel, P. R., Mitchell, S. M, Sowers, T. E., II, & Thyne, C. L. (2008). Bones of contention: Comparing territorial, maritime, and river issues. Journal of Conflict Resolution, 52(1), 117–143.Find this resource:
Jo, D.-J., & Gartzke, E. E. (2007). Determinants of nuclear weapons proliferation. Journal of Conflict Resolution, 51(1), 167–194.Find this resource:
Jones, D. M., Bremer, S. A., & Singer, J. D. (1996). Militarized interstate disputes, 1816–1992: Rationale, coding rules, and empirical patterns. Conflict Management and Peace Science, 15(2), 163–213.Find this resource:
Klein, J. P., Goertz, G., & Diehl, P. F. (2006). The new rivalry dataset: Procedures and patterns. Journal of Peace Research, 43(3), 331–348.Find this resource:
Kroenig, M. (2009). Importing the bomb: Sensitive nuclear assistance and nuclear proliferation. Journal of Conflict Resolution, 53(2), 161–180.Find this resource:
Kupchan, C. A. (2010). How enemies become friends: The sources of stable peace. Princeton, NJ: Princeton University Press.Find this resource:
Maoz, Z., & Mor, B. D. (2002). The strategic evolution of enduring international rivalries. Ann Arbor: University of Michigan Press.Find this resource:
Marshall, M. G., Gurr, T. R., & Jaggers, K. (2016). Polity IV project: Political regime characteristics and transitions, 1800–2015. Retrieved from http://www.systemicpeace.org/polityproject.html
Mitchell, S. M., & Thies, C. G. (2011). Issue rivalries. Conflict Management and Peace Science, 28(3), 230–260.Find this resource:
Rapkin, D. P. (1999). The emergence and intensification of U.S.-Japan rivalry in the early twentieth century. In W. R. Thompson (Ed.), Great power rivalries (pp. 337–370). Columbia: University of South Carolina Press.Find this resource:
Rasler, K., Thompson, W. R., & Ganguly, S. (2013). How rivalries end. Philadelphia: University of Pennsylvania Press.Find this resource:
Reiter, D., Stam, A. C., & Horowitz, M. C. (2016). A revised look at interstate wars, 1816–2007. Journal of Conflict Resolution, 60(5), 956–976.Find this resource:
Rider, T., Findley, M., & Diehl, P. F. (2011). Just part of the game?: Arms races, rivalry, and war. Journal of Peace Research, 48(1), 85–100.Find this resource:
Singh, S., & Way, C. R. (2004). The correlates of nuclear proliferation: A quantitative test. Journal of Conflict Resolution, 48(6), 859–885.Find this resource:
Thompson, W. R. (1995). Principal rivalries. Journal of Conflict Resolution, 39(2), 195–223.Find this resource:
Thompson, W. R. (1999). Why rivalries matter and what great power rivalries can tell us about world politics. In W. R. Thompson (Ed.), Great power rivalries (pp. 3–28). Columbia: University of South Carolina Press.Find this resource:
Thompson, W. R. (2001). Identifying rivals and rivalries in world politics. International Studies Quarterly, 45(4), 557–586.Find this resource:
Thompson, W. R. (2003). A streetcar named Sarajevo: Catalysts, multiple causation chains, and rivalry structures. International Studies Quarterly, 47(3), 453–474.Find this resource:
Thompson, W. R. (2015). Trends in the analysis of interstate rivalries. In R. Scott & S. Kosslyn (Eds.), Emerging trends in the social and behavioral sciences (pp. 1–14). Hoboken, NJ: John Wiley.Find this resource:
Thompson, W. R., & Dreyer, D. R. (2012). Handbook of interstate rivalries. Washington, DC: Congressional Quarterly Press.Find this resource:
Valeriano, B., & Powers, M. (2016). Complex interstate rivals. Foreign Policy Analysis, 12(4), 552–570.Find this resource:
Vasquez, J. A. (2009). The war puzzle revisited. Cambridge, U.K.: Cambridge University Press.Find this resource:
Wayman, F. (2000). Rivalries: Recurrent disputes and explaining war. In J. A. Vasquez (Ed.), What do we know about war? (pp. 219–234). Lanham, MD: Rowman & Littlefield.Find this resource:
Womack, B. (2011). Asymmetric rivals: China and Vietnam. In S. Ganguly & W. R. Thompson (Eds.), Asian rivalries: Conflict, escalation, and two-level games (pp. 176–194. Stanford, CA: Stanford University Press.Find this resource:
Zinnes, D. A. (1976). The problem of cumulation. In J. N. Rosenau (Ed.), In search of global patterns (pp. 161–166). New York: Free Press.Find this resource: