The ORE of Politics will be available for subscription in late September. Speak to your Oxford representative or contact us to find out more.

Dismiss
Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS (politics.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 16 August 2017

Theories of Civil War Onset: Promises and Pitfalls

Summary and Keywords

Empirical research on civil war onset has been largely dominated by two approaches: a correlational or “correlates of civil war” approach which seeks to identify country-level characteristics associated with a higher likelihood of civil war outbreak, and a bargaining approach which starts from the assumption that warfare is costly and which views civil conflict as a by-product of bargaining failures. Correlational and bargaining studies of internal conflict onset have reached an analytical plateau because they fail to specify the precise mechanisms that yield civil warfare instead of a different type of violent or nonviolent outcome. An alternative, contentious framework is advanced for studying civil war onset. This framework situates the conflict event within a larger cycle of contention and specifies the mechanisms through which civil conflict is most likely to occur. According to this contentious perspective, civil wars are commonly produced by the combination of one structural condition—a state crisis of authority and/or legitimacy—and the interdependent effect of two mechanisms—radicalization and militarization. Through theory development and vignettes from a handful of civil war cases, the article makes the case that the contentious approach holds promise for elucidating how exactly civil conflicts break out. Despite holding initial explanatory power, the contentious theory of civil war onset advanced herein awaits more systematic empirical testing.

Keywords: bargaining theories of war, civil war, contentious politics, correlates of civil war, empirical international relations theory, ethnic conflict, mechanisms, violence

Civil wars are conventionally viewed as instances of coordinated, sustained violence between political organizations that are subject to a common authority at the outset of hostilities (Kalyvas, 2007; Levy & Thompson, 2011). According to this widely accepted definition, civil war encompasses coordinated, sustained violence between a government and a rebel organization or between two or more insurgent groups. Civil wars are the most common forms of conflict in the post-1945 environment. Although relatively rare occurrences compared to the number of peaceful years, civil wars can be particularly destructive: they take lives and tear apart communities, destroy physical infrastructure and human capital, arrest socio-economic development, and leave behind indelible marks on collective psyches.

If they are so devastating, why do civil wars break out? This article examines the evolution of empirical research on the causes of internal conflict onset. Particular attention is given to two dominant approaches to internal conflict outbreak: (a) a correlational approach, originating in the early 1970s, which comprised a first wave of empirical civil war research and which sought to identify country-level characteristics that tend to be associated with a greater likelihood of observing civil war onset, and (b) a bargaining approach, originating in the late 1990s, which characterized a second wave of empirical civil war research and which focused on the strategic interaction between disputants and viewed internal conflict outbreak as a product of bargaining failures. The purpose of this article is not to review the main findings in the vast civil war research program1 but, rather, to discuss major contributions and limitations of these two dominant empirical approaches to internal conflict. In light of the shortcomings that beset these frameworks, the article proposes an alternative, contentious approach that may provide a fuller understanding of the processes through which civil wars break out.

The Correlational Approach to Civil War Onset

The first wave of empirical work on civil war onset focused almost exclusively on the structural conditions that make countries more prone to experiencing civil war or, more precisely, on those factors that allow domestic opponents to overcome collective action problems and mobilize against internal rivals (Olson, 1965). Rather than engage in complex theorizing of actors’ strategic interactions and of the dynamic processes through which internal conflict might ensue, this first wave of empirical research viewed civil warfare through a static, correlational lens. The “correlates of civil war” wave gained visibility and prominence as a mirror image of the “correlates of war” approach that examined the conditions associated with the outbreak of interstate war, which, by the early 1990s, dominated much of the empirical work on international conflict. At the beginning, the “correlates of civil war” literature was fairly amorphous, but it was soon dominated by two broad classes of explanations: grievance-based explanations and greed-based explanations. Grievance-based accounts (Gurr, 1970) focused on background conditions or government actions, such as repression, discrimination, poverty, income inequality, ethno-linguistic fractionalization, that create grievances at the individual or group level and galvanize collective action. Greed-based frameworks concentrated on rebels’ opportunity costs for engaging in violence (Collier, Hoeffler, & Roehner, 2009; Collier & Hoeffler, 2004; Fearon & Laitin, 2003). Adherents of this approach typically examined factors that encouraged rebel predatory behavior, such as the presence of lootable mineral resources and the existence of an oil dependent economy, or conditions that favor or make insurgency more feasible, like state weakness, rough terrain, external support from other countries or ethnic diasporas, recent outbursts of civil conflict or political instability, large population, state age (i.e., newness of a state), or regime type (i.e., the risk of civil war breakout tends to be higher in anocracies, transitioning regimes that display both authoritarian and democratic characteristics). Although they relied on various datasets and statistical models, the “correlates of civil war” studies produced a “near consensus that poverty, large populations, a low level of economic development, a prior history of civil war, and political instability increase a country’s risk of civil war. There is also some evidence that a dependence on natural resources, the existence of ethnic diasporas, concentrated populations, rough terrain, and anocracies are positively associated with the outbreak of civil war” (Walter, 2009, p. 244).

The “correlates of civil war” literature sought to uncover variables consistently linked with civil war rather than advance unified theories of internal conflict onset. As a consequence, it suffered from a litany of shortcomings: (a) The theoretical connections between certain country-level characteristics, such as income inequality, and civil conflict were not so clear cut (Blattman & Miguel, 2010); (b) Ad-hoc theoretical explanations were advanced to account for correlational patterns gleaned from the data; (c) Power and survival as primary motivations for armed group behavior were given insufficient treatment (Vinci, 2006); (d) Time-invariant covariates, such as gross domestic product (GDP) per capita, could not adequately explain the variability in civil war onsets (Young, 2013); (e) The root (underlying) and proximate causes of civil war were not easily detectable (Blattman & Miguel, 2010); (f) Outcomes were over-determined—too many mechanisms were proposed to account for statistical patterns; (g) Empirical tests did not always distinguish between rival theoretical explanations; (h) Measurement problems were abundant; (i) Endogeneity was pervasive but not properly addressed; (j) Interaction effects were insufficiently explored; (k) Most such studies had trouble explaining the variation in civil war outbreak across countries that shared similar characteristics (Walter, 2009). Overall, the first wave of empirical studies on civil war onset provided a fairly opaque view of exactly how internal conflicts emerge because they paid little attention to how rebel organizations coalesce and splinter, how they interact with the government, other insurgents, or third parties, or how this interaction unfolds to produce varying violent or nonviolent outcomes.

The Bargaining Approach to Civil War Onset

The second wave of empirically based theories of civil war onset was dominated by rationalist accounts that regarded conflict as an outcome of bargaining failures. Given the dearth of complex theorizing in the “correlates of war” literature, bargaining theories of conflict were enthusiastically embraced by internal conflict scholars because they managed to address at least some of the problems plaguing correlational studies. Bargaining approaches view civil war as a puzzle: Given the existence of less costly alternatives to settling disputes, why do actors resort to violent costly behavior? As a subset of rationalist accounts for war, bargaining approaches envisioned conflict as a by-product of three factors: informational asymmetries, credible commitment, and issue indivisibilities (Fearon, 1995; Powell, 2002; Reiter, 2003). Initially developed to explain international conflict, the bargaining logic was quickly applied to civil wars, especially to ethnic conflicts (Walter, 1997, 2002, 2009). Bargaining accounts for civil war onset start from the assumption that war is costly and that domestic opponents should prefer a settlement that offers them exactly what they would obtain at the end of hostilities. As with international rivals, the problem is that three recurrent conditions make that type of settlement unlikely. One such condition is information asymmetry—a situation in which actors hold private information about their military capability and determination to fight against their enemies. Private information is “knowledge an actor possesses that is not available to the other” (Lake, 2003, p. 83). Information asymmetries occur because opponents have rational incentives to misrepresent information about capabilities and resolve. Walter (2009, p. 246) holds that informational asymmetries are particularly acute in civil war situations for at least two reasons: (a) Because rebels operate clandestinely, governments often have trouble monitoring their capabilities and strategies; (b) Because insurgents are generally weaker than the government, they have strong incentives to misrepresent their military prowess and determination to fight in order to bolster their bargaining leverage.

Bargaining theories suggest that, even in those situations where informational asymmetries may not be severe, war could still break out because antagonists cannot credibly commit to abide by an agreement. Problems of commitment tend to arise with expected shifts in the relative power distribution between disputants and with divisions within the organizational structure of each actor—veto players with a say over government policy and radical factions within an insurgency may oppose a proposed settlement and continue fighting (Cunningham, 2011). Commitment issues tend to be severe in internal conflicts because expected shifts in the power distribution in governments’ favor following negotiated settlements, which generally provide for insurgent demobilization, offer incentives for these governments to renege on promises. As Walter (2009) points out, “governments can offer to reform the political process, share power, or transfer autonomy to competitors, but these weaker competitors will have little ability to penalize a government should it fail to follow through” (p. 246). Finally, a third bargaining explanation for war onset relates to the type of issue under dispute. Certain issues, such as control over territory, may be indivisible in that they cannot be split among opponents (e.g., the status of Jerusalem). Issue indivisibilities may arise because of inherent properties of the issue under dispute (Hassner, 2009) or because of actors’ inability to commit to a certain distribution of the disputed good (Goddard, 2009). Hence, in the latter case, issue indivisibility emerges as a consequence of bargaining failures, not a cause thereof.

Overall, bargaining models offer a more realistic view of the processes that might lead to inefficient behavior and civil war. By focusing on the strategic interaction between adversaries, they better capture the dynamic character of internal conflict onset. Bargaining frameworks mark an important departure from the static correlational studies that merely identified state-level characteristics associated with a higher likelihood of civil warfare. While also taking into account structural conditions, bargaining approaches privilege, and rightly so, actors’ agency—especially when the unitary actor assumption is abandoned in favor of the more realistic view of actors comprising of multiple subgroups or factions with divergent preferences and interests. Although they provide a better understanding of internal conflict dynamics compared to the largely atheoretical “correlates of civil war” approaches, bargaining frameworks themselves suffer from several shortcomings. For instance, they do yield falsifiable predictions, “but few articulate the precise empirical tests that would distinguish among alternative mechanisms” (Blattman & Miguel, 2010, p. 17). Also, the ability of empirical methods to test a bargaining model “appears fundamentally limited because of the high information demands of the most direct tests of the bargaining model” (Reiter, 2003, p. 37). Furthermore, these models do not clearly distinguish between types of bargaining failures that produce civil warfare from the types of bargaining failures that produce different kinds of violent or nonviolent outcomes. The remainder of this article discusses some areas where the explanatory power of bargaining frameworks is limited and advances an alternative approach to studying civil war onset that may address at least some of the shortcomings plaguing the still dominant bargaining accounts for civil conflict.

Pitfalls of Bargaining Explanations for Civil War Onset

Bargaining accounts of internal conflict have limited explanatory power in those circumstances where there is little, if any, strategic bargaining or interaction between rivals prior to the outbreak of violence. Consider the onset of civil war in Libya (2011) in the aftermath of Arab Spring protests that were sweeping the larger Middle East and North African (MENA) region. In this case, bargaining between the regime of Muammar Gaddafi and the various militias comprising the Libyan insurgency was limited or, some would argue, virtually nonexistent. Rather than an outcome of bargaining failure, the war seems to have been produced by a situation in which the antagonists (i.e., the rebels and the regime) had rational incentives to fight: both the insurgents and the government were confident that they would prevail militarily and, as a consequence, did not make any effort to prevent the onset of hostilities. In a conventional bargaining scenario, one would have expected to see the Libyan regime and rebel leaders devoting significant effort to “put serious offers on the table—offers that have a positive probability of being accepted” (Fearon, 2013, p. 2). Instead, on the eve of the Libyan civil war, neither side advanced any serious offer for a negotiated settlement. As Fearon (2013, p. 3) notes, the position adopted by the antagonists seems to have been: “Let’s just fight for a while to see if we can win outright, or demonstrate that we must ultimately be given better terms.” If anything, the internal conflict that broke out in Libya in the wake of the Arab Spring revealed an initial preference for fighting instead of bargaining.

Actors’ preference for fighting that was visible in the case of the Libyan civil conflict poses an important challenge to the war inefficiency assumption that underpins bargaining explanations for warfare. The main theoretical premise underlying bargaining, and, more generally, rationalist accounts for conflict is that war imposes large costs on all actors involved in the dispute relative to an alternative peaceful outcome. From this vantage point, war is thought to be inefficient in that there always exists an ex ante bargain that avoids the costs of conflict and leaves all parties better off. The problem is, according to scholars operating within the rationalist tradition, that informational asymmetries, credible commitment problems, and issues of indivisibility prevent opponents from reaching an ex ante deal that gives them exactly what they would have obtained at the end of the conflict but that would have allowed them to avoid the costs of war (Fearon, 1995; Walter, 2002; Walter, 2009). The circumstances surrounding the outbreak of the Libyan civil war depart from the assumption that war was perceived to be inherently costly or inefficient by the government and the rebels. Clearly, the Gadhafi regime and its opponents expected a higher utility from warfare than from peace. If some actors, as was the case in Libya, do indeed display an affinity for conflict, then the costly war edifice on which conventional rationalist theories are built begins to unravel. Libya is not an isolated case where actors’ behavior might suggest a preference for war. In fact, the range of circumstances where strategic actors might regard war as less costly than the peace alternative is larger. Coe (2012), for example, identifies three common sources of costs in peace: predation (extraction of tribute by one actor from another); imposition (imposition of punitive measures, such as sanctions, by one actor upon another); and arming (“costly attempts to shift the balance of power”). Coe (2012) argues that predation explains the American War of Independence, imposition explains the outbreak of civil warfare after the end of the Gulf War, and arming explains the Iraq War. Coe’s work offers a welcome corrective to standard rationalist analyses of conflict because it suggests that, in some circumstances, outcomes other than war—predation, imposition, arming, proxy conflicts, restrictions on foreign investment—can be equally, if not more, costly. In his view, “all these things are the result of underlying commitment problems and asymmetric information of varying severity, and they are all partial substitutes for each other; war occurs only when it is the most cost-effective option. So, the mechanisms for bargaining failure that rationalist scholars thought were explanations for war are only partially so—we also need to know why war was chosen over other costly behaviors” (Coe, 2012, p. 6). This argument has important theoretical and empirical implications: if the range of actors’ costly actions goes well beyond war, then standard bargaining frameworks are internally inconsistent and empirically inconclusive. The internal inconsistency stems from the inability of existing theories to explain why bargaining failures produce war rather than a different type of outcome. The empirical inconclusiveness derives from the type of data conventionally used to test the theoretical expectations produced by bargaining models of war. If bargaining failures can yield outcomes other than war, the empirical tests that do not include these possible outcomes are biased and, thus, largely inaccurate.

This problem is particularly salient in studies of civil war where the range of possible outcomes that result from bargaining failures can be quite extensive. Bargaining breakdown between domestic actors can generate various violent or nonviolent forms of contention: demonstrations, genocides, politicides, revolutions, uprisings, purges, rebellions, riots, strikes, protests, acts of civil disobedience, nonviolent resistance, or other forms of social and political disorder. Civil wars rarely break out in a contentious vacuum; rather, internal conflict is typically embedded in larger cycles of domestic contention that often display violent and nonviolent manifestations of unrest (Tarrow, 2007; Tilly & Tarrow, 2007). For example, the ongoing Syrian civil war (2011–) erupted amid a cycle of contention that began in March 2011 with peaceful manifestations inspired by the Arab Spring revolutions in Egypt and Tunisia. The minor power-sharing concessions offered by the Assad regime did not placate the opposition; as a result, peaceful protests quickly diffused across much of the country. What began as nonviolent antigovernment mutiny quickly morphed into violent regime repression, radicalization of opposition demands, and insurgent military mobilization, primarily along sectarian lines. In a matter of months, the country plunged into a protracted and violent civil war that shows few signs of abating. The factors that precipitated the outbreak of internal conflict in Syria cannot be fully grasped if we analyze it in isolation from other forms of contention—nonviolent protest, government repression, insurgent violence—that preceded it. Conventional bargaining accounts provide incomplete answers to the civil war puzzle in Syria because they fail to specify which types of bargaining failures produced the different types of outcomes—nonviolent protest, government repression, insurgent violence, civil war outbreak—across the entire cycle of contention that started with the Arab Spring–inspired demonstrations. Stating that a particular type of outcome occurred because bargaining between actors failed is theoretically indeterminate. For bargaining frameworks to carry satisfactory explanatory power in analyses of internal conflicts, they need to specify the precise mechanisms that produced civil warfare rather than other types of outcomes across the broader contentious cycle in which the conflict event is embedded. Put otherwise, bargaining accounts of civil war alone are incomplete; explanations of other violent or nonviolent forms of contention that predated the onset of internal conflict are also necessary.

Situating civil warfare in a larger framework of contentious events has important theoretical and empirical implications. Theoretically, if internal conflict is embedded in a broader cycle of contention that displays various types of collective action at different stages, then a unified bargaining framework may abstract too much away from the processes that produce the outcome of interest. Bargaining theories help illuminate civil war onset only inasmuch as they are also able to elucidate other types of events that can be observed within a given contentious cycle. A contentious politics angle de-exceptionalizes civil war and locates it within a larger repertoire of collective action (della Porta, 2013, p. 15; Rule, 1988, p. 170). To argue that bargaining failures led to civil war, it behooves the analyst to explain the processes through which bargaining failures might have produced different forms of violent or nonviolent collective action before the eruption of the conflict. In other words, the observable implications of bargaining theories need to be traced along the entire contentious cycle, not only cross-sectionally at the point of civil war onset. For instance, a comprehensive bargaining account of the breakout of violence in Syria needs to explain not only why the collapse of bargaining between the government and the opposition led to violence but also why bargaining failures produced other types of contentious outcomes that preceded the eruption of hostilities. More generally, we still lack theoretically consistent answers to two key questions: What kind of bargaining failures are more likely to produce civil conflict than other types of contentious events? And under what circumstances might similar bargaining problems produce different contentious outcomes?

Looking at civil war through a contentious lens has empirical consequences as well, especially for large-N analyses of internal conflict onset. If one starts from the theoretical premise that civil war is embedded in a wider cycle of contention that includes multiple stages at which the strategic interaction between actors may yield varying violent and nonviolent outcomes (Chenoweth & Stephan, 2011; Demirel-Pegg, 2014), then the conventional units of analysis—the state-year, the conflict-year, or the dyad-year—are no longer appropriate. Dominant, bargaining-centered studies of civil war suffer from a severe selection bias because they typically exclude not only instances of non-violent mobilization (Kalyvas, 2006, p. 32) but also episodes in which bargaining failures produced forms of contention other than warfare, such as opposition purges, pogroms, or civilian riots that are observable across multiple cases like the ones in Nagorno-Karabakh in Azerbaijan (1992) or Tamil Eelam (1983) in Sri Lanka. In both of these instances, civil war emerged out of an escalatory process of contention that started with interethnic pogroms in Sumgait and Baku in the former case, and with a Sinhalese riot that led to the destruction of the Jaffna Library in Tamil Eelam in the latter case (DeVotta, 2004; de Waal, 2003). The point about including acts of violent and nonviolent collective action in models of civil war onset is not theoretically or empirically trivial because the same mechanisms posited to have produced civil war in existing bargaining theories could also be related to cognate phenomena that are not properly observed and measured and, thus, are routinely excluded from model specifications. Even in the absence of proper theorizing on civil war as an outcome of a contentious process that displays various forms of violent or nonviolent collective action, large-N bargaining studies—which typically rely on civil war data from the Correlates of War (COW) or the Uppsala Conflict Data Program/Peace Research Institute Oslo (UCDP/PRIO)—should at the very least include other types of contentious events that occurred during the analytical time frame to guard against the bias inherent in selecting only on the civil war outcome.

A Contentious Politics Approach to Civil War Onset

If civil war is viewed as an emergent phenomenon situated within a complex process of political contention, then the processual, dynamic nature of internal conflict onset might be better captured if one adopts the “cycle of contention” or the “contentious spell” as the unit of analysis. Cycles of contention and/or contentious spells encapsulate the whole range of violent and nonviolent forms of collective action that occur within a bounded time frame (Tarrow, 2011, 2012; Tilly, 2001). In the social movements literature, a cycle of contention describes a “phase of heightened conflict across the social system, with rapid diffusion of collective action from more mobilized to less mobilized sectors and [with] a rapid pace of innovation in the forms of contention employed” (Tarrow, 2011, p. 199). How exactly can contentious cycles be captured empirically? A cycle of contention can be operationalized as a succession of “interrelated collective actions and reactions to them whose aggregate frequency, intensity, and forms increase and then decline in rough chronological proximity” (Tarrow, 2012, p. 134). The observable indicator of the beginning of a contentious cycle would be the occurrence of organized collective action either by the government or by a nonstate actor. On the regime side, such organized collective action could take the form of repression, for instance. On the opposition side, organized collective action could initially manifest itself in the form of protests or riots that may or may not escalate to violence, for example. The end of a cycle of contention/contentious spell would be empirically marked by government victory, rebel victory, peace agreement, or the lack of nonviolent or violent forms of contention for a certain period of time. Contentious cycles can be observed at the country level, where violent and nonviolent forms of contention can be traced within the respective country for a given time frame demarcated by the onset and termination of mobilization. Cycles of contention can also be tracked at the dyadic level, where a government-nonstate actor dyad, or a nonstate actor-nonstate actor dyad can be observed throughout all stages of the mobilization process.

By locating civil warfare within cycles of contention, one can better ascertain how civil wars might escalate from various types of violent or nonviolent contention or how they may later mutate into other contentious phenomena (Cunningham & Lemke, 2014). Della Porta notes that actor competition tends to intensify during contentious cycles, “as social movement organizations multiply and then split over the best strategies to adopt, some of them choosing more radical ones” (della Porta, 2013, p. 75). Placing civil warfare within larger cycles of contention puts the analyst in an ideal position to explore the mechanisms that connect the different stages of the cycle.

Contention, Mechanisms, and Civil War

The idea of embedding internal conflict within contentious cycles implies that civil war onset is a process rather than a one-shot event that occurs in isolation from the broader context in which actors interact. This perspective bears resemblance to the Tillyan political process approach, which envisages mobilization as an outcome that emerges gradually out of actors’ contentious interactions (Tilly, 1978). Looking at the contentious cycle/spell as the unit of analysis can provide a firmer grip on the mechanisms that generate civil war outbreak. Mechanisms “refer to intermediary steps between conditions and outcomes” (della Porta, 2013, p. 24). According to Hedström & Ylikovski (2010, p. 54), mechanisms are central to causal inferences: “the knowledge that there is a mechanism through which X influences Y supports the inference that X is a cause of Y.” For example, poverty—a consistent predictor of civil war onset in correlational studies—does not in and of itself cause internal conflict; “it is mechanisms that provide the link” (Ron, 2005, p. 444). Conversely, “the absence of a plausible mechanism linking X to Y gives us a good reason to be suspicious of the relation being a causal one” (ibid.).

Some argue that a mechanism-centered approach is necessary for future progress on civil war research (Checkel, 2013, p. 24). Indeed, a fuller understanding of internal war onset implies knowing not only whether certain factors are associated with a higher likelihood of observing civil warfare but also how exactly these factors operate to produce the outcome. “Correlates of civil war” and bargaining-centered explanations remain incomplete without a specification of mechanisms that describe how independent variables are connected to the pathways through which internal conflict breaks out. Mere associations between variables conceal an empty shell of causation because they do not tell a compelling story about why covariations run in certain directions rather than others (Gerring, 2010, p. 1502). As Bennett (2013) notes, a focus on causal mechanisms does incur an important cost—the loss of parsimony, but this might be necessary for grasping the complexity of civil warfare. After all, “a mechanism is an irreducibly causal notion. It refers to the entities of a causal process that produces the effect of interest” (Hedström & Ylikoski, 2010, p. 50). Hence, a mechanism-based approach to civil initiation is likely to offer a more realistic view of the causal processes that can lead to internal conflict (Blattman & Miguel, 2010; Checkel, 2013).

A mechanism is the connective fabric between variables of interest (Tarrow, 2011, p. 186), the pathway or process by which an outcome is produced (Gerring, 2010, p. 1501). According to Tilly (2001), mechanisms “form a delimited class of events that change relations among specified sets of elements in identical or closely similar ways over a variety of situations” (p. 25). A mechanism is ontologically distinct from a variable—unlike variables, mechanisms are not properties of the units of analysis. Rather, “they reside above and outside the units in question and describe the relationship between the units of analysis” (Falleti & Lynch, 2009, p. 147). Causal mechanisms are rarely observable in the literal sense of the term (Gerring, 2007, p. 245). They are “portable concepts that explain how and why a hypothesized cause, in a given context, contributes to a particular outcome” (Falleti & Lynch, 2009, p. 1144). From this vantage point, causation is inextricably linked to mechanisms: to argue that X causes Y means not only to find an association between X and Y but also to explicate the process through which changes in X induce changes in Y in a particular direction. Additionally, an analytically useful mechanism is one that “must be general enough to be portable across different contexts but may produce different results in analytically nonequivalent contexts; mechanisms alone cannot cause outcomes. Rather, causation resides in the interaction between the mechanism and the context in which it operates” (Falleti & Lynch, 2009, p. 1146). Thus, a “mechanistic” view of causation implies theorizing about both the process through which X affects Y as well as the context in which such process unfolds.

Mechanisms are frequently invoked by political scientists, sociologists, and economists alike in their attempt to elucidate various phenomena. In traditional economic theory, Adam Smith’s “invisible hand” of supply and demand market forces is the quintessential mechanism that ensures an efficient level of production, consumption, and distribution of goods. Hirschman (1970) argues that two widespread mechanisms affect the fate of inefficient organizations, regardless of whether they are firms or states: “exit” and “voice.” Members of declining organizations have two options to address organizational inefficiency: they can select “exit” and leave the organization in favor of a better alternative or they can choose “voice” and attempt to alter the operation of the polity. “Exit” and “voice” are not independent of one another; rather, they operate in tandem. As Hirschman (1970) himself states “the actual level of voice feeds on [the] lack of opportunity to exit” (p. 34). A “threshold” mechanism lies at the core of many sociological accounts for social mobilization. For example, Granovetter (1978) notes that sustained collective action requires the participation of critical number of participants—a “threshold”—that signals to other potential participants that the risks of engaging in collective behavior are acceptable. Once a protest threshold is reached, the rate of participation tends to increase dramatically because more individuals will fall in line if they see a large number of protestors joining in at high rates. Kuran (1989) dwells on Granovetter’s “threshold” mechanism to explain the individual rate of participation in the East-European communist revolutions of 1989. He advances the idea that authoritarian regime disapproval is not sufficient for mobilizing large numbers of people for collective action against the government. In repressive regimes, collective action is hampered by a high threshold for personal participation in collective action. However, after a critical mass of participants manages to sustain collective action for longer periods of time, dissatisfied individuals are highly likely to join the protest. As Kuran puts it, in the context of the 1989 revolutions, sustained collective action from a critical mass of protestors was the “spark” that set alight the revolutionary “prairie fire.” Another landmark study that relies on the “threshold” logic to explain mobilization for collective action is Rasler (1996) who investigates the escalation of popular mobilization in the 1978–1979 Iranian Revolution. Her main argument is that government’s inconsistent use of both accommodation and repression lowered the threshold for participation in collective action by wide segments of the population (Rasler, 1996, p. 144). Because of inconsistent government policies of repression and concessions, dissent intensified both across segments of the population and across regions (through a spatial diffusion effect), which eventually led to the overthrow of the Iranian Shah in February 1979.

The main virtue of a mechanism-based approach to civil war resides in the ability to illuminate the processes through which covariations occur—mechanisms tell the “story” behind the association between variables. Although mechanisms display latent explanatory properties (i.e., they can explain how variation in X produces variation in Y), they are not panacea for comprehensive causal accounts. Mechanisms are “not some sort of magic bullet for causal inference” (Hedström & Ylikoski, 2010, p. 54). Gerring (2010) argues that a mechanism-based approach to social behavior can suffer on several accounts. One is that “specifying a causal mechanism is sometimes a highly speculative affair. The posited mechanism may be highly specific but at the same time remain entirely unproven, and perhaps highly dubious. This is what prompts some partisans of causal mechanisms to insist that mechanisms not only be specified but also tested empirically wherever their veracity is in doubt” (Gerring, 2010, p. 1505). The important question is, then, whether mechanisms can really be tested. How does one test a “story”? Gerring (2010, pp. 1499–1511) states that empirical tests of mechanisms are indeed desirable “but often unrealistic” because for the same association between X and Y, one can find a “litany of theoretically plausible mechanisms.” Another problem posed by a mechanism-based approach is that concepts underlying a mechanism are sometimes vague and, thus, hard to operationalize. They “either resist measurement or are liable to many plausible indicators” (Gerring, 2010, p. 1510). While X and Y are likely to be observable and measurable, the territory between X and Y, by contrast, “is often a morass” (Gerring, 2010, p. 1517). A third challenge is that, as discussed above, the same mechanism can have varying effects on an outcome depending on the context in which it operates (Gerring, 2010, p. 1511). In these situations, it would be impossible to “say anything definitive about the causal mechanisms at work, unless one has correctly identified the circumstances in which X is positively, or negatively, associated with Y” (Gerring, 2010, p. 1511).

These shortcomings notwithstanding, tracing the mechanisms through which civil war emerges within contentious cycles is likely to provide greater analytical purchase than static correlational analyses or theoretically indeterminate bargaining frameworks. The next section lays out an alternative, mechanism-based framework for studying civil war onset. The framework suggests that many instances of internal conflict outburst display an underlying or structural condition—a crisis of authority and/or legitimacy—and two interdependent mechanisms that may produce sustained, organized violence: radicalization and militarization. Radicalization signals a change toward more extreme actor tactics, whereas militarization marks a commitment toward the adoption of more violent forms of contention.2 Radicalization and militarization can mark two shifts: (a) the shift from loosely coordinated contentious action toward organized rebellion where actors systematically use violence in pursuit of their strategic goals; or (b) the shift from coordinated collective action that employs nonviolence to organized rebellion that relies on violence as its main tactic.

A Mechanism-Based Framework for Civil War Onset

A common pathway through which civil warfare breaks out within a given contentious cycle/spell involves the presence of a state crisis of authority and/or legitimacy and the escalation of actor interactions through two interlinked mechanisms: radicalization and militarization (see Figure 1).3

Theories of Civil War Onset: Promises and PitfallsClick to view larger

Figure 1: A common process leading to civil war outbreak

A necessary, but not sufficient, condition for civil war outbreak is the existence of a crisis of state authority and/or legitimacy. This is the underlying or structural characteristic that marks the beginning of a contentious cycle/spell and functions as the sine qua non of many, if not most, civil wars. Crises of authority and/or legitimacy appear when there are multiple power contenders who challenge the sovereign actor (the central government) and undermine its functions. When crises of authority and/or legitimacy occur, we are dealing with a fragmented body politic in which challengers constantly try to undercut each other’s strategies using nonviolent or violent means. This internal competition may “arise from ideological conflict, from competition for space in a static organizational sphere, or from personal conflicts for power between leaders” (della Porta, 2013, p. 79). During such crises, parallel structures emerge that rival or seek to replace the sovereign government as the sole purveyor of authority. Essentially, crises of authority and/or legitimacy are commonly associated with a declining capacity of governments to perform key functions of sovereign statehood (violence monopoly; extraction; redistribution). These types of crises can arise in various situations, but they are more prevalent in authoritarian regimes, in newly formed or weak states, in ethnically diverse societies, or in the aftermath of imperial or federal collapse.

In the contentious politics literature, a state authority/legitimacy crisis provides the political opportunity structure for organized collective action (McAdam, Tarrow, & Tilly, 2001). Political opportunity structures emerge in the presence of factors that enhance (or inhibit) prospects for mobilization (Meyer & Minkoff, 2004). Political opportunity structures obtain when exogenous or endogenous conditions open up the space for violent or nonviolent contention among domestic actors. If we look closely at the theoretical assemblage on civil war onset, we cannot help but notice that existing theories, especially the “correlates of civil war” approaches, overwhelmingly focus on the political opportunity structure phase of the civil war contentious cycle/spell—in other words, they mainly elucidate why a crisis of authority and/or legitimacy appears in the first place.

The state-level characteristics examined in “correlates of civil war” studies help illuminate the circumstances under which crises of authority/legitimacy are most likely to emerge rather than when civil war is most likely to break out. Although the “correlates of civil war” literature is large and fairly diverse, it typically focuses on recurrent state-level variables that can be clustered in a handful of categories (Dixon, 2009): demographic variables (e.g., ethno-linguistic fractionalization, religious diversity/polarization; population density, and minority concentration); geographic-environmental variables (e.g., topographic characteristics, environmental degradation, region, and neighborhood); natural resource variables (e.g., oil, diamonds, and other lootable resources); socio-economic variables (e.g., GDP growth, trade, investment, income inequality, literacy rate, mortality rate, exclusion, and discrimination); political variables (e.g., regime type, conflict history, rivalry, and new state). Many of these covariates have been shown to be systematically linked to civil war. Does that mean that they are causally connected to internal conflict onset? Skeptics would dispute this assertion on both theoretical and methodological grounds. As mentioned previously, “correlates of civil war” approaches suffer from multiple shortcomings: (a) Predictors of civil war are largely static or slow-moving; (b) Proxies for concepts such as minority grievances are rather crude; (c) The mechanisms linking state-level attributes to civil war outcomes are poorly specified; (d) Many findings are contradictory; (e) The statistically significant correlational patterns gleaned from these studies often come out as artifacts of measurement and modelling choices. More substantially, however, if we perceive civil warfare as an outcome of a contentious process that begins with a crisis of authority/legitimacy, the state-level features that are purported to explain the outbreak of internal conflict only account for when we are most likely to observe an authority/legitimacy crisis in the first place.4 Thus, they are able to elucidate only the initial stage of the larger contentious process that may or may not escalate to civil warfare. Exploration of the mechanisms that eventually lead to internal conflict cannot be satisfactorily achieved with these frameworks.

An example that illustrates these problems is the civil war in Nagorno-Karabakh, an Armenian-controlled enclave of Azerbaijan that has functioned independently of Baku since 1992. The Nagorno-Karabakh conflict broke out in the midst of a cauldron of domestic instability precipitated by the collapse of the Soviet Union and the mobilization of rival ethnic communities (King, 2009). Right after declaring independence in December 1991, the newly formed Azerbaijani state suffered from mutually reinforcing crises of authority and legitimacy. In the chaos that followed the unravelling of the Soviet Union, Azerbaijan emerged as a weak new state that, in the initial stage, scrambled to amass a military that would be effectively deployed to enforce a violence monopoly over its sovereign territory. This happened at a nebulous time when the central government’s authority was challenged by Armenian irregulars who had already begun to carve out areas of exclusive rule in Nagorno-Karabakh and its surroundings. At the moment of its (re)birth in 1991,5 Azerbaijan also suffered from a severe legitimacy deficit: it was a poor country in transition with inchoate political institutions which were poorly equipped to absorb dissent or offer credible guarantees for protecting the rights of the territorially concentrated Armenian minority (Toft, 2003).

Some may argue that civil war outbreak in Azerbaijan in the early 1990s falls squarely in line with prevailing theoretical expectations in “correlates of civil war” studies. After all, in this body of literature, factors that seem to have contributed to civil war onset in Nagorno-Karabakh, new state embroiled in regime transition, poor and weak country, concentrated minority that lacked credible security guarantees, mountainous terrain that facilitates rebellion, feature prominently as robust predictors of conflict outbreak (Fearon & Laitin, 2003). Nonetheless, one should avoid falling into the trap of post-hoc rationalizations. The civil war outcome in Nagorno-Karabakh was not foreordained, but it originated in an environment of acute authority and legitimacy crisis wherein processes of radicalization and militarization rapidly escalated into violence. Civil war did indeed break out in the background of a legitimacy and authority crisis, but only after the opponents radicalized their demands to the point where compromise became unrealistic and a security dilemma led to a spiral of militarization.

In the absence of radicalization and militarization, crises of authority and/or legitimacy may not generate civil warfare; instead, they may produce different types of outcomes along the contentious cycle. For example, strikingly similar conditions to the ones in Azerbaijan were present in Sri Lanka when the country gained independence in 1948: newly formed, poor, weak, transitional state dominated by a Sinhala majority that could not credibly commit to offer legal protections to the Tamil minority concentrated in the northeast of the island; discriminatory policies against minorities; rough, jungle terrain that favored insurgency. Yet, civil war did not immediately break out. Instead, in the Sri Lankan case, we observe various types of outcomes along the contentious cycle that began with a crisis of authority and legitimacy upon the country’s independence in 1948 and culminated in civil war onset toward the end of 1983: exclusion of Tamils from the civil service, academic, and military sectors; non-violent protests and acts of civil disobedience organized by Tamil groups; violent repression by government forces against Tamil protestors (Ramaswamy, 2007). Civil war onset occurred only in the context of radicalization and militarization that had engulfed the country by early 1980s. Beyond the examples of Azerbaijan and Sri Lanka, it is abundantly clear that a crisis of authority and/or legitimacy is a necessary, but certainly not sufficient, condition for observing internal conflict onset. In the contemporary period (post-1945), many countries have suffered from some kind of authority or legitimacy crisis; however, civil warfare has been a relatively rare occurrence. This understanding of radicalization and militarization as key mechanisms leading to civil war onset lies close to della Porta’s conceptualization of mechanisms as “chains of interaction that filter structural conditions and produce effects” (della Porta, 2013, p. 24). Therefore, a mechanism-based approach to internal conflict outbreak gives attention to both structure and agency. Civil war is an emergent phenomenon in a process where actors’ agency constantly interacts with structural characteristics of the environment in which they operate. To elucidate civil war onset, it is not sufficient to explain when a crisis of authority and/or legitimacy is most likely to occur; it is also necessary to examine the ways in which radicalization and militarization combine to produce an escalatory spiral that results in organized collective violence between the government and insurgents or among rebels themselves.

State-level “correlates of civil war” approaches can elucidate how collective action obstacles to social mobilization can be overcome and why crises of state authority and/or legitimacy may occur, but they have a hard time explaining when we are most likely to observe radicalization and militarization producing civil warfare. In their turn, by capturing the strategic interaction between actors at the dyadic level, bargaining frameworks can integrate more effectively those factors (e.g., divisions within the government and/or the rebel movement; external military support for the state or the rebellion; third-party interventions) that are typically associated with the radicalization and militarization mechanisms. However, bargaining theories do not clearly specify whether radicalization and militarization occur because of informational asymmetries, commitment problems, or issue indivisibilities, or whether radicalization and militarization create or exacerbate informational asymmetries, problems of commitment (Jenne, 2007), and issue indivisibilities (Goddard, 2009; Hassner, 2009). Thus, these frameworks remain theoretically opaque: their underlying logic produces empirically indeterminate expectations. Particularly problematic remains their inability to specify why different kinds of bargaining failures produce different types of violent or nonviolent contentious outcomes or why similar kinds of bargaining failures generate various forms of violent or nonviolent contentious outcomes within a given cycle of contention.

The focus on radicalization and militarization within mechanism-centered frameworks allows us to see how actor interactions transition from nonviolence to violence or from low levels of violence to high levels of violence. Nonetheless, this analytical task is complicated by conventional conceptualizations of civil war in the empirical literature which may obscure or conflate various forms of violent contention with internal conflict (Florea, 2012; Lake, 2003). Current empirical studies typically embrace the UCDP/PRIO definition of conflict as a “contested incompatibility that concerns government and/or territory where the use of armed force between two parties of which at least one is the government of a state results in at least 25 battle-related deaths” (Melander, Pettersson, & Themner, 2016).6 The casualty-based conceptualization is methodologically appealing because it provides the consistency needed for large-N analyses (Canestaro, 2016). However, the rationale for including arbitrary violence thresholds as observable indicators of war quickly withers away if civil war onset is examined within a broader contentious cycle that displays diverse manifestations of violent or nonviolent collective action. The casualty-based threshold raises important theoretical and analytical concerns. Was violence targeting minority Armenians in Sumgait (February 1988) and Baku (January 1990) a pogrom, civil war, or something else? What distinguishes civil warfare from other forms of domestic organized violence that produce 25+ or 1000+ “battle-related deaths”? The casualty-based conceptualization strips away the dynamic nature of civil conflict because it prevents us from empirically capturing the wide variation in the types of violent and nonviolent forms of collective action within a given contentious cycle/spell. If the casualty threshold is retained at the most reliable criterion by which civil wars are identified, it would be quite difficult to accurately track the key escalatory mechanisms of radicalization and militarization along a contentious cycle/spell because various forms of violent collective action may be artificially bundled under the general category of civil warfare where, in fact, they may capture violent phenomena other than civil war. Therefore, violence and civil war need to be analytically decoupled: civil warfare is best conceptualized independently of the level of violence (Florea, 2012; Kalyvas, 2006).

Embracing a contentious politics, mechanism-based approach to internal conflict onset yields several theoretical and methodological payoffs. Theoretically, such a perspective affords a better understanding of the steps or processes that might lead to civil war outbreak. A contentious view on civil war outbreak allows the analyst to unpack the pathways through which less intense forms of domestic strife might escalate to internal conflict and to ascertain the variation in the types of violent and nonviolent outcomes resulting from actors’ interactions. As suggested herein, a common pathway through which countries become embroiled in civil warfare involves the existence of a crisis of authority and/or legitimacy as an underlying condition and two subsequent, interconnected mechanisms, actor radicalization and the militarization of the dispute.7 This combination creates a dynamic that is most likely to bring about internal conflict. Methodologically, embedding civil war in a larger repertoire of domestic contention entails adopting the contentious cycle/spell as the unit of analysis. This has important implications for data collection practices because theories of civil war onset, duration, and termination need to be tested on the whole range of nonviolent and violent forms of collective action observed within a given contentious cycle/spell.8 While more labor intensive, this alternative approach to civil war data collection and analysis will likely alleviate concerns about unit homogeneity that plague existing practices (e.g., current civil war codings, especially those used in the UCDP/PRIO dataset, may actually capture forms of violent collective action that are different from civil warfare).

Conclusion

Civil war is not an independent event that occurs in isolation but an outcome of a process of political contention that commonly displays various forms of violent and nonviolent collective action. To gain a fuller understanding of how civil warfare may grow out of the interaction between a government and an armed nonstate challenger or from the interaction between two or more armed nonstate challengers, it needs to be empirically investigated within a contentious cycle/spell. One common pathway that leads to civil onset involves the presence of a crisis of state authority and/or legitimacy (an underlying or structural cause of internal conflict outbreak) and the interplay of two mechanisms (proximate causes of civil war onset), actor radicalization and militarization of the political contest, which may be visible at different stages in the contentious cycle/spell. Examining internal conflict outbreak within a contentious ecosystem allows the analyst to trace how civil war may emerge from the escalation of militarized or nonmilitarized interactions between domestic adversaries. By viewing civil war as part of a larger process and not as a one-shot event, scholars can better trace the range of actors’ behavioral repertoires in a given contentious cycle/spell.

A contentious approach to civil warfare integrates important insights from the literature on social movements and political mobilization, provides a more realistic picture of how internal conflict breaks out, and addresses important limitations in the correlational and bargaining studies that have dominated existing empirical research. Such an approach allows the investigator to focus on the process that produces civil war, not just on the event itself. At the same time, placing civil warfare within a contentious framework requires substantial rethinking of our theories and methodologies.9 Theoretically, the precise steps through which the mechanisms of radicalization and militarization interact to produce civil warfare, as opposed to a different form of political contention, still need to be elucidated. Methodologically, the next task for scholars of internal conflict and violence is to look at the cycle of contention as the unit of analysis and explore empirically the strategic and nonstrategic factors that influence the violent or nonviolent outcomes observed at each stage in the cycle.

References

Bennett, A. (2013). Causal mechanisms and typological theories in the study of civil conflict. In T. J. Checkel (Ed.), Transnational dynamics of civil war (pp. 205–230). New York: Cambridge University Press.Find this resource:

Blattman, C., & Miguel, E. (2010). Civil war. Journal of Economic Literature, 48, 3–57.Find this resource:

Canestaro, N. (2016). Towards a practitioner-centric definition of civil war. Civil Wars, 18, 359–377.Find this resource:

Checkel, J. T. (Ed.). (2013). Transnational dynamics of civil war. New York: Cambridge University Press.Find this resource:

Chenoweth, E., & Stephan, M. J. (2011). Why civil resistance works: The strategic logic of nonviolent conflict. New York: Columbia University Press.Find this resource:

Coe, A. J. (2012). Economic origins of war and peace (Unpublished doctoral dissertation). Harvard University, Cambridge, MA.Find this resource:

Collier, P., & Hoeffler, A. (2004). Greed and grievance in civil war. Oxford Economic Papers, 56, 563–595.Find this resource:

Collier, P., Hoeffler, A., & Roehner, D. (2009). Beyond greed and grievance: Feasibility and civil war. Oxford Economic Papers, 61, 1–27.Find this resource:

Cunningham, D. E. (2011). Barriers to peace in civil war. New York: Cambridge University Press.Find this resource:

Cunningham, D. E., & Lemke, D. (2014). Beyond civil wars: A quantitative examination of causes of violence within countries. Civil Wars, 16, 328–345.Find this resource:

della Porta, D. (2013). Clandestine political violence. New York: Cambridge University Press.Find this resource:

Demirel-Pegg, T. (2014). From the streets to the mountains: the dynamics of transition from a protest wave to an insurgency in Kashmir. Mobilization, 19, 309–327.Find this resource:

DeVotta, N. (2004). Blowback: Linguistic nationalism, institutional decay, and ethnic conflict in Sri Lanka. Stanford, CA: Stanford University Press.Find this resource:

De Waal, T. (2003). Black garden: Armenia and Azerbaijan through peace and war. New York: New York University Press.Find this resource:

Dixon, J. (2009). What causes civil war? Integrating quantitative research findings. International Studies Review, 11, 707–735.Find this resource:

Falleti, T. G., & Lynch, J. F. (2009). Context and causal mechanisms in political analysis. Comparative Political Studies, 42, 1143–1166.Find this resource:

Fearon, J. D. (1995). Rationalist explanations for war. International organization, 49, 379–414.Find this resource:

Fearon, J. D. (2013). Fighting rather than bargaining (Unpublished manuscript). Stanford, CA: Stanford University Press.Find this resource:

Fearon, J. D., & Laitin, D. (2003). Ethnicity, insurgency, and civil war. American Political Science Review, 97, 75–90.Find this resource:

Florea, A. (2012). Where do we go from here? Conceptual, theoretical, and methodological gaps in the large-N civil war research program. International Studies Review, 14, 78–98.Find this resource:

Gerring, J. (2007). Is there a (viable) crucial-case method? Comparative Political Studies, 40, 231–253.Find this resource:

Gerring, J. (2010). Causal mechanisms: Yes, but . . . Comparative Political Studies, 43, 1499–1526.Find this resource:

Gibler, D. (forthcoming). Combining behavioral and structural predictors of violent civil conflict: Getting scholars and policymakers to talk to each other. Forthcoming in International Studies Quarterly.Find this resource:

Goddard, S. E. (2009). Indivisible territory and the politics of legitimacy. New York: Cambridge University Press.Find this resource:

Granovetter, M. (1978). Threshold models of collective behavior. American Journal of Sociology, 83, 1420–1443.Find this resource:

Gurr, T. R. (1970). Why men rebel. Princeton, NJ: Princeton University Press.Find this resource:

Hassner, R. E. (2009). War on sacred grounds. Ithaca, NY: Cornell University Press.Find this resource:

Hedström, P., & Ylikoski, P. (2010). Social mechanisms in the social sciences. Annual Review of Sociology, 36, 49–67.Find this resource:

Hironaka, A. (2005). Neverending wars: The international community, weak states, and the perpetuation of civil war. Cambridge, MA: Harvard University Press.Find this resource:

Hirschman, A. O. (1970). Exit, voice, and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard University Press.Find this resource:

Jenne, E. K. (2007). Ethnic bargaining: The paradox of minority empowerment. Ithaca, NY: Cornell University Press.Find this resource:

Kalyvas, S. N. (2006). The logic of violence in civil war. New York: Cambridge University Press.Find this resource:

Kalyvas, S. N. (2007). Civil wars. In C. Boix, & S. Stokes (Eds.). The Oxford handbook of comparative politics (pp. 416–434). New York: Oxford University Press.Find this resource:

King, C. (2009). The ghost of freedom: A history of the Caucasus. New York: Oxford University Press.Find this resource:

Kuran, T. (1989). Sparks and prairie fires: A theory of unanticipated political revolution. Public Choice, 61, 41–74.Find this resource:

Lake, D. A. (2003). International relations theory and internal conflict: Insights from the interstices. International Studies Review, 5, 81–89.Find this resource:

Levy, J. S., & Thompson, W. R. (2011). The arc of war: Origins, escalation, and transformation. Chicago: University of Chicago Press.Find this resource:

McAdam, D., Tarrow, S., & Tilly, C. (2001). Dynamics of contention. New York: Cambridge University Press.Find this resource:

Melander, E., Pettersson, T., & Themner, L. (2016). Organized violence, 1989–2015. Journal of Peace Research, 53, 727–742.Find this resource:

Meyer, D. S., & Minkoff, D. C. (2004). Conceptualizing political opportunity. Social Forces, 82, 1457–1492.Find this resource:

Olson, M. (1965). The logic of collective action. Cambridge, MA: Harvard University Press.Find this resource:

Powell, R. (2002). Bargaining theory and international conflict. Annual Review of Political Science, 5, 1–30.Find this resource:

Ramaswamy, V. (2007). Historical dictionary of the tamils. Lanham, MD: Scarecrow.Find this resource:

Rasler, K. (1996). Concessions, repression, and political protest in the Iranian revolution. American Sociological Review, 61, 132–152.Find this resource:

Reiter, D. (2003). Exploring the bargaining model of war. Perspectives on Politics, 1, 27–43.Find this resource:

Ron, J. (2005). Paradigm in distress? Primary commodities and civil war. Journal of Conflict Resolution, 49, 443–450.Find this resource:

Rule, J. B. (1988). Theories of Civil Violence. Berkeley: University of California Press.Find this resource:

Sambanis, N. (2004). What is civil war? Conceptual and empirical complexities of an operational definition. Journal of Conflict Resolution, 48, 814–858.Find this resource:

Senese, P. D., & Vasquez, J. A. (2007). The steps to war: An empirical analysis. Princeton, NJ: Princeton University Press.Find this resource:

Tarrow, S. (2007). Inside insurgencies. Politics and violence in an age of civil war. Perspectives on Politics, 5, 587–600.Find this resource:

Tarrow, S. (2011). Power in movement: Social movements and contentious politics (3d ed.) New York: Cambridge University Press.Find this resource:

Tarrow, S. (2012). Strangers at the gates: Movements and states in contentious politics. New York: Cambridge University Press.Find this resource:

Tilly, C. (1978). From mobilization to revolution. Reading, PA: Addison-Wesley.Find this resource:

Tilly, C. (2001). Mechanisms in political processes. Annual Review of Political Science, 4, 21–41.Find this resource:

Tilly, C., & Tarrow, S. (2007). Contentious politics. Boulder, CO: Paradigm.Find this resource:

Toft, M. D. (2003). The geography of ethnic violence: Identity, interests, and the indivisibility of territory. Princeton, NJ: Princeton University Press.Find this resource:

Vinci, A. (2006). Greed-grievance reconsidered: The role of power and survival in the motivation of armed groups. Civil Wars, 8, 25–45.Find this resource:

Walter, B. F. (1997). The critical barrier to civil war settlement. International Organization, 51, 335–364.Find this resource:

Walter, B. F. (2002). Committing to peace: The successful settlement of civil war. Princeton, NJ: Princeton University Press.Find this resource:

Walter, B. F. (2009). Bargaining failures and civil war. Annual Review of Political Science, 12, 243–261.Find this resource:

Wood, E. J. (2008). The social processes of civil war: The wartime transformation of social networks. Annual Review of Political Science, 11, 539–561.Find this resource:

Young, J. K. (2013). Repression, dissent, and the onset of civil war. Political Research Quarterly, 66, 516–532.Find this resource:

Notes:

(1.) For such reviews, see, inter alia, Blattman and Miguel (2010), Dixon (2009), Sambanis (2004), and Walter (2009).

(2.) Tarrow (2011, p. 207) holds that radicalization marks a shift in social movement organizations towards “the extremes and/or the adoption of more disruptive forms of contention.” Wood (2008, p. 550) conceptualizes militarization as “the supplanting of local forms of governance with new forms that reflect the influence of armed actors.”

(3.) Obviously, this process is probabilistic, not deterministic. Internal conflicts may emerge through alternative pathways; however, the contentious cycle described here captures the event sequence that is most likely to produce civil war outbreak.

(4.) As Gibler aptly suggests, these factors routinely associated with civil conflict “may be a product of structural conditions and not at all associated with the conflicts themselves” (Gibler, forthcoming, p. 1).

(5.) Azerbaijan was an independent country from 1918 until 1920 when it was incorporated into the Soviet Union.

(6.) The Correlates of War (COW) dataset requires a much larger casualty threshold to code for the presence of war: at least 1,000 battle-related deaths per year.

(7.) This processual view of civil war onset bears some resemblance to the “steps-to-war” theory in which territorial issues are the underlying causes of interstate war, and practices of power politics, such as militarization through alliance making and military buildup, are its proximate causes (Senese & Vasquez, 2007).

(8.) Cases currently coded as civil war recurrence may actually capture the variation in violence within the same contentious cycle (Florea, 2012; Hironaka, 2005).

(9.) One recent study that embraces a process-based view of civil war is Young (2013). Young views internal conflict onset as an outcome of an escalatory process which sees high levels of state repression and opposition dissent.