Making Sense I 43
Making Sense
Conflict Theory and the Caucasus
The search for a single factor or a set of factors that explains everything is comparable to the search for the Holy Grail—noble but futile.1
In this book conflict is understood as a process in which two or more parties attempt to pursue their interests, which are perceived as mutually incompatible, by directly or indirectly seeking to reduce the other party's capacity to achieve its goals. Conflict is thus competition over resources and symbols, conducted in the framework of institutionalized procedures. Conflict is therefore the normal state of affairs in society. By contrast, violent conflict is understood as a process in which two or more parties attempt to pursue what they perceive as incompatible interests by trying to seize or destroy the property of the other party and/or to injure or eliminate the other party with the help of organized violence.2 As a rule, violent conflicts are conducted outside all those institutions that a society provides and practices for dealing with conflict. From such a definition, the key question that conflict research needs to address is when and how conflict spins out of its institutional embed-dedness and escalates into violence.
The emergence of violence is often perceived as a chaotic, irrational eruption. Wars "break out," violence "erupts," and social order "collapses." Such perceptions aptly capture the feelings of witnesses or victims of organized violence. However, again contrary to everyday perceptions, it is deeply misleading to depict violent conflicts as irrational, disordered, and chaotic processes. Undoubtedly, violent conflicts cause horrific devastation. Nevertheless, the normative rejection of destructive violence is no substitute for the analysis of the rational strategic calcula-
tions of those people organizing violence, investing in violence, and hence sustaining violence.
Social scientists cannot describe, let alone explain, events unless they are equipped with sufficient theory. Telling the stories of war and peace thus inevitably opens up the debate on the appropriate theoretical approaches that may be applied in order to collect, order, and interpret the "material" of which these stories are made. Behind these approaches are different and often competing epistemological positions that inform how knowledge of the observed social processes are gained and verified.
In this chapter, the specific analytical tools applied to the case studies in chapters 4 to 7 are discussed against the backdrop of recent conflict theories. The chapter starts with an overview of the findings in recent quantitative literature. This strand of research aims to identify those general factors that increase a society's propensity for violence in general, using a sample of most or all internal wars after 1945. Among the most prominent factors seen as increasing the risk for war are state weakness, low level of development, existence of resources for financing rebellion, complex patterns of ethnic geography, and terrain favoring rebellion. Next, these global propositions are complemented by a closer look at the microlevel of organized violence—by treating the organization of violence as a problem of collective action and identifying those fuctors and processes that are necessary for organizers of violence to overcome the many problems associated with collective, sustained violence. Finally, linking the two and expanding the explanatory power of both, an institutionalise approach to conflict research is introduced.
Risks for War: Global Propositions
Since the late 1990s, a number of quantitative comparative studies have sought to systematically identify those factors that make societies in general more prone to violence. In contrast to qualitative approaches, which are better suited to providing idiosyncratic, highly specific explanations for a small number of cases, these studies seek to uncover the ^cneralizable factors that have an effect on the likelihood of war and which can be systematically traced across many cases. While sociological definitions of conflict and violence primarily focus on the processes 1 Iwt lead to the escalation of conflict, the quantitative strand of research literature focuses on structural factors that account for higher risks of
A 1
44 I Making Sense
conflict. Typically, such studies are based on a sample that encompasses all civil wars from 1945. The dependent variable is usually the occurrence or nonoccurrence of "internal conflict" (civil war, insurgency, internal war, or rebellion). These studies then test the effect of various independent variables such as "level of development" "regime type," "ethnic and religious fractionalization," and "mountainous terrain" on the dependent variable: that is, on the occurrence (or the length or the scale) of internal conflicts.
Since these approaches rely on statistical analysis for inferring causality, they require that their objects of interest—the war and the factors causing the war—are defined and "packaged" in such a way that it becomes amenable to statistical measurement and comparison.3 Consequently, in terms of these studies, societies are at war or they are not, it is an ethnic war or it is not, the parties to the war are the government or they are not (in which case they are the rebels), the rebels are motivated by greed or they are not, and so on. This reductionism defies the tradition of sociological writing on conflict, with its emphasis on the dynamics of conflict; the shifting and blurring of boundaries; and the social construction of identities, motivations, and patterns of interaction as a result of conflict, not as its precondition, or independent variable. Furthermore, statistics may identify associations between a given factor and an outcome, but they usually do not illuminate the causal mechanism that explains this association. For example, one of the best established findings of quantitative conflict research is that a low GDP (gross domestic product) is associated with an increased risk of civil war, but which mechanisms actually cause this outcome are still hotly disputed. Despite these limitations, the general propositions that quantitative research have to offer provide a useful framework for comparison and are thus a rewarding starting point for this investigation.
The findings from recent quantitative literature point to six factors that increase a society's risk for internal war: a low level of economic development, a sudden loss of administrative capacities, opportunities for financing rebellions, recent previous wars, a complex ethnic geography, and terrain that favors rebellions. In the following sections, these findings are summarized, and their relevance and applicability in the context of the Caucasus are briefly discussed.4
Making Sense I 45
Level of Economic Development
There is broad consensus that a society's level of economic development (usually measured by growth of per capita income, or by per capita income, or by average energy consumption) considerably influences the risk of internal conflict. Very poor societies are more prone than less poor societies to internal violence. Wealthy societies, by contrast, appear immune to organized violence. Collier et al. show that the level of per capita income, growth rate, and the structure of the economy—namely, dependence on primary commodity exports—have a significant effect on the risk of internal war. They estimate that doubling per capita income approximately halves the risk of rebellion.5 Similarly, Fearon and Laitin showT that, on average, reducing per capita income by $1,000 is associated with 341 percent increase in the annual odds for the onset of civil war.6 Hegre et al. found that in states moving from a very low to a marginally higher level of development, the risk of conflict also increases. If the level of development surpasses a certain threshold, roughly equivalent to the level of development found in Bhutan or Haiti in 1990, then the risk of conflict falls. Highly developed states are around eight times less at risk of conflict than are states with the highest risk level of development.7 This might explain why, out of 117 internal wars since 1945, some 34 have taken place in sub-Saharan Africa and 3 3 in Asia, but only two occurred in the West.
Although there is much evidence that countries with low levels of economic development are at a typically higher risk of internal wars, it is not clear whether there is an actual causal link between a low level of economic development and civil wars; moreover, if there were such causality, the exact mechanism by which it works is not clear. Different studies have interpreted the correlation between a low level of economic development and a high propensity for civil wars in different ways. For example, Fearon and Laitin assume that the risk for war depends crucially on the government's capacity for effectively policing rural areas and combating insurgencies; hence, they interpret a low level of GDP as a proxy for the government's capacity to organize counterinsurgency.8 Other authors see low levels of GDP as a sign of economic hardship, which lowers recruitment costs for rebels, or as an indicator that poverty may spark violence. NStably, these different interpretations, which iire based on the same statistical findings, correspond to a quite different motivational structure for rebels. The first two interpretations see
1
4 6 \ Making Sense
rebellion emerging when there is the opportunity to do so; the latter sees rebellion as being motivated by grievances. Both interpretations are consistent with statistical analysis of the data.
One of the benefits of such broad propositions is that it allows different regions to be put in global perspective. However, by global standards the Soviet and post-Soviet Caucasus probably does not qualify as a region marked by low levels of development. Even acknowledging the massive economic problems with which the Soviet Union was struggling in the last decade of its existence, the Caucasus of the 1980s must still be seen as a relatively well resourced region in comparison with other regions of the world. In 1990, the GDP in Georgia was approximately five times higher than the average of countries in the world's poorest quartile, even though the economy was stagnating during much of the decade and even shrinking toward the end (Table 3.1). In addition, the slow growth of the official economy in the early and mid 1980s was partly compensated for by the blossoming of the informal "shadow" economy. According to both anecdotal evidence and Soviet public perception, the shadow economy reached legendary levels in the Caucasus. The most conservative estimates assume that the size of the shadow economy was the equivalent of at least 30 percent of the official GDP in the last years of the Soviet Union (Table 3.2).
Real GDP is generally perceived as being the single most important indicator for level of development, but even if other measures of development are applied, it would be difficult to label the Soviet Caucasus as an underdeveloped region. While the benefits that the Soviet Union had to offer to its citizens considerably lagged behind what it had promised and could not compete with the West, its achievements were still impressive. Health care was widely available and free at the point of service; enrollment in secondary schooling was almost universal, because
table 3.1
Development of GDP Per Capita, Constant 2000 $US .
	19 So	rSSy	'99°		2000	ioo;
Armenia	u.a.	n.a.	7,940	461	620	1,12S
Azerbaijan	11.a.	u.a.	1,250	488	655	1,182
Georgia	1,670	1,986	1,493	458	648	971
Low and middle income (world)	864	899	961	1,035	1,190	1,435
Low income (world)	254	271	310	338	390	476
Source-. WDf Online: World Development Indicators, online database (World Bank) n,a,, not available.
Making Sense I 47
TABLE 3.2
Size of the Shadow Economy Relative to the GDP, Minimum Estimation (%)
	'99°	(average)
Armenia	31	83
Azerbaijan	28	30
Georgia	33	86
Source: Yair Eilat and Clifford Zinnes (zoai), "The Shadow Economy in Transition Countries: Friend or Foe? A Policy Perspective," World Development 50(7], pp. 12.33-1154.
in the Soviet Union secondary school education was, with a few exceptions, free and mandatory; and a social safety net for pensioners was in place. While there was hidden unemployment, particularly in such economically backward regions as Dagestan, Nagorny-Karabakh, and South Ossetia, it was only with the violent escalation of conflict that unemployment really became a mass phenomenon. In sum, the Soviet Caucasus was by no means a region that suffered from a low level of development. Explanations that draw on atypical low absolute levels of development thus cannot easily be applied.
Likewise, Donald Horowitz's argument that economically backward, peripheral ethnic groups are more likely to rebel than groups that are relatively economically advanced cannot easily be applied.3 Naturally, there were differences between the union republics in the level of economic development, with the Baltic republics being at the upper end and the Central Asian republics at the lower end. The differences between the Caucasian republics, however, were rather small. Moreover^ the Soviet system put a high premium on equalization, and the Soviet redistribution system ensured that the differences in living standards across the population were remarkably small, especially in comparison with the dramatic gaps between social strata that have emerged all over the post-Soviet space.
To be sure, secessionist groups did refer to economic discrimination in order to legitimize their secessionist claims. For example, the Karabakh Armenians and the Abkhaz, in particular, complained that they were economically disadvantaged by the unloved republican centers. It is difficult to find substantial proof for these claims, however. Abkhazia was i\ rather affluent region within Georgia, mainly because of its privileged position as an all-Soviet kurort (tourist resort) and its flourishing i-xporr of tea and citrus fruits. Karabakh was certainly an economically
48  I  Making Sense
backward region compared with urban centers such as Baku and Yerevan; according to economic indicators, however, it was more prosperous than most other regions in Azerbaijan. The perceived economic neglect was therefore more informed by what Karabakh could have been within Armenia compared with what it was within Azerbaijan.
Other indicators such as urbanization and education also show relatively little variance across groups within the Caucasus. But even if the North Caucasus, as a largely rural and little industrialized region, could have qualified as economically backward in relation to the central regions (mainly urban centers in Soviet Russia), this does not explain why only one group, the Chechens, embarked on organized violence, while other groups in the North Caucasus did not. Moreover, it also fails to explain why the Abkhaz, arguably a peripheral but a relatively prosperous (thanks to the thriving shadow economy) group in Georgia, opted for secession, whereas the poorer Ajars did not. Hence, neither an absolute low level of economic development nor a relatively low level of development (compared with other groups) is a factor that offers much explanatory power in the Caucasus. While expectations and fears over future economic development played a prominent role on the agenda of political actors and followers in the Caucasus, the economic situation, as it was, does not lend support to arguments that stress either economic discrimination or a general low level of development as explanations for internal wars.
Regime: Strength, Type, and Change
States are at risk when they are newly independent or when they have recently suffered political instability at the center. Fearon and Lai-tin estimated that the odds of a civil war starting are five times greater in the first two years of a state's independence than in other years. Political instability at the center also dramatically increases the risk of war. According to the same authors, a country that has suffered from political instability has a 67 percent greater chance of internal violence in the following three years than a country that did not suffer political instability.10 Likewise, as noted above, there is a broad consensus that a low level of GDP per capita, which is interpreted as being an indicator of low administrative capacities, has a significant influence on the risk of violent conflict.
Making Sense I 49
States that have undergone a rapid change from autocracy to democracy, or vice versa, are also at risk. According to Hegre et al., on average, regimes are 3.55 times more at risk of civil war on the day after a change of regime. This effect fades over time: one year after states have undergone a regime change, they face only a 1.89 times greater risk; after six years, they are no more likely to suffer violence than states that have not undergone regime change. Interestingly, the direction of regime change does not play a significant role—moves toward both autocracy and democracy are risky.11 Contrary to commonly held assumptions, democratic states do not seem to be better protected than authoritarian states from the risks of internal wars. Collier and Hoeffler found no evidence that more democratic regimes were less vulnerable to organized violence than less democratic regimes, and they take this result as further evidence that the grievance script provides less explanatory power than the opportunity script. Fearon and Laitin reached similar conclusions, finding that the onset of civil war is no less frequent in democracies than in nondemocratic regimes, once they controlled for level of income.12 However, they found clear evidence that regimes that are neither fully democratic nor fully authoritarian suffered from a markedly greater risk of internal conflict.13 The same finding was reported by Hegre et al., who determined that authoritarian regimes and institutionally consolidated democracies are far less vulnerable to conflict than mixed systems or transitional regimes are. Very authoritarian and very democratic regimes display a similarly low conflict risk, while intermediate regimes are four times more susceptible to conflict. Semidemocra-tic and semiauthoritarian regimes combine a limited amount of public freedom with a limited amount of repression. This mixture can encour-nge conflict because the permitted public freedom makes the organization of resistance and the expression of protest against repressive politics possible.14
Instability and regime change certainly epitomize political developments in the Caucasus. All states in the region experienced political instability at the center, embarked on a rapid change from autocracy to democracy, and turned into weak transitional regimes en route. However, the Caucasian paths from political instability to violence show some peculiarities. The rapid devaluation of administrative capacities went hand in hand with national mobilization and a widening of the political space (liberalization). This conjunction of national mobiliza-
50 I Making Sense
tion, liberalization, and diminishing administrative capacities opened up various avenues that led to organized violence. As the Soviet state became weaker, it could no longer guarantee that competition between elites took place in line with institutionalized procedures, since it could no longer punish competitors who did not abide by the rules. Elite competition, as with most other aspects of politics, became "disembedded," as it was no longer played by the rules. In Chechnya, Azerbaijan, and, to some extent, Karabakh, nationalist counterelites, backed by mass support in the streets, captured the state by force. Likewise, state weakness also opened up windows of opportunity for realizing secessionist aspirations. Local elites in Abkhazia, Ossetia, Karabakh, and Chechnya wanted to seize the political opportunity that the weakening of the central state offered and opted for secession. The secessionist drive of these groups was without doubt further pushed by the lack of capacity and of political will of the newly ruling nationalist elites in Russia, Georgia, and Azerbaijan to offer credible security guarantees. Hence, it was not only the window of opportunity that triggered secessionist policies but also, especially in the case of Abkhazia and Karabakh, the lack of credible commitment of the new states to the protection of their national minorities. The structural weakness that prevents states from offering credible guarantees has been referred to as the commitment problem, and it certainly had an important role in the conflicts between Georgia and Abkhazia and between Azerbaijan and Karabakh.15
Financing the Rebellion
With the economic turn in conflict research, the financial viability of organized violence has come to the forefront, and researchers have sought to identify the important sources of income that finance rebellions. Potential sources are "lootable" natural resources such as oil, gems, and coca or opium poppy, as well as the taxing of illicit economic activities by rebels (for example, drug smuggling) and donations from a foreign patron or a wealthy diaspora. Both qualitative and quantitative studies suggest that the production of oil is associated with the onset of internal wars, particularly separatist conflicts.1* With Collier and Hoef-fler taking the lead, it is now broadly accepted that a high dependence on natural resources, primarily oil, makes a state more prone to internal violence.17 However, while the link between oil and the risk of violent conflict is well established, the causal mechanisms are predictably
Making Sense I 51
opaque.58 The existence of natural resources may provide an incentive for looting, it may motivate rebels to capture the state, or it may trigger secessionist movements that hope to gain control over the resources. A different line of thought, associated with the "resource curse," suggests that a wealth of natural resources often leads to distorted economic development, high levels of corruption, and poor governance, which may, in turn, fuel conflict more indirectly.19 Also, natural resources may foster rentier states with weak ties to their societies and weak administrative capacities.20 This may then lead to another factor that increases the tisk of war—namely, the lack of state capacities.
Evidence from these case studies suggest that such lootable commodities, while not causing wars, help to prolong wars because rebels reinvest the profits from selling these commodities in the organization of violence. For example, in Afghanistan, the trade in gemstones and opium was an important source of income for the Northern Alliance of Ahmed Shah Massoud, but it cannot be claimed that trading in gemstones or drugs led to the onset of the war.
Finally, there is also evidence, both from case studies and from statistical analyses, that a large diaspora increases the risk of internal violence.21 The donations and support from diaspora are often a crucial .source of income for rebel movements. Well-known examples are the Tamil Tigers and the Kosovo Liberation Army, both rebel movements that were able to obtain, often by criminal means, material support for the cause from the diaspora.
In the Caucasus, only Azerbaijan and Chechnya possess natural resources of any significance—in both cases, oil. The diaspora argument, by contrast, is of wider applicability, if the definition of diaspora is relaxed. In the strict sense, only the Armenians possess a significant diaspora, and it was certainly crucial in helping to finance the war efforts. When the meaning of diaspora is broadened to include such diverse phenomena as the Chechen community living in Soviet urban centers (primarily Moscow and St. Petersburg) or ethnic kin groups living in neighboring states (such as the Ossets in South Ossetia who supported the struggle of the Ossets in North Ossetia, and the many volunteers I'mm the North Caucasus who supported the Abkhazians' struggle iigainst the Georgians), the support of the diaspora was in all cases of considerable relevance for the financing of war, once it had broken out. There is little evidence that the support of a diaspora was instrumental 111 explaining the outbreak of war.
52 I Making Sense
Making Sense I 53
Previous Wars
War breeds war. A factor that makes a society significantly more prone to organized violence is previous organized violence. Societies that have experienced conflict in the last five years are more prone to slip back into violence. Statistically, the risk of renewed war is twice as high as the risk of new war,22 although this effect fades over time. After five years without war, the country has an "average" risk of. war. This observation allows for different interpretations. Renewed conflict could point to hatred and motives for revenge, accumulated and strengthened during the first war as a core trigger. Alternatively, the previous conflict could have further weakened state capacities by damaging the economy and sapping political institutions. This, in turn, would make a renewed rebellion cheapen since the government's counterinsurgent capacities would be reduced.
Another possible causal mechanism links renewed violence to the fact that in post-conflict societies, the organization of violence remains "cheap" because the logistics of war (weapons, command structures, networks, and maybe fear) are still widely available, and because widespread unemployment makes recruiting young males as fighters cheap.23 In the Caucasus, two episodes of renewed war may serve as a test case for this proposition. First, there is the series of subsequent wars within Georgia; second, there are the two Chechen wars. In the case studies, the processes and mechanisms by which the previous war contributed to renewed wars are traced in detail.
Ethnicity: Diversity and Inequality
Perhaps the most original contribution of large-w, quantitative crosscountry studies, running counter to widely held popular beliefs, is that ethnically or religiously diverse societies do not have a higher risk of violent conflicts than do more ethnically homogeneous societies.24 Collier and Hoeffler as well as Hegre et al. demonstrated that ethnically frac-tionalized societies are more secure than more homogenous societies. They hypothesized that in diverse societies, consisting of numerous ethnic groups, it is more difficult for organizers of violence to recruit along ethnic lines, simply because the pool for recruits is smaller.25 They also found, however, that this only occurs when the largest ethnic group is less than 45 percent of the total population. When the largest group
makes up 45 percent or more, the risk of violence significantly increases. The authors argue that this is because groups that form a relative or absolute majority are tempted to use their numerical dominance to discriminate against smaller groups, which, in turn, stirs discontent. "Ethnic dominance" is therefore seen as a major factor in increasing the risk of war.
Ethnicity per se is thus not a causal factor. However, once a society has become unstable, ethnicity takes a prominent role in shaping the dynamics of violence. In 51 percent of all internal wars since World War II, at least one party to the conflict recruited mainly along ethnic lines; in 18 percent of all wars, recruiting was also done at least partly along ethnic lines. Only roughly one-third (3 r percent) of all wars cannot be labeled ethnic wars.2* Furthermore, once tensions escalate into war, there is evidence to suggest that ethnic wars tend to be longer than wars not fought along ethnic lines.27
Last, economic and cultural discrimination against ethnic groups has been seen as the root cause of many conflicts.28 Ted Gurr and his research group (in the Minorities at Risk Project) indicated that economic discrimination against minorities constitutes a major risk factor.29 A number of cases, such as Peru, Colombia, Kosovo, and Macedonia, strongly support this assertion.30 Quantitative research has also tried to test the "worldwide" validity of the discrimination argument. Neither Collier and Hoeffler nor Fearon and Laitin found a significant correlation between inequality within countries (measured by the Gini coefficient) and the risk of war. However, neither study was equipped to measure income inequality between ethnic groups because the inequality within the population of a given country is what was measured. These measurements are therefore blind to possible inequalities between different ethnic groups.31 Consequently, the argument that inequalities between groups may breed violence cannot be discarded, although this does not imply that ethnicity itself is a cause of conflict. Instead, it implies that commonly held grievances within an ethnic group can motivate conflict. It also implies that ethnicity can facilitate mobilization and recruitment in the course of conflicts, and that an ethnic conflict is likely lo breed more mutual negative stereotypes, which makes a resolution more difficult to reach.
In the Caucasus, four cases3of war occurred in an ethnodemographic situation that would qualify as ethnic dominance. In Chechnya (73 percent), Karabakh (76 percent), South Ossetia (66 percent), and Abkhazia
54 I Making Sense
Making Sense I 55
(46 percent), the largest ethnic group clearly outnumbered other groups. In all cases, it was the titular group (Chechens in Chechnya, Karabakh Armenians in Karabakh, Ossets in South Ossetia, and Abkhaz in Abkhazia) that wanted to secede, thus initiating a downward spiral that led to violence. Interestingly, however, in was not necessarily the titular group that was the numerically dominant group. In Abkhazia, the titular group formed a clear minority with only 17 percent of the overall population. Any explanation of conflict processes in the Caucasus that draws on "ethnic dominance," therefore, has to explain how, in at least one case, the initiative for secession came from a minority group. Ethnic dominance as an explanation is also weakened by the converse cases. Ingushetia (70 percent}, Kabardino Balkaria (48 percent), North Ossetia (53 percent), and Ajaria (63 percent) would also qualify as areas of the Caucasus with ethnic dominance, but all have avoided violent conflict. Hence, while ethnic dominance may increase the risk of conflict, it does not always lead to conflict. Therefore the mechanisms that helped defuse tensions related to ethnic dominance are of key interest.
Since ethnic dominance points to an ethnodemographic situation which, in turn, is formed by existing borders and by the ascription of membership to a specific ethnic group, any explanation that is based on ethnic dominance must refer to how "ethnicity" was institutionalized in the Soviet Union. Soviet ethnofederalism provided certain ethnically defined groups with particular territories and vested these territories (the union republics, the autonomous republics, and the autonomous ob-lasts) with political institutions. These political institutions proved to be a powerful organizational resource that made mobilization easier along predetermined ethnic lines. The political institutions of the Soviet Union thus created a path dependency for splits along ethnic lines, which may or may not have occurred without this institutional framework. This then points to the fact that ethnicity per se is never an explanation for conflict; rather, the way ethnicity is institutionalized and how this institutionalization becomes contested in periods of rapid social change explains conflict. Seen in this way, it becomes clear that ethnicity is not an a priori given but a social construct that emerges through history, and that "ethnicity" is contingent on both the discourse of a group about its own identity and on the political institutions in which certain aspects of this ethnicity are enshrined. Both are subject to change over time. By contrast, quantitative research, with its high premium on simple catego-
ries, tends to take ethnicity as a given attribute to a group and is therefore guilty of implicitly perpetuating a primordial position.
In a brilliant and polemical article against the use of oversimplified typologies, Valery Tishkov warns that "the basic methodological weakness of such theories of conflict analysis lies in their vision of groups as collective bodies with needs and universal motivation—not as situations, feelings, or acts of speech."32 Adopting such a perspective allows us to reconcile the fact that many internal wars (all but one in the Caucasus, and almost two-thirds of all internal wars since 1945) have a clear ethnic dimension, while at the same time ethnicity per se does not seem to be a cause of conflict: most multiethnic societies are at peace, most ethnic minorities do not rebel, and ethnically diverse societies are no less stable than homogeneous societies. However, ethnic boundaries, while not a cause of conflict per se, become reinforced or even reinvented during conflict. The salience of ethnicity can therefore be the result of the "ethnicization" of conflict. Cultural difference becomes important in the course of conflicts, for it is the material from which the barriers between groups are built. Cultural difference is the quarry, the stone from which serves to build walls between groups. These stones, however, are continually being moved, and cultural difference is a product of this continual work at building barriers; it is, so to speak, a permanent building site.
Furthermore, cultural difference is relational—the perception of the one group depends on how the "other" group is perceived and is constructed accordingly, and this perception changes precisely under the impact of conflict and violence. The perception of cultural difference is more a consequence than a cause of conflict. If groups clash, they deepen existing cultural differences or even invent new ones. This engenders a narrowing of cultural possibilities. Conflict brings about the destruction of multiple possibilities of identification: the number of available fits is reduced, the flexibility of the remainder diminishes, and the perception of the other is strengthened. Once such a process starts, it matters little whether the differences perceived by the participants are deeply rooted in cultural everyday life or are superficial and newly invented. Group conflicts can equally well be organized on the narcissism of small differences as on a border between civilizations. The conflicts in the Caucasus provide evidence for how ethnicity was reconfigured and perceptions and relations between groups became subject to change, once groups engaged in conflictive competition.
56 I Making Sense
Making Sense I 57
Finally, ethnicity, which is, after all, embodied in cultural norms of how things are done, may be a proxy for available social capital. Ethnic ties can be used to grease the social relations that are essential for collective action and, consequently, for organized violence. Ethnic social capital, while by no means being the only or the most important social capital, makes recruitment and financing easier and helps in finding sponsors abroad. After all, diaspora support, which has been identified as a major source of funding for organized violence, works precisely through ethnically defined social capital.
Terrain
The statistical data indicate that mountainous terrain increases the risk of conflict and wars tend to be longer when they are fought in mountainous countries. This finding obviously resonates well with the suggestive image of the noble but fierce mountain peoples, a cliche that, sadly, still inspires writers to write books like Caucasus: Mountain Men and Holy Wars.33 The interpretation of this finding is less romantic, however. It is assumed that mountainous terrain affords the rebels a military advantage by providing shelter and hideouts and that mountains increase the government's costs of controlling the territory. Both may increase the opportunity for organizing violence. The connection between terrain and the risk for internal violence has not attracted much attention, perhaps because it is hard to imagine a viable policy prescription against the negative effects of mountains, at least outside the Soviet Union under Stalin. In 1944 Stalin deported many of the indigenous mountain peoples, and the Soviet Union invested considerable resources in resettling mountain dwellers in the plains.
In the Caucasus, "mountainous terrain" as a factor offers a treacherous parsimony. The fact that all violent conflicts in the Caucasus took place in mountainous regions seems to confirm that a high risk of conflict is associated with terrain that favors rebellion. However, there are good reasons for not putting too much trust in the simplicity of this argument. First, there is little to be gained from this observation since the Caucasus is a very mountainous region, but, as stated earlier, not all potential hotbeds have escalated to violent conflicts. So, while the mountainous terrain cannot be excluded as a factor in increasing the propensity for conflict in some cases, it is not a universal explanation. Second,
a closer look at the actual fighting indicates that the strong correlation between physical geography indicators and conflict must be interpreted with caution. The character of the terrain obviously did not matter much at the start of the struggle for power, as the conflicts were centered around the capitals and not in the mountains.
In the case of the Georgian wars, terrain only became important when the western province of Mingrelia became the key theater of violence. Terrain was also not particularly relevant in the South Ossetian war, which was fought primarily around the capital, Tskhinvali. In fact, high mountains were more of a problem for the rebels since the vital connection with North Ossetia was blocked during winter. Even the war in Abkhazia was not influenced that much by the mountains and forests since it was fought primarily in the narrow corridor along the coast, with very little guerrilla activity. Likewise, in Karabakh, terrain initially favored not the rebels but the Azerbaijanis who had occupied strategic positions in the mountains overlooking Karabakh's capital of Stepanakert. The war in Chechnya followed a similar pattern. Initially, the conflict centered on the capital Grozny and a few other urban centers in the plains of Northern Chechnya. The mountains in the South of Chechnya served as a safe haven for regrouping and logistical support. Terrain grew more important when the Russian army started to control the plains and the rebels were pushed to their strongholds in the mountains. Hence, while it is plausible that terrain has affected the duration and, above all, the tactics of the rebels, there are no simple linkages between terrain and the onset of war.
From Risk to War: Organization of Violence
Having risk factors does not automatically lead to organized violence. This is because organizers of violence, notwithstanding their motivations and objectives, face massive difficulties. First, they must recruit members to join an organization, the activities of which impose great risks and considerable personal costs on members. Second, they must sustain this organization by providing internal cohesion, organizational structure, and discipline. Third, they must finance the expensive logistics of organized violences vehicles and fuel for transportation must be provided; weapons, ammunition, food, and medical supplies must be
58 I Making Sense
purchased; rewards for the fighters and especially for the leaders must be distributed; and perhaps some material support for the families of killed and injured fighters may be granted.
In facing recruitment, leaders of organized violence have to overcome what Mancur Olson has famously called the "free rider problem."34 Since many of the collective benefits of a "rebellion" victory will be realized independent of individual participation within the rebel movements, and because the high costs and risks involved in participation provide very good reason not to join the movement, organizers of violence usually find it difficult to recruit individuals to their organizations.
In theory, there are a few ways to overcome this collective action problem. Motivation for the "right cause" alone, while it certainly plays an important role, is usually not enough to recruit members to rebel movements, and it is certainly never enough to sustain the movement.35 To overcome collective action problems, organizers of violence have to make use of the material and social endowments that they have at their disposal.36 In other words, they have to materially reward participants (or at least promise future rewards), and they have to mobilize participants by making use of the bonds of shared norms.
Material endowment can come from different sources. Organizers of violence may generate profits from the extraction of natural resources such as oil or diamonds;37 from donations from the diaspora;38 from taxing the population; from control over segments of the legal, illicit, or illegal economy; or from sponsors abroad. The amount and the type of material endowment to which organizers of violence have access depend partly on the natural resources of the country and partly on its economic structure and on its neighbors. None of this can be easily changed.
Social endowment refers to the social ties of trust, to general or specific reciprocity, and to the mechanisms of social control that exist within certain segments of society and that make the coordination of collective action easier. A famous concept that subsumes these social qualities is "social capital," which, since Robert Putnam's study, is most often positively associated with being a precondition for democracy.39 However, social capital can go both ways, and networks, reciprocity, shared norms, and trust are also valuable ingredients for the organization of violence.40 Under certain circumstances, social capital can be activated and used for organizing social action.41 To be sure, there is nothing inherent in the concept of social capital that predetermines that this collective action should be directed toward community development
Making Sense I 59
rather than organizing raids on the neighboring village.42 Social capital is therefore a concept that captures social ties within a society suited for overcoming collective action problems, but it does not state how and to what ends these ties are used.
Thus, the feasibility of organizing violence depends on both a specific mix of material rewards and the invocation of social capital, and both are highly contingent on the given social, political, and natural conditions. In theory, material and social endowments could be used in a compensatory manner—that is, a lack of material resources could be compensated for by tapping deeper into existing social capital, and large assets in social capital could compensate for lacking material resources. While it is possible for organizers of violence to acquire new sources of material resources, however, existing social capital is limited, and it is very difficult to build new social capital in a short period of time. Therefore, in practice, social and material endowments are not quite compensatory; whereas organizers of violence may succeed in opening up new sources of revenues, they must rely on what social capital is already there. The significant factor is thus how successful organizers of violence are in making use of existing social capital, and under what circumstances society is willing to put its social capital at the service of the organizers of violence.
Social capital may reside in ethnic, religious, or communal ties, or it may have been accumulated through repeated interaction in socio-professional or cultural networks. Once more, counter to conventional wisdom, ethnicity does not necessarily imply a "higher" or "specific" amount of social capital available within a group. As shown in the following chapters, the social capital on which the various organizers of violence in the Caucasus relied varied from case to case, but all of the social ties that were used were of a much more specific fabric than just "ethnicity." This again highlights the fact that ethnic groups are not collective actors. What brings agency to ethnic groups are small but dense networks that stem from communal ties, kinship, and socioprofessional interaction.
Recruitment is only the beginning of organized violence. Organizers of violence also have to think about the sustainability of their organization. Sustained organized violence requires internal cohesion, the ability to recruit new members, and the capacity to finance the logistics of organized violence. The difficulties involved are similar to, but nevertheless distinct from, those associated with the recruitment stage. Any rebel
6o t Maktng Sense
Making Sense i 61
movement that wants to sustain its existence needs at least some success on the battlefield, otherwise it will soon lose its credibility. To a large degree, military success depends on internal cohesion and discipline. Thus, organizers of violence need to move beyond recruitment and need to build up internal sanction capacities. If they do not succeed, the movement will not be able to operate successfully on the battlefield, and it will probably fragment into smaller, barely coordinated units without a unified command structure. Such a dispersed structure may be quite adequate and successful for waging guerrilla warfare, especially when mainly terrorist tactics are applied, but it is not adequate for winning on the battlefield against an army. As the experiences in Georgia and especially Chechnya dramatically proved, highly fragmented "rebel" organizations also hinder the emergence of efficient state institutions, which, in turn, lead to new waves of organized violence.
Sustaining violence depends on a constant supply of revenue. Many theories of rebellions assume that the success of rebel movements depends on the distribution of rewards only after victory—for example, after the state has been captured or the province has become de facto independent. Such models look at the rewards for the fighters (which, if the fighters are highly motivated and confident, could indeed be postponed until after victory), but they overlook the fact that organized violence requires a constant investment in the logistics of war. Consequently, there is a need to generate profits not only after the war but during the course of violence. This necessity affects the dynamics of violence. Most notably, it facilitates the emergence of a situation that can be called a "market of violence."43 Because sustaining violence is expensive, organizers of violence engage in economic activities that characteristically combine legal business, organized crime, and warfare. Organizers of violence then transform into entrepreneurs of violence.4'1
Entrepreneurs of violence create an economy that tends to be integrated into transnational networks of trade and investment, and they engage in the trafficking of drugs and weapons, in kidnapping and extortion, and in taxing the shadow economy. Profits are reinvested or kept in offshore banks. Gradually, short-term economic interests replace long-term political ones, and entrepreneurs of violence may become interested in avoiding significant battles and sustaining profit instead. Such a situation can be understood as a market of violence: as an economic area dominated by civil wars, warlords, or robbery, in which a
self-perpetuating system emerges linking nonviolent commodity markets with the violent acquisition of goods.45
Once a market of violence is established, there is a strong rationale for the warlords to stabilize the status quo. If government officials receive a share of the revenues from the market of violence, or are themselves acting as warlords, they have an interest in prolonging the violence at low levels. In such cases, sustaining low-level violence with reduced risks becomes a rational objective of both the "rebels" and the "state." This understanding contradicts the view of prolonged war as a communication problem or as the result of anarchy. This market is characterized by the lack of large battles because neither side is ready to commit substantial resources to winning the war. All violent conflicts in the Caucasus developed for a certain period into markets of violence in which combat operations were combined with profitable economic activities. In Georgia, markets of violence did not last long in comparison with such classic cases as Afghanistan, Lebanon, Sierra Leone, and the Democratic Republic of Congo. Chechnya, in contrast, has been completely torn apart by rival entrepreneurs of violence. The establishment of statehood in Chechnya went awry because the successful field commanders were more interested in perpetuation of the market of violence than in restoration of the state. In the permanent struggle for power between the "rump state" and the violent entrepreneurs, the remaining state institutions were dismantled, institutions capable of containing conflicts were devalued, and the rump state was deprived of the resources required to crack down on the private organization of violence. As a consequence, Chechnya sank into anomie and internal conflict, which, among other factors, provoked the second Russian invasion in
The shift from a mainly politically motivated war to a partly or overwhelmingly economically motivated war has important consequences. Kntrepreneurs of violence have a vested interest in prolonging the functioning of their "markets." They may thus obstruct or undermine any political settlements, even if the terms seem to take into account the stated objectives of the movement. In other words, political solutions, which may work for the settlement of politically motivated conflicts, do not work once the conflict has developed into a market of violence.
Unsurprisingly, this shift from'a politically motivated war to an economically motivated market of violence is not reflected in the official
Řa I Making Sense
Making Sense I 63
statements and discourses of the organizers of violence. Circumspection is therefore required regarding the motives they profess. The definitive-ness of these discourses, mostly based on wrongs suffered and the quest for justice, as well as exaggerated friend-foe distinctions, may be deceptive. The differentiation between the normative framing of the conflict and the rules guiding the incentives of relevant actors is essential. This is not to say that the normative framing of a conflict by the parties and other interested observers is just a fancy story of good and evil told by profit-maximizing warmongers. Making sense of violence (perpetrated or endured) is a crucial function of normative frameworking and has strong implications for the chances of post-violence rapprochement and healing. Of even more immediate practical influence for the organization and channeling of violence is the mobilizing power of the normative story. In terms of mobilizing young people to fight, kill, and die it may come second only to the organization of fear.
Finally, the organization of violence depends on leadership. Notwithstanding the suggestive power of rational choice theories, which are inclined to explain the organization of violence as a process that is very much guided by the invisible hand of universal cost-benefit calculations, leadership is crucial. However, telling the stories of the leaders of organized violence in the Caucasus enters normatively slippery terrain. What they all have in common is a skill for organization, a certain amount of social and material capital that they were willing to invest in organized violence, and a charismatic enough personality that appealed to followers. But whereas some became presidents (such as Robert Kocharian in Armenia), others ended up as the most wanted terrorists in Russia (such as the Chechen warlord Shamil Basayev) or were killed in Russian covert actions (such as the first president of Chechnya, Dzokhar Dudayev). Musa Shanibov, a military leader in the war in Abkhazia, retired to his previous life as a university professor in his hometown of Naltchik in Kabardino-Balkaria.
Arguably, the fact that some organizers of violence became politicians, whereas others were killed and still others became terrorists depends not only on the personality, the political convictions, and the actions of these leaders but also on their success on the battlefield; on the political economy of international relations, on which the fate of the struggles for independence depends; and on dominant public opinion (which, for that matter, most often follows the political economy of international relations, albeit often with a certain time lag).
Assigning normative labels to the leaders of organized violence is not the primary objective of this book. Of more relevance is following their various career paths and highlighting similarities and differences in the wny they put their organizational skills and their social and material endowments in the service of organizing violence. Organizers of violence Acquired specific know-how in very different environments. Organizational skills, for example, were acquired in the Soviet army, in the police force, or in the criminal underworld. Social capital stemmed from their well-known and respected families or from their prestige as nationally famous athletes, especially wrestlers. Some had none of this but nevertheless became effective commanders, such as Samvel Babayan, a car washer who became Karabakh's most efficient and ruthless military leader and later defense minister of the unrecognized enclave. It is telling that the sources of their material resources were less diverse than the sources of their social capital. All of the successful organizers of violence were either already quite active in the shadow economy of the late Soviet Union or they quickly learned how to tap into it. Finally and perhaps most important, at a certain stage in their career, all organizers of violence were invited by politicians to participate in the state. These attempts to co-opt commanders into state structures backfired in many cases and ended with the collapse of any remaining state structures.
Tracking the specific trajectories of organizers of violence—from their origins in Soviet-style networks of social interaction; through their time as independent field commanders and violent entrepreneurs; to defense ministers, presidents, or terrorists—helps to unpick the stories of war and peace in the Caucasus.
Expanding the Theory: Institutions and Conflict Analysis
The conjunction of a specific ethnic demography institutionalized by the Soviet Union and the weakness of the central state account for an atypi-cally high risk of conflict in the Caucasus. These risks were then amplified by what could be called pathological social constellations such as atypically high opportunities for either capturing or secession from the state, a massive commitment problem on the side of the new states, and a climate that was highly favorable to ethnic entrepreneurship. This added up to a breeding ground; for the organization of violence, and, in four polities, groups capable of waging war emerged. However, some
64  I Making Sense
polities could opt out of organizing violence, whereas others, despite sharing most characteristics with those that spiraled into war, could riot. For example, South Ossetia, Abkhazia, Karabakh, and Chechyna waged war, whereas Ajaria, Dagestan, Ingushetia, Kabardino-Balkaria, mi Karachai-Cherkessia avoided violence. Of particular interest here ai;e the circumstances and conjunctures that helped some polities to success fully overcome the fatal problems associated with state weakness, while other polities proved unable to rebuild a functioning statehood from the rubble of collapsed Soviet institutions.
Similarly, ethnic entrepreneurship in the Caucasus was potent enough to capture the state, but in Chechnya, Georgia, and Azerbaijan, it was not potent enough to sustain it. Some leaders who came to power on the tide of nationalism have been unable to retain power. Abulfaz Elchibey in Azerbaijan, Zviad Gamsakhurdia in Georgia, and Dzokhar Dudayev in Chechnya all lost presidency and power, and then the state, relatively rapidly. All three left behind failed states wracked with escalated ethnic conflicts, and the reconstruction of these states has still not been fully achieved. There were some factors that enabled the careers of ethnic entrepreneurs to develop but hindered successful state building precisely in those polities in which ethnic entrepreneurs were initially most successful.
In addition, the general propositions of conflict research fail to illuminate how the various armed movements in the Caucasus solved the problems associated with recruiting and financing once tensions had been transformed into organized violence. Consequently, the following areas also need to be explored: how followers were recruited, and how recruits were turned into fighters; how fighters were armed and paid, and from where the weapons and finances came; what the specific mix of material resources and social capital that organizers of violence successfully exploited was, in order to build up a military organization; and, generally, how the logistics of war were addressed, and what specific factors proved beneficial to or hindered this endeavor.
The analysis of these issues is by necessity context-specific since the trajectories, if not the structural preconditions, of an individual conflict unavoidably depend on social, cultural, and political factors and constellations that are in all likelihood highly idiosyncratic. The theoretical lenses of this study therefore need to be expanded in a way that allows the idiosyncrasies of a specific conflict process (its trajectory) to be captured, while at the same time not losing sight of the structural factors
Making Sense I 6;
iliii: accounted for the risk propensity of a given society and the patho-social constellations that made the organization of violence more likely. Hence, the story of "actors who seek to fulfil their dreams, make alliances, learn from one another and make mistakes"46 needs to be included, and, since social action is always to a large extent path dependent, those factors that structure the actions and interactions of relevant actors need to be described carefully. In short, the integration of an action-theoretic paradigm into a structuralist paradigm is needed. Bridging the actor-structure divide requires a focus on the interactive strategies of actors operating within institutional settings that, at the same time, enable and constrain these strategies. Such an approach has become known as actor-centered institutionalism47 and is closely associ-ilted with the new institutionalism in social science.48
In analyzing the trajectories of conflict processes, institutions—which ;ire, in the broadest sense, simply "the rules of the game"—become the most relevant unit of analysis. They are crucial for an understanding of nctors and their strategic interactions because actors depend on socially constructed rules to direct their action. Institutions constitute actors, actor constellations, and the mode of interaction within this constellation. Moreover, they define who is able to participate in a particular political arena, they determine what actors think is both possible and desirable, and they shape the strategies that actors choose in pursuing their preferences. Institutions are rule-bound and repetitive patterns of human interaction. They are "the humanely devised constraints that shape human interaction. In consequence they structure incentives in human exchange, whether political, social, or economic."451
Focusing on the institutional framework of a society therefore allows an understanding of how the expectations, fears, and preferences of actors are shaped, and how these preferences then translate into the strategies that actors pursue—the strategic choices that trigger counter-strategies by other actors and culminate in an interaction which then may or may not lead to violence. Besides incorporating an actor-centered perspective into the analysis, an institutional approach also brings the great benefit that it is not blind to locality or to history. The institutional framework of a given polity will inevitably have a distinct local flavor, as institutions emerge and get rooted during a long historical process, shaped by cultural, geographical, and political idiosyncrasies. This is even true for the Soviet Union, the seemingly isomorphist institutions of which were locally grounded and blended with local informal
66 I Making Sense
Making Sense I 67
rules of the game.50 Focusing on the specific institutional framework of a society thus grasps the specific local conditions that have shaped preferences and strategies, and it is precisely this highly differentiated local institutional framework that accounts for variance in the trajectories of conflicts. In that sense, the institutional framework of a specific polity can be understood as the crucial intervening variable that explains variance in outcomes across cases that seem to have many, if not most, structural similarities.
Neo-institutionalist approaches, in a multitude of nuances, figure prominently across all disciplines of the social sciences and cover a fascinatingly wide array of issues. For example, scholars have used institu-tionalism to explain economic performance, variance in welfare policies across European states, large-scale historical change, the behavior of bureaucracies, and the success of rural development projects. Political scientists in comparative politics analyze political institutions to explain domestic politics, and researchers in international relations investigate the effects of institutions (regimes) on international cooperation. However, institutionalist approaches do not figure prominently in the field of conflict research.51 There appears to be no work in the field explicitly centering on an institutionalist approach, despite the enormous heuristic and explanatory power that institutionalist approaches offer to the study of conflict processes.j2
Similar to what has been said earlier on social capital, there is a certain ambivalent quality to institutions: institutions do not only defuse conflict, they can also breed conflict, indeed, both outcomes—organized violence and the avoidance of it—are equally the result of a society's institutional framework.
Institutions can defuse conflict, thus providing societal stability, via two main mechanisms: institutions can provide procedures to accommodate the diverging interests of competing social groups, thus reducing the motivation to organize violence, and they can provide coercive capacities (monitoring and sanctioning capacities), thus reducing the likelihood that actors actually pursue rule-breaking strategies. Hence, the feasibility of organizing violence is reduced. Both the accommodating function and the coercive function of institutions helps to "embed" conflict. Conflicts are then dealt with (processed) according to known rules and procedures, and any defiant behavior from actors is negatively sanctioned. This reduces the likelihood of conflict spinning out of control and turning into organized violence. However, this embedding of
conflict processing should not be confused with a conflict-free society. Conflicts may still exist, but they are routinely dealt with in a procedural way, and actors usually adhere, either because the threat of negative sanctions for defiance are credible or because the benefits of nonviolent conflict processing exceed the potential benefits of unilateral, non-rule-abiding behavior, or both.
The accommodating capacity of institutions makes it unlikely that actors actually want to organize violence, and the coercive capacities of institutions make it difficult (expensive) for actors to do so. The accommodating capacities are often attributed to democratic institutions. Democratic procedures ensure that different groups in a society have a voice in shaping the politics of the polity they share. Political institutions designed to protect minority groups help to accommodate tensions between different groups, and complex power-sharing arrangements can balance the various interests of different groups.53 None of the accommodating functions—representation, minority rights, and power sharing—need to be provided by formal political institutions, enshrined in the constitution or the legislation of a country. These functions can equally well be provided through informal, unwritten institutions. Informal mediation, informal mechanisms of power sharing, or even informal, yet routinely practiced buying off or cooptation of political contenders by the dominant elite may all add to the stability of a society. For example, within the Soviet Union, informal power sharing between ethnic groups within an autonomous region or territory was routinely practiced. An informal system of quotas that assigned a more than proportional share of key posts within the republic to members of the titular nation was in place in all ethically mixed regions where the titular nation was a minority.54 In multiethnic regions, such as Kabardino-Balkaria and Dagestan, the ethnic composition of the elite was carefully balanced in order to avoid one group achieving a dominant position.55
The coercive capacities of institutions—that is, the ability to monitor and sanction norm-defying behavior—is usually attributed to the power of state institutions: specifically, to the state's administrative capacities and its legitimate monopoly of violence. It is thus the law-enforcement bodies of a state that actually ensure that the rules of the game are adhered to. By this logic, a sudden weakening of administrative capacities is seen as dramatically increasing the risk of organized violence. As with the accommodating functions ofj institutions, the coercive function need not be exercised by formal institutions. Arguably, many regimes derive
(>H I Making Some.
their stability not from strong state institutions but, rather, from informal yet highly institutionalized patron-client relations.1"' Such informal patron-client networks may substitute for state institutions, but, more often, they merge with state bureaucracies to form hybrid arrangements, which are nevertheless stable and quite effective in providing regime stability, as the longevity and persistence of many patrimonial regimes aptly demonstrates.
Institutions do not always defuse conflict; indeed, under certain circumstances, they may breed conflict. By a twist of fate, they may turn into a subversive power that actually undermines the stability they have helped to preserve in the past. Institutions do not change easily. Even though the rules of the game are usually stable, and actors rely on these rules in their expectations and strategies, this does not mean that all or even most actors are content with this institutional equilibrium. Certainly, at least a significant coalition is, but for those who are not content with the current situation, the risks and costs associated with changing the rules may simply be too high. They may fear the uncertainty that will come once the actual set of rules becomes defunct, or they may be locked in a power asymmetry, which makes them unlikely to succeed in changing the rules against the will of the dominant elites. Nevertheless, institutional change happens; indeed, it is one of the constants of society.
Institutional change may happen incrementally, as a result of shifting coalitions and the changing bargaining power of actors, or it may be triggered by external shocks to the institutional framework.57 Rapid and exogenously induced institutional change is by default conflict prone. A weakened institutional framework may prompt the disembed-ding of conflict-—conflicts are no longer "processed" according to shared norms or procedures, because one or both parties to the conflict seek to achieve their objectives by force. A weakened institutional framework also leads to increased competition, because actors see an opportunity to redesign the institutions in their favor. In such a situation, the competition between actors turns into a "winner-takes-all" game, and the increased stakes lead to increased risk taking, which, in turn, increases the risk of violent escalation. Moreover, a change of the institutional framework will also lead to a change of the distributional consequences of institutions: As a result of changes in the institutional framework, some social actors will get more, while others will lose out, and this will affect the power balance. Since actors base the decision of
J
1
3
Making Sense I 6y
whether to pursue a risk-avoiding or risk-seeking strategy on their perceptions of their relative strength vis-a-vis other actors, a sudden change in the distributional function of institutions may dramatically increase flic propensity for actors to engage in high-risk strategies, thus increasing the propensity of conflict and violence.
An institutional approach is a heuristic device that allows the identification of relevant critical junctures in the trajectories of conflicts, and as such it requires an inductive research strategy. The reason for this inductive strategy is straightforward and dictated not by methodological reasoning but by the sheer complexity of social life. The universe of specific arrangements of institutions is infinite, and societies across space and time have produced highly diverse institutional arrangements. But whereas the universe of institutional solutions may by infinite, the functions that an institutional framework fulfills or fails to fulfill with regard to conflict processing (respectively, conflict escalating) are not; the institutional framework in general shapes incentives and constraints of social actors. More specifically, the distributional function of institutions determines who gets what; hence, the relative power of actors within a society is a function of the distributive consequences of the institutional framework. Whether an actor will pursue a defensive, risk-avoiding or aggressive, risk-seeking strategy therefore depends on its relative power vis-a-vis other actors and thus, ultimately, on the institutional framework.
Summing up, the distributional consequences of institutions and their capacity to embed conflict by accommodation or coercion are highly important for the analysis of conflict trajectories. It is the specific institutional framework of a society that accounts for, among other things, how ethnicity is institutionalized, what social capital is available to groups, how leaders mobilize and control followers or fail to do so, and how states become consolidated. Hence, it is by focusing on the specific institutional framework of a society (and its changes) that the global propositions of conflict theory and the microlevel foundations of organized violence can be woven together into a coherent narrative.