European Research Council

Conceptualizing Civil War

Civil war is still not conceptualized in Social Sciences. In recent years, various studies have questioned this omission and, more broadly, the general difficulty of addressing the subject of armed conflict (Lutz 1999, Joas 2003, Joas & Knoebl 2013, Wimmer 2014). This has led some to advocate a return to the classic authors who thoroughly studied war in the late 19th century (Huebner 2008, Malesevic 2010, Nolan 1970).

The difficulty in thinking sociologically about war resides in the object itself. Firstly, part of the academic publication is driven by strategic agendas (i.e. normative concerns) focusing on counter-insurgency and terrorism, and remains on the side-lines of the social sciences. Secondly, a significant part of the literature is concerned with the causes of civil wars and the post-war period, in particular the post-conflict literature. Between the time before and the time after, the actual time of the war itself tends to disappear. Finally, one trend in research, particularly anthropological, focuses on the war-peace continuum (Richards 2004, Naepels 2013), significantly minimising the structural effects of both violence and the withdrawal of the state.

In the end, few studies have put forth general hypotheses regarding the specific processes of civil wars, in particular the implications of the loss of state monopolies. An analysis of the scientific output regarding this field leads us to highlight several key weaknesses: the methodological shortcomings of a quantitative approach, the difficulty of qualitative approaches to produce comparative studies and the analytical limitations of current research on rebel institutions.

1. The limits of numbers. Through the 1990s, mathematical formalisation associated with quantitative studies gained currency. These approaches, based on the American databases started in previous decades, especially Ted Gurr's Minorities at Risk and J. David Singer's Correlates of War (COW), collected data on civil wars. Despite their technical nature, these two projects, COW in particular, were highly influential, notably by imposing definitions, benchmarks and other criteria used ever since in all quantitative research.

"Currently, about a dozen research project have produced civil war lists based on apparently divergent definitions of civil war, but there is less pluralism here than one might think. Most projects do not conduct original historical research and depend heavily on COW. The result may be replication of errors due to the original COW coding rules and uncertainty about whether different definitions generate different results" (Sambanis 2004).

In this vein, Paul Collier and Anka Hoeffler (2000), followed by James Fearon and David Laitin (2003), wrote the two seminal studies that have structured the discussion, introducing an opposition between greed and grievance. Starting from different assumptions, these researchers correlated the occurrence of civil wars with the presence of natural resources, the nature of the political system and the stage of economic development. Therefore poverty and institutional weakness constitute the two key explicative variables put forward by most economists to explain the occurrence and duration of civil wars (North et al. 2009; Chassang and Pedro i Miguel 2009; Acemoglu et al. 2010). Other studies have also found a correlation between inequality and civil wars (Esteban and Schneider 2008, Cederman et al. 2013).

These quantitative studies have been used to empirically validate hypotheses based on Rational Choice Theory and Game Theory. Thus, following the application of rational choice theory by James Fearon (1995) and modelling of "bargaining" by Harrison R. Wagner (2000), Barbara Walter claims "reputation-building" is determining regarding the duration of violent separatist conflicts (Walter 2009). Equally, we find adaptations of realist theories of international relationships in quantitative studies of civil war, for example around "bandwagoning" as a factor in shifting alliances (Christia 2012).

These studies have been criticised for their lack of rigour and multiple biases affecting their samples (Kalyvas 2006), the definition of the model, and the use of Rational Choice Theory (Walt 1999, Cramer 2002) and suffer from serious limitations. Such approaches define Civil War based on quantitative thresholds (more than 25 or 1000 dead) for which the empirical and the theoretical justification is unclear. The state of the international system is rarely taken into account, which leads to the assemblage of case series that are historically arbitrary. For instance, the current stability of borders, like the Cold War alliance system until the 1990s, is a decisive constraint on armed actors and changes the nature of civil wars. Armed conflicts take place in areas where data collection is particularly difficult. Researchers specialized in specific conflicts have questioned the validity of the quantitative data collected in their areas, for example Ibrahim Abdullah (1998) and Roland Marchal and Christine Messiant (2002). Conflicting conclusions emerge depending on how dependent and independent variables are selected. Thus, quantitative studies have published contradictory correlations between resource scarcity and civil war, on one hand, and resource abundance and civil war, on the other (Mildner, Lauster and Wodni 2011).

In response to these criticisms, some American political scientists, in a sub-field largely dominated by quantitative approaches and Rational Choice Theory, have integrated qualitative elements (Tarrow 2007). Inspired by the Mancur Olson's opposition between stationary and roving bandits (1993), Jeremy Weinstein opposes a model of predatory and opportunistic rebellions practicing greed, to the politically-involved activist rebellions which mobilise grievances. Comparing four conflicts, he concludes that the availability of resources is negatively correlated with the ability of the armed movement to take into consideration the people's demands (Weinstein 2007). Stathis Kalyvas (2006) explains the spatial distribution of violence in civil wars by the degree of territorial control exercised by insurgents. The book downplays the ideological aspect of the struggle and centres the analysis at the micro level, looking at private conflict. Kalyvas built two models, a "Hobbesian" privatisation of violence by political actors and a "Schmittian" model of polarisation/politicisation of private disputes. Finally, Elizabeth Wood shows that the long-term mobilisation of farmers in the Salvadorian armed conflict is not a by-product of the economy, but of the political culture produced by struggle (Wood 2003). Weinstein, Kalyvas and Wood have since opened new avenues of research on the mode of organisation, the level of territorial control and social transformations, which confirm the importance of focusing on the processes involved in civil wars. From this research, a scientific community has emerged, without a European equivalent, around the Order, Conflict and Violence Program at Yale University (Kalyvas et al. 2008).

These approaches present certain limitations. Weinstein objectifies armed movements, considering them immutable in time and space, and as a result neglects socialisation and engagement processes. Kalyvas does not take into account the meaning that perpetrators give to the violence they produce, actions that may be local in their actual organisation, but also part of a broader engagement and set of values. Finally, Wood has proposed a stimulating look at the reconfiguration of social networks during civil wars, but the emphasis on the identification of dependent and independent variables is problematic (Wood 2004). Indeed, civil wars are periods of accelerated transformations where the identification of cause and effect is complicated by the impact of feedback. A comprehensive approach of societies living through civil war would resolve this problem and respond to the fundamental objection of quantitative researchers: the absence of a comprehensive theory based on qualitative methodologies (Hegre and Sambanis 2006).

2. Searching for a qualitative comparative theory. While qualitative approaches exist in the United States, they remain marginal among the academic community. In fact, Collier, and even Kalyvas, insist repeatedly on the limits of these approaches: speeches are not reliable, and actors can dissimulate, deliberately or not, their real motives (Collier and Hoeffler 2000, Kalyvas 2006). In practice, the tradition of qualitative research has been pursued mainly by European researchers without little dialogue with quantitative researchers.

Faced with a proliferation of civil war in the 1990s, studies using thick description have opened up new perspectives. Anthropologists were led to study the conflict that took place in their field using ethnographic tools, especially in Sri Lanka (Nordstrom 1997), Mozambique (Geffray 1990), Guatemala (Stoll 1993) Sierra Leone (Richards 1996) and the Côte d'Ivoire (Chauveau and Bobo 2003). Collectively, this work constitutes a major contribution to a process-based analysis of civil war. Carolyn Nordstrom highlights the coping strategies of people living in war zones, Christian Geffray the importance of the grounding of the armed movement in its social environment, and David Stoll the constraints imposed on civilians by armed actors. Richards highlights the centrality of generational conflicts and Chauveau and Bobo the importance of internal migrations and narratives on indigenousness. More recently, the work of Hilde Waage (2006), Sverker Finnström (2008) and Henrik Vigh (2006) usefully emphasised the debrouillardise (resourcefulness) of social agents, their interpretation of events as well as their social navigation in order to illuminate the trajectories of individuals in an environment of uncertainty.

In parallel, Africanists have made significant theoretical breakthroughs, mainly through the study of decentralised modes of governance (Bayart et al. 1999, Daloz and Chabal 1999). Such analyses are particularly useful to the examination of civil war. Work on the socialisation of fighters (Schlichte 2009, Reno 2011, Banegas 2012, Debos 2013) effectively articulates social transformation and the strategies of the actors. In the Islamic world, a vigorous debate has emerged on the role of social capital amidst crises in Lebanon (Seurat 2012) and Afghanistan (Dorronsoro 2005). Recent work also usefully underlines the relationship between social structures, ideology and violence in Islamist movements (Bonnefoy et al. 2011), which may be linked with certain studies of Southeast Asia (Gayer and Jaffrelot 2008). In addition, the routinisation of violence has been particularly well analysed in Latin American studies (Green 1999, Kooning and Kruijts 1999).

However, despite the empirical depth of these approaches, they are still considered in isolation as case studies, even when presented as a collection. If Nordstrom rightly emphasises the creativity of the people in situations of armed conflict and the role of transnational actors, her general work (2004) does not specify how each of these two elements function in practice. Similarly, Richards correctly points out that civil wars are "social projects", produced by groups for whom it is important to "understand the character, organisation and beliefs [...] and their impact on other supporting groups, resisting or victimised by their activities [...]. In short, war is inescapably sociological" (Richards 2004: 4). Yet the collective work he leads is primarily a collection of case studies. Finally, in the Journal of Agrarian Change issue he co-edits with Christopher Cramer, Richards helpfully highlights the issue of access to land and more generally of land capital in civil war (Richards and Cramer 2011), building on the thoughts of Christian Lund on the control of land by sub-state actors and the fragmentation of sovereignty (Lund 2011). But again, the areas studied are not systematically compared.

This literature establishes the foundations of a sociological approach to civil war. We draw upon this literature to develop an innovative theoretical framework. European approaches to civil wars have until now lacked theoretical publications that could provide the basis of a comparative research (Picard 2006, Imbusch and Veit 2011). Indeed, a methodologically rigorous comparison will need to go beyond drawing parallels between different, even numerous, cases, and one-off applications of a theoretical model.

3. The rebel governance The most recent American theses on civil wars endeavour to overcome the inherent limitations in the work of Weinstein, Kalyvas and Wood, focusing on the meso-sociological level: the internal organisation of armed groups and the dynamics of the pre-war period. Zachariah Mampilly (2011) and Ana Arjona (2010) have shown that the degree of development of existing institutions is a determining factor in the ability of insurgents to set up an administration. In addition, Paul Staniland (2014) and Sarah Parkinson (2013) note the importance of the pre-existing social relations in the cohesion and organisation of armed groups. Nevertheless, these studies, by focusing on armed groups, do not take into account the transformations that affect the whole of a society. Moreover, the search for continuity through the pre-war and the civil war period can disguise effects that are specific to the withdrawal of the State.

Since then, several studies have tried to rectify this bias by focusing on the governance of the population by armed groups (Kasfir 2005, Bakonyi and Stuvoy 2005, Risse 2011). Drawing on cases of Somalia (Menkhaus 2006), Congo (Tull 2003 Vlassenroot and Raymaekers 2008), Sri Lanka (Stokke 2006), India (Barruah 2005) and Colombia (Taussig 2003), researchers have shown that, far from being purely areas of disorder, civil wars are areas that experience a significant political reconfiguration. Even if these studies do not provide a general model, they question the link between different scales of analysis, meso, micro, and macro. In this vein, my project provides a modelling of global changes that affect societies facing civil war, with the objective of creating a theory that is qualitative, comparative and empirically testable, thus addressing a key objection of quantitative researchers.

The theoretical framework

For the purposes of our research, civil war will be defined as a situation of withdrawal or disappearance of the State when challenged by one or more armed groups. The objectives of the latter can be numerous, sometimes global (Al Qaeda) sometimes parochial (Nigerian delta militias), and can change over time. In addition, we will focus on situations that last, as opposed to sometimes violent but brief political crises, (the Tunisian revolution for example) that do not have structurally comparable consequences. Following on from the important work done by Pierre Bourdieu, this project prolongs and alters his conceptual framework to bring a better understanding of volatile situations of conflict. Concepts developed in a pacific context cannot be transposed uncritically to analyse civil wars. My previous research has now led me to reformulate certain theoretical propositions.

One of the recent achievements in social sciences is the comprehension of non-routine situations within a framework of contentious politics (McAdam, Tarrow and Tilly 2001) and multisectoral mobilisations (Bourdieu 1984, Dobry 1986). Nevertheless societies prior to the outbreak of civil war are, on the whole, structured differently. The increasing autonomy of social sectors (social, political, religious) (Weber 1978) is regarded as a characteristic of modernity (Luhmann 1995, Giddens 1984, Bourdieu 1980). However, non-democratic societies, in which most civil wars occur, are often characterised by an over-pervasive security apparatus (Baczko, Dorronsoro and Quesnay 2013) and intricate state-party relations (Dorronsoro and Gourisse 2014). Moreover we add to the three types of capital (cultural, economic, social) established by Bourdieu the concept of identitarian capital (Dorronsoro and Grojean 2014).

Armed challenge to the state-monopoly brings about a radical and non-anticipated reconfiguration in three significant areas: the value of (different types of) capital, the formation of rival institutions, and rationality in everyday life (deliberation, identities, life trajectories, routines). These three themes correspond to three levels of analysis - macro, meso, and micro - that for the sake of clarity, I will now outline starting with the macro level.

1. The changing value of capital In the context of a "general economy of practices", the State is the guarantor of the value of capitals and relations between sectors (religious, judicial, political…) (Bourdieu 1980: 209 and later, 2012). However, any withdrawal of the state does not, on the whole, provoke the disappearance of these different types of capital, simply the reassessment of their value. The ability of social interactions to continue beyond the collapse of central institutions reveals the ability of certain social sectors to self-organize through local, informal or transnational regulation. For example, in the religious field restructuring can occur around transnational institutions. More generally, three phenomena can be observed where the State has lost control: the emergence of new capital, an often-abrupt change in the value of existing capital (identity, cultural, economic and social) and circuits of conversion of these capitals.

1. The State's withdrawal marks the end of State monopolies - a "reverse sociogenesis" -, which leads to the (re)creation of new types of capital. As an example, the disappearance of national armed forces permits the appearance of a new military capital. The formation of this specific capital occurs through the conversion of existing skills or capital, for example, economic or social (Dorronsoro 2005). In addition, the formation of a new capital affects the relative value and the potential conversion of the remaining types of capital of social agents.
2. Certain types of capital are particularly affected during civil war. This is particularly the case of social capital (Baczko, Dorronsoro, Quesnay 2014) and identitarian capital (Dorronsoro and Grojean 2014). In both cases, the withdrawal of the State indicates by contrast its role in what is commonly referred to as "spontaneous social life". For example, identity (ethnic or religious) can be considered as a collective capital. Indeed, identity cannot be reduced to territorial or social belonging, a category of ethnographic description or popular perception; it also reveals varying access to resources. The role of the state, as articulated by its public policies, is central to understanding this unequal access. This hierarchy is in part naturalised by the state (Anderson 1982, Bourdieu 2012), even if the state is not able to retain control of the patterns of thinking in a world of circulating ideas and transnational mobilisation. A sudden denaturalisation of inter-group hierarchies occurs when the state is challenged or disappears. Therefore, actors have to deal with psychological adjustments (see C) and may employ violence in order to restore, or to challenge, the hierarchy hitherto accepted.
3. The withdrawal or collapse of the state alters the way the various types of capitals are converted. Indeed, during peace, the State is usually both the operator and the guarantor of this conversion. Civil war reveals how capital is converted when State intervention disappears. On one hand, new conversion circuits appear. For example, politico-military actors have the resources to accumulate economic capital through taxation or alliances with various elites (e.g. landowners) (Dorronsoro 2005). On the other hand, State barriers of various kinds (economic, legal) that prohibit or make it costly to transfer capital from one field to another tend to disappear, and hence facilitate capital conversion. For example, religious elites, previously excluded from politics, convert their religious legitimacy into a political one by becoming party leaders in Afghanistan (Dorronsoro 2012).

2. The genesis of institutions The end of a state monopoly creates the opportunity for the formation new institutions. This process permits us to observe the genesis, the workings and the transnational dimension of institutions in situ.

1. Three interlinked social processes permit the establishment of alternative institutions: the conversion of activist or party-based capital, objectivisation (Berger and Luckmann 1966) and resource extraversion. First, social capital, activist or party-based capital is mobilised to form new institutions. Because activist networks are inter-linked, they occupy a central place in the formation new institutions. In this context, institutions appear as objectivised social capital as we demonstrated in Syria (Baczko, Dorronsoro, Quesnay 2014). The genesis of rational-legal organisations is often to be found in the mobilisation of networks that are informal and based on emotional ties. The objectivisation of these new institutions often follows, by mobilizing symbols and closure procedures. We observe, for example, the appropriation of State attributes (forms, uniforms, official maps) by armed groups as diverse as the PKK, the SPLA, the RCD Goma and the Tamil Tigers. The judicial field can rely on pre-existing and legitimate institutions, particularly within an Islamic framework, or other traditional practices. Finally, extraversion is the rule because of the scarcity of local resources and the frequent incapacity to self-organize. As a result the dependence on external resources (diaspora, foreign countries, transnational networks) is essential to the development of new institutions, as we have shown in Syria and Afghanistan. This observation confers an even greater significance to the exceptions (the PKK in Turkey, the Islamic State in Syria and Iraq, the FARC in Colombia), highlighting that bureaucratization is crucial to explain the ability of an armed group to mobilize resources.
2. Two important observations emerge from the study of rebel institutions. First, multiple competing institutions (State-related and those associated with armed movements) transform the workings of traditional monopolistic institutions, notably in terms of security and justice. Monopolies can be local, but in the absence of complete territorialisation, complex situations arise, notably cooperative relationships of accommodation (formal or informal), and collaborations to manage certain indivisible goods, such as electricity in Aleppo (Syria). Second, the networks related to these institutions are consolidated in a process of mutual validation. For example, Taliban courts, enjoying functional autonomy succeeded in objectifying judiciary activities in the eyes of the population. This legitimized the administration and military organization which, in turn, empowered them in their newly gained functions (Baczko 2014).
3. Rebel institutions often bring organisational innovation. The extraversion of resources involved plays an important part in this innovation. The dynamics of civil wars generally go across state borders and most armed movements have political leadership outside the territory. This helps to explain why organisational models of the armed movements are sometimes at odds with local history and pre-existing authorities. This transnational dimension has important consequences for governmentality, particularly when NGOs or IOs import their own models of population management. For example, the organization of camps often leads to changes in social hierarchy; whether it is the status of women or the legitimate modes of authority. Moreover, the decentralized functioning of NGOs goes against the armed movements' attempts to impose a monopolistic authority and can lead to recurring tensions.

3. Rationality and daily routines During civil war, the daily practices and perceptions of actors change in four areas: identity, routine, decision-making, and biographical trajectory.

1. How do civil wars change identities? In the absence of State guarantee, the hierarchy within and between different ethnic or religious groups are subjected to brutal changes. Individuals face complex situations, and deploy varied tactics to navigate the new status quo: notably the concealment of identity, and the choice of one affiliation over another. These situations mobilise powerful affects and show that civil wars are situations that polarise and over-simplify the multiple identities of an individual or a group. Changing alignments of religious and ethnic identity and political affiliation forces individuals to reconsider and reconstruct their identity in a newly over-simplified manner. In this way, and as a result of the current conflicts, despite deep doctrinal differences and a troubled history, the Twelver Shias, Alawis and Alevis are often perceived and now tend to see themselves as belonging to the same group in Lebanon, Syria, Iraq or Turkey.
   Furthermore, the effects of violence also transform the relative power dynamics between age groups and between men and women. Contrary to the assumption of Elias, violence does not appear to be a step backwards, but more often the result of the socialisation into an institution (political party, army), conforming to an ethos of virility or religious duty. The use of violence values certain skills and social positions, which makes it a resource that young men are more prone to use. Women's activism (in the PKK for example) can work to slow, to a certain extent, the overwhelming domination of young men (Grojean 2013).
2. Routines are an important aspect of daily life. The majority do not constitute instrumentally rational behaviour (Berger and Luckmann 1966). But in a civil war, individuals can no longer follow institutionalised routines, because the uncertainty of violence forces them to anticipate a new set of consequences, each time potentially dramatic, in every one of their actions (Green 1999). Most daily activities (sending children to school, participating on the black market, quarrelling with a neighbour) have potentially serious consequences. For example, how now to choose the authority to whom one should present a potential grievance?
   Do individuals decide differently in times of crisis? Civil wars reveal to what extent decision-making depends on a specific context (Vigh 2006). The ideal-typical actor has three characteristics. First, she calculates the risks and likelihood of success. Second, she arbitrates between different ends, individual or collective. Finally, she deliberates, that is to say she makes her decisions after discussions in which he clarifies his preferences. The site of deliberations is strategically important because it is at the same time a place of information exchange, collective calculation of risks and benefits and project comparison.
3. Non-routine contexts transform the calculations, sociability and goals of individuals. The propensity to act is affected by unusual contexts in three ways: a hyper-assessment of every situation, group dynamics and the definition of a common project. Firstly, the lack of precedent and the risks involved mean that actions carry uncertainty and require therefore increased time and energy as early as the preliminary discussion stage. Individuals show greater attention to the consequences of their actions as institutionalised routines, which minimise uncertainty at an individual level, shrink or disappear. However, minimising the risks taken does not mean fewer mistakes are made, due to the high degree of uncertainty and the inexperience of the actors. Secondly, the deliberations of small cohesive groups often lead to a higher level of risk-taking. Indeed, the social psychology literature shows how under certain conditions decisions tend towards radicalism. Finally, during these discussions, stakeholders define the meaning of the conflict. These exchanges allow the construction of a common view on the legitimate means and the nature of the claims. Deliberation transforms the subjective perception of political opportunities, regardless of objective events (Baczko, Dorronsoro and Quesnay 2013).
4. Finally, civil war involves biographical disruption, pushing us to consider the partial loss of predictability in terms of social trajectories. In particular, the existing social capital of individuals fluctuates enormously when confronted by two opposing trends. On the one hand, the majority sees its social capital decrease and consequently its ability to gather information. On the other hand, activists, socially hyper-active, increase their social capital and thus increase their access to information. However, in certain contexts, the level of engagement cannot be explained in terms of conversion of existing capital or as part of the continuation of the pre-existing social capital. Local contexts (sub-national) are often decisive. When looking at the background of activists prior to war, individual participation in activist networks appears to be on the whole unpredictable.