The Exhaustion of a Paradigm

5 By adopting RCT, neopositivist works execute a theoretical power grab whose radicalism has several consequences. First, this way of thinking explicitly rejects a century’s worth of social science research on the pretext that the works it produced do not tally with scientific criteria. Rejecting this intellectual heritage has created the conditions for a sort of professional amnesia that translates into a tendency to reinvent extant concepts, as demonstrated by certain recent works on transnationalism and socialisation. [6] Next, the lion’s share of studies on civil wars these days is produced by researchers with limited or no direct knowledge from fieldwork or even from the secondary literature dealing with the countries in question. Young researchers are explicitly discouraged from gathering data by observation and non-structured interviews. The result is that the neopositivists, bereft of contextual knowledge, frequently are incapable of putting forward sociologically relevant causes to explain the correlations found. The American Political Science Review recently furnished an example in the form of a disastrous study on the demarcation line in use during the German occupation of France, in which the authors failed to acknowledge the presence of a railway line, although it was, one would have thought, the obvious reason for the Resistance’s sabotage actions. [7] Beyond this one questionable article, the more general problem is that leaving historians out of the referee process leads to texts being accepted solely on the basis of the formal validity of the methodology employed.

6 Then, the epistemic closure of the neopositivist movement has the effect of shielding its findings against all substantive criticism. It has been pointed out for a long time, by Donald Green and Ian Shapiro in the United States or Raymond Boudon in France, among others, that RCT, the anthropological model favoured by the neopositivists, suffers from a design flaw. [8] The neopositivists shut off the discussion by challenging the validity of critiques that do not accept their premises, at the risk of making themselves look highly partisan. Thus, in 1996, Robert Bates, professor of political science at Harvard and president of the Comparative Politics section of APSA (American Political Science Association), called for drawing a distinction between “social scientists,” who line up behind RCT and “area specialists,” whose output is reduced to a literary form. [9] With respect to civil wars, recurrent critiques have been levelled against RCT since the late 1990s. Here, the debate over greed vs. grievances, growing out of the studies by Paul Collier and Anka Hoeffler as well as those by James Fearon and David Laitin, exemplifies this epistemic closure. [10] Despite the theoretical and empirical criticisms levelled at these works, [11] their categories are regularly revived, for instance by Jeremy Weinstein when he contrasts predatory rebellions (oriented towards greed) with politically engaged rebellions (mobilised by grievances). [12] Similarly, most of the analyses put forward to explain the duration and occurrence of civil wars follow the same pattern of contrasting poverty and institutional weakness [13] with inequalities. [14]

7 Finally, and this is probably the crucial point, the advances made in understanding civil wars thanks to research that relies on the RCT paradigm have been remarkably limited, particularly considering the scale of the resources expended. [15] Specifically, game theory has only very marginally contributed to clarifying the reality of the conflicts. Barbara Walter, in applying game theories to civil wars, reworks three factors previously proposed by James Fearon in his 1995 paper on the occurrence of international conflicts: asymmetric information, “commitment problems,” and issue indivisibilities. [16] From this she concludes that governments confronted with separatist movements employ violence as a means of snuffing out other secessionist demands. This conclusion, which has limited scientific interest, is—more importantly—impervious to proof. Indeed, unfazed by this non-sequitur, the author extrapolates the strategies of states from the observation of individual behaviours in laboratory settings. [17] Even research that is more empirically based presents similar limitations. In his classic book, Stathis Kalyvas examines the spatial distribution of violence in civil wars as a function of the degree of microscale control exerted by the insurgents. [18] His theory of a joint production of selective violence by fighters and civilians rests on a radical reductionism derived from his hypotheses about rationality. Notably, because of the “urge to survive,” the individual is supposed “to be good at” calculating threats and opportunities but unaffected by emotions and the past and incapable of anticipating. [19] Kalyvas perceives very well that his hypotheses are extremely restrictive but suggests, based on anecdotal evidence, that civilians overestimate the stability of armed groups’ control and that they make their decisions essentially according to local information and events. It therefore follows that violence produces obedience among those who suffer it. [20] However, there are many examples of resistance to violence, of the porosity between civilians and militaries, of the importance of governance among armed actors, of a capacity for anticipation, or the importance of national or international information. [21] In fact, any consideration of the sociohistorical context implies a renunciation of the RCT. Thus, Elisabeth Wood argues that, in El Salvador, the long-term mobilisation of peasants in the armed movement does not depend on the economic situation, but on the political culture produced during and by the struggle. Unable to get around the Olsonian paradox inherent in the RCT paradigm, she draws on the theory of social movements. Hence, in a specific case she revisits the findings of Tarrow, McAdam, and Tilly. [22] The attempt to revert to the RCT in the appended formal model forces a non-operationalisable definition of personal interest: “defiance, an intrinsic motivation, and/or pleasure in agency.” [23]

An Arbitrary Reduction of Legitimate Objects

14 A falsely rigorous methodology and a problematic conception of rationality translate into a naturalisation of research objects and an arbitrary discounting of legitimate subjects. Generally, the neopositivists claim that their categories objectively describe actions. Thus, acts of violence are often studied through the more easily measurable category of homicides, which excludes other forms such as injuries, and despite the difficulties that the social qualification of these acts poses. In practice, the neopositivists generally adopt the classifications produced by international and Western institutions. The literature on refugees is particularly instructive in this regard, since the statistics and hence the definitions of the international agencies (United Nations High Commissioner for Refugees, International Organization for Migration) are adopted without debate. [32] Similarly, the distinction between civilian and combatant is a reification of categories taken from international law. In the real world, a man fights for part of the day, an individual shelters a fighter, a judge serves the insurrection: in a civil war context, are they civilians? It is telling that these questions, ignored by the neopositivists, do not escape the actors: thus, the Taliban movement and the Human Rights Unit of the United Nations Assistance Mission in Afghanistan (UNAMA) engage in repeated discussions on the status of policemen when the latter are not engaged in combat operations. Here, too, the uncritical adoption of categories injects a normative bias in lockstep with the institutions that produce them.

15 Legitimate research objects are selected depending on how feasible quantification is and are co-produced with the institutions that provide the data. Thus, after Kalyvas’s book, works multiplied that dealt with the occurrence of violent civilian deaths and the degree of control over territory, because of the measurability of these objects rather than their intrinsic interest. Similarly, the forming of armed groups is reduced to enquiries into cohesion, and the process of engagement to a mere decision. [33] In addition, accessing institutional databases (World Bank, the American army) most often presupposes that those institutions get to co-define the research object. [34] Publications are legion that pose questions directly inspired by these institutions: for instance, the effects of aerial bombardment on the support for Western intervention in Afghanistan and the role of cell phones in the insurgent attacks in Iraq—both studies funded by the U.S. Air Force. [35] Similarly, the conclusions drawn by Andrew Beath, Fotini Christia, and Ruben Enikopolov—to wit that involving women in the distribution of aid has a partial effect on their social position; that bypassing the local elites in aid distribution gives villagers a sense of participation without increasing effectiveness; and that distribution of aid does not help win “hearts and minds” except where violence is low-level and of short duration only—are those of an audit of a World Bank programme. [36] Besides the ethical questions that a de facto participation in counterinsurgency programmes can raise, this alignment with institutional objectives discourages theoretical discussions in favour of a normative and technicist approach.

16 Finally, these works leave the impression that they systematically disregard the most interesting sociological enigmas in favour of objects whose main interest lies in their being quantifiable. Emotions and values are treated at the margins and in maladapted analytical frameworks. Hence, Kalyvas and Walter treat norms and emotions as prediction errors in their models. [37] Roger Petersen highlights the importance of emotions and social norms, but ignores their social construction and reduces them to their measurable consequences, thus emptying the initial observation of all relevance. [38] Another failing of neopositivist works is the refusal to take an interest in ideologies, storytelling, and, more broadly, imaginaries. This results in the key articles and papers in the field skipping over these questions, as if one could analyse moves by the PKK (Partiya Karkerên Kurdistan), the Islamic State, or LTTE (Liberation Tigers of Tamil Eelam) without considering their universe of meaning. We had to wait until 2014 to see an article conceding, in very general terms, that ideas and ideologies do play a role in civil wars. [39] In summary, everything points to an exhaustion of the paradigm: a sub-field that is more and more technicist, rarely contributes to general sociology, fails to revisit its hypotheses, and, moreover, whose outputs are frequently repetitious and trivial.

Monopoly or Economy of Violence?

18 Most often, the definitions of civil war rest, explicitly or not, on a Weberian conception of the state as holding a monopoly on legitimate violence within a given territory. [40] Civil war therefore would be the loss of that monopoly following a violent challenge that ends in a situation of divided sovereignty. Without denying the heuristic interest of this definition, it is based on a conception of the state that can obscure the complexity of the objects it subsumes. We propose to introduce an alternative conceptualisation in terms of the economy of violence, defined by the relatively stable interactions (competition, cooperation, delegation) between actors capable of using violence or threatening it. These actors differ (individuals, clans, criminal organisations, state institutions, militias) and encounter one another around social, economic, and political stakes. The economy of violence as we conceive it does not presuppose an economic rationality of the actors and, moreover, says nothing about this social order being functional or legitimate. [41] This definition departs from the Weberian approach in three aspects: legitimacy, monopoly, and the role of the international.

19 First, this definition does not assume that state violence is legitimate, since this assumption is an empirical and theoretical impasse. [42] In particular, discovering why individuals obey the state in routine situations is often impossible. Empirical work under authoritarian regimes is generally fraught with difficulty and it is scarcely possible to decide between different hypotheses (fear of repression, interiorisation of dominance, a more or less negotiated equilibrium between the state and social groups). Moreover, crisis situations do not reveal the individual’s “real” preferences because these are not stable. [43]

20 Second, no state disposes of a complete monopoly on violence. Some states coexist with an important level of private military capital: [44] tribes, social or geographical margins, vigilantes, criminal organisations. [45] In certain countries, there is even a trend towards developing private military capital (Mexico and Central America). Certain types of violence are also accepted or tolerated: against women, children, marginalised populations, etc. What is more, the state is never a unitary actor and conflicts between state institutions are on occasion anything but trivial. Nevertheless, institutions of the state play a key role in how the economy of violence is organised, including by confining the actors, for instance in the illegal economy or in the territorial margins. [46] From this perspective, a collective work in progress shows how, in situations of high levels of violence in Mexico and Central America, the existence of informal rules allow a reproduction of social routines. [47] Criminal actors that challenge the state directly are punished severely, for example, the Medellin Cartel in Colombia or the Cosa Nostra in Italy. Furthermore, an armed group may employ violence to grab power or as a tactic in negotiating with the regime in place. In the case of the Democratic Republic of Congo, after the Pretoria Agreement (2003) armed groups regularly entered the political system during negotiations that functioned as sites for converting military capital into political or economic capital. The opposition followed a different logic than during the preceding period, when what was at stake was the overthrow of the regime. More broadly, this case can be understood as a variation on certain historical cases, notably the Ottoman Empire, where armed dissidence by fringe elements was the opening gambit in the process of negotiation. [48]

21 For our part, we choose to limit ourselves to situations where the existing economy of violence is at stake (regime change or secession). Hence, it is not the level of violence that characterises civil wars for us—unlike the thresholds defined by the neopositivist—but the attempt to install a different economy of violence. This leads us to two observations. On the one hand, these are political actors explicitly organised as such, which differentiates them from the more informal practices of resistance that partake in an economy of violence. [49] On the other hand, we should not presuppose a fixity of intentions and goals. The aims of an armed group are mutable— sometimes global, sometimes local—and the same movement can transform its goals over time or pursue several simultaneously. Yet, there are few examples of passing from a logic of political contestation to a logic of economic accumulation—or the reverse—, even though certain cases (Sierra Leone, Liberia, Democratic Republic of Congo) seem ambiguous in this regard. Indeed, the 1990s thesis of a criminalisation of politics, as distinct from a movement’s use of illegal financial sources to achieve its political goals, smacks more of ideological discourse than rigorous research. [50]

22 Third, civil war is too often thought of as an an-historical phenomenon, an “internal war,” whence the frequent juxtaposition in the literature of examples drawn from the Peloponnesian War, the English Civil War, the Thirty Years War, or the Vietnam or Iraq wars. And yet Max Weber describes the formation of the state as a historical phenomenon, which should lead us to question the validity of a transhistorical definition of civil war. Besides, whereas Weber conceives the formation of the modern state as an individual trajectory, we see it as inseparable from the emergence of the inter-state system. Civil war is therefore an object that can only be set in a historical context by questioning the nature of the state and the international system during a given period. The stakes and dynamics in a civil war depend on the international system at a given time, whence the difficulty of making comparisons over extended historical periods. The effects of the international system on civil war can be difficult to isolate for a single case, but they become visible when civil wars are examined in series. In other words, contemporary civil wars exhibit structural resemblances that indissociably are effects of the international system and of the nature of contemporary sovereignty. It is therefore prudent to confine our study, at least tentatively, to a particular characteristic of the international system, namely the specific relationship to territory and international borders following the waves of decolonisation.

23 What are the systemic characteristics of contemporary civil wars? The first is that they are transnational phenomena: the quasi-totality of armed movements make use of a sanctuary in a neighbouring country, refugees are the universal products of these conflicts, and the non-military actors (international organisations, NGOs, private enterprises) step in systematically. The existence of a sanctuary, the role of transnational political organisations, and the management of refugee camps play key roles in the possible outcomes for an armed movement. The second trait is that the stakes orienting the practices of the parties in conflict are broadly defined by the international system. In fact, contemporary civil wars do not entail, as a rule, a modification of international borders; annexations, rare as they are, in practice are never internationally recognised, and secessions remain infrequent. [51] Since the end of decolonisation, territories have been more stable than states. In fact, contrary to the Eliasian model, in which the political centre progressively defines its territorial spread, [52] borders today have largely stabilised. Therefore, the state, even if absent, imposes itself as the key stake in the struggle, including for secessionist projects. In a much more structuring manner than commodities, resources linked to state formation such as control of the capital city, of currency, customs duties, and international recognition are critical. Even if they have little chance of success, authentically transnational dynamics (the Islamic State) are all the more interesting to analyse in that, a contrario, they permit an understanding of the dominant logic. [53]

24 This set of constraints and resources largely conditions the interactions of the actors battling for the state. Any stabilisation of the relations between armed groups, barring an external guarantee, can only be temporary, since the survival of one actor ultimately depends on it controlling or neutralising the political centre, which implies the elimination or assimilation of its competitors. The economic interests of military actors are not enough to create the necessary conditions for a stable division of the territory and its resources, which confirms the criticism levelled at theories that reduce civil war to the predatory exploitation by violent actors who avoid or, at least, minimise confrontation. On the contrary, anticipation of a durable reconstruction of the state brings with it an intensification of violent competition for control of territory, state institutions, and international recognition. The development of a bureaucracy is a key element in this struggle, since the ability to accumulate resources strongly impacts the probability of an armed group’s survival.

Attempt at a Definition

25 Setting aside the neopositivist paradigm, what do the social sciences tell us about situations of political violence that deeply affect the social fabric? Faced with multiplying civil wars during the 1990s, some researchers turned—perhaps more than before—to studying societies at war. This is especially true of anthropologists who saw the conflicts play out on their fields in Mozambique, Sri Lanka, Sierra Leone, or Bosnia. [54] Sociologists and political scientists tried to account for the transformations in societies confronted with political violence, for instance in Algeria, Afghanistan, the Democratic Republic of Congo, Bosnia, or Angola. [55] Numerous works on specific aspects of the transformations of societies rent by civil wars are extant, for example, on socialisation, violence, property, gender relations, segmentary organisations, and migration. [56] However, the fertility of these approaches should not obscure the difficulty of generalising from them, since the cases are not connected in a systematic manner. A methodologically sound comparison does not result from looking at cases in parallel, however many in number, or from a one-off theoretical borrowing, but from discussing them in a general theoretical framework. In this regard, the works of Georgi Derluguian, Klaus Schlichte, and Koen Vlassenroot are all the more important as they set out the bases for a comparativist approach. [57]

26 From this perspective, revisiting Bourdieu’s conception of the state could constitute a theoretical breakthrough. The state can be understood as a central authority for defining the relative value of various species of capital and relationships between fields. [58] Its retreat therefore entails a revaluation of the relative value of the various forms of capital, which draws our attention to alternative institutions that may guarantee them and to the reorganisation of certain fields. For example, the religious field may restructure itself around transnational institutions; the institutions linked to an armed group may (or may not) guarantee the value of economic capital and impose a new legal order. It should be noted that the societies researched may be structurally different from the Western societies in and for which the fundamental concepts of sociology were originally formulated. For instance, Bourdieu’s list of forms of capital (cultural, economic, and social) can be expanded by adding identity capital. [59] The relative weighting of forms of capital—made visible in their conversion—is not identical from one society to the other. Likewise, while the specialisation of fields is practically universal (Niklas Luhmann, Anthony Giddens, and Pierre Bourdieu), we do not assume any specific degree of independence for them vis-à-vis the rest of society. [60] In particular, the degree of autonomy of fields in non-democratic regimes is conditioned by the transversal role of the security apparatus and party-state relationships. In communist regimes, for example, the ubiquitous nature of the party limits the autonomy of fields (political, economic, cultural, and so on). At certain times in Turkey, the conflation of the party in power with the state has yielded the same result; [61] in Syria, the multiple security services played a similar role; [62] in Libya, the Gaddafi regime deliberately weakened institutions to the point where Libyan society functioned largely on an informal basis. [63]

27 Ultimately, we define civil war as the coexistence on the same national territory of competing social orders engaged in a violent relationship. By “social order” we mean an economy of violence, relative values of types of capital, and relations between fields, at varying degrees of institutionalisation. The national territory therefore sees the coexistence of different identity regimes, legal systems, and property rights. Competing social orders are most visible when armed movements control a territory, yet other modalities may prevail, such as day/night division of population control. [64]

A New Economy of Capital

29 The existence of competing social orders on a national territory implies the emergence of several economies of capital. Civil war thus affects social structure in three ways: the appearance of new forms of capital, an often-severe fluctuation in the value of existing forms of capital, and, lastly, new circuits for converting one type of capital into another. To begin with, the state’s withdrawal engenders the (re)formation of certain forms of capital, notably the development of military capital, made possible by the conversion of pre-existing competences or forms of capital. The new centrality of military capital changes the value and conversion of all types of capital, especially the modes of economic accumulation and identity regimes.

30 Second, economic capital, social capital, and identity capital are particularly affected during civil wars. In these three instances, the retreat of the state shows its role in what is a contrario routinely seen as spontaneous social functioning. To start with, the fighting changes the value of economic capital, especially because of currency and real estate price fluctuations. Next, territorial control by an armed group translates into a change in social capital, [65] for example, because gender relations or the performativity of the affiliation with segmentary groups (families, clans, tribes) are transformed. In general, individual social capital fluctuates widely in opposite directions: on the one hand, the majority sees its social capital shrink and, consequently, its ability to act. On the other hand, the militants, inserted into highly mobilised networks, increase their social capital. [66] Last, identity capital is impacted by a brutal denaturalisation of hierarchies between groups. To be Hazara in Afghanistan, Hutu in the east of the Democratic Republic of Congo, or Zaghawa in Sudan does not imply the same thing before and after the outbreak of civil war and by region.

31 Third, the withdrawal of the state shifts the circuits for converting one form of capital into another. Barriers of various kinds (economic, legal) that prohibit or impose a cost on crossing from one field to another tend to disappear, thus easing the conversion of capital. For example, in Afghanistan, starting in the 1980s, religious capital was converted into leading positions within political parties. [67] Additionally, in many cases, social capital appears essential to economic and military accumulation (Afghanistan, Syria, Libya). We may therefore hypothesise a greater convertibility of social capital into economic or military capital in situations of civil war. Moreover, the division of the national territory engenders the emergence of internal borders, which poses the problem of the differing value of forms of capital between regions. On the economic level, while internal barriers often become more constraining than the international borders, channels between territories can develop around certain resources (food, oil, gas, drugs, etc.). At the same time, the existence of identity regimes orients migration channels, as in the case of Afghanistan’s Hazaras. [68]

Competing Networks of Institutions

32 Control over a population by armed groups opens a space for setting up new institutions. [69] The prevailing perspective understands these institutions (rebel governance) from the perspective of controlling the population and providing public services. [70] We shift our attention towards the formation of fields—rather than an individual history of institutions—and the production of society, especially through law.

33 To begin with, the reconstruction of sectors of activity—health, education, justice, religion, economy, security—encounters three limits: objectivation, a dearth of resources, and competition. First, the objectivation of new institutions happens through recruiting specialised personnel and implementing set procedures often copied from state institutions (forms, maps, uniforms, official IDs). This can be observed among armed groups as disparate as the Taliban, the PKK, the Rally for Congolese Democracy-Goma, and the LTTE. The courts can rely on pre-existing institutions, including religious ones in the case of Islamist movements, or notables. Next, the dearth of available resources creates conditions of dependency on the outside (diaspora, foreign countries, transnational networks). Also, putting in place new institutions is complicated by the presence of competing institutions (linked to other armed groups), which alters the usually monopolistic functioning of sectors like currency, security, and law. Local monopolies may form, but, unless exclusive control is achieved, complex situations arise that compel forms of cooperation (formal or not) and accommodation, as in operating certain indivisible goods like electricity or water, or the recognition of legal documents between otherwise competing legal systems. Furthermore, beyond a single institution or even a specific sector, some movements succeed in re-establishing differentiated fields. This network of institutions consolidates itself by cross recognition between sectors, which favours their objectivation and, perhaps, their legitimacy. Justice here appears as a defining element due to its structural effects on other fields. Indeed, relatively sophisticated and independent legal systems affect the entire administration and activities in the territory controlled by the armed movement. To cite an example, the Taliban courts—because they enjoy functional autonomy and have, in part, achieved their enterprise of objectivation—legitimise the administration and military organisation, which, in turn, helps them function. [71]

34 Next, the institutions diffuse modes of governance and so participate in producing society. First, the administration of armed movements redefines the social fabric. For example, in handling conflicts over land or marriages, the courts affect property and relationships between genders. [72] More generally, the institutions transform social hierarchies by the classifications they use during their daily activities. Who will get priority treatment in a hospital—children or the aged, fighters or civilians, men or women? Further, civil wars are generally times when new models of governance are established. Given their lack of resources, armed groups generally welcome outside organisations (or homegrown ones organised on the same principles). Some armed groups allow, or facilitate, the presence of NGOs and international organisations that can take over entire swaths of everyday administration (the Sudan People’s Liberation Movement in South Sudan, Afghan commanders in the 1980s). Other movements, like the PKK or the LTTE, mobilise their network of sympathisers abroad to collect resources and organise NGOs that act on the interface between the armed movement, the population, and international actors. [73] Even the movements that a priori appear to be closed maintain contact with transnational networks, such as the madrasas in the case of the Taliban and diverse Salafist preaching organisations in the case of the Islamic State. The policies of these organisations can have important effects; for example, the physical layout of refugee camps induces changes in the status of women and in the modes of authority and personal standing. Moreover, the decentralised functioning and competition between these actors hinder the attempts at monopolisation by armed movements, leading to recurring tensions. These dynamics explain in part how government by armed movements sometimes represents a break with local history.

Dispositions, Skills, and Decisions

35 We counter the neopositivist model of the isolated individual with an agent who is socially grounded (gender, class, dispositions to act, and bodily hexis) and involved in daily interactions. From there, we consider three analytically distinct questions: socialisation, political context, and situation of interaction. [74] The agent calculates the risks and odds of success, but this strategic aptitude depends on his socialisation and available resources (information, time, money) enabling him to read the context and, more immediately, the situation of interaction. The agent deliberates, meaning that he generally makes his decisions following interactions during which he forms his judgment. Moreover, he arbitrates between different ends that may be collective or individual, altruistic or egoistical, which implies ethics and values. Finally, in contrast to a static vision of individuals and their preferences, the context transforms the dispositions, competences, and decision-making processes of agents. [75]

36 First, the emotional intensity of commitments, the multiplying traumatic experiences, and the participation in total institutions work to modify the initial dispositions of some agents. [76] To begin with, the exercise of violence, suffered but also committed and witnessed, has a transformative effect on the psychic economy, [77] as do total institutions, armed movements (PKK, LTTE, FARC (Revolutionary Armed Forces of Colombia), Islamic State), prisons or, to a lesser extent, refugee camps. [78] Next, the denaturalisation of the most mundane categories of daily life (religious or ethnic affiliations, gender, etc.) leads individuals to reassess their identity. These transformations translate into biographical ruptures that are difficult to predict from the social positions actors occupied before the war.

37 However, not all individuals experience such radical transformations of their dispositions to act, which may survive the institution that produced them because of continuing social recognition and socialisation. For example, some judges or military personnel display signs of their professional socialisation (bodily hexis, manner of speech, vocabulary, etc.). Probably the most striking case is that of doctors or members of the clergy, as social recognition of their competence endures even after the collapse of the health sector or religious field. Furthermore, some individuals who are less ideologically engaged or less exposed to violence maintain their dispositions largely unchanged—which reminds us of the great contrast in individual experiences across regions and periods. It is questionable whether the inhabitants of Kabul before and after 1992, or those of Latakia and Aleppo in Syria, really lived through the same war. Meanwhile, civil wars produce radically new interactions, in which existing dispositions to act do not allow a “transparent” adaptation to situations. Such difficulty to adjust is an immediate and brutal form of the hysteresis of the habitus. As a rule, this concept applies to socio-economic transformations (such as rural exodus) that also occur in civil wars, which involve accelerated migration and urbanisation. [79]

38 Second, the war leads to, and sometimes compels, competences differentiated by the frequently changing positions of agents: body techniques, [80] mastery of objects and tools, specific social skills, and the interpretation of situations. This is evident in the case of civilians turned fighters (knowing, handling, maintaining weapons, tactical skills, physical discipline) [81] but in reality it affects the entire society. For example, restrictions affect ways of cooking; fighting compels concealment in order to move and to protect oneself, and the interpretation of sounds and smells; the incidence of casualties leads to learning simple first aid techniques (possibly in a care structure); militancy leads to mastering specific communications and administrative tools. As has often been noted, some recycle past competences: for instance smugglers, urban gangs, secret societies, or ex-military personnel. More broadly, organisational abilities, developed against or outside official institutions, are regularly put to use in crisis situations.

39 Finally, decisions are social processes that are impacted in crisis situations. These processes are complex, but we can approach them through three dimensions—calculation, emotional investment, and collective deliberation—all of them affected by the onset of civil war. [82] Here it is worth emphasising that situations can vary significantly, from disorganised phases with inclusive discourses to front line war between territorialised armed groups relying heavily on a network of institutions. The hypotheses we suggest evidently do not apply to all situations; we aim primarily to characterise non-routine situations in which institutions are weakened. First, incertitude is heightened when the institutions are no longer capable of playing their role as reducers of uncertainty. Yet, routines are an important dimension of daily life: most actions in fact are non-strategic behaviours. [83] Furthermore, crisis situations lead to a denaturalisation of the social world and a mismatch between “incorporated structures” and situations of interaction. This tension is unconducive to a “regression to the habitus” as hypothesised by Michel Dobry. [84] Agents therefore become caught up in a calculating hyperactivity, because the risks are mounting and institutional routines attenuate or disappear. The actors are forced to anticipate the consequences of their actions and to inform themselves. [85] The most mundane actions, like sending the children off to school, going shopping, or quarrelling with a neighbour, can have serious consequences. For example, which jurisdiction do you appeal to for filing a complaint? Then also, the emotional investment is especially intense: agents may adhere to a cause and the stakes are existential (personal or family safety). Finally, individual decisions generally have a more pronounced collective dimension than in routine situations. The group dynamics become key in the exchange of information, risk and reward assessment, and making sense of events. Moreover, in certain phases, civil wars may see the reconstitution of routines and predictability as agents sometimes decide to submit to an institution (an armed movement, or a refugee camp) and uncertainty can also be reduced by the formation of new institutions or the stabilisation of front lines. [86]