Dynamics of Violence: Processes of Escalation and De-Escalation in Violent Group Conflicts [1 ed.] 9783428499571, 9783428099573

Citation preview

Dynamies of Violence

Beihefte I Supplements to "SOCIOLOGUS" Zeitschrift für empirische Ethnosoziologie und Ethnopsychologie A Journal for Empirical Ethno-Sociology and Ethno-Psychology Herausgegeben von / Edited by Georg Elwert

Heft I Number 1

Dynamies of Violence Processes of Escalation and De-Escalation in Violent Group Confficts Edited by Georg Elwert Stephan Feuchtwang Dieter Neubert

Duncker & Humblot · Berlin

Die Deutsche Bibliothek - CIP-Einheitsaufnahme

Dynamics of violence: processes of escalation and de-escalation in violent group conflicts / ed. by Georg Elwert ... - Berlin : Duncker und Humblot, 1999 (Sociologus ; Beiheft; 1) ISBN 3-428-09957-5

Alle Rechte, auch die des auszugsweisen Nachdrucks, der fotomechanischen Wiedergabe und der Übersetzung, für sämtliche Beiträge vorbehalten © 1999 Duncker & Humblot GmbH, Berlin Fotoprint: Berliner Buchdruckerei Union GmbH, Berlin Printed in Germany ISSN 1438-6895 ISBN 3-428-09957-5 Gedruckt auf alterungsbeständigem (säurefreiem) Papier entsprechend ISO 9706 9

Acknowledgement A book Iike this is always a collective project. In this case the publication is based on two conferences held in Berlin and Mainz in 1995 and 1996 with the title "Feud, war and genocide". During these conferences around 40 papers were presented. This book is a selection with a focus on the "dynamics of violence". However, conference papers dealing with other subjects inspired the outline of the book lind the papers presented here. Therefore, we would Iike to thank all participants for their contributions and those who participated in the planning and organisation ofthe conferences. We want particularly to name Dr. Artur Bogner who was an active and stimulating member of the working group that prepared the conferences and who influenced the development of the project's central idea. The conferences were a joint activity of the "Section development sociology and social anthropology" of the "German Society for Sociology", the "BerlinBrandenburg Academy of Sciences" and the "Forum against ethnic violence". We are grateful to those institutions that supported the conferences financiaIIy: "Berlin-Brandenburg Academy of Sciences", "Friends of the Mainz University" and "Erika-Sulzmann-Foundation". The latter also supported the publication of the book. Last but not least, we would Iike to thank Sandy Gray, David Shaw and Carolin Leutloffwho supported us in copy-editing and proofreading. Berlin, London, and Mainz Georg Elwert Stephan Feuchtwang Dieler Neubert

Contents Georg Elwert. Stephan Feuchtwang. and Dieter Neubert

The Dynamies of Collective Violence - An Introduction ............................................ 9

I. The Logic of Violen ce - Theoretical Approaches Trutz von Trotha

Forms ofMartial Power: Total Wars, Wars ofPacification, and Raid. Some Observations on the Typology ofViolence ...................................................... 35 Peter Waldmann

Societies in Civil War ................................................................................................. 61 Georg Elwert

Markets of Violence ................................................................................................... 85 Jürg Helbling

The Dynamies of War and Alliance Among the Yanomami. ................................... 103

11. Auto-Regulation ofViolence and Escalation Erdrnute Alber

Violent Conflicts in West-African Borgu on the Eve of Colonisation ..................... 119 Stephanie Schwandner-Sievers

Humiliation and Reconciliation in Northern Albania. The Logics of Feuding in Symbolic and Diachronie Perspectives .............................................. 133

8

Contents

Dieter Neubert

Dynamics ofEscalating Violence .......... ..... ........................ ................ .......... ............ 153

III. De-Escalation and Treatment of the Consequences of Violen ce Tim Allen

War, Genocide, and Aid. The Genocide in Rwanda ..... ........ ...... ........ ........ ............ .. 177 Gerhard Grohs

About the Role of Churches in the Peace Process in Africa and Central America............ .... .. ...................... .. .... .. .. ............. .. ......... ....... ........ ............... 203 Heilre Schmidt

Neither War Nor Peace: Making Sense ofViolence ................................................. 211 Ivo Strecker

The Temptations ofWar and the Struggle for Peace Among the Hamar of Southern Ethiopia........... .. .. ........................................ ...... ........................... ..... .... 227

Bibliography ........ ...... ................ ........ .. .. ....... .. ........ .. ... ........................ .... .. .... ... .... .. ....... 261 Contributors ......................................................................... .......................................... 290

Tbe Dynamics of Collective Violen ce - An Introduction By Georg Elwert, Stephan Feuchtwang, and Dieter Neubert "Collective violence" is part of social reality all over the world. This is reflected in social science. Recent debates in social sciences on ethnicity, or social movements, or democratisation, or the transition from Communism, all must also deal with violent conflict. However, in these debates violent group conflicts are seen as phases or elements of other social processes. Even in peace and conflict research, collective violence is analysed in terms of its causes and consequences and not as a social phenomenon as such (against this kind of sociology see von Trotha 1997 and Nedelmann 1997). Collective violence is still a neglected subject in sociological and anthropological research in its own right. Engagement in violence is not a regression to atavistic instincts as is so often held in over-convenient analyses. lt is rather a narrowing ofthe available forms of action and at the same time it is a strategic choice. Once we understand violence to be an everyday possibility ofthe implementation ofwill, we can go on to analyse its social channels. Violence is always channelled. Particular targets, victims, weapons, battlefields, and times can be fixed beforehand, other targets and forms of battle can be excluded. The concepts of social channels of violence imply that violence is to the same extent also constrained (cf. Sieferle 1998, Althoff 1998). In feuding, which is the most severely channelled form of violence, weapons, the potential victim, the place of combat (in New Guinea often just the one fighting ground), and the time can be prescribed. Even the procedures to end hostilities are culturally defined norms. The end of the battle is part of the program. Research into such containers of violence, such 'inhibitors', can help to implement peace projects. Hopes that the end ofthe cold war would bring peace to the world have been deeply disappointed. The contrary has been true. Ofthe former Soviet bloc only the countries of East-Middle-Europe have undergone a mainly peaceful transition. At the southern fringe of the former empire a multitude of violent local conflicts have been the heritage of Soviet supremacy. Africa, even during the cold war, was burdened by numerous violent conflicts. With the end ofthe cold war, international military support for the conflicting parties has diminished, but wars have not ceased. The civil wars in Somalia, Liberia, Sierra Leone,

\0

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

Tajikistan, Rwanda and Congo/Zaire, to name but a few, started after the end of the cold war. To know why violent group conflicts are such an important element of our world we must try to understand the conflicts themselves as social phenomena. Then the analysis of collective violence can be integrated into more general analyses of society. The objective we follow here is to understand potential conflict situations, to see what are the processes oftheir escalation and what are the regulating structures and mechanisms of violent conflict. In short, our objective is to deal with a sociology and a social anthropology of collective violence. In one sense this book is a battle against psychologisation and culturalisation. A crucial sector for the understanding of "collective violence" are processes of escalation and de-escalation. Escalation as a topic includes analysis of phases of collective violence once it has been initiated, whether towards greater violence or just towards endemic violence. Escalation can be understood as a distinction between more and less regulated violenc~ practised by the same people but in different circumstances. Escalation is a move, for instance, from feuding between descent groups to treating them as strangers and raiding them. It can also be conceived on an historical scale, as a shift in the kind of wars waged, in particular, from wars of mercantile and agrarian empires to the wars ofmodem states and their civil wars. De-escalation means the reverse processes of limiting collective violence. Completely successful de-escalation leads to reconciliation and long-Iasting peace. But de-escalation also includes reduction of collective violence to a lower level, or to the establishment of greater control of violence. 1. Types of Collective Violence We look at different forms of collective violence and at clearly defined cases. They deal with a diversity of forms and modes of collective violence that vary also in historical terms. The examples stretch from feud, raid, civil war, war of colonial pacification, war between nation states up to genocide. Every case has its peculiarities. Therefore, the analysis of every case of collective violence has to start in an historical context, though some of the characteristics of war suggested by von Trotha and some of the economics of violence outlined by Elwert may be applicable in a great range ofhistorical contexts. None of the cases discussed in this book include the two world wars in our century. They would not be good cases for our study because they are, from an anthropological perspective, a cluster of composite wars with quite differentiated characters (compare, for example, the Finish with the Yugoslav war). They would have to include different forms of collective violence, extend over different parts of the world following different developments and political and military objectives. They would be complex social events and, to take them as a whole would hinder a systematic understanding of collective violence.

The Dynamics of CoJlective Violence - An Introduction

11

Something else we have not done is provide a typology of wars and violent conflicts because this would not render the transformative or sliding potential of each of such a "type". But some distinctions are nevertheless helpful. One is the crucial distinction between collective violence or war as anormal state of affairs, namely the regular and endemic exertion of force not only for protecti on and survival but also as an economy, and war as an exceptional state of affairs, in which the normal economy includes readiness to wage war but not warring itself. Examples of collective violence as anormal state of affairs inc1uded here are cases of recurring raids, feuding or warlordism or long lasting civil wars (Waldmann, Elwert, Helbling, Alber, Schwandner-Sievers, and Strecker). In these cases we see recurring escalation and de-escalation. They follow a typical pattern of alternating appeasement and violence. The level of violence oscillates between outbreaks of violence and phases of controlled violence influenced by social forces that strengthen the oscillation. Therefore we will refer to them here as cases of "oscillating violence". We use also the word 'warre' für normal as distinct from exceptional war, borrowing from the seventeenth-century English spelling of Hobbes and following the example of Marshall Sahlins (1968, chapter I). Typical examples of violence as an exceptional state of affairs, or "warcatastrophes" in a systemic (not a normative) sense, are wars between nation states or a civil war that leads to an end, either by victory or peace negotiations. Local conflicts might also follow this pattern. Escalation and de-escalation are single processes that come to an end with the end of the war or conflict. We will name such cases "exceptional collective violence", "exceptional violence" or "war-catastrophe". On a theoretical and comparative level von Trotha deals with exceptional violence in total wars and colonial wars of pacification. An extreme case of exceptional violence is the Rwandan genocide (Neubert). Beside the crucial distinction between "oscillating violence" or "warring" and "exceptional violence" or "war-catastrophes" AIIen's paper reminds us of the distinction between external and internat collective violence. External collective violence points to "wars between internationally recognised nation states". Internal collective violence refers to conflicts "occurring within a social grouping". Civil wars are a special case of internal collective violence where the objective of the conflict is the control of the national government. Other types of internal conflicts take place without government involvement and without the objective of control of the government. Examples are raiding, feuding or fighting on a limited local level in a nation state (SchwandnerSievers, Strecker). External war has effects on the internat situation and produces internal war conditions. The occupations of France, Italy and Norway by Nazi-Ied forces of Germany produced Quisling and partisan divisions, not to mention the warring divisions of resistance themselves as they did in China under Japanese occupation.

12

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

Collective violence whether oscillating or exceptional mayaIso be differentiated according to the role of civilians or non-fighters. Firstly, there are two types of conflicting parties: groups including women and children versus groups including only fighters without a relationship to the population. Secondly, civilians may be protected and spared by a special warrior ethic or rules like the Geneva convention (though this normative protection works less weil in practice than in name); or civilians are preferred targets for attacks; or they are mobilised as supporters in the war. In cases where complete groups including women and children are involved in the fight every group has its own fighters for protection who can be trusted. Violence is based on political or tactical reasoning (Helbling) or on ideology (von Trotha). And it is used in the real or imagined interest ofthe group. Examples are the Yanomani (Helbling), and the conflicts in Southern Ethiopia (Strecker). I A more frightening situation is the existence of pure fighter groups. The difference between enemy and own population is absent or diminished. Fighters are a threat not only to the enemy (fighters and families) but also to their own people. Their violence can be commercialised and used for private purposes, or they can be mobilised in a politics ofhatred. They threaten unprotected civilians, as in the Rwandan genocide (Neubert), or in cases ofwarlordism (Elwert, Waldmann). The articles presented here are not a judicious representation of all these different cases of collective violence. The focus is mainly on internal conflicts with a bias toward oscillating violence. Oscillating violence is over-represented because this type contradicts the modem ideals of everlasting peace and a complete control of violence. The cases analysed here show that oscillating violence is possible because it is embedded in an historical context and in a society and an economy. The Rwandan genocide (Neubert) is a case of escalation from oscillating to exceptional violence. Waldmann's analysis of Colombia, on the other hand, shows a process of stabilisation of what might have been an exceptional civil war into astate of oscillating violence. 2. Warring or Oscillating Violence as·a Normal State of Affairs It has been part of the myth of the modem system of states that oscillating violence is a thing of the past. Yet, as some of our studies show, oscillating violence is still important in modem nation states.

We can specify at least three of its conditions: - small distance between conflictual parties or an intensive contact through trade routes,

I Arthur Bogner in his paper for the first of our conferences in Berlin pointed out that the process of ethnic we-group formation and collective violent conflicts are closeJy interwined (see: Bogner 1998).

The Dynamics of Collective Violence - An Introduction

13

- no monopoly of violence (before the establishment of states or after their disintegration), - space at the periphery or in niches of empires or nation states which are open to the use ofviolence. Yanomami warring (Helbling) suggests the importance of increasing densities ofpopulation within a given dependence on land. Warring is greatest where the Yanomami are most densely settled. The availability of land allows for an avoidance of raids and wars; conversely, the risk of separation from a crop and the unavailability of new land for sowing a new one increases dependence on raids and wars. Another historical condition suggested by Helbling is that the intensity of warring increased with trade in iron tools and weapons. Trade in any case accompanies Yanomami provisional alliances. In Borgu, as suggested in the chapter by Alber, trade itself was the precolonial condition of warring for booty and the control of trade routes. It was a condition which arose in the fifteenth century and ended with the imposition of European colonial states in the nineteenth century. Peasant cultivators needed protection by and from a class of warriors and warlords who Iived on their tribute or booty (the term to use depends on the phase of imposition, replacing a previous protection). Booty included the taking of slaves to be sold, whereas the Yanomami aimed to kill men and capture women, a human reproductive war rather than a war over trade and tribute. A further distinction between the two is that Borgu warlords formed a political hierarchy reaching from village chiefs to a lord in the most important trade station in Borgu, whereas the Yanomami are acephalic ("segmentary"). Borgu economic diversity was far greater and translated into religious and ethnic distinctions: apart from the majority Baatombu and Boko peasants, there were Gando peasants, Muslim town traders and craftsmen, and Fulani cattle herders and their Gando slaves, all under the authority of Baatombu military families who took slaves from villages of both their own and Gando language groups. Two such different cases suggest that astate of war as a norm rather than as an exception may have been usual in many different regions and different economies. We can certainly add empirical evidence to Rousseau's refutation of Hobbes' description of the state of warre (that is to say oscillating violence) as a natural, rather than a social state. But is warring unusual in the longest era of human social history, which is that of hunting and gathering? Does warring start with shifting cultivation and pastoralism, and intensity with trade, though these are not sufficient conditions for it to occur? Warre, in any case, is not an era in human history, it is a condition which can exist in several kinds of production and trade economies. If we were to return to the Yanomami and guess that raiding existed before the Spanish and subsequent encroachments on the territory which included them, we would then have to include hunters and gatherers as possible protagonists in oscillating violence. But let us turn instead to the Hamar who are hunt-

14

Georg E1wert, Stephan Feuchtwang, and Dieter Neubert

ers, pastoralists and have, in this century, increasingly engaged in shifting cultivation. Strecker's chapter is about peace-making and the Hamar idea of war caused by madness. When they have a sec ure subsistence base, peace for the Hamar is the best and the rational state of affairs. But Hamar also entertain the idea of well-conducted raids or wars, and sober leaders have conducted them. Peace seems to have included such c10sely regulated violence. Successful raids were celebrated as great achievements, triumphs of reason, sobriety, superior knowledge and luck (Strecker, personal correspondence). In other words, drought, the import of arms, the stresses of Iiving on the margins of astate and its markets, the insecurity of a subsistence base have increased 'mad' war, escalating beyond sober and regulated raids. There had been oscillating violence but it has become increasingly mad. Another example of an escalation of oscillating violence can be seen, not only in the extension by the emergent Albanian state ofthe bese into a patriotic rather than a feuding tradition, but also of the loyalties of bese into something similar to mafia among North Albanian migrants to foreign towns and cities (Schwandner-Sievers). An arguable hypothesis is that oscillating violence exists in the large fringes of empires and on the borderlands and marginal conditions, sometimes in the heartlands of states. Let us consider empires of oscillating violence. Before the North European colonial orders imposed their peace upon Borgu there were other states of oscillating violence in other kinds of empire, not only in Africa but in Asia and in Europe itself. The pax Romana, the pax Mughal, the pax Sinica and so on, not only had their central agrarian territories. They also had huge fringes. One of them included the feuding patrilines of the northem mountains of Albania in the Ottoman empire. But these do not appear often in the standard histories. What the great chronicles of Roman, Chinese, Mughal, Holy Roman, Papal, Venetian or Ottoman empires, or those of the Romanov, Habsburg, Bourbon and other houses record are extensions and continuities of bounteous peace through marriage, succession, cunning, diplomacy, military prowess, and holy virtue. Their chronicles are histories of exceptional war and conquest, and of established peace as seen from centres ofrule, which are centres ofthe greatest mustering ofmilitary force and of protected plains and valleys of agrarian production. These centres are sacked by rivals, or drained of resources in order to wage war. But siege and sacking are episodes of devastation, which are exceptional. On the contrary, seen from their hinterlands and from the shifting frontiers of imperial rule, the establishment of peace and the episodes of violence are not so dramatically contrasted, un less the point of view is from the paths of great warriors. These large hinterlands should be described as permanent states of raiding and brigandry, tlaring up into local wars, in short ofElwert's warlord economy. The Yanomami were indirectly connected to an agrarian empire broken up by conquistadors, then distantly connected to a post-colonial state which steadily decreased the land available to them for shifting cultivation. They are a case of oscillating violence at the edge of a nation state. Borgu warlords represent

The Dynamics of Collective Violence - An Introduction

15

oscillating violence in mercantile colonies whose most peaceful territories were the agrarian lands of the most powerful military organisations. In either case, fading edges of colonies and mercantile empires touch and overlap with the edges of other empires. This makes for less secure portions of trade routes, maritime or landed. Peace may have been violently imposed by modem colonial states but these very modem states were also centres of (slave) trade with their own hinterlands offar less security. They not only contributedto the insecurity - they fought off at other moments - they also profited from it. Oscillating violence is seen as the weakness or absence of an enforcer of peace, and therefore, as Waldmann puts it, as a necessary readiness to organise in self-defense. Surely this can also be historically specified as a condition placed at a distance from centres of the exertion of strong states and their trading and productive interests. One contemporary equivalent is the weakly maintained state financed by a trade in arms and in drugs, trades which find their biggest agents and outlets in the margins of strong states. Waldmann 's example is the state of Colombia, and its trade of coal, oil and cocaine with the USA and Europe. We couId add Somalia (see Schlee 1995) Myanmar (Burma), Afghanistan and the adjoining fringes of Thailand, China, Laos and Vietnam, or the borderlands of many other states' territories. Recurrent resort to violence is easy to describe. A long lasting peace is not in the interest of the conflicting parties. The level of violence may change, but as long as both parties can keep their power the oscillation between violence and limited peace or protection will carry on. The question is how oscillating violence is supported or regulated. Two patterns of regulation are presented here. The pattern of preventive strike combined with changing alliances and feasting (Helbling, Alber) and the pattern ofmarkets ofviolence and protection (Elwert, Waldmann). Borgu and Yanomami are obviously very different cases of warring as a permanent state of affairs. Yet there are features of the dynamics of violence which they share. One is that the territorial extent of protection depends on the actual exercise ofviolence. Territorial control has constantly to be reasserted by the exercise of raids. Another is that the capacity to wage war is always tied to a strategy of alliance and feasting. Alliance is a strategy to gain strength in relation to a third party, and is therefore always prone to the seeking of alternative advantages by partners. Yanomami alliances are usually between unequal partners, from which the weaker seek ways of removing themselves. Helbling develops a game-theory analysis ofthis strategy. He also highlights their wellfounded suspicion of treachery. Like Scottish clans until the seventeenth century imposition ofEnglish rule, the Yanomami use feasts for alliance as lures to trap and kill an enemy.2 The Yanomami retain their reputation in anthropology 2 The most famous Scottish treachery feast was also the last, in which those Campbeils who were allies of the English invited the MacdonaIds to a treacherous feast in 1662. .

16

GeoTg Elwert, Stephan Feuchtwang, and DieteT Neubert

for an extremity of distrust and ferocity. Nevertheless, they also needed alliances as a tactic for gaining strength against a third party. Feasts and alliances in pre-colonial Borgu are described by Alber for their function of converting booty into prestige and prestige into might. But the Yanomami strategy of alliance may weil be applicable to Borgu warlords. In reverse, the ceremonial competition of feasting and gifts, though more elaborate in Borgu, is as much a feature of the state of warre among the Yanomami. The basic pattern of markets of violence and protection is equally cIear. The capacity for violent attacks incIudes the ability to protect others against violence. Those who have the capacity use it in two ways. First, they appropriate goods and money by raids. Second, they seil protection. This may be protection from other armed groups but it is also protection from their own attacks. Violence and protection are commercialised goods, they are a commodity. The business of a violen ce entrepreneur is based on the oscillation between the threat of an attack and the hope for peace - warring creates its market dynamics. Borgu warring shares elements of the Yanomami pattern and that of violence markets. That there is no or only an incomplete monopoly of violence characterises, as mentioned above, contexts conducive for warring. "Monopoly of violence" "Gewaltmonopol", a term coined by Max Weber {I 922) as a defining parameter for the concept of state, designates not only an important difference between power structures but also between potentialities for collective violence. If the use of coercive force is monopolised, there is a chance for stabilising realms free of violence. The German term "Gewalt" was translated by Gerth & Mills (see Weber 1948) and Secher (see Weber 1962) by "force", in German "Zwang", possibly because "Gewalt" also has a connotation of authority. The modem carreer of the concept is, however, Iinked to its further development by Norbert Elias (I978) who stresses the importance of realms free from arbitrary violence ("gewaltfreie Räume") as a correlate to the monopoly of violence. For the population this creates an attraction toward and legitimacy of this power model and thus builds up a basis for mass mobilisation. The monopoly of violence allows also for the unchallenged build-up of destructive force and its concentration upon extern al wars which will be the subject ofthe next section.

3. War Catastrophes, Exceptional or Escalating Violence Authority systems without a monopoly of violence can as weil generate a lot ofviolence - especially an internat one - but they are less likely to be caught in a violent competitive race with neighbouring powers (see Janssen 1991). Under condition of an effective monopoly of violence peace appears as a normality and war as a catastrophe. Such a war is also a catastrophe in the systemic sense ofmathematical (catastrophe-)theory, insofar as it creates new structures which

The Dynamics of Collective Violence - An Introduction

17

cannot revert to the fonner ones by a simple change of a parameter as is the case with oscillating violence. Von Trotha's chapter on war covers the whole of ancient, medieval and modem European history, including its imperial wars. It makes a sharp distinction between war-catastrophes (exceptional violence) and what we call oscillating violence or warring. Von Trotha is mainly concerned with what he calls total war, or war in its own tenns. War as such has one main aim, wh ich is defence from or defeat of an enemy, or at least elimination of the threat of an enemy. The enemy is its object. Oscillating violence, or warre, on the other hand always has other and more limited objects. Von Trotha cites the ancient Greek martial ethics of fighting to the last free man as characteristic of total war. To Aristotle and his Enlightenment followers a slave is one who prefers life to liberty. The other characteristics of total, or exceptional war are equally ancient and true of non-European traditions of war. One is collective mobilisation and organisation to do material hann to an enemy, wh ich demands a rigid demarcation of inner from outer morality, with strong ideological mystification and mythology. Another is glorification of violence and indifference to the suffering of its victims, ineluding those of the warriors' own society. G lorification is also a heightening of imagination and sensuality within fantasies of violence and an obsession with threat and superiority. A further characteristic noted by von Trotha is the reorganisation of time, in which war becomes its own priority and condition. We will return to this important feature, which is the defining characteristic of exceptional war. But here we should note again the distinctiveness ofwar from warre (oscillating violence). Warre shares some of the ethics and mythology of honour and shame which total war engenders. Both are organised and in that sense collective. But warre is targeted and limited by the aim of economic gain or the protection of subsistence and reproducti on by raids, whereas the object of exceptional war is defeat of an enemy. Between war and warre, von Trotha places the colonial wars of European pacification, selecting for examples Gennan occupations ofwest and southwest Africa. He asks whether these wars ofpacification were total wars. Conquest of territory may be seen as a limitation of total war, and von Trotha coneludes that wars of pacification which are a kind of conquest are therefore not total. They are exceptional but not total. But he leaves this conelusion open. Their object is a territory, not an enemy. On the other hand the subjugation of a population to something like if not actually a population of slaves, is elose to being total. Catastrophic war or exceptional violence is escalated violence. So a central question for analysis is how escalation starts. Exceptional wars, whether civil or extemal, whether total or colonial, seem to start with a deelaration or, if internal (civiI), with a sudden escalation, a volcanic explosion of violence. Waldmann's example is the bogotazo of 1949 in Colombia, Neubert's is the four months of genocide in Ruanda in 1994. Of course there are reasons and long periods building up to the explosion of pent-up hatred and fear, just as 2 Sociologus, Beiheft I

18

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

there are political manoeuvres and ultimata before the declaration of an external war. When we focus on the dynamics and phenomenology of collective violence, we find this moment described by participants as a sudden switch. Neighbours who have been friends for decades, even centuries, become enemies over the course of a few weeks. It is the poignant moment recorded in the film by Tone Bringa in a village near Sarajevo (We are all neighbours, Granada Television, 1993). An old lady calls out to another old lady with whom she had, all her adult life, drunk a cup of coffee in the morning, but the other old lady will not hear her. Neither of them can explain what has happened. Neighbours have become a threat on different sides in militia defence organisations. The propaganda which has bred such paranoia over weeks or months is part of the moment. It is a switching of attention from ordinary ambiguities, divisions and tensions to stark fear, victimhood, and heroism, about which anyone has images and intuitions fed by stories and rumours, if none have been directly experienced. It is also a switch ofleadership to a war-like militancy. Above all, war is a switch of identification processes to a restricted and simplified one of loyalty to a fight leadership. The dynamics of war are also the dynamics of group identification. Here we can learn from the chief characteristic oftotal war noted by von Trotha: that it is a rupture which reorganises time and the orders of past, present and future into a situationalised immediacy. Exceptional war inverts the primacy of the everyday as ordinary and long-term to the everyday as extraordinary and catastrophic. In the new time of war, everything is set in the drama of a present threat to life. The noise of a plane passing above is not simply the trafflc of the air; it could be the herald of a rain of fire. To fetch water is to risk being shot. The suddenness of attack as experienced does not hide its social construction. Violence has to hide its societal architecture. Violence is staged suddenness. Surprise eases the success of violence. In order to be efflcient violence should not be announced. The early warning period should be shorter than the time necessary for the planning of defence. If strategic motives are hidden, or covered behind motives outside the realm of discourse, then resistance as weil as a potential peace-creating dialogue are rendered more difflcuIt. That violence appears as an outbreak is intended. Staged as a spontaneous reaction or as an expression of emotion, the practice of vioIence reduces its need for Iegitimisation. A strong force engendering legitimacy and total mobilisation is ideology (von Trotha, Neubert cf also Orywal 1997: 40 on values and conceptualisation). Ideology reorients political thinking, changes categories of good and bad and motivates people to sacrifice their own interests and lives for the "nation", the "tribe" or another imagined community. Escalation is driven by hope of a 'just victory". Exceptional violence creates new memories of collectivity. It may, in the retrospective tense used by a fight leadership, be a call for sentiments of unity sanctified in a hallowed and natural past. Sometimes war is a technique

The Dynamics of Collective Violence - An Introduction

19

by a beleaguered leadership to divert attention from internal divisions, and to recreate a given social solidarity. Certainly it is not 'given'. On the contrary, war is a moment in the creation of a new social totality. For convenience, we refer to the groups as if they were continuous and we name them as if they remain the same. But war is a particularly vivid process of identification and reidentification. A society's unities are mobilised by being defined against enemies who threaten life and collective existence. Such processes of identification are particularly powerful and memorable precisely because they are created around a contingency of dying and killing. Each day could be the last. The shared victimhood, shared sacrifice, shared privation, create a kinship. When it is a modem war, a war of mass involvement, whether internal or externaI, the process creates boundaries of identity on a larger scale than any other or any previous segmentation, differentiation or creation of ethnicity. Political loyalties have been engendered with the same passion as confessional or ethnic identities, as is evident from revolutionary civil wars (Mexico, Colombia and Sri Lanka are examples mentioned in the articles here). All these identifications are localised and to some extent also ethnicised, as claims or counter-claims to the same ethnicity, or as local traditions. These become the big loyalties of war catastrophes. But we must also remember that every civil war is also a war of petty vengeances, running along alliines ofsocial division and envy. By the time of the switch out of war time, only some of the population and parts of the territory of astate will have been mobilised into the violent dramas of death and collective identification. Even the genocidal explosion in Ruanda was concentrated, so that, as Neubert remarks, areas of the territory were not immediately drawn into it. But the longer the time of war, the more likely its inclusion of hostile and threatened identification will spread, so that, as in Ulster, lines are drawn across every town and village, or between towns and villages, as well as between city neighbourhoods. In the switch which ends an exceptional war back to a time of banal normality, its divisions, ambiguities and blurring of boundaries, the deaths and memories of that other time, the time of war, become collective mythology and history to which future moments of heroie identification and defence will turn for whatever creation of collective identity a militant leadership will mobilise them. That future will be another mobilisation, against an 'ancient' enemy redefined in a new political situation. This perspective on collective violence as an exceptional state of affairs contradicts the analysis of causes of war. Asking for causes of war assurnes that a certain set of causes led inevitably to escalation. In such a perspective war is the automatie result of special conditions. Here exceptional violence is seen as occurring at the end of a process of escalation. This process does not proceed automatically, it is influenced by the people who take part and try to control events. Starting from a certain situation in so me cases the process ends in war, in other cases not. The process is open to change, participants can change di2*

20

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

rection and may return to a non-violent situation. Analysis can show a situation that bears the potential of escalating violence. We can point to elements that push escalation and to elements that hinder or inhibit escalation. But we cannot definitely predict whether a war will start or not. Examples of recent cases show that violence can be fenced in. After the break-up of the Soviet empire many conflicts led to war (for instance in Chechenia, Georgia). Other conflicts did not escalate to a war. The declaration of independence of Lithuania provoked the Soviet Union. The conflict escalated and the Red Army marched up to force the newly established Lithuanian government to retreat. The Lithuanian people came to support their government, practically without arms. The Red Army started to attack but they realised that the Lithuanians were ready to enter a fight. The more prudent forces in the Soviet government drew back and stopped the attack. The conflict was resolved in complicated negotiations. Nearly the same pattern of a conflict escalating to the first shots and a lastminute turn to negotiations happened at the failed coup d'etat in Moscow in 1991. The lesson is c\ear. In principle, an escalating conflict can be de-escalated even when the violence has already begun. However, de-escalation gets more and more complicated when the fighting has already begun, when exc\uding boundaries have been defined and the conflicting parties are hoping for victory or fear total defeat.

4. The Motives and Aims of CoIIective VioIence Those who associate violence with limited rational control and high output ofadrenaline misconceive the social character ofthe dominant form ofviolence in human communities: strategic violence. Strategic violence was a very important development in the evolution of hominids. In this evolution spontaneous exertion of violence lost its place to instrumental and strategie violence. Instrumental violence is violence diligently meditated and prepared. Strategic violence is violence whieh uses other persons in an organised plan. For these persons the use of violence is a highly risky choiee. What then brings a man to make this choiee? Trying to answer that question it is seductive to follow declared intentions. This practice - often observed in anthropology can be a pitfall. We become integrated as actors in a staged game where there is a need for a phenomenologieal distance. Violence uses the difference offa~ade and structure in order to create staged suddenness and, thus, to paralyse the construction of resistance. Camouflage is not only a military tactic but also a basic skill in the social construction of violence. Constellations of violence are characterised by processes of self-organisation by which the motives of the actors get changed. Violence can create its own motives. This is a problem for research since most actors are able to define a motive for action as a goal even when their action itselfwas erratic and contingent.

The Dynamics of Collective Violence - An lntroduction

21

The analysis of violence is easier in those cases where we face stable patterns of violence. Stable patterns appear whenever there are feedback processes which strengthen types of violence based upon socially organised chains of action. There is a division of labour in these chains of action wh ich create an image of atomised activities. On the other hand the concentration on specific targets, which are amassable parameters, allow interpersonal understanding. These parameters are power, prestige and goods. It is that violence wh ich stabilises power, which allows the attribution of prestige and which organises the appropriation of goods which commands our specific interest. The more decisive contested resources (power, prestige or goods) appear, not just for the well-being but for the very reproduction of the social and economic situation of someone, the more probable is the decision to use violence. These resources may appear scarce but they can be accumulated. Let us compare the case of prestige: if it is battle which produces prestige, then a war can increase the attribution of this reputation. If the war or a warring situation lasts, then the type of reputation called heroic first may increase its relative weight in respect of other types of reputation. This can but does not have to produce an inflation ofhonours. The death ofthe heroes may have an anti-inflationary result. War is of course also a means of extending power. This appears to be a banal statement insofar as one understands war as a zero sum game. It can then only be to gain what someone else in the same process loses. But war can also be pictured as a creation of power, rather than as a taking of power. Clearly there are possibilities of creating power bases through the delegation of command during war. On the other hand, the resources for wh ich people go to war may be so unlike those which our own society considers to be essential that we overlook them. It can also happen that such grounds for war are never stated, although they are salient. They are obvious to the actors concerned, but it could be a matter of shame or stupidity to speak openly of them, for these material grounds are overlaid with the ideology in wh ich a larger circle of alliance is couched. We will name the cases of land rights, employment quota and marriage. In many states there are today de facto no secured property rights in respect of land. Access to land in agricuIture is linked to a so-ca lied ethno-national identity. Land security is only given when the inhabitants, their chief and an officially recognised administrative unit (chiefdom in former British colonies, oblast in the former Soviet Union) coincide. If anyone decides to link the village to another administrative unit or to define another group of inhabitants as the autochthonous population (titular nation in the Soviet Realrn), then there is a risk of being expelled, of becoming victims of ethnic c1eansing. Violence appears then as the only means of self-help. Another conflict laden theme is the imposition of quotas for employment according to ethnic affiliation. When the most important employer is the state,

22

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

then its quota system in the allocation of work be comes an issue upon which livelihood depends. When jobs, particularly those in administration, are also means of access to scarce resources, legally or illegally, the allocation of jobs becomes a particularly fraught question. Whoever is a gate-keeper on this allocation of jobs becomes the patron of a clientele. Insurrection by those threatened by exclusion or a war ofrivals for such positions can result. In endogamous systems (in which a marriage partner must come from the same designated unit), the presence of strangers who do not respect this marriage circle can seriously reduce the chan ces of a young man or woman to find a mate. To fight against such a threat can become a vital war aim in itself. These reminders of motives which can remain under the surface should not be taken as monocausal explanations (and should in any case only be put forward with empirical evidence). Usually more than one interest is involved. Configurations of violence lasting over a long term, implicating a large population, can only be sustained when a plurality of outcomes, for different groups of interest might be accomplished by it. When a desired outcome can not be accomplished, a sufficiency of interests can still have been mobilised to continue the violence. A need for resources does not automatically lead to violence. On the contrary, the norm is to seek ways which involve the least risk. Every society offers structures for reaching compromises to gain access to scarce goods. Three of the most important institutions for reaching such compromises are markets, law courts and parliaments. Violence becomes likely only when recognised procedures for reaching compromise fail or lose their legitimacy. There is now a broad and amorphous spectrum of what can be described as procedures for reconciliation - round tables, mediators, and so on - which have become very popular. But their popularity can obscure the basic elements of a sociological concept of a procedure to end conflict (Luhmann 1993). Two crucial conditions distinguish such a procedure from a ritual or a palaver: one is that for the time and space necessary for reconciliation, negotiations must be kept apart from the operation of power (the power-holders must not be allowed to usurp negotiati on with the argument of violence); the other is that resolutions must be consequential, breaches must have sanctions. Sanctions can weIl be in the form of shame, if this has economic and social consequences for the person who has brought shame upon hirnself by a breach of the accords. For this reason it is important that to be effective, the reputation ofthe mediators themselves should have a high stock. Failure of institutional means for reaching compromises in contlicts is a main feature of the situations of violence analysed in this volume. One might retort that some of the cases analysed here refer to democratic countries. The existence of the institutions of the rule of law at the heart of a country, however, is not necessarily decisive. The question remains wh ether their rule of law is effective at the peripheries ofthe state.

The Dynamics of Collective Violence - An Introduction

23

Violence needs weapons. Where they are lacking, peace may simply be the result of weakness and war can be kindled by a new availability of weapons. The end of the cold war did not make it more difficult to acquire weapons. In some countries it is the state itself who is the supplier, others allow the arms trade into their territories. In Europe and Central Asia it so happens that it could be dangerous to have Russia, Pakistan and Greece as neighbours for this reason. (For diplomatie reasons they are sei dom cited as dangers). Weapons are not the only requirements for the regular practice of violence. One of the most important conditions for strategie violence is the effectiveness of internal sanctions. When a we-group (religious, ethnic, ideologieal, regional, clan) develops its own sanctions to shield itself off from a central state, there is already a potential for civil war. We think that the construction of such a capacity to exert internal sanctions is one of the most important indicators for a near outbreak of violence. This sanction capacity is also one condition for creating command structures. Command combines latent sanction with communication. Command structures are easier to build up when social networks with reciprocal obligations existed before the outbreak of violence. That means that c1ientelism on any basis (not necessarily religious or ethnic) is a building material for the creation ofthe social structure for violence. Last but not least, we have to come back to propaganda as a tool for violence.Violence always implies risks. Propaganda can reduce the perception of risks, can play down moral restrietions and social inhibitors. Propaganda can change the structure of relevance, that means, the - subjectively feit - relative importance of specific issues increases. This can change the references for the evaluation of decisions, the framing. Ifthe framing presents a situation as being a threat of loss, people are more inclined to use violence than if it is portrayed as being a chance to win something of equal value (Gosztonyi 1998). 5. CoIIective Violence in Modernity The cases upon wh ich all the analyses in this book, even that of the precolonial Borgu, are drawn come from what is usually called 'modem' history, in other words, history of the global system of emergent nation-states and of capitalism. The centuries-Iong process of nation-formation and the turning of mercantile into advanced capitalism have intensified trade and the trade in arms in particular. They have also intensified pressure on land use. They could therefore also have intensified warring in their borderlands. These cases therefore invite us to re-think the search for a social artifice to establish a permanent state of peace, which European political philosophy began and which led to the theory of the state as a legal monopoly of violence. The global system of modem states inaugurated in mercantilist Europe in the fifteenth century was also an exacerbation of piracy and protectionism, the most violent of which was the trade in slaves. It was itself a system of states and an economy of violence

24

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

(Tilly 1985). That violence was (1) monopolised by a corporation wh ich called itself state and (2) that then this monopolised violence was subjected to and bridled by a system of rules called a legal state, which produced the "legal monopoly ofviolence". It was the result oftwo different historical processes. To be clear, we should not confuse "modem history" with "modernity". "Modernity", starting with the European Enlightenment, led to a new philosophical thinking that produced new normative standards for modem society. These promises of modernity were for the first time formulated in the American War of Independence and the French Revolution as "freedom, equality, and fraternity". Every one ofthese terms is also fraught with the tensions ofhypocrisy, of racist, or gendered, or class exclusions, and more exceptions than rules in the selective application of human rights. The promises of modernity are not only not always kept, modernity also created techniques and logistics that made war extremely efficient and gave the option of total destruction. This destructive power ofmodern war also made it possible to pursue such extensive targets as the complete destruction of the enemy. Total war, as described by von Trotha may be an ancient child but it is a modem adult. Modernity is Janusfaced. It not only created these changes and a modem science that perfected the destructive power of war. At the same time, scholarly thinking provides an analysis of collective violence and the categories and normative legitimacy for a critique ofthe development ofthe global system. The modernity of total war and of the colonial wars of European nationstates is the modernity of astate, a territory with firm boundaries based on its population as a mass to be mobilised or to be used. Von Trotha notes that the horne population becomes war material. Modem war is demotic and anonymous, not just cruel. The weapons of modem war, weapons of mass destruction, confirm this political fact. The weapons of ultra-modem war, laser targeted, are a callous terror which co-exist with the terrorising sniper. Their accompanying euphemism of 'collateral' casualties contrasts with the named lists of the horne side's martyrs. On the one side, everyone is an individual, on the other an anonymous category. The sniper is the human side of a bureaucratic war machine, acting out the sadistic fantasies of sport, vengeance and triumph which the clinical rationality oflaser targeting denies. When we turn to civil war, we are again in a specifiable historical context. Sy definition civil war is impossible without the modem state. Lessons from previous wars can teach us about civil war. The war wh ich devastated Germany for thirty years 1618-1648, a war between monarchies and confessions, is cited by Waldmann because ofits dynamics - war breeds war. Waldmann's point is that modem, civil war can also breed warring, oscillating violence when the warriors become paid mercenaries. Sut note that civil war is the most modem of conditions. Tim Allen in his chapter produces figures for the exponential rise of civil wars since 1945. Similar figures could be cited for the rise in criminalised violence which has become an enduring state of life in more or less

The Dynamics ofCollective Violence - An Introduction

25

contained regions of nearly all the major cities of the world. Somewhere between criminalised neighbourhoods, bandit areas and civil war are outlaw neighbourhoods or forest regions which are more less politicised bases for militia, as the history ofwarring in Los Angeles, Colombia, Sierra Leone, Liberia and Ulster seem to indicate. The word 'civiI' alerts us to a situation in which long periods of peace maintained by the conventions of inter-state relations are broken. Those conventions are in any case cross-border, inter-territorial agreements linked by increasing numbers of dislocated populations and their networks built up over periods as long as three centuries or as short as three weeks. Their civil breakdown reminds us that the political arrangements we are used to are not fixed. Insides are broken into and tumed inside out. New· insides are created, often with extreme violations of physical boundaries. The re-inscriptions of territory and the violations of the bodies of residents of the Yugoslavian federal state, the zones of terror and counter-terror of Sri Lanka, the weak frontiers of central Francophone Africa (Burundi, Rwanda, and Congo/ZaYre) and of Liberia and Sierra Leone are some of the recent reminders wh ich come up in the articles of this volume. Another aspect of modemity is the secularity of the modem state. The effect this has on the possibilities of conflict and de-escalation has been the subject of Bruce Kapferer's work (1988), inc\uding a presentation he offered to the first of our conferences. Kapferer stresses a discursive condition of nation-states: their rationalising incorporations of symbolic violence in the myths and rituals which are their heritage. He focuses upon the hierarchical rational-bureaucratic order exemplified by an extended analysis of Sri Lankan civil war. The civil conflict is the long-term result of European colonial and capitalist imposition. But the potential conflict was consolidated and more fiercely totalised in the post-colonial state which resulted from the struggle for independence and the necessary formation of its own nationality out of a revival of pre-colonial myths, sacred places and occasions. A modem state with its powerful regional and centralising apparatuses undertakes a totalising process in what was hitherto a loose encompassing of local traditions. They are tightened into anational tradition. The very same process turns group identities into potential rivalries, now with reference to the national whole and its state. Political differences and social exc\usions, as well as regional differences, make fragmentation and the remaking of totality and continuity an issue and a condition of the body politic. Continuity and fission which were issues of descent line and biographical identity, under some mythical or cosmological authority wh ich transcends them, have now become issues of anational, political community. Kapferer's point is that the propensity to violence in precolonial traditions was contained by an encomp~ssing spiritual authority. Spiritual authority contained rites and myths of exorcising demons by demonic force, stories of kings and princes fighting demonic chaos or reasserting dignity

26

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

against humiliation by a demonic enemy. They were encompassed by a transcendent sanctity of inclusive beings with spiritual purity; stories of kings becoming monks. In the pre-colonial traditions, religious authority, ultimately of spiritual purity, was lent to a secular power. Now the order is inverted as weIl as tightened. Bureaucratic state power supports religious authority. Its ultimate sanction of force mobilises the traditions of demonic power to annihilate demonic enemies and establish order, in the name of Buddhist (in Sri Lanka's case) non-violence, wh ich is the state's national tradition. But this very same religio-national tradition serves as justification for the use of annihilating power. Remythologised and taken more literally in the manner of a rationalised religion, the demonic myths and rites are taken into a far more strictIy ordered territory and hierarchy, in which a cosmic state is now also populist and bureaucratic. The crises of such astate include accusation of a regime for corrupting its tradition, and the necessity for its renewal. Renewal includes the use of the same demonic myths acted out as a terrorising pursuit of purity by both sides, by the forces both of the state and of its revitalising opposition. Religious authority becomes an untranscendable battIe between good and evil over the same population and territory. Under these conditions, in our words, secularism breeds nationalism and a civic religion ofviolence. Seen as a critique of contemporary culture such arguments seem to underline the thesis that modemity goes together with an increase in collective violence. Such a simplified point is not intended here. The effects of modemity are more complex (cf. Dinges 1998). We have already pointed out that in many parts of the world before the modem age oscillating conflicts were common. This is true for Europe as weIl as for Africa, America and Asia. At the fringes of already existing empires oscillating violence was anormal state of affairs. The modem nation state developed more effective techniques of war. Using their military strength some modem nation states built up colonial empires (Britain, France, Holland, Germany, Belgium, Russia, Japan). Colonial wars of pacification were obviously extremely violent. But after the establishment of a monopoly ofviolence there were long phases of colonial peace. Population growth in Africa is not only a result of improved hygiene but also of pacification and its stabilising effects on the economy including food production. Collective violence in modemity has developed an exceptionally destructive force that is efficientIy used and supported by martial ideologies. At the same time its normative promises include a strong orientatiop towards peace and the state's monopoly of violence has created wide spaces of controlled violence, and made collective violence an exceptional case. Modemity's increased risks of collective violence go together with increasing chances to negotiate peace.

The Dynamics ofCollective Violence - An Introduction

27

6. Processes of De-Escalation De-escalation has a different meaning in cases of oscillating violence, where violence is anormal state of affairs, and those where violence is an exceptional and escalated state of affairs. Oscillating violence incIudes roles for deescalation, including procedures for the negotiation of peace and protection. Feasting with the Yanomami, the payment oftribute or payments for protection in Borgu or buying protection in markets of violence are routine affairs. A limited peace can be reached easily using well-known ways. Easy peace also reduces the threshold of using violence in conflicts. De-escalation as apart of warring is always Iimited in time. Oscillating violence does not give cIear points of departure for long lasting peace. The next preventive strike or the fear of losing ground for a violence-based economy is the threat that inhibits peace. In situations of oscillating violence long-Iasting peace is the exceptional state of affairs that requires a complete change in the social conditions and structures. This often happens only by the imposition of peace by a superior force of an empire, or more recently of a colonial state. In other words, it is an historic de-escalation. In its own terms and time, however, de-escalation of any instance of oscillating violence is an alliance of warring parties to enhance the capacity to raid or make war on another target. De-escalation of violence as an exceptional state of affairs is possible at two different points: I. as preventive de-escalation before the eruption of open violence and the start of a war; 2. at the end of an already violent conflict. Deescalation is in either case anormal state of affairs. In potential or actual violent conflict de-escalation is confronted with the problem of "throwing away the victory" or the "unjustified defeat". As long as war is seen as leading to complete victory violence once started is measured by the standard of victory. A de-escalation incorporates the potential of violence to regain the "thrown away victory" or to undo the "unjustified defeat". When a peace is negotiated we should see distinct processes of peacemaking. There is: 1. The peace of the victors. They are the heroes and control the defeated, politically and economically. This is de-escalation by force. 2. The peace settlement (Frieden des Ausgleichs) that is agreed upon by both conflicting parties negotiating on a similar level. Political and economic power is subject to negotiation and incIudes an element ofbalancing. In these types ofpeace there may be different roles for the war heroes. I. War heroes are the new (perhaps even the old) powerholders. In case of a peace of victors only the victorious are heroes. In a peace settlement the war heroes ofboth sides hold power. 2. War heroes are ex-combatants with no more use for society. They are symbolically honoured but power is taken by civilians. In the case of internal

28

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

wars even losers may gain access to power. Sometimes the combatants are already discredited during peace negotiation, because of a widespread warweariness or because the combatants have become a threat to the civilians of their own party. The best chances for de-escalation in cases of exceptional violen ce (escalating violence) are before the switch to an eruption of violence. The switch creates c1eavages that are hard to bridge. The South African "Commission for Reconciliation" has a chance to succeed because the widespread violence did not lead to a total war or to a genocide. To a remarkable extent, opponents ("them") were not involved in violence. This makes it c1ear that after a genoeide de-escalation is an enormous challenge without guarantee of success. The still complicated relationship between Germans and the Jewish world community is an example of the long-lasting quality of such an escalation, particularly since it was entirely one-sided. The studies in this volume give us an insight into the social processes of deescalation. De-escalation of exceptional violence is similar to de-escalation of warring that breaks out of the cyc1e of violence and limited peace. What do our studies suggest? No de-escalation is non-violent. One type is the assertion of superior force, reducing the threatening forces to a point where they are immobilised by a reformed or renewed state. A case is the Nationalist and then the Communist ordering of what had been the brigandry and the fragmented warlord polity and economy of China. In another case which we call the pluralising tradition, the opposing forces divide into separate entities, implying an agreement to regional autonomy brokered by another state or by international ageneies. Neither deescalation is secure. But then since every de-escalation is anincorporation or a marginalisation of the use of collective force, no de-escalation is sec ure or nonviolent. Marginalised into a soeial exc1usion on a border zone, the defeated factions and the ethnically isolated may re-mobilise. De-escalation is often a mere toning down of armed collective violence, de-escalation into banditry or warre, as in the border zones ofRwanda, Burundi, and Congo/Zaire. Lasting de-escalation is the "soeial incorporation" of violence. Brokering peace after violence requires agencies of recognition. It works through appeals to professed ideals and the very notions of honour by which warlords and their forces glorify their acts. Warlords and their ambitious, external rhetoric, are bound into institutions of c1ientage and support. Their orders of prestige also include, beside courage and feroeity, those of protection, service, loyalty, hospitality, generosity. Peace brokers prompt them out of their defences against fear and from their armour of aggression into a recognition of suffering and grief. The recognition is often itself only a front put on for those who check and remind from outside. But such checks sustain the possibility of de-escalation. Often brokers are outsiders such as returning emigrants, external patrons, or religious authorities if, as Grohs points out, those authorities have established a

The Dynamics of Collective Violence - An Introduction

29

credibility with both or all sides. Grohs also makes the important point that the ultimate recognition must come from a legal and secular authority which has the sanctions of international law and of force. On the other hand, such recognition can be used to mobilise forces for further violence, particularly when the sanctioning regional or world powers recognise only one side (as Grohs shows in his examples ofthe USA in EI Salvador and South Africa in Mozambique). War catastrophes split, with an incommunicable gap, the heroic mythology of the purposes of war from the experience of the pointlessness of suffering. This is the splitting which makes possible the immunisation against pain noted by von Trotha. During wartime the pointlessness of injury, death and killing is expressed as an undercurrent, a subversion which becomes treachery when it emerges into the open, and yet it has the potential for healing and the crossing ofthe borders offriend and enemy, with their excluding identifications ofGood and Evil. When soldiers for whatever reason allow themselves to experience as they would in normal time the pain they see and intlict in battle, they suffer 'shell shock'. The undercurrent of "normal perception" is transformed into a medical symptom. Afterwards when their perpetration or witnessing of defilement and pain comes back to them as memory, it can, in contemporary medieine, be diagnosed as Post-Traumatic Stress Disorder. The psychiatry of memory, or rather of its return after denial, presently deals with the splitting by recognising it through a psychology ofpost-trauma counselling. That is a pecuIiarly Western and medical return to normaIcy and its future. When offered by international agencies, it can be entirely inappropriate to situations which have their own, local means of re-integration into a time of normalcy. By whatever means, however, de-escalation starts by re-integrating and transforming the mythology ofwartime into ordinary commemoration and genealogy. The discursive space for the brokering of peace and the social incorporation of violence is one of "accountability" which has to replace violence itself as the currency of negotiation. It is the opening of discursive and memorial, bodily, verbal and ritual spaces, building bridges between fight leaders and talk leaders, in which collective suffering can be recognised, responsibility for it assigned, but the perpetrators included. This involves forgetting as weil as remembering, duplicity as weil as transparency. Transparency and duplicity cannot be separated. Translation and negotiation entail both the duplicity of mystery, which is hidden and privileged knowledge, and trust in the procedures of accountability between the sides brought into negotiation. De-escalation is Iikely also to involve further violence, that of accusation and vengeance. The recuperation of social relations ruined by insecurity and distrust must often be worked out through rituals of accusation and compensation. Local traditions of exorcism, or of witchcraft, of divination, of spirit possession, of gossip and of other kinds oftribunal (as analysed by Heike Schmidt) are a necessary part of de-escalation and cannot be replaced by historicalor

30

Georg Elwert, Stephan Feuchtwang, and Dieter Neubert

truth commissions and international courts of justice which work at another, inclusive level (Werbner 1995 and Allen 1998). However, we have to be careful. An instrumentalised use of local traditions may lead to reverse effects. Schwandner-Sievers shows that in Albania local concepts were used by the state and by other brokers for control or reconciliation in local conflicts. But this strengthened the ethic of oscillating violence. Instead of long-Iasting peace the result was warring and the already mentioned risk of lowering the threshold of violence through easily accessible peace. The problem of further violence in processes of escalation mayaIso be handled by repression or even suppression of war memories. After the second world war the Nuremberg trials dealt only with a small number ofprominent Nazis. Other powerful figures of the Nazi regime tled Germany or changed their identity. The bulk of those who took active part in the Nazi atrocities managed to integrate themselves into German post-war society. It needed more than twenty years and a new generation of Germans before suppressed questions of guilt were discussed in public. The turning point was the Auschwitz trials in Frankfurt/Main which Iinked the deeds ofthe past to legal norms and thus effectively made them crimes. Collaboration with Nazis in the occupied countries, especially in France and the Netherlands, has successfully been suppressed to date. Only very recently has this part of French and Netherlands history been discovered. An opposite case is Rwanda. There are now more than 90.000 people accused as participants in the genocide (Neubert). This effort of juridical reappraisal of genocide overloads the weak juridical system and has led to an atmosphere of denunciation. Processes of de-escalation are always fragile. Peace could break down and violations occur at any moment of de-escalation. The violence of healing will inevitably also involve the organisation of force. To inhibit its use, a discursive vacuum must be filled turning violence into a less restricted discourse. Like the talking eure - from somaticisation to verbalisation - peace-making is the incorporation of violence, a work turning violence to oratory and to rituals of aftlicti on and conflict, from sectarian to contested histories, from defensive fear of death to recognition ofrival claims to truth. Another danger to de-escalation is the difficulty of demobilisation, and the disarmament of combatants in civil war. Any long-term de-escalation must find means by which the young men of militia whose only loyalty is to fight leaders and each other can find a Iivelihood beside that of marauding or mercenary employment. The temptation of non-participant agents is to act in response to media exposure of the horrors of war. Emergency intervention and the imposition of a negotiated peace is often necessary. But it can weaken the possibility of longterm de-escalation, particularly when it imposes institutions of peace-making and conciliation which ignore local institutions and the potential of the parties involved in the conflict to switch to talking. Here we must mention the efforts

The Dynamics of Collective Violence - An Introduction

31

of humanitarian aid in violent conflicts, as analysed by Allen. The objective of this aid endeavour is simple. If it is not possible to stop violence immediately, at least the suffering of civilians should be prevented or alleviated. Sut these well-intended activities often support the conflict. The aid is integrated into the war economy and in cases ofwarlordism the aid itselfand the protection sold to aid agencies become new, contested economic assets. Even when aid agencies organise the help themselves, as in Southern Sudan or Mozambique, they can diminish the chances for peaceful reconstruction. They undermine the sovereignty of the states and their administrative capacity, which is necessary for a post-war development. This is not to mention those agencies that are more interested in their public image than in locally adapted sound aid programmes. In conc1usion, de-escalation requires the strengthening or revival of institutions of accountability, conciliation, or stand-off, with effective sanctions at all levels, internal, state, and international. Rules and institutions for conflict resolution which are widely recognised and necessarily also sanctions wh ich enforce rules, compromises and judgements are conditions sine quo non for a social incorporation of peace. Our studies offer no solutions wh ich can be abstracted from local and historical contexts. They do indicate some general historical contexts for endemie or oscillating collective violence. They also point out strategie, instrumental, ideologieal, and emotional conditions wh ich move and sustain the dynamics of collective violence. These imply the necessity for movements of large-scale social, economic and political reform to reduce the possibilities of economies ofviolence and the emergence ofmilitant, fight-politicians. They do not offer a vision of the end of collective violence. Their vision is one of de-escalation to more controlled and regulated violence, the substitution of acts of violence by institutions and rites of accusation with recognised forms of compensation and sanction. They are part of the modernity which offers itself the sanctions of force for peace and which at the same time condemns the unique horrors of modern mass destruction. They add to it the imperative of acknowledging nonEuropean and local means of controlling and regulating violence.

I. The Logic ofViolence - Theoretical Approaches

Forms of Martial Power: Total Wars, Wars of Pacification, and Raid Some Observations on the Typology of Violen ce

By Trutz von Trotha

Introduction

Power is the faculty to prevail against extraneous forces (cf. Popitz 1992: 24). Violence, as Heinrich Popitz also emphasises, is one version of a form of power, he calls it "action power" ("Aktionsmacht") or what we mightjust call "power".) The following remarks aim to further develop the concept of power by studying three types of violence, which I have named "martial power" ("kriegerische Aktionsmacht"): raid, war of pacification, and total war. I will begin with the concept of war (I) and then look at ditTerent types of war with a focus on wars of pacification and total wars (2-3). By asking if the German colonial wars in Africa were total wars, I will continue by outlining some important ditTerences between wars of pacification and total wars (4) and with some comparative observations on raids (5). Finally, retuming to the methodological and theoretical questions from the beginning I will give a few hints as to the place of my observations and conceptual etTorts within a 'theoretical eth-

I The translation of "AkJionsmacht" by the term "power" does not really carry the precision ofthe German term, which is based on Popitz' differentiation between various forms of power and the German difference between .. Macht" ("power") and .. Herrschaft" ("domination", "institutionalised power"). The term "AkJionsmacht" insists on the notion of human action and grounds the theory of" Macht" and "Herrschaft" upon a theory of action. Nevertheless, I use the term "power" here to save the English speaking reader from a vocabulary which seems to be rather Teutonic and thus implies other misunderstandings.

The conceptualisation of violence as a form of power is very distinct from the conventional concept wh ich opposes violence to power (cf. Arendt 1970) and has led the study of violence off the track. For the concept of "Aktionsmacht" see Popitz (1992: 24-25,31-39,43-47). I should mention, that all translations from French or German are mine. 3*

36

Trutz von Trotha

nography' of violence which, in my view, is the methodologie al and theoretical core ofa sociology ofmartial power (6). The general theoretical aim of my considerations is to contribute to a paradigmatic change from a sociology ofthe causes ofviolence and war to a socio/og)' of violence and war based on the theory of social action which starts with anthropological questions conceming basic human faculties and vital human dependencies. 2 0ne of the reasons for this change of perspective is the idea that before studying the so-called 'causes' of violence we must have a precise sociological understanding ofwhat violen ce and war are. The analysis ofviolence and war is not very weil developed within the conventional sociological discourse. Many theories speculate about dubious 'causes' of violence and war without ever having made clear or having understood what the celebrated 'explanandum' is, which is to be 'explained' (cf. the critical remarks of Sofsky 1997, 1996: 49). In addition, by placing the emphasis on descriptive and analytical-typological questions instead of the causa I question, I wish to stress the limits of the mainstream etiology of violence and war and its scientism. In my view, the latter is a dead-end road for theory and research, because it cannot come to terms with the processional character of violence and war. Contrary to the mainstream theory and research, the sociological anthropology of violence and war, as I understand it, is a sociology of the structural and processional Iogics of violence and war (cf. Trotha 1997, 1996; Trotha and Schwab-Trapp 1996: 56-57). The more specific aims of my notes are to clarii)' some of the prominent forms ofmartial power in Africa, especially African colonial history. ] hope that my observations directly bear on the debate about the character of colonial wars in Africa and, in particular, on the German colonial wars in Africa which are still prominent candidates for the concept oftotal war (cf. Boemeke et al. 1999; Trotha 1999). 1. What is War? War is a specific form ofmartial power (,kriegerische Aktionsmacht ,), which I understand to be that form of power which is based on the collective and more or less organised use of violence. The violence is directed against another co 1-

2 When referring to 'anthropology', I have in mind the German tradition of anthropology in the sense of Max. Scheler (1962), Helmuth Plessner (1981) and Amold Gehlen (1966, 1963). The specific framework is instead the 'sociological anthropology' of Heinrich Popitz (1992, 1989, 1980: 1-19) wh ich is theoretically rather different from the philosophies of the three scholars. I should add that the term 'sociological anthropology' is not to be found in the work of Popitz. It was coined by myself to summarise Popitz' theoretical endeavour.

Fonns ofMartial Power: Total Wars, Wars ofPaeification, and Raid

37

lective and includes first and foremost military or paramilitary actions. Martial power is grounded on the combination of two basic anthropological forms of power: violence (,gewalttätige Aktionsmacht ') in various forms, and material harm. The destruction ofthe enemy's material resources is one ofthe most evident elements of war, and in many, especially modem cases, it captures the imagination of later generations long after the dead have been forgotten. 3 The essence of war is the use of 'absolute violence'. lt means to kill the adversary. But in war, absolute violence goes along with "total violence" which brings together the glorification of violence, the inditTerence towards the sutTerings of the victims, and the mechanisation ofviolence. 4 As part ofits collective and organised character war is directed against collectivities. It destructs or appropriates the material resources ofthe adversary; limits or destroys his political autonomy or hurts physically and materially the enemy to such a degree that the whole society ofthe enemy collapses or the enemy even stops existing as a human group.5 In short, war is the collective and organised use of the combination of three basic anthropological forms of power: material harm, and absolute and total violence. The combination of total violence, collective organisation and the violent confrontation of collectivities includes the strict distinction between "the internal and extern al group ('we' and 'them') and a related separation ofthe internal and external morals" (Mühlmann 1972: 475--476). In war, people follow two kinds ofmorals, one which applies to the members of one's own group, and one which is valid for the confrontation with members of the 'enemy' group. Thus, war articulates most sharply a basic feature of society, which always draws a line between members and non-members, between the "internal and external morals".6

J Contrary to the annihilation of people, in most cases the destruction of the material world is never absolute, thus keeping alive the relationship between the past, present and future. It is the ruins which remind us - and which are the basis of the archaeologist's contribution to society's memory and contribute to establish that kind of somehow abstract relationship to past wars whieh is part of our relationship to the material world. Of course, this is much more pronounced in modem non-oral and seeularised societies wh ich place so much emphasis on "data setting power" ("datensetzende Macht"), based on technology. For the concept of"data setting power", see Popitz (1992: 30-31, 160181 ).

4

For the eoneepts of"absolute" and "total violence" see Popitz (1992: 52-78).

5

The last ease is what might be called 'genocide'.

6 The strict distinction between 'we' and 'them' is direetly related to what I call 'basic membership conflicts', whieh are at the core ofethnic and raeist confrontations (cf. Trotha 1998).

38

Trutz von Trotha

War is an interactive process - to take up sociologically a military truism. The course and character of war is shaped by the process, wh ich confronts more or less organised groups, socio-cultural orders and actors. Wars have a history. War is a particularly dynamic phenomenon. Any theory of war has to take into account the proper dynamics of armed conflicts. The dynamic character of violence and war implies a sociology of speed which is a fundamental feature of power and domination (cf. Sofsky 1997; Trotha 1994: 69-79).7 An important element of the dynamics of war is the re-organisation of time and of the orders of the past, present and future. Violence is an order of the immediate time, of situationalised time. It is the here and now which counts now I have to prevail for good and evil. Violence might even produce the co 1lapse of any time structure and reduce it to the very moment the violent attack hits (cf. Sofsky 1997). War is the collectively generalised order of situationalised time. The situationality of war is even present in the historical and cosmological myths, in the 'basic story' of political cultures and the 'basic tales' of everyday life, in the propaganda of war. 8 It is the underlying point of reference. In all these forms of narration and discourse war is released from its situationality and transformed to an order, the time structure of wh ich cannot be caught up by the actors. But it is typical of war that exactIy in those moments when everything is dominated by the immediacy of war, the war propaganda with its heroic slogans of 'etemal war' is increased to a thundering roar. There seems to be a positive relationship between an increase of the immediacy of war and the efforts within the legitimised discourse to transcend the situationalised time structure of war by placing the ongoing war and its actors within the timeless struggles of a fathomless history of glory and sometimes even of doom. With the situationalisation of everyday life the past loses its relevance. What was does not count any more. Especially when its scale increases, war is a rupture. It produces a rift between the present and the future - and sometimes it is an abyss - and frees us from the past. We strike the 'liberating blow', we 'set out' or whatever the metaphors might be, which characterise the specific unconditionally of the social process under the conditions of war. The social proc-

7 I do not go into those forms of war which are characteristic of many societies of hunters and gatherers and slash-and-bum agriculturalists marked by extreme ritual isations of the violent confrontations between warring groups (see also footnote 9). In addition, the dynamics of war were rather limited during many periods of the European Middle Ages, ofwhich the Hundred Years' War is an example.

8 By 'basic story' I mean that construction of the history of a society and culture which embodies the dominant and legitimate construction of the past. 'Basic tales' are those constructions of history, which are told in everyday Iife. The latter contribute to the construction of the basic story and translate the basic story into the order of everyday Iife (cf. Schwab- Trapp 1996; Trotha 1995a).

Forms ofMartial Power: Total Wars, Wars ofPacification, and Raid

39

ess in the shadow of war is dominated by the present - and particularly when the imagination is totally captured by the soundings ofhistory. The re-organisation of time is not Iimited to the past. It includes the future. War increases the unpredictability of social action and thus denies one of the most important functions ofnormative order (cf. Trotha 1974: 63-128). Typically, we can differentiate wars according to the ways the unpredictability of violent conflicts is handled. For instance, slash-and-bum agriculturalists try to limit the unpredictability ofwar by ceremonial strategies,9 in industrial societies people count on organisational power, an extremely developed technology of war and logistics. Guerrilla war, however, relies on the intensification of unpredictability. By the situationalisation of everyday Iife and the structure of unpredictability, war reverses what phenomenologists call 'normalisation' or the 'taken-forgranted' quality of everyday Iife. As with any violence, war 'denormalises' and 'dramatises' the social process and the cultural order. Perhaps, war is the most dramatised social process. This does not mean, even in the case of total war, that all mundaneness and thus the ordinary pursuits of ordinary people are completely destroyed (cf. Chickering 1999). A society at war is still a society with all kinds of conflicts, ambitions and occupations, which, seen from the perspective of fighting men involved in a deadly confrontation, might be qualified as mundane. In fact, in war 'normal' Iife and the 'extraordinary' wh ich is part of war, normalisation and dramatisation exist simuItaneously. I would even argue that in order to fight a war, it is necessary to preserve the structure of everyday life to a certain degree. But on the other hand, in war normalisation and dramatisation, the breakdown and the construction of mundaneness are in conflict with each other. Wars and phases of wars differ according to the degree by which everyday Iife is destroyed. Wars, especially some forms of civil war or violent ethnic conflicts, can even be transformed into a new form of everyday life. Such a transformation has far-reaching consequences. It changes the course of war, wh ich typically gets institutionalised in many ways and on different levels of everyday life, and deeply affects the way children grow up and spend

9 In the discussions during the conference in Berlin, which this paper was originaHy prepared for, Bernt Glatzer (Berlin) stressed the ritualisation of war in 'primitive society' implying that this ritualisation limits the use and the consequences of absolute and total violence which, as I have said, are fundamental features of war. From my readings of the literature, it seems to me that one should not overestimate these forms of ritualisation for which the field battle of the Dani of the high lands of Irian Jaya is the paradigm. Even for the Dani, the field battle is only one way to practise war and, in addition, it is only a specific phase in the whole process ofwarring (cf. Hanser 1985: 124-138). Nevertheless, I agree with Glatzer that my concept ofwar is only of limited value for the analysis of specific phases of war characterised by extreme ritualisation of martial power.

40

Trutz von Trotha

their youth. The contlict between nonnalisation and dramatisation is shaped by cuItural constructions of war. Some cultures regard violence and war as 'normal', others as a complete break from the 'nonnal' state of society. In more complex societies we will find that there are also contlicts between different and opposing cultural constructions ofwar. Situationalisation, dramatisation, and the unpredictability of action are related to features of war which have been dealt with copiously in literature and poetry. In the sociology and anthropology of war they are almost non-existent: war as a sensual experience and as aspace ofthe imagination. Violence, and above all war, are realities of feelings, which sometimes subdue human beings. They are also the world of emotions, of unlimited feelings. It is ofparamount importance that in this emotional world violence is a bodily and sensual experience - a feature which it shares with sexuality which typically has an important place within war. Violence is the quintessence of sensual experience. War is a world of shouting and yelling, people and objects move fast; even quiescence may be nothing but the calm before the stonn; in the deadly silence nerves are strained to the limit. In war, people wrestle with one another, they smash each other's heads, bon es crack, blood gushes forth, buildings go up in tlames and crash. War is a world of physical endurance, hardship and deprivation. It is also the realm of the feast and crapulence of the victor. In war, people wait and feel bored for hours and hours as weil as perfonn extreme mental and physical accomplishments, sometimes for many days, even weeks or months. The aItemating experiences underscore even more the sensual dichotomies of war. War is a world of extreme sensual experiences, many times changing literally from one minute to the next. The way people handle this sensuality of war, and the resources of meaning cuItures offer to people dealing with the pressing sensuality ofwar, are important elements for achieving victory or heading for defeat. They intluence the way people deal with the tasks they are confronted with when the war is over and the time has come to build a new order out ofthe ruins ofwar. 1o The body and the senses are at the centre of violence and war, and they are indispensable for what has been called the "fascination of violence" (cf. Steinweg 1983). The fascination of violence and war is the fascination of the sensuality of violence. People experience the body and have feelings and emotions, 10 Recently, in a remarkable article Birgitta Nedelmann (1995) has brought attention to the fact that the sociology of violence has failed to deal with violence as a bodily experience and has particularly failed to describe and analyse adequately the bodily quality of violence. For any theory of war, the description and analysis of the sensual and bodily quality of war is aprerequisite. The pioneering study of John Keegan (1975) about the "face ofbattle" has set standards, which will not easily be matched, but might help sociology and anthropology to develop a genuine sociological theory of war, based on a theory of action.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

41

which go beyond the limits of the ordinary. It is a world of the present and action, of conflicting feelings, but also of emotions, which seem to leave behind the tiring challenges of ambiguity and make us believe that the world is plain and elear. 11 The sensual experience which war produces is directly related to a fact which Popitz highlights in his theory ofviolence and offers one ofhis most penetrating and resultant insights: violence is connected to what Popitz calls the "boundless imagination" of human beings. 12 "Violence is not only something that happens or which happened ... , it is also something which might happen: the feared violence ofthe other, the triumph of one's own violence, which we so dearly wish. This horizon of possibilities goes beyond anything which might be calculated. Imagined violence lingers in daydreams and . nightmares ofall kinds." (Popitz 1992: 51)

Violence captures our imagination. Violence is obsessive. There are not many spheres of human experience outside war, which are so laden with fantasies of threat and superiority. The history of war is also a history of the social organisation of such kinds of fantasies. To its youngest product belongs the history of propaganda. In war, the obsessive fantasies are linked to a reality of experiences of which it has been said, since the times of Thucydides, that they go beyond anything humans can imagine - a statement which is not correct, but tells us a lot about war, and the relationship between war and the human imagination. 13 The fascination of war is the result of a reality, where the boundless human imagination and the sensual experience of human actors come elose to the imaginary point, where both meet. The present sociology and anthropology of war is neither a theory of bodily harm, of the sensuality and imagination of violence; nor an analysis and a theory of that ongoing aspect which is at the centre of war: dying, killing and, in particular, being killed. It is one ofthe most astonishing facts that the sociological and anthropological theory of war does not deal with the fact that human beings have to be brought to kill others and even more to let themselves be killed. The sociology and anthropology ofwar - and peace - must seek answers to the question: How do societies and cultures manllge to bring their members to prefer violent death to life? In this context, a lot can be leamt from John 11 The unambiguous and categorical aspect of war has deeply influenced the history of the Western intellectual in the 19th and 20th who fancied himself in writing vain and irresponsible apologies of violence and war (cf. the despairing critique of Enzensberger 1994: 66--70). 12 The Gennan original is called the "Uferlosigkeit unserer Vorstellungsfähigkeit" or, with a slightIy different meaning, the "Produktivität der Vorstellungskraft" (cf. Popitz 1992: 51-52,124-129).

13 With the rise of modem atomic, chemical or biological warfare we might indeed have come to the limits of our imagination, which are expressed by the use of images, which we reserve for the apocalypse.

42

Trutz von Trotha

Keegan - and rrom hirn about the Greeks, who in the history of western civilisation have created the murderous martial ethic of fighting to the last man. It is an ethic which is rather strange to the culture of 'primitive war' (cf. Keegan 1997,1995):4 But the culture ofwar is only one dimension to differentiate wars and societies. No less important is the social organisation of war, especially the social organisation of fight and battle. It is the analysis of the connection between the cuIture and social organisation of war, which, for instance, allows us to approach the dramatic difference between the wars and battles of the Middle Ages and the early modem times and modem total war, represented by the First World War, in which thousands of soldiers let themselves be massacred in the trenches of the German-French front. The French and German soldiers did not make off as had been the case over centuries when defeat seemed to be imminent. I summarise the most important elements of the concept of war outlined above: collective mobilisation and organisation of violence, material harm, killing of the 'enemy', glorification of violence, indifference towards the sufferings ofthe victims, the mechanisation ofviolence, the reorganisation oftime, the unpredictability of action, the situationalisation and dramatisation of everyday life, the bodily and sensual qualities ofwar, the obsessive fantasies of one's own violence and of the violence of the other, of threat and victory, and the cuIture and social organisation of killing and dying. 15 I suppose that this concept of war applies to wars in very different civilisations. The elements of war highlighted by the concept are found in wars of slash-and-bum agriculturalists and tribaI societies as weil as in wars of chief14 This martial ethic is a key to the Greek legitimisation of slavery, because the slave is a person who prefers life to death, slavery to Iiberty.

15 1 should add that this concept ofwar consciously avoids any presuppositions about the 'causes', the 'aims' or 'functions' of war, wh ich are usually further qualitled by calling them 'final', i. e. by speaking of 'final' cimses, aims or functions (cf. Orywal 1996: 41). The latter are the deceptions of a research about war, which desperately searches for 'causes' ofwar, and does not realise, that the aims ofwar can be as various as the occasions and grounds of war, the fighting people, the victims and the constellations of conflict. The etiology of war misses the processional character of war, and of violent conflict in general. Wars do not have 'causes' nor 'variables' in the sense of conventional deductive theory. Wars do not erupt Iike volcanoes. But that is exactJy the underlying image of causation cherished by conventional theories, which aim to discover as many 'factors' as possible to define and measure, so to speak, the 'pressure' , wh ich 'finally' leads to the 'outbreak' ofwar. 1 call this theory the 'steam-boiler-model' of war. Instead, wars develop. Wars have a genesis and a course. It is the genesis wh ich determines if there will be a war at all, and it is the course of war wh ich decides about its results. The analysis and theory of war are the analysis and theory of processes, not 'causes' ofwar.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

43

doms and between industrialised nation-states of the 20th Century. In addition, types of war differ according to the social forms different societies give these elements, the meanings which these elements have in different cultures, and the aims for which wars are fought. In the next part, I will outline four types of war: total war, the total war of highly industrialised societies, the 'limited' and 'unIimited' (or 'genocidal') war of pacification. When discussing the different types of war, I will not deal with all elements referred to in the general concept of war. I will only demonstrate briefly, that, on the basis of this concept, it is possible to distinguish certain types of war, which are historically meaningful and empirically relevant. 2. Total War in Stateless Societies and Highly Industrialised Nation-States Total war is not an invention of the industrial age. On the contrary, it was practised by many comparatively egalitarian societies and chiefdoms, at least in Melanesia and the Americas, if we take the following peculiarities as being characteristic of total war: the war tends to involve all members of the society engaging in warfare, and at the same time is directed without distinction against all members of the 'hostile' society. I should Iike to include further features, which are perhaps to be found in other forms of war, yet when radicalised, they result in their own particular type ofwar that is total war. There is the boundless cruelty with which opponents, who are captured, are slaughtered. This cruelty expresses the radicalised or total indifference towards the suffering of the victims as weil as the rigid demarcation between inner and outer morality. Oue to the strict, ethnocentric opposition between 'us' and 'them', the opponent tends morally to be no longer considered a member ofthe human race. Above all, total war is typically a 'holy war'. It has a religious and/or an ideological basis. In wars oftribal societies and other stateless societies with animist religions, this missionary trait, which in particular goes along with monotheism, is typically missing. Total war is aimed at the complete destruction of the opponent or his merciless expulsion from the area inhabited or laid claim to by his aggressor. 16 Total war in stateless societies can result in the destruction of villages and the expulsion of the members of entire societies. In comparison with the total war of industrialised nation-states, it does, however, have clear limitations, which

16 In anthropology the ethnic tribes of East and Northeast New Guinea are particulariy weil kt.1own for this kind of practice (cf. Hanser 1985; Hal/pike 1977; Mühlmann 1940; for the Tauade, a particularly interesting case of New Guinea, cf. Turney-High 1971).

44

Trutz von Trotha

make it appear more closely related to the 'limited wars', which have been prevalent in the history ofmankind. The total war of industrialised societies (or with the means of industrialised civilisation) is unlimited. This lack of limitation refers to the quantitative size of the population, which can be mobilised, on the one hand, and to the number of people on the 'enemy' side, who are affected by the war, on the other. The distinction between soldier and civilian is lost (for a critical view ofthis aspect, see Chickering 1999). Everyone becomes part of a 'war machine'. Everyone becomes a target of absolute violence. In a specific sense, war becomes a territorial war as it basically involves all people in a specific area, both as soldiers 'in the field' and at the 'horne front', and as victims (cf. Morgenthau 1963: 212, 308; Jouvenel 1972: 182). The incomparable degree of mobilisation in the total war of an industrialised society is partly due to the banal fact that nation-states include remarkably large numbers of people. However, what is more important is the fact that industrialised nation-states have an incomparable and historically unique organisational power over the people in their territory, whereby the latter become known as 'the masses' and, in the case of total war, simply 'human material' (Jouvenel 1972: 13-20). This transformation ofhuman beings into 'material' signifies that in total war the indifference towards the sufferings of the victims spreads to the members of one's own society. The power of organisation includes the industrialisation of the organisation of military resources, the extreme division of labour in war management, and, above all, the comprehensive militarisation of a society at war. Thus, the comprehensive militarisation of society expresses the high degree of integration of central power into industrial society.17 The industrialised organisation ofwar involves a technology ofwar, which is known to overshadow everything, which mankind had known until the revolution ofweapon technology in the last third ofthe Nineteenth Century. Since the first atom bomb was dropped, weapon technology has been threatening the world with a 'war of total annihilation', 18 in wh ich a "new way of killing, human behaviour of a new kind" has arisen, as Heinrich Popitz (1992: 74) so rightly emphasises. 17 For the concept ofthe 'integration ofpower', see Popitz (1992: 233-260). One should add that the militarisation of society inc1udes even peace, insofar as, for example unlike war in the Middle Ages, the pn:paration for war is an important part of society (e. g. in the form ofthe arms industry). In this respect, however, attention should be drawn to the basic differences between totalitarian dictatorships and liberal democracies, which have institutionalised checks against the radical militarisation of society. 18 The idea that the end of the Cold War has liberated us from the threat of atomic war could become an illusion with apocalyptic consequences.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

45

Total war in the industrial age is a war of systematic innovation of lethaI technologies. The process of innovation is akin to a competition, wh ich is centred around the invention oftechnologies ever more deadly, precise and able to kill an ever increasing number of people. In this sense total war in an industrial society is a war of science, especially natural sciences and engineering. The unlimited character of total war in an industrialised nation-state is deeply embedded in the radicalised missionary trait of total war. It is a war of irreconcilable ideologies. The ideologies are incorporated in the aims of war, the way war is waged, and the sum total, in which violence is glorified and the indifference towards the suffering ofthe victims is legitimised. 19 3. 'Limited' and 'Unlimited' Wars ofPacification: The Ca se of German Colonial Wars in Africa There is no state-building conquest without wars of pacification - disregarding the more monstrous case of genocide. Thus, colonialism as a phenomenon of most various forms of state-building processes on agIobaI scale is a paradigmatic case to study wars of pacification. I limit myself to the German colonial wars in Africa. Imperial Germany alone waged thirteen wars in Africa between 1884 and 1908 simply to subdue their African colonies and keep them under control. 20 We shall never know how many people died as a result ofthese wars and 'punitive expeditions' or their immediate or delayed consequences. The number of African dead can only be guessed at. They amount to hundreds of thousands. 21

19 These ideologies are important distinguishing features of the types of total war in industrialised civilisation. Particularly noteworthy is the distinction between the 'war of total world domination' of National Socialism and the 'war of total defence' of liberal democracies.

20 This figure includes only the more important confrontations. The more or less comprehensive "punitive expeditions" amount to hundreds. In Gennan Southwest Africa between 1891 and 1897 more than 60 larger "punitive expeditions" were sent out by the Gennans (cf. Gründer 1985: 154). Between 1895 and 1899 the official record of the police troops in the small and, from a military point of view, comparatively unproblematic colony Togo Iists 18 "march es", "expeditions", "military" and "punitive expeditions" (cf. Sebald 1988: 172-173). 21 The recording of death was more ace urate for the temporary victors in these anned confrontations. Regarding purely military personnel (i. e. without taking into account the "police troops") by 1918 7,200 officers, non-commissioned officers (NeO) and soldiers had died. But, while African soldiers made up the greater part of the army, the numbers of African dead were only estimated. In the wars of Gennan Southwest Africa 1,849 Gennan officers and soldiers lost their Iives, many of them through sickness and

46

Trutz von Trotha

The colonial conquest and suppression of armed opposItIon took place within the framework ofthe idea of occidental state power, which, in the eyes of the conquerors, was to be established, carried out and maintained in its essential features, i. e. in the form ofa modem territorial state (cf. Trotha 1994: 12-15). The colonial wars of conquest were not satisfied with the control of trade, terms of tribute and other economic and politieal considerations. The despotie and armed intruders did not undertake conquest in order to expel the people who had settled there, nor did they consider the area as a supply of slaves, who could with impunity and without risk of dangerous and expensive opposition be plundered for human booty from time to time. 22 Instead, they used martial power for the establishment of colonial domination (,koloniale Herrschaft'). Historiography has therefore introduced the concept of the 'pacification phase' for the first phase of colonial conquest. The concept highlights the statebuilding framework of colonial conquest and looks at it from the more or less utopian result, from the imposition of the state monopoly of violence and the establishment ofthe modem state. But it is somewhat euphemistic. It suppresses the fact that pacification is a violent imposition of rule. Partly, this violence is structurally rooted in the vicious circle of violence control: in order to curb violence, the use or at least the threat of violence is aprerequisite; in order to enforce the state monopoly ofviolence, the state builders crush all those who do not submit to their monopolistic claim. Consequently, pacification typically consists of wars of pacification. It is the old, familiar, violent conquest, and massacre is one of its characteristie features (cf. Trotha 1994: 37-44). Yet, in spite ofcases ofmassacres, wars ofpacification are generally 'limited', because they are aimed at subjugation and not at mass killings or even annihilation of the conquered. As long as the subjugated submit to the rule of the arrogant and powerfid intruders or even seize upon the chances the conquerors might offer, the violent demonstration of martial power remains limited. In short, it is the dream of state-building whieh limits the unleashing of martial power - with the qualification that the conqueror does not solely consists of settlers, who wish to wipe out the natives. other injuries, which were not a direct result ofwar (cf. GanniDuignan 1977: 126, Table 15; Kämpfe der deutschen ... 1907: 187; see also Bridgman 1981: 164). 22 Although when confrontations became more intense, an increasing number of voices among the conquerors were heard calling for the complete destruction and expulsion of Africans in certain areas. In addition to the Chief of Staff, Count Alfred von Schlieffen, the Commander in Chief of the German troops in German Southwest Africa with dictatorial powers, General Lothar von Trotha, was one of this number. But Trotha's political and belated failure - since the Herero and Nama paid for it by being almost completely wiped out - is arefleetion of the perspective in which the colonial wars took place: they were aimed not at annihilating or driving away the people in the colonies but at subjugating them to the state rule of foreign conquerors; cf. the excellent and impressive account of Uwe Timm (1985: especially 29-30).

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

47

Pacification tends towards the delimitation of violence, to 'unlimited wars of pacification', if those who are subjugated, do not accept the imposition and decide on armed resistance. In contrast to the ideas which for a long time have determined the image of colonial conquest, in the last twenty years the fact has come to the forefront, that the colonial assumption of power was an ongoing struggle on the part of the conquerors against the most varied forms of resistance on the part ofthose who had been colonised; for a lengthy period oftime armed resistance was one of these (cf. Trotha 1994: 79-84, 411-434; Ranger 1987). Accordingly, colonial pacification became a bloody proof of the superiority of the colonial conquerors. It followed the logic of pacification and was asserted and strengthened by the occidental war culture, which was expressed so radically and brilliantly in the seminal work of Clausewitz (1991). Clausewitz accurately summarised this logic in his ideas about the "first" and "second interaction", which govern war (ibid.: 18-20). They proceed along the lines that violence becomes unlimited, that the use of violence "must be taken to the limit", (ibid.: 19) whenever the opponent does not accept defeat, and "I (must) fear that he may defeat me, that I ... (am) therefore no longer my own master, but he is the one who (enforces) his rule on me, as I do on hirn" (ibid.: 20). The differences in the pacification process of Togo and the colonies of German Southwest and East Africa provide sorry evidence for this tendency towards the totalisation of war in a process of pacification, which encounters more or less continuous armed resistance. While people in Togo by and large gave up the strategies of collective armed resistance after the turn ofthe century in favour of other forms of opposition, the German conquerors in German Southwest and East Africa were faced with armed resistance, which was not only aimed at driving the intruders from the country. It was resistance, which followed the pattern of guerrilla war, which was no longer completely at a disadvantage as far as weapon technology was concerned, whose leaders made hardly any military mistakes given their resources, and consequently led very quickly to losses hitherto unknown on the German side. The Germans were afraid oflosing the war (for the Maji-Maji war, cf. Gwassa 1972: 140; for German Southwest Africa, cf. Bridgman 1981). In addition, a comparison ofthe 'Iimited' war ofpacification in Togo and the 'unlimited' wars of pacification in German Southwest and East Africa indicates an important difference, which is to be found in the socio-economic and political differences between these colonies, and refers to the 'settier problem' mentioned above. Togo was a trading colony, where, after the failure of the plantation policy and the political defeat of its supporters, a policy prevailed which was directed towards the native, rural family economy, "the promotion and development 0/ Negro cultures ", as the director ofthe Imperial Colonial Office, State Secretary Bernhard Demburg (1908: 226; emphasis i. orig.) expressed it in one of his major speeches on colonial policy. However, the situation in the settlers' and plantation colonies in German Southwest and East Africa was

48

Trutz von Trotha

completely different. There, the insatiable hunger for land by settlers and, above all, plantation owners, their constant need for cheap labour and their relentless racism, which plantation owners and settlers typically adopt and esteem, were the rule of the day. The existence of settlers led to a change in contlicts, not only between the rulers and the ruled, but also between the groups of rulers themselves. Unlike in a trading colony, the conquered in colonies of settlers and plantations see themselves confronted with the life-threatening alternative of slavery or annihilation, and the network of opposing interests within the group of conquerors is far more varied. In German Southwest Africa, these opposing interests were to be found in the contlicts over the policy of the Governor, Theodor Leutwein, and in the opposition between Leutwein and Lothar von Trotha (cf. Bridgman 1981; Bley 1968). The interests, ideologies and behaviour of the settlers and plantation owners serve to sharpen the antagonism of the "colonial situation" (Balandier 1982). The ideological bases of the conquerors' sense of superiority become particularly radicalised. In an unlimited war of pacification the totalising tendencies of the logic of the circle of violence join the racist radicalisation of the sense of superiority, which the members of the colonial society, especially settlers, plantation and even mine owners, share. The results are a military strategy of massacre and scorched earth, which exposes the victims to starvation and is part of a more general dehumanisation ofthe opponent. The latter prevails in the destructive objectifying of the 'enemy' and in an external morality, in which either the rules of internal morality are no longer valid or are so compromised, that nothing remains ofthem. Not only are the infamous proclamation of General von Trotha to the Herero on the 2 October 1904 and the ensuing exchange of correspondence with his superior, the Chief of the General Staff, Count Alfred von Schlieffen, proof of this; in much the same way the people and the whole area of the Maji-Maji uprising were similarly wiped out. Faced with the famine in the East African territories which had been destroyed by the burnt earth policy, Captain von Wangenheim (quoted in Gwassa 1972: 140) said: "That's quite right, the fellows can just starve. We'll make sure that we have enough to eat. If I could, I would even stop them from planting anything. That's the only way to make these fellows tired of war.,,23 The unlimited war of pacification in the colonies of land-hungry settlers and paraslave-owning planters has a genocidal tendency. How far this genocidal tendency can prevail depends in colonial wars to a great extent on the policy the colonial headquarters in the 'mother country' have adopted, and what kind of metropolitan lines of contlict there are. Under conditions of long-Iasting military conflict there is an incisive change in the structural 23 According to areport in the Kingonsera Chronicle of February 1906. Captain von Wangenheim's words were reported by Captain Richter in Songea; Gwassa did not make this very c1ear; cf. Seeberg (1989: 79-80).

Forms ofMartial Power: Total Wars, Wars ofPacification, and Raid

49

relationship between metropolitan headquarters and colonial governrnent. The colonial govemment loses a large part of its independence, the ''true rulers" of the colonial empire, to use a term coined by Robert Delavignette (1939), that is the members of the administration in the colonies themselves and, first and foremost, the Govemors, become directly dependent on the political confrontations which go on in the politicallife of the 'mother country'. The forms of this shift in the balance of power between the colonial administration in Europe and the governrnent in the colony are many. There is a metropolitan politicisation of colonial rule. This politicisation strengthens the ideological side of armed conflict and increases the amount of direct intervention from the metropolitan colonial headquarters. In German colonial history all three developments are exemplified in the role, which General von Trotha played vis-a-vis Govemor Leutwein in German Southwest Africa, and in the socalled 'Hottentot election' in 1907. In this respect, it is possible to say in the categories of a political-military 'dependency theory', that wars of pacification, which are long-Iasting and escalate, are 'peripheralising processes' of colonial rule. The metropolitan headquarters acquire more decision-making power. The colony is involved in a process of becoming a militarily and politically 'peripheralised area'. In brief, in settlers' and plantation colonies there is a tendency for the limited war ofpacification to develop into an unlimited one. Above all, when there is a long-Iasting armed resistance by those subjugated to the rule of the conquerors, war becomes peripheralised, and the waging ofwar is more closely linked to the situation of political conflicts in the 'mother country'.

4. German Colonial Wars in Africa: Were They Total Wars? What differences exist between the colonial wars of pacification and the total war of industrialised nation-states? Where are the tendencies towards the totalisation ofcolonial wars ofpacification to be found? There is the degree of mobilisation of the population, which has to be considered. In the case of the German colonial wars, it remained Iimited. On the German side, the mobilisation was far from being that of total war. Disregarding the narrower political conflicts, the population in Europe remained untouched by the war. For service in Africa it was possible to limit recruitment to volunteers in Imperial Germany. The limited character of the wars was maintained even at the height of the success of the African resistance, when the tendencies towards unlimited wars grew stronger, and there were moves to set up militias of members of the European colonial society (cf. Seeberg 1989: 65). Moreover, in the colonial wars of pacification there were basic structural limits to the direct militarisation of the society to which the conquerors belonged. Thus, the political context of colonialism, the limits of the peripheralisation 4 Sociologus, Beiheft I

50

Trutz von Trotha

process, the limited organisational power of the colonial rulers, and the wars themselves, which continuously proved the superior strength of the conquerors, all prevented the process of the militarisation of society from being put into practice. The degree of mobilisation of the affected African and German populations obviously varied a great deal. On the African side in certain regions and at certain times, more or less large parts of the population were involved. For example, at the height of the Herero and Maji-Maji uprisings the war involved the members of the affected societies completely. For the Africans, resistance had the features of total war in stateless societies. At the same time, however, the violent resistance ofthe Africans retlected the lack of inter-ethnic and intertribai organisational power of the African societies. It could therefore be considered limited from the point of view of the conquerors, provided that interethnic and inter-tribal organisational power did not cover the entire area of conquest at any one time and thus never endangered the conquest itself. The picture of the African side is much more complex, since the degree of mobilisation depended on the political structures. On the one side, it was affected by the degree of centralisation of the societies involved in the fight or considering resistance against the German conquerors, on the other side, by the contlicts existing between different autonomous political units. Nevertheless, in the struggle against the German conquerors, both in German East and Southwest Africa, there were not only trans-local and trans-tribal relations, but also the pattern of guerrilla warfare entailed large parts of the population being drawn into the war, and that the dividing line between 'civiIian' and 'soidier' disappeared. 24 Thus, on the side of the German conqueror the colonial wars were surely Iimited wars of pacification, whereas on the side of the Africans the wars had a totalising tendency with respect to the degree of mobilisation they involved. The situation is different when the victims are considered. In comparison with the degree of mobilisation, the approximation to the features of total war between the German conquerors and the Africans who offered armed resistance tend to be the other way round. The deadly defeat or even annihilation ofthe Germans in the colony was beyond any realisable goal of African resistance. Even at the height of African military success, the major military aim of the Africans could be no more than to force the Germans to forego their military conquest and leave the country (cf. Gwassa 1972: 136). In order to attain this end, armed African resistance typically did not indiscriminately attack all whites who lived in the colonies. In German Southwest Africa, Hendrik Witboi or Morenga Iimited any extension of 24 That caused the Governor of East Africa, Adolf Graf von Götzen, to speak of a "people's war" (cf. See berg 1989: 72).

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

51

the categöry of victims, which the German strategy of massacre and scorched earth suggested, by repeatedly mentioning and earrying out the praetiee of sparing women, children and prisoners of war. Nevertheless, members of European colonial society were still very much at risk and fell victim to African attacks. With the development of unlimited wars of paeification, on the German side the divisions between 'civiI population', soldiers and prisoners of war beeame increasingly less distinct, not only from a tactical, but also from a moral and strategie point of view. In addition to the numerous massacres which were committed by the conquerors in all wars of pacification and the strategy of scorehed earth, which the Germans used particularly in the Maji-Maji war, the strategy and policy of General von Trotha are once again exemplary. Von Trotha not only drove the Herero to their death after the battle of Waterberg, but also developed the idea ofan irreconcilable "racial struggle".25 The eonquerors brought to bear the occidental, industrially based technology of war. The superiority of European weapon technology always changed armed conflict into massacre for the Africans, whenever, and this was mostly the case, they were unable to equip themselves sufficiently with European weapons and

25 The different ways Gennans and Africans fought the colonial wars demonstrate that in one and the same military conflict the actors may fight very different types of war. That is a typical feature of war, especially in cases of military conflict between insurgents and government troops or in what might be called modern 'intervention wars', in which foreign armies fight insurgents in another country usually on behalf of the officially recognised government.

The simultaneity of different types of war raises very difficult and highly complex conceptual and methodological questions. It means, for instance, that a sociological analysis of war must reckon with the definitions by which the actors involved qualiry an ongoing military conflict. These definitions are not only relevant in view of the legitimising strategies of the parties, but also with respect to the objective characteristics of war. The case ofthe Gennan colonial wars in East Africa is a good example ofthe kind of complexities the analyst has to take into account. From the Gennan perspective, the Gennans fought a war against 'the' Africans, which implies the idea that there is a unity of 'Africans', which, in fact, did not exist considering that 'the' Africans constituted a world of great political, cultural, social and even economic heterogeneity. Consequently and due to the Iimited inter-ethnic and inter-tri baI relations of their opponent, which are part ofthe heterogeneous African world, the Gennans did not consider themselves to be faced with total war (as two villages or nations can wage a total war against one another). Instead, they thought themselves to be fighting local groups of insurgents. Thus, from the point of view of mobilisation, what developed into total war for individual African societies was understood by the conqueror to be regionally and ethnically Iimited conflicts; this is the reason why mention was made of "bands" and uprisings of individual "tribes". 4*

52

Trutz von Trotha

did not adopt the tactics of guerrilla warfare. 26 In addition, the technical improvement ofiogistics (particularly the use ofthe telegraph) gave the Europeans important tactical advantages. Nevertheless, the technological side of the form of the colonial wars of pacification should not be overestimated (cf. Wessling 1989). The supply of the most modem weapons was largely limited to the European officers and NCOs. The Africans in the German colonial armies had to be satisfied with antiquated guns (cf. Gann and Duignan 1977: 119). Equally important was the military organisational power of the conquerors, who waged their wars with a professional army commanded by highly trained and motivated officers. Above all, neither on the German and certainly not on the African side the wars were 'war machines'. The more the Africans countered the Germans with the tactics of 'bush war', i. e. guerrilla war or guerrilla-like tactics, the less likely was the outcome ofthe battle determined by technologies of deadly weapons. On the contrary, the Germans forced the unconditional surrender of the Africans by using non-technological war techniques, namely the strategies of 'scorched earth' and hunger. The restricted importance of weapon technology is also documented by the fact that the German losses, as in any traditional war, were to a great extent not due to immediate armed conflict but rather to sanitary and hygienic conditions and fatal diseases (cf. Wesseling 1989; Gann and Duignan 1977: 119). Accordingly, the wars of pacification were hardly an incentive to invent and develop weapon technology. The Germans used the technologies which they had at their disposal and deemed sufficient for victory in a bush war. Specific innovations, which were not even technological in the strict sense of the word, such as the deployment of a camel troop, were limited. With the important exception of the 'Maxim' machine-gun, the totalising tendencies of the war were rather more political and ethical than technical in nature. For example, the Germans used dumdum bullets in order to psychologically terrorise their opponents. Clear tendencies towards the totalisation of colonial wars are found instead in instances where limited wars of pacification developed into unlimited ones. That was the case in German East and Southwest Africa. The unlimited war brought to the fore those tendencies of the war culture of the modem European nation-state in which war is presented as unrestricted victory, annihilating defeat, decisive battle and outstanding military leadership; again, Clausewitz's theory of 'pure war' has provided excellent examples of these ideas. In the strategy of exposing entire populations to certain death by hunger and thirst, unlimited wars of pacification are radicalised to the limit and take on the genocidal trait oftotal war. 26 David Killingray (1989: 153) points out, that the Europeans, especially after the Brussels Conference of 1890, were relatively successful in preventing the Africans from equipping themselves with modem weapons.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

53

The treatment of the defeated approached genocidal warfare. Men, women and children were deported, intemed in camps, which were called "concentration camps" following the example of the English Boer War, and died there of hunger and diseases, which they did not have the strength to res ist due to the appalling conditions of the camps (cf. Bridgman 1981: 165). In German Southwest Africa not only the political and social, but also the economic structures of the Nama and Herero were destroyed in the interest of an enslaving 'worker policy'. In the unlimited wars of pacitication, the ideological context of war came to the forefront. Ideologies themselves became increasingly irreconcilably determined by the ideology of 'racial war'. The sense of racial and civilised superiority and the contempt for the cultures and life styles of the Africans, which were an integral part of the idea of the whites' civilised mission and had their social equivalent in the colonial situation, developed their destructive potential. They were transformed into the conviction that resistance of the colonised peopies should not only be broken, but that complete destruction of peoples, cultures and social organisations, where armed resistance was rooted, was necessary. The ideologically based complete inditference to the sutferings of the victims tumed into total violence. At the end of the war in German Southwest Africa a policy of controlling the defeated was implemented, wh ich accorded German Southwest Africa the awful privilege of being the only proof of Hannah Arendt's thesis that the foundations of totalitarianism could be seen to be emerging in the colonial policy of Africa (cf. Gründer 1985: 124; Arendt 1986: 307-324).27 However, unlike the two world wars, where the National Socialist war of destruction became the epitome of modem total war, the ideological totalisation of war always remained restricted even in unlimited wars of pacitication. The interests, attitudes and ideologies, not only on the metropolitan side of colonial policy, but also in the colonies themselves were too beset with conflicts. Baron Albrecht von Rechenberg, Govemor of German East Africa, attempted to overcome the devastation left after the Maji-Maji war by an ambitious colonial political programme, which wished to have nothing to do with settlers' and plantation colonies and was intended to make the colony a "country for merchants, Indian traders and native cultures" (quoted in Gründer 1985: 163). Also in German Southwest Africa, totalitarian dreams were shattered when they came up against the reality of colonial rule and its economy, which depended on being drip-fed by the Imperial Budget and consequently saw itself bound to win over the Africans as independent producers. This was impossible in a forced labour economy, as the Secretary ofState, Demburg, repeatedly pointed out.

27 Fortunately and admittedly, the goals were purely illusory in view of the possibilities of colonial rule.

54

Trutz von Trotha

War ofpacification is aimed at the conquest of and rule over territories and people. Total war of conquest is characterised by extreme indifference to the sufferings of people and by complete indifference towards those who live in the conquered territory. lt is aimed at the conquest of certain areas and, provided it does not immediately wipe out the inhabitants ofthese areas, it does not wish to rule over people, but rather to enslave them, to rule over people who have been radically dehumanised. The colonial wars of Imperial Germany in Africa remained within the limits of wars of pacification. But where they became unlimited, features appeared which belong to the reality of a total war of conquest.

5. What is Raid? For centuries the 'raid' or 'razzia .28 has been shaping the relationships between societies, groups within societies, and the political and economic landscape of vast areas. It is a pretty stable form of martial power, at least as stable and important as war. lts history and sociology has not yet been written. Within the framework of the observations and thoughts of this paper, my aim is very modest. I will only look at sorne features of the razzia in comparison to what I have said about war thus emphasising the idea that the raid is a proper form of violence different from war as weil as from other forms of collectively organised or 'individual' violence. My remarks rely basically on material based on West African camel or horse raids and the maritime razzia ofthe Mediterranean (cf. Klute 1994a; 1994b; Spittler 1984; Earle 1970; see also Keegan 1995: 265306). Contrary to war, the raid is usually not a collective and organised use of the combination of material harm, absolute and total violence. Instead, it is an organised use of the combination of two forms of power: material harm and absolute violence. It is not constitutive for the razzia to be organised and carried out collectively. The collective form as, for instance, in the case of the yearly raid ofthe nobility of the Anufom of northem Togo against their neighbours (cf. Trotha 1994: 256-259) is only one of several forms, whereby the boundaries between war and raid become blurred. All the more, the razzia is the realm of organised groups of more or less adventurous men led by martial entrepreneurs. The entrepreneurs organise and maintain the raid, even in the case of the Anufom, where only some forms of razzia were collectively organised; they belong to the kind ofpeople the Maltese or Moslemic corsairs ofthe Mediterranean and other 28 For the sake of Iiterary variety, I use interchangeably the tenns 'razzia' and 'raid'. The word 'razzia' is of Arabic origin meaning 'attack', 'military undertaking' as weil as 'raid', 'marauding expedition', 'plundering', 'predatory excursion'.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

55

areas of the high seas were recruited from; they were the men, who by the way of razzia set out on their violent road to farne, power, and above all wealth. The organised, but yet not collective, character of the raid is weil demonstrated by the line the Tuareg nomads draw between a war and a razzia: in war even the slaves and freed slaves are obliged to participate, while in the case of raid the decision fell to them (cf. Spittler 1984: 151). The more individualised as weil as entrepreneurial character of raid is weil expressed by the English word for corsairs, who were called 'privateers'. Thus the maritime razzia, which has deeply shaped the collective conflict between the Islamic and Christian worlds, between the Ottoman Empire and the expansive states or parastates of northem, middle and southem Europe, was the business of martial entrepreneurs more or less politically and economically licensed by their overlords. In view of the collective character of organised power, war and raid may even represent two opposite forms of martial power: on the one side, total war with its allembracing mobilisation ofthe members ofsociety, on the other, raid tied to the initiative of daring and greedy individuals. No less important are the facts that razzia is not bent on total violence and establishes a clear hierarchy between absolute violence and material harm, where material harm definitely ranks before absolute violence. Raid is a proper form of economic action as weil as violent power. It is an action at the intersection of economic action and violence; it is a combination of economic action and violent power, it is violent commerce or martial robbery.29 Giving preference to material harm guided by material aims, the raid is determined by rules, which limit material harm and absolute violence - similar to limited wars of pacification. Razzia is regulated violence, which includes that the degree of unpredictability is less dramatic than in war. Raid can even be regulated to a degree, which approaches the level ofthe feud ofthe late Middle Ages in Europe, wh ich was highly formalised and a proper legal institution (cf. Srunner 1973). As Earle (1970) has so weil shown, there were, for instance, rules which regulated who was entitled to practise the maritime razzia, who may be taken as prisoner among the crew and the passengers of a ship which had

29 Claude Uvi-Strauss (1943) once conceptualised war as "the result of commercial transactions wh ich failed" (quoted in Klute 1994b: 1). With respect to war, this notion is a kind of a theoretically misleading economics. More valuable is Uvi-Strauss' notion when we analyse raid. But there still remains a euphemism. It concems the economic market, for which Uvi-Strauss claims that it denies war. Indeed, that may be the case as, for example, Napoleon Chagnon (1971: 144-149) has documented so weil in his research about the Yanomamö (cf. Trotha 1986: 7-8). But that is only one of several relationships between war and commerce. There is also simultaneity between commerce and war; commercial transactions may prcpare the grounds of war, a case which not the least wars ofthe colonial conquest stand for, as weil as unbalanced commercial transactions may be an occasion ofwar.

56

Trutz von Trotha

been seized, how to deal with the cargo of a friendly ship, which was either sent from or destined for a hostile customer. There were insurance systems and a whole network of institutions which supported those who wanted to buy back the men, women and children, who had been captured - including the slavemarkets of important ports of the Mediterranean. Individualised, grounded on and maintained by economic interests and other institutionalised personal motives, not centred in absolute violence, and regulated to a high degree, the raid can forego more or less elaborate political and cultural legitimisation. As long as the chances for economic success, social farne and social mobility, which may go with them, seem to be rather reasonable, razzia gets round the central questions ofwar, which, as mentioned above, address the problem of killing and dying. Certainly, there is no power without legitimisation; this holds good particularly for violent power (cf. Popitz 1992: 17-20; for a possible critique, see Spittler 1976). But within the limits of this general need for legitimisation also accompanying razzia, the instrumental character of raid emphasises the interests of martial entrepreneurs and their overlords. For example, the maritime razzia could even be maintained while it was officially condemned - a case which is rather at odds with the needs of legitimacy ofwar. 30 Unlike war, raid is not directed towards the annihilation of people. On the contrary, similar to limited wars ofpacification and to protection rackets,3) raid is confronted with the problem not to destroy its own base by plunder and destruction to such a degree that the victims leave the area for good or look for less dangerous routes not threatened by deadly or otherwise harmful attacks. 32 As Earle documents, and my own studies about razzia in pre-colonial northern Togo seem to underscore, this problem was not solved for many periods in the history of the razzia. Raids either devastated the privileged areas of attack leaving them behind with no future hope for more human and material booty, or forced the people to migrate to areas which were out of reach of the violent attacks or allowed for better chances of defence against the ferocious onslaught of martial entrepreneurs. 30 I cannot deal here with the special case, wh ich has come to be known as 'dirty war', and can be defined as a war, which is not officially legitimised. I will only call attention to the question conceming the character of 'dirty war', which might not be a war at all, and may belong to the realm of terrorism and counterinsurgency and similar forms ofviolence. 31 In fact, raid can function like a protection racket, as Erdmute Alber has shown so convincingly in her contribution to this book.

32 Not giving priority to absolute violence, but to material harm, razzia can be understood as a system ofredistribution ofmaterial wealth (cf. Alber in this volume). Consequently, it is restrained by the fact that it has to watch over its destructive consequences. Where there is no more to redistribute, the razzia collapses.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

57

This problem has never really been solved definitively, because it is intensified by a feature of the razzia which it shares with most forms of violence, but which it accentuates strongly. Most violence means movement. That is valid, in particular, for the basic anthropological form of violence, and related to the feature of speed I have talked about above. 33 In line with the basic anthropological features of movement and speed, razzia is foray and invasion. The predators move fast, they seem to suddenly come from nowhere; they plunder and take everything which is not red hot or nailed down; part of the booty is squandered in luxury on the spot, because raiders can only remove what they can load on their ships, horses or camels. Most of the time, they vanish as fast as they show up. Razzia always includes highly developed logistics of movement. But it is a special kind of movement. It is 'careless movement' (,sorglose Bewegung~, as I call it. The raid does not care about tomorrow. Its realm is the present. This orientation at the present delimits the razzia from another form of power, from bureaucratic domination. Bureaucratic domination is the most developed form of 'stationary power' (,stationäre Herrschaft~. Consequently, it is careful power, even more so in its modem version. It is solicitous about and obliged to the future. Tied to its local environment, bureaucratic domination has to care about its resources and the world of its subjects, which have to be modelled according to the requirements of bureaucratic administration. Instead, domination by raid enjoys the carelessness of those who have the freedom to move. As long as not all major resoutces of an area are plundered and destroyed, the raiders can move on - to bring death and destruction to the next village, the next port (for further details, see Trotha 1994: 61-79).34 With its carelessness, razzia is Iinked to certain cultural patterns, among which the chivalrous virtues and vices of nomads and corsairs are suited for the creation of legends, and which characterise the self-image and stereotypes of others of nomads and corsairs. Characteristically, these self-images and hetero33 Certainly, there are fonns of violence Iike certain f~nns of torture or the industrialised killing of the Nazi annihilation camp, which do not realise the basic anthropological fonn of movement and increased speed. But even these cases are the result of phases of violence based on movement and heightened speed, for instance, when the victims are hunted down, rounded up and transported to the camp (cf. Sofsky 1997; in fact, Sofsky uses the tenn 'razzia' for these kinds of violence. Though he notes, that this hunt for humans is rather different from the phenomena I am talking about here, his tenn, which is taken from the context ofpolice, reminds us about the similarities, which exist between the old historical fonn of razzia and the modem 'razzia' of the police, especially, ifthe modem police razzia is not regulated by laws of due process and other constitutional safegards). 34 To use a somewhat risky metaphor, modem bureaucratic domination is puritan, razzia is hedonistic.

58

Trutz von Trotha

stereotypes are organised on the grounds of dichotomies, wh ich, as 1 have mentioned, belong to the realm of violence: cruelty versus pity, magnanimity and prodigality versus mercilessness and greed, passion versus cold-heartedness, whereby the coldness of the heart is just another expression of the consuming fire inside the restless raider. Different from war by the high value it places on material harm and the fact that it is relatively more regulated, delimited against bureaucratic power by the principle of careless movement, razzia is a form of violent power between war and bureaucratic domination. Through its violence and destructive consequences, razzia is a relative of war to which it can turn into, especially, if the victims mobilise resistance. But for many centuries, it also became a stable element of two forms of domination: of despotie rule and the kind of domination 1 call 'paramount power' (,Oberhoheit'), which is the way the traditional forms of empires (, Großreiche ') normally were ruled beyond the borders of the territorial centres of domination (cf. Trotha 1994: 287). Typically, despotie rule uses a variation of the razzia35 in order to exact taxes and to recruit soldiers. It usually goes along with paramount power, which limits itselfto control more or less vast territories without interfering with the inner affairs of the territories under control. There, raids are the means to secure a general political and economie control, and to demonstrate the might ofthe ruling centre. Situated between war and bureaucratic domination, razzia does not know either beginning or end. "It was an eternal war", as Earle says (1970: 3; emphasis i. orig.). War is determined by the cycle of war and peace. (cf. Mühlmann 1940) Like the feud, the order of raid keeps people under more or less continuous strain. When the raiders pass out of sight of the robbed and injured, and the people leave their hiding places, the bitter and poignant knowledge remains that the raiders will return some day - maybe next year, when it is time to harvest again, in the ominous double-meaning of the word. For raiders as weil as their victims, the end of the razzia is only a reprieve within a recurring cyc\e of mar. tial power. 6. Towards a Theoretical Ethnography of Violence

The preceding analytical efforts hope to contribute to the development of a 'theoretieal ethnography' of violence, as 1 call it, whieh I regard as the methodological and theoretical core of the sociology of violence, which the sociology of martial power is part of. The theoretieal ethnography of violence is an answer to futile causal questions of the conventional sociology of violence and war. In emphasising the replacement of causal questions in favour of questions 35 It is a variant, which lies between the razzia 1 am mainly tal king about and the police razzia of modern states.

Fonns ofMartial Power: Total Wars, Wars ofPacification, and Raid

59

which focus on analytical description, the theoretical ethnography of violence substitutes the etiology ofviolence and war for a "thick description" ofviolence or martial power in the sense of Clifford Geertz' (1994) paradigm, and, at the same time, stresses the theoretical task of developing sociological concepts of forms, processes and 'orders of violence' (cf. Trotha 1995b). The theoretical ethnography of violence starts from the premise that the key to the phenomena of violence is found in the forms of violence itself. To understand violence means to study violent practices in great detail; in order to establish theories of violence, one cannot help but analyse violence, what violence itself expresses, makes possible, and gets going. What is at stake is the logic ofviolence, i. e. the basic processes and elements of violence and their relationships, the sufferings of the victims, the perceptions of the actors and sufferers, their thoughts, feelings, and emotions, the relationships between the violent actors, their accomplices and victims. Substituting the etiological questions with descriptive and analytical questions, the theoretical ethnography of violence turns against the relatively static and deterministic view of human reality and gives priority to a dynamic and also historical perspective of human affairs, in which social actors and social action enjoy some kind of privileged status within the theory of violence. In line with what has become common knowledge in sociology, one could say that the theoretical ethnography of violence emphasises the processional and constitutive character ofhuman action. The theoretical ethnography of violence is anti-reductionist, directed towards processes, based on 'conceptual coding', and aims at the discovery of the structural and processional 'logics' of social and cultural forms of violent power. The theoretical ethnography of violence is anti-reductionist in two ways. As a variation of thick description, it is microscopic. It follows the pattern of ethnographic studies of micro-worlds of social actors. It takes a social event and analyzes the social actions in view of all the relevant social arrangements and social relationships, symbols, emotions or ideas, which constitute the logic of social action and the meaning of the event. As mi~roscopic description it focuses on social processes, because processes constitute social events and are the very nature of social action and interactions. The description is 'thick', because it tries to disco ver a maximum of social and cultural relationships which are essential for the understanding of the social event at hand, which is to be conceptualised as a "total social fact" in the sense of Marcel Mauss (1978: 11-19). In addition, it is anti-reductionist, because thick description has the task of reconstructing the social and cultural logics of social events and the actions by which they are produced by way of looking for those elements which transcend the realm of the event and relate them to more general processes and structures, of which they are part. To say it with the categories of a remarkably artificial and troublesome sociological debate: thick description connects the micro- and

60

Trutz von Trotha

macro-Ievels of social and cultural realities. Thus, theoretical ethnography proceeds from a thick mieroscopic description to a rich macroscopic analysis. The key to relate microscopie description and macroscopie analysis is 'conceptual coding', to modify a term of "grounded theory" (cf. Strauss 1990: 17, 17-36,55-81). That is, what I have modestly tried to do in the preceding notes. Being mostly identieal to what Strauss said about coding, conceptual coding attaches great importance to 'inductive' methodology. In line with ethnography and phenomenological analysis, and contrary to the 'deductivism' of the categorieal analysis of conventional general sociological theory and of the construction of 'models', it starts with a detailed analysis of empirical phenomena. It sets great store by the emic categories of actors and the definitions of situations on the part of the actors. But it differs from c1assieal ethnography and phenomenology, in that the final analytical aims of conceptual coding are neither rieh description nor the reality of how it is constructed by actors in an interpretative process. In conceptual coding, the aims of the former are only steps within a process of categorising observations and conceptualising the relationships between observed phenomena. Conceptual coding is directed towards the production of general sociological concepts, a set of etic terms, which are able to define and make c1ear the general aspects of partieular social events. In the same way, it slightly differs from Strauss' concept in so far as it starts with anthropological questions, underlines the etic categories of the ob server, and aims at the discovery of the structural and processional logics of social and cultural forms, which in our case is martial power. Conceptual coding does not content itself what Strauss regards as the goal of grounded theory, which he describes a little bit ambiguously as "a theory that accounts for a pattern of behaviour which is relevant and problematie for those involved". (ibid.: 34; emphasis by TT) Instead, right from the beginning to the end of a particular theoretical effort, conceptual coding is theoretical by posing theoretical, i. e. anthropological questions, pursuing them by the way of thick description, and ending up with concepts which are historically meaningful and tied to the meaning of social actors, but are objective and general in the sense that they represent the view of the observer and thus reconstruct the objective logies of forms of social action and of cultural meanings. For the sake of brevity, the preceding observations and thoughts left out to a large degree the level of mieroscopie description, the core of thick description. Instead, I concentrated on the more abstract level of general concepts in order not only to sort out some of the major dimensions of various forms of war and raid, but to emphasise the fact that for understanding and, to use this ominous word, 'explaining' war and raid we do not have to look for the 'causes' of violence. On the contrary, what we have to do is to reconstruct patiently what war and raid are in the first place.

Societies in Civil War By Peter Waldmann

Introduction I

Although civil wars constitute dramatic and painful life experiences for the people affected and even though no major region on earth has not witnessed in some way this form of collective catastrophe, it is a striking fact that social sciences have given only little attention to this subject. We are not talking of a lack of monographs on civil wars such as those of North American or Spain, but what is missing, are systematic, comparative studies, where pervasive and general features of this form of war have been examined. 2 One reason for this situation might be the difficulties that arise when one tri es to define the phenomenon 'civiI war'. If we take for example the current definition of civil war, that is a "conflict carried out with means of violence by single groups within a state" (Noack/Stammen 1976: 29), the question arises as to the differences to other forms of violent conflicts within astate, such as revolutions, uprising or coup d'etat. The reasons for the lack of comparative work becomes more evident when we resort to the Anglo-Saxon definition which specifies that we can only speak of civil war, if a considerable number of people are involved in a long-standing conflict within astate (Bogdanor 1987: 106). As we can guess ftom these features, civil wars represent phenomena of their own kind, often prone to run over their initiators and to get quickly out of contro\. It is possible only to a Iimited extent to explain them with the methods of causal and final analyses which are traditionally applied in social sciences. The roots and original motives of long-standing civil wars are often difficult to detect in retrospect; and even more ftequently, in the course ofthe conflict they loose their significance, I Slightly modified and enlarged version of an article on the same topic published 1995 in "Zeitschrift ftir Politik" 1995, Heft 4, 343-368. The translation has been done by Adrian Waldmann and Alice Walter.

2 An exception to this is a book compiled by Robin Higham (1972). Parting from case studies, this book tackles more general questions. Nevertheless it seems to us that the term "inner war" that was coined by the North American violence school, is of little use, since it is vague and undefined; see Eckstein ( 1964).

62

Peter Waldmann

as the war generates its own causes, which drive it on and on. A case in point is the Thirty Years War, at the end of which not only the common people, but also some of the political leaders no longer knew how it had originated and why it had been fought (Burkhardt 1992: 1). As suggested by the last sentence, the difficulties of identifying reasons and motives of a ci vii war are often related as much to its end as to its origin. It happens rarely that a specific purpose is achieved or that a distant goal is visible for which it seems worthwhile fighting. Ironically, at the end of these energy and money consuming conflicts their agents are very close to where they were when it started. In most cases civil wars have neither winners nor losers (Hanf 1990: 414). The pact or the joint resolution, which grants its formal end, often resemble those agreements made before the outbreak of the hostilities, wh ich were then later arbitrarily breached. We thus can establish that civil wars have to be understood as phenomena in their own right. Detached from their origins and motives, they follow their own dynamic which is driven forward by a form of violence freed from political constraints. Civil wars can culminate and at some moment loose momentum, without it being possible to predict the moment when they are ready for peace. 3 It is hardly possible to manipulate them from outside. In the following article we will try to sketch an outline ofthis dynamic. We will thereby focus upon two types of civil war: Those with a social revolutionary and those with an ethnic conflictual background. The examples of civil wars conditioned by social revolutions are all taken from Latin America, as is the case with the Mexican Revolution (1910 to 1920, although it was named revolution, this conflict resembled a civil war during most of its period), the civil wars in Colombia (with interruptions, from 1948 to date) and in Peru (1980 to 1990).4 The case material on those civil wars which have been motivated by ethnic conditions sterns mostly from Europe and the Middle East. It includes the Lebanese conflict (1975 to 1990), the Northem Irish conflict (1969 to 1998) and the conflicts in Bosnia Herzegovina (1992 to 1995).5 On the concept of"readiness for peace" see Zartman (1985; 1991). It might seem problematic to call the Colombian situation from 1948 to date a civil war. Some authors allow only for the ideologically and politically heated contlicts ofthe Fifties to be labelled in this way. The reader will however be able to see that we start from a concept of civil war that also includes violent but non-political forms of contlict. In this sense, indeed, Colombia has been caught up in a civil war for decades. 3

4

5 It is obviously impossible to know all cases as much in detail as this type of comparative essay requires. Partly, as is the case with North Ireland and Lebanon, the author draws on a former in depth study on a related subject. See Waldmann (1989). Over the last years, he has also extensively researched on Colombia and Peru. Apart from his own knowledge, he could rely on excellent monographs on single civil wars or revolutions. We mention especially the extensive study by Hanf(l990) on civil war in Lebanon and Tob/er (1984) as weil Mansilla (1993). As for the Yugoslavian contlict and particularly

Societies in Civil War

63

First we will brietly comment on some of the recurring conditions of civil wars. Then we will focus on its immediate consequences. The. central part of the essay deals with the frightening auto-dynamic of violent processes. This auto-dynamic has repercussions on all other structural spheres of society. We shall look at their transformation. Finally we will return to the initial question on the lacking function, meaning and sense of civil wars. 1. Recurring Conditions of Civil Wars Although causal analyses do not serve to explain adequately the self perpetuating character of civil wars, it is worthwhile asking, to what extent civil wars are embedded in the historical context oftheir country. The case studies in point show certain recurring constellations at their beginning, which obviously have some sort of causal relationship with those enduring contlicts. These initial constellations seem to represent "necessary" but not "sufficient" conditions, without which it is very unlikely that episodes of violence turn into civil wars. Let us mention three of these conditions: the absence of astate monopoly of violence, historically speaking long-standing social tensions and processes of economic and social modernisation, which stimulate new expectations, generate fears of social descent and increase a general feeling of unrest and discontent. There is no country in South America where the state has been able to impose an effective monopoly of physical power. This is even the case in a seemingly peaceful state like Uruguay which possesses an established two party system and a functioning legal system. But it is all the more valid for the Andean countries Peru and Colombia which are haunted by internal contlicts and unrest (Mansilla 1993: 54; Leongomez 1994: 17). Particular1y in the latter country the state administration exerts only control over apart of the national territory, most of it being located in the centra1 highlands. On the slopes of the Cordilleras, in the valleys and in the lowlands further inside the country, the police and the military have to share their power and controlling function with other armed groups such as guerrilla organisations, peasant militias, and gangs of drug traffickers (Le Grand 1986; Alarc6n 1992). Peru has recently been less frequently an arena of bloody contlicts than its northern neighbour. However, also here, the periodical uprisings of highland Indians and the long-lasting guerrilla crusades indicate that the central government in Lima is far from having subjected and tamed its hinterland.

the war in Bosnia-Herzegovina, the reports and analyses by M. 1. Calic (1993) from the Stiftung Wissenschaft und Politik prooved very useful. I also want to thank here my assistant Martin Roesiger, who was very helpful in finding and making accessible remote material for this comparative study.

64

Peter Waldmann

The absence of a coercive state monopoly becomes evident not only from the numerous cases of explicit unrest. It can also be sensed during longer periods of apparent peace, when the sovereign order for peace seems to be respected. Such periods do not mean that people have forgotten the right of self defence and that they no longer are prepared to make use of it. The experiences in Western and Central Europe have shown that this is only the case when external disempowerment goes along with mental processes of disciplining which in certain cases can stretch over centuries. 6 At the periphery of Europe, in Northern Ireland and Yugoslavia, this disciplinisation has obviously either not taken place or has had only a limited influence on people's attitudes. Particularly the example of these two countries show how little one can trust the apparent appeasement and the abandonment of violent forms of conflict resolution, even when peace has reigned for decades. The second conclusion from the case studies is that civil wars do not break out from one day to the next, and are not generated from a social vacuum, but generally have a background of social, cultural and economic tensions. This is particularly striking with ethnic conflicts. For, even if in certain cases violent clashes between different confessional groups might have provoked astonishment and shock amongst the citizens of these states, it had been clear to them before that they were far from constituting a unified nation. R. Rose found from a research he carried out shortly before the beginning of the troubles in Ireland that Catholics and Protestants were as alien to each other as they always had been. Their everyday habits, religious orientation and political stance were so different from each other that one could weil have feit obliged to speak of two nations that were locked into the same state (Rose 1971: 74ff., 113, 179). In Yugoslavia the Tito regime did not weaken but only at best covered up animosities between the three main ethnic groups (Serbs, Croats, Muslims). Their hostiIities were additionally fed by a change in the constitution in the early Seventies, by which the territorial properties of the three ethnic groups were forcefully locked into a confederate system of order. 7 When this confederation began to break up as a result ofthe demise of Socialism, inevitably the republic of Bosnia-Herzegovina became an object of conflict, with the three ethnic groups colonising it together. It was already at that time when the society broke up into three parties, long before the political haggling would start and be perverted into a violent confrontation. Entrepreneurs, the mass media, the political parties and even the church, resorted exclusively to the criterion of religiousethnic affiliation.

In the case of Latin America, the situation is more complicated. Although on this continent the social structure had also been deeply disturbed, the profile of the conflicts was not as clear cut as it is in ethnically segmented societies. Gen6

See Oestreich (1969: 179-197); Breuer (1986); Rassem (1983); Mansilla (1992).

Oschlies (1992: 31); Calic (1993: 24, 28). See also in Calic 's study the quotation of the Bosnian leader ofthe Serbes Radovan Karadzic: Everybody hates everybody here. 7

Societies in Civil War

65

erally regional disparities intersected and overlapped with social ones. In Peru, for instance, we encounter a widespread distrust by the population in the provinces towards the metropolis of Lima. Limei'ios are accused of absorbing economically, demographically and culturally the hinterland and of looking down on people ofthe provinces as ifthey were second class citizens. This feeling of damage and discrimination was widespread among the provincial bourgeoisie and the Indian lower classes and was one of the main reasons why the guerrilla organisation Shining Path was able to rally support and operate in some regions with its Maoist slogans such as "hem the cities from the countryside". 8 A case which is certainly representative for Latin America is the Mexican Revolution. Originally triggered by discord between riyal upper class groups, it was also fed by regional motives of conflict and social revolutionary impulses which determined in their specific blend the profile of the "revolution" (Tobler 1984: 201 ff., 206, 228). Whereas the absence of astate monopoly and the presence of social and economic contlicts constitute relatively unchanging basic conditions, the third condition, that is processes of social modemisation, represents a particularly dynamic element ofthe conflict constellation. The manifold, intertwined developments of urban modemisation, of general increasing educational opportun ities, of spreading mass media, of economic transformation and of a general change in life style, which as a whole are often labelIed as processes of "modemisation" generally tend to heighten existing tensions. However, in the Fifties and Sixties scientists had assumed the opposite. They were of the opinion that the above mentioned processes would rather help to blur and overcome traditional antagonisms, that they would make people more similar to each other and bring them closer together. Meanwhile people have sadly experienced, particularly in the so-called developing countries, that instead, waves of modemisation are more prone to catalyse socio-political conflicts. For they shake a balance gained with effort and strain, without compensating the alteration they have provoked with a new element. Since these processes of modemisation never affect all individuals and groups of a society to the same extent, but tend to favour some and to disfavour others, since they raise hopes of socially ascending and foster fears of descending, they easily trigger a cycle of mutually stimulating misunderstandings and threat feelings, which drives towards an aggressive discharge. 9 One such case is the Northem lrish conflict. The protest marches of a sm all group of Catholic academics and students who were no longer willing to bear the discrimination of their confessional group, led to a new outbreak of the contlict in the late Sixties. This type of Catholic middle class intellectuals had 8 On the "shining path" (span. sendero luminoso) see the studies by Mansilla (\993); Degregori (1990); Favre (1984); McClintock (1984) and Waldmann (1989b). 9 On the most recent discussion about processes of modernisation and development read Menzel (1992).

5 Sociologus. Beiheft I

66

Peter Waldmann

never existed before. They had only emerged in the course of a systematically expanding welfare state after World War II. Catholic doctors, teachers and social workers, provoked, with their claim for equal treatment in all areas, vigorous defence reactions amongst the Protestant working class which feared for its possessions. Thus the spiral of violence began to move again (Waldmann 1989a). 2. Immediate Consequences ofCivii Wars Latent conflicts and animosities are one thing, open confrontations are another. Civil wars can start innocuously with a little shooting or an arrned robbery. Often however, there is an eclat at their beginning, a social explosion, where accumulated feelings of anger and hate abruptly explode. For instance, the Colombian "bogotazo" in 1949 was such a "bang", which by some authors has been described as the biggest urban upheaval in Latin America to that date (Sanchez 1985: 211, 219). It was the on set ofthe Colombian civil war's first and particularly bloody phase, which was by and large marked by the discord between the two major parties. 1O At the beginning ofthe conflicts a sudden change ofthe extemal appearance of these societies occurs. The ethnic-religious and socio-cultural tensions that before had been present in everybody's mind but had not deterrnined everyday life, now suddenly become visible and turn into the dominant principle which subjects all other spheres of life. A profound restructuring and reordering of society takes place as much in amental and political as in military-geographical respect. The most accurate observation on this bizarre mental metamorphosis after the outbreak of a civil war, was made already 2400 years ago with the occasion of the Peloponnesian War (Thukydides 1995: 262): "In times of peace and under happy circumstances cities and people have better convictions, because they are not caught up in involuntary coercive situations. The war however, wh ich restrains the effortless needs of everyday life, swings the coercive whip of violence and guides the passions of the masses, as the occasion dictates it. Thus the cities were shaken by party struggles and the example of those who had started it, compelled their followers to ever bigger extravagances and encouraged them to make use of unheard means of clever party deals and satisfaction of revenge". Here is already a clear reference to a dynamic of its own and a polarisation of what Thukydides calls "party struggles", once it has taken 10 At the time, following the assassination of J. E. Gaitan, the charismatic leader of the Liberal Party, thousands of workers and poor people streamed from the suburbs of Bogota into the city and attacked everything that in their eyes represented power and establishment: government and administration buildings, banks, churches and monasteries.

Societies in Civi! War

67

possession of social Iife. In an extreme case it can provoke, according to Thukydides, a change of all values: "Recklessness was now considered sacrificed standing up for one's friends, wise restrain was now disguised cowardice, those who kept measure were taken for effeminate, those who resorted to rational thinking, were basically considered lazy and comfortable, but those who beat blindly, were considered real men, etc." (Thukydides1995: 263). Our case studies provide plenty of proofs for these sharp observations which are today as valid as they were in their time. Therefore, it is of secondary importance, whether the escalation of conflicts is caused by rivalry among political parties or by the animosities of wide sectors of the population, whether it is driven by cold calculation or by heated emotions. One can state that in all cases, after the outbreak ofhostilities, conciliatory voices pledging for patience, tolerance and for willingness to talk, rapidly loose influence and audience. Instead, the time has come for fanatics to demand radical solutions. Their line of argument always follows the same basic pattern; it appropriates the increasing fear wh ich in the face of growing tensions takes possession of all groups,1I and is voiced more or less in the following way: One is running the risk of being oppressed and discriminated; the only way to escape this menacing destiny is to preempt the enemy and to resort to armed attack. Occasionally one can find testimonies of the shock which invades the peacefully minded sections of the society in the face of a seemingly irreversible escalation of the conflict. One such example is M. Cehajic, a respected and wealthy Bosnian Muslim, who wrote to his family after his deportation by the Serbs into a camp and shortly before his assassination (Gutman 1994: 167): "Since May the 23rd, when they came into our house to pick me up, I have Iived in a different world. It seems as if everything that has happened to me was a nasty dream, a nightmare. And I simply can not understand how such a thing is possible." The inability to understand in the face of the toughness of suddenly erupting contrasts, as voiced in Cehajic's letter, is particularly characteristic of academics and intellectuals. However, these groups of shocked people represent generally the social minority. Wide sections ofthe population from all social strands applaud the escalation of conflict or at least believe it to be inevitable. Activists from militant milieus rapidly join all sorts of mushrooming militias. The motives for joining a revolutionary army, a guerrilla organisation or an association ofmilitia can be manifold. For some ofthe voluntary "fighters" the prospect of using it as an outlet for their rage on defenceless civilians virtually triggers a rush for power and blood. One gets this impression when reading for example

11 Hanf (1990: 15) emphasises at the beginning of his monograph that fear is an essential driving force, which invades a\l actors involved in the conflict and drives them towards aggressive actions.

s*

68

Peter Waldmann

the reports on assaults of Bosnian Serbs on Muslims in Bosnia-Herzegovina.1 2 Not only did the aggressors imprison and expel their alleged enemies from their long established domiciles, but they thought of all sorts of tortures and humiliations before they killed their vietims. Another pattern ofmotivation can be found in Northern Ireland. The attacks of Protestants against Catholics, who had dared to settle on their "territory", showed features of a quasi sacred purgatory act, as if the gradual intermingling of confessional groups were equivalent to a profanation of the Protestant belief (Waldmann 1989a: 86). Emotional arbitrariness and ideological fanaticism are however not always the main driving forces of the bloody conflict. As has been pointed out by H. W. Tobler, in the case of the Mexican revolution, in the end the material reasons were the ones which made brigades and officials eventually join the rebels (Tobler 1984: 217, 224). The concrete advantages, which may derive from partieipation in military campaigns (whieh are mostly predatory acts), should generally not be underestimated when comparing them to the often emphasised political and religious motives. We have thus already tackled the second aspect ofthe sudden polarisation of forces after the outbreak of a civil war, wh ich is the military-geographie one. It is this component whieh, strikingly enough, makes contlicts irreversible, as it does not allow after the beginning of a contlict to return to the status quo ante. Each of the conflicting parties takes quick and determined possession of specific regions, with the consequence that a territory previously controlled and united by the state is soon broken up into several pieces. This process of spatial division is partieularly visible in the case of ethnie contliets. The ingloriously famous measures of "ethnic cleansing" taken in the Yugoslavian contliet are by no means exclusively typieal to the Yugoslavian situation. Instead, these processes of segregation can be found everywhere, when ethnic societies clash violently together. 13 The general tendency is to build homogenous territorial blocks whieh are safe against attacks as much from ethnic minorities inside as from outside. 14 The blocks are formed from territorial bases, in whieh the respective ethnic group is demographically dominant. These plans are all the more difficult to put into practice, the higher the ethnically mixed zones are in number and the more closely interwoven the different ethnic groups live together are. Particularly, bigger cities which have grown as a eonsequence of a consecutive intlow of different ethnie groups, turn after the outbreak of hostilities into a foeus of eonflicts. The heightened animosities provoked by the spatial eoncentration of different groups of the population can, in some eases, only be contained by walls that divide a city into

13

Gutman (1994: 69); Rathje/der (1992: 45); Zü/ch (1993: 65). Chapter 4 in Wa/dmann (I 989a).

14

This process is particularly weil iIIustrated by Hanj(1990: 418, 429).

12

Societies in Civil War

69

several sectors, as is the case in Belfast where the wall is euphemistically called "peace line". Generally the desire of each group to consolidate its territory as quickly as possible generates an aggressive hecticism when civil wars start. Later on, when prospects of gaining territory become exceptional - for instance in the case of extemal help -, since the different groups no longer allow themselves to be run over, then initial attacks give way to a strategy of defence (Hanf 1990: 431). The division of a country into different zones each respectively controlled by a civil war faction is of far reaching importance to each single citizen. Oepending on where he Iives and works, he may feel safe with his family or may be weil advised to leave his horne and to move to the horne territory of the group to which he belongs. Civil wars always provoke massive migrations. Ouring the Thirty Years War, hostilities led to a major migration of the rural population into the cities, where the walls gave a feeling of safety (Schreiner 1985). In the first phase of the Colombian war, which was marked by the conflicts between the to big parties (liberal and conservative) the safety of a person depended decisively on whether "his" party was militarily speaking dominant in the area where he was living. Liberal minded farm holders whose lands were located in a conservative area, were keen to move away quickly and to rent or seil their land. Apparently two million Colombianes, that is approximately a sixth of the national population at the time, moved house because of political persecutions between 1949 and 1953 (Sarmiento 1986: 267; Mansilla 1993: 155). This figure shows that the proportion of an ethnic group or an entire population which is expelled or voluntarily seeks a new horne during a civil war can be considerable. Ouring the Lebanon conflict about half of the population, that is roughly 1.5 million people, moved temporarily or permanently away from their original horne. Many of them were forced to migrate again later in the conflict (Hanf 1990: 439). In the case of Bosnia Herzegovina the estimated figures of migration are similarly high with 1.5 to 2 million expulsions within 2.5 years, out of the whole population of 4.5 million inhabitants (Rathfelder 1992: 54; Calic 1993: 55). When the zones of dominance of the conflicting factions in a civil war are clear, the options for the migrating masses are c1ear and compulsive: They must leave the area where they represent only a vulnerable minority and they have to move to the respective rural or urban area where either their party, their confessional or ethnic group constitutes a majority able to defend itself. However, it sometimes happens that there are no c1ear-cut border lines and zones of influence, since several groups claim the same territory. All commentators agree that unclear conditions ofthis sort carry the greatest risk for the population here

70

Peter Waldmann

Iiving, as people are at risk of subjugation and exploitation by both sides or of being punished because of alleged alliance with the enemy.15 The amount ofviolence used in civil wars does not escalate continuously. Its characteristic unpredictabiIity (a result of its arbitrariness) finds one of its most important expressions in its discontinuous development. In the same way as certain villages remain completely unaffected while others are badly haunted by the horrors of the war, civil wars have phases in their development where the dynamic decreases strongly and then suddenly escalates again forcefully pushing towards a new climax. These waves may be determined by interventions of extemal powers, but they can also be explained to some extent as an expression of the circular auto-dynamic that violence has in civil wars. The latter will be dealt with in the next chapter. 3. Auto-Dynamics of Violent Processes Our thesis conceming the intrinsic dynamic of unleashed violence in civil wars is not new. This idea can be found in most monographs on civil wars such as those by H. W. Tobler, E. Hanf, H. C. F. MansiIla and others. 16 However, the question arises as to what the terms "process of becoming independent of violence" or violence as an "end in itself' specifically mean. Can one already say that violence is becoming self-sustaining when civil war armies or guerriIla groups start to claim taxes in a region they have occupied? Has violence become an end in itselfwhen battles are carried out, although their original political and ideological goals have already lost importance? Is it tenable to state in general terms that violence is practised only for its own sake as the term "end in itself' suggests? In order to answer such questions it is useful to draw a scheme ofthe different levels on which violence gradually becomes detached from political restrictions and taboos. Such a scheme can only represent ideal types. The concrete cases underlying our considerations do not necessarily belong to only one level but contain elements of several levels or move up and down the "ladder of levels". The starting point is generally a situation where the state possesses the monopoly of violence wh ich inevitably drives political dissidents with radically diverging political ideas to resort to the use of weapons in order to achieve their goals. This starting point is also fictitious. The assumption that violence is merely used as an indispensable means to a specific end, renders it easier for us 15 A good example ofthis is certain high land regions in Peru, which at the beginning ofthe 80s were successively occupied by guerrillas and state security forces. Because of the vexations to which they were exposed on both sides, most ofthe province's population migrated into the safe anonymity of the bigger cities, see Peter Waldmann (I 989b: 184). 16

Tobler (1984: 225); Harif(1990: 416); Mansilla (1993: 148).

Societies in Civil War

71

to pursue and to understand the process of independisation of violence. Thereby, we distinguish three levels of escalation of violence. We shall refer to them as "Independisation of the apparatus of violence", "privatisation of violence" and "commercialisation of violence" . The transition from a merely serving function to the first level, the self independisation of violence, occurs almost innocuously and automatically. The functional logic underlying this development can be described in the following way: Political movements of resistance require long term planning in order not to disappear without effect. For, the further the goals ofthe rebels are away, the less Iikely it becomes for them to reach these goals in one go. This again means that the enemy will adopt measures to face the menacing danger and to arm himself against it. Therefore one can assume that the conflict will be of long duration and will consume huge amounts of energy and means on both sides. The rebels stand a chance of surviving such a conflict only if their leaders operate cleverly, ifthey are weil equipped and supplied, ifthe material and logistical infrastructure is made certain, that means in short, if they are weil organised and if they proceed in an organised fashion. It is however, a wel1 known fact that organisations lead a Iife oftheir own and that they have their own interests, mainly to preserve themselves. Be it a terrorist cel1, militia, a guerrilla movement or a revolutionary army, all these organisations of violence tend to develop in the same way. Once they have been created they tend to develop a dynamic oftheir own and tend to degenerate to apparatuses ofviolence. I7 The moment that triggers this process is generally the financial need of the rebellious movement. If it starts in a province wh ich has its own fees and troops as was the case in the North Mexican movement, then it is already enough that the rebels increase both (Tobler 1984: 20 I, 206). If outspoken enemies of the insurgents dwel1 in the region taken by the rebels, their land and belongings can be confiscated and its gains can be used for the revolutionary army. When such easily accessible, political and morally unproblematic resources are absent, the rebels are often forced to resort to dubious methods for raising riIoney. That happens when they claim taxes from their "fol1owers" in "their" areas,18 when they better their financial situation with bank robberies, kidnapping or bribery or when they engage in shady businesses such as gambling or drug trafficking. Even if those unorthodox forms of organising money might seem justified to their followers as a means and as a response to the difficult situation, they still have two defects which intluence fatally the further development: Firstly, it is a burden which is not accepted by the respective population but is in fact im17 The author has systematically pursued this process of gradual independisation of the apparatus of violence in the case of the Basque ETA. See Waldmann (1990: IO I). 18 The introduction of aso called "revolutionary tax" ranges generally among the first measures taken by a rebellious army on ce it has taken possession of a territory. Apart from covering its financial needs it serves to emphasise its claim for sovereignty. See Hanf(1990: 157); Waldmann (I 989b: 159).

72

Peter Waldmann

posed because of the power monopoly of the rebelling organisation. Furthermore, these unconventional financing practices forecast a mingling of political and private forms of violence wh ich eventually lead to a complete blurring of the distinction between these two forms. Ifthe contlicts have been going on for too long, all rebellious organisations, no matter what appearance they have, must face the question as to how they can justify their existence. 19 The responses to this problem can vary. One answer consists in permanently reaffirming that the danger has not diminished for the group. The argument then goes, that as long as the revolution has not finished and the ethnic possessions have not been confirmed, a counterattack with unpredictable consequences could happen. The best conditions to perpetuate the privileges of irregular fighting fo0ups are situations of balance of power between different civil war parties. 0 If the counterparts shows first signs of weakness or seems even willing to give in, a quick provocation makes sure that it will defend itself. Thus in fact, under the pretence of mutual attack, a "balance ofhorror" is established in which different militant groups help each other to stay alive. A less macabre possibility for rebellious organisations to legitimise their existence consists in assuming quasi-state functions. The logic of this step is obvious: On ce these groups have appropriated an important part of State power by protecting a population group from an external enemy, they also seek to replace the weak or inactive official state in other spheres, for instance in the control of public security and order. The IRA for instance, has assumed in the catholic areas of Belfast and Londonderry important police-Iike and judicial functions. It punishes criminals, resolves family conflicts, supervises the traffic and takes truanting schoolchildren back to school (Schulze-MarmelingiSotscheck 1989: 252; Darby 1990: 83,94). Finally rebellious organisations can respond to the growing alienation from the bases by ignoring and repressing possible protests. This is also the easiest response since it applies the principle most familiar to those organisations which is that of making use of coercion instead of respecting the voluntary consensus. By so doing, they approach the second level of escalation of violence. One can generally observe that the elose connection of organisations of violence with the supporting popular sectors of the rebellion, be they of social revolutionary or of ethnic nature, ensures that the long term goals do not get out 19 For the Basque Country, the increasing ambivalence of the population which originally supported the rebels without reservation, has been studied in Waldmann (1990: 157). Also in the case ofthe Northern lrish IRA the problem is similar. Burton (1987: 82, 88).

20 Mansilla (1993: 148) speaks in this context of a "catastrophic balance"; Hanf (1990: 429) calls it "unstable stalemate".

Societies in Civil War

73

of sight. 80th things are typical of the first step of escalating violence. Even though the apparatus of violence is established and proceeds ruthlessly and without measure, on the first level the obligation towards the high goals of the movement remains unaffected. In contrast to that, on the next step this connection breaks up or becomes at least very unstable. Violence becomes available for various purposes which are not necessarily of political or social nature. Violence becomes "private". There are several indicators that hint at the fact that civil wars have reached the second step of escalation: - The tensions within the political or ethnic war parties increase, often battles between rivalling factions break out which exceed the scope the conflict had with the original enemy. Guerrilla leaders leave their troops and offer their armed services to everyone who pays enough for· them. The entire militarypolitical constellation starts flowing, alliances are quickly established and dissolved depending only on tactical considerations. The original claims and long term goals for the sake of which the struggle had been started, loose importance. Thus a grotesque imbalance is generated between the relatively modest claims and the effort and money consuming campaign ofviolence which is meant to help these claims to be fulfilled. The population is no longer taken seriously in terms of social backing and is not included in the political decision making process. Instead it becomes an object of exploitation, bribery and forced recruitment. In this new context a basic distinction is no longer made between population groups that belong to the enemy's and to one's own faction. The latter are also looted and illtreated occasionally. - Violence is used overtly and wildly, without restrain for private purposes. Many individuals use it to make money, but this is not all. Desire for revenge, envy and jealousy can also be triggering moments for acts of violence. The border line between political violence and ordinary violence associated with criminality becomes increasingly blurred. The same armed group can at one moment appear as a group of fighters for liberation and at another time it may act as gang of bandits. As a consequence of organised coercive violence, real feudal courts come into being, which live on services and taxes of their "subjects". The right of the strongest becomes the dominant means ofsocially imposing one's will. Most features we have presented here are present as much in the Mexican revolution, as in the civil war in Peru, Colombia, Yugoslavia and Lebanon. We will limit ourselves to show them in the case of Lebanon. The conditions in this state during the civil war have repeatedly been apostrophised by T. Hanf as a rule of mercenaries in allusion to the Thirty Years War (Hanf 1990: 423). He observes that the militia, originally recruited from different confessional groups have meanwhile become independent and have developed a style and procedure very similar one to another. They have repeatedly split apart and are incessantly engaged in bloody conflicts with each other, they have taken con-

74

Peter Waldmann

trol of administration and parties, they claim taxes from their respective confessional group without bothering about their steadily declining popularity. In short, they have established themselves as a "state within astate". S. Khalaf goes even further by asserting that after a decade of war, violence has penetrated all pores of Lebanese society (Khalaf 1987: 238). This is evident from the arbitrariness with which objectives and alliances amongst the militia change, but even more so from the fact that violence has become a common means of making and imposing decisions outside ethnic conflicts. He points to the alarming increase of crimes of violence of alI sorts which stretch from vandalism to robbery and murder. More and more frequently, armed gangs of thieves enter the scenario and take away from the citizens the few things they have rescued from the civil war. As we can see from assertions made by both authors, the expansion of violence on this level of civil war generates strong feelings of disapproval and of indignation. A horrific vision of a society a la Hobbes is created where everyone is permanently wary of everyone, but where it is also of great importance that many citizens still vividly remember the times when the norms of behaviour were dictated by law and not by violence. In the third stage of escalation and diffusion of violen ce, this memory is by and large extinguished. In Colombia which is a paradigmatic case for this third level, scandals related to violence are not longer conceived as such. And this is the case, aIthough more than 20,000 people die of unnatural causes or murder in this country every year, more than anywhere else on earth. 21 Many attempts have been made to explain this steady augmentation of violence in Colombia. The most recent theory argues that drug trafficking is responsible for it, since it has transformed the use of violen ce into a business. 22 There might be a bit of truth in this argument, but the drug cartels constitute, as we shall see, only one amongst many organisations of violence. It is likely that in the course of 50 years of permanent political and social conflicts people have got so used to violence as a means of imposing one's will that they tolerate and use it as if it were something normal. When comparing Colombia to other countries which are also not free of violence, we can identify three features which distinguish the Colombian situation from others. It is first the high number of actors of collective violence, second the trivialisation and third the closely linked commercialisation of violence. As for the first feature, in the case of Colombia one is already impressed at first sight by the abundance of groups which are involved in one way or another in violent activities and who mostly make their living with it. Let us

21

See Frankfurter Allgemeine Zeitung, 4th of July 1994.

On drug traflicking and its influences on the Colombian society see Jaramillo ( 1988) and KrauthauseniSarmiento (1991). 22

Societies in Civil War

75

briefly list the most important ones amongst them. 23 Apart from the military and the police which are instances of violence recognised by the state, group violence is extorted by, paramilitary groups who allegedly fight in the countryside against left wing sympathisers; para-police death squads who proceed in bigger cities against criminals and marginal groups (homosexuals and prostitutes); mercenaries and bodyguards in the service of drug traffickers; guerrilla organisations; criminal gangs ofyouths; groups of citizen-vigilantes who chase young criminals; militia who have been hired by farmers for their own protection. This enumeration gives an idea of how violence, on an organisational level, permanently reproduces and renews itself: A closed oligarchie system of party control, which excluded politieal outsiders from partieipation in the politieal decision taking process, generated violent protest at some point. Since this protest was not listened to, it condensed to a guerrilla organisation which operated in the countryside and in the cities. In the countryside it imposed taxes on farmers who defended themselves by forming militias for self-defence. In the cities they trained young "fighters for freedom" in the use of weapons. Once the guerrilla groups got on the defensive, these fighters used their ability to offer their services as bodyguards and killer bands to the newly emerging drug cartels. Besides, it is interesting to note that the Colombian law not only allows for this diffusion of violence, but stimulates it conscientiously by ordering its armed forces to assist the formation of self-defence groups. The ubiquity of actors of violen ce and of violence in general in Colombia has led to a situation were violence, which would be restrained by prohibitive norms and affective barriers in other societies, has here tumed into a "normal" instrument of imposing one's will, a means that provokes no public outrage. Nothing illustrates this everyday, unspectacular feature of even massive actions of violence more appropriately than the proliferation of massacres in this country.24 The term massacre refers to murders which include at least four victims. Between 1980 and 1992, 1030 massacres have taken place throughout all provinces of Colombia. Most of them happened at night in rural areas and caused the death of entire families, most of whom were peasants. What strikes most, is the instrumental character ofthese collective killings. Only rarely does accumulated emotions or political or ideological fanaticism co me into play. By far the most frequent variant are massacres driven by "social" or "economic" motives. The possibility of having someone killed by commission constitutes another important difference from the proceeding phase. While the use of violence on the second step of escalation opened new opportunities for personal enrichment and acquisition of certain privileges, has it now become a commercial service, a 23 See Heinz (1989: 351, 351). 24

On the following paragraphs see Uribe ( 1994).

76

Peter Waldmann

professional business. A person who is after someone else's life no longer has to kill hirn in person, but can commission someone to do the job for hirn. In Medellin alone there are dozens of "bureau's", which live rrom such commissions (Osorno 1993: 72, 85). It is sufficient to give them a photo of the prospective victim and to pay in advance half the agreed price, which may vary considerably depending on the rank and the bodyguard situation of the indicated person. The "bureau" takes care of everything else. Several thousand professional killers, called sicarios, live in Medellin. Most of them are youngsters between 13 and 25 years ofage, who dream ofthe great "Coup", the murder that will make them rich overnight. 25 The Colombian case shows that the spectacular deviation of violence away rrom the political sector as it is typical for the second step, is followed in a further step by the profanation of violence and its pervasion of human interaction in every day life. Therefore the stepwise sequence of diffusion of violence must not be understood as a continuous process of independisation and of loss of function, but should instead be looked at as a process of spiral and dialectic development: Violence, initially contained in the political area, transcends in a first step the borders of its restrictions ("privatisation"), in order to achieve the second step, a predictable, fixed value, in the sense of a purchasable service, on the exchange market of social relations. It goes without saying that this development is not inevitable and that its process is neither fast nor goal oriented when taking off. The multi-step scheme serves to illustrate that it would be misleading to deny a civil war any dynamic, to look at it only as a meaningless massive extinction of people which follows no rules and carries on until all participants are tired and return to the negotiating table. Civil wars have a dynamic of their own, which explains itself rrom the particular logic of self perpetuating and expanding processes of violence. And again these processes of violence exert their impact on the structures of society, economy, politics and culture. These impacts will be tackled in the following chapter. 4. Repercussions of the Auto-Dynamics on other Spheres of Society Opinions agree, for the most part, with respect to the economic consequences of ci vii wars: Almost without exception they are considered negative. The Thirty Years War seems to be a warning rrom historical precedence. Not only did it leave behind a dramatically diminished population, but also empty treasuries burdened with debts, destroyed cities, looted and destroyed monu-

25 Apart from the book by Osorno (1993), which is based on several interviews made with sicarios, their self testimonies compiled by an anthropologist are extremely recommendable: Salazar (1990).

Societies in Civil War

77

ments and devastated stretches of land, wh ich often took decades to be colonised again (Schreiner 1985: 68). However, we have to make distinctions. Not all regions of a country are affected to the same extent by a civil war. Civil wars vary according to their degree of intensity. Depending upon the scope and the density of the fighting, they interfere more or less deeply in the network of production and commerce of a country. The economic structure of astate is also of importance. A small country specialised in commerce, financial services and tourism, as is the case in Lebanon, suffers far more from the enduring bloody contlicts than the geographically vast Mexico with its still undifferentiated economic structure at the time ofthe revolution. The impact ofthe civil war in the case offormer Yugoslavia and particularly Bosnia-Herzegovina was virtually catastrophic (Reuter 1993; Büschenfeld 1994; Calic 1993: 16). In this case we have to keep in mind that the Yugoslavian economy was already declining before the outbreak ofthe disturbances, and that this emergency situation soon got dramatically worse in consequence of the intemationally imposed embargo. Independently from this, it has to be stated that a civil war that rages so furiously and destructively, as was the case in this core republic located in a breaking up Balkan state, inevitably paralysed the entire economy. The industrial production, commercial and service sector, banks, transport, everything was unable to function, only the shadow economy tlourished. Similar but less devastating news came for a long time from Lebanon. In contrast to Yugoslavia, this country had been known for decades as a sort of "Switzerland" of the Middle East, because of its diligent inhabitants, and because of its privileged position as a centre for commerce and finance. The fighting soon led to the loss ofthis privileged position. The damage left behind by the military contlicts inc\uding the destruction ofthe infrastructure (harbour, streets, airport), the division of a formerly tlourishing capital city and the depopulation of vast areas of the hinterland, was too great. At the end of the struggle, that had lasted 15 years, the state had huge debts, since the militia now kept its tax income for itself, the Lebanese currency had falIen to a fracti on of its former value, the trade balance was negative and the average living standard was half as high as in 1975 (Hanf 1990: 449). The list of states whose economies have been badly damaged by civil wars could be extended. Not all of them had to pay as badly for the wars as Yugoslavia and Lebanon. The negative economic consequences of the enduring unrest in Northem Ireland and Peru, two not very weaIthy countries, are probably bearable. In Mexico, H. W. Tobler writes that the revolution has only slightly reduced the average living standard, since the lack of production in some sectors could be compensated by increased production in other economic sectors (Tobler 1990: 478). The case of Colombia shows a totalIy different picture, which confirms its exceptional position with respect to its development of violence. Colombia is

78

Peter Waldmann

one of the economically most stable and dynamic countries of Latin America in spite ofthe internal conflicts that have been going on for decades. 26 In contrast to most countries of the region it has a positive trade balance, the state debts are contained, the GSP grows steadily, as does the flow of foreign investment. In short, a\1 economic indicators have an upward tendency. This does not mean however, that without the bloody internal conflicts economic growth would not have advanced faster. The question arises as to whether the overall positive balance is perhaps a consequence of a general immunisation of the economy towards acts ofviolence. 1fthis is the case, then we can hardly expect any impulses from Colombian economic cireles that would contribute to diminishing the level ofviolence in the country. When comparing the evaluation of economic consequences with the assessment of social consequences of civil wars, the latter turn out to be much more diverse. Some authors make references to the levelling-out effect, others hold occasiona\1y that civil wars are no equalisers, but on the contrary that they make differences between rich and poor grow. Those are probably not real contradictions, both assertions may be correct depending on the area we look at. If we look at the public sector, one can hardly doubt that it is always badly affected by civil wars and that a\1 citizens suffer from this decline in the quality of public life. When for instance, an artillery attack provokes a power cut, water has to be rationed, central trafik lines are elosed, public schools elose or because of lacking state control gangs of burglars increase their activities, then the entire population suffers, no matter wh ich social elass people belong to. At the beginning, the wealthier amongst them manage, perhaps, to partly compensate for the negative repercussions with increased private investment (private teaching, bribery); however, because ofthe steady impoverishment ofthe entire society which ineludes all social elasses, this initial advantage melts quickly away in the course of a long lasting civil war. The decisive levelling-out factor derives from the threat that one may become a victim of the violence that menaces everyone. In the face of the unpredictability of violence of civil wars no one is safe from this threat, as long as he stays in the country. The scenario changes when one looks at the private "sphere" of distribution of possessions. In this case social inequality is generally greatly enhanced by civil wars. This means for the major part of the population a considerable reduction of goods to buy, a deterioration of living standards, and in some cases terrible poverty. These negative consequences can be observed throughout all strands of society, even though they may vary gradually. With the collapse of 26 Stockmann (1989: 451); see also Mansilla (1993: 45, 145). The same combination of high violence intensity on the one hand and of economic dynamic on the other hand was already typical ofthe Violencia-time. Compare Sanchez (1985: 209-258, 247-250).

Societies in Civil War

79

the state and ofthe economy the traditional upper class suffers a noticeable loss of chances of development. If a person belonging to the upper class does not decide to leave the country, then he has to be content with a considerable reduction of income and at least temporarily, with a loss of political power. The same applies to the middle classes, which additionally often lack the means för a quick and smooth change of domicile abroad. Particularly affected are the "new", academically trained middle classes, who are normally employed in big private or public organisations and who depend on regular salaries. They are often forced to look for new jobs which are far below their academic qualifications. The most affected by the troubles of the war are the lower class groups, even when civil wars officially claim to introduce their liberation and to improve their material standards. They generally lack material reserves and do not possess the mental flexibility to quickly adapt to the new situation and to adjust to it appropriately. The price they have to pay for the conflict, be it in terms of expulsions or in terms of wounded and dead people, is far higher than the losses other social sectors have to suffer. Nevertheless, in all societies affected by civil wars there is one social group which knows how to take advantages of the troubles. 27 Some of its members stern directIy from the leadership of newly formed militia associations and mercenary groups, some of them emerge from the shadows of violent conflicts and are composed of speculators, merchants with confiscated or cheaply purchased land and civil servant cadres for the newly created administrative units. Most of these newcomers who stern from lower middle classes, do not fear to take risks, are unscrupulous and have in common a talent for improvisation and organisation. These features, and the shameless way of showing off their newly acquired wealth, is probably one of the main reasons why they are disliked by the majority ofthe population and negatively described in the literature, beneficiaries of the war, members of the Mafia, typical representatives of a prospering shadow economy, who were not ashamed of taking advantage of the general misery for their own benefit. Seen from an objective perspective, we can simply state that civil wars do not differ basically from other forms of accelerated social change: They increase the mobility of social ascent or descent. This is even more true for societies with rigid quasi feudal social structures, as was the case in Spain before 1936, where the civil war suddenly opened, for entire strands ofthe population, new opportunities to move socially.28 In the context of these processes of social mobility the question arises as to the transformation of power structure in the course of civil wars. With respect 27 For Mexico, Tobler (1984: 206, 449), for Lebanon Hanf(1990: 456); for Yugoslavia Oschlies (1992: 39); Gutman (1994: 161); Calic (1993: 70); for Colombia, Jaramillo (1988: 63); Hobsbawm (1985: 13-23). 28 For Spain, L6pez-Casero (1982: 342). Particularly in the Colombian case the emancipating and mobilising effect of collective violence has been emphasised. See for instance Hobsbawm (1985: 18), and Sanchez (1985: 248).

80

Peter Waldmann

to the conflicting parties we need to distinguish extemal and interna I aspects. As for the extemal aspect concerning the relations between the different factions ofa civil war, the changes ofpower resulting from years ofwrestling with each other are smalI. As we mentioned initially, at the end of the war the conflicting militias, armies and guerrilla groups are often not far from the point where they had started. 29 The decisive military advantages are generally achieved in the first phase when energies are still fresh and chances great to surprise an unprepared enemy who can be comered. Later on a "power stalemate" establishes itself, a term which is frequently used in civil war literature. 3o Apparently all ci vii war parties are from a certain moment no longer interested in extending their territorial assets by undergoing considerable risks, but seek to maintain them. Civil war militias are excellent fOT defence, but weak aggressors, particularly when they have to fight an enemy who is also struggling for survival, instead of a defenceless population. Paradoxically, the important changes of power in civil wars are not gen erated by the relationship between the different conflicting parties. Instead they arise from changes of relationship within these parties. This result is paradoxical, because the traditional elites of power who had provoked the conflict in order to obtain a little advantage, turn out to be generally the main losers. Some reports and analyses even give the impression that the ruling elite changes completely. They say that traditional decision takers are discharged since their skills as mediators have become redundant in the face of an armed conflict; they are replaced by a new ruling elite which recruits itself from military associations and is closely connected with the new military cadres. Such analyses might jump to conclusions too quickly. It might be true that the traditional political elite gets relegated to a second position, while instead 'homines novi' acquainted with questions of war tactics come to have a say in their place. But traditional power elites are a very tough species, which is definitely very difficult to eliminate. After some time, when people have become tired of fighting, these traditional elites can recover political territory they had lost. In the long run, a fusion ofthe new and the traditional upper class takes place. Only those political elites survive who take into' account the interest of both sections of the newly constituted upper class. We shall now finally turn to the cultural and moral developments of civil war societies. Opinions of scientific literature in this area are generally clearcut and agree in condemning the consequences these conflicts have on people's emotions and ethics. One example for a comparatively moderate judgement is the passage quoted from H. C. F. Mansilla's book on the civil wars in Colom-

29

For Colombia, Leongomez (1994: 29); for Lebanon see Hanf(l990: 414).

30 Hanf( 1990: 432); Leongomez (1994: 29) refers to a "negative power stalemate" in Colombia: All political-military forces paralysed each other, no one was able to impose a constructive solution to the problems.

Societies in Civil War

81

bia and Peru: 31 "The general insecurity, the diminished agricultural production, the fall of prices for houses, grounds and land ownership in the fighting zones as weil as the apparent devaluation of human Iives, open the way for a collective demoralisation, since now the future seems bleak and live chances precarious. Benefits from savings, the necessity of a predictable economic and social behaviour and even the value of family arid friendship links are thrown into question. This painful relativising of central norms and guiding images, which are not being replaced by new value concepts, drives the militarised population into a deep socio-cultural and ethnic crisis". Most authors make similar assessments. Can one therefore conclude that periods of civil war are times of moral decay? Such generalisations are dangerous, as one can see from studies on Northem Ireland which prove that one can hardly speak of a lowering of moral standards and of a relaxation of social control within the two confessional groups after 1969 (Waldmann 1989a: 343). When posing the question on the development of morals and ethics it is important to distinguish between the social norms of behaviour on the one side, and principal ideas of good and bad on the other side. A certain loosening of social control in civil wars is also evident from growing rates of criminality which go along with the conflicts. 32 The much criticised increasing egotism, the ever growing rudeness of social forms of behaviour, the proliferation of fear, of distrust and of a 'survival of the fittest' mentality, they are not only a sign of moral vulnerability of people in general terms, but are forcefully proposed as strategies of adjustment to changed conditions, as is the growing insensibility against human pain. The other question is whether the people who are affected by such external pressures also change their deep convictions. This reaction may occasionally occur, but cannot be taken as a general rule. We know from other situations when norms are extremely shaken, as is the case in times of hyperintlation (Waldmann 1987: 367-392, particulariy 389) that they do not alter deeper rooted concepts of value and guiding ideas of behaviour, but on the contrary, that they tend to cement existing values. It is just as difficult to erase in times of chronic violence the memory of civilised, peaceful forms of social behaviour as it is impossible to eliminate from the collective memory of societies which have ne ver experienced a permanent state monopoly of violence and the option to resort to armed self defence. At least, it is not possible as long as generations exist who have not grown up in times ofviolence. 33

3\

32

Mansilla (1993: 152); see also chapter XI in Khalaf(l983). For the contlict in Bosnia Herzegovina, see for example Gutman (1994: 169-173).

33 Opinion poils in Lebanon have shown that this is not pure speculation. On one hand they make plain how strong feelings of apathy, social atomisation, distrust and forms of conduct on the verge to iIIegality have shaped lifestyles and attitudes towards Iife in Lebanon, but likewise, they show that traditionally prioritised values such as high

6 Sociologus. Beiheft 1

82

Peter Waldmann

5. The Lacking Function, Meaning and Sense of Civil Wars Do civil wars have a function or not? This question has also been of concem to the authors in the monographs mentioned here. Hanf intimates at the end of his work that the most important consequence of the war could be the rise of a totally Lebanese nation. Burkhardt states, by way of summary in the introduction to his study on the Thirty Years War, that it was not a war between states but a conflict that led to the construction of states. Other authors maintain that the civil wars they have studied had no sense at all (Hanf 1990: 754; Burkhardt 1992: 26). The two main topics we encounter when dealing with the consequences or functions of civil war are the constructions of states or nations. However, they leave one troubling question open: What do these ex post facto interpretations tell us about the particular nature of civil wars? Why was the creation of astate or a nation in some cases conditioned by a civil war and why was it not in others? And particularly: How do we explain civil wars which do not fit the functional scheme "creation of state" or "creation of nation", as is the case in the Northem lrish civil wars or in the never-ending bloodshed in Colombia? Coming back to our initial hypotheses, we propose not to look at civil wars from the point of view of causes or consequences. Instead we seek to understand them as systems of their own kind, whose dynamic is mainly determined by the logic of processes of expanding and perpetuating violence. It is typical ofthese systems that violence flows over from the more narrow political sphere of state to other social areas seeking to subject them to its own mechanisms of coercion, obedience and execution. As a result of it, various forms of social amalgam arise which are mostly unstable. Nevertheless, we have also found cases where violence has permanently established itself outside the political sphere. These cases concem exactly those two countries, which most clearly do not fit the scheme of functional attributions to civil wars: Colombia and Northem Ireland. In Colombia violence has become a purchasable service. It is subject to money and has acquired a calculable position in the market. Profaned, standardised and commercialised as it is, it might be despised as a means of imposing oneself in society. But it is impossible to foresee at the moment how it could be removed from the market in the future and how it could be put under exclusive control ofthe state. Northem Ireland represents in many ways a contrast to Colombia: instead of an immeasurable extension of violence it displays a limited use of this means; far from being trivialised, violence in Northem Ireland has on the contrary efficiency, the great respect for experience, readiness to take risks and chances, as weil as the close relation to family and religion have not lost their former importance. Khalaf (1983: 238); Hanf(l990: 581).

Societies in Civil War

83

become highly ritualised and sacred. On the other hand, both countries have in common that the state has no control over violence. The archaic rules of revenge which detennine the relations between both confessional groups in Ulster, as much as the excessive violence in Colombia show utter disdain for all principles of state order.

Markets ofViolence l By Georg Elwert

Introduction

Civil wars and conflicts dominated by warlords seem to be irrational events. The very arousal ofthis impression is in itself systematic. The protagonists and their representatives cite venerable traditions of hate, revenge and religiousmoral obligations as justification for their activities. However, these justifications, which sometimes seem well-founded in western eyes, are no more than a smoke-screen concealing the actual events which are predominantly based on economic motivations. The long-term pattern of these markets of violence are based on rational i. e. comprehensible economic behaviour. Emotions, such as hate and fear, are instrumentalised in this context, but they do not determine the structure of the process. The leading actors - and not necessarily the fighters who in addition to pecuniary gain are also motivated by other purposes - are primarily characterised by their economic motivation. They are active within both the internal markets for blackmail and receipt of stolen goods and external markets for gold, weapons or drugs. To state it in two sentences: there is a group of conflicts which we will call markets of violence. These markets of violence exhibit a self-stabilised structure and owe their reproduction to a profit-oriented economic system which combines violence and trade as a means of access to commodities. The starting conditions of such violent conflicts also imply other factors. The dis integration of the monopoly of violence can be triggered by aseries of varied motives and situations. The longer the violence lasts, however, the stronger the compulsion to rely on economic imperatives. In many cases, the continuation of violence can only be explained in terms of the existence of markets ofviolence.

I My thanks for criticism, the permission to quote from unpublished reports or advice go to Borut Brumen, Susan Cox (translation), Felix Elwert, Krist6f Gosztonyi, Ulrich Hiemenz, David Keen, Sabine Knödlstorfer, Jan Koehler, Dirk Kohnert, Carolin Leut10tT, Jean-Luc Lory, Dieter Neubert, Günther Schlee, Klaus Schulte, Trutz von Trotha, Peter Waldmann, Albert Wirz and Thomas Zitelmann.

86

Georg Elwert

Ouring the 1980s and '90s, markets of violence could or can be found in Somalia, Ethiopia, Sudan, Sierra Leone, Liberia, Northem Mali, the Central African Republic, Chad, Mozambique, zarre, Angola, Lebanon, the Caucasus, Afghanistan, Tajikistan and Northem Burma. In Rwanda, Columbia and Bosnia only some periods bear these characteristics. Acquisitive behaviour similar in nature to the markets of violence also plays a role in some other violent situations (cf. Rwanda as analysed by Neubert in this volume). It is important, however, to note that in all these countries or regions areas exist or existed where the use ofviolence did not regularly occur. 1. Markets of Violence as the Reproductive System of Warlords Markets of violence are understood as economic fields dominated by civil wars, warlords or robbery, in wh ich a self-perpetuating system emerges which links non-violent commodity markets with the violent acquisition of goods. It is the profit implied in the entwined violent and non-violent forms of appropriation and exchange wh ich is the guiding principle of action. A self-perpetuating economic system emerges beneath the surface of moral, world-view and power contlicts - even in the absence of such attempts at legitimisation. It is possible that, irrespective ofthe conscious motives ofthe actors (such as freedom, honor or revenge), only those actors who pursue economically profitable strategies irrespective oftheir intentions - will survive. Every social system implies violence. In situations of social order - "normal" times - there typically exists a monopoly of violence. (Even without a monopoly of violence there are possibilities for violence control. Segmentary societies can produce a social control which channels violence into the legitimate and manageable forms of feud, expulsion and ordeal. Even unregulated violence may lead to the establishment of routines, it will not, however, result in the formulation ofrules.) "Violence open areas", by contrast, lack a predictable and contained channelIing of violence. Violence spills into the open and becomes pervasive. The existence of violence open areas is a precondition for the emergence of markets of violence. Thereby it is irrelevant whether the inception of the market of violence was driven by economic or other motives. In fact, it is mainly political and not economic motives which predominate in the initial face ofthe type ofviolent situation referred to here. In the majority ofthe cases analysed here the market of violence started with the breakdown of a monopoly ofviolence, the "disintegration ofa state". In markets of violence contracts between the warring parties can be breached. Nevertheless, this system can be described as a market in the descriptive sense of economic anthropology. In other words, the exchange under this system is based on the exchange value ofthe goods and unlike in reciprocal systems, long term personal trust between economic partners is not required. The system described here in which the breach of contract, blackmail and theft

Markets ofViolence

87

are the mle could not be described as a market in the normative sense of economic science. Areas in which there are no mies which build channels for the use of violence are "violence-open", open to unbridled violence. Even if unregulated violence may lead to the establishment of routines, it will not, however, result in the formulation of mies, since the breaching of contracts is common. Violence can even destroy the clientelist relations within a warring party between warlords, their chiefs and mercenaries. Tbe possibility of murder for the sake of murder cannot be excluded in areas open to violence, even if such murder would appear uneconomical to most actors. Thus, to summarise, a market of violence is a field of activity which is mainly characterised by economic aims, in which both robbery and barter and the related activities of collection of ransoms, protection money, road tolls etc. feature. Each actor has a number ofbasic options ranging from theft to trade. The generals, princes, militia chiefs and party leaders who lead the troops in such conflicts are known in the literature as warlords 2• Warlords are understood as entrepreneurs who use deliberate violence as an efficient tool for achieving economic aims. These "entrepreneurs" differ from normal entrepreneurs in that they also use violence - although not exclusively - as an instrument for the generation of revenue. This image of persons primarily acting on an entrepreneurial level is confirmed by the - few - academic reports 3 emerging from the centre of such conflicts. We noticed that the same circle of actors (as currently in Ethiopia) may continue to operate as peaceful entrepreneurs after the conflict. This view does not coincide with the stereotypes largely propagated by the media, wh ich cite emotions and tradition by way of explanation. Unlike pub brawls, modem wars necessitate strategic planning and logistics. The killing cannot be sustained without cool and calculated planning for supplies of weapons, munitions, food and fuel. The planning of strategic action and military logistics require a cool head and not the sustained evocation of emotion. Gosztonyi observed in Bosnia that in the initial phase of the conflict risk-seeking and even mentallY disturbed persons played an important role; later, however, their places were taken over by rather cool actors - managers ofviolence. Markets of violence can emerge as the first structure of "modemity". An example from precolonial Africa is provided by the reports of Henry Morton Stanley on 19th Century Congo. He followed in his travels the paths which had 2 This term is used here without any acknowledgement of its connotations with respect to the civil or criminallaw status ofthese persons. 3 Some ofthese reports cannot be published or quoted with reference to their authors in order to protect on-going research. I owe most ofthem to the research project "Konflikt-Treiber, Konflikt-Schlichter" financed by the Volkswagen Foundation and to the "Projekttutorium Anthropologie der Gewalt" financed by the Freie Universität Berlin (see also Koehler and Heyer, eds., 1998).

Georg Elwert

88

been established be the Arab slaveraider and slavetrader Tippu Tip. Robbing and trading slaves, ivory, bees wax, brass and prestige goods were central for the reproduction of the traders/raiders. Sedentary peasants had to enter the military structure as mercenaries, serfs or as self-defending communities, if they wanted to avoid slavery4. The appearance of some of the fighters in these conflicts, barefoot and armed with cutlasses, is sometimes perceived as "medieval". That we are facing something "modern", is not immediately evident. The "modernity" ofthese wars is given by the internal structures of market and command and by their necessary linkages to the industrial world. The new perspective which we would like to present here is not so much concerned with the revelation of economie motives behind conflicts5 or the demonstration of the economic success of individuals in situations previously characterised by destruction. It is concerned with the demonstration of the systemic character wh ich lies behind conflicts which appear to be chaotic. Thus, we will now turn our attention to the regular forms of economie behaviour wh ich can be observed in these areas. After this we will focus on the external economic forms of stabilisation and, finally, on how such markets of violence are formed. The - many - causes of the emergence of the markets of violence do not explain their durability; the latter lies in their systemic nature. 2. Partial Markets and Market Strategies as Elements of an Economic System The decision whether to steal or obtain certain goods through commercial transactions is basically always open in the markets of violence. The warlords need to win time so that they can make strategie decisions and keep options open. Thus, the actors are confronted with a strategie triangle of violence, trade and time. The warlord makes caIculated moves between these poles; he carefully considers and optimises the costlbenefit relationship. Given that it is often not recognisable as such, this caIculating rationality is explained below. We are not suggesting that emotions have no structurallY formative role to play in the social process. In the markets described here, however, optimisation processes are based excIusively on economic imperatives. This does not, however, eliminate the possibility of emotions making a significant impact on the 4 See Stanley 1872:86, 90, Tippu Tip in Bontinck (ed.) 1974 and especially Frank McLynn 1991 and 1992, to whose analysis I am indepted. It is tempting to analyse early modern Europe and the period preceding the establishment of stable capitalism structures there under the same perspective; cf. Tilly 1986 on states and the economy of war and Puhle 1996 on the pirates and their merchant partners operating against the Hanse in the North and Baltic Seas.

5

On the importance of economic motives see Jean and Rufin (eds.) 1996.

Markets of Violence

89

cost aspect ofthe individual instruments and thus influencing the cost structure of the market. The production of violence itself is also based on economic imperatives. The aim is to reduce the cost of the violence 6 • Weapons, munitions and fuel are indispensable resources and it is not advisable for the warlords to save money by skimping on quality. lt makes more sense to economise on the fighters' fees. Thus, "marauding", i. e. systematic theft by soldiers is an obvious form of the reproduction ofworking capacity. We even find that a market actually develops for the marauding itself. In other words, it is possible to participate in organised robberies for a fee. In Yugoslavia, volunteer organisations made use ofthe support of"weekend soldiers" who on Friday afternoons for a suitable fee and the loan of weapons were bussed to the front to return with their rich pickings on Sunday evening (E/wert 1997:90).7 Among them were also sadists looking for sexual pleasure. This did not change the character of the whole enterprise as long as they paid their fees. In cases where population density is too low to make marauding a profitable activity, "donations" of food for the troops are important. Food aid has been used to supply troops in Liberia (Montelos 1996), in Lebanon (Picard 1996), Afghanistan (Doronsorro 1996) and Sudan ( Keen 1991, 1994, Prunier 1996). We received unpublished reports about refugee camps which have been converted into barracks for troops awaiting deployment. The camp Kakuma (Kenya) is even said to host three liberation armies simultaneously. The lowest estimation of aid appropriated by warlords for the Sudan gives more than 20 %" (Prunier 1996: 376). The refugee camps in Congo established in the 1960s under UN supervision (see e. g. Mummendey 1997) bore considerable responsibility for the recruitment of the militia wh ich continued to ex ist in the country (henceforth known as Zaire) until the 1990s. The parties succeeded in creating their armies by terrorising their "tribai brothers" by means of their party police or militia which acted as an instance of coercion. Such developments are favoured by the collusion of warlords and organisations active in the distribution of the aid (which do not wish to jeopardise the flow of aid as a result of critical reports) and the absence of monitoring of the dynamics of a conflict (beyond the mere counting ofvictims and the needy).8

6 Mohammed Aidid of Somalia had to pay US $ 40,000 per week in 1994/95 to sustain his militia - N. N. 1995: 275.

7

A German official informed me about a parallel case observed in Rwanda 1997.

Some social scientists have, however, denounced this collusion as shown by Allen in this book (cf. also Jean 1996 and Duffield 1997). 8

90

Georg Elwert

Another way in whieh the warlords can reduce their overheads is by exploiting des ire for social prestige 9 . Young men can easily be won over as highly motivated "volunteers" if involvement in dangerous acts (e. g. multiple robberies) means they wilI be finally considered "real men", or if the exercise of violence is one of the conditions of an initiation ritual. This is typical for East Afriean generation- and age-cIass systems in astate of warring (see e. g. Strecker in this volume); but it applies also to some milieus of Georgia and other countries ofthe Causasus (as shown by Koehler 1998). In the short tenn - and this is particularly important in the initial phase of the development of a market of violence - the promise of justice and freedom, or to be more precise the hope of resolving latent conflicts or overtuming despotie rule can be sufficient to mobilise volunteers, i. e. unpaid actors (cf. Wirz 1984 on African civil wars). The generation of fear is a partieularly cost-effective fonn of mobilising troops. Thus, propaganda is an important instrument of production in such conflicts. From an economie perspective, "senseless violence" can find a meaning in this way (see Geffray 1992 and Weissmann 1996 on Mozambique). The fear of retaliation on the part of the victims leaves no option open but to join an anny or support it for one's own protection. Fear of revenge stabilises the system. Fear of becoming a victim of violence oneself can also motivate one to participation in preventive attacks. These are generally not based on strategic planning, can escalate rapidly and end in the slaughter of neighbours who may previously have had cIose relationships. One main difference between markets of violence today and those in the last century lies in the way in which electronie propaganda can reach large populations more quickly and cheaply and can thus convert the fear of people to mass participation in military actions (which are economically profitable for a small minority) on a scale previously unknown. In Rwanda the radio station "mille collines" created and amplified fear and then transfonned these energies into organized genocide (MeilIassoux and Behrend 1994). In Bosnia it was the TV which created among Catholics the fear of pecoming vietims of - probably non-existent- Islamic fundamentalist troops. This fear prompted villagers to strike "in pre-emption" their Muslim neighbours with whom they hitherto had friendly relations (Bringa 1993). The automation of revenge and fear of revenge creates single appartenance where multiple memberships (e. g. based on language or religion) would previously have left people neutral towards the militant actors.

9 The general importance ofvalue systems and prestige is maintained also in violence markets although economic motives act in the foreground. Two layers of motivation are needed if violence is to build a stable pattern (see Elwert 1997). On the ro1e of value systems in violent contlicts see Orywal 1996.

Markets ofViolence

91

This new unanimity means that massacre can take place on a large scale; there is no need to omit individual houses during firing. The militarily meaningful friend/foe distinction can now be ideologically reformed and take the form of "ethnic c\eansing". The fear of being dec\ared an enemy and thus becoming a potential victim, or fear of revenge motivates civilians to assume the role of unpaid "c\eansing agents". The trade side ofmarkets ofviolence is not less dominated by considerations of efficiency.1O The most important resource for trade is money or a similar object of exchange. Exchange goods which are easy to transport are particularly important. For this reason, trade in valuable objects is of disproportionate significance in the market of violence. Diamonds, gold and precious stones enable the transfer of large values in a single movement. 11 • Theft and enslavement (e. g. in Mozambique and Sudan) are particulariy obvious in this context (Geffray 1992, Weissmann 1996: 323, Prunier 1996: 380). Theft is rarely a matter of direct acquisition. Anyone who hauls off refrigerators, video recorders and cars (or to use a more blatant example, removes gold teeth as witnessed in Bosnia) usually does not do for personal consumption but is reacting to the demands expressed by the receivers of such stolen goods. The drug trade is also lucrative - trading in Somalia, for example, in the drug qat and in Afghanistan in hashish. The same applies to weapons. Wherever markets open to violence have developed, trade in these valuable commodities may even be attracted to the area, as can currently be observed in Tajikistan, Afghanistan (Doronsorro 1996: 175) and Sri Lanka (Labrousse 1996b:487--488). Professional protection of transport of goods by war lords and possibly even entrepöt trade (temporary storage in a "safe" location) can be particularly attractive services for all kinds of illegal trading such as drugs and weapons smuggling. Smuggler's routes deviate into areas controlled by warlords. The combination of drug trade and weapons trade can help to extend the trading networks (Labrousse 1996b: 473). In some markets of violence there is a commodity sold which the buyers do not recognise as such: the resource in question is tbe sacrifice of human lives for the sake of ideology. Battles are fought in pursuit of an ideal and accompanied by detailed reporting. Lives on both sides of the divide are sacrificed for one product which is ofvalue to emigrants or foreigners interested in the fate of the world. "The Free West", "the Socialist World Revolution", "the Honour of Our Nation", "Saving our Faith" are the issues wh ich appear to be at stake. 10 The profit interest implies that violence can be put to a secondary rank if there is a chance for more profitable trade. The Bosnian warlord Stela excelled in this combination of fight and trade. Sometimes the shooting was interrupted at midnight to allow the exchange of money and commodities (unpublished report 1997). 11 For precious stones from Cambodia see Lechervy 1996: 213-218, for emeralds in Colombia see Waldmann in this volume and Krauthausen 1996.

92

Georg Elwert

Special attention to this commodity - commercialised sacrifice - through the establishment of a special troop which is trained in rhetoric and international communication (see Zitelmann 1997 on Ethiopia).12 It was in this way that Jonas Savimbi succeeded in selling his battles in Angola as both sacrifice for the Maoist world revolution for socialism and then later for the defence ofthe free West against socialism before turning his attention again to his main activity in Angola's and Zaire's diamond business.

An intermediary activity between trade and theft is the collection of protecti on money, also referred to as duties. The taking of hostages (see Labrousse 1996a: 396-397, 403 on Colombia and Peru)has assumed major significance. Today the guarding of goods is a particularly relevant phenomenon. Diamond and gold smuggling in contemporary Congo/Zaire, trade of the drug qat in Somalia, emerald smuggling in Colombia and, last but not least, convoys carrying food aid in Sudan (cf. Allen 1996), Somalia (cf. Duffield 1997), Liberia and Bosnia have made this "protection" the warlords' most important source of income during certain periods. Gold trading and especially protection or extortion of protection money for the gold trade was the speciality of the warlord Kabila who has been active for over 30 years and has now reached new farne in Zaire/Congo. The extortion of protection money is a flourishing trade in "disaster areas" (see Allen in this volume). An aid organisation paid US § 1,000 in 1994 to a Liberian warlord for each lorry with medical equipment which he allowed to pass through his territory (Germund 1994). There is one resource needed for successful warlordism, which is not obvious as such: time. It is possible to observe an accumulation not only in the means of violence and money but also in time options. Time options are created first and foremost by ensuring means of subsistence. Such guarantees can take the form of contracts for food supplies (also from aid organisations) or delivery of tributes from enslaved farming communities (see Geffray 1992 in Mozambique and Labrousse I 996a: 40 I in Peru). Time options are also created . through pacts with competitors or opponents. And finally, time is won through negotiations (cf. von Trotha's concept of "ZeHraub" 1994: 425). It is possible to negotiate about non-aggression, delivery of supplies or pacts without necessarily being primarily (or even "secondarily") interested in the non-aggression, goods or pact in question. What is important is winning time. Mediators who fail to understand this part of the strategie game become nervous when faced with stalling tactics. Some internationally noted attempts at mediation ride on concealed propagandist juggling with violence or exchange of goods and also time. Is it better to rob someone, extort a contribution, exchange goods or block everything and play for time? One plays for time, allows the non-aggression to be bought and then calls it taxation, duty or financing the peace process. The 12 Angoustures and Pascal report (1996: 509) that Kurdish troops received in 1992 half ofthe budget from sympathizers' contributions.

Markets ofViolence

93

wasting of time can weaken or simply unnerve the other side making military success easier to achieve. This insistence upon the economic level, therefore, does not imply that emotions could not play an important role as creators of structure. What is stated here, is only that the processes of optimisation follow economic imperatives. This does not exclude, as shown above emotions from appearing in the market as a structuring force on the cost side. 3. Stabilising Markets of Violen ce

A deregulated market economy develops in fields of violence, areas open to violence. The opportunities hoped for determine who cooperates and who attacks whom. It is not traditional values ofreligious communities, ethnic groups or clans who oppose each other in these "ci vii wars" but economic interests. Values are inward-Iooking, they aim at cohesion - interests are outwardlooking, they imply competition. Economic systems 13 based on the use and generation of violence become established, systems which repeatedly erect smokescreens of political legitimisatiqn and are thus happy to assume and play back culturalistic interpretations. 14 The economic imperatives imply that optimal use should be made of the resources available. These resources are not necessarily "economic in nature". Thus, an element of social structure which is instrumentalised can be useful precisely because of its non-economic character. Allegiance is such aresource. Although the Somali war split all the big clans and clan aIIiances (which implies that it cannot be called a "war of clans"), loyalty within the smallest subIineages was maintained. 15 Reliable allegiance reduces the cost of contro\. 16 Secret societies and a code of conduct formed in prisons and punishment camps can provide for allegiance. Secret societies built in ilIegality maintain cohesion within the troops (see e. g. Koehler 1998 on Georgia). The Croatian warlord Mladen Neletilic, called Tuta, referred (1997) explicitly to the common criminal past of his battalion caIIing it "prison gang" (kaznejenicka snaga). But even a non-economic loyalty breaks down ifthe troops are no longer fed. Economic imperatives cannot be by-passed. 13 The concept of system implies reproduction. A general perspective which links reproduction to violence is offered by Balandier 1986 and by Bazin and Terray 1982.

14 Prominent in this culturalist discourse are phrases about differences in basic values and about historically founded foeship. Such statements act as a deterrent for foreign intervention. 15

On the Somali war see Schlee 1995, Lewis 1997, Heyer 1998.

16

On ethnic loyalty as a resource see Gallagher 1997, Zitelmann 1997 and Lewis

1997.

94

Georg Elwert

One consequence of violent situations is that the opportunities for income are moved out ofthe traditional economic sectors. Trade, industrial production, peaceful commercial activity and agriculture enter crisis and completely collapse when they become dependent on continual external inputs (see Keen 1991 on Sudan). Wages and income in these sectors slump. Invested capital is devalued. In many cases, the only way that those dependent on wages and the self-employed can survive is to become soldiers and/or marauders. Entrepreneurs are weil advised to invest their liquid capital in the formation of a troop and purchase of weapons. Thus, it is not surprising that Somalia's warlords were previously - in times of peace - mainly involved in wholesale trading or as political entrepreneurs (also as partners in development cooperation). The market of violence becomes established as alternative income sectors come under pressure and for the most part lose their opportunities for reproduction with workforce and capital being absorbed by the - relatively - higher wages and profit opportunities in the violent market sector (see e. g. on impoverished refugees in Lebanon: Picard 1996: 86 and on deinvestment in Bosnia: Bougarell 1996: 239). Autostabilisation also involves efforts in the symbolical-ideological sphere. Use by this trade as conspicuous brutality (see e. g. Richards 1996 on Liberia and Sierra Leone). The ideological self-presentation which forces violence to the forefront is intended to stabilise the position in the market of violence. It facilitates inter aUa the sale of "protection". The selection of victims is based on highly complex calculations. The warlords need trading partners, supporters and neutral powers. It is helpful to allow the violence to follow cIear, symbolically delineated, lines to enable these actors to feel safe. Religion, urban or rural costume, regional accents etc. are suitable as a basis for distinction here. They arouse the impression of ethnic or religious confrontation. However, no trading partner or associate can be sure that he will not fall victim to the covetousness of yesterday's ally. Given the high risk of betrayal of leaders by troops, the former try to revive proven patterns of authority. The emphasising of belonging and distinctions combined with the invoking oftraditional authority often feature as innovations within the markets of violence which can create new ethnonational units and thus contribute to the stabilisation of the violence along cIearly defined boundaries. This may engender either oftwo outcomes: a) the progression to cIassical cease-fire between commanders who are representative and capable of entering an agreement or b) it can also lead to the outbreak of cIassical war. Thus, the end of a market of violence is not necessarily the beginning of peace. The widespread hope that the majority of the population wh ich is the victim of the violence and the economic limitations can rebel against the warlords can only materialise if a monopoly of violence is established from outside wh ich enables an independent jurisdiction, the holding of free elections and what is often neglected - the circulation of free and dependable information. It

Markets ofViolence

95

is not in the interest ofthe warlords to allow this as long as their trade remains profitable. If an end to the violence is actually achieved, a shiftin values often occurs within the population which now places a taboo on violence. Thus, whoever establishes a monopoly of violence - whether by democratic or nondemocratic means - can be sure of a high degree of legitimacy (cf. Labrousse 1996: 390 on Colombia and Waldmann in this volume). 4. The Origins of Markets of Violence How do markets of violence come into being? A monopoly of violence does not disintegrate suddenly. We just perceive it only when it is too late. It crumbles with increasing speed as a result of violations such as robbery and the violent despotism of (Iocal) rulers. In most cases it disintegrates from within. In other words, state intervention contravenes the notion of legitimate use of violence which exists among the people and thus legitimates counter-violence or an imitation of this despotism at the lowest level. This despotism is a characteristic ofthe command state, "the brother ofthe command economy" in which the authority present has priority over laws, contracts and other written regulations (Elwert 1990). That arbitrariness of the state agents - namely police and military - played a central role in the origin of African civil wars has been shown by Albert Wirz (1982). Arbitrating institutions and the security of the legal system can prove to be too weak if confronted with a rapid economic development. The idea of "self help" which is intended as a means of achieving justice is then an obvious option. This "self help" often takes violent forms, particularly in the vital area of land rights: the expulsion of sections of the population in Nigeria and Central Asia led to the emergence oftemporary or permanent fields ofviolence. How self-defence can slide into warlordism is shown by the example of the Gadabursi in Somalia (Marcel Djama 1992). This was not evident from the beginning. Groups of herders had long since bought fire weapons in order to secure their rights to water places independently of clan courts or state justice. This development was approved by the state since this "privatisation of law or contract enforcement" reduced its expenses. The government even pushed this development which alleviated its tasks. This development, promoted by young men, was for them an emancipation from the old clan system and its justice of old men. For a time the use of weapons remained at a low level. Sut then the frontier to Ethiopia was closed. Without the state taking any effective action to defend the herders' interests, access to water and to cheap food (from "disaster relief aid") became difficult. Not only the nomads suffered from this situation but also the traders who had exported their sheep and cattle to Yemen and Saudi Arabia. Now these very traders procured weapons in bulk for the nomads in order to enable them to seeure the reproduction of their herds by use of violence. Thus the militia of the Gadabursi was created. This militia was soon to

96

Georg Elwert

discover that weapons facilitate the acquisition of food, that taking hostages (also traders), asking street tolls from food aid transports and selling protection to drug smugglers are no less profitable than herding. Liberation movements which merely aim to transfer the monopoly of violence to other hands - their own - actually create a second pole of violence (see Waldmann in this volume). When violence comes from both sides, the monopoly of violence collapses quickly and areas open to violence emerge. In the course of further development, the economic motives for the exercise of violence can become dominant in these fields of violence. This does not deny that politically motivated wars of liberation exist. However, the transition to an economy of violence is fluid. On the one hand, the guerrillas gain legitimacy from the violent state despotism, a despotism which serves the individual profit of the officials. On the other hand the long-term financing of weapons supply necessitates that the war sustains itself from the war (Meillassoux 1990). Thus, ideologically motivated fighters may become economically motivated warlords who continue to promulgate the old ideology for the sake of greater legitimacy. In local societies which are not subject to central state control an economic or technological imbalance - e. g. gold-digging or a demand for sc outing services to help robbers, or the supply of cheap and efficient weapons - can lead to failure in the internal control of violence. In the case of the Ugandan civil war, the Ik, who were previously mainly active as hunters and gatherers, saw the opportunity to enrol as scouts. This resulted in the partial disintegration of the internal moral control; the old, women and children were left behind and suffered from hunger (Turnbull 1972). In the highland of New Guinea, as Lory reported, feuds became war when automatic firearms became available and replaced spears, bows, axes and cudgels. In Somalia, some militias, as shown above, emerged from nomadic self-help units which needed to obtain access to wells and food, received free or cheap fire-arms and soon discovered that the taking of hostages, protection and blackmail were also profitable activities. In Afghanistan, cheap arms contributed to the transformation from feuds to warring (Centlivres 1997). The temporal coincidence of democracy movements and the increase in internal violence has led some observers to the conclusion that one is the cause of the other. In actual fact, however, both are a consequence ofthe end ofthe Cold War. Large powers and former colonial powers now enjoy less legitimisation for intervention when violence or apower shift occurs in their (former) areas of interest. They now allow local actors to prevail in situations in which they would previously have taken immediate action. This can give rise to a situation whereby free elections take place without the monopoly of violence being in the hands of the state. It occurs that parliaments de facta have no legislative authority, as the public servants do not feel bound by the laws voted by the parliament. We think that such situations should not be described as democratisation in the sociological sense. Violence control, especially by means of a

Markets ofViolenee

97

monopoly of violence, is fraiI; institutions of arbitration and eonflict resolution are weak; the future is insecure, given the low reliability of laws. This frailty of law results in the fact that the gradual conflict (the more/less) loses ground and the alternative conflict (the either/or) without options for compromise becomes dominant. In these situations clear solutions which eliminate alternatives (and the persons representing them) seem more plausible. The willingness to compromise would appear dangerous. The willingness to engage in self-help through the use of violence increases. This can occur both in conjunction with elections and where elections are suppressed. The dis integration of the monopoly of violence is a necessary but not sufficient condition for the emergence of markets of violence. The future warlords must be at least familiar with the rules of the market economy. This explains, for example, why in the Caucasus and Central Asia the warlords emerge mainly from the circles of former black market activists. Know-how in low-Ievel violence is helpful. This explains the attraction that mujahidin (fighters) trained in the Lebanon (financed among others by Libya or in Afghanistan (financed among others by the USA) feit for other markets of violence. 17 Furthermore weapons, munitions and fuel must be available at acceptable prices. The decline in the price of weapons which can be observed in Africa since the early 90s may have contributed to the spread of markets of violence there. And finaIly, a certain level of exposable resources must also be available. Extreme poverty of the victims prevents the emergence of markets of violence in situations ofviolence (e. g. in the mountain regions of Afghanistan and Tadjikistan). A concentration of wealth and excess to legally exploitable resources promote the emergence of these markets (e. g. precious metals and stones wh ich have sustained the markets of violence in Zaire since 1960s). Thus, even if the other parameters remain unfavourable, positive economic development can favour the development of markets of violence. When areas open to violence and market economy coincide there may be positive interaction between them, i. e. the market interests expand the fields of violence and market interests are increasingly realised in areas open to violence. Thus, the self-stabilising system ofthe market ofviolence is generated. 5. The Demise of Markets of Violence Markets of violence "only" ever exist for a few decades. There are several factors wh ich lead to their eventual demise.

17 It might as weil be that strategie know-how from periods ofwarring transmitted by oral tradition was helpful for the aetual markets of violenee in Congo/ZaTre, Sudan (Prunier 1996: 351) and Tsehad (cf. Oppenheim 1902 on the warlord Rabea in the 19th eentury).

7 Sociologus. Beiheft 1

Georg Elwert

98

Everything is allowed in the radical free market economy, including breach of promise and the betrayal of one's kin l8 • Thus, anything that can stabilise relationships of trust is particularly in demand. Associations based on religion, language, origins or (secret) associations are created. Within the radical free market wh ich allows alI, violence and breach of word, efforts are made to mark out small areas of internal morality, areas in which violence is eliminated and trust established. The invention of an ethnic group or religious community is one way of doing this. In Somalia and Afghanistan, for example, the concept of religious community ( 'umma ') came to the foreground. Such moral groupings can lead to underground transformations which can elude observers focussed on violence for a considerable time. At the beginning ofthis century, underground secret societies flourished in China which was tom and divided by warlords (cf. Sheridan 1983, Ch'i 1976). They included groups with high ethical claims and one Christian lay mission even succeeded in becoming established within some ofthe armed troops. One ofthe secret societies wh ich flourished in this underground was based on a particularly successful combination of extreme obedience, the concept of a peaceful Utopia and efficient military organisation: the Chinese Communist Party. Many observers deemed it anachronistic due to its moral rigour. And yet this was the very characteristic which proved the key to its success. The establishment of a monopoly of violence, wh ether it sterns from such an internally emerging movement, from an outside intervention or from the victory of one warring faction 19 terminates markets of violence but this is not the only possible outcome. The dis integration of a market of violence can, of course, also lead to the outbreak of a c1assical war situation as discussed above. This often has particularly unpleasant consequences for the populations involved. In the "normal" market of violence, due to the cost of the violence, there are actually fewer victims than in c1assical wars (e. g. in many phases in the development of the war in Afghanistan). The warlord sees death in terms of its financial cost. The deployment of regular forces in a "normal" war can result in a drastic increase in the number of victims. When the violence is used to serve non-economic aims (such as territorial rule of an ethnic group or the prestige of a nation), when ethnic cleansing becomes an aim along with theft and blackmail, the nature of the activity changes and the number of victims increases. 20 Why should a warlord resist the temptation to put hirnself in the service of such a

18

See Sch/ee (1995) on the fragmentation ofkin groups in the Somali war.

An interesting case for this is presented by Vo/ker Janssen 's (1991) transposition of Norbert E/ias' (1978) model ofthe monopoly process /Tom Europe to Ethiopia. 19

20 To serve a "higher cause" is not detrimental for a warlord as long as his profits are not endangered.

Markets ofViolence

99

"superior" object if the business continues to be profitable for hirn (as demonstrated by the development ofthe conflict in Bosnia)? No market activity is purely based on motives ofprofit, subsistence and consumption. Market events cannot be consolidated without secondary motivation?1 Secondary motivations - such as the des ire to create a new ethnonational community - can become primary motivations which then establish a new reference. This was also a European experience: the capitalists of the mercenary armies of the Thirty Years War became both power-accumulating statesmen and peaceful entrepreneurs after the end of the war. A similar process is witnessed in Lebanon today. Concomitant to the shift of motivations the exhaustion of resources contributed to the establishment of (relative) peace. Markets of violence do not develop and exist in a vacuum. They grow out of self-organising social systems which by virtue of their nature depend on an exchange with their environment. The forms of this exchange are, themselves, subject to evolution. Given that markets of violence largely destroy the internal institutional structure and production potential of a country, they are considerably dependent on clients, suppliers, banks and other services outside their area. It is easy to overlook the importance of the service sector, because its actors shy away from publicity. Markets of violen ce need external infrastructure. Somali warlords, for example, make extensive use of educational centres, special hospitals, banks, insurance companies, stock exchanges and commerce arbitration courts in other continents (North America and western Asia). Exchange across the borders of its own system is, therefore, one of the Achilles heels of markets of violence. Blockades, i. e. the prevention of this exchange, can destroy them from outside. It is, however, difficult to find examples of such successful blockades (cf. Hufbauer 1990). Mozambique, after South Africa ceased to provide the external infrastructure for violence, became such a case. Somalia cannot currently be used as an example of such a development. The transition from markets of violence to other forms of violence, or to more peaceful states, can, therefore, be achieved through the exhaustion of resources, the imposition of external blockades and internal shifts in orientation. The transition period is characterised by ambivalent activity. It is as though players of agame seated together at the same table interpret the moves differently. Players judged as "irrational" by "normal" standards because theylook for power and/or peace instead of profit, can prevail if they manage to obtain control over a scarce resource or synchronise their actions with the effects of an external blockade or the exhaustion of internal resources.

21 This can be seen in our society, for example, in the prestige awarded for academic work which is a condition for the academic labour market.

7'

!OO

Georg E!wert

6. Epilogue: Are There Political Options to Promote Peace from the Outside? Markets of violence will continue to emerge wherever there is a coincidence of areas open to violence, exploitable resources and corresponding markets, and where the economic environment remains neutral or favourable. Violence markets are more likely to increase than decrease in number in the near future, if 1. wealth increases, 2. weapons become cheaper and 3. monopolies ofviolence remain unconsolidated. Are there political options to promote peace from the outside? It is possible to jeopardise seriously economic interests (smuggling, investment) and the supplies ofweapons and fuel through blockades. Such blockades are admittedly difficult to implement. They involve extensive manpower and technological resources on the borders of the war zones and necessitate that the personnel who carry out the border policing be capable of using weapons in the completion of their duties (unlike in Yugoslavia). Economic interests are not only realised by smuggling but also by robbery, blackmail and hostage-taking. The proceeds of that are commonly invested in safe third party states. Therefore such blockades must also be implemented as blockades of services, particularly the movement of money. This could be seen by neutral states as interference in their "freedom of banking".

In some situations, the market ofviolence can only come to an end ifthe inner-state monopoly of violence is guaranteed by external powers for a limited period of time. This policing (in the sociological sense -a non-military task) has four prerequisites: firstly, the arms and logistics resources available to the intervening powers must be equal to those ofthe internal actors; secondly, the use of sanctioning violence must be based on legal state regulations and enjoy institutional back-up; thirdly, weapons control must be achieved; fourthly, everyday conflicts must be arbitrated if violence in the form of self-help by people "c1aiming their rights" is not to flare up again. The linking up with local concepts of justice and local institutions is unavoidable here if the aim is to lay the foundations for the endogenous development of the rule of law by native powers. All this, however, is based on the assumption of socio-cultural competence on the part ofthe intervening party. This is seI dom in evidence. The most common form of intervention in conflicts is currently the dispatch of doctors and nurses to the war zones and the distribution of food aid and clothing to the victims of violence. Given that the people are assembled in camps, the task of recruiting troops is therefore made easier rather than more difficult for the warlords (as actually demonstrated by the example of Rwandan refugees in Congo/ZaIre). The establishment of refugee camps, which act as barracks for the warlords and are supported through the distribution of food, is thus counter-productive as a form of intervention. Decentralised supply concepts would be more suitable. The fact that some aid supplies will be diverted from the intended recipients cannot be completely avoided. It would, however,

Markets ofViolence

101

be possible through monitoring to avoid aid supplies being used to benefit warlords more than their victims 22 • The necessity ofthe rule of law, of a control in the use of state violence and of dependable media also demonstrates the needs for development cooperation in the post-conflict phase. Of particular urgency is the support of the development of local institutions for conflict resolution. "Truth commissions", which negotiate the institutionalisation of common norms on both sides of a conflict can represent an important preliminary stage in this process. Forms of perpetrator-victim exchange can also promote the development of these institutions and the development of a shared awareness of the law. In this instance in particular, it is important to resolve confrontations by seeking compromises. To achieve this the notion of "honourable conflict" needs to be changed. The more/less gradual conflict needs to take the place of the either/or conflict (alternative conflict). It is an open question whether this can be achieved by linking up with endogenous concepts oflaw. In this context help must also be available for the victims. If this is seen 10cally as solely the business of the international donors it can fatally undermine the efforts to find an exchange between perpetrators and victims and the reactivation or preservation of social responsibility for widows, orphans and the disabled. Since states of war reward the values of soldiers, there is a particular danger that peace will only be seen as a temporary cease-fire. In spaces of unregulated violence, a violence specific order of prestige often becomes established which ties young men's opportunities to acquire honour and status to acts ofviolence against outsiders acts of revenge for "wounded honour". As long as this order of prestige remains, areserve army continues to grow in peacetime. This is not the only possible outcome in such situations. It is also possible to open new paths to social recognition - a difficult but not insurmountable tasks for development policy. The promotion of entrepreneurial ambition in peaceful forms can transform potentially destructive energies and provide both profit and prestige. As violence is a product of men's social organisation, so are markets of violence. Thus there is a chance to undo them. 7. Conclusion

Markets of violence are highly profitable social systems which remain stable over several decades; their purpose is both violent appropriation and exchange of goods. They generally originate in conflicts of a non-economic nature. The continuation of violence is, however, based on economic motives or uncon-

22 Social science monitoring has to include also the tlow of goods, alliance dynamics, forms of recruitment, development of autonomous means of reinforcement.

102

Georg EI wert

scious economic behaviour. These markets can collapse as a result of the monopolisation of the violence, the exhaustion of inner resources or a cessation in the availability of extemal resources. From the perspective of the protagonists, the warlords, violence can be used to maximise profits to such an extent that it is on par with other economic strategies. The fact that the balance sheet is far from positive in its effect on the overall system is irrelevant for the functioning of the system. In economic terms: the negative welfare effect does not impede the reproduction of the market. It takes decades for devastation to take its toll. This inherently rational economic behaviour can continue as long as warlords are able to exercise their power without the support ofthe majority. Expanding economic potential, falling arms prices, the widespread existence of commandstate structures and a continued reluctance on the part of foreign powers to intervene make an increase in such markets ofviolence probable.

Tbe Dynamics of War and Alliance Among tbe Yanomami By Jürg Helbling

Introduction Since the 1970s - after the decline of functionalism and under the impact of the wars of independence in the Third World - anthropology has regained its interest in war in tribai societies I. Indigenous wars, as can be observed in Amazonia, in the highlands of New Guinea, in East Africa and elsewhere, are wars of "our contemporary ancestors" (Service 1968), i. e. they are premodem wars occurring nevertheless in today's world of states and global economy. These wars, however, do not aim to achieve secession from astate or to gain control over astate. Rather tribaI wars are armed conflicts between largely autonomous local groups and/or between coalitions, to which they line up (Otterbein 1973i. This articIe explores the dynamics of war and alliance among the Yanomami in the Mavaca-Orinoco area of Venezuela between 1930 and 19803 • The Yanomami, numbering about 20.000, are scattered over 200 to 250 villages of on average 100 people with variations between 40 and 200 inhabitants (Chagnon 1983: 79; Lizot 1971: 139). They practise shifting cultivation with plantain, manioc and other crops as a mainstay; in addition, they do some hunting, gathering of wild plants and fishing (Lizot 1971; 1977; JohnsoniEarle 1987, chap. 5). The village territories have a radius of about 10 km. Every 5--6 years the groups move between 50 m and 1 km further on, to build a new shabono 1 I use the (problematic) concept of tribaI society in a purely descriptive sense as a regional population of cultivators, pastoral nomads and sedentary fishermen, who live in autonomous local groups and who entertain kinship relations, exchange women, fight each other, enter into alliances and celebrate feasts (Rappaport 1968 on regional population, Sahlins 1968). 2 Cf. Otterbein (1973); Hallpike (1973); Harris (1977); Clastres (1977); Hanser (1985); Ferguson (1984) (1990b), McCauley (1990); Ross (1993) on war in tribai societi es in general.

3 On the Yanomami cf Chagnon (1968; 1983; 1992); Lizot (1971; 1977); Ales (1984); Herzog (1990); Zerries/Schuster (1974); Biocca (1972); Ferguson (1995).

104

Jürg He1bling

(Lizot 1971: 149f.). The population density averages only 0.2 P/sqkm (Harnes 1983: 415, 425). The relations between local groups are characterised by war and alliance. Allied villages engage in trade and feasting. However, alliances are always conflictive: allies try to outsmart each other and can turn into enemies virtually overnight. When conflicts emerge, they first take the form of club fights between individuals. EventuaIly, however, they escalate to a collective clash between the two parties (Chagnon 1983: 171ff.; Ales 1984: 92-97). Warsusually take the form of surprise attacks aiming at killing as many enemies as possible while minimising their own losses, and at capturing women. Wars are often of remarkable cruelty and perfidy, and hostilities often culminate in treacherous feasts (nomohoni): here one group tries to convince an allied group to invite a third group which together can be overwhelmed and massacred during the feast (Chagnon 1968: 138f.; 1983: 152-168; Zerries/Schuster 1974: 197-236). Although no precise statements about wars per local group per decade can be found, the war frequency seems to vary according to region and time: the groups of the Namowei and Shamatari bloc in the west (in the Mavaca-Orinoco area) wage war more frequently than the groups of the other population blocs. Among the former the settlements are larger and mortality in war is higher. Up to 30% ofadult men die in war (Harnes 1983: 424), and Lizot (1989: 106) writes that "toutes les communautes de la montagne sont en guerre les unes contre les autres..4 • According to Lizot (ibid.: 103) the intensity of war decreases with increasing acculturation. Contradicting theories on the causes of war have been proposed: among them conflict over scarce game (Harris 1977; 1984), over trade goods in short supply (Ferguson 1992), because women are scarce (Chagnon 1968; 1988), in order to maintain the local groups' sovereignty (Chagnon 1968; Clastres 1977; Lizot 1989) and for revenge (Lizot 1989; Chagnon 1983). I will not discuss in detail here all these different theories because this has already been done extensively.5 1 will however, just briefly discuss some ofthem. It is highly questionable whether game and women are scarce cf. Lizot 1971: 149-168; 1977: 190202; Chagnon 1983: 57, 85f., 119; Ferguson 1989). But even if resources, such as' game were scarce, these are consequences of an already existing bellicose environment rather than causes of war (Ferguson 1989; Helbling 1995; 1996). It is war itself that forces local groups to adopt an expansive "reproduction strategy" (high growth rates) to increase group size as weIl as to intensifY production (hunting and farming) in order to recruit allies, thus leading to a shortage of local resources (Ferguson 1989: 186). Although, war may strengthen the 4 For new developments in the Yanomami area, especially on the influx of gold diggers and the impact of epidemics etc., cf. Helbig et al. (1989) and Chagnon (1992, chap.7).

5 Cf. Ferguson (\ 989; 1995); Harris (1984); Lizot (1977; 1989) and Helbling (1996) on the Yanomami.

The Dynamics ofWar and Alliance Among the Yanomami

105

cohesion of a 10cal group, this does not explain why groups wage war and incur high costs and risks of being defeated or annihilated. Furthermore revenge for previous killings or a cultural disposition for aggressive behaviour only make sense in an already warlike environment and therefore cannot explain it (Ales 1984: 97f.). Most ofthe above-mentioned theories do not take into account the context of states and the economic world system, in wh ich indigenous wars have always occurred (FergusonlWhitehead 1992). These wars always take place in the "tribaI zone" in which state and nonstate societies interact (ibid.: 5, cf. Wolf 1982; 1987). According to Ferguson and Whitehead, the competition for scarce import products (iron tools and weapons) - introduced by expanding states leads to an intensification ofwar in the tribaI zone. "What has been assumed to be pristine warfare now seems more likely to be a reflection of the European presence" (1992: 27). Important as the context of expanding states and world economy may be, it seems nevertheless imperative to analyse the internal logic of indigenous war, without falling back on rhetoric like "pristine wars". Thus a theory of tribaI war is needed that takes into consideration both the inherent logic of war and alliance, as weil as the wider political and economic context in its historical dimension. The theory of strategie interaction (game theory) could be of some interest here because it links the structural aspects with the strategie perspective of the actors. While the institutional approach explores the (changing) structural framework and incentive structure in whieh local groups pursue their interests, the strategic approach elucidates how actors decide and what (unintentional) consequences their strategic interaction may have for the institutional system (Axelrod/Keohane 1986: 252). I will first turn to the structural conditions conducive to war and deal with the models in game theory that analyse the strategic interaction of politieal actors. In the second part the dynamics of war and alliance among the Yanomami will be discussed. 1. War in a System of Anarchy

According to Koch (1974; 1976), war between 10cal groups in tribaI societies can be explained by the fact that there is no overarching institution, such as astate, which has the power to sanction agreements and to prevent the violent settlement of conflicts between local groups. According to Sahlins (1968: 4-13), tribaI societies are in "a state of war", a situation in which war is an imminent possibility at any moment. Koch and Sahlins take up the argument of Hobbes who explained endemic warfare by the absence of a common government (Hobbes 1968, chap. 13, 15, 17). Thus, war is not the result of an aggressive disposition of men competing for women and higher status, as maintained by Chagnon (1990), but of an all-pervasive uncertainty, of a fear of being annihilated and thus of the anxiety about survival. Furthermore, local

106

Jürg Helbling

groups do not fight eaeh other beeause of any positive effeets war may have on soeiety, but rather beeause there is no overarehing politieal strueture to prevent and stop these fights (Hallpike 1973: 455). The multipolarity of the politieal system, whieh eonsists of autonomous loeal groups, is a neeessary eondition of war but not suffieient in itself: hunter-and-gatherer soeieties also laek astate but usually groups do not wage war against eaeh other (although there may be some interpersonal fighting). Thus we must look for a seeond eondition. Contrary to hunters-and-gatherers, shifting eultivators have long-term interests in their fields; resourees are eoneentrated loeally and yield predietable outputs; monopolising a territory, therefore, pays off (Dyson-HudsoniSmith 1978). A loeal group may leave its fields only by ineurring high opportunity eosts (Harnes 1983: 397, 408ff.) beeause it is bound to lose its harvest and thus risk starvation until new fields elsewhere yield erops (Gross 1983: 436). Henee, loeal groups are unlikely to withdraw from armed clashes but have to adapt to the requirements of war (ef. Fried 1967; Helbling 1987; 1992). Among the Yanomarni and other warlike shifting eultivators, it is not the eompetition for searee resourees but the monopolisation of loeal resourees and the high opportunity eosts of mobility that make war highly probable or even inevitable. But still, not all tribai societies are warlike and not everywhere is war fought with the sarne intensity. The intensity of warfare depends on the frequeney of eontaet, which in turn depends on settlement density. Thus, eonfliets intensify with deereasing distanees between groups. On the other hand, epidemies and massaeres in the eontext of colonial expansion reduee settlement densities, and formerly warlike societies eease to wage war (Ferguson 1990a: 242 on the Panare, Piaroa ete.). The settlement density also influenees the eosts of mobility: mobility is espeeially limited where settlement density is high and where loeal groups are faeed with "social eireumseription" (Carneiro 1973: 171 ff.; Chagnon 1973: 250; 1983: 147, 153). Where settlement density is low, the intensity of war, too, is lower, and the eosts of mobility ean be redueed by having fields in different plaees. This will reduee the risk of starvation if the group has to reloeate after a military defeat (Lizot 1971: 152ff.). Where settlement density is higher, the intensity of war, too, is higher and loeal groups are inereasingly foreed to improve their military strength (Ferguson 1992: 213f.). Equally, feasts for allies are more eostly and, therefore, the fields tend to be larger and hunting more intensive (Lizot 1971; 1977). Thus, loeal groups fight, frrstly, beeause no overarehing institution sueh as the state ean prevent a violent settlement of eonfliets between politieally autonomous groups and, seeondly, beeause groups are relatively immobile and eannot withdraw from eonfliets with other groups without ineurring high eosts. Furthermore, the intensity of eonfliets and warfare inereases with the frequeney of eontaet and with settlement density.

The Dynamics ofWar and Alliance Among the Yanomami

107

Having described the structural conditions leading to war in tribai societies, we will now focus on the local groups pursuing their interests and interacting strategically within the existing framework. The logic of strategic interaction can best be analysed by the models of game theory. In a first step we will try to explain why the Yanomami fight each other so fiercely instead of settling conflicts in a peaceful way. In terms of game theory we have to explain why an iterated prisoners-dilemma game does not have a co-operative solution but evolves into a destructive zero-sum game. In a next step we shall deal with the complex interaction between enemies and allies by introducing a n-personzero-sum game. 2. War as the ResuIt of Strategie Interaction Local groups can be interpreted as collective actors both in war and alliance. Each local group consists of individuals of different position according to sex, age, status in the web of kinship and marriage etc. and, therefore, with divergent interests Chagnonl968: 131, 143f.; 1983: 27f., 129f.; Lizot 1971: 144ff.). Furthermore, a local group may be divided into two or more political factions each led by an ambitious leader with his own interests. Factional conflicts have to be dealt with by a village leader (tushawa) who is sometimes successful but in the long TUn cannot prevent the splitting ofa local group (Chagnon 1983: 6, 26ff., 152ff., 186). In spite of these facts, a local group can be treated as a collective actor because it usually acts in unison both in war and alliance. The warlike interaction of local groups can be described in terms of game theory as a prisoners'-dilemma game, the basic logic of which can be elucidated in the following way: Two local groups would benefit from settling their disputes peacefuIly, since costs and risks of violent conflicts could thereby be avoided. Let us now assurne that group A is willing to settle its disputes with group B in a peaceful way. However, because no superior force sanctions such an agreement, group A can not be sure that group B also prefers a similarly peaceful way of conflict settlement. Group B could increase its chance of military success by only pretending to stick to an agreement but secretly preferring a warlike strategy. In this way group A, not being prepared for war, could easi1y be defeated or even annihilated, as in the case of a treacherous feast (nomohoni). Because breaches of contract cannot be prevented under the given conditions and because it is too risky to rely on the peacefulness of the other groups, each local group has to avoid the risk of its defeat, or even annihilation. Groups, therefore, prepare themselves for war by increasing their fighting strength in order to deter possible attacks or, even better, to attack first in order to anticipate the attacks of other groups and to destroy the enemies. Hence, local groups wage war, paradoxically, for defensive reasons, i. e. because they cannot trust each other, and because a peaceful strategy would be too risky. Consequently, a Yanomami explaining the reason why they fight each other is cited as having said: "We are fed up with fighting. We don't want to kill any

108

Jürg Helbling

more. But the others are treacherous, and one cannot trust them." (Pfeiffer, J. in Horizon, Jan. 1977) Even in an iterated prisoners'-dilemma game with many players, a cooperative (peaceful) strategy such as tit-for-tat (tft) will not evolve. A tftstrategy promises to co-operate, as long as the other co-operates, but threatens with retaliation if the other no longer co-operates (Axelrod 1987: 32). According to Axelrod such a co-operative strategy is profitable if actors (must or want to) interact for a longer time period and, therefore, attribute a sufficiently high value (measured as discount parameter) to other's reaction to one's own action 6 : Each group will co-operate (settle conflicts peacefully), if the expected damage by retaliation is larger than the expected advantage of an attack. At least two reasons speak against such a peaceful strategy. In the first place a tft-strategy requires that the actors are informed about the intentions of the others and can communicate their own intentions to the other actors. However, as information about strength and intentions of other groups are always incomplete, the pretence of false intentions and the dissemination of rumours is possible and profitable (Biocca 1972: 142ff.). For this reason a tftstrategy is difficult to handle and risky, for each actor can misjudge the other actors and be misjudged by them (HirshleiferlMartinez Coll 1988: 381 f.).7 The risks involved may be reduced by stereotyping the other "players" according to past experiences (by considering the worst case) as weil as by acquiring a deterrent reputation in order to communicate one's own intention efficiently (Axelrod 1987: 132ff.). It is precisely in this context that aggressive behavioural ideals and a "tough guy" reputation are important and make sense8 : A "tough guy" threatens to strike back at the slightest provocation regardless of losses 9 . However, such a strategy has not only a deterrent effect, but escalates threats further because the other groups are convinced of the dangerousness and aggressiveness of the adversary and of the necessity of corresponding countermeasures. An aggressive behavioural ideal, therefore, encourages risk-prone behaviour which leads to mutual provocations and preventive strikes, as does 6 The discount parameter measures the influence of the other's future reactions to one's own action in the present (the "shadow ofthe future") and serves the calculation of the accumulated pay-off: The higher the discount parameter, i. e. the smaller the value loss of the next pay-off compared to the present, the larger the accumulated value of the long-term pay-off (cf. Axelrod 1987, parts I and 11). 7 The probability of misunderstanding increases and the probability of conditional cooperation such as tft decreases with the number of actors (Axelrod/Keohane 1986: 234f.). 8 On the waiteri complex at the Yanomami cf. Chagnon (1968: 112), Lizot (1989: 107). However, men who behave too aggressively are feared and hated by the other villagers as in the case ofRohariwe and Fusiwe (cf. Biocca 1972; Ales 1984: 108f.). 9 This irrational moment - attacking regardless of losses - may be rational in the context of a deterrence strategy (Deutsch 1968: 179-184).

The Dynarnics of Warand Alliance Among the Yanomami

109

the labeJling of other groups as potential enemies (sorcerers and malefactors), whom one has to mistrust. The escalating effects of mistrust, preventive violence and violent reputation are c1early reflected by the Yanomami conceptions of iIIness and death 10. Offensive mistrust and preventive violence are attractive strategies, because they allow the local groups to rely on their own strength, instead of depending on the uncertain· intentions of the other groups and on solving the security problem in complicated and risky negotiations (Vasquez 1993: 35; Evens 1985: 93ff.). A second reason for the failing of a co-operative tft-strategy lies in the very nature of the game itself. A tft-strategy pays off in a ceremonial gift exchange because it is always possible to stop exchanging gifts if the other refuses to reciprocate (cf. Görlich 1992). However, if the survival of local groups is at stake, a tft-strategy would be too risky because it would be too late to retaliate after a military defeat (Riker 1962: 31, 174; AxelrodIKeohane 1986: 232). In order not to be taken as cowardly and weak and, therefore, to be attacked, each group has to inspire respect bybehaving ferociously and aggressively and to demonstrate that it will leave no provocation unanswered, but strike back inexorably (Chagnon 1977: 41, 203; Ferguson 1992: 223). Although each local group would like to behave peacefully, it nevertheless has to attack pre-emptively because if it does not attack at a favourable moment it risks being attacked at an unfavourable moment (Waltz 1960: 5)11. Thus, the logic of massive deterrence, retaliation and pre-emptive attack leads to an uninterruptable chain of retaliation and counter-retaliation. Because no group can be sure that the other groups will renounce violent means when pursuing their interests, each group must reckon with the others using military force. This security problem can only be solved by arming and trying to gain military superiority, each group threatens simultaneously the safety ofthe other groups who try in turn to gain military superiority (Levy 1989: 224ff.; Otterbein 1988). Thereby, the prisoners'-dilemma game transforms into a destructive zero-sum game in which onegroup wins military superiority, if the other loses it (HirshleiferlMartinez-ColI 1988). What counts is, to win and not to lose, i. e. "to stay in the game" (Riker 1962: 22). Because the costs of moving away are 10 The Yanomami believe that deaths are caused by the sharnan of another village. Therefore, the sharnans of their own village have to employ counter-magie in order to defend it and to preventively attack the other groups. By making neighbouring groups responsible for death and iIIness, reasons to take revenge are permanently generated (Chagnon 1968: 112, 127f.; 1983: 176; Ales 1984: 99). However, only ifthe relations with another village are al ready in a bad state, is it accused of having used sorcery, and only if the group is strong enough to overwhelm the enemy will it take revenge (Lizot 1989: 105f.). Ifthe enemy group is too strong, their purported responsibility for a death is simply "forgotten" (Ferguson 1992: 223f.; Chagnon 1977: 118; Lizot 1989: 31 f.). II However, the extent of threat and therefore the necessity of preventive attacks depend on the spatial distances between the local groups.

110

Jürg Helbling

usually too high, each local group will maximise its military strength. The military strength of a group essentially depends on the number of its warriors. Hence, local groups pursue an expansive "population policy".12 3. War and Alliance The military strength of a local group, however, not only depends on its size and on the detennination of its warriors but also on the number and the reliability of its allies. Local groups have to fonn alliances against common enemies, and it is the common enmity towards third parties that makes (conditional) co-operation between allies both necessary and possiblel 3 • Alliance partners expect more from an alliance against third parties than from a war against each other, however each of them wants to avoid being cheated; hence, the mode of the alliance must be negotiated (Schelling 1960, chap. 3; Elster 1989, chap. 14). Wars and alliances between local groups may be described as an N-actors-zero-sum game: a zero-sum game between war coalitions, a bargaining game within a coaIition characterised by both conflictive and cooperative aspects i. e. the logic of the prisoners' -dilemma game (Barth 1959; Riker 1962: 81ff.). Each group wants to belong to the victorious coalition (being large enough to defeat the common enemy) but also to maximise its share ofthe spoil (land, women, strategic position etc.) within the coalition. A group looking for allies will try to concede as little as possible but it must concede enough to avoid losing its ally (Riker 1962: 121). In addition, the risk that an alliance partner will secretly desert, ally with the enemy and organise treacherous feasts has to be faced. If an ally has many enemies and, therefore, is desperately seeking an alliance, one can be sure that he will not commit treason (Chagnon 1968: 120). However, the propensity to fonn an alliance with a weak group is low if a joint victory is doubtful. Thus the Hekurawe-teri, when they were threatened by the powerful Shamatari, invited the Karawe-teri for an alliance feast, but the latter declined with thanks for obvious reasons (Biocca 1972: 49). In any case every group will profit from a (momentary) weakness of an ally and press hirn into an unequal exchange ofwoman and goods etc. So, even between allies the relative military strength is decisive: each group looks for allies who are weaker, but strong enough to jointly defeat common enemies. The bargaining power of a local group depends, firstly, on its relative strength (size and unity) and on itsdetennination; for instance, an aggressive 12 In order to increase the size of a group's fighting force, as many adult women as possible must be available in the group: through local endogamy, through unequal exchange of women with allies as weil a s through wife stealing from enemies. According to Harris (1977, chap. 4 and 5) female infanticide serves to raise the relative share of warriors in a group. 13 Among the Yanomami adjacent villages are either allies or foes; indifferent villages become either enemies or allies sooner or later (Chagnon 1983: 170).

The Dynamics ofWar and Alliance Among the Yanomami

111

group threatening to retaliate on the slightest provocation will influence the bargaining process to its own advantage. Secondly, the bargaining power of a local group depends on its alternative alliance options as weil as on its relative threat situation. A local group that is more threatened by a common enemy and therefore is more dependent on its ally has to make more concessions (more costly feasts, unequal trade and unequal exchange of women). Bargaining always requires information and communication (Riker 1962: 79). However, because strength and threat, readiness for concessions and alliance options of other groups are never fully known, information can be used strategically, that is, by withholding information, spreading rumours, using false pretences etc. (cf. Biocca 1972: 142-148). For the same reasons, misjudgements concerning the strength of opponents and the support from their allies as weil as the reliability ofone's own alIies are frequent, as the example ofthe treacherous feasts c1early shows. The bargaining process between allies of about equal strength can be described as a prisoners' -dilemma game: threats, promises and side-payments (Riker 1962: 109ff., 120) as weil as the groups' relative position in the regional balance of force must be considered. If the "shadow of the future" is long enough (high discount parameter), i. e. the common interests are strong enough, then a co-operative tft-strategy, i. e. an alliance, is adopted. The impression of being weak, however, must not only be avoided in relation to enemies but also in relation to allies because allies would immediately try to profit from it l4 . On different occasions - trade, feasts, marriages, wars - the reliability of allies is tested (Zerries/Schuster 1974: 197-215; Chagnon 1983: 148f.; 1968: 121f.). Meetings between allies seldom take place in a harmonious mood; on the contrary, they are always accompanied by mistrust and provocation, threat and cheating. These tensions may erupt at any time into open enmity ranging from club fights to treacherous feasts and surprise attacks (Chagnon 1968: 132ff.; Biocca 1972: 135ff.; Ales 1984: 92-97). Thus, a "tough guy" image is the best reputation for a local group, even when dealing with its allies. Because allies have to be deterred and intimidated by aggressive behaviour, additional conflicts are created in turn, and because each ofthem suspects the other ofbreaking away or even planning a betrayal first, both will try to anticipate the treason of the other and to break away first, as soon as the balance of force and the situation ofthreat has changed (Chagnon 1983: 147ff., 170ff.). The mutual interdependence between allies is not always symmetrical but may take a marked asymmetrical form: Iftwo allies are ofunequal strength and varyingly threatened, the better-positioned group will take advantage of its weak ally: the Momariböwei-teri and the Reyabobowei-teri were surrounded by 14 Cf the alliances between Bisaasi-teri and Monou-teri, between Momariböwei-teri and Reyabobowei-teri (Chagnon 1968: 153ff.), between Bisaasi-teri and Korohi-teri (Chagnon 1983: 165-168), between Namowei-teri and Mahekodo-teri (Biocca 1972: 156-164).

112

Jürg He1bling

enemies and thus largely dependent on an alliance with the Bisaasi-teri and Monou-teri. The latter groups were quick to force the fonner groups into an unequal exchange of women. The weaker allies had to accept this in order to prevent a disastrous defeat by a powernd enemy. Alliances are even more asymmetrical if a defeated group takes refuge with its ally. In this case the stronger group has even more freedom to exploit its alliance partner, as is seen in the example of the Kreibowei, who joined the Mahekodo-teri after having been decimated in a treacherous feast. AIthough the Mahekodo-teri supported them in waging war against the Möwaraoba-teri who, together with the Iwahikoroba-teri, decimated the Kreibowei in a nomohoni, they nevertheless forced the Kreibowei to cede many women, not without having threatened to kill all men in order to take all the women. Thus it is understandable that the Kreibowei tried to leave the Mahekodo-teri as soon as possible, to cultivate new gardens in a secure place and to become an autonomous and independent group again (Chagnon 1968: 119f.; 1983: 147-153; Ferguson 1995: 251-255, 305). A weaker ally is Iikely to behave according to the logic of the chieken game because the disadvantages of a non-alliance by far exceed the possible disadvantages of being allied. The stronger alliance partner, behaving according to the logic ofthe prisoners'-dilemma game will not break the asymmetrical alliance as long as it profits from it (unequal woman exchange, support by weaker partner in war against third parties), i. e. as long as the discount parameter is sufficiently high. 4. The Process of Escalation and De-Escalation The basic instability of an n-person-zero-sum game sets in motion a dynamic process of war and alliance. Alliances are always ambivalent and precarious; an alliance can quickly change its character and turn into enmity, accelerated by the suspicion that the other group may already have changed its intentions. The allies of today may easily become the enemies of tomorrow, or they may have become enemies already and may be secretly planning a treacherous feast. The relative strengths and balance of force between groups may change following victories, defeats and demographie hazard as weil as macro moves, thus leading to a constant regional restructuring of relations regarding enmity and alliance (Ales 1984: 103ff.). This also changes the pay-off matrices: the "shadow of future" may shorten (the discount parameter decrease) and a co-operative strategy in a prisoners' dilemma between allies may transfonn itself into a zero-sum game. Whether a local group decides on war or peace in a given situation, depends on its relative strength, i. e. its fighting strength and the number and reliability of its allies. The fighting strength of a local group depends on the number and detennination of its warriors, on its leadership and its armaments. The number and reliability of its allies depend on mutual concessions and commonality of interests, on their relative strength and threat situation (Brown 1994: 67-79;

The Dynamics of War and Alliance Among the Yanomami

1\3

Levy 1989: 243, 268-278). In order to start a war, there must be a sufficiently high chance ofthereby improving one's position in relation to the status quo. A group (or a coalition) will attack, if it considers itself to be stronger than its adversaries (Ferguson 1995: 293, 300, 302f.) or if it hopes to hit the adversary before the latter becomes stronger and finally too strong 1S . Sm aller groups may also start a war in order to gain respect from a stronger adversary. They try to compensate their military inferiority by foolhardy hit-and-run attacks and thus demonstrate their determination - regardless of losses - to maintain their sovereignty against enemies as weIl as allies (Chagnon 1983: 147, 182ff. on the Monou-teri). Local groups consider, therefore, the costs and advantages of war and its alternatives as weIl as the probability of a victory (Levy 1989: 240-245, 279-289). Not only the strength of a group and the number of its reliable allies but also the willingness of a group or its warriors to take risks are decisive with regard to war and peace (Bueno de Mesquita 1980: 383): While so me waiterimen are prepared to take high risks, other group members tend to be more cautious; village factions argue about whether to wage war or not (cf. Biocca 1972: 112-119 on the Namowei-teri).16 Hostilities run through different stages of escalation: they extend from the destruction of fields and the killing of individual enemies, ambushes and hitand-run attacks up to the siege of hostile villages, wars of annihilation, or persuading other groups to organise a treacherous feast (nomohom) in order to massacre their enemies. Attacks aiming at killing as many hostile men as possible and robbing their women are the most frequent form of war among the Yanomami. This mode of war has the inherent tendency of escalating hostilities still further (Chagnon 1983: 175; Biocca 1972: 26-34).17 On each level ofescalation a group has to decide on how it should react: escalate, wait and see or withdraw. The weaker group has the alternative, either of being annihilated, of getting the support of allies or fleeing nearby its aIly. However by deciding on the latter option, it would lose its fields and would be dependent on its ally until the new fields yield, an option wh ich is both costly and risky, even ifthe macro move costs may be reduced by having plantain fields in different places (Chagnon 1983: 71-78; Gross 1983: 436). Because hostile groups move away from 15 Thus the Shamatari attacked the Hekurawe-teri at a moment when most men were out for several days on a hunting expedition (Biocca 1972: 50-55; Ferguson 1995: 256 on the attack ofthe Shipariwe-teri against the Mahekoto-teri). 16 Sometimes village members plot against their far-too-aggressive leaders who provoke unwanted wars, as in the case of Rohariwe, tushawa the Konabuma-teri (Shamatari) who was killed by Fusiwe with whom Rohariwe's e10se relatives collaborated (Biocca 1972: 171-189; on Rohariwe and the tragic end ofFusiwe who killed hirn cf Ferguson 1995: 226-236). 17 This is in marked contrast to the Maring in the highlands of New Guinea among whom hostilities proceed through different levels of ritualised violence, and where the process of escalation may be interrupted - if the mutuallosses are about equalised - on each level (Vayda 1976).

8 Sociologus. Beiheft 1

114

Jürg Helbling

each other - enemies usually settle about half a day to three days away from each other (Ales 1984: 110) - the spatial distance between them will gradually dampen their hostilities, and conflicts de-escalate. Furthennore, their bellicosity sometimes decreases because allies refuse to participate in a raid as they are secretly intending to steal women in the absence of their allies (Chagnon 1983: 181ff.; Ferguson 1995: 291f.). While bellicosity decreases with increasing distance, new conflicts may emerge between allied groups settling in cIose range (from 2 hours to a day; Ales 1984: 103ff.). They will quarrel about the. distribution of the spoils (women, trade products), like the Karawe-teri and the Koroshiwe-teri (Biocca 1972: 26-34). Allies test each other's strength and detennination, look for new allies and try to obstruct each other's alliance policy (Chagnon 1983: 178). They fuel internal conflicts - competition between factions and between faction leaders - in order to weaken each other: Thus the Bisaasi-teri had to support the Monou-teri against the Patanowä-teri who in turn tried to win the Bisaasi-teri as allies by attacking only the Monou-teri and by weakening the pro-Monou-teri faction in Bisaasi-teri (Chagnon 1983: 122ff., 178). Mutual mistrust and the fear that the other has already allied with the common enemy will contribute to a further escalation ofthese conflicts. Allies will start to fight each other, or one of them will try to massacre the other in a treacherous feast. If the alliance partners finally become enemies, both ofthem will move to their new allies. Sometimes indifferent groups or even fonner enemies - foes of foes - may become allies and start to trade and to hold feasts together. Although Chagnon (1983: 170) writes that indifferent villages are often suspected of bad intentions - this explains why attempts at alliance between such groups often fail - he nevertheless reports on successful pacts, especially after such a group has split into separate groups, and one of them has shown interest in a new allylS. Factional conflicts within a village may lead to the break-up of a group; one of the factions moves away, but settles near the old village with a view to cooperating in the face of common enemies, or they become enemies forever as in the case ofthe Namowei-teri and the Bisaasi-teri. After the breaking up oftheir village and several attacks by the Bisaasi-teri, the weakened Namowei-teri moved near the Tetehei-teri where they were later fiercely attacked by the Bisaasi-teri, their fonner co-villagers, who increased their military pressure and adapted a tactic ofsiege (Biocca 1972: 135ff, 196-235). Due to the fundamental unreliability of allies and the constant threat by enemies in an unstable and insecure environment, each group strives for military superiority in order to prevail against its enemies but also to assert itself in dealings with allies. This fundamental volatility of both alliance and enmity 18 Cf. the unsuccessful attempts to form an alliance between the Kreiböwei and the Möwaraoba-teri who later even fought each other (Chagnon 1968: 151;1983: 1fT., 152) as weil as between Bisaasi-teri and Mahekodo-teri (Chagnon 1983: 156, 161-168).

The Dynamics ofWar and Alliance Among the Yanomami

115

may be largely explained by the (somewhat restricted) mobility of local groups depending in turn on low settlement density: Each group has the (somewhat costly) opportunity to move away if threatened by superior neighbouring groups or after a fission of the group. Therefore the tactieal goal of each group consists of not being expelled but of expelling (if not annihilating) its enemies by war, but never of seeking a truce or compensation payments (Biocca 1972: 26ff., 50ff., 198ff.; Chagnon 1983: 215-237). By having the opportunity to stop interaction with a neighbouring group by moving away, the probability of a continued alliance decreases (decreasing discount parameters) and mutual mistrust grows. Moreover, insecurity and mistrust between allies are not overcome by a large-scale exchange of gifts and women among the Yanomami, whieh would make alIiances more stable and binding. Thus, the politically motivated macro moves away from superior enemies to nearby one's allies, with whom new conflicts emerge in turn, generate the destabilising dynamics that make the alliance relationships short-Iived and enmities more erratic and inexorable than in other societies. 5. Summary

Although indigenous wars are not pristine wars but take place in anational and global context that influences the course and intensity of wars in the tribaI zone, they cannot be explained only by reference to this wider context. Important as expanding states and the world economic system may be, it seems nevertheless imperative to analyse the internal logic of indigenous war. Both the wider political and economic context in its historieal dimension, as weil as the inherent logic ofwar in the tri baI zone, should be taken into consideration. War can be explained structurally by the multipolarity of the political system consisting of autonomous local groups, i. e. the lack of an overarching institution, and by the fact that shifting cultivation limits the mobility of local groups so that they cannot prevent an armed conflict by simply moving to another place. This socio-cultural constellation forms the conditions under wh ich the strategie interaction of local groups takes place. War and alliance can be analysed according to models of game theory. Even in an iterated prisoners' -dilemma game a co-operative tft-strategy will not evolve because, firstly, it would be too late for an unprepared group to retaliate once it had been defeated, and because, secondly, the reputation of a "tough guy" - the best way to deter even with a tft-strategy - fuels the escalation of conflicts and leads to provocation and preventive attacks. Thus, a bellieose strategy is chosen because a peaceful one is too risky; this means that groups opting for war could do better at the expense of groups who adopted a peaceful strategy. A policy of deterrence, of strength - caused by a defensive concern for survival - will paradoxieally lead to escalation and a warlike settlement of conflicts. Each group strives to gain military superiority at the expense of the other groups that in turn strive for superiority. Thus, the prisoners'-dilemma 8*

116

Jürg Helbling

game evolves into a destructive zero-sum game. I have argued that the complex constellation of enmity and alliance may be analysed as an n-person-zero-sum game: a zero-sum game between coalitions, a prisoners'-dilemma game (sometimes a chicken game) within the coalitions between allies. Groups form an alliance if they do better in an alliance than in fighting each other, although each of them tries to outsmart the other. A specific quid-pro-quo has to be agreed upon by both partners. In this bargaining process a reputation as a tough guy pays and does not lead to a break-up of the alliance as long as the mutual interests are strong enough (high discount parameter). However, a change in the relative strength andlor threat by third parties may lead to a transformation of the constellation: allies become enemies, enemies become allies; the prisoners' -dilemma game between allies transforms into a zero-sum game between enemies or vice versa. Because of this fundamental volatility of both alliance and enmity, each group strives to be as independent and strong as possible by becoming larger and recruiting more warriors than its enemies.

11. Auto-Regulation of Violence and Escalation

Violent Conflicts in West-African Borgu on tbe Eve of Colonisation By Erdmute Alber

Introduction

In pre-colonial Borgu I, raids and feuds happened on a daily basis. Due to the constant threat of being raided, travelling traders considered the region unsafe. 2 Stories of wars and raids, of slave and cattle raids, violent succession struggles and caravan raids are still very much present in the memory and narratives of today's Baatombu. 3 The Baatombu praise the boldness oftheir ancestors in war and describe men who excelled in raids and feuds as being held in high regard. Nowadays, forms of appropriation similar to raids are still topical in discussions about the present situation. Often the answer to the question, what distinguishes a powerful person, is that he can take anything he wants, or that this or that man is so strong that he cannot be refused anything he demands. Such statements are based on the notion that the exercise of power is, to a large extent, based on the exercise of violence. 4 Those men who are described as perI The social and geographical area of historical Borgu is not identical with the province of the same name in the North of today's Republic of Benin. Historical Borgu comprised about 70,000 km 2 ofpresent Nigeria and the Republic ofBenin and stretched from the River Niger in the East to the Atakora Mountains in the West. Cf. Lombard (1965) and Kuba (1996). The material outlined in this paper is based on 18 months of research in the field and in archives between 1991 and 1995. This research took place exclusively in the Benin part of Borgu which had been colonised by the French. Cf. also Alber (1994 and 1997a). I thank Sabina Matthay for the translation of this paper. 2 Cf. Lander, J. and R. (1832); Clapperton (1829). 3 In today's Department du Borgou, about 420,000 Baatombu (Universite Nationale du BeninIMinistere d'Education Nationale (ed.) (1994: 116) represent approximately half of the population, they are the largest and were - due to the military superiority of their warlords - the dominant group in the ethnically pluralistic group of historical Borgu. 4 Popitz (1992: 223) introduced the term "basic legitimacy" (.. Basislegitimität ") of power in order to describe those processes which make people bow to the power of the powernd, come to terms with it or accept it. Different from Max Weber 's concept of three types of legitimate rule, for Popitz the subjective part of those who accept the

120

Erdmute Alber

sons who can take anything they want are often chiefs; this is noteworthy because today their power is rather Iimited. The concept of power behind such problematic statements therefore refers not only to recent structures but also to historical experiences. The close connection between recent concepts of power and those of war and violence is surprising considering that raids and open warfare were banned in western Borgu by the French at the beginning of colonisation. Baatombu, as weil as French officials, claim that raiding and feuding ceased during the first two decades of this century.5 If some of the Baatombus' concepts of power refer to raids and feuds even today, after 70 years of pax ga/lica, these must have been such central processes in pre-colonial Borgu that it is worthwhile investigating them. In this paper, I will try to explain forms and functions ofwarlike conflicts in Borgu on the eve of colonisation. 6 I want to prove that an emic concept of the exercise of power of the Baatombu is based on two modes of power: violence and wealth wh ich can be converted into each other by a cycle of (violent) appropriation of goods and of transformation of wealth into prestige. This concept still exists, though the cycle was altered by the ban on open raids. For pre-colonial Borgu, this not ion ofpower is a central indicator for the explanation of the perpetuation of violent conflicts: sincepower is not a stable factor but must be permanently proved in warlike actions and by the (violent) accumulation of goods, raids and feuds are reproduced and perpetuated. They were important parts of a system of redistribution in which goods and prestige were converted into each other. All social groups were involved in these warIike conflicts, either as victims or because - in various measures - they profited from them. The relative stability of multi-ethnic social co-existence in pre-

power and intemalise it is decisive (1992: 227). The term "basic legitimacy" therefore seems to be especially useful for the investigation of emic concepts of power. Von Trotha (1994) showed in his analysis of colonial Togo that violence can represent an important legitimisation of power. 5 According to Baatombu sourees, the last violent conflicts took place in the second decade ofthis century. In the monthly reports ofthe colonial commanders of Borgu the last arrest of a man who contravened the raiding ban was registered in January 1921 (Archives Nationales du Benin, Porto-Novo [ANB] 22, Rapport Mensuel Trimestriel [RMT] January 1921). 6 Wherever, for reasons of brevity, I speak of pre-colonial Borgu in this paper this means Borgu on the eve of colonisation. Pre-colonial historical transformations are not the topic of this paper nor is the question of whether the late 19th Century should be regarded as an exceptional situation due to violent invasion attempts. Cf. Lombard (1965: 65ff. ) and Kuba (1996: 325ff. ).

Violent Conflicts in West-African Borgu

121

colonial Borgu therefore was not threatened by violent conflicts, on the contrary, these conflicts contributed to stability.7 In order to explain these theses empirically, I shall begin with abrief description of the various groups and the warlords in multi-ethnic Borgu. After that, I will describe the various forms of warlike conflicts and the involvement of individual groups. In that part, I will also outline mechanisms to regulate conflicts. Then I will deal with the second mode of power at the Baatombu, i. e. the redistribution of wealth by means of gifts. Finally, I will describe the above mentioned cycle connecting those two modes of power. This paper is based on a small number of written documents from the precolonial times and on documents from early colonial tim es as weIl as on the picture of pre-colonial times drawn by today's Baatombu in discussions and narratives. 8

1. Warlords in the Multi-Ethnic Borgu On the eve of colonisation, Borgu was a multi-ethnic society dominated by Baatombu and other peasant groups - mainly Boko and Mokolle - with similar social structures. Along the caravan routes a town-dwelling Islamic population settled who mainly lived on caravan trade. There existed protective alliances between them and the warlords. They also traded locally in horses, cloth, luxury goods and slaves in Borgu. 9 Apart from the Islamic traders, there were cattle-holding Fulani and Gando, the slaves ofthe Fulani and the Baatombu. 1O 7 Max Gluckmann (1966: 16ff.) points out that feuds and violent conflicts can be stabilising factors in a society. His programmatic statements can be verified by applying them to pre-colonial Borgu. 8 Jones (1990) points out that with the colonisation of Africa not only the societies themselves, but also their interpretation of the pre-colonial past changed. From this perspective one could ask if Baatombu stories about their history can really be regarded as historical sources or if they should be treated exc\usively as testimonials of the interpretation of the present. Transfonnations of the interpretat·ion of the past, though, do not come about at random, but are the result of processes whose historical reconstruction is in principle possible: Existing or new elements are changed, integrated or suppressed. In this case, raids and feuds are - due to the pax gallica - certainly not elements of collective memory which were generated by colonial transfonnation. The fact that they survived the colonial period underlines their importance in pre-colonial Borgu. Secondly, I find no difference between Baatombu stories and written records which indicate that today's Baatombu representations of pre-colonial feuding and raiding are fundamentally different from those at the beginning ofthe century.

Cf. Dramani-Issoufou (1981); Kuba (1996: 233ff.); Lombard (1965: 80ff.). Even today, Baatombu distinguish between Gando, Yobu, Mareyobu and so on, according to their origin and the kind of relationship of the fonner slaves to the powerful. I avoid these complex distinctions by using the collective tenn Gando which is also used in Borgu. 9

\0

122

Erdrnute Alber

While Fulani and Gando had no regional power structures of their own, the Baatombu and Boko peasants were dominated by regional and local warlords who were Baatombu and Boko themselves and who held most of the village chief posts. Their power was mainly founded on violence exercised at raids and feuds and - in connection with this - on the protection-booty complex. Before describing the warlike violence, I would like to outline two more characteristics of the warlords: Their power was based solelyon providing protection for the peasants, not on providing goods or resources necessary for agricultural reproduction. They were neither responsible for the distribution of land - of which there was a lot anyway - nor for organising agricultural work processes. Agricultural production was organised autonomously within the household. The exercise of power was therefore Iinked to violence and violent protection, its range was Iimited and it remained relatively unstable so that it could not become the rule, i. e. "Herrschaft" as defined by Max Weber. For example, there were no fixed territorial structures: areas of influence constantly had to be re-defined and were always shifting. The peasants had quite a lot of possibilities to escape these warlord-chiefs: they could "go into the bush", establish new villages or evade the demands for dues by disappearing. In short: exit options (Hirschman 1970) were effective patterns of reaction to the specific forms of the war lords ' exercise of power in sparsely populated Borgu. Secondly, the Baatombu warlords had relationships of power to other groups, especially to the Fulani of the region. The Fulani had no regional chiefs of their own but had protective relationships to "their" respective Baatombu warlord who provided protection from raids by other warlords and who received dues in the form of cattle for this protection. This specific form of power, which is based on protection and booty, was therefore applied interethnically. It was not restricted to the Baatombu, but characterised and stabilised the social structure ofBorgu as a whole and provided one possible form of . inter-ethnic relationship. 11 2. Feuds

On the eve of colonisation, the power of Baatombu warlords was founded on the exercise or the threat of violence. Its instability and limited range was due to the specific forms this violence took, i. e. raids and feuds. Both are warlike actions with limited local range which - unlike bureaucratic rule - are unable to exercise control over relatively large periods of time and relatively large areas. Jl Boesen (1995: 421) rightly points out that research up to now has mainly dealt with the "political" relationships between Fulani and Baatombu. But the concrete and individual relations between Fulani and Baatombu were not only not "completely or even to a large extent determined by political dependency", they "could on the contrary be specifically informal and free ofrole constraints".

Violent Conflicts in West-African Borgu

123

The demands of a Baatombu warlord could be pushed through concretely and violently only at the place where his power of enforcement actually was. In precolonial Borgu there was no bureaucracy or any other comprehensive institutionalised form of representation which could assert his interests in his absence or which could influence local decisions. Not only were there no central authorities for the whole of Borgu 12 , local conflicts did not develop into wars which involved the whole society, either. For a detailed description, I will now outline the various warlike actions and define types of warlike actions, starting with feuds. Countless Baatombu stories about succession struggles deal with violent conflicts between two rivals for succession to a post like that ofthe sinaboko of Nikki (cf. footnote 10 on the sinaboko). Apart from stories of poisoning and witchcraft, there are reports about open warfare between two warring parties who both claim the same position. While raids served the purpose of appropriation ofbooty, feuds were fought to eliminate or weaken a rival. Feuds were often fought within one clan or between relatives. In stories, two classificatory brothers often fight each other and each of them is supported by their relatives on the matemal side. Conversations with Baatombu also show that kinship is not automatically regarded as an area within society which is free of violen ce and which provides cohesion. On the contrary: especially the constant competitive struggles for posts were often fought between relatives or members of the same clan \3. Rules of succession such as the privilege of the eIder brother were in practice rarely obeyed. In Borgu, violent conflicts took place within the society, they did not follow ethnic or territorial boundaries between Baatombu and strangers and they did not stop at relatives or clans.

12 There are regional chiefs whose area of influence is larger than that of a village chief. There are for example the saka of Kandi or the sinaboko ofNikki. At his court the largest and most important festival of western Borgu takes place, the Gaani festival. Though these more central chiefs did have a lot of prestige their area of command did not transcend the place. Cf. Lombard (1970). By integrating local chiefs as "chef superieur" in their colonial administrative structures, the French obviously overestimated these chiefs' power of command. The colonial records continuously complain about their lack of power (cf. RMT September 1900 in: ANB, series 22). \3 As a counter force to hostility and competition among relatives, the Baatombu know institutions which are meant to create cohesion, above all the adoption of children (children are not raised by their physical parents but by their classificatory siblings or parents). The adoption of a child is meant to create especially solid relations between physical parents and adoptive parents.

124

Erdrnute Alber

3. Caravan Raids Raiding as the other fonn of warlike conflict did not aim at eliminating rivals but was a way of capturing booty and therefore a way of acquiring wealth. According to the type of booty, caravan, cattle and slave raids can be distinguished. Caravans crossed Borgu mainly from East to West, they transported and traded cloth, salt, cola nuts and luxury goods. Along the central trade route which ran from Bussa to Nikki, Paraku and Jugu (Djougou), trade was relatively intensive although Borgu was considered unsafe. The caravans stopped at stations wh ich during the years had come into existence at larger places, they stayed a few days and paid dues to the powerful to ensure their protection. This protection was quite effective. The Lander brothers (1832: 201) describe how they met representatives ofthe sinaboko ofNikki in Bussa after thy had raided a caravan carrying cola nuts. When the King of Bussa intervened, the people from Nikki had to return their booty to the traders. The warlords of Borgu were responsible for the insecurity along the trade route through Borgu and they also received dues from caravans passing through the area. This reveals a central element ofthe raid, namely the above mentioned relation between protection and booty: the warlords received payment for the protection they provided to the caravans against the raids from other warlords, but they also threatened the protection areas of those other warlords. In this way, they perpetuated the need for protection and also provided it. They were constantly creating and renewing protective relationships as a central source of power. The fact that even the sinaboko of Nikki commanded or ordered raids shows that raids were regarded as a nonnal and central part of the power of the chiefs in pre-colonial Borgu. According to the Baatombu, the caravans were usually accompanied by peopIe who knew where potential attackers might hide. For atoll, travellers obtained a piece of bark from the warlord of the region on which his symbol was scratched. With this they passed through the region unscathed to the area of influence ofthe next warlord. Victims of caravan raids were rarely people from Borgu but strangers travelling through the region. The fact that the creation of protective relationships could really help to avoid the loss of goods proves that caravan raiding as a social practice of redistribution of resources was more than simply an arbitrary action. In order to minimise the danger to one's own caravan, one had to follow certain ruIes, e. g. provide certain persons with resources such as protection monies. Caravan traders were forced to follow these rules, otherwise they would fall victim to a raid in which the raiders took an amount of money or goods which was higher than the protection payment.

Violent Conflicts in West-African Borgu

125

4. Cattle Raids Caravan raids were mainly carried out against people who did not live in Borgu, but cattle raids were directed against cattle holders settling in the area, mainly Fulani. The warlords and their retinue came suddenly and unexpectedly, attacked Fulani farmsteads and took away cattle and other easily transportable booty. Among the Baatombu, there is the widespread view that Fulani have a special relationship with their cattle, which is inexplicable to the Baatombu. They describe this relationship as extremely caring and it also has magical connotations. During one of my conversations with an old man from the little village Tebo in November 1993, the conversation turned to "the Fulani": "The Fulani, he said, have the magical ability to sense danger for their animals. They sensed for example when strange warlords were approaching. Then they directed their cattle rapidly to the Baatombu village where their warlord lived. In the village war drums were immediately beaten, the warlord gathered his warriors and ordered a counter attack".

This and other reports show that the protection-booty relationship structures cattle raids too. The old man's story refers to the special relationship between the Fulani and "their" warlord. It emphasises that organised warlike defence of the Fulani and their cattle was not organised and carried out inter-ethnically (therefore the episode does not take place at the Fulani village), but that - as soon as the Fulani were threatened by raids - they went to the vilIage of the warlord who then defended them with Baatombu warriors. From a different perspective, this inter-ethnic relationship supports the theory that the protection-booty complex, which is based on violence, is a central source of power in pre-colonial Borgu. The internal akephaly of the Fulani in pre-colonial Borgu l4 has often been emphasised; in a multi-ethnic society wh ich is characterised by violent contlicts this akephaly can only exist if the necessary military protection is provided by an extra-ethnical factor. Warlike violence as the central source of power thus remained a monopoly of the Baatombu warlord. In both groups' system of values this monopoly had advantages: the Baatombu warlords did not have to reckon with institutional rivals from the Fulani society who might dispute their monopoly. The extra-ethnic monopolisation of violence implied the support of akephalous structures on the part of the Fulani and this complied with the Fulanis' high esteem of equality which has repeatedly been emphasised by the Fulani research. While the warlords provided protection to their "own" Fulani by threatening other warlords they created and perpetuated the need for protective relationships and therefore the basis oftheir role. 14 On the akephaly ofthe Fulani cf. Lombard (1965: 132ff. and 173), Bierschenk and Forster (1987), Bierschenk (1989; 1993), Boesen (1994). The colonial administration appointed chiefs from the Fulani themselves (Bierschenk 1993).

126

Erdrnute A1ber

The price for this delegation was the cattle and the dues the Fulani had to pay to "their" protector. Even today people often claim that the money and the food for the lavish Baatombu festivals were actually provided by Fulani and Gando. 15 According to reports by Baatombu, the Fulani as a matter of principle were never involved in the actual warfare while the Baatombu and their Gando actively fought as warriors under the war lord. Apart from building relationships with powerful war lords, the Fulani, like the peasants, could choose exit options to deal with the omnipresent violence. These included a change of location which meant building new protective relationships with other warlords. 16 5. Slave Raids Slave raids as the third form of raids were directed against all people in Borgu and the adjoining areas, with the exception of the Fulani who were ne ver enslaved. There were raids on Yoruba-speaking groups in the South and the East ofBorgu, on the Hausa and on small groups in the area oftoday's Atakora province. But most of the slave raids happened in Borgu itself. Many planned and organised attacks were directed against Gando villages. Another type was the abduction of single persons or whole families to other places where they were settled as Gando for a master. Apart from this slave trade, occasionally slaves were sold across the borders ofBorgU. 17 The fact that the Baatombu also took slaves within their own group is significant for the social structure of Borgu. This is a marked contrast to the Fon, the great slave hunters and southern neighbours of Borgu, and of serious consequence theoretically. The slave raids of the Fon were directed exclusively against members of other ethnic groups. Thus, internal cohesion and structures of a centralist state could develop wh ich were based on organised slave hunts. With the Baatombu, inner-ethnic cohesion was systematically minimised by general insecurity and threats, Lombard' (l965: 353) rightfully argues that this

15 These frequent answers also refer to a stereotype which does not necessarily correspond to reality. Fulani and Gando were mentioned to me far more often as the source of income before regional festivals such as enthronement ceremonies than money from cotton growing. But a considerable part of this money was invested in these festivals and caused rising bride prices in some regions (Bierschenk 1987). 16 The reports of colonial commanders from the early colonial period are fuH of co mplaints about the great fluctuation in Borgu. Fulani often crossed the Anglo-French border in Borgu. But the frequent change of place within French-ruled Borgu also troubled the administration (cf. ANB, series 29, letters of7.8.1924, 16.10.1922,28.7.1927 et al).

17

Caravan stations along the trade routes were places ofslave trade.

Violent Contlicts in West-African Borgu

127

was a reason why local power could not develop into astate with central structures. Occasionally, even relatives were enslaved in warlike contlicts. I will now explain the important structural elements of slave raids by means of a story which an old woman from Tebo told me casually in October 1993 about her grandmother. This grandmother was the aunt of Sero Toro Swaru, one ofthe most powerful warlords in the history ofTebo, who was sinaboko of Nikki from early colonial times, i. e. from 1901 until his death in 1914. The episode must have taken place shortly before colonisation, at a time when Sero Toro Swaru hirnself was involved in raids and counter raids. The old woman toldme: "Once my grandmother had gone into the bush to collect wood, when she was surprised by a raid. Sero Kpera and his people stood in front of her, took her prisoner, carried her ofT and were going to seil her as a slave. When Sero Toro Swaru heard about this, that his aunt had been taken prisoner by Sero Kpera, he set ofT straight away, went to see hirn and threatened hirn. He said that Sero Kpera had carried Sero Toro Swaru 's aunt ofT and if he didn 't let her free straight away, he would be in trouble with hirn. So Sero Kpera set my grandmother free straight away."

In order to understand this episode,· one needs to know that Sero Kpera came from a village not far from Tebo. He was a nephew of Sero Toro Swaru and with the help of Sero Toro Swaru's younger brother later became hirnself sinaboko ofNikki. The episode illustrates that the threat of raids was very much present in the peasants' every Iife. The message conveyed by the story is that everybody could be affected by it everywhere even while engaging in such a banal activity outside the village as wood collecting. Secondly, raids here do not affect "others" but take place within the ethnic group ofthe attacker, indeed within his kinship group. This episode is not typical for larger, organised slave raids, though. According to the Baatombu, these took place in far-off regions of Borgu, since they had too many social relationships and obligations to people in the immediate surroundings. Raids in the immediate surroundings were smaller, they were often directed against a certain warlord in order to do hirn damage. The episode told in the story must have been such a raid. 18 Thirdly, the aunt is saved through her (protective) relationship to a powerful person, namely Sero Toro Swaru. He in turn can expect to receive protection payment and dues for his conduct. 18 I do not suppose that stories historically happened in exactly the way they are told. They do however reveal basic patterns and repeated actions and contlicts. But this old woman was always surprisingly accurate where personal details were concerned. Many details of her description of stories about Sero Toro Swaru correspond in a surprisingly precise way to written records ofthe colonial administration.

128

Erdmute Alber

Fourthly, specific knowledge about adequate behaviour in violent contexts is required: only after the abduction and by means of a threat, Sero Kpera did find out that the victim of the raid is the aunt of Sero Toro Swaru. Thus he had unwittingly committed a transgression wh ich he corrected as soon as he knew about it. Part of the violence was obviously mitigated; in this way raids and feuds in general took place in a rational manner and did not waste many resourees. Fifthly, this story hints at an element of revenge and therefore at the selfreproducing structure of raiding. If Sero Kpera had not set the aunt free he would have been in trouble with Sero Toro Swaru: raid is then answered with raid, feud is answered with feud. Sixthly, the story shows that the warlord really feels responsible for his proleges. In other stories it is constantly pointed out that it is a disgrace for hirn if "his" people are abducted by others. The informants point out that the warlord needs the other peasants because they make up the foot soldiers for raids and because they supply other forms of support, e. g. by means of their magical abilities and knowledge. 6. Redistribution at Baatombu Festivals

All population groups in multi-ethnic Borgu were affected by raids and feuds: Fulani and Gando were affected by cattle raids, travellers and Islamic traders were victims of caravan raids, slave raids were directed against Baatombu and other peasant groups and against Gando. As warriors, Baatombu and Gando actively took part in conflicts and could become warlords themselves. The warlords on the other hand represented no isolated autonomous group who pushed their interests through by means of violence only. They were on the contrary specialists in warlike conflicts and as such lived in the peasant villages with wh ich they were closely connected by relatives, marriage, friendship and social obligations. They could not keep the booty they took in raids exclusively for themselves, these goods were redistributed to a relatively large part of the population. Along with violen ce the redistribution of captured wealth was a pillar ofthe warlords' power. Finally, I would like to describe this correlation. Surprisingly, Jacques Lombard's description of the Gaani festival in Nikki, the largest festival in western Borgu, starts with the description of violent conflicts. He writes: "In principle, the raids started one month before the festival. At that time trafik between the caravan stations was halted and no strange trader dared to travel through Borgu." (Lombard 1965: 33)

Bierschenk (1989: 46) also notes that in pre-colonial times the Fulani always had to expect the loss of part of their herds before larger festivals. Both authors

Violent Conflicts in West-African Borgu

129

take the correlation between festivals and raids as a matter of course so that it seems unnecessary to check it. It is banal to note that there was a special need for goods and food at times oflarger festivities which could be met by raids. Even today, there are systematic collections of goods before larger Baatombu festivals which hint at the old system of(protective) payments and raids: the Gando take their contribution in the form of cattle or foodstuffto Baatombu villages. Relatives of the persons organising the festivals visit hamlets of Gando and Fulani and return to the Baatombu village heavily laden. Before the Gaani festival the whole region around Nikki sends foodstuffs to the sinaboko which are redistributed during the festival to feed the persons present. 19 Apart from the special need for foodstuffs to feed guests, at some Baatombu festivals there are other forms of ceremonial redistribution of wealth: at funeral ceremonies for outstanding persons (goo yero), inauguration ceremonies (di yaaru) and at the Gaani festival in Nikki, ritual distribution of gifts is the c1imax of the festival. The proceedings are observed by the audience, whilst money, c10thes and cloth are given to those minstrels who are present. In Nikki, where all important personalities and candidates for headman posts take part in the gift ceremony, this ceremony is extremely competitive: the persons bestowing gifts try to appear more generous and richer than their competitors. This ritual giving of presents at Baatombu festivals has a potlach-like character and also has "agnostic traits" as described by Marcel Mauss (1968: 24). It is conspicuous that the persons giving presents are either persons newly elected to a post (at the di yaaru) or candidates for posts (at funeral ceremonies of chiefs). In the latter case, the competition between the candidates is more than obvious, because the contest for the most generous (and the most prestigious) gifts opens the contest for the succession to the dead persons' post. Even today, all important officials of the region, candidates for posts and other important persons attend the Gaani festival at Nikki every year on horseback and c1ad as warlords for the ritual distribution of gifts. Immediately afterwards, the persons in the role of warlords enact spectacular martial shows on horseback. Recipients of the gifts are the minstrels. In return for the gifts, they ritually praise the warlords and proclaim their glory. At the giving ceremony, goods are immediately "translated" into prestige. 20 This prestige is important for recruiting followers needed for raids to the warlord and in order to gain an important post. 19 By redistribution, I - Iike Sahlins (1965) - do not mean that as much is given away as had been gathered at the points of concentration. As a matter of fact not everybody concemed profi ted from such redistribution at Baatombu festivals. 20 Steiner (1977) coined the term "translation" for the transformation of basic commodities into "ritual values" Iike the prestige of the person bestowing gifts upon others at festivals. Cf. Meillassoux (1968) who points out the aspect of destruction of value at the generation ofprestige by gifts to the West African ministers.

9 Sociologus. Beiheft I

130

Erdmute Alber

The ceremonial distribution of gifts to the minstrels is one example of the distribution of gifts; there are other examples outside ceremonies wh ich I will not describe in any more detail in this paper. Normally, lavish distributions of gifts from Baatombu chiefs to important visitors and to their retinue were part ofthe normative standard. Chiefs had to exhibit their power through gifts while simultaneously receiving large payments of dues and gifts from their subordinates. In this way, warlords also saw to it that their retinue - for the purpose of war and prestige - was equipped with horses, weapons and c1othes. 21 The recipients on their part normally distributed part of the gifts to "their" people?2 The number of gifts was a sign of the wealth and the power of the person gi ving them. Gift transactions were popular and important topics of conversations about an official because they were a measure of their power: the more someone is able to redistribute, the more goods he is able to appropriate. I have described the ritual public redistribution of gifts to the minstrels at festivals as an example for other gift transactions because the central aspect of the generation of prestige is especially c1ear. The minstrels are mouthpieces of the powerful and disseminators of knowledge, as such they are the social specialists in prestige 23 and therefore predestined to receive publicly displayed and ritualised distributions of gifts.

21 Even today, the present of a horse to a subordinate raises the prestige of a person and is extremely desired. Today, horses epitomise prestigious objeets. Sinee they have lost their military neeessity, horses are regarded as pure luxury beeause grooming them requires a lot oftime. Only people with the time andlor staff for this keep a horse. Even now horses are never used for agricultural purposes. 22 In order to illustrate the ehannels of redistribution, a fanner once told me, if he brought a eow to the sinaboko he could do with it what he wanted. But the sinaboko was morally bound to give part of it to each of his dignitaries and other important persons. None of these though could just eat their meat alone, but had to leave part of it to family and village. "Maybe", the peasant said, "I sit here in TEb::l and finally eat part ofthe hind leg ofthe cow I sent to my sinaboko to Nikki." 23 My definition of prestige follows that of Erdheim (1973) who defines prestige as the knowledge of society of the exemplariness of an individual rather than of exemplariness itself (1973: 26). Firstly, he stresses thereby that prestige is no stable value but must constantly be aequired and eonfinned. A single spectacular warlike action is not sufficient, instead a warlord must prove himself again and again. By reference to knowledge, Erdheim secondly emphasises the stage-managed part of prestige: only demonstrative wealth, publicly noticed and judged gifts and the public relations work of the griots seeure prestige. Only Erdheim 's eonnection between prestige and exemplariness, which is based on the notion of homogenous common values, is doubtful. Warlords can have prestige without their power being regarded as exemplary by everybody.

Violent Conflicts in West-African Borgu

131

7. Raids and Gifts The connection between raids and gifts can be summarised as folIows: a warlord acquires goods and money at raids and dues by threat of violence, their amount depends on his success and on his military strength. His wealth is redistributed by way of gifts and is thereby transformed into prestige. This is necessary to bind his retinue to the warlord or to gain new followers. The more followers a warlord can mobilise, the more successfully he can organise raids and feuds. Prestige is therefore transformed into a warlord's military strength; this enables hirn to undertake new raids to capture goods which are redistributed in the form of gifts, as described above. Wealth is thus constantly transformed into military power and vice versa. This connection is equivalent to an emic concept ofpower. The Baatombu have no transcendent abstract term for power but there are various terms wh ich refer to power which is based on concrete media. Most important among these concepts are gobi and dam. In many conversational contexts gobi is simply translated as "money", but it often means power based on wealth. Dam on the other hand can be translated as "force" in many contexts and means power based on physical violence. The sentences "u gobi m:>" ("he has money") and "u dam m:J" ("he has (physical) strength") are employed almost synonymously as standard answers to questions like why this or that man holds a prestigious position or office. 8. Conclusion

In order to explain the perpetuation of raiding and feuding as a constant situation of social normality in pre-colonial Borgu, I have tried to describe emic concepts of power. Power means wealth and physical strength for the Baatombu. These are closely linked and can be "translated" into each other. They are resources necessary to install and to maintain domination. Violence exercised in warlike conflicts and wealth proven by gifts must be demonstrated again and again in order to support the warlords claim to power and his capacities. This norm establishes "raid" as a permanent state of affairs in pre-colonial Borgu. Up to now political anthropology related to Africa - with the exception of the investigation of ethnic views of concepts like "own" and "strange", "we" and "they" which was initiated by ethnicity research 24 - does not often deal with emic concepts. 25 The extrapolation of emic concepts of power, however, Lentz (1993) summarises modem ethnicity research with reference to Africa. Schatzberg 's (1993) comparatively general study which refers to the whole of Africa is an exception; he describes the "consumptive" aspect of power in Africa with methods derived from discourse analysis. 24

25

9*

132

Erdrnute Alber

can be a useful contribution to understanding power structures. "Power" does not always mean the "chance to push one's own will through against opposition within a social relationship", as Weber (1985: 25) wrote. The warlords in Borgu were not so much interested in influencing social decisions in their respective "areas of power" but in the acquisition of prestige through violence and gifts which led to posts and honours, without denoting exercise of power in the sense of decision-making. 26 Investigating emic concepts could enhance the scientific discussion of warlike contlicts. This would have to be done by way of comparative research. The comparative exploration of emic concepts of power and emic typologies of violent contlicts as weil as emic notions of peace as a central theme would then be the next step in research.

26 Dahl 1957. Cf. Luke 's (1972) excellent study on criticism of the one-dimensionalism of decision-making.

Humiliation and Reconciliation in Northern Albania The Logics of Feuding in Symbolic and Diachronie Perspectives

By Stephanie Schwandner-Sievers

Introduction "Northem" or "High Albania" is the region with the highest mountains and inost remote valleys in Albania. Here, scarce economic resources have always correlated with high labour and poverty emigration, called kurbet and a family - based subsistence economy.\ In 1991192, after the collapse of the Albanian communist system, blood feuding occurred in this area to an extent that had been known only in pre-communist times. The actual figures range between 600 known feuds (Albanian Daily 1996) and about "70,000 people actively involved in feuds throughout Northem Albania" (Vickers 1997: 9). The number of cases given depends on whether individuals or families involved are counted, although some observers claim that these numbers tend to be exaggerated for reasons of sensational joumalism? The Albanian expert Ismet Elezi stressed that the phenomenon was limited to the Northem regions of "MirMiti!, Kurbin, Malesia e Madhe (including the Dukagjin area) and Puke.,,3 The Northem regions ofHas, Kukes and Tropoja could be added. 4 While conducting fieldwork in this area in 1992 and 1993, among those interviewed were randomly selected young men in a village main square, individuals convicted of feuding in a prison near the North Albanian town Lezha, different male family representatives in isolated mountain villages and, in addition, village policemen and prison guards. Despite varying positions of power all frequently explained that one handles one's business according to kanun. It I For diseussions of eeologieal feuding theory and demographie eorrelations, cf. Otterbein (1993: 140f.); Helbling (in this volume).

2 There are no exaet figures. Offieial Albanian statements tend to lower, foreign journalists to raise them. 3

Taken from an interview with ProfIsmet. Elezi, 13 Sept. 1996.

4

In eonversation with Mare Clark 1997.

134

Stephanie Schwandner-Sievers

is this magic word kanun that is referred to today in the North, when people want to legitimise physical violence and killing and to claim self-regulation and the monopoly ofviolence in opposition to the state. Kanun is also used in peace negotiations and for discussions concerning customs and proper social behaviour. There are private reconciliation missions attempting to settle feuds by referring to kanun. Since 1995 the "national reconciliation mission" headed by Pjetr Ndreki from Shkodllr became the best known in the international media. Up to September 1996 his mission is said to have settled 188 blood feuds in the North and Northeast Albanian distriets. Yet, unofficial reconciliation missions organised by local Catholic priests or traditional tribai authorities were active as early as 1991.5 The current usage of kanun raises the question of whether the application is the same as was known and documented around the turn of the century when Albania was still under Ottoman rule (until 1912). In feuding today there is a proper distinction between "cases with memory", which would entail a "debt to give" or "obligation to take blood" for disputes of the grandfathers, and "cases without memory", which would be the newly emerged cases. 6 Officially, kanun and feuding was said to have been completely abolished in communist times. However, ifmajor elements of kanun really did survive the nearly fifty years of Albania's communist dictatorship numerous questions arise: How were they perpetuated in communist tim es? Which elements of kanun really persist today? What kind of a role did it play under communist rule and after, in the period of the so-called transition to democracy under president Berisha? Are there any causes wh ich explain the situation in early 1997, commonly referred to as anarchy? Can a retrospective orientation, aimed at local appeasement by referring to kanun, lead to a future without feuding? By approaching these questions from an anthropological point of view it seems that reference to Northern Albanian kanun today is an important discourse, yet referring to a mythic past. It is promoted and interpreted differently by specific interest groups to deal with present day conditions. The actively used reconciliation ritual, for example, is embedded in this retrospective discourse despite some newly invented features. However, historical as weil as contemporary proverbs and rituals related to kanun discourse, such as the reconciliation ritual, can be read and underscored through an analysis of their inherent symbolism. This will demonstrate the persistence of an underlying dualistic value orientation and clarify one of the many dominant ideologies which one can find in different realms oftoday's Albania. This is true not only for Northern Albania where these values are socially relevant and determine 5 In 1993, in Abate village, Dukagjin, we copied a list compiled by a reconciliation mission consisting of nine members who represented different fis ("tribes" or "lineages"). This group had successfully sett1ed nineteen feuds between August 1991 and August 1993 by referring to kanun ritual. 6

Distinctions found in the region of Miredita, 1992.

Humiliation and Reconciliation in Northem Albania

135

status politics, but a polarised ideology is also evident in the speech and actions at the highest levels of state govemment. However, it does not match many elite Albanians' Westem-oriented ideals and abstract categories of democracy, of independent institutions, pluralism, equal rights and the respect of any citizen's dignity. Instead, it works with ideas of absolute loyalty and treachery in kinship or personal relationships and the alienation of "the other." People are divided into friend and foe through employing "tradition" as a moral code that justifies categories of inclusion and exclusion, for example, in access to new resources. Critics are effectively excluded since part of this code is to allow doing harm to those not loyally bound to one's reference group. In a diachronic perspective it can be demonstrated that from Ottoman, through communist times and into the present day, the respective state organs have used and are using kanun concepts to maintain control, thereby generating these dualist (polar) conceptions in society which support regional and political clan structures, familism and patronage. Looking for latent tendencies a very specific North Albanian concept of reconciliation, loyalty and faith, called "besa" will be introduced. 7 According to Christopher Boehm, reconciliation is a central element of feud. He defines feud as "deliberately Iimited and carefully counted killing in revenge for a previous homicide, wh ich takes place between two groups on the basis of specific mies for killing, pacification, and compensation" (Boehm 1984: 194).8 In war or feud "taking the blood" of a strong enemy will increase the perpetrator's strength by substituting formerly "given blood" which is then a symbolic incorporation of foreign blood.9 The North Albanian "blood-peace", pajtimi te gjaqeve, substitutes these requests "for blood" ofthe living and ofthe deceased kin: By creating classificatory "blood" kinship it successfully interrupts the reciprocal killing. 1t is a powerful incorporation ritual as weil, forcefully guaranteed through besa, this concept ofhonour, protective bilateral obligations and loyalty between former hostile non-kin groups.

7 Besa (indef. bese) for a first working definition, implies as weil the oath for reconciliation, the reconciliation itself, as the holy obligation of the ritual, it guarantees and the new bounds created between former blood enemies, also truce, alliance. Bese-lidlye means besa-bound, besa - unity. "Besa" etymologically derives from "oath", Albanian be, beja (Cabej 1988: 206). 8 He investigated the neighboring Montenegrin people's moral conceptions in change: "For ... the tribesmen who Iived without any centralized govemmental power up to 1840 - blood feud was not merely acceptable and legitimate; for them it was a moral necessity that a man (or a clan) takes vengeance, if adecent social status was to be maintained." (Boehm 1984: 66) For his contemporary informants "the taking of vengeance was still a reasonable and eminently moral form of social action" (ibid.: 62; cf. Otterbein 1993: 136). 9 I am very grateful to Prof Peter Waldmann for inspiring these points in a comparative discussion in Siegen, May 1997.

136

Stephanie Schwandner-Sievers

Yet, feuding does not simply follow a predictable and Iimiting "blood algebra", evident in this often cited rule ofthe kanun: "a head for a head - blood for blood" (cf. Fox 1989: 172). Naturally, violent interactions involve strong sentiments wh ich in Northem Albania's past and present easily led to extended and unpredictable feuds. They oscillate between a family's self-estimation of status and integrity, the aim to be respected by the others and by the actual village public opinion. Therefore, statements Iike: "If they kill one of us, we'll kill ten of them," 10 are not simply intended to deter, but to display high status and offensive capacity. In short, these sentiments are part of the "dialectics of honour" (Bourdieu 1976) which are inherent in kanun symbolism and based on a particular belief in the role ofthe ancestors to protect and guard the family. The "dialectics of honour" dynamically determine social hierarchies (cf. Giordano 1994: 179f.). Independent of any position of advantage defined through senior lineage descent, the past and present "homo hierarchicus" (Dumont 1970) of the Northem Albanian mountains on the one hand seeks to demonstrate and consolidate high status in his every-day behaviour, wh ich means proving a family's honour through the men's capacity to defend, to kill, or to guarantee someone's protection (bese) at the right time. On the other hand, a lower status has to be assigned to the others. The active practice of humiliation is an effective instrument for this "task" although this is never openly admitted in kanun discourses. 11 lt is principally carried out by young men who still have to prove their "warrior" attitudes. The most effective humiliation and the strongest offen ce to honour imaginable is deliberately breaking someone's bese. According to an earlier documentation of kanun conceptions, a broken bese implied that "the stamp of dishonour remains in the family, generation after generation up to the seventh faqe ['face', here: 'generation', 'degree')" (cf. Gje~ov 1933: 110). Active humiliation is a dangerous undertaking, because it might evoke a violent reaction by the victim. He may try to invert status relations again if he does not accept a bottom place in village society. Perhaps these means to dynamically define status playa stronger role today than in pre-communist times, when status dispositions defined through descent had not yet been vigorously destroyed. The concepts of honour and humiliation, of altemating appeasement and violence, however, were reproduced by the ideology and actions ofthe highest communist officials.

10 Recorded in 1992 from a Miredita family; the grandfather, now deceased, had offended another family in pre-communist times. They now expected vengeance because the victims just recently had publicly announced this intention. Similar multiplicative statements were already reported by Hahn, I. (1854: 176). Regionally, a rule '''two fOT one' ... when a man killed his social superior" existed, proving social differentiation (Hasluck 1967: 392). 11 The focus on this point was inspired by SchreineriSchwerhofJ( 1995).

Humiliation and Reconciliation in Northem Albania

137

Historically, the Albanian ritual device of settling feuds, best!, became internally relevant on a macro-scale, when an external threat or a common political interest required strategie alliances to be formed. In the early 1990s, thousands of Kosovo Albanians, facing Serbian hegemonie aggression, united in mass reconciliations, called best!/idhje, initiated by students and priests and led by the ethnographer Anton Cetta. However, such "general best!" wh ich historically frequently governments and state representatives attempted to impose were usually looked upon with suspicion in the Northern face-to-face society. As will be shown later, in contrast to the assumptions in the idealist kanun discourses, they turned out to be more or less short-Iived. The fading of a state-imposed general best! would lead to increased internal feuding and to dis integration of the state. A broken' best! alliance between territorial or kin groups would result in either renewed bilateral feuds or intertribaI war. A single villager who broke a collective best! guaranteeing safety for particular spaces, time periods or occupations, like on a village road, in the hunting season, for the public field guard etc., was usually punished by communal expulsion of both hirn and his entire family (cf. Hasluck 1967: 388, 398; Fox 1989: 86 ff., 168 ff.). Violating a man's or his family's best! was considered to be a private affair between the involved "houses", i. e. families. Today's Northern Albanian promoters of a new kanun discourse, oriented in the past, ass urne that former customary law guaranteed a balance of killing and peace and the predictability of reactions to norm violations thereby insuring social control in the mountain villages. Yet, who are these promoters? In anthropology of law one universal form of institutionalised law is recognised as "abstract mIes that ... are found as a set of verbalised ideals in the repository of the minds of knowledgeable individuals in an illiterate society" or, in literate societies, codified in written form (Pospisil 1974: 19).12 Yet there are some problems concerning these kanun specialists, their knowledge, their legitimacy in Northern Albania today and also concerning a "modem" kanun that left the remote areas. 1. The kanun-Discourse

In North Albania reference to kanun varies by specific names given to it which are derived from mythical, medieval, prestigious tribai mlers and forefathers who are believed to be the founders ofthe customary law. There is the so-called "kanun of the mountains", or the "kanun of Skanderbeg" or the "kanun of Leke Dukagjin" which people proudly feet composes part of their identity. A kanun of Southwest Albanian Laberia, kanun of papa zhuli or of idriz sulli was actively referred to until the mid-fifties (Godin 1956: 184). However, there is no recent evidence of any aims for revitalisation or of a new 12

Pospisil paraphrasing LlewellyniHoebel (1941: 20).

138

Stephanie Schwandner-Sievers

kanun discourse in the Southern parts of Albania, except when scoffing at Northern "backwardness,,13 (cf. Schwandner 1996: 77ff.). In the Northern mountains the kanun customary law is believed to be passed on orally from generation to generation, codified in common proverbs and sayings. The most prominent early printed collection of North Albanian local proverbs was published in 1933 by the Franciscan Seminary of the Northem town Shkod~r (Skutari). [t had been collected by one of the patriotic missionaries, Padre Shtjefen Gje~ov (Gje~ov 1933).14 He arranged the proverbs in paragraphs and artic1es, creating the image of a fixed codification. However, regional variations persisted conceming the details of proper behaviour in feuding and the common guidelines conceming how to save, restore or prove honour and pride. 15 In the mountains, single copies of Gje~ov's printed kanun were hidden from the communist authorities. Today, a second inexpensive edition has been published by the re-established Franciscan Seminary of Shkod~r, and between 1993 and 1996 it was available in many book kiosks (Gje~ov 1933). This same publisher, the Franciscans, continues the patriotic ethnographic tradition. Recently they have published, for the first time, a co 1lection of the "kanun of Skanderbeg" which ethnographically covers an area slightly further South (Kruja, Mati) (Illia 1993). Of course, these printed sources might have a fundamental impact on the revitalisation of customary law. But field experience has proved that these texts are not referred to, and reliance on local community elders, who through personal, pre-communist experience are believed to still know the "old kanun", is preferred. Unfortunately, the information these elders have, after fifty years of communism, may be inaccurate or they themselves may already be senile. This was seen when a ritual blood-peace-making ceremony, which related to a single case of attempted homicide and conducted under the guidance of a community eIder, failed. Some participants later explained that this failure was partially a result of the fact that the old expert had "forgotten" that the kanun rule "blood for blood" implies that no reconciliation can be required unless the parties . involved feel that there has been a fair "balance of blood". This means that a minimum of one similar attack in revenge must have taken place before negotiations may start. 16 In this case, kanun became relevant as a tool to solve a

13 Von Godin found striking similarities comparing Northem and Southem kanun variants (Godin. ibid.). My research in South Albania (1994, 1995, 1996, conceming ethnicity), mostly in the urban area of Korea, simply showed that values like besa and maleness are explieitly glorified in some rural eontexts. 14 There is an English translation available whieh unfortunately does not provide many ofthe "colourful" metaphors used in Albanian sayings (Fox 1989).

15 Best variation eollection in Hasluck (1967). 16 6 Dee. 1992 in a new slum-like suburb of the Northem town of Shkoder where many young families settled who emigrated from the hard eonditions in the mountains.

Humiliation and Reconciliation in Northem Albania

139

situation of latent threat, It was, however, re-invented in vain because of contradicting interpretations. 2. Different Generations Contesting kanun It is particularly the older generation that remembers the "better" precommunist times, especially if they were members of a high status family and subsequently lost positions of inherited authorityl7 which were previously guaranteed by kanun laws. In 1991/92 everyone faced a "Iaw vacuum in the moment of political change, a general psychological crisis" (Elezi 1996). The revitalisation of a pre-communist mode of orientation became a rational choice while alternatives were lacking in the mountain peripheries. In this desperate situation, despite the fact that many would have preferred a functioning rule of state law, these former triballeaders and their descendants emerged as an active interest group though not recognised by central state politics. In communist Albania it had been a top priority to fight the "traditional patriarehaI family" (Alia 1989). Thus, these families had been among the most persecuted, condemned to poverty and humiliation. With political liberation, these people, not compromised through being former communists, realised the best opportunities in village struggles for power. They aimed to invert power relations, to renew their prestige and to claim back inherited private property rights by referring to kanun and a past social order (cf. PichIer 1995: 74ff.). Notably, it was in the Northern mountain villages where the egalitarian precepts of the land distribution reform of 1991 18 were not accepted and sometimes caused new contlicts. Albanian state officials and international consultants, involved in land privatisation programmes, understood that "the force of beliefs about the sanctity of private property are very strong, so that in some villages people who were excluded from the land distribution process because their families had no claim to land from before collectivisation often accepted their fate without complaint, since they believed they had no legitimate claim" and consequently, they emigrated (StanfieldlKukeli 1995: 3). If two parties did not come to a peaceful agreement, land disputes and competing systems of power and prestige became new reasons for new feuds. Sometimes, kanun discourse was used to peacefully settle questions about the borderlines of pastures. Elders like Ded Noja from Theti, for example, instructed their fellow-villagers about an old ritual to settle land disputes and to confirm circumferences: After negotiations, a stone is Reasons for the failed reconciliation ritual were explained differently by the c1erical ritual leader Emesto Troshani (see below). 11 Organized according to the principles ofpatrilinearity, seniority, and ability (in order of prevalence). 18 Law 7501 on Land of July 19, 1991, Albanian Ministry of Agriculture and Food (MOAF) conceming the division and distribution of agricultural land fonnerly organized into cooperatives ... and state fanns ... (StanfieldlKukeli 1995: I).

140

Stephanie Schwandner-Sievers

thrown to mark the border while swearing "by the mind [or: "flesh"] and spirit of my ancestors", an agreement supposed to be so holy that doubt about it would entail blood. 19 Presumably because of their social responsibilities, this kind of clear kanun advice attracts many family heads, who generally are in their forties and fifties, especially if a kanun reference offers a solution to new threats. It is always such a "lord of the house" of the offender's family who asks for the elders' assistance and longs for reconciliation in order to secure the physical integrity of his family. Then, as shown above, kanun and the elders become an important device. The Franciscan priest Troshani who had taken part in the failed ritual recounts the opposite attitude of the victim, a man around thirty, who had denied the offender forgiveness: It is this age-class "that does not respect the elders, especially ifthey left the mountains and their people because ofpoverty. In the villages, in contrast, there are never problems conducting blood peace successfully. This is because the elders, by their authority, guarantee the obedience of kanun and assure peace in the community." This statement seems quite idealistic taking into account that the elders lost any ultimate power in enforcing kanun rules. 20 Nevertheless, it is this age-group in the Northem villages which is apt to emphasise the kanun 's value as a collective moral authority. According to Emest Troshani, another reason for the decline of effectiveness in kanun practices can be seen in twenty five years of state imposed atheism (1967) which destroyed the recognition of the sacral components of kanun formerly guaranteeing respect. 21 Elezi stressed that, "it is these forty- and fiftyyear-olds [brought up in these rigorous times] who know very little of kanun. They heard it as a word, but they don't know it and they don't respect it.,,22 However, it seems that many of this middle generation are apparently taking part in the new kanun discourse. They might have difficulties constructing the discourse, since the rituals were not practised and therefore not sustained since the Fifties, yet, in case of emergency, they still hope for its suggestive power. Quite often these new threatening situations seem to be created by adolescent males and young men. Contesting status they display aggressive male 19 BoeklSchwandner (1994), a film documentation; for historical similarity cf. Gjec;ov (1933: 67).

20 Mare Clark is at present collecting material on the decline of kanun reference in the pre-communist period in the region. Edith Durharn has already described how locals expressed the desire for a strong and just government, cf. for example, Durharn (1909: 55).

21 Troshani was interviewed the day after the failed reconciliation ritual in the town of Shkodl!r, December 1992. Obviously, there is a nostalgia for a "pure" kanun by any priest in Gje'Yov's tradition whose kanun guaranteed also respect for church representatives. Additionally, it should be stressed that the influence of TV accounts for new orientations. 22 In an oral interview, 1996, as mentioned in footnote 3.

Humiliation and Reconciliation in Northem Albania

141

behaviour and easily overreaet to anything eonsidered an insult. Elezi eonfirmed that in regard to the last pre-eommunist eases, the elders "were the most powerful supporters, but the aetual killing in feud was pushed by young men aiming to be less looked down on" (Elezi 1983: 204).23 Robbery was doeumented as a main proving ground for young "warriors" in history (Mjedia 1901: 357; cf. Fox 1989: 158f.). This form of violence was channelled aeeording to lines of the kanun proverb: "the wolf lieks his own tlesh, but eats that of others" (Fox 1989: 122). Edith Durharn, an early travelling anthropologist of this area, eonfirmed that " ... robbery from another tri be with which no besa exists [was] rather a virtue than a erime" (Durharn 1909: 204). On the other hand, somebody who dared to " ... steal from the house or the knapsack of the mik (friend) [was] considered 'black faeed' [i. e. 'dishonoured'] by kanun, and he was excIuded from taking part in the men 's assembly (Gje~ov 1933: 136). By the same token, violenee imposed upon foreigners is still eonceived of as absolutely legitimate, if the strangers have not been ritually incorporated into the kinship-circIe, that is to say recognised as a "friend". In 1996 a rural North Albanian Ioeal politician 24 wondered about the official reasons for Socialist leader Fatos Nano's imprisonment (now prime minister): "Why is he still imprisoned? He did not embezzle our money but that of the Italians!" We also leamed from local policemen that they would never interfere with a local dispute and feuding ifthey themselves were from the same area (Bock! Sehwandner 1994). An AIbanian acquaintanee travelling by lorry to the North for some business was hijacked by a band of armed young men. Fearful for his Iife he mentioned the name of an intluential Northem Albanian friend. His shoeked assailants asked hirn why he did not say before that he was a man of honour and therefore protected by "bese". They then invited hirn to a spontaneous feast and slaughtered a sheep in his honour. 25 However, it may weil have been less the fulfilment oftraditional ideals than new wealth and money which made the common friend to whom both parties referred so impressive to this predatory group. Today, in contexts beyond the villages, new institutions with "modemised" codes ofhonour have developed. 26 There are gangs ofyoung men, negleeted by state education and without officiaI employment, that are primarily involved in different new types of violence. Some engage in trading aetivities involving 23 Unfortunately Elezi differentiates only between men more than fifty years old and less than fifty. 24 25

Interviewed while doing pre-election observations for the OSCE. Recorded 1997.

26 For the point of a "modemised knnun" processing a possible "brutalization" I am especially grateful to Prof Giordano, Universite de Fribourg, Switzerland, for stimulating discussions in February 1996, especially regarding cases of Albanian feud in Switzerland and Germany.

142

Stephanie Schwandner-Sievers

drugs, weapons and women in the areas bordering Albania, Montenegro, Kosovo and Italy.27 The discourse of kanun is available in these new contexts as an excuse for murder (cf. NZZ 1995), but for obvious reasons, research is hard to conduct in these cases.

3. Status Politics, Exclusion, and Inversion Those classified as outsiders against whom violence might be legitimately committed, do not only exist outside the village. The "dialectics ofhonour", by defining social hierarchies inside the village, analogously determine the insiders and outsiders of the community. People who in the public opinion are considered dishonoured are socially "dead" by kanun (cf. Fox 1989: 130) and seem to become open targets for public mockery and humiliation. 28 For example: "[A] man slow to kill his enemy was thought 'disgraced' and was described as 'Iow class' and 'bad'. Among the Highlanders he risked finding that other men had contemptuously come to sleep with his wife, his daughter could not marry into a 'good' family and his son must marry a 'bad' girl. ... [Further South] he paid visits at his perit; his coffee cup wasonly half-filled, and before being handed to hirn it was passed under the host's left arm, or even his 1eft le§ to remind hirn of his disgrace. He was often mocked openly" (Hasluck 1967: 391).2 Consequently, proving high status means to acquire social capital ("honour") which guarantees access to various resources like advantageous arranged marriages or power in village politics that are denied to people of low status (cf. Schneider 1971). In 1993 in the Dukagjin region we were told about persisting symbols of low status: a dishonoured family was said to be required to cover their bowed heads on their way to church. This open display of lost honour forces the victims to attempt a status inversion, because "by his successful crime the murderer had proved himselfthe better man, and the victim could not endure this inferiority and was bound to strike back in kind" (Hasluck 1967: 392). There are two options in kanun discourse used to invert a low status into new reputation after a murder. The frrst is to ritually forgive "if you see fit" which usually can only be done if some time has past after the last murder or if it has been recognised that the guilt was partly shared by the victim. Forgiving must imply the opportunity for the victim's family to compensate the honour 27 Smuggling weapons was perfectly compatible with early kanun as weil, cf. Sie-

bertz (1910: 94-98).

28 An impression MirMita 1992: stones were thrown in someone's windows. These dehumanising practices make it so puzzling if people, Iike the former Wehrmacht officer Peinsipp, glorify the mountaineers' pride and honor. (Peinsipp 1985). 29 Cf. Fox (1989: 136); Durharn describes the symbolic pressure of the social action when raki was only passed behind the back to the "outcast" (Durham 1909: 41).

Hurniliation and Reconciliation in Northern Albania

143

(or integrity) deficit by the demonstration of extreme generosity, provided in a ritual "play with time" (cf. Bourdieu 1976: 32) and the ritual humiliation of the originaloffender in order to equalise status. Historically, these demands were met in all reconciliation ritual variations which, for example, forced the guilty petitioner to kneel down for hours with a heavy ritle on his neck (Hahn 1854: 177; Hasluck 1967: 404ff.). Yet, modem priests forget about this symbolism when today they ask the victim to humbly kneel down while swearing vengeance on a Christian Bible and cross. 30 The successful reconciliation, i. e. an inversion of foes to classificatory "friends", kanun promoters seek to confirm by pushing the former opponents to form a blood brotherhood (Bock! Schwandner 1994), thereby indicating equalised status and community integrity. The second option leaves no doubt that the original victim has "washed his blackened face" (Gje~ov 1933: 101, 111) as it is still commonly expressed. A kanun proverb gives the formula of how to clean a face, that is, according to persisting perceptions, "blackened": "The soap of a man is the gun-powder" (cf. Cozzi 1910: 664). According to idealising kanun discourses recorded in Miredita 1992 the violation of a family's bese leaves no choice for reaction other than the taking of blood in contrast to the two options after a murder. Since it is considered the strongest offence imaginable to a family's honour imaginable it counts more than the ass'assination of one of its members (cf. Fox 1989: 13 Off.). To explain bese, an informant referred to the following proverb: "Who gives his bese will even slaughter his son for keeping it.,,31 4. "Besi!" or the Honour of Protecting a "Friend"

The Albanian term for a "friend", mik, derives from Latin amicus. An old and still known proverb attributed to kanun says that: "An offence against the father, the brother, and up to a cousin without heirs you may forgive, but you cannot forgive [any harm done to] ... your mik" (Gje~ov 1933: 116; cf. Fox 1989: 136). It is this artificially constructed, non-kin relationship that must be protected and secured. The term "mik" stands for all the kinds of persons who are related to the kin group in a specific way. That is, as former strangers and therefore potential enemies, they were incorporated into the kin group by ritually cleansing them from the dangerous outside world. This was partly accomplished through certain types of humiliation of the initiate. All these rituals of incorporating the mik if organised according to kanun still follow a left to right, West to East (sunset and sunrise) directed motion, which can be interpreted as a symbolic inversion of the chaotic and foreign to the ordered, known world

30 Observed at the failed reconciliation ritual in the town of Shkoder, 1992, cited earlier.

31 According to the head of an extended farnily interviewed in his horne in the rnountains ofthe region Fan, Miredita, 1992.

144

Stephanie Schwandner-Sievers

reflecting the social status inversion. 32 So we find a whole set ofpersons designated as mik (pI. miqe): al1 agnate in-laws, a former enemy of blood feud after reconciliation, and a guest. The designees al1 belong to one category ofpersons: actual1y they are incorporated al1ies and classificatory kin. The word in Albanian for the specific relationship to a person cal1ed mik, which includes al1 expectations and elements assigned to this specific, institutionalised, non-kin relationship is bese. It is a "total social phenomenon" covering political, legal, economic, religious, moral and social aspects. 33 It does not translate easily into only one foreign word. Translations range, according to the context and the segmentary level ofrepresentation ofthe uniting partners,34 from "peace treaty", "alliance", "truce", "blood-peace", "reconciliation", "reconciliation guarantee", to "hospitality", "honour of the house", "safetyguarantee", and also "oath", "word of honour" or the "given word", "loyalty" and "trust". In kanun discourses, the violation of a bese serves again as an explanation for initiating a feud and for its extension to other families. It always involves more than one party: if a guest is killed, for example, his blood relatives, as wel1 as the host, will have to claim blood. If a woman is abused or killed, her agnates as weil as her husband seek vengeance. Violations of a family's bese include treachery, insult or physical violation of one's guest or wife, both considered to be under the protection of the house; public questioning of one's capacity to keep the given word or defend one's honour; and of a village's bese violation of an agreed upon limited or temporary truce or alliance. A man who meets al1 requirements of the highest, ideal-type values is cal1ed a besnik: he is honest and honourable, brave, reliable, faithful and strong. Besnik also stands for maleness, a besnik is synomynously a "hero" (trim) who might look at anyone with a "white face". Pabese means the opposite, a man without bese who is treacherous, unfaithful, cowardly, weak and "black-faced" (Cf. Lambertz 1958: 136).

32 In Schwandner (1996: 88) some of these North A1banian rites de passage, Iike marriage, guest ritual and reconciliation are described as observed in 1992 and 1993. 33 Davis, referring to Emile Durkheim and Mareel Mauss, proposed assigning the eollective oath in Albana the status of a "Mediterranean soeial fact" (Davis 1977: 13).

34 This ean be understood in eoneentrie circ\es aeeording to a patrilinear segmentary soeiety organization, for example: the lord ofthe house and the house's guest, the elders ofthe brotherhood or village and the village guest (the priest, or others), e1ders of different tribes or valleys agreeing on an alliance, ete. Regionally, kinship and/or territorial eriteria are either synonymous or weighted to varying extents.

Humiliation and Reconciliation in Northem Albania

145

5. "Black-Face" and "White-Face" - The Logic of Symbolic Coercion The symbolic colours black and white indicate social status in Northern AIbanian every-day language and provide a value orientation in order for violence to be used legitimately. To explain the underlying meaning ofsymbolism ofthe white ("honour") and the black face ("dishonour") it would help to look at an old North Albanian ancestral or guardian spirit, a polymorphous being who is still unconsciously referred to in many common sayings as weil as in old mountain legends. This being once served to represent every segment of the patrilinear constructed society in the sacral space. Totemistically, there was one such spirit for every individual, for the house (or the extended family) and the clan. These spirits were believed to hold assemblies like actual people do (cf. Georgescu 1963: 81; Kaser 1993: 109; Lambertz 1922: 35). They were called "are" which relates to the ancient Greek guardian spirits or goddesses hore. These terms are still used in Albanian, as weil as in ancient Greek, as synonyms for "time" or "fate". In ancient and recent Balkan folk beliefs, three such spirits predict the fate of a new born child on the third day after birth, also they gather in assemblies to discuss human fates. The Albanian fairy tales and legends prove abilateral responsibility of protection and fate, a bese between man and are. The are of a man who behaves like a besnik will be beautiful, strong and white, called "white-ore ". But somebody who fails to meet the ideal requirements of behaviour, a pabese, will have an inert, black are, called "black-ore" (cf. Lambertz 1922: 33f.). Since an ore's behaviour might be fundamentally disastrous, it was also called by a taboo name in order not to awaken its attention unnecessarily. This was, respectively, either "black-face" or "white-face" (ibid., cf.