Why do people so often display such intense feelings for other people, faraway strangers whom they have never met and most likely never will encounter? These strong affections at a distance may well be a quite recent phenomenon, the emotional complement in individual experience of major social transformations.
Identifications and Disidentifications
In a 'companion piece' to this article I argued that identification is a cognitive and emotional process in which people increasingly come to experience others as similar to themselves (de Swaan, 1995).
The earliest biogenetic and sociogenetic context of identification was based on kinship in survival bands . Only with the spread of sedentary agriculture did a second sociogenetic matrix of identification emerge in the context of the peasant village: proximity. These two, 'blood' and soil', for millennia remained the main strands of identification, and to this day they provide the mobilizing metaphors for invoking sentiments of likeness on a much vaster scale: 'children of one father', 'brothers in arms,' the 'fatherland' or the 'mothertongue', and 'love thy neighbor as thyself'. Intensely felt identifications beyond the pale of family and village remained quite rare until well in the modern era of nation-building, the emergence of class consciousness and race ideologies.
There may have been early exceptions, such as the solidarity that linked quite distant members of dynastic and aristocratic networks, or among dispersed monastic orders or in the ranks of large armies. All along the believers were exhorted to identify with their distant brothers and sisters in the faith, just as the inhabitants of the realm were reminded of their duty to be loyal to their ruler and his many, widely spread subjects. All these long distance identifications were couched in the metaphors of kinship an proximity, but for the vast majority of the common people they carried only incidental and limited emotional meaning.
Identification is the emotional complement of group formation. It entails the affective realization that others are similar to oneself, and belong to one's own group, and that still other people are different, do not belong, and must therefore be excluded. Negative tendencies, i.e. sexual, aggressive or self-serving inclinations that are denied in oneself and one's peers are ascribed to these outsiders. This combination of denial and ascription is called 'projection' in psychoanalytic theory. In a more elaborate version, it is assumed that the projected characteristics are then vicariously experienced by those who previously had projected them unto the others, in a process called 'projective identification.' The present term 'disidentification' refers to this process of denial, projection and, as the case may be, vicarious experience .
In this paper, I concentrate on the latter process, the process of social exclusion and of emotional disidentification with its accompanying affect: hatred . The discussion will focus on the developments leading up to the genocide of 1994 in Rwanda. In a concluding section the social-psychological account will be placed in its political context of state disintegration and external military threats.
Identification and disidentification are not each other's opposites, rather they occupy two sides of an emotional triangle, with at its base ignorance and indifference. The vast majority of human beings are completely unknown to one another. Throughout history, most people were unaware of the very existence of the greater part of their fellow human beings, and only in the past few centuries this has begun to change at an accelerating pace. But even if a vague awareness of the presence of other people in distant lands prevails, indifference reigns, sometimes coupled with diffuse fantasies not very strongly held. Thus, for people to even begin to have any feelings at all about distant strangers they must first find out about their existence, and, if next they are to hate them, they must first be told about their hateful characteristics.
All this would be a rather vacuous and sterile exercise if there were no actual interdependencies between the one group and the other. The abject habits of the Ludimango, the ways in which the men exploit the women, their cannibalism, the roasting alive of deer and fowl, the teasing to death of eels, cats and bulls, the torturing of adolescent boys, the excision of young girls, the burning of widows, none of it will much excite other groups as long as they have no business with the Ludimango. It is the transformation of social relations that brings with it the transformation of sentiment, in a double movement of identification and disidentification that supersedes prior unawareness and indifference.
The increase in scope of these emotional concerns corresponds with the increase in scale of social relations, that is relations of conquest, conversion, trade, and rule but above all, with the extension of conflict. Conflict is an ambivalent game. It entails similarity of strivings and difference of interest. When European merchant slavers and their African raiding accomplices hunted down, deported, sold and exploited tens of millions of Africans, they did so precisely because they needed them not as raw material, or as fodder, but as workers, as human laborers, as sentient, competent human beings who would understand orders, anticipate reward and punishment, master the demanding crafts of plantation agriculture and the subtle skills of domestic service.
In other words, the slave traders and the owners at once had to identify with their victims as similarly equipped human beings, and to disidentify from them as beings with similar sentiments, of moral value, with a human soul. Slavery was a concomitant of the rise of the modern world system, i.e. the emergence of the great Atlantic trading triangle between Europe, Africa and the Americas. To transform Africans into slaves required massive social and emotional work. They had to be recognized as human beings and to be excluded as 'cattle', as 'apes' (still the breed closest to man), as quasi-animals, as sub-human beings. At the very best they were considered 'childlike', that is 'not yet' completely human, but undeniably on their way to full humanity under the tutelage of stern but benevolent guardians.
The theme of slavery is raised here in passing only, to demonstrate the two double movements that constitute the dynamics of the widening scale of emotional concerns: the increase in scale of social figurations, that is the emergence of the triangular Atlantic trade, that goes with the emergence of sentiments of widening scope; and, the twofold socio-emotional process of identification and disidentification that transcends unawareness and indifference. Plantation slavery was in a sense 'the original sin of modernity': one great training ground for mass deportation, massive exploitation and extreme deprivation, increasingly legitimated with racist theories. It provided the first instance of the socio-emotional work of identification and disidentification on a transcontinental scale.
The Extending Scope of Political Authority
The empirical referent of this study is the preparation for genocide in the Rwanda of the early nineties. Initially, Western opinion perceived the mass killings in Rwanda as a spontaneous and catastrophic outburst of long simmering 'tribal hatred.' There is now ample documentary evidence that the mass extermination of Rwandan citizens was the culmination of a carefully prepared, well-organized, bureaucratic campaign, using modern means of mass communication, propaganda, civil administration and military logistics.
The categorization of Rwandan (and Burundi) inhabitants into 'Tuutsi' and 'Hutu' underwent a series of changes in the course of the past century, and entailed a multiplicity of overlapping meanings that are almost impossible to disentangle, variable over time and from one community to another. The encompassing social context was one of political and economic transformations towards a larger scale of interdependence. The last point is discussed first.
It is true that long before the European conquest by German invaders of the lands that are now Rwanda and Burundi, quite stable and intricate political systems prevailed in the region of the Great Lakes. Thus there existed 'Kingdoms' that exerted some kind of authority over large parts of the region (Sagan, 1985: 3-58; Lemarchand, 1996 ; Newbury, 1988; Prunier, 1995). However, these so-called central Kingdoms did not exert much impact on daily life in the villages and held little sway in the minds of their subjects. Bäck (1981: 30) concludes:
Rwanda was represented by most earlier authors as a ‘sacred kingdom’ or as an ‘absolute monarchy’. In the light of more recent empirical research it appears that the role of the umwaami and the court directly associated with him may have been more limited even during the nineteenth and twentieth century when the monarchy succeeded in increasing its influence considerably...
And Trouwborst (1991: 99) writes: 'For all these reasons it is difficult to maintain without qualifications that the political economy of the intralacustrine states had a centralized character and could be considered as a centrally administered whole.' The author continues: 'My conclusion therefore is that as regards expenditures also, redistribution had a very decentralized character. It was true that the king stood at the center of the state, but this was true basically only in a symbolic sense.' (idem: 102). Trouwborst (1991: 105) concludes, with some qualifications: 'A political economy in the sense of a public administration of the interests of the state and its subjects scarcely existed.' Clearly, in the past century and a half, the region of the Great Lakes went through a major social transformation, much accelerated during the colonial and postcolonial era, from essentially segmented and decentralized polities and economies to the much more centralized states and markets of contemporary Uganda, Rwanda and Burundi. But this process, far from being exceptional on the African continent, was the normal process of state formation and economic development there (Cf. Bayart, 1989).
Thus, on its own, the increase of scale in the social organization of Rwanda can not serve to explain the extraordinary (although not unique) events that occurred there. It constitutes one, necessary, condition for these developments. The social transformation, as the present argument goes, forms the context for a transformation of emotional concerns . The next step requires a discussion of the evolving meanings of the major opposition that set the Rwandese against each other: the conflict between the 'Tuutsi' and the 'Hutu'. Here, only a rather schematic review can be presented of what remains a baffling complexity of evolving and cross-cutting social relations (Cf., however, Newbury, 1988 and Malkki, 1995).
The Outsiders' View - the Vicissitudes of Scholarly Opinion
Contemporary scholars are unanimous in rejecting earlier interpretations of the conceptual pair 'Tuutsi'/'Hutu' as referring to fixed racial categories. But paradoxically, this is what the terms have come to imply in contemporary political discourse, influenced as it was by earlier scholarly writings.
This may well be one of the most ironic and tragic examples of a situation being realized as a consequence of the definitions and expectations of those that are part of it . Early missionary and ethnographic authors chose as their informants almost exclusively court aristocrats who identified themselves as 'Tuutsi' and suggested that their peers had always been in command as a hereditary ruling group. This most likely was an ex post fiction of the sort that established oligarchies are wont to provide (and to believe themselves after one or two generations).
The early German anthropologists identified this hereditary ruling class with certain physical characteristics, a 'somatic normtype' (Hoetink, 1962), such as a tall frame, a high forehead, narrow, elongated hands and feet, a long and thin nose, in comparison with the 'Bantu's' of the 'Hutu' peasant class with a flat, broad nose, thick lips and a short and stocky build. The members of the ruling group shared the image that the foreign experts presented of them.
Contemporary 'Tuutsi' reject the binary, hereditary division between 'Tuutsi' and 'Hutu', preferring to use political and socio-economic categories, but this may well be a 'public stance' . And their very denial of inborn differences between the two groups is what most enraged the ideologues of 'Hutu-power.' These images are reminiscent of the racialist theories about the French aristocracy, which was said to be of 'Germanic' stock, tall, blue-eyed and blonde, as contrasted to the common people of France who were portrayed as swarthy, short and stocky and of Celtic origin, according to the racist theorists of the day, foremost among them de Gobineau (1853). The German anthropologists speculated that the original population of the Lake kingdoms (Uganda, Rwanda and Burundi) were of Bantu or 'Negroid' origin) while the 'Tuutsi' were assigned to the 'Ethiopid' or 'Nilotic' stock that descended from 'Hamitic' or 'Semitic' roots .
The Belgians adopted the German scholarly view and carried it into administrative practice. In accordance with their policy of 'local self-rule', they favored indigenous officials as local administrators, many of them recruited from the court aristocracy that was identified with the 'Tuutsi' (even though a considerable proportion was considered to be of 'Hutu' origin). From 1926 on they put in place a municipal civil registry containing filecards of all citizens with a photograph and the mention of their ethnic affiliation . Contemporary scholars have completely refuted the categorizations of their predecessors. First of all, no evidence has been found so far for an invasion by 'Tuutsi' pastoralists and for their settlement among the Bantu peoples of the Great Lake region. Second, given the present, almost completely identical linguistic, religious, and cultural practices and a distribution of physical traits that does not correspond with the current categories, scholars tend to the opinion that 'Tuutsi' and 'Hutu' must have belonged to one and the same cultural and endogamic entity for many centuries, and that they may well belong to the same genetic grouping . Lucien Bäck's (1981: 17) summary of current expert views on the Rwandese deserves to be quoted at some length:
"The population is generally divided into three distinct groups: Hutu (accounting for 82.7 percent of the total population in 1956), Tutsi (16.6 percent) and Twa (0.7 percent)... The spatial distribution is unequal. In the north and the northwest of the country the overwhelming majority is Hutu. Elsewhere, the Tuutsi and most Twa live interspersed with the Hutu. Despite peculiarities pertaining to the regions and the groups, all Rwandese can be said to share a common culture. This is particularly confirmed by the fact that they all speak a common language with minor local variations, ikinyarwanda, which is closely related to ikirundi, the idiom spoken in neighboring Burundi. The three groups have been said to differ in their physical traits. The Hutu are generally described as 'Bantu', the Tuutsi as 'Ethiopids' and the Twa as 'Pygmoids'... However, anthropometric data cannot systematically confirm this description... The division into Hutu, Tuutsi and Twa thus seems to be an essentially social reality" .
Bäck (1981:18) concludes his summary of recent findings on the history of the Rwandese population as follows: '...the fact that Hutu and Tuutsi speak one language and share a common culture suggest that they must have lived together for much longer than merely a few centuries, if they are not actually of common stock.' And as late as 1988, the ethnic divide was by no means the single and universal cleavage in rural Rwanda. Even in 'mixed company', peasants allowed themselves ethnic jokes, 'inoffensives plaisanteries', as village politics pitched the locals against outsider officials, rather than one ethnie against another. But this was soon to change under the impact of ethnic conflagrations in neighboring Burundi and propaganda in the Rwandese media .
An Interlude on Identification of the Others
The discussion of exterior differences between social categories is - rightly - considered distasteful in contemporary polite society. But this is not an occasion for polite discourse. Emotional identification with and affective disidentification from other people presuppose first of all identification of the others. The affective and moral categories require a prior cognitive, or pseudo-cognitive, construction. In Rwanda, up to the present, this occurred in everyday interaction on the basis of reputation, experience, intuition and impressions. From the 1920s on, this informal practice was complemented with administrative techniques for identification: civil registries, identity papers, passports, photo's and so forth. But this did not foreclose public discussion, forever questioning who were 'Tuutsi', who were 'Hutu', and how to tell the difference between them. On the contrary, as we shall see, precisely the most fanatic proponents of a hereditary division between 'Tuutsi' and 'Hutu' are obsessed with the possibility of 'Tuutsi' posing as 'Hutu' in order to confuse and divide them.
Apparently, even today, even for people whose life vocation consists of applying the distinction, it is very difficult to decide by looks alone. In stead, they rely on identification cards, which are based on the civil registry, and at the same time they deplore that these have been tampered with all along. Now, all divisions that have an effect in society, even those that closely correspond to readily observable somatic distinctions, and that may well be based on genetic differences, are 'essentially social.'
And on the other hand, there may be clear, physical characteristics that are 'inherited' from one generation to another and that do not necessarily have any genetic basis at all: body length is a good example, as it is closely correlated to superior nurture. Until very recently, and in many regions of the world, adequate nutrition was a privilege of the rich, who passed on their fortune and their height to their offspring. A clear skin, a firm glance, an upright posture, a resounding voice, a vigorous stride, in brief, an entire 'habitus' that seems to completely characterize physical appearance, and especially to mark the contrast between the appearances of the mighty and the lowly, may be passed down from parents to children and yet lack all genetic foundation . Thus, there may be significant, inherited differences in physical appearance between social groups, that do not originate in different genetic stock, but in socially inherited differences in wealth, prestige and power.
Through the social process of sexual selection, that is by selective mating, these social differences may in the end cause genetic divergence between the various endogamic groups . Finally, a single somatic norm type may very well be absent, and nevertheless the social group may still be recognizable, because of 'family likeness,' both in the literal and the Wittgensteinian sense (cf. Blok, 1975). There may even be a series of mutually quite different somatic types, which each are considered characteristic for the group, each of them representing a node in the network of family likenesses and possibly originating in a specific intermarrying network. The latter seems so be the case with European Jews: they do not look alike at all, and yet some do strike the interested observer as 'very Jewish', that is as displaying a family likeness to one of the dozen or so of 'Jewish types' he has memorized in the form of a Gestalt . Thus it may very well be the case that some 'Tuutsi' look quite 'Tuutsi', which means at the same time, unlike 'Hutu', without this implying a previously different genetic stock. There may well be several quite distinct 'Tuutsi' types. Every 'Tuutsi' type would be the result of socially inherited differences in nurture and socialization, reinforced by selective inter-marriage, with minor but visible genetic differentiation as the end result.
This would explain that Rwandese can sometimes successfully identify 'Tuutsi' and therefore also 'Hutu', but that often they can not decide or make mistaken assignments. It would also explain a major paradox in the 'Hutu power' propaganda: on the one hand 'Tuutsi' are said to be very different from 'Hutu', also in appearance, for reasons of genetic inheritance, but on the other hand they are accused of forever trying to pass as 'Hutu' for sinister reasons of their own and succeeding quite well at it. Since this is such a central theme in the socio-emotional work of 'Hutu' propaganda, it seems necessary to take a carefully reasoned agnostic stand on the issue, typographically rendered here by the consistent use of quotation marks around the terms of the conceptual pair 'Tuutsi'/'Hutu'.
'Hutu's' and 'Tuutsi's': from personal ties to general stereotypes
But what did the terms refer to initially, before they were absorbed in the European discourse of physical anthropology and then re-introduced in the Interlacustrian political vocabulary? These early connotations are very difficult to reconstruct, as there are hardly any written sources from precolonial times that could provide some idea of what the terms onnce meant. The few remaining informants who lived through the early colonial period and still remember the older meanings of the terms, have unwittingly absorbed the subsequent connotations that became current under Belgian and postcolonial rule.
Nevertheless some cues to the persisting multiple meanings of the conceptual pair exist, thanks to Liisa Malkki's superb report of her research among 'Hutu' refugees in Tanzania in 1986 (Malkki, 1995)  and Catharine Newbury painstaking interviews with elderly inhabitants of the Kinyaga region (Newbury, 1988). Without doing full justice to Malkki's subtle account, it is possible to distinguish at least three layers of meaning. The first takes the existence of 'Tuutsi' and 'Hutu' as hereditarily distinct categories for granted:
''It is evident that the Tutsi appear in the mythico-history first of all as foreigners - historically recent arrivals ‘from the North’, ‘from Somalia’, or ‘from the Nile’. It was often claimed in this connection that the Tutsi were really ‘Hamites’, who did not belong in the land of the ‘Bantu’ Hutu. (Malkki, 1995: 68).''
The 'Tuutsi' invaders conquered the land that once had belonged to the 'Hutu'. But they accomplished this appropriation not simply by violent means. And here begins a second layer of meanings of the 'Tuutsi'/'Hutu' pair: They, the 'Tuutsi,' did so by trickery: '...the Tutsi, possessed of "innate cleverness" in the art of deception, tricked the original inhabitants of Burundi into servitude by the gift of cows.'(Ibidem). Very roughly, the term here refers to a clientele relation or debt bondage between a cattle-owning pastoralist who lends some of his cows to an agriculturalist who pays for it with part of the produce and with labor services. The relation seems to be embedded in a network of political authority that turns the economic exchange into a lasting bond of servitude.
"..before the Tutsi came, the Hutu were not Hutu at all; they were simply abantu which ... signifies in Kirundi ‘the Bantu peoples’ or simply, ‘human beings’. …. The name Hutu, the refugees said, was imported by the Tutsi from their home in the north and means ‘slave’ or ‘servant’. Thus... ‘we became their slaves." (Malkki, 1995: 71).
That is what the Burundian 'Hutu' in Tanzania made of it for Malkki, and there is considerable historical evidence to corroborate the prevalence of such unequal relations of exchange and deference enforced by political authority and military ascendance, also in Rwanda in precolonial and colonial times. Newbury points out that in Kinyaga, in south-west Rwanda, clientelage initially imposed collective obligations on the lineage, but these umuheto ties were gradually replaced by ubuhake relations that involved individual obligations and claims. These bounds became increasingly oppressive and exploitative under colonial rule (Newbury, 1988: 115-140). Apart from an 'ethnic' or 'cultural' connotation, at least in Burundi, the term 'Hutu' had and still has an additonal, overlapping meaning:
''In the latter sense, Hutu refers to a ‘social subordinate’ in relation to someone higher up the pecking order...’social son’ is perhaps even more accurate, since it denotes not just social inferiority but a measure of affectivity... Thus a Tutsi cast in the role of client vis-à-vis a wealthier patron would be referred to as ‘Hutu’, even though his cultural identity remained Tutsi. (Lemarchand, 1996: 10)'' .
At this level, the 'Tuutsi'/'Hutu' pair denotes a specific, contextualized, political and economic relation, somewhat like, say a landlord and a tenant, a creditor and a debtor, or a master and a servant. The exchange occurred between a pastoralist and a peasant, and it entailed the use of cows against the rendering of labor services. The notion of a specific, local, even if unequal and resented, face to face relationship between two persons was transformed into the concept of a generalized and decontextualized relation between two timeless, irreconcilably hostile categories .
This was accomplished by introducing the 'mythico-history' of the conquest by the 'alien Tuutsi's' and their subjection of the 'indigenous Hutu's'. This layer of meaning, while referring to a distant past, was used to transform the second layer of meaning, that refers to a much more recent and actually remembered past, into a third layer of meaning that should serve to re-interpret present experience: in this latest meaning all 'Tuutsi' forefathers had been conquerors and exploiters of all 'Hutu' predecessors, and if some of them were not, they could not really have been 'Tuutsi'. All present 'Tuutsi' are out to regain their political and economic predominance by whatever means possible and if they are not, they can't be 'Tuutsi'. This must necessarily be so, it is not contingent on the former relations among 'Tuutsi' and 'Hutu', it is not even a consequence of past invasion and oppression, it is the inexorable outcome of the 'Tuutsi' essenc .
In this most recent layer of meaning, all 'Tuutsi' share common characteristics, both in their outward appearance and in their inner psyche. Their most important trait is lust for power, which determines all their strivings. But surprisingly, some 'Tuutsi' characteristics make them superior to the 'Hutu', according to the same 'Hutu-power' ideology: 'Tuutsi' are more intelligent than 'Hutu'; they are more loyal to their own kind; and 'Tuutsi' women are more attractive. However, these properties are only exploited in the service of the 'Tuutsi' thirst for power. Therefore, 'Tuutsi' use their intelligence only to deceive the 'Hutu', they are 'devious', or 'sly' . And it was not so much with naked power that they subjected the 'Hutu', but with deceit that they stole their land. (Malkki, 1995: 68-73). The 'Hutu' describe themselves as 'simple', 'frugal' and 'honest' people, but also as 'primitive', 'naive' and 'uncivilized' .The mutual loyalty of the 'Tuutsi' only serves them in their open and covert power struggle. Equally, the beauty of their women is exploited by luring 'Hutu' men into marrying them, so that sooner or later they can be made to spy upon their husbands and betray them (Malkki, 1995: 82-8) .
The Mobilization of Fantasy
Malkki's 'Hutu' respondents from Burundi in the Tanzanian refugee camps of the mid-eighties may have been 'naive', dilettante ideologues, but almost word for word their themes were raised again by the professional reporters of the 'Hutu power' review Kangura and their colleagues of Radio Mille Collines in the Rwanda of the early nineties. The agitators were by no means alienated from the common people, but voiced notions that had circulated widely among the 'Hutu' for many years. From the first issues on, Kangura reprocessed the familiar themes of the 'Tuutsi' as devious manipulators, of the 'Tuutsi' women as treacherous seductresses, all in the service of the conquest of power. But, in addition, in these texts reverberates the repetitive drone, the insistent hyperbole of hypnotic rhetoric:
Every Hutu should be aware that the Tuutsi woman, wherever she may be, works in the pay of her Tutsi nation. As a consequence, every Hutu is a traitor - if he marries a Tuutsi woman; - if he lives with a Tuutsi woman; - who hires a Tuutsi as his secretary or supports her (...) Every Hutu should know that every Tuutsi is dishonest in business. He has no other goal than the supremacy of his nation. (Chrétien, 1995: 39-40) .
In the text there is one major opposition, that between 'Tuutsi' and 'Hutu,' and a minor opposition, between men and women. These oppositions admit of no exception and are all-embracing, as the repetition of the word 'every' drives home relentlessly. A new element is introduced: 'The Hutu must stop pitying the Tuutsi.' (Chrétien, 1995: 40). This is an almost literal adhortation to disidentification, which nevertheless assumes that identification did occur previously, at least among some 'Hutu'. At this point, the disturbing invective 'inyenzi,' surfaces, to be translated by the familiar 'cafard', meaning 'cockroach.' Cockroaches keep on coming, they keep on eating, and the peasant must keep on killing them . At this point, the 'Tuutsi' have been transformed into a general category, abstracted from any specific local episode or historical experience, dissociated from any particular acquaintance, decontextualized and detemporalized. Identification of the 'Tuutsi', disidentification from the 'Tuutsi' and avoidance of all identification with the 'Tuutsi' - those are the necessary conditions for establishing a 'Hutu' identity. Through projection, all evil but still human characteristics have been assigned to the 'Tuutsi', by exaggeration they have been demonized into superhuman proportions of evil, and finally, through dehumanization they have been transformed into vermin. The process of disidentification is complete, it has gone even beyond hatred and achieved a level of dispassionate destructiveness. And yet, even at this point, a complementary process of identification continues, creating an equally abstract category of 'Hutu', who should be bound by mutual loyalty, regardless of place or rank:
''Every Hutu should consider another Hutu as his brother. If tomorrow one of the volcanoes were to erupt, the Hutu from Rukiga could come and live in Nduga and become by this very fact one of the people there. And if he expected a famine in Nduga, a Hutu from Nduga can go and stay in Rukiga and becomes one of them. But no matter what he does, a Hutu can not become a Tutsi, nor the inverse. (Chrétien, 1995: 98).'' .
Such were the messages that were incessantly relayed by the new medium, radio. In the late eighties, the government had distributed receivers in large numbers among the peasant population (Chrétien, 1995: 57) . As elsewhere in Africa, people carried them about wherever they went, closely pressed to their heads. Quia ex auditu fides, non ex visu, in the words of that great propagandist, Martin Luther: Faith comes from hearing, not from seeing.
The New Standard Model of Ethnic Conflict - The Explanatory Context
The widening scope of identification and disidentification was a necessary condition for the subsequent genocide, but as such not sufficient. A full explanation would require a complete account of the economic, political and military constellation and a precise analysis of the relation between the 'Hutu'-power' movement and the political regime in Rwanda . In this case, it would be frivolous not to try to distinguish fears of an actual military invasion on the one hand, from fantasies about the demonic nature of the 'Tuutsi's' on the other hand. At the time, the RPF, the army of 'Tuutsi' refugees from the massacres of 1959 and subsequent mass murders, who had lived in exile in Uganda and fought there in the victorious rebel ranks of Museveni, did constitute a real military threat at the Northern borders of Rwanda . And, of course, all that time the 'Tuutsi' minority in neighboring Burundi had remained in government and continued to control the army. The oppression of the 'Hutu' majority in Burundi resulted in mass killings, culminating in the 1972 'genocide' of 'Hutu's' and re-occurring to this very day (Lemarchand, 1996: 76 ff.) . In addition, there was the justifiable fear that the 'Tuutsi's' who had remained in Rwanda might form a fifth column, once the invasion from the North began. However, the target population of a genocidal hate campaign need not constitute a threat in any 'real' sense: the Jews in Germany did not, the Kulaks might have been expropriated in stead of exterminated, and it is hard to see how the victims of the Cultural Revolution in China threatened the Chinese Communist regime. In this sense, the relative autonomy of fantasy is vindicated. But the dialectics of identification and disidentification is played out within a specific political context of domestic disintegration and external threat.
The dynamics of this process may be summarized as the 'new standard model of ethnic conflict'. Developments in the transnational system of states erode the monopoly of violence of the state in a particular territory. If they succeed in mobilizing supporters behind some banner, this will be experienced as a threat by those who are excluded from the group assembled under that new label, since the state can no longer effectively protect them from violence. If under conditions of anarchy and anomie one witnesses others banding together as 'Serbs,' one better find shelter with kindred 'anti-Serbs', even if they have to be re-invented in the process as Bosnian Muslems or Bosnian Croats . But such labels can only be invoked effectively if they are elements, among others, of a collectively constructed past . But apart from the political context of a disintegrating state monopoly of violence, there was in the Rwandan case also the economic circumstance of an extreme scarcity of the one indispensable economic resource: land. Rwanda, with some 300 inhabitants per square kilometer, is among the most densely populated areas in the world, and what is more, there is hardly an alternative to agriculture. This lends all conflict of interest an especially explosive zero-sum quality: the land that one person wins is necessarily lost by another. It thus becomes hard to imagine that mutual compromise and consent between rival sides might ever profit all parties concerned. This is the material base for the extreme 'either'/'or' character of mutual group perceptions .
Once the dialectics of identifications and disidentifications is in full play, fantasies mutually exacerbate one another. At this point, activists will resort to violence for their own purposes: to oust members of the opposite group from rewarding positions, to take over their houses, shops, jobs and land. This is the 'rational action' aspect of the spiralling violence. Rape is an essential, symbolic and sexual, part of this expropriation. All these violent acts will vindicate the worst fears of the other group and prompt it to respond in kind. Every incident is magnified and resounds in the fantasies of the parties concerned. The presence of hostile neighbor states and the threat of invading armies reinforces and accelerates the process. Every outside intervention that should pacify the warring factions is perceived as an act of war by at least one of them. This compels outside forces either to stay out of the fray entirely or move in with massive force and subdue the entire population with military means. If no external power is willing and able to accomplish this feat, only inertia or the fatigue of battle can end the fighting. Everyone in Rwanda (and Burundi) had sufficient cause to fear violent attacks, once the state could or would no longer ensure safety. The 'Hutu' government in Rwanda was trapped between the 'Hutu-power' bands on the one hand and the threat of the refugee army in the north. The 'Tuutsi' regime in Burundi, almost completely controlled by the army, could not afford to relinquish its minority monopoly without the danger of being swept aside by the 'Hutu' majority.
Each country lived in fear that what had occurred in the neighboring state might next be perpetrated there, with the help of the adjacent regime. Over the years, the circles of identification and disidentification widened from the scope of village and lineage to generalized categories of 'Tuutsi' and 'Hutu' on a national and even transnational scale. With each round of violence new memories supplied the raw material for subsequent fantasies about the nature of the complementary group. Thus the stage was set for the next conflagration.
In Rwanda, the relation between the 'Hutu-power' movement and government circles was intricate but intimate. The militia were covertly supported by the state, even if sometimes publicly rebuked. In the end the government condoned the genocide and the army participated in the mass killings. But the massacres had been prepared long before in an orchestrated campaign which closely followed the fault-lines of pre-existing and widely-spread patterns of identification and disidentification that had been evolving for almost a century.
Abram de Swaan, Theory, Culture & Society 1997 (SAGE, London). Vol. 14 (2) p. 105-122.