FIRST WORDS

1This study has moderate objectives. It uses existing perspectives on violence in general, as a background for focusing on violence in the urban context. The major thesis is that towns and cities, as the hubs of political, economic and social processes, as well as social change, provide the locale for analysing the usually complex and interrelated forms of violence. The structural and spatial inequalities often found in towns, and the growth of slums and un-and underemployed sub-cultures, provide the milieu for the proliferation of violence, particularly against the existence of relatively high levels of opulence. By their very nature therefore, cities provide an enabling environment for violence, arising from individual and mass frustration caused by relative deprivation, political repression, and systemic injustice.

2For these reasons, the urban framework of revolution has been emphasized by many scholars, but this is only one part of the totality of violence. The urban framework enables political forms of violence, such as revolution and also non-political forms, such as criminal gang violence and armed robbery to be studied as phenomena deriving essentially from the same milieu. The South African case in focus here demonstrates this very well. With its long history of apartheid and discriminatory urban policies, sustained by one of the most repressive systems of state violence ever known in human history, the cities of South Africa show how violence in all its ramifications has come to constitute one integrated whole. Such an integration of violence can only be approached holistically and this is the major advantage of an urban focus. This is, however, only an introductory and exploratory study. Its purpose is to identify and demonstrate the ways in which urban violence can be most meaningfully studied.

1. Introduction

3Violence is central to the South African body politic, as is clearly reflected, not only in its pervasiveness, which has been described in terms of a culture of violence (Marks and Anderson, 1990) but also by, amongst others, the large volume of literature on the subject, the large number of researchers involved in studying its various ramifications, and the large number of research institutions, non government organizations, and various other bodies involved one way or another in projects aimed at understanding, preventing, and resolving problems of violence. At the government level, the concern has been no less substantial, especially during the transition to the post-apartheid polity, during which, rather than decrease as was hoped, violence escalated to unprecedented levels.

4From whatever perspective one looks at it, South Africa is a structurally violent country. It is commonly agreed that this is the result of the country’s long history of apartheid, a discriminatory system perpetuated by South Africa’s government and their armed forces, which excluded the non-white majority from its share of state power and entrenched social and economic inequalities between whites and non-whites in all sectors. The militarization of the apartheid state made violence somewhat inevitable, though there were other important predisposing factors, as will be seen. Because the policy of apartheid was all embracing, every instance of violence in South Africa was treated as an aspect of political violence.

5Political violence is the focus of most scholars of violence in South Africa. This focus appears all-embracing because most forms of violence are associated with apartheid and the struggle to end it. It tends to lay too much emphasis, however, on manifestly political variables which obviously cannot sufficiently explain acts of criminal violence such as rape, murder, gangsterism and armed robbery, which are much more pervasive than strictly political acts of violence (cf. Marks and Anderson, 1990; Simpson et al., 1991). What is clearly required is an all-inclusive framework which attributes to the various forms of violence the different emphases in explanation due to them, even while treating them generically as consequences of the same social (dis)order.

6The focus of this study – urban violence – advances such an all-inclusive approach. The approach takes its cue from the empirical fact that urban areas are centres of violence. If we are to fully understand the phenomenon of violence as having various forms which constitute an interrelated whole, we should locate this analysis in the urban context, just as many students of guerilla warfare and revolution in the Third World have, with success, located their analyses in the rural areas. This is not to say that violence is exclusively an urban phenomenon or that it bears no relationship to rural violence – in fact if nothing else, there is a great deal of rural-urban linkage in violence, and scholars like Skocpol (1979) argue that without peasant revolts urban radicalism cannot succeed in revolutionary transformations – but to emphasize the fact that it is in the urban areas that factors which are predisposed to violence are most dramatically played out.

7Concern with urban violence is not new, but it has not been sufficiently addressed in studies of Third World countries in general, and Africa in particular. Most analysts recognize the urban concentration of violence but few studies exist which view it as an urban phenomenon, while those who approach it from the urban perspective do so only in relation to explaining revolutions and as the location of social movements (cf. Farhi, 1990; Walton, 1979; Gugler, 1982). The approach here has a different orientation: the urban setting is taken as the locus of violence and its analysis. In other words, the importance of the urban area for our purpose is that it is the centre of all forms of violence, and not only as the breeding ground for revolutionary violence which is the popular focus in the literature. To repeat the overall objective of doing this: the urban context provides the framework within which conceptually different aspects of violence (including revolutionary violence) can be analysed as essentially aspects of the same generic phenomenon.

2. Violence as a Conceptual Variable

2.1 Clarifying the meaning of violence

8Although there seems to be a consensus that violence is any act that involves a threat to, or destruction of, lives and/or property, there is no agreement on how the concept should be defined and measured. As a value-laden concept, violence is subject to cultural, ideological, and religious idiosyncrasies. As du Toit (1990) has observed, it is essentially a contested concept. There is disagreement over whether violence is an objective or subjective phenomenon, over what to include and exclude as violence, and how to classify its various forms. Dating back to Thomas Aquinas’ notion of legitimate revolt against tyranny, violence has usually been defined in relation to the state which has led to distinctions being made between ‘legitimate’ or ‘justified’ violence and ‘illegitimate’ or ‘routine’ violence.

9Most political scientists have focussed on the former, in terms of revolution, rebellion, civil strife, internal war, and political conflicts, and have tended to disregard the latter category whose major distinction is that it is not directly, if at all, related to the state. Shupilov (1981) distinguishes ‘progressive’ violence which he regards as the struggle against capitalist oppression and tyranny, from ‘destructive’ violence which is mostly of a criminal kind and often involves ‘the personal aspirations of the individual pursuing his own personal aims’. Although not quite as explicit, Gurr (1970) limits his definition of violence to political violence against the state, which he classifies into three types: turmoil, i.e., relatively unorganized and spontaneous political violence with substantial popular participation (riots, strikes, demonstrations and localized rebellions); conspiracy, i.e., highly organized political violence with limited participation (assassination, coups d’etat and small-scale guerilla wars); and internal war, i.e., highly organized violence with mass participation designed to overthrow a regime or the state (large-scale terrorism, mass-based guerilla warfare and revolution).

10These distinctions directly affect what is defined as violence; the trend being, as indicated, to disregard non-political violence as undeserving of analysis by political scientists or better left to psychologists and other social scientists. Such narrow disciplinary delineations do not sufficiently address the epistemological problems involved in conceptualizing violence, for example, the question of its morality, and do not reflect its diversity and complexity as found in the urban area. They cannot suffice for South Africa where the distinction between political and non-political violence is blurred. What is needed is an all-inclusive definition. According to Galtung (1981), the best way to arrive at such a definition is to identify what is common to all phenomena termed as violent. He suggests two commonalities: destruction of lives and property and, on a more abstract level, anything avoidable that impedes human self-realization. The latter, being inherently subjective, is liable to difficulties in observation, but shows that violence is not always a directly observable phenomenon. The destructive aspect of violence is easier to discern, though its psychological aspects, like the trauma of a rape victim, cannot be quantified. Some authors regard death as a measure of destruction, but not all forms of violence involve death.

11Although Galtung’s point is very important, his identification of commonalities to all forms of violence in terms of the end, or consequence of, violence leaves out the area of the means of violence. Indeed, it can be argued that violence is not the only cause of destruction or the only avoidable thing that impedes human self-realization and therefore, that the distinguishing characteristic of violence is not its ends but its means. All forms of violence involve the use of force, whether this is legitimate (as Max Weber says the state’s use of force is) or illegitimate (as is the case when force is applied by the murderer or robber). Indeed, Gurr (1970), says political violence is subsumed under force which he defines as ‘the use or threat of violence by any party or institution to attain ends within or outside the political order’. The use of force is, for us, more important than the consequences of that use; though, for a well-rounded conceptualization, attention should be paid to both the means and ends of violence.

12Many authors in fact combine means and ends in defining violence. For example, Hibbs Jr. (1973: 8), says violence ‘implies the use of physical force and is generally evinced by the destruction of property, the killing or wounding of people, or the use of riot control equipment’. Similarly, Degenaar (1990: 71), defines violence as ‘the intentional application of extreme force against X in such a way that it is destructive of objects and physically injurious to animals and persons’. The main advantage in emphasizing means in the definition of violence, is that it includes the state as a principal actor in violence, something the focus on ends does not do. Another advantage is that it avoids the evaluative problem of whether violence is legitimate, moral or legal by making its definition more objective and empirical. As an empirical construct, it leads the researcher to ask the relevant question: Why and when does violence come into play? Any focus which is strictly on ends could generate a similar question but, by its very nature, it is more oriented towards justification.

2.2 Typologies of violence

13Violence has been classified in several ways, for analytical purposes and to reflect the diverse forms it takes. The bases of classification have ranged from the level of actors involved to the extent to which violence is endemic or incidental in a polity. Thus, at the level of actors, violence has been classified into intra-personal, inter-personal, inter-group, and inter-societal where inter-group and inter-societal violence involves small and large groups and tends to be more politically consequential, and intra-personal and inter-personal violence are regarded as having more of a criminal nature. Another classification which employs the same criterion distinguishes between mass or collective violence and isolated individual violence, which is mostly criminal (cf. Hibbs Jr., 1973; Welch Jr., 1980). Most students concentrate on collective violence because of its more manifestly political dimensions. This is an approach which does not reflect the totality of violence and the interrelatedness of the various levels at which it exists. As has been argued, the urban situation provides the best framework for such a holistic approach.

14Another criterion that has been used to classify violence is its relation or bearing to the State, which underlies the political/non-political typology. The violence that counts for many political scientists is that which Hibbs Jr. (1973: 7), says should meet three conditions: (a) it should be anti-system; (b) it should be politically significant by being actual and not a mere threat to the status quo; and (c) it should involve mass political action. Concern with rebellion, revolution, internal war, terrorism etc., clearly reflect the anti-state slant of the popular literature on violence.

15There are major problems with this typology. In the first place, there is a suggestion that violence that is not directly related to overthrowing the state or ruling class is not politically relevant, which is not the case put forward in this study. Secondly, by narrowing the political to anti-state action, some authors fail to give sufficient attention to the fact that the state itself is a repository of violence, following Weber’s definition. Domenach (1981:35-36), emphasizes the role of the state in violence thus:

… the state … is precisely the authority that makes of violence an institution having no obligation to conform to any moral or legal norm, for it is always prepared to use the maximum force if it considers its survival threatened... Whether we see the state absorbing violence or unleashing it, it is always tied to violence.

16Political violence can only be properly understood if the interface between state violence and anti-state violence is addressed, especially because, in some cases, anti-state violence is a response to state violence. The liberal democratic wisdom that the state is neutral or that its use of violence is legitimate has been questioned by students of state repression, terrorism and genocide, as well as leftist scholars who have continued to regard the state as an instrument of class domination (cf. van den Berghe, 1990; Stohl and Lopez, 1986; Bushnell et al., 1991). The argument presented below is that the role of the state is critical in any explanation of violence in whatever form.

17The final clarification that needs to be made on the political/non-political typology is about the tendency of authors like Shupilov to lump together all non-political forms of violence under the rubric of criminal violence. It should be obvious that not all non-political violence is necessarily criminal. There are, for example, categories of economic violence such as that perpetrated by capitalist exploitation and class domination; cultural violence such as entrenched male dominance and ritual violence; and various forms of social violence, such as those perpetrated by the mass media which are not necessarily criminal. Quite apart from these, supposedly criminal acts of violence such as murder and armed robbery can be justified, as they were in South Africa for a long time, as a continuation of the political struggle at the personal level.

18What these imply is: (a) that, to be properly analysed, the category of non-political violence needs to be further disaggregated into social, economic and cultural forms and not simply lumped together as criminal; (b) that even purely criminal violence has sometimes to be related to the political context of violence to be properly understood; and (c) non-political violence should not be dismissed as necessarily negative or deviant. As a UNESCO Panel of Experts pointed out:

… it is no longer possible to study violence as an exclusively negative phenomenon seen in terms of aggressive behaviour. It also [has] to be seen as a method of pursuing positive interests by other means, or as a response made in reaction to a less visible negative violence present in the whole of the social structure. (Domenach et al., 1981: 258)

19Another typology looks at the extent to which violence is endemic or incidental to the polity as its major criterion. Here, the major distinction is between structural and non-structural violence. Structural violence has a relative degree of permanence, being deeply embedded in the social structure and political order. Typically, a society in which there are deep-rooted socio-economic and political inequalities, in which the state is not only controlled by a tiny segment but used to perpetuate relations of dominance by repressing minority groups, and one in which the people are denied their rights to freedom and integrity by force, in short, a society in which the state relies exclusively on the use of force to ensure compliance, can be described as structurally violent. Clearly, the apartheid state in South Africa was structurally violent, as were other colonial states, particularly those of the settler kind. Non-structural violence on the other hand is not endemic, and is found in societies in which inequalities and other injustices are not entrenched in the social structure, and in which the state permits citizen participation and relies on voluntary consent rather than force, which is used sparingly and only in situations where the survival of the state is seriously threatened.

20As an analytical construct, the structural/non-structural typology serves the very useful purpose of allowing us to differentiate highly violent from less violent societies. But even so, the categories must be understood as relative, because there is no society where one form of inequality or another does not exist; though, admittedly, inequality is more entrenched and upheld by state power in some than in others. The point in making this observation is to warn against the common error of assuming that only societies structurally disposed to violence can be violent, or that such violence is automatic. For one, there are countries where structural inequalities have been carefully managed so as not to lead to violence. For another, inequalities or discrimination create situations of violence only when they are politicized and become the basis for political and social struggle. As Welch Jr. (1980: 34), put it, ‘It is inequality perceived as inequity …that gives rise to collective action, including collective political violence. Inequality (real or imagined; individual or group) becomes the basis for resentment when differences are perceived as unjust’ (also see Galtung, 1969 for a similar view). So, even in so-called structurally violent societies, we still need to know how inequalities were transformed into inequities. This is the key to structural analysis.

21The final typology we shall consider, which is of immediate relevance to this study, is that which classifies violence as rural or urban. The point in this classification is not that there is a category, ‘urban’ violence, distinguishable from ‘rural’ violence, the way some modernization, scholars have interpreted it (cf. Hibbs Jr., 1973), but that the essence of violence is basically the same in both rural and urban areas, the only notable difference being perhaps the superior implements of violence (weapons, mass media, and institutional support) available to urban dwellers. In this paper, the essence of the classification lies in the differences between urban and rural areas due to location: by its very nature, the urban area with its concentrated population, weakened and dislocated cultural and social controls, industrialization, concentration of government institutions and so on, is more prone to ‘structural’ violence than the rural area. The greater diversity of social relations in urban areas also means that one is more likely to find a greater representation of the various forms of conflict and violence there, than in the villages. This is not to deny the linkages that exist between violence in rural and urban areas, but to show that drawing a distinction between the two locations enables the researcher to focus on the totality of violence in either of them, thereby giving violence its necessary ecological framework.

22Typologies basically serve to delimit the researcher’s scope of study, and guide his theoretical approach. To complete our discussion of violence as a conceptual variable, we shall next consider the various theoretical explanations that have been given for violence.

2.3 Theoretical perspectives on violence

23The reason for seeking explanations for violence hinges on the assumption that violence is purposive action. In other words, the only way to make ‘sense’ of violence both from the point of view of its users and the analyst, is to see it as aimed at the attainment of certain objectives. This notion of rationality ties in very well with larger questions of morality and legitimacy and the search for the justification of violence. Arendt (1970), says, ‘a theory of revolution can deal only with the justification of violence’. Within these larger contexts, violence for its own sake, assuming there is something like that, does not make sense. But to say that particular acts of violence do not make sense is not the same thing as saying they cannot be explained. They may not make sense to the analyst but, to the actors involved, they have significance. Therefore, in seeking to explain violence, there is an overwhelming need to pay as much attention to the sense in which it appears to the analyst as to the actors themselves. That balance can be found when violence is situated within its overall social context and sense is made of the explanations offered by the actual perpetrators of violent means.

24There are several explanations for violence which, depending on the nature of study and the approach adopted, may be regarded as contending or complementary. For a study such as ours, which approaches violence holistically and which has no narrow limitations as to discipline or level of analysis, the explanations will be regarded as complementary. These explanations can be classified according to the unit of analysis, whether individual, small group, large group; the levels of analysis and theory, which could be micro, medium or macro; as well as the conception of violence, e.g., as only political, legitimate or progressive; all of which are underlined by disciplinary ‘specializations’. In the discussion which follows, particular attention will be paid to the relevance of these explanations to the South African situation and to how various scholars have sought to apply them.

25At the individual and micro levels or, as some prefer to call it, the intra-personal and inter-personal levels of analysis, the major explanations of violence are bio-social, psychogenic, and psychological. The most po’pular perspectives here are the frustration-aggression thesis which attributes violence to frustration, which breeds anger and finally leads to aggressive behaviour (cf. Dollard, 1939; McNeil, 1959; Laborit, 1981; Idineberg, 1981); the bio-medical perspective which attributes violent behaviour, especially in cases like rape and murder, to certain psychiatric disorders in the individual; the revolutionary personality perspective which attributes revolutionary potential to elements of personality development in childhood and adolescence, and has been used to study revolutionary leaders like Lenin (cf. Rejai, 1980; Rejai and Phillips, 1979; and Wolfenstein, 1971); the relative deprivation perspective which holds that when people habitually experience a discrepancy between their expectations in life and their actual achievements, they are very likely to engage in violence (Gurr, 1970; Davies, 1962, 1969); and the systemic frustration perspective which, as a variant of relative deprivation at the macro (structural) level, says violence is most likely in transitional societies where the ‘revolution of rising expectations’ is matched by the ‘revolution of rising frustrations’ (Feierabend et al., 1969). These differing perspectives have been applied to the South African situation by several authors, though the most popular perspective is relative deprivation.

26In general, these micro-level theories have been criticized for laying too much emphasis on the individual’s personality to the virtual exclusion of the social structure, and for providing inadequate explanations for collective and widespread violence (Coser, 1967; Seegers, 1991). The relative deprivation perspective is distinct in these senses because, unlike the others, it does not attribute violence to individual personalities, and is basically aimed at explaining collective violence. In fact, Gurr developed the thesis to explain the three types of political violence which he identified, namely, turmoil, conspiracy and internal war. We will dwell briefly on this perspective because it is directly relevant to our concerns, and has been a popular point of departure for most studies of violence in South Africa (cf. Manganyi and du Toit, 1990).

27The relative deprivation thesis is best stated in the words of Coser (1967:59):

[It is] deprivation that arises not so much from the absolute amount of frustration as from the experienced discrepancy between one’s lot and that of other persons or groups that serve as standards of reference. Whether or not superordinate groups or persons are taken as standards of reference by subordinate groups and individuals depends, at least in part, on whether the unequal distribution of rights and privileges is considered illegitimate by them.

28Therefore, relative deprivation is most likely to result in violent uprising when the discrepancies are considered illegitimate, as was clearly the case in apartheid South Africa and situations of national liberation struggles. In other instances where relative deprivation is accepted as legitimate, as for example, in caste-based societies or strictly religious systems where inequalities and deprivations are rationalized, it is less likely to produce violence. Coser further points out that relative deprivation is more likely to result in violent action “when normative restraints and traditional expectations have been shattered’ (1967: 69). This is particularly true of the youth:

The breakdown of tradition creates in the young two seemingly contradictory expectations: the fear that the gradual advancement in the age hierarchy is put into question, and the hope that it is no longer necessary to wait the requisite number of years for the rewards of maturity. Insecurity about the future and hope for the present leads to behaviour, that so far had been future-oriented, to be replaced by present-oriented activity. (Coser, 1967: 70)

29This insight explains in large part the critical role of the youth in violence, particularly in South Africa after the Soweto massacres of 1976, when they took over the leadership of the anti-apartheid struggle from the politically neutralized elders. This ascendancy of the youth is believed to have intensified the violence in the 1970s and 1980s, as the older members may in all probability have been less rapaciously violent.

30Although relative deprivation is, by definition, an individual attribute, the perceptions of deprivation have to be generalized to produce a sense of community before it can produce mass action. In other words, ‘A group rather than a person must be presented as the victim of inequity’ (Welch Jr., 1980:36). The politicization of discontent which Gurr (1970), paid little attention to in his original formulation of the relative deprivation thesis, is therefore a critical factor in applying this explanation. In the case of South Africa, the ANC, PAC, UDF, the Black Consciousness Movements of the 1970s, the students organizations, the trade unions, as well as the various community and civic organizations, all intensely politicized the inequalities and relative deprivations of the apartheid era.

31Even with all these, the actions of the state determine, in the final analysis, whether or not violence will ensue: If the state is open to negotiation and prepared to effect desired changes, violence may yet be averted or minimized; but if it seeks to perpetuate the status quo by force, as the state sought to do in South Africa, it makes violence inevitable. Two other factors have been identified as intervening between the anger caused by relative deprivation and the recourse to violent action (Seegers, 1991). The first is the coercive balance which involves an assessment of the balance of violent power available to the state and to the individual or group by the actors: If the state’s power is overwhelmingly superior, violence is not likely, or at best, it will take the form of underground terrorist or guerilla activities; but if the state’s capacity is adjudged to be weak, as was the case with the rebel attacks in Liberia, Sierra Leone, Uganda, and Rwanda, a recourse to violence is most likely.

32The second is the justification of violence, which is a social structure variable. Cultures, ideologies and religions hold different attitudes to violence which could serve to legitimize violence or make it less likely. In South Africa, the recourse to violence received the active support and encouragement of the strategic non-state elite which led the anti-apartheid struggle: In 1985, the ANC declared that ‘police and soldiers must be killed!’ That same year, Oliver Tambo said that violence was to be intensified against black councillors and the police; between 1983 and 1993, about 953 policemen were killed. In 1986, Chris Hani praised the youth for ‘clearing our townships of collaborators, puppets and agents of the regime’ while, in an interview, Alfred Nzo, then Secretary General of the ANC, said the organization supported whatever method the people used to destroy enemies, including ‘necklacing’ (Sunday Times, London 14 September, 1986). Even some church organizations rationalized violence, a point which is very significant in a country which prides itself as Christian. The rationalization of apartheid by the Dutch Reformed Church is well known, while, from the perspective of the challengers, clergy from twenty-two denominations produced the ‘Kairos Document’ in 1985 which condemned blacks who collaborated with the apartheid state and said ‘intimidating’ them ‘was self-defence’ rather than violence, since only oppressors could be said to be violent.

33So, relative deprivation has to be related to the larger social and political contexts to provide an adequate explanation. The real problem with the perspective, however, is methodological. How do we measure relative deprivation? How do we know when a person or group feels relatively deprived? The way it is, the explanation is offered only after violence has occurred, something of a data-fit, although it is possible to proceed on the basis of observable inequalities to predict a violent outcome as many authors did in the case of South Africa. But in doing so, i.e., in using aggregate data, absolute deprivation is made to take the place of relative deprivation. This methodological drawback does not, however, reduce the explanatory power of the relative deprivation perspective as long as it is not used in isolation from other systemic factors.

34At the medium and macro levels, there are explanations in terms of the group and society as a whole. One of the well established views in this category sees violence as being intensified where societies are in transit, i.e., undergoing the processes of modernization and industrialization (cf. Graubard, 1964; Kerr et al., 1960; Olson, 1963; Kornhauser, 1959; Huntington, 1965; Deutsch, 1961; Almond and Coleman, 1960). One of the key areas of the modernizing process, which Deutsch encapsulates in the concept of social mobilization, is urbanization. Increasing urbanization has been said to be responsible for the escalation of conflicts and violence in societies in transition. As Hauser (1963: 212) puts it:

[rapid urbanization promotes] internal disorder, political unrest and government instability fed by mass misery and frustrations. The fact that the differences between the ‘haves’ and ‘have nots’ peoples within nations have become ‘felt differences’ and that we are experiencing a ‘revolution of increasing expectations’ have given huge urban population agglomerations an especially incendiary and explosive character.

35Another important dimension of the modernization-transition perspective on violence is that it identifies increased demands for greater political participation as one of the key factors for increased conflicts, especially where the political system is unable to cope with the demands. Generally, the modernization-transition perspective has been criticized for its system stability bias and for overstressing the role of modernization in political conflicts. Indeed, Seegers (1991: 218) points out that supporters of the apartheid regime in South Africa tried to justify their totally repressive policies of the 1980s on the grounds of ensuring system stability. None the less, the perspective draws attention to the need to examine the role of aspects of social change such as increasing urbanization, rural-urban migration, increased political mobilization of the urban dweller, and how these affect the social structure.

36The class-based Marxian perspective explains the inevitability of conflicts in terms of the irreconciliabity of class antagonisms, the oppressive character of the capitalist system, and the fact that the state is an instrument of class domination. Violence is the major instrument of the proletarian revolution – Marx regarded it as the labour pangs of the revolution. Many authors have applied Marxian explanations to violence in South Africa with remarkable success, not only because the country has one of the most advanced capitalist economies in the African continent, but also because it has clear-cut class cleavages which coincide with racial and ethnic cleavages. Nevertheless, care has to be taken not to shield the autonomous factors of racism and apartheid by subsuming them under capitalism as some authors have done.

37The Marxian perspective has three important advantages for the purposes of this paper. First, it directs attention to the state as an actor and object in violent struggles. In fact, Marxian scholars like Shupilov (1981) regard only violence directed at overthrowing the capitalist state or in pursuance of the class war as worthy of attention. We have already argued that other forms of violence need to be analysed as products of the same social milieu. Secondly, in the writings of Karl Marx, although it is not explicitly stated, urban areas were regarded as the points of capitalist concentration – of industries especially – and, therefore, as the centres of violent conflict. Although latter day Marxists like Lenin and Mao Tse Tung revised the original thesis to accommodate peasant and rural-based revolutions, the fact remains that the urban area where the state’s apparatus is concentrated has to be taken over before the revolution can succeed. The final advantage of the perspective is the methodological emphasis it places on history, the economy and the social relations to production. As Shupilov (1981: 142) puts it, The social content, the political purport and the aims of violence are directly dependent on the social relations it reflects and defends’.

38There are several other medium and macro perspectives which dwell more on the processes, strategies and structures of forms of collective violence, especially revolutions (cf. Brinton 1938, 1968; Johnson, 1973; Hutchinson, 1978; Beckett and Pimlott, 1985; Rejai, 1977; Tilly, 1978; Skocpol, 1979; Skocpol and Somers, 1980). These explanations mainly have to do with factors making for successful revolutionary outcomes and are particularly useful for identifying the structural factors present in conflict situations as well as those which make social movements (or revolutionary movements) more likely to achieve their goals. For example, in addition to classifying the groups in a polity into four – government i.e., the group which controls="true" sources of coercion; polity, i.e., the group whose demands are routinely satisfied; challengers, i.e., the group whose demands are not routinely satisfied; and non-contenders, i.e., the group of passive citizens – which facilitates the identification and analysis of potentially violent groups, Tilly (1978) explains the success or failure of collective (revolutionary) action as dependent on (a) the extent to which interests and goals are shared by members of the group, (b) the strength of organizational structures, (c) the amount of resources available to the group, and (d) the extent to which the group is opposed or assisted by others.

39From Tilly’s insights, we see that violence has to do more with the relations among groups than any particular psychological constellation, that governments are principal actors in violence, that urbanization and industrialization are important determinants of inter-group relations; and, that to the extent that there are challengers in every system, violence is endemic in all political systems. In relation to South Africa in particular, Tilly’s insight on group interaction is a useful framework for analysing the so-called black-on-black violence which, on the face of it, involves conflict between supporters of the African National Congress (ANC) and the Inkatha Freedom Party (IFP). As Tilly points out, group interaction takes place within the context of the struggle for power. It is also important in analysing the solidarity of the anti-apartheid struggles, and the extent to which internal and external support to the liberation movements made them successful.

40Skocpol’s structural perspective, to take another example, is also very important for, amongst others, helping to factor the contributions of the international system into an analysis of violence and revolution. Skocpol not only posits that the state possesses relative autonomy and is not merely an instrument of class domination but that it is also a critical actor in revolutions. However, its weakening or collapse is due not only to international pressure but from internal crises (from below), mainly from countryside peasants who form coalitions with other groups. Although Skocpol was more concerned with countryside-based revolutions, her structural perspectives provide a lot of insight into studies of urban-based revolutions as Farhi (1990) has shown in her brilliant study of Iran and Nicaragua. For the South African case, which is our immediate concern, the factor of international pressure and crisis-from-below is very useful in explaining the final demise of the apartheid state.

41One theory which provides some explanation for violence or at least a framework for analysing it, and is relevant to all levels of analysis, but has not received much attention, is the ethnic perspective. The only notable references to ethnicity in the mainstream literature on violence are in terms of its being an organizing principle for forging community solidarity for purposes of collective rebellion (Welch Jr., 1980); and its relevance in the analysis of genocide and state-based ethnic domination (see the various articles in van den Berghe, 1990). Why has ethnicity, which is a well acknowledged source of political conflicts, not been fully explored as an explanation for violence? One reason could be the concentration on revolutions or legitimate political violence. But, like religion, which students of violence recognize as an ideological factor in countries like Iran and India, ethnicity has been known to be the basis for political violence both by the state and its challengers.

42Another reason could be the controversial contention that ethnicity is a dependent rather than an independent variable. But even the worst critics of ethnicity agree that there are some residual categories of conflict which arise from ethnocentricism, language, autonomy, national self-determination, and so on. In any case, ethnicity has been found to be a major element of conflict in multi-ethnic states. What all this means is that greater attention needs to be paid to the ethnic perspective for the light it throws on inter-personal, inter-group, and inter-societal conflicts and violence. In doing so, however, ethnicity has to be understood within the wider socio-economic and political contexts and related to the other forces making for violence.

43Fortunately, efforts have been made to interpret violence in South Africa in terms of ethnicity for the reason that, as Adam (1990) describes it, the apartheid state was an ‘ethnic state’, i.e., a state which was perpetuated as the exclusive preserve of only one group. In addition to the white-black distinctions which are mainly racial, there are important cleavages among the white population (Afrikaans, English, Portuguese) as well as among the non-whites (Indians, coloureds and the various Black groups, notably the Xhosa and the Zulu) which have important consequences for the matrix of violence. Adam (1990) and Adam and Moodley (1993) have analysed how the apartheid state manipulated these cleavages, especially among the blacks, to keep the opposition divided, and attributes the so-called black-on-black violence largely to these manipulations (for arguments of apartheid as opposed to ethnic origins of these conflicts see Taylor, 1991).

44Adam’s arguments stand in interesting contrast to those of du Toit (1994) who regards recent black-on-black violence partly as a continuation into the present of the past of black-on-black violence which followed the expansionist policies of the Zulu kingdom in the late 19th century; from which, as it were, the colonialists saved them! While du Toit’s historical excursus is an important methodological element, he fails to recognize the larger context of the supposedly new form of violence and to give the manipulation thesis its due weight. As Simpson et al. (1991: 10) explain:

[Black-on-black violence is] a part of South Africa’s racial legacy [and] so clearly fails to pierce the veil of apartheid as to make its regular usage appear to be ideologically manipulative. Far from offering an explanation of violence, this term does more to disguise causation by reinforcing the camouflage of racial stereotypes.

45But this is not all that the ethnic perspective can help to explain. It can also be applied to analysis of the Afrikaaner demands for a volkstaat and recognition of their right to self-determination, which have underlain the terrorism of the extreme right Afrikaaner movements since 1990.

46This discussion of theoretical perspectives on violence does not exhaust all existing perspectives, but it does point to the rich variety of explanations that have been and can be offered. As was stated at the beginning of the section, for a study of urban violence which hopes to deal with violence as a complex whole, the various explanations are complementary rather than mutually exclusive. Thus, group and class-based explanations as well as structural ones which deal largely with political and collective violence are to be balanced with those at the micro level, which dwell almost exclusively on psychological factors. The point in doing so is to be able to explain the various forms of violence, whether derived from the same sources or not; and not, as some authors do, to offer one set of explanations for the special category of political violence and another for other forms.

2.4 A brief note on the state of the study of violence in South Africa

47It is no exaggeration to say that every social scientist in South Africa is involved, one way or the other, in the study of violence. An analysis of published works on violence in major journals and textbooks shows very clearly that the South Africa case is the most studied in Africa, while a further analysis of the subject matter of works published by South African social science scholars shows that one out of every three publications is either on violence or some directly related subject.

48This balance sheet is also impressive for the well developed multi-disciplinary and collaborative research orientations that include such disciplines as medicine, linguistics and religious studies, which are not usually part of the social science action set (cf. Manganyi and du Toit, 1990; Chidester, 1992). The vast array of the dimensions of violence which have been covered is remarkable. In addition to the more conventional concerns of political and criminal violence, a gender school of violence has crystallized (cf. Vogelman, 1990; Vogelman and Eagle, 1991) as has a youth and children’s school (cf. Bundy, 1987; Nasson and Samuel, 1990), while efforts have been made to re-create explanations of violence from the pre-modern times (du Toit, 1994). Indeed, the dynamism of these studies cannot be better reflected than by the concern already being devoted to the transition to post-apartheid violence (cf. Everett and Sisulu, 1992; Adam and Moodley, 1993).

49In spite of this heavy scholarly involvement and what may be considered the ‘saturated’ study of violence in South Africa, some gaps can be discerned in the literature. Some of the gaps arise from the disproportionate concentration on political violence, which as indicated earlier, is the result of the historical centrality of violence by and against the apartheid state. Indeed, virtually all studies of violence, in whatever form, allude to the political dimension which seemed to make ‘sense’ of violence when oriented towards the overthrow of the apartheid state. The limitation of such a reductionist approach is aptly borne out by du Toit’s (1994) recent attempt to make ‘sense’ of post-apartheid violence outside the struggle to end apartheid.

50While it might be necessary to justify violence in terms of its bearing on the political struggle to end apartheid, this should be taken as only one of the possible political variables. There is the other equally important variable of the struggle for power itself; afterall, the struggle to end apartheid was not to be an end in itself, it was a means by which a particular segment of the black majority hoped to capture their share of state power and thereafter compete with others to keep it. There is also the whole question of participation and empowerment which involves competition among groups outside the ambit of the state. Violence in South Africa, however, is not all political; in fact, if deaths and destruction of property are taken as a good measure of the intensity of violence, then it can be argued that violence such as murder and armed robbery, which are not directly political, are far more serious than political violence.

51This is not to deny the centrality of political violence, as even the murderer, in South Africa, justifies killing on the grounds of fighting discrimination and structural inequalities, but to say that these forms of social and criminal violence should not be subsumed under political violence as some scholars have done. It is only by separating them that the linkages to political violence can be more meaningfully studied. The call therefore, is not for separation of analysis for its own sake, but to ensure that other forms of violence are given proper attention and studied as integral parts of the matrix of violence. As argued before, a focus on the urban locale of violence enables us to do this most successfully.

52In terms of dimensions, the study of violence in South Africa has been fairly exhaustive, notwithstanding a tendency to link the various forms to political motives. There have been studies of different aspects of political violence as they relate to the struggle against apartheid (cf. Levin, 1987; Cock and Nathan, 1989; Johnson, 1988; van Vuuren et al., 1983; Saul and Gelb, 1986; Mckendrick and Hoffman, 1990; Manzo, 1992). These have included studies of rape, murder, robbery, gangsterism and other forms of criminal violence (cf. Hansson and van Zyl Smit, 1990); of taxi wars, township wars, the so-called black-on-black violence (cf. Meer, 1989; Khosa, 1992; Taylor, 1991) of the role of security forces, including the police, in the escalation of violence (Cock, 1990; Cawthra, 1992); of conflict resolution and other aspects of violence management; of the roles of urban social movements and trade unions (Swilling, 1986; Seekings, 1988; Lodge et al., 1991); of the roles of students and youths; of the historical origins of violence, especially of the earliest forms of political protests.

53However, some important aspects of the crisis of violence have either been completely neglected or understudied. One of these is right-wing violence which has been on the increase since February 1990, when the president at the time, F.W. de Klerk, renounced apartheid. The Human Rights Committee reported in 1990 that, following the unbanning of the ANC, there were at least 45 right-wing attacks which resulted in over 26 deaths and 138 injuries. In the period immediately before the 1994 elections, right-wing groups carried out bomb attacks on Jan Smuts airport in Johannesburg and several other places. Best described as counter-revolutionary, because their objectives revolved around an opposition to black rule and the demand for an autonomous Afrikaaner republic, right-wing violence as perpetrated by the AWB and other organizations has taken the form of terrorist bomb attacks, assassinations and murders. The critical place of right-wing violence in the overall context of violence is underscored by its long history (the Broederbond operated for a long time to advance Afrikaaner interests violently); the alliances right-wing bodies formed with the IFP and other anti-ANC black factions; the association of key officers of the defence, security and police forces with right-wing activities; the migration of right-wing elements from Germany and other parts of Europe to South Africa to give support to the cause of ‘racial purity’; and the resolve of AWB leaders to continue terrorist activities. Right-wing violence certainly deserves as much attention as that which has been given to black-on-black violence in recent times.

3. Urban Violence

3.1 Why urban violence?

54The pre-industrial and agrarian nature of most Third World countries where revolutions or rebellions have taken place – China, India, Kenya, and Zaire, to take the example of cases studied by Welch Jr. (1980) – has made the rural area the setting for most studies of collective violence. The focus on the rural locale has been justified by Skocpol (1979) who argues that without peasant and rural support, urban radicalism in agrarian societies cannot ultimately bring about revolutionary transformation. In an important article aptly titled The Urban Character of Contemporary Revolutions’, Gugler (1982) points out that Skocpol’s rural framework is best suited to the study of one type of revolution, namely, the national war of liberation which is typically mass-based and does not threaten the existence of the metropolitan powers. Even so, he points out, rural-based revolutionary movements have often been led by urban radicals and activists who work underground and supply arms, finance and food. For example, in Cuba, between 60 and 80 per cent of the guerillas including Fidel Castro and Ernesto Che Guevara, were drawn from the urban populations. In any case, Gugler says, rural movements often need urban solidarity to succeed because:

loss of control over rural areas is not a sufficient condition to persuade a national government to relinquish power to a revolutionary movement. For such a movement to succeed it must confront the government in its urban location. Control over the capital city is usually of crucial importance. (1982: 68)

55Gugler’s main point is that, unlike national liberation movements, which are typically mass-based, ‘revolutionary movements’, which he defines as movements that employ extra-legal means in challenging a national government, often in alliance with elements outside government and security forces, are mostly urban-based. From his comparative study of the revolutions in Cuba, Bolivia, Iran, and Nicaragua, Gugler (1982: 66) concludes that ‘the rapid urban growth experienced by most Third World countries can be taken to presage the age of urban revolutions’. Another remarkable study of revolutions as urban phenomena is Farhi’s (1990) comparative study of the Iranian and Nicaraguan revolutions. She makes basically the same points as Gugler does, but for her, the key to an urban focus lies in the central role of the state in the process of industrialization:

Given the central role of the stale in the industrialization processes of most peripheral formations, it is only logical to assume that political conflict takes the form of interaction between the state and organized urban groupings. This is because in the urban setting, all classes demand access to the collective material condition of daily life, and the state increasingly intervenes to resolve or ameliorate the contradictions ensuing from these conflicting demands. (Farhi, 1990: 18)

56Furthermore, she says, the concentration of schools and students in the urban areas and the large population of the urban poor, which provide a critical reserve for revolutionary mobilization, make urban areas the hotbeds of revolutions.

57Although our concern is with violence in general rather than revolutions, the foregoing provides a useful introduction on why the urban area should be the locus of studies of violence. Recognition of this fact is not new. Marx (1957) talked of the ‘idiocy of rural life’ and the inability of peasants acting on their own to undertake class action. Modernization scholars of social change and transition have highlighted the role of urbanization in the intensification of conflicts in modernizing societies. The social movement school of urban sociology, identifies the city as the locus of political conflict arising from the state’s intervention in the provision of social services (Castells, 1977, 1983; Saunders, 1979; Dunleavy, 1980; Lowe, 1986). Castells posits that the city provides the arena for challenges to the dominant capitalist order; specifically, it enables social movements which are harbingers of social change to ally with industrial workers to change the configuration of power to a socialist one.1

58The key question in focussing on urban violence is of course: Why urban? There are three answers. The first is that as the fulcrum of social, economic and political processes, the city provides an enabling environment for conflicts and violence in all its ramifications. Secondly, fundamental mistakes in urban policies and planning create conditions which make violence endemic to cities. The rise of slum areas where ‘deviant’ sub-cultures prevail, acute shortages of housing which encourage overcrowding and informal settlements, insufficient policing, lack of recreational facilities, and neglect of the peripheries of cities in the provision of essential services, are some of the consequences of planning failures. Where these are a matter of deliberate urban policies rather than policy failure as has been the case in South Africa, the potential for endemic violence is increased.

59One perspective which directly alludes to the urban condition in explaining violence is the Chicago school which attributes criminal, especially gang violence in cities, to the conditions of slums and the development of deviant sub-cultures (cf. Yablonsky, 1962; Cohen, 1955; for studies of the consequences of slums in African cities, see Obudho and Mhlanga, 1988). South Africa has its own share of urban gangs which, in confirmation of the urban slum hypothesis, are found in the townships and non-white neighbourhoods. The third reason for focusing on urban violence is methodological, which is that the urban framework makes it possible to analyse the various forms of violence as ensuing from essentially the same social structure. For example, within an urban framework of homelessness and unemployment, armed robbery, rape, and mob action can be interpreted as deriving from the same generic frustration. In the case of South Africa, it particularly helps to draw the linkages between political and non-political violence. Since we have already dwelt considerably on the methodological advantages of focussing on urban violence in the preceding section, we shall concentrate on how the city, as the hub of activities, provides an enabling environment for violence.

60The city, especially the capital city, is the centre of all activities in a country. It is the seat of government and its various agencies, including the security agencies. As such, any action which expects to achieve significant results, especially those which aim at political changes, have of necessity to take place in the city or involve action in it. The concentration of security forces in the city also enables the state to ‘mastermind’ the city, and gives it a stake in its survival; the state’s action in this regard, for example, by policing a riotous crowd, only serves to engender violence. Moreover, their proximity to the state government also changes the outlook of urban dwellers towards their grievances: it tends to make their demands ‘immediate’ and achievable, which explains the readiness of aggrieved urban populations to rise up in arms.

61Second, as the centre of development in the country, the capital city attracts large populations for a variety of reasons; for employment in industries, government establishments and private enterprises; for the ‘good life’ which is absent in the countryside – electricity, potable water, recreational facilities, and the myth of ‘endless opportunities’; for educational choices and so on.

62One aspect of the attractiveness of the urban areas thesis, which has been well studied by students of social change, is the effect of rural-urban migration on urban conflicts. The popular notion at a time was that the city radicalized the migrant, but this has been found to be untrue (see Weiner, 1967; Nelson, 1969; Portes, 1972; Cornelius Jr., 1975; Kazemi, 1980). From his European studies, Tilly (1969) concluded that, in the short run, rapid migration from rural areas and the growth of large cities acted to dampen rather than spur violent protests, and gave two reasons for this: (a) migration withdrew discontented men from communities in which they already had the means for collective action and placed them in communities where they had neither the collective identity nor the means necessary to join forces with others, and (b) it took considerable time and effort for the individual migrant to assimilate in the large city and thus to join others in staging collective action. In other words, it is only after the migrant has been integrated into the urban social movements that his propensity for violence increases.

63Another popular explanation for the high incidence of violence in the cities is that which alludes to the social discontinuity which attends rapid urbanization. As Weiner (1960: 173) puts it, ‘large numbers of rootless, crowded, and often unmarried urban dwellers are easily prodded to violence and readily organized by political groups’ (also see Hauser, 1963). The high rates of divorce and broken families, the weakening of traditional and cultural social controls, the anonymity, individualism and, in many cases, alienating tendencies of urban life, all belong to the social dislocation variable. Alongside social dislocation there is the widespread frustration of the urban poor, who were attracted from the countryside by the hope for a ‘better life’. Coming to the city stimulates high expectations, but more often than not, these expectations are quickly replaced by feelings of relative deprivation as the system proves incapable of providing jobs, houses, higher incomes, affordable education for children, and so on. The frustration tends to be more pronounced as the urban dweller sees around him members of the upwardly mobile classes, the affluent and powerful, and finds himself asking why he cannot be like them. As was pointed out earlier, the extent to which there is a general acceptance of these inequalities as legitimate or illegitimate will determine the likelihood of frustrated people acting individually or collectively to demand redistribution.

64A final structural element that makes urban areas hotbeds of violence is the concentration of civic organizations, as well as the large reserve of violence-prone segments of the population in towns and cities. Trade unions, professional associations, human rights organizations, women’s organizations, ethnic associations, students’ organizations and several other civic organizations are located in the cities where they aggregate and articulate collective interests. The strikes, lock-outs, consumer boycotts, rent boycotts, riots, and demonstrations, for which many urban areas have become notable, are organized by these groups.

65The cities are the bases of militant and terrorist groups, and are also the targets of their attacks. Then, in terms of an urban ‘army’, there is always a large concentration of the poor, homeless and unemployed who are easy recruits for criminal and political violence, and whose participation usually turns peaceful demonstrations into bloody confrontations with the police. Students and the youth constitute a special category of this army, especially in polities where structural inequalities are regarded as illegitimate. They have been known to engage in collective action to challenge those in authority over specific educational policies on behalf of society. In South Africa, the youth and students virtually took over the anti-apartheid struggle in the mid-1970s, and escalation in the violent character of that struggle can be partly attributed to this fact. In societies in which inequalities are structurally entrenched, these civil society groups engage in various forms of related struggles. In such situations, as was the case in South Africa where the civilians, students organizations, trade unions, self-defence units, liberation movements and armies, and even prisoners organizations joined forces to fight apartheid, the ensuing violence had basically the same roots. It is in such an integrated sense that violence should be studied, and the urban locale provides the framework for doing so.

66Without explicitly stating so, most analyses of violence in South Africa have been about urban violence. Most violent activities take place in the towns, especially the townships, and data required for analysis, as compiled by the police, human rights bodies and other monitors of violence is more readily available for urban, than rural areas. This is due partly to the neglect of the black countryside under apartheid. Studies of violence in the black townships and of the labour, youth, gangs, and security dimensions of violence are particularly notable in this regard (cf. Lodge et al., 1991). Most of these studies are, however, set within the framework of ‘revolutions’ and anti-apartheid struggles. Nevertheless, there have been a few studies of violence as a specifically urban phenomenon or of urbanization as a facilitator of violence (cf. Swart, 1983; Smith, 1992; Lemon, 1991). For example, Saul and Gelb (1986) raise the ‘urban question’ in terms of how migrant labour policy, inadequate provision of housing, and the fact that the urban areas were the ‘nodes of white economic power’ made the cities the ‘arenas of war’. Swilling (1988) adopts a similar approach, but pays particular attention to social movements and why state terrorism could not dislodge liberation struggles (also see Swilling, 1986 and Swilling and Shubane, 1991).

67Seekings (1988) also emphasizes the relevance of the urban social framework in analysing violence in South Africa because much of the violence is group based. The critical questions are how groups organize and mobilize support. The urban areas have been the main arenas for social change, violent change etc., because they were the repositories of the evil and injustices of apartheid. But these studies have been more interested in collective political violence as it related to the anti-apartheid struggle. A more holistic approach is the methodology here.

683.2 The structure of South African urban areas in the historical context

69Having dwelt on the necessity and benefits of the urban violence framework, let us now consider the nature of urbanization in South Africa as a background to the more empirical aspects of this study. Officially, urban areas are defined as towns and cities with ‘urban facilities’ like water, electricity, communication facilities, schools, recreational facilities, etc., and which are under one form of local management or another (Cilliers and Groenewald, 1982). The number of urban areas in South Africa in this sense has risen astronomically from 45 towns in 1845 to 603 in 1951, a development which has been attributed to gold and diamond mining and the development of the railway system which rose from 3.2 kilometres in 1861 to 11,044 in 1910, 22,428 in 1967, and 23,447 in 1980.

70Relative to most other Third World countries, South Africa is highly urbanized, with over 50 per cent of the total population living in towns and cities. Saul and Gelb (1986) point out that the country has the highest levels of urbanization and proletarianization in sub-Saharan Africa and therefore the greatest revolutionary potential. According to the 1991 census, 78 per cent’ of the country’s population, excluding the former homelands (Cape, Natal, Orange Free State, and Transvaal), lived in urban areas, while 57 per cent of the population in the formerly non-independent homelands (Gazankulu, KaNgwane, KwaNdebele, KwaZulu, Lebowa, and QwaQwa) lived in urban areas, while 59 per cent did in the formerly independent black homelands (Transkei, Venda, Bophuthatswana, and Ciskei) (SAIRR, 1994: 85-86). From these figures, it is clear that South African urbanization is uneven, due largely to the years of apartheid policy which restricted the movement of blacks to urban areas and entrenched unequal development between the homelands and the ‘real’ South Africa.

71According to A Revised Urbanization Strategy for South Africa, an official document published in 1992, there are two major areas of uneveness in urban development. First, there is racial uneveness: whites (90 %), coloured (80.2 %) and Indians (94.5 %), who do not have villages of origin as do the blacks, are the most urbanized, while a large proportion of blacks continues to live in the rural areas. The proportion of black urban dwellers is only expected to rise to 49 per cent in 2010, although in absolute numbers they still form the vast majority of urban dwellers. The second area of uneveness is the concentration of the vast majority of urban dwellers in the Pretoria/Witwatersrand/Vereeniging (PWV) area. This area, which has been called the country’s ‘primary city’, because it is the industrial, commercial, and administrative centre of the country, had a population of 7.5 million or one-third of the total urban population in 1990. Expectedly, the PWV region has the greatest number of problems in terms of housing, provision of social services and racial inequalities; and has historically been the nerve-centre of political and non-political violence.

72The PWV is followed in size by the Durban/Pinetown/Pietermaritzburg axis which has a population of 3.4 million, the Cape Peninsula with 2.5 million, and Port Elizabeth/Uitenhage with 900,000. The official report also said that there were accelerated migrations from rural areas and from urban to bigger urban areas. The bigger urban areas are better cases for study because they best dramatize the complexities of violence. This is not to suggest that towns in the former homelands have been less violent – the townships in KwaZulu/Natal have been the most violent in the country since the transition to democracy. It is expected that a more substantial study will examine violence in these towns where the so-called black-on-black violence has mostly been played out.

73The most notable historical fact about these urban areas, which made them the centres of violence, is spatial segregation along racial lines, and the concentration of urban development in terms of housing, jobs, and provision of services in the white areas to the utter neglect of the black townships, and to a slightly lesser extent, coloured and Indian areas. Even police protection was restricted to whites only, in most cases, while urban management councils administered only white areas. Most towns and cities have two sharply contrasting parts: a white core, which is well planned and provided with essential services; and non-white peripheries which are further segregated along racial lines. The structure of Pretoria is typical of all other cities, with the exception of those in the former homelands. The core part of Pretoria is made up of Verwoerdburg and Akasia which are almost exclusively white, while the peripheral townships are Atteridgeville and Mamelodi (Black), Eversterus (coloured) and Laudium (Indian). The origins of spatial segregation date back to the system of racial discrimination which was established early in the country’s history, and specifically to the Natives Land Act of 1913 which severely limited the parts of the country where blacks could own land to less than 10 per cent; the Native Land and Trust Bill of 1936 increased this to about 13 per cent.

74But discrimination and segregation did not become systematized until the National Party took control of state power in 1948 and introduced apartheid (literally ‘separateness’). Existing discriminatory rules and practices were stepped up and new ones were introduced. In 1949, inter-racial marriages and sexual relations were declared illegal; the Natives’ Representative Council which previously advised on black affairs was abolished; and black workers were removed from the protection of employment insurance laws. In 1950, the Population Registration Act was enacted which classified all groups according to race. This was followed by the Group Areas Act which demarcated residential and business areas in all cities and towns along racial lines. The Group Areas Act which remained in force up until the 1980s led to the forced evacuation of the ‘wrong’ races from their lands and their resettlement in new locations.

75Other laws of segregation followed: the movement of blacks into and out of towns became highly restricted; the Bantu Education Act of 1953 removed education for black children from the department of education, arts and science, and placed it under the department of Native Affairs; the Reservation of Separate Amenities Act (1953) provided for the segregation of public facilities along racial lines; the Industrial Conciliation Act (1956) empowered the labour minister to designate jobs according to races and excluded black trade unions from collective bargaining; the Bantu Self-government Act of 1959 divided the black communities into 8 (later 10) territories; etc. The height of separate racial development was reached in 1971 with the enactment of the Bantu Homelands Constitutional Act which empowered the central government to establish self-governing homelands along the lines of the Transkei homeland model which had been established in 1963. Citizens of those homelands which became ‘independent’ – the TVBC states – lost their South African citizenship, which strengthened the restrictions on their movements to urban areas.

76Of the several laws of discrimination, two which were most consequential for urban violence were the Group Areas Act and the restrictions on black migration to the cities. Under the Group Areas Act, blacks were restricted to the townships where permitted black workers stayed, in many cases without their families. The townships, which are best characterized as satellites around the major cities, were essentially labour reserves where migrant labourers and other black workers who had been issued a pass and met other requirements lived. The largest and most historically famous township is the South Western Townships (Soweto) collective which comprises 12 continuous townships where, according to Odendaal (1991: 14) ‘Nearly one in ten black South African lives and works’. The influx of blacks into cities was restricted by the Blacks (Urban) Areas Consolidation Act 25 of 1945. The Act prohibited alien (non-South African) blacks from prescribed urban areas, and stipulated conditions which qualified non-alien blacks to stay in the cities. These requirements included (a) that he had been domiciled in the area for an uninterrupted period of 10 years during which he had worked uninterruptedly for one employer or, failing that, that he had lived in the area for 15 years, and (b) that he or she was the legal spouse or child of persons who met condition (a) and (c) that he or she had been granted permission to live in the area by a labour bureau.

77These rules produced the ‘urban blacks’, i.e., those who qualified for residence. All other black migrants were only allowed to stay in hostels without their families for the period when their labour was required. These were the ‘migrant labourers’. But the restrictive laws, zealously enforced as they were, could not stop the influx of blacks into the cities. Squatter camps or informal settlements, as they are also called, sprang up in the peripheries of most townships and cities to house the overflowing and homeless migrants. It was partly in acknowledgement of this failure and also the pressure from the business community which needed more relaxed labour laws that, following the recommendation of the Riekert Commission in 1979, the policy of ‘orderly urbanization’ was introduced. The essence of this policy was ‘a widening of the official definition of ‘settled urban’ Africans to embrace sections of the de facto urban African population in and around the cities’. The aim was ‘to limit the growth of established townships in the core metropolitan areas and encourage homeless township families, as well as some squatter families, to move to new residential areas called deconcentration areas … established on the peripheries of the metropoles’ (Hindson and Swilling, 1988).

78Pass laws were abolished, but the government descended heavily on ‘unapproved accommodation’ which led to the demolition of squatter camps in Langa, Crossroads and other places in 1986. Those uprooted were allowed to relocate to townships where they could erect the same shacks if they liked.

79The administration of black townships was in perpetual crisis. They were initially administered by adjacent municipal authorities, while some had elected black advisory boards and councils. In 1971, the administration of black townships was transferred to regional administrative boards in the Department of Cooperation and Development, but with little funding from the central government, these townships were neglected, as we discuss below. When, in 1971, they were removed from the jurisdiction of the local authorities for white municipalities and placed under black urban community boards and, from 1977, under elected black councils which were disenabled by legitimacy crises, their decay was complete.

80It is not possible to analyse the historical antecedents of the urban problems in South Africa exhaustively. Fortunately, these have been well covered in the historical narratives. What is of greater concern in this paper is the consequence of these historical processes in terms of the structured inequalities they created in the cities, which is arguably the basis for the widespread violence in the country. What strikes any first time visitor to Cape Town, Port Elizabeth, Johannesburg, and other major cities is the sharp contrast between the white suburbs and the black and coloured townships and squatter camps and the wide gaps in standards of living. As many South African commentators put it, the city is a combination of the first and third worlds (cf. The Commonwealth, 1986). The structure of violence also closely follows these lines of division. It is the townships and squatter camps that are the centres of violence. The white suburbs have, over the years, been virtually immune from serious violence, except inter-personal violence like murder, rape and robbery and isolated instances of terrorist attack both by the liberation comrades and the right-wing. One reason for this is that the white suburbs are well planned and their residents have been protected all along by the police, while the townships have not.

81The urban whites, Adam (1990: 130) says, are more worried about how to prevent ‘third world’ encroachments on their ‘first world’ privileges and retain their affluence than physical violence which poses no great threat: ‘So far, the daily life of most whites has hardly been affected by [violence]. Behind their fences and barred windows in exclusive suburbs, the good life continues’…Life in white suburbs is in sharp contrast to that in the townships and informal settlements where residents live in daily fear of violent attacks: If the gangs are not at war or unleashing terror, young girls are being raped; there are endless wars among taxi unions in which innocent commuters lose their lives; robbery and theft have become accepted ways of life; jungle justice ranging from public flogging to necklacing (i.e., placing a tyre around a person’s neck, dousing him with gasoline/kerosene and setting him on fire) is meted out to supposed criminals and public enemies by a so-called peoples’ court; and so on. Apart from sheer physical inequalities, differences in the quality of life and provision of services are even more serious.

82Let us look at a few indicators. Although the figures are not exclusively for urban areas, they reflect the disparities well enough. I should also add that the figures tell us nothing of intra-racial group inequalities, which are equally high, especially amongst the blacks and the coloureds. For example, while the average annual household income for the wealthiest 20 per cent of blacks rose from 24,780 rands in 1975 to 34,243 rands in 1991, that of the poorest 40 per cent of blacks declined from 3,048 rands in 1975 to 1,784 rands in 1991 (see report of study by the Human Science Research Council and the University of Natal in Sunday Times, 27 March, 1994). Such rising intra-group inequalities exacerbate feelings of relative deprivation, but it is the more underlying inequalities produced by the years of apartheid that we are concerned with.

83In the area of health, only 7 per cent of the blacks were covered by medical aid schemes in 1991, whereas 69 per cent of the whites, 33 per cent of the Indians and 29 per cent of the coloureds were covered. Infant mortality rates for the period 1985-1990 were 74 per cent for the blacks, 47 per cent for the coloureds, 16.4 per cent for the Indians, and 10.8 per cent for the whites. Life expectancy figures were 70.3 years for the whites, 64.8 years for the Indians, 58.5 years for coloureds, and 57.5 years for the blacks. Child welfare budgets in non-homeland urban areas in 1993/94 went 50 per cent to coloureds who constituted only 15 per cent of the non-homeland population, 15 per cent to whites who form 23 per cent of the population, 8 per cent to Indians who are 5 per cent of the population, and 26 per cent to the blacks who constitute 57 per cent of the population.

84In terms of education, most of the primary, secondary and tertiary institutions are in the towns. Pupil-teacher ratios for the racial groups in 1984 were 1: 23 for Indians, 1: 26 for coloureds, 1: 19 for whites, and 1: 41 for blacks; the percentages of unqualified teachers were 2.3 in white schools, 8 in coloured schools, 11.9 in Indian schools, and 16.9 in black schools; and the percentages of students who gained entrance into university were 9.3 for blacks, 16.1 for coloureds, 45.0 for Indians, and 48.7 for whites (HSRC, 1985). Moreover, while the state funded white education adequately, schools were not only in short supply for black students, they were nowhere near the white schools in number.

85In terms of housing and amenities, the number of people, almost all of whom were black, living in informal settlements in PWV was over 2.8 million in 1992, while about 2,500 were homeless in central Johannesburg alone; about 80 per cent of the blacks living in greater Cape Town lived in informal housing; less than 20 per cent of houses in black townships had access to potable water and electricity in 1993. In terms of wages and means, comparative monthly salaries for public servants in the central government in June 1992 were 2,149 rands for black workers, 2,756 for coloureds, 4,004 for Indians, and 3,991 for whites, while in universities and technikons it was 2,014 rands for blacks, 3,112 for coloureds, 3,664 for Indians, and 5,435 for whites; in Port Elizabeth, a study conducted in 1993 showed that 95,000 black households lived below the poverty line of 1,200 rands per month compared to 22,600 coloureds, and 8,300 white households, while national urban averages were 32 per cent for blacks, 23 per cent coloureds, and 1 per cent whites (SAIRR, 1994). A recent study of urban policy in South Africa has rightly concluded that if these trends continue, ‘most urban blacks in the year 2000 will be poor, incapable for economic reasons of constructing anything but the most rudimentary shelters for themselves, have low levels of formal education, and poor access to health and other urban facilities’ (IDRC, 1992: 1-2).

3.3 The mobilizers of violence

86For proponents of relative deprivation, these structural inequalities which were greatly reinforced by discrimination against blacks and other non-whites, who were prohibited from entering or using ‘whites only’ beaches, hotels and bars, schools and so on, provide an enabling environment for violence and revolutionary action, especially because they are the product of an illegitimate system. But as we pointed out earlier on, these background conditions are only a necessary, not a sufficient reason for violence. The factors which facilitate violence or what I refer to here as mobilizers provide the key to accounting for it. These factors are discussed under the following headings. I should stress by way of reiteration that apartheid was most dramatically played out in the cities and that, consequently, the violent consequences of the system were most visible in them.

3.3.1 State violence and repression

87Apartheid was not a system to which the non-whites, especially the blacks, acquiesced; the system was entrenched and sustained by force. The extent of state violence can only be appreciated when it is realized that the sophisticated defence and security systems which the state developed and which makes South Africa one of the major defence centres of the world, were not in response to any serious external threat but were put in place to crush the liberation movements and other opponents of the regime. Put simply, the apartheid state was a military state. From the very beginning, urban areas, where the consequences of apartheid were most directly felt, were the centres of resistance. Initial resistance to racial discrimination and inequalities by organizations like the ANC (which was formed in 1912) were passive and peaceful. Even after apartheid became fully entrenched in 1948, the ANC only agreed to boycotts, strikes, and other forms of civil disobedience under pressure from its more radical youth league soon after the pass laws and other laws were put in place.

88The ANC’s major strategy remained passive and non-violent. In 1955, along with coloured and Asian anti-racist groups, it produced the Freedom Charter in which it demanded non-racial government. The state’s response was to detain and imprison the leaders and enact more laws to repress the organizations and bar public assembly. The intolerant character of the state was unacceptable to more radical and militant elements who, led by Robert Sobukwe, broke away from the ANC to form the PAC in 1958. It was the anti-pass laws demonstration which the PAC organized on March 21 1960, to which the state responded with the Sharpeville massacre that marked the turning point in the history and character of the liberation struggle in the townships. Following the massacres, the government banned the ANC and PAC, and arrested most of their leaders, while many went underground and others fled to neighbouring countries. Thereafter, the armed wing of the ANC – Umkonto we Sizwe (MK) – and PAC’s Poqo embarked on the sabotage of public utilities while organizations like the African Resistance Movement, a white group, joined in blowing up train stations, bridges, and radio towers. The police infiltrated the groups and made several mass arrests – by the end of 1965 about 1300 people had been sentenced to average jail terms of 7 years. When at his famous trial in 1964 Mandela stated that the regime had made peaceful settlement impossible, he was merely stating the obvious: The long reign of political violence had been launched.

89Violent struggle was inevitable because the laws practically abolished all forms of political organization and agitation. Press reports of police actions were prohibited; the Riotous Assembly Act of 1956 prohibited the gathering of 12 or more people; the Sabotage Act of 1962, which carried sentences ranging from 5 years imprisonment to death, defined sabotage as any wilful act which threatened public safety and health or interfered with public utilities. The Suppression of Communism Act, which was replaced by the Internal Security Act of 1976 imposed prison sentences ranging from 1 to 10 years on anyone who advocated, defended, advised, or encouraged any object of communism; the minister of justice was empowered to ban anyone engaged in activities considered dangerous to law and order from particular towns, or restrict him to certain localities for a period of up to 5 years. The Terrorism Act of 1967 spelt out penalties ranging from 5 years imprisonment to death for any activity broadly defined as likely ‘to endanger the maintenance of law and order’ and empowered the police to arrest and detain anyone suspected of being a terrorist for an indefinite period pending investigation; etc.

90It was the turn of the students and the youths to take on the state in the 1970s and 1980s, after their parents had been silenced. Leaders like Steve Biko organized black consciousness and student bodies which conscientized the youth and mobilized them to revolutionary action. Armed with the slogan ‘No education before liberation’, the comrades as the youths came to be known, were prepared to take on the state and engage in actions which furthered their goals, including murder and gangsterism. In 1976, they demonstrated their ability to undertake mass action with the Soweto riots whose immediate cause was resistance to the use of Afrikaans in their schools. The riots spread quickly to Cape Town and other major cities with students forcing black workers out of work and crippling the industries. The police responded very ferociously: after 3 months of continuous rioting, according to the Cillie Commission which was appointed to investigate the riots, 575 people had been killed, 3907 injured, and thousands arrested. Again in 1984, there was another violent nationwide students/youths uprising. This time, widespread opposition to the introduction of a tricameral legislature provided the enabling environment, but students who went on the rampage in Atterridgeville, near Pretoria where it began, were protesting against corporal punishment and sexual harassment. The protests spread throughout Transvaal, Eastern and Western Cape townships where the opposition to tricameral legislature was most serious. The protests were crushed by the defence forces, which in Sebokeng township, engaged the youths in a street ‘civil war’. Youths were killed, imprisoned and detained. It is important to point out that the determination of the youth to overthrow the apartheid state kept most of them out of school for the period beginning from 1976. In fact, though schools in the homeland territories continued to have a semblance of normal education, those in Soweto and other PWV townships were unstable. It is not surprising therefore, that one of Mandela’s first appeals on becoming president was for the township youths to return to school. This was heeded, as schools in Soweto recorded 100 per cent attendance for the first time since 1976 (Sunday Times, 29 May, 1994). For the period they were away from school, the students were virtually out in the streets providing recruits for self-defence units, criminal gangs and other violence-related groups.

91Increasing anti-state attacks which followed the strategy of making the country ungovernable were adopted by the ANC and UDF. This led the state to declare a state of emergency and ‘Operation Total Onslaught’ in the 1980s. These steps brought the defence and security forces fully into the battle to sustain the apartheid state; as a prelude, the South African Defence Force’s (SADF) official mandate was amended in 1976 from ‘service in the defence of the republic’ to ‘operations as the defence of the republic’. In October 1984, army units patrolled Soweto streets for the first time and 7000 of them were deployed under ‘Operation Paliet’ to carry out house-to-house searches in the township of Sebokeng. Their roles were later extended to include eviction of rent defaulters and stoppage of school boycotts and strikes (Cock, 1990).

92The deployment of state violence was not, however, restricted to the apartheid state. In the transition to the post-apartheid state, the policy of declaring troubled spots unrest areas and imposing states of emergency, which brought affected areas under virtual military rule continued to be used, ostensibly to protect public safety. In March 1994, a state of emergency was declared in KwaZulu Natal which led to the deployment of SADF soldiers to the province. Battle-ready troops of the short-lived National Peace-Keeping Force and SADF have also been deployed to East Rand townships of Tokoza and Athlone. The use of the military to quell violence may be taken to be a measure of the seriousness of the situation, but it tends to exacerbate conflicts because some township comrades still see the soldiers as defenders of apartheid and supporters of the enemy. Indeed, SADF soldiers have been accused by both IFP and ANC members of supporting the other, while the NPKF was unsuccessful in its peace-keeping operations largely because IFP members saw it as an ANC instrument. So, rather than reduce violence, the deployment of troops has in some cases exacerbated it.

3.3.2 State manipulation of challengers

93Alongside the stepping up of its military strength, the state tried, with varying degrees of success, to break the ranks of the comrades and liberation organizations. From the beginning, a policy of encouraging ethnic divisions, especially between the Zulus and Xhosas was pursued. In furtherance of this, accommodation in the hostels was allocated according to ethnic origin. After the formation of Inkatha, whose leader, Mangosuthu Buthelezi, was favourably disposed towards the state, Zulu hostel dwellers were encouraged to oppose mass actions like stay-aways, consumer boycotts and demonstrations, and there have been allegations that the police supplied hostel dwellers with guns to fight the comrades. It is instructive to note that violent clashes in the townships have mostly taken the form of clashes between hostel and township dwellers. The state also infiltrated the townships through the vigilantes who were informants and lackeys of the security agencies planted to identify activists and destabilize underground liberation movements (cf. Haysom, 1989; Levin, 1987). The 1970s and 1980s saw battles between comrades and vigilantes in which many of the latter were necklaced. At some stage, the association of vigilantes with Inkatha exacerbated the conflicts between ANC and Inkatha supporters.

94The most popular notion of state manipulation and complicity in violence is what has appropriately come to be known as the third force theory. Simply put, the ‘theory’ attributes escalating black-on-black violence to destabilizing activities of the apartheid state using the police and defence forces. Specifically, proponents of the theory such as Kane-Berman (1993), allude to the arming and instigation of IFP cadres and hit squads to attack and kill ANC leaders and members. Others, however, argue that the slant of the theory is too pro-ANC and suggest a need to examine how state forces have also been used against the IFP (Adam and Moodley, 1993: 131). This manipulative strategy was originally aimed at weakening the liberation movements and keeping the majority of the blacks permanently divided, but it has since been exploited to serve the interests of the elites in their struggle for power. In particular, third force activities have been linked with right-wing elements who controlled the semi-independent military intelligence agencies like the Civil Cooperation Bureau which masterminded the killings of liberation activists – as has been revealed by inquests into the murders of Goniwe and Lubowski and Webster. Allegations by Nelson Mandela of third force activity operating within the police and security agents gained credibility in August 1990 when, soon after the death knell for apartheid was sounded, there followed unprecedented train massacres, attacks by hostel dwellers and drive-by shootings on the Reef and KwaZulu/Natal. The government promptly denied the allegation and accused Mandela of wanting to cover up ANC’s role in the violence. But the release of the Goldstone Commission on the Prevention of Public Violence and Intimidation reports in April 1992 and March 1994 put paid to all the speculations and denials. The Commission found evidence of the involvement of top officers of the South African Police (SAP) in large-scale destabilization schemes involving the training and deployment of individuals and hit squads, gun-running, and fomentation of violence, in collaboration with the IFP. The report specifically talked of the creation of ‘a violent murder machine’ deployed in train and hostel violence. Another investigation conducted by a task force of the Transitional Executive Committee into violence in KwaZulu found evidence of collusion between the KwaZulu Police and SAP in hit squad activities which dated back to 1986 when the South African Defence Force trained a nucleus hit squad in the Caprivi Strip under the guise of preparing them for VIP protection.

3.3.3 The rise of militant civic organizations

95The critical role of civic groups in urban revolutions and violence has been emphasized by the social movement school of urban sociology as well as proponents of the strategic and structural schools of revolution. Level of organization, leadership, available resources, ideological coherence, and linkage with other groups have been identified as some of the key variables affecting the ability of social movements to undertake revolutionary action. The banning of the ANC and PAC led to the growth of several militant and civic organizations which emerged to cater for narrow and general interests. Students and black consciousness organizations were particularly notable in this regard. Such groups included the South African Students Organization, Black Peoples’ Convention, South African Students Movement (which organized the Soweto riots), and the Azanian Peoples Organization. There were also civic organizations whose nucleus was the Soweto Committee of Ten which was formed to demand rights to land and other facilities for non-whites in the townships. Civic organizations organized rent and consumer boycotts among others, and joined forces with the youths and workers in other mass actions. Then in 1984, the United Democratic Front (UDF) was formed to coordinate the activities of the various social movements, acting as the internal arm of the ANC which continued to wage the battle from outside. A special category of these militant organizations was made up of the trade unions which were involved in resistance and liberation struggles from the very beginning. Their struggles were particularly consequential because they were directed against both the state and capital. These organizations were not, however, allowed free rein. In February 1988, for example, UDF, AZAPO, and sixteen other organizations were banned, while COSATU was prohibited from political activities. Their leaders were not spared either. As the security forces descended on these organizations and detained, tortured and killed their leaders, political violence became more legitimate.

96Right-wing organizations, which have defended the apartheid system, opposed black majority rule and, since 1990, have demanded an autonomous Afrikaaner republic, constitute another category of militant groups in the urban areas. The major right-wing organizations (whose main strategy is terrorist attacks) include the Afrikaaner Weerstandsbeweging (AWB), the Boere Weerstandsbeweging, and the Wit Wolwe. Although they are not as mass based as the black social movements, these organizations have recorded great successes in terrorist activities, especially in the bombing of public places, because their members include well-trained police and military officers both serving and retired. They also possess adequate financial backing from wealthy farmers and superior know-how. Thus, it has been possible for them to organize a series of private armies and a network of radio stations which facilitate easy mobilization (as was clearly demonstrated when Radio Pretoria was used to mobilize the Boere army for action in Bophuthatswana in March 1994).

3.4 The legitimation and generalization of violence

97Earlier in this study, we discussed how the approval given to acts of violence by the liberation movements and their leaders legitimized violence. For example, the slogan of APLA, the armed wing of the PAC – One settler (Boer) one bullet – gave exoneration for the killing of whites, while Winnie Mandela’s famous match box and petrol speech provided the justification necklacers needed. Even the leader of the IFP, Mangosuthu Buthelezi, who argues that the organization did not believe in armed struggle, nevertheless rationalized violence by its members on the grounds of self-defence. But it was not only black liberation movements that approved of violence. The Afrikaaner right-wing pronounced death to ‘kaffirs’ and ‘slaves’ and generally, across racial lines, there were cross fires of violence. The Wit Wolwe, a right-wing Afrikaaner nationalist group declared in 1993 that it regarded ‘every black man as an enemy’, and later claimed responsibility for the killing of a 5 year old girl in Nigel on the East Rand in retaliation for the attacks on whites in Eikenhof, south of Johannesburg. That same year, there were rumours that the Afrikaaner Weerstandsbeweging (AWB) had created a ‘retribution unit’ to carry out reprisals for the murders of Afrikaaners. In the campaigns for the historic 1994 elections, the National Party played on fears of racial violence in the Western Cape by alleging that under an ANC government, the slogan would be ‘Kill a coloured, kill a farmer’. In addition to these generalized rationalizations of violence, the successes of armed liberation struggles in neighbouring Zimbabwe, Angola, and Mozambique, with which the liberation struggle in South Africa was intricately linked, for which reason South African defence forces carried the suppression of its challengers to these countries, convinced the comrades of the rightness of their actions. These justifications for violence extended to otherwise criminal acts such as robbery, rape and murder, and it is interesting to see that demands for post-apartheid indemnity by members of the liberation movements and right-wing groups have been based on the political objectives of violence.

98The role of the mass media in the generalization of violence should be emphasized. The newspapers, radio and television, through the wide coverage given to violence and, in some cases, the racial partisanship of coverage, have fanned the embers of violence in several ways. Radio Pretoria and other right-wing radio stations, for example, broadcast programmes of racial hatred against blacks, and were used to mobilize Afrikaaner commandos for action in Bophuthatswana. Some authors have also identified movies and other programmes on television which glorify violence as mobilizing factors (cf. Halloran, 1981; Posel, 1990).

3.5 The politicization of the police and security forces

99The police is the state’s agency responsible for crime prevention and public safety. It is therefore expected to play a critical role in minimizing violence. But this was not the case in South Africa where the police force was a major instrument of the perpetuation of the apartheid state. The police were thoroughly militarized; its major function was to destroy the liberation movements, which gave it little time to address issues of crime prevention (for changing roles of the police, see Cawthra, 1992). But even where it did, it had a duty to protect only whites, and its presence in the townships and other non-white neighbourhoods was for the purpose of arresting enemies of the state or terrorizing the people. Consequently, even if it wanted to, the police, not to talk of other security forces, lacked the credibility or legitimacy to control violence in the townships. In fact, the policeman was seen as a symbol of apartheid, to be attacked and killed.

100This problem has persisted in the post-apartheid era, even after the police force has repented of its ignoble past and warmed itself up to the people as their ‘friends’. According to police sources, there were 4,165 cases of unrest-related violence against policemen, resulting in the deaths of 279 police; 35 per cent of these were in Witwatersrand, and particularly the East Rand, which has become the centre of black-on-black violence. Within the first quarter of 1994, 606 policemen were attacked on the Reef alone, resulting in 55 dead and 124 wounded (The Argus, 9 June 1994). Acknowledging the historical context of these attacks and appealing for a change of perception, the police spokesman said The political playing field has changed drastically and the police are now in the service of a democratic South Africa. Attacks on the police are, therefore, totally unacceptable’ (The Argus, ibid.).

101It will take a long time for the township dwellers to accept the police as politically neutral as they continue to be dominated by white officers. There are still allegations of racial discrimination in their approach to policing, of illegal methods like assassination, torture and surveillance which were carried over from the days of a fully militarized force. Their complicity in violence, revealed by the Goldstone Commission referred to earlier, has not helped the issue of their credibility.

102The major effect of the absence of police protection in the townships was the emergence of autonomous or parallel protection and judicial structures. There are the Self-Defence Units (SDUs) which are associated with ANC cadres especially in Soweto, and the Self-Protection Units (SPUs) formed by IFP cadres in the hostels and KwaZulu/Natal townships, both of which are weapon-carrying armies. The SDUs and SPUs have been in the forefront of township wars, as well as battles with police and security forces. In-fighting within the units has also been a major source of conflict. For example, many of the deaths from violence in Tokoza township are attributable to factional fights within the SDUs (The Argus, 17 May, 1994). Next we have the people’s courts which exist on most streets in the townships. These clandestine courts, which have sometimes been used to deal with enemies and political foes, try all cases ranging from rape and assault to sabotage and murder and deliver judgements which range from public caning to necklacing. Community and neighbourhood watch associations have also been formed to organize collective protection for residents of particular areas. In Cape Town, these groups have been involved in the fight against criminal gangs. Lack of police control has also given rise to gangs, various drug pushers, and warlords who offer protection to owners of shebeens (speak-easys) and other township dwellers. Taxi wars among rival members of rival taxi unions also exist because there is no police protection. Since the taxi business was unregulated for a long time, the unions took the law into their own hands and settled disputes by street/taxi rank shoot-outs and other violent means.

3.6 Social dislocation

103Several authors have pointed to the social dislocations caused by the urbanization policies of the apartheid state as a major instigation of violence. The acute shortage of houses in cities for non-whites, especially blacks, is one of the root causes of social dislocations. The hostel system which prevented migrant workers from living with their families, took from these enforced ‘bachelors’, the controls="true" and inhibitions of normal family life. Similarly, the forcible evacuation of non-whites under the group areas act and their relocation to places where accommodation was in very short supply destabilized family cohesion and control. In particular, many authors have pointed to the eviction of coloureds from District 6 in Cape Town as being responsible for the proliferation of gangs in the Cape Flats. Then there were the inter-generational crises which followed the ascendancy of the youths as leaders of the liberation struggles. Johnson (1988: 96) alludes to the generational consciousness, analogous to class consciousness, which underlay this ascendancy as arising from extreme poverty, little prospect of future employment and a better life, and the intensive politicization of society. Driven from home by poverty and inadequate accommodation, and from school by boycotts and demonstrations, the youth found solace in peer group activities outside the control of parents who had been relegated to the background in the new perception of power. Most of them had guns and, after the deaths of several friends and loved ones, death did not seem to mean much anymore. Driven by anger to fight an unjust system, they became the forces of violence. As the Sowetan put it in an editorial:

Let there be no illusions about this: the youths in the townships are in an unstoppable mood of anger…having seen disregard for their lives…most parents and a number of very respectable leaders are aware that they have lost control of the anger expressed by township children. (cited in Johnson, 1988: 112)

3.7 The struggle for political power

104Central to most political conflicts is the struggle for power. The South African situation has been no different. The liberation struggles and attendant political violence were not only aimed at crushing apartheid; they also aimed at the capture of state power. Right-wing terrorism is aimed at preserving Afrikaaner predominance and power. But the most dramatic link between the struggle for power and violence is to be found in the violent conflicts between the IFP and the ANC. The mobilization of Zulu nationalism through cultural revivalism, conscription into the IFP and the manipulation of their king; the establishment and training of a secret Zulu army, as well as collusion with state security and police forces were all part of the strategy of IFP leader, Mangosuthu Buthelezi, to extend his sphere of influence beyond KwaZulu and challenge the ANC’s inevitable domination of state power. The success of the IFP’s ethnic mobilization accounts for the readiness with which hostel dwellers and militant groups in KwaZulu/Natal are prepared to confront ANC supporters, sometimes in the name of Zulu survival. As one of the survivors of the March 1994 clash between ANC and IFP demonstrators in Johannesburg said in a television interview, they were told to march for Zululand, not IFP. But underneath Zulu nationalism is the IFP’s struggle for power, and especially Buthelezi’s desperate efforts to keep his political constituency intact. The ANC challenge to his hegemonic control of KwaZulu has therefore been met with serious violence, notwithstanding the fact that members of the ANC in the area are themselves Zulu.

3.8 A society in arms

105Most adult South Africans have personal guns. While a good number of them do so legally – they have firearm licences – most of the guns and other assault weapons in the townships were acquired illegally from various sources. Most of them are weapons which were given to soldiers of MK and other liberation armies. It was decided in April 1994 that these unlicensed weapons should be turned in to designated assembly points. Some were smuggled across the borders from Mozambique and Angola where civil wars have raged for several years. Others were obtained from the police and security forces which the Goldstone Commission found to have been involved in third force activities, and whose officers are claimed by criminal gang leaders in Cape Town, to be the main sources of weapons. Then there are the local arms factories and dealers in arms and ammunition whom the PWV provincial government has accused of fuelling the violence in the townships to maximize their profits. According to a spokesman for the government, ‘Gun and bullet sellers view peace in the area as a death knell for their flourishing business’ (The Argus, 5 July, 1994).

106Assault weapons, especially AK47s, are easily available in the townships, and can be bought for as little as 50 rands. The range of weapons is frightening, and gives an indication of the intensity of the violence. For example, in a raid on three migrant workers’ hostels in Vosloorus on the East Rand in May 1994, the police seized 13 AK47 rifles, 2 detonators, 25 tear gas canisters, 3 hand-grenades, over 1,000 rounds of ammunition, 12 magazines, and 2 RPG rockets (Sunday Times, 15 May 1994). The second interim report of the Goldstone Commission in 1992 identified the assemblage of firearms in hostels, squatter camps and townships as a key factor in the escalation of violence. There are also, in addition to these, traditional weapons, mainly clubs, spears, machetes, axes and defence shields which are carried freely in public gatherings and protest marches, mostly by Zulus.

107It is not only in the townships that illegal firearms are to be found. Policemen are attacked daily and their guns stolen, while soldiers and policemen have been alleged to sell reserve weapons, bought from the forces, to members of the public. Private security firms and so-called private armies constitute another major source of weapons, which though legally acquired, have nevertheless been put to uses that exacerbate violence. For example, some of the security firms are associated with right-wing elements, and have been known to be involved in training and arming terrorist squads. Then there are the private armies which are allowed free rein. In addition to APLA and MK which are now integrated into the new South African National Defence Force, there are the right-wing armies, notably the Wenkommando which is the para-military wing of the Afrikaaner Weerstandsbeweging (AWB) which AWB leader Eugene TerreBlanche said was a self-defence unit established to protect Afrikaaners and their property against the ANC and APLA. Other right-wing armies include the Boer Republican Army which carried out the bombing of a Johannesburg’s magistrate court in 1993 and organized to defend the Lucas Mangope government in the former homeland of Bophuthatswana in March 1994; the Oos Transvaal Boerekommando which broke away from the AWB’s military wing in 1992; and the Pretoria Boerekommandogroep which operates along the lines of a community guard (SAIRR, 1994). The command structures of these right-wing armies are controlled by retired generals and former ‘crack officers’ of the South African Defence Force. The existence of such armies, including the SDUs, and the relative ease with which firearms can be acquired and carried by individuals and groups, explain the quick degeneration of common conflicts into violent ones. At the inter-personal level especially, the use of guns to settle matrimonial and family disputes and even commonplace disagreements in drinking places is the direct consequence of the fact that South African society is an armed society.

108The foregoing mobilizers of violence provide general explanations for urban violence against a backdrop of factors which were examined in the previous section. In particular, they explain why violence has tended to be endemic and generalized, and why a holistic approach is necessary. To conclude this section, the point should be made that once it has crossed the threshold, violence tends to become generalized. Guns that were acquired for liberation for example, can very easily be put to criminal uses, while latent violent tendencies get full expression. This has certainly been the case in South Africa.

4. Aspects of Urban Violence in South Africa

4.1 A methodological note

109Violence is a rather elusive phenomenon for empirical study. Even where the background and facilitating factors are known, it remains an unpredictable phenomenon. Not only is it difficult to know ahead of time that violence will occur (it is easier to predict collective violence like students’ riots than criminal violence like armed robbery) but even when one is present when the violence actually occurs, the underlying factors are rarely open to direct observation. This is why, as I have emphasized, a contextual analysis which hinges on the historical approach is necessary. But this is not the only problem involved in the empirical study of violence. A far more serious problem has to do with the fact that every incidence of violence has its own unique history and context, meaning that there is a limit to the extent to which it can be generalized. The methodological implication of this is that background and mobilizing factors are not always sufficient to explain specific cases of violence.

110According to Schlemmer (1991), we still need to know the factors that ‘trigger off specific cases of violence. This is why the case-study method is well-suited to the study of violence. It enables the researcher to empirically anchor contextual analysis by relating background/facilitative factors, which are usually played out in the specific cases, to the more immediate factors, and to search for more complete explanations. But in selecting the cases, care has to be taken in order to adequately reflect the major forms of violence present in the society to be able to deal with the complex forms they take. A focus on exceptional cases could create problems of overdeterminism. It is against this background that we focus on two cases of urban violence in South Africa:

• so-called black-on-black violence as represented by the Johannesburg massacre of March 1994
• criminal gang violence in the Cape Flats

4.2 Case studies

4.2.1 The Johannesburg Shell House massacre

111On March 28th 1994, the IFP planned to hold a rally at the Library Gardens in central Johannesburg, about eight blocks away from Shell House where the ANC has its national headquarters. The rally was part of the show of strength through mass action engaged in by IFP and ANC supporters as campaigns for the April elections gathered momentum. Although IFP leaders were still adamant at that stage that they would not take part in the elections, this did little to reduce the power struggle they had been having with the ANC since the 1980s. The March rally which involved over 40,000 IFP supporters was staged to demonstrate support for the call on the Zulus by their king to uphold the kingdom’s ‘sovereignty’ declared by him. As the IFP supporters, who arrived by train and bus from the hostels on the Reef, marched through Plein Street where the ANC headquarters is located, all hell broke loose. Shots were fired at them from Shell house and adjoining buildings; including Customs House, according to eye witnesses. ANC officials claim shots were only fired by their security forces in self defence after the marchers had opened shooting. The truth about who started the provocation may never be known, although official investigations were still being conducted by the police at the time of this writing; but as a result of the cross-fire, at least 52 people were killed, 11 of them at the hands of ANC security officers, and over 100 injured.

112The Johannesburg massacre as it came to be known, aptly demonstrates one of the characteristics of violence referred to earlier, namely, its ultimate unpredictability. While the escalating violent conflicts between IFP and ANC supporters in KwaZulu/Natal and the Reef could have led one to expect conflicts, the fact that the IFP march/rally was supposed to be a peaceful one, for which permission had been granted officially, and moreover, that the police force was closely monitoring it, made a violent outcome unlikely. Who ever expected the marchers to march on Plein Street or, as ANC officials claim, attempt to attack their headquarters? Did the ANC security officials open fire on the protesters in self-defence? or Was the action premeditated, as suggested by accounts of survivors who claimed that those who opened fire on the protesters were well positioned on adjoining roof tops? Were ANC security agents the only gunmen, or were there others who cashed in on the confusion? Why was the police force whose presence was quite visible not able to control the situation by, for example, preventing the protesters from marching through Plein Street? These are empirical questions which can only be answered as suggested earlier, by linking background and mobilizing factors to the immediate factors at the time of the massacre. Tentative explanations along these lines are discussed in the following paragraphs.

113In terms of background, the Johannesburg massacre was the high point of violent conflicts between supporters of the ANC and IFP which had reached unprecedented levels in the first quarter of 1994. The townships of KwaZulu/Natal were the main centres of violence. Between January and April over 1000 deaths were recorded in those townships, with 311 deaths in March and 552 in April, doubling the monthly averages for 1993; these conflicts spilled over to the PWV townships where mostly Zulu and IFP-supporting hostel dwellers confronted ANC supporters. This spill-over was partly the consequence of attempts to extend the IFP’s sphere of influence beyond KwaZulu/Natal. To do this, confrontation with ANC supporters was inevitable, given the long history of struggle for supremacy between the two groups in the townships. A study group of the Human Rights Committee, which monitored the escalating violence, attributed it to well planned strategies by IFP factions opposed to the election. The strategies involved training and arming ‘commandos’ (mostly members of the SPUs, at secret IFP military camps), deployment of hit squads into ANC areas, disruption of ANC rallies and preventing the ANC from using campaign venues (Sunday Times, 5 June, 1994). In another report, the Transitional Executive Council also pointed to the role of terrorist groups trained by the KwaZulu police with the assistance of senior officials of the KwaZulu government and the South African Police who helped to procure weapons – AK47s, Makarovs, Uzis, pump action shotguns, limpet mines, pistols, and Rl rifles.

114The background to the unprecedented ‘black-on-black’ violence in the run-off to the election is provided by the intensified struggle for political power between the ANC and IFP; particularly because the IFP looked certain to lose the privileges enjoyed, when it was the only major black ‘nationalist’ group allowed to operate in the country in the days of apartheid. The close association of its leadership with the apartheid regime, and its objection to the armed liberation struggle (which objection naturally was actively encouraged by the regime) put it on a collision course with ANC comrades in the townships. In fact, since the late 1970s, when IFP supporters sabotaged mass actions like workers’ strikes and consumer boycotts and vigilantes and unpopular black councillors became associated with the IFP movement, there have been violent clashes with ANC supporters. The decision by the IFP initially to boycott the April elections; the fear of marginalization under an ANC government which made it demand not only regional autonomy but even a sovereign Zululand; the formation of an alliance with right-wing groups, the adoption of the strategy of ethnic mobilization which involved exploiting the Zulu king’s spiritual, traditional and cultural controls="true" over the people, and the preparation for war with the ANC, all intensified violent conflict between supporters of the two black parties.

115Of these mobilizing factors, that of ethnic mobilization fits in with the stereotype of African ‘tribalism’. By tying IFP fortunes with those of the Zulus as a people, Buthelezi reinforced this appearance. Indeed, one survivor of the Johannesburg massacre said in an interview that they were told to march for the love of Zululand, and not as IFP supporters. Also, the trainer of IFP commandos at a military camp near the Mozambican border, who was formerly an intelligence officer with the South African Police, said the fight was for Zulu accommodation in the new South Africa, an objective he very clearly linked to party politics: ‘White rule is finished here, but instead of moving to a multiparty state, we are heading towards a one-party state. Why should the Zulus be left defenceless?’ (The Argus, 18 April 1994). So, was the violence which led up to the Johannesburg incident ethnic? To the extent that many Zulus saw the conflict as one between them and the Xhosas who predominate in the ANC, there is a sense in which it could be so regarded. But to the extent that the ANC enjoys strong support among the Zulus – in fact the violence in KwaZulu/Natal is between Zulu supporters of ANC and IFP – and is not an ethnic movement, and given the long history of conflicts between members of the two groups, the ethnic explanation to be meaningful, has to be situated within the larger context of inter-elite struggles for power, and divisions accentuated by the colonial state.

116The Johannesburg massacre was facilitated and aggravated by other immediate factors. One was the fact that both the protesters and the ANC security officials were armed with traditional weapons (assegais and spears) and assault rifles. Another was that the police failed to prevent the protesters from carrying arms and marching through Plein Street, where the ANC headquarters is located; and holding its rally so close to Shell House. According to the ANC, the policemen who had been present all along simply disappeared at the critical moment that the marchers began to shoot at security men, and only re-appeared after the massacres had taken place. The lack of credibility of the police was definitely a key factor in its inability to prevent the massacre.

4.2.2 Gangsterism in the Cape Flats

117Most major towns and cities have formal and informal criminal groups formed by youths mostly residing in slum areas, where they are socialized into ‘deviant’ sub-cultures. Indeed, gangs are one of the most distinctly urban forms of violence, as a consequence of factors which are entrenched in urban areas. They are generally regarded as being formal or informal anti-social or delinquent groups, which frequently engage in conflicts and crimes (cf. Klein, 1971; Cartwright et al., 1975). According to Smith (1988: 14), gangs are different from mobs and other fleeting or loose delinquent associations whose members are usually unknown to each other. The distinct mark of gangs is their relative permanence and the development of peculiar sub-cultures. The existence of gangs and their propensity for violence are facilitated by systemic factors like family dislocation, unemployment, alienation, restricted opportunities for upward mobility, legitimation and generalization of violence, accessibility to firearms, and the ineffectiveness of, in some cases, collusion with the police. As all of these factors have been present in the South African situation, it is not surprising that there is a proliferation of criminal gangs in most cities, especially in the townships. Many of these gangs were originally formed by ‘street youths’ who dropped out of school (most comrades especially in PWV townships abandoned school in the 1970s and 1980s), could not remain in over-crowded homes, and were out of parental control. From being groups of naughty neighbourhood boys and perpetrators of petty theft and rape, these gangs steadily graduated into formidable groups of criminals which terrorize neighbourhoods, lure and conscript young men and women into their ranks. They are behind most urban neighbourhood crimes including drug and gun-running, have protective connections with the police, engage in shoot-outs with rival gangs, and kill opponents with impunity (see Pinnock, 1984; Scharf, 1989). They have developed very sophisticated forms of organization which include language and dress codes, and have spread their tentacles even to the prisons (for prison gangs, see Haysom, 1981). The existence of these dreaded gangs which operate in the open and in defiance of police and security controls="true" is one of the most problematic aspects of urban violence in the country today. This is especially so in Cape Town, and specifically the Cape Flats,where over 80,000 of the youths are estimated to belong to gangs (Pinnock, 1984).

118There are over 150 gangs in Cape peninsula, while in the Cape Flats there are no less than 50. Pinnock (1984) classified these gangs into five categories:

• Corner kids which are mostly informal gangs, and whose members are teenagers involved in street life
• Defence gangs which are more formal, organized by territory,, engage in defence of their territorial jurisdiction against intruders and perpetrate theft and robbery
• . Reform gangs which are groups of ‘graduates’ of reformatories and constitute the largest gangs, like the Cape Town Scorpions and Born Free Kids
• . Mafia gangs which are big time groups engaged in robbery and drug running and providing protection for shebeens and shops – some Mafia gangs like Cisco, Yakis, Stalag 17 and Mongrels were formed in District 6
• . Syndicates, which are the most well established and sophisticated gangs, and engage in drug and gun running, receipt and sale of stolen goods, armed robbery, and assassination.

119The internal structures of these gangs vary from gang to gang, but most have ‘generals’, judges, teachers (who initiate new members into gang ‘policies’), soldiers, and killers (NICRO and Institute of Criminology, 1990). They are organized according to area or neighbourhood branches whose heads are answerable to the overall boss. The most visible gangs in the Cape Flats include the Sexy Boys gang, Dixi Boys, American Gigolo, Hard Livings, Mongrels, the Americans, and Clever Kids. Of these, the Hard Livings led by the Staggie twins, Rashied and Rashad, the Mongrels, and the Americans constitute the most formidable, and they have been involved in the struggle for supremacy. These struggles have taken the form of shoot-outs even in court premises, assassination, and revenge robberies.

120As indicated earlier, these gangs operate openly; but not openly enough to enable a researcher study their internal structures and dynamics, which are closely kept secrets. Leaders and members are known to the community and police who have sometimes had to resolve inter-gang conflicts. Parents in Cape Flats even refer cases of assault and rape involving gang members to their ‘kangaroo courts’. A film entitled Beloved Country: Cape of Fear has been made from gang activities in the Cape Flats, and it features the Hard Livings. The Staggie twins confess to killing people, especially members of rival gangs, and to supplying guns. According to Rashied: ‘I am the supplier. I give all over the peninsula guns. Even to the Mongrels who are our enemies. We mainly do business with guns’ (South, 20-24 May 1994). Rashied claims in the film that policemen not only supply them arms, but provide protection and sometimes fight for them: ‘Do you know how many policemen shoot for us in gang fights? Do you know how many policemen bring us guns and how many come here to sell their service weapons to us and then go back to the police station to report them stolen? They definitely bring us guns and after they aid shebeens and drug merchants, they bring the tablets (mandrax) to us and we pay them’.

121The beginnings of criminal gangs in South Africa have been traced to the emergence of urban war lords’ in the early 1900s, rapid urbanization, and the growth of townships and slums (cf. la Hausse, 1991; van Onselen, 1985). With specific reference to the Cape Flats, it is commonly agreed that the forceful evacuation of coloureds from District 6 to the Cape Flats under the Group Areas Act marked the turning point in the development of gangs. In District 6, the skollies or street kids and other youths organized criminal groups whose activities were serious enough to warrant the formation of a special police squad to fight the menace in 1946 (by the end of the 1930s, there were as many as 2,600 Cape coloured youths in prison). The police, which was already politicized, was unable to deal with the situation. This led to the emergence of more formidable community-oriented gangs like the Globe gang which provided protection from the skollies for shop owners and other residents for agreed fees, although the gang was involved in drug running and robbery. The situation changed in the 1970s when District 6 dwellers were forcibly relocated to the Cape Flats where housing was in acute short supply and families became disoriented. As one source explains, ‘Forced removals tore communities like District 6 apart, destroyed neighbourhoods, and resulted in the breakdown of the extended family … gangs replaced the extended family as a support network for the children and the youth’ (NICRO and Institute of Criminology, 1990: 3). Making a similar point, Pinnock (1984:53) says that as a result of the forced removals, ‘the working class in Cape Town was like a routed, scattered army, dotted in confusion about the land of their birth …The only defence the youths had was to build something coherent out of the one thing they had left – each other’

122But the factors associated with the forcible relocation to the Cape Flats should be regarded as aggravating factors. There are many more underlying systemic factors which explain the emergence of gangs in District 6 as well as in Soweto and other townships which did not experience such forced removals. Two sets of factors have been proffered. The first set comprises the socio-economic or systemic factors. In the literature on gangs, these factors are emphasized by adherents of the Chicago school who identify slum and ghetto factors of rapid population change, poverty, poor housing, breakdown of social control and adult crime in the rise of gangs (cf. Thrasher, 1927; Shaw and McKay, 1931); the sub-culturalists who see gangs as response to status frustrations experienced by lower class adolescents (cf. Cohen, 1955; Clowad and Ohlin, 1960); and the cultural deviance school which treats delinquent culture as a normal part of middle class culture. In the South African case, the systemic factors have been listed in a NICRO (1990) working paper as: lack of recreational facilities where the youths can ease tension, e.g., Mannenburg which has over 70,000 residents had only one recreational centre – the Silvertree Youth Centre – for a long time; unemployment; unequal distribution of wealth; poor educational standards; lack of political direction; repression of civic organizations by the state, which encouraged underground movements; poor town planning; and uncontrolled media glorification of violence.

123Similar explanations were offered by the Minister of Law and Order in Parliament in 1990: The formation of gangs is the result of deeply rooted socio-psychological, socio-economic and social problems over which the police have no control… the predisposition to crime can only be transformed by parental education and community institutions such as the school and the church (House of Representatives Hansard, 18 April 1990: 919-926). These environmental factors, as stated, underplay the historical context of apartheid and skip the important factor of police complicity. With regard to the latter, in addition to alleged collusion with the gangs, to which reference has already been made, the neglect of police protection for the non-white population and their preoccupation with defending the apartheid state prevented the police from giving the gang problem the attention it deserved. As the Western Cape command has found out, criminal gangs can only be dealt with through the cooperation of the community which could not be guaranteed while under apartheid. It has launched ‘operation gang burst’ which, involves close linkages with community and neighbourhood associations and raids on gang members. But the suspicions of collusion between some policemen and the gangs still prevents most Cape Plat residents from relying on the police; some in fact think the military should be called in.

124Did the proliferation of gangs have anything to do with political violence or was it related to the liberation struggle? The Cape Flat gangs did not have such orientations, though most of them regard blacks as a threat to the coloureds; in fact one of the gangs is called kaffirslagtters – black slaughterers. Even the gangs in the black townships had a negative influence on the liberation struggle. Mokwena (1991) shows how the gangs in Soweto actively obstructed civic organization activities and targeted ANC activists, leading at one stage to suspicions that they were sponsored by the government to destabilize the townships and weaken liberation movements.

125The second set of factors comprises the psychological or personal variables. These emphasize, as the major explanations for gang formation, individual pathologies resulting from faulty upbringing, unfulfilled personality needs, and masochist tendencies. Thus, the attraction of youths to gangs in the Cape Flats has been attributed to the following ‘services’ provided by these gangs: security, brotherhood, friendship, sense of belonging, identity, expression of masculinity and aggression, expression of care and concern, money, and status enhancement. Because these youths lack family support, they fulfil their need for belonging, survival and self-protection from the gang (NICRO, 1990). Mokwena (1991) also sees gangs in terms of ‘survivalist youth culture’. The gangs in Soweto he says, are attractive to youths because they promise quick money, and successful leaders are idolized. These explanations are useful because they help to explain why not all youths exposed to the same stimuli would join gangs (it is assumed for example that a child brought up under strict parental control and who enjoys acceptance, love and protection from the family is unlikely to join a gang), but they have to be related to the environmental factors discussed earlier which relate upbringing to inequalities and the frustrations of relative deprivation. This is the only way criminal violence can be related to other forms of violence as products of the same social milieu.

126While the study of criminal gangs underscores the urban location of violence, it also demonstrates some of the major problems involved in certain aspects of the study. Because of their very nature, the operations of gangs are shrouded in secrecy, and the researcher is treated with suspicion by both members of the gang and the police. In the case of the Cape Flats however, the existence of various organizations involved in the resolution of gangs and gang-related problems, notably NICRO and community anti-gang associations as well as an anti-gang unit in the police force proved quite helpful. These need, however, to be balanced with information that the researcher can gather from former gang members and, in some exceptional cases, actual gang members.

5. Towards a Framework for Analysing Urban Violence

127To conclude this pilot study, I shall summarize the major methodological points that have emerged by proposing a framework for the study of urban violence. The following key points and questions provide the dimensions of this framework, Its objectives are to discover how urban violence can be analysed holistically and empirically.

5.1 The nature of urbanization

128The first concern is the historical context of urban areas. What is the history of urban areas? What policies and planning processes have shaped them? Have these policies been continuous and unchanging? How do they relate to the overall character and orientations of the state? The second area of concern is the spatial and socio-economic condition of cities, or what may be called the structural factors. Are there marked inequalities among the different areas of the city? In particular, are there slums? Are these a consequence of planning failure or deliberate policy? What is the situation of housing, employment and social services delivery? Is there overcrowding? Is homelessness acute? Where do the homeless stay? What is the city’s rate of growth and what is its capacity for absorbing more migrants?

5.2 The mobilizing factors

129What are the factors which make urban conditions more likely to predispose its inhabitants to violence? What in particular is the level of relative deprivation? Is it at the individual level, where it is most likely to lead to non-collective violence, or has it been transformed into inequity such as allows for the mobilization of collective action? What is the structure of stratification and social mobility? Are the inequalities and deprivations perceived as legitimate or illegitimate? Are there ‘deviant’ and rebellious sub-cultures? Are there gangs which transmit these cultures? How are these gangs organized, and what are their major activities? How is the civil society composed? Are there militant organizations? What is the character of the state: Is it despotic or democratic? What are the prevalent strategies of violence control? What is the role of the police? What is the social attitude towards violence as an interest-begotten strategy? Are there justifications which make it acceptable?

5.3 The history and milieu of violence

130Violence is basically a historical phenomenon. Whether in political or criminal form, every situation of violence has a threshold which, once passed, makes it become generalized. At what point does violence get out of hand and why? How are the various forms of violence related? Within the urban framework, as I have argued, it should be possible to analyse these various forms as emanating from basically the same social structure.

5.4 Case studies

131Case studies provide the best empirical way to study urban violence in any depth. It enables the researcher to relate background and facilitative factors to concrete cases, realizing as we have pointed out, that no matter how much the background factors tell us, they still are not sufficient to explain why violence occurs when it does and the way it does. Another reason for the suitability of the case study method is that it is difficult to study all or even many aspects and instances of urban violence at once. The totality of violence as emphasized earlier is a very complex one, though, following the typologies which we discussed, it is possible to disaggregate them. When disaggregated, it is found that no matter how related they may be, certain types of violence are more prevalent than others. The researcher has first to find out what the prevalent forms are, and then, carefully choose cases which represent them. In doing so, he should be guided particularly by the need to separate the popular from the neglected cases. In the case of South Africa, he must resist the temptation which the literature provides to focus on political violence.

Acknowledgement

132I want to express my gratitude to the French Institute for Research in Africa (IFRA) Ibadan, and its Director, Georges Hérault, for the research grant which made this study possible, and to my colleagues at the University of Ibadan, Jinmi Adisa and Isaac ‘Wale Albert, who assisted with the initial design for this pilot study. I also wish to thank the University of Cape Town (UCT) for offering me a visiting position in Political Studies. This afforded me a base from which to conduct this study. At UCT, I enjoyed the assistance of Annette Seegers, Andre du Toit, Janet Sandell, Sonia Weber, and Sharon Moore of the department of political studies, and Rana Cochrane and Jenny Radloff of the Institute of Criminology. I thank Katoh Toshinosoke, Khanyisile Fangopa and Victor Gumbi for their research assistance. Finally, I want to thank Amen, my wife, and Osahon, Noyosayi, and Eseosa, our sons and daughter who contributed the sacrifice of separation from me during my sojourn in South Africa. I dedicate this study to Eseosa, our little darling, who came when I was away.