About this site

This resource is hosted by the Nelson Mandela Centre of Memory, but was compiled and authored by Padraig O’Malley. It is the product of almost two decades of research and includes analyses, chronologies, historical documents, and interviews from the apartheid and post-apartheid eras.

Dimensions of Turmoil

An approach to civil violence

According to figures supplied by the South African Police, 3 195 lives were lost in politically linked unrest in South Africa between October 1989 and February 1991. Very recently there has been an intensification of the turmoil, marked particularly by conflict between migrant workers in single men's hostels and youth groupings in township areas on the Witwatersrand. From the beginning of 1991 to the time of writing some 770 people have died,-roughly 200 more than in the equivalent period in 1990.

The argument which will be put forward in this position paper is that this violence has a multiplicity of "causes", the patterns of which have changed over time as the effects of earlier factors, in a sense, become new causes. In this way the violence becomes self-reinforcing. However politically expedient it might be, it is very doubtful whether "blaming" any single party, factor or agency for the violence is valid or useful. This analysis will suggest that a complex problem with interrelated "causes" cannot be easily solved and that it requires attention at a number of levels.

Furthermore, the factors in the mix of elements contributing to the violence are probably of widely-differing types:

i. General background conditions: factors which render communities or categories or people vulnerable to the effects of more immediate and specific factors referred to below.

ii. Predisposing factors: a number of economic, social and political factors which create solidarity; competition between groups or focused hostility towards particular people.

iii. Triggering events: factors of an immediate and current kind, generally quite visible and often generously covered in the media.

It will be argued, therefore, that short-term policies cannot be expected to address the diversity of factors which form "chains" of causation. Responses to the violence in the short-term must concentrate on.

- achieving negotiated agreements between persons in authority over all factions involved in the violence, so as to impose constraints on the behaviour of followers, as far as is possible;

- breaking the vicious cycle of triggering events, by introducing mediating, pacifying or diversionary measures which will prevent current effects from becoming new "causes";

- counteracting as many of the predisposing factors as possible in order to reduce levels of hostility in violence-prone areas.

Factors in the violence

General background conditions

These factors are merely noted in passing because, despite their seriousness, there is very little that can be done about them in the short term. Most of these factors are developmental issues and development action always takes a long time to alleviate frustration. It might even happen that, by raising expectations, development programmes intensify frustrations for a period. Obviously, this is not intended to imply that development should be deferred; it simply means that development strategies cannot normally address any immediate crisis of violence.

The type of factors referred to here are problems like:

- high levels of unemployment and an increase in dependency on employed workers;

- the dislocating effects of the rapid urbanisation and population movement which has occurred particularly since 1986 with the abolition of the so-called Influx Control legislation*;

- problems in the family structure of poor urban communities which lead to youth group and gang formations replacing the more established sources of social identification, authority and discipline in communities;

- pathologically high levels of social violence (not unrest or political violence) which erode normative restraints on killing and assault. In the period from October 1989 to February 1991, when some 3 200 "unrest" deaths occurred, there were some 26 300 recorded murders. Where one type of violence is prevalent, other types of violence are more likely to occur as well*.

- shortages of social resources like accommodation, land for settlement, educational opportunities and the like.

A perusal of the factors listed above will confirm that they are not amendable to short-term corrective action.

Factors predisposing people to violent competition or hostility

This is the level at which many of the most relevant problems have their effect. Unless at least some of these factors can be addressed, a considerable propensity to violence will endure no matter what other steps are taken. Unfortunately, not all these problems can be alleviated in the short run.

Factors at this level cover a wide range, and include the following.

Intergenerational, rural-urban and class tensions in the relationship of migrant workers to their urban social surroundings

One of the major questions to answer is what the reasons may be for the very intense, massive and violent responses of migrant hostel dwellers in the very recent unrest. The argument presented here is that there has been a build-up of tensions over a long period which are now having their effects.

The social alienation between rural migrant workers and township residents is well known*. The sociologist Herbert Vilakazi has more recently discussed this in relation to violence*. It is difficult to do justice to Vilakazi's eloquent essay. Interpreting his analysis very broadly draws attention to a quite fundamental clash of social orientations between migrants from rural areas and urban township groupings.

On the one side one has the partly urbanised, partly rurally orientated, less educated and less skilled marginal working class of migrant workers. They value the rural tempo of life, they are still subject to clan discipline, with its respect for age, seniority and ancestors. A strong identification with area of origin and its social networks exists. Many are economic target-workers with a strictly functional interest in the cities. The concern with job security and with financial responsibilities to their families at home is overwhelming.

Vilakazi goes further to elaborate on their social and political culture: "yes, they are slow, steady and cautious ... but once their pride is pricked, provoked or treated contemptuously, peasants and semi-peasant people can easily turn as wild and violent as a ... mad ox ... (flinging their) whole being into the fight without the slightest fear ...".

In shrill contrast are the township youth and young adults; sons and daughters of the urban-insider industrial class or the "slum kids" from squatters areas. Since the early seventies, activism and social codes have developed which are rooted in urban youth subculture. The banning of major organisations removed the guidance of senior political leaders. Youth activism developed an aggressive quality of its own. The migrants, as Vilakazi sees it, are "turned off, puzzled and infuriated by the impatient, impetuous tactics of the youth ... instructing adults in a manner totally outside the traditional code governing relations between young and old".

The tensions arising in this contrast of orientations are not new. We may recall the conflict between youth and migrant workers in the "Soweto" conflicts of the late seventies, or the confrontations between so-called "vigilantes" (also with migrant origins) and urban political formations in the Western and Eastern Cape, as well as early conflicts between UDF-linked youth and local headman in Natal shack areas in the early eighties.

This basic social tension will not be alleviated very easily. Even if the hostel system could be done away with quickly (see below), the migrant groups would most probably regroup and cluster together in shack areas, would be less exposed to official surveillance, and all the social tensions suggested above would persist. One must remember that prior to the construction of modern hostels, networks of migrant workers lived in so-called "yards" or informal compounds in places like Sophiatown, Alexandra and Lady Selborne.

In any event, no immediate dismantling of the hostel system is possible. These are some 140 000 hostel beds in the Pretoria-Witwatersrand-Vereeniging area with up to twice that number of residents.

Recent sharp increases in retrenchments and unemployment after earlier rapid growth

The noted authority on rebellions, Ted Robert Gurr*, on the basis of a mass of international evidence, formulated a social "equation" of factors predisposing towards violent action. Among the factors in the equation are four prominent variables which he calls "inflexibility of value stocks", "rate of change in value position", "reduced value opportunities" and "differential increases in other groups' value positions". Gurr uses the words "value stocks" to refer to interests and associated opportunities. Applying it to the local context, one may sketch the following dynamic as it applies to black urban communities.

Since the mid eighties migration to urban areas has increased. Work opportunity expanded with the 1986-1989 mini-boom. This has been followed by a sharp contraction in opportunity but within a context of hugely increased demand for urban employment. Alternatives to formal employment are limited since informal activity is increasingly "over-traded" at present. Although not perhaps fully conscious of it, urban youth and unemployed adults, Zulu migrants, other migrants and Xhosa-speaking shack dwellers are all in heightened competition for very fixed and limited opportunities. Gurr's equation fits perfectly: expectations and opportunities rising in 1988, followed by opportunities falling in late 1989 and 1990, faster for some than for others, and with many people arriving rapidly in a situation of increased privation with few or no alternatives.

Gurr would have predicted some turmoil, particularly since some of the competing categories are "cohesive" groups, like migrant workers, whom Gurr regards as additionally predisposed to violent reactions, because of solidarity and a shared sense of deprivation.

Spelling out these dynamics may not be relevant for very short-term strategies (it is impossible to "repair" the effects of an economic downturn aggravated by labour-saving strategies in industry). These factors have a medium-term relevance, however. South Africa cannot afford "boom and bust" economics. People proposing that the economy be "kickstarted" must be very sure that the growth can be sustained without inflation and a subsequent need to cool an overheated economy. Counter cyclical strategies of job-expansion as part of planned and finite development projects may also be called for.

Faction fighting traditions

Migrant workers often originate from districts where- a long and established tradition of "feud" exists. The Msinga area of Natal is merely one of the better-known examples. In the seventies some of the rural faction fighting spilled over into places as remote as Hillbrow and the Johannesburg CBD. Non-Zulu workers in mining employment have also been exposed to the phenomenon. McNamara records that between 1974 and 1986, at least 141 inter-factional or inter-ethnic clashes occurred in mining employment. Over 300 migrants died and more than 2 000 were injured*. McNamara's analysis points to the salience of competition between communal groups of workers for recruitment and employment advantages and anxiety over the possibility of other groups augmenting their relative advantages as the more critical cause of violence.

What is being argued is that the basic response patterns seen in the current violence are well-established reactions to heightened tensions over interests. The massive "over-reactions" of armed hostel dwellers in the violence have to be seen in this context.

Mass action and mass mobilisation

The strategy of "mass action" or "mass mobilisation" of the ANC and its linked and affiliated organisations is a complex topic which cannot be debated here. Jeffery identifies no fewer than 17 different goals of mass action*. It has often been argued that, given the earlier suppression of political activity in South Africa, "mass action" was one of the few effective strategies left to the mass-based movements.

Whatever the case, the concern in this paper is to assess its effects in predisposing communities towards violence. In late 1989, the "Mass Democratic Movement" approved a resolution at its Conference for a Democratic Future, to bring pressure to bear on "homeland" leaders to resign and to bring similar pressure to bear on black town councillors. Jeffery adds "other aims, not publicly acknowledged but suspected ... are to destroy the power of Chief

Buthelezi and Inkatha". Youth movements contributed to the mounting tension: "Mobilise, organise for the final offensive ... all youth to battle ... all youth to the frontline" are quotes from press reports on the 1990 conference of the South Africa Youth Congress.

Thus there was mounting activism seen inter alia in the form of consumer boycotts and stayaways, and, according to Jeffery, attacks on councillors in the first seven months of 1990. According to a list published in Ilanga*, 13 assassinations of middle-level Inkatha leaders occurred in 1989 and early 1990. The Inkatha Institute also details 22 attacks on Inkatha members between 29 January and 5 April 1991*. Many more details could be given. The effects on the social climate are all that is relevant here.

Among traditional Zulu circles a sense of mounting embattlement could be discerned. In March of 1990 King Goodwill Zwelethini called on his people to honour their culture "with your Zulu might" and he called on Zulu traditionalists to "act swiftly and decisively". Inkatha was also accused of mounting coercive counter-recruitment campaigns particularly in Natal.

Reactions on the ground in the townships are of interest. In late 1988 this author conducted a personal interview-based survey in conjunction with Market and Opinion Surveys (Pty) Ltd among a random sample of 1 016 people in the Pretoria and Reef townships (the results referred to here have not yet been published). Some 51 per cent of both recent and long-established migrants to the city declared themselves to be "very dissatisfied" with the enforcement of rent and service charge boycotts. About a third of established city people also disapproved strongly. John Kane-Berman of the SA Institute of Race Relations has just released the results of a March 1991 survey among 905 people in all metropolitan townships in which a third of township residents said they had been forced to participate in consumer boycotts and in stayaways. More relevantly, four out of every five Inkatha Freedom Party (IFP) and AZAPO supporters claimed that they had been forced to support these actions*.

The concept of mass action implies that the "mass" (ie, a large majority) will voluntarily support the action taken in their name. Where this is not the case, as is evident from the results of the surveys quoted above, "mass action" has two effects. One is on the target it is directed at. The other effect is to polarise communities in which it occurs. This polarisation is not necessary visible. Township residents opposed to mass action would be too cautious or fearful to react. Migrant workers living in cohesive, socially separated groups, however, are not exposed to the same constraints. Among them mass action will lead to counter-reaction.

One must add to the polarising effects of the strategy the fact that Zulu migrant workers identified with the King or with the IFP would have begun to feel their identity threatened. Meetings and political sing-songs held on commuter trains caused great discomfort. A state of emotional "siege" began to take hold. The actions of semi-political gangs like the "Gaddafi" and "Zim-Zim" gangs in Reef townships, for example, increase the tension and stimulate the emergence of counter-gangs. For migrant workers present in the city for specific economic reasons, the social environment became very threatening and hostile indeed.

A very recent opinion-survey by this author*, shows the following:

- while the ANC, as an umbrella organisation, has very clear majority support, organisations linked to it (however loosely), like the civic associations, street committees and youth organisations associated with much of the mass action, do not enjoy majority support (some 40 per cent and even less in Soweto);

- the support for these mobilising organisations is virtually non-existent in the migrant hostels;

- in regard to strategies aimed at putting pressure on system-linked functionaries like town councillors to resign, urban township communities are split down the middle, with at least about half opposed to such action (without necessarily supporting the councillors).

Mass action, and the assumptions underlying it, also have the effect of promoting hegemonic political control of territory. By no means only the ANC or civic associations are at fault here. Similar accusations can be levelled at the IFP in Natal and at AZAPO/BCM in the Eastern Cape in the past. Control of territory must inevitably engender hostility.

Summing up one might say that, at a time of mounting anxieties over jobs and survival, "mass action" has contributed to the build up of tensions and stress among the rank and file. One accepts without reservation that political organisations have the right to mobilise to achieve political effects and to attract attention to grievances. At the same time, however, where the strategies involved do not carry the endorsement of very clear or overwhelming majorities in the townships, polarisation and increased hostility between political "camps" must inevitably be the result.

The alleged bias of the police

In the survey by this author referred to immediately above, township residents were asked which groups they felt were "mainly responsible for starting the violence". The IFP was mentioned most frequently. (A recent media survey by the Community Agency for Social Enquiry (CASE) tends to bear this out)*

At almost the same level of mention as the IFP, however, were the politically and semi-politically oriented youth formations, but third in order of mention was the South African Police. This raises the issue of the possible partiality of the police, which is constantly raised in the press. There has also been much mention of the possibility of some form of stimulation of violence by a so-called "third force" with suggested connections to security agencies. According to the CASE media survey referred to above, the press pointed to 18 per cent of acts of aggression as being apparently due to security force action*.

These accusations may or may not be valid. In the case of the ANC and the IFP there has been some acknowledgment by their respective leadership of a role in violence (Mr Thabo Mbeki of the ANC said that his organisation bore some responsibility for violence*, and the IFP has admitted certain attacks)*. Spokesmen for the South African Police, however, have consistently rejected the accusations and asked for the submission of firm evidence; evidence not merely based on hearsay, on opinion or rumour. The police have acted against individual members on the basis of evidence on some occasions (one policeman was suspended recently for comments made in front of the media). This author has no evidence to go on. The following points may be relevant to the topic, however.

In a research paper, Olivier of the HSRC suggests that police actions have been shown to aggravate situations of violence*. It is, however, fairly universal for the police to be accused of brutality and bias against certain categories of people. It has been a familiar accusation during race riots in the UK, for example. At a personal level, personnel in law enforcement agencies may very well have particular sympathies and dislikes. In South Africa this possibility has to be taken seriously for the following reason.

According to figures released by the South African Police, between 1 October 1989 and 30 April 1991, 85 policemen were killed and 811 were injured in political unrest-related incidents. According to the Survey of Race Relations, 1989/90, the homes of 522 policemen had been attacked in the first eight months of 1990*. Hence the police themselves are under attack, and, most importantly, dominantly from one political grouping involved in the violence.

If police partiality is a problem it must be addressed very urgently. It cannot be fully addressed without at the same time giving attention to politically motivated attacks on the police themselves. It is problematic to fully expect any category of actors to be impartial if the people involved perceive themselves to be victims.


An ethnic factor has already been alluded to in the earlier discussion of the responses of Zulu migrant workers. Very recently there have been reports of an upsurge of "tribalism" among Xhosa-speaking squatters on the Reef. In one incident there was an overtly tribally based attack on a neighbouring squatter camp by Xhosa migrants when they heard that a cohesive group of Zulus was living there*.

Ethnic identity is what one may term a "floating" phenomenon. It is usually an element of personal re-assurance, pride in identity and a social locating factor. When interests are threatened, or survival is at stake, it can readily float to the forefront of self-assertion and become a powerful basis of mobilisation. Once communal identity is mobilised in this way, it is more powerful in its effects than the material or social interests which may stimulate it. Like religious identity, ethnicity has significance for self-esteem, and this contributes to the intensity of its effects.

Therefore it is incorrect to see antagonistic ethnicity as an inevitable prior factor. It is perhaps more often than not a consequence. Once activated, however, it focuses hostility on "out groups" very sharply, can be deadly in its effects and does not abate for a long time.

It would be naive not to expect ethnicity to become activated in the current violence. It is going to make the violence immeasurably more difficult to deal with.

Trigger effects (blaming the right people at the wrong time and the wrong people at the right time)

With precipitating or "triggering" effects it is unnecessary to discuss their particular characteristics. Once the predisposing factors are present, almost any aggravating incident can trigger a reaction. It can be a fight over a woman, a shebeen brawl, an individual assault, a rumour (rumours of impending attacks are currently rife in the townships) or even a chance insult. Some years ago in the Durban Indian suburb of Chatsworth, a group of African soccer-players singing songs on the back of a truck sparked a mass panic in the suburb and thousands of residents fled to the police station.

There has, however, been somewhat of a pattern in many of the precipitating events. The seasoned Natal journalist, Graham Linscott has written perceptive articles which illustrate the problem*. His accounts demonstrate the difference in "styles" of conflict between the "youths" or "Comrades" on the one hand and

IFP-linked groups on the other. This difference in the way violence is organised gives rise to the impressions which make blaming one group more than others very easy.

Linscott describes the situation in a way which accords with my own impressions of earlier conflict in Natal. The "youths", often but not necessarily first, operate as small, anonymous and highly mobile clusters of people harassing particular opponents or people on transport routes to and from areas of IFP concentration. Individuals are assaulted or killed, women and children can be endangered, road blocks are erected, tyres burnt on roads and from time to time the house or car of an IFP notable is burnt. Tension builds up in IFP-dominated communities. The response of people associated with the IFP is almost always more visible to the public and the media. What appear to be planned attacks are launched on areas in which the youths are suspected to be concentrated. Large numbers of armed men are deployed and immense destruction is caused, scores are killed and refugees can run into thousands.

On neither side are the antagonists necessarily formally part of, or officially sanctioned by the major political parties. The association is made by observers. IFP and ANC-linked groups, however, blame each other for initiating the violence. Earlier in Natal, the Mass Democratic Movement developed a strategy of obtaining court orders against Inkatha. Inkatha spokesmen, on the other hand, pointed to the harassment of Inkatha youth in schools and objected to residents having to run gauntlets on roads patrolled by youth or Comrades on their way to and from work and town.

This example is perhaps more relevant to Natal than to the Transvaal, but it illustrates a point. Once all the predisposing factors are in place and hostilities have started, violence becomes self-perpetuating in a process of reaction and counter-reaction. Very frequently, though, the events which capture the attention of the media are the larger-scale reactions, and the preceding sequence of events which stimulated rising tension is underplayed, more often than not unintentionally.

Precipitating factors may not be very important as "causes", but they require urgent attention because they perpetuate hostilities. There is a tendency among superficial commentators to identify them as the major factor, usually in the context of blame. Hence we have a litany of opinion "blaming" the IFP, the ANC, The police, vigilantes, "Third Forces", "warlords", "chiefs", gangs, and others. The agencies so blamed are at one and the same time agents and victims of violence. Superficial moral condemnation is an extremely unhelpful (but unsurprising) aspect of our current violence.

Some more general comments, conclusions and suggestions

Certain issues have not been dealt with under the factors mentioned above, although they are implicit in some of the arguments put forward. One such issue is political competition and the possibility of some kind of encouragement of violence by political leaders aimed at weakening opponents.

Political competition is obvious, and it is a perfectly predictable factor at a time of transition to a new political dispensation. We must bear in mind that, notwithstanding the claims of some political spokespeople, South Africa's negotiations are not about the terms of capitulation of any major political grouping. There is, in a sense, a power stalemate. The negotiations are, and will increasingly be, competitive. All major parties involved intend to negotiate to secure strategic advantages in the new system. Those parties which are not (yet?) part of the negotiation process, like the Conservative Party, the PAC and AZAPO, are absent perhaps precisely because they do not see much opportunity for securing advantageous positions in the new negotiated system, given their political missions and goals and their relatively small support bases.

The competitive nature of the interaction among the parties participating is quite clear and needs no exemplification. It is for this reason that so much "position play" takes place outside of-negotiations. Attempts are constantly being made to increase leverage in negotiation by political pressures extraneous to the negotiation itself. Jeffery, referred to earlier, includes this among the goals she lists for mass action. Strategists within the ANC, like Ronnie Kasrils, are quite blunt about it. Quoting Mao Tse Tung (apologetically) Kasrils and Khuzwayo argue that "You cannot win at the negotiating table what you have not won on the battlefield", and their view of mass action is to "open the doors to a transfer of power", to establish coercive leverage; "the negotiating table is not the real battleground"*.

The same concern about leverage obviously exists with other parties. It has been suggested in numerous ways in debates that the IFP considered itself to be in danger of being sidelined because of the high-key attention given by government to the ANC; here one refers particularly to the Groote Schuur and Pretoria "minute" meetings and associated and subsequent interaction. Conceivably it too found the strategy of leverage outside the negotiation process a relevant issue for consideration.

It is easy to jump from this kind of logic to a conclusion that political parties and agencies of government have hidden agendas which may involve the violence. This approach would define the violence as "conspiracy", to one degree or another.

All is possible in so complex a process of transition as South Africa. This author, however, does not have any information sufficient to back-up this kind of proposition.

Furthermore, it has to be pointed out that the dynamics of social action in stressful situations are seldom clear cut and systematically directed. Ambiguity is always a factor in social action. Furthermore, certain forms of behaviour or action can persist not because they are imperatively determined or deliberate but perhaps because they do not critically inconvenience the parties or agencies which might have the power to stop them. Then again, some forms of social action, and violence is a good example, occur in the crevices, nooks and crannies of social life and institutions, and it is conceivable that those local groupings immediately responsible for the violence cannot be controlled. Even the one party which has made official statements suggestive of social pressure of a kind which can evoke violent reaction (see references earlier to Jeffery and to Kasrils and Khuzwayo on the ANC) is not necessarily ordering the specific kinds of pressures which evoke violence. In the study of this author referred to earlier, township residents pointed to youth action as a cause of violence but they tended more especially to identify youth elements not directly connected to the ANC.

The point being made is that if one assumes "conspiracy" and the real process is instead one of "slippage", one cannot make an analysis which is appropriate as a basis for ending or curbing the violence. "Conspiracy" theories tend to land commentators in a never-ending process of oscillating blame,, which ends up simply stimulating more antagonism.

This author's view is that the available evidence suggests that one should "blame" all or "blame" none. The real task is to understand the violence in such a way as to combat it. And the violence is obviously very complex.

This brings this argument to its conclusions. Many years ago, Berelson and Steiner, after reviewing comparative evidence, came to the conclusion that "Human conflicts cannot be settled by removing the original source",* and one might add, if one can find the very original source at all. Civilian violence is one issue where it is appropriate to treat the secondary or predisposinq factors and symptoms.

Many of the suggestions made in the media debate on violence are useful. These include the need to establish monitoring and mediating mechanisms to achieve as close a surveillance as possible of the participating parties by objective committees and institutions. Here one must emphasise the word objective, because in a polarised society like South Africa, most "civil society" and even the media are not independent of political viewpoints. Some further comment will be made on this issue below.

These suggestions need to be expanded. Before doing so one or two brief substantive conclusions are required:

1) By now the violence, in its impacts on all relevant parties,

is self re-enforcing. In other words the people on the ground, irrespective of whether their actions were instructed, tolerated or overlooked by political superiors, are by now self-motivated and are gaining some reward. One cannot assume that they can be fully controlled. Any intervention must deal directly with them.

2) This point notwithstanding, there are aspects of strategies or responses in all parties and agencies involved which could re-enforce tensions, whether deliberately or inadvertently. All the agencies have their rationales for the responses. This includes police who may be slow to intervene where they might otherwise have been more effective; it certainly includes mass action and consumer boycotts at a time of exceptional economic privation and it includes assumptions, deliberately or unconsciously made by some participants, that the best method of defence is concerted attack.

Public sanction of such responses needs to develop more muscle,

3) An aspect of the action not dealt with thus far is the emergence of aggravating symbols. Zulu "traditional" weapons have become a symbol even though they are quite evidently not the weapons used to do most damage. Concentrations of workers in hostels, in part, have become a symbol, despite the fact that in Natal, concentrations of shack-dwellers, communities of tribesmen and large gangs of youths can act with equal cohesion and lethal effect. A group of youth "toyi-toying" has also become an aggravating symbol, despite the fact that most destructive youth action occurs with great stealth and without choreography. It is obvious, for example, that Zulu migrants in shack accommodation will do no less damage with hastily-pulled fence-poles than with spears. Youth in neighbouring shack areas can buy soft drinks next to a filling station and raise a dozen houses with petrol bombs ten minutes later. It may be politic to deal with the symbols, but a symbol is of equal value, negatively or positively, on both sides. Attacks on or actions against offensive symbols give equal offence on the other side. One must guard against the violence leading to a futile contest of symbolic reparations.

Against this background, then, it might be appropriate to consider the following as guidelines for addressing violence:

a) Since political competition and tensions between political movements are an obvious and important element in the violence, the current emphasis being placed on negotiations between political leaders, and on agreements or "pacts" relating to political conduct are obviously of crucial importance.

b) Because much of the violence is beyond the control of leaders, top level negotiation is not the only answer. Equal emphasis should be placed on establishing active, full-time small task forces of communicators and mediators in each local area to interact directly with and listen to the grievances, fears, interests and misperceptions of people of the kind who are active in the violence. The first step to peaceful social integration occurs when aggrieved local groups are noticed and taken seriously.

Hence consideration should be given to the establishment of an independent fund under a group of representative trustees to engage, pay and train small, appropriately selected local task forces of say six people under a non-aligned leader. They might spend their full time listening to the local leaders and activists, providing feedback to parties and the authorities and giving unbiased and fairly conclusive reports to the press and important observers.

c) The media must strengthen their capacity to editorialise with authority and effect on the behaviour of parties. All parties and agencies in South Africa are sensitive to the media. Speculative, impressionistic and tendentious articles or programmes often make good reading or listening and this publishing freedom must be protected. It might be appropriate, however, for the media to develop strategies (perhaps panels of objective experts?) to deliver telling and categorical assessments from time to time and to spend resources on obtaining as many valid facts as possible. News items quoting "residents" by the roadside should, perhaps, always be contextualised. As far as possible news items on violence should always contain quotes from all protagonists in the violence.

If it is possible for the churches, in their comments on violence, to develop similar balance, depth and merciful objectivity, so much the better.

d) Any action taken against groups whose participation in violence might be in defence of interests or self-esteem must be done in such a way as to make them feel part of the solution. This would apply, for example, to proposed action against the hostel system, which migrants in Tembisa and Diepkloof are already mobilising to defend. The same principle should apply across the board.

e) The frequent accusations levied against the police highlight the need for the system of local monitoring committees involving the police and parties active in a particular district to be revitalised. The ANC has not nominated a full contingent of representatives to these committees thus far. This author has been told by prominent ANC members that one reason for this is that the police are not trusted. This might be precisely a very good reason for participating fully in such committees, however. It is quite clear that a basis for full, multi-party participation in monitoring committees with the police must be negotiated very urgently at high level. Possibly an independent convener of such local committees would solve some of the problems.

At present there is a lull in the violence, but past patterns suggest that it may be temporary. The problem has become so complex as to defy most types of corrective action for quite a while. For this reason it is imperative that all possible influence be exerted on all parties to continue negotiating at a political level despite the violence. This is most important of all because many of the factors underlying the current turmoil, uncertainty and violence are activated precisely by the stresses of transition from an old to a new political order.

This resource is hosted by the Nelson Mandela Centre of Memory, but was compiled and authored by Padraig O’Malley. Return to the Nelson Mandela Centre of Memory site.