Olivier, J. (1991). The South African Police: Managers of conflict or party to the conflict? Paper presented at the Centre for the Study of Violence and Reconciliation, Seminar No. 1, 27 March.


Presenter: Dr Johan Olivier

Johan Olivier is a Researcher at the Centre for Conflict Analysis at the Human Sciences Research Council.

This research was sponsored by the Anglo American and De Beers Chairman's Fund and the Human Sciences Research Council. Opinions expressed here and conclusions drawn are those of the author and not necessarily those of the sponsors.

Date: 27 March 1991

Venue: University of the Witwatersrand, Johannesburg, South Africa


This paper discusses the role of the South African Police during instances of civil unrest. First it provides a definition of violence and then proceeds to call for a definition of the term "violence" which is not restricted to instances of physical harm. State-supported violence is then defined. The issue as to whether the police should be seen as managers of conflict or whether they are party to the conflict is then discussed with particular reference to the South African situation. Research findings which show how actions of the police at instances of civil unrest contributed to an escalation of civil unrest in this country are then reported.

On 4 May 1970 National Guardsmen opened fire on Kent State University students protesting the United States' invasion of Cambodia, resulting in the death of four students and the wounding of nine others.1 On 4 June 1989, after seven weeks of peaceful pro-democracy demonstrations by students on Beijing's Tiananmen Square, members of the People's Liberation Army opened fire with AK-47 assault rifles. An estimated 5 000 citizens died and thousands more were injured.2 On 8 October 1990 Israeli police opened fire on a crowd of Palestinians on the Temple Mount who were, according to the Israeli version, throwing rocks at Jewish worshippers at the Western Wall. Nineteen Palestinians died and 140 were wounded as a result of police action.3 At the time of writing this paper, the Soviet military was busy with a systematic assault on citizens of the Baltic states of Latvia, Estonia and Lithuania.

What do these incidents have in common? It is clear from accounts of these events that the degree of violence used by the agents of the state by far outweighed the degree of violence used by protesters.

Consider the event which took place on 21 March 1985 in Langa near Uitenhage:

Fourteen people were shot dead and many were wounded when police opened fire today on a crowd of more than 3 000 in Langa township near Uitenhage … . Police opened fire with semi-automatic rifles and shotguns when an order to disperse was met with stones and other missiles. (Argus, 21 March 1985)4

This event reminded us of the Sharpeville incident on 21 March 1960 when police opened fire on a crowd protesting the pass laws. The Sharpeville incident left 69 people dead and 180 wounded5 - many were shot in the back and victims included women and children. More recently police action on 26 March 1990 at Sebokeng and other townships in the Vaal Triangle, which left 12 dead and 281 injured as a result of police action,6 once again highlighted the use of violence by the state and its agents.

What are the limits of force the state may use to enforce its laws? Putting it differently, what constitutes state-supported violence? What is the role of the police during instances of civil unrest? This paper seeks to answer these questions.

In the next section I shall define violence and state-supported violence after which I shall briefly refer to the debate on whether the state is a party to conflict or only a manager of conflict. I will then proceed to discuss the role of the police during instances of civil unrest with particular reference to the South African situation.

Defining State-Supported Violence

Before proceeding to defining state-supported violence we need to have a working definition of the term violence. It is important to recognise that the meaning of the term violence may vary widely with the user and with the context. While its meaning has traditionally been restricted to instances of physical harm, it is increasingly used to incorporate meanings quite distant from the notion of physical harm.

Garver's (1970) definition, which views violence as any action which infringes upon the autonomy and dignity of individuals, is an example of a broader definition of the term. Van der Merwe (1989:16) supports this view when he defines violence as:

… the application of force, action, motive or thought in such a way (overt, covert, direct or indirect) that a person or group is injured, controlled or destroyed in a physical, psychological or spiritual sense.

What is important to recognise is the fact that violence has many spheres. These spheres include the physical, the psychological, the economic, the social, the structural, and the political.

I would therefore like to argue for a definition of 'violence' which is not restricted to instances of physical harm, but one which also incorporates instances of psychological/emotional harm inflicted upon victims in the different spheres mentioned above. One individual can, for instance, verbally (and violently) abuse another without physically attacking the person. Degenaar (1980) also opts for a broader definition of the term violence.

Degenaar (1980:14) defines violence as "… extreme force wilfully carried against X, which violates X because it does not show respect for the value of X." This definition highlights three key dimensions, namely

  • Extreme force which violates the integrity of the victim
  • The intentionality behind the violence, and
  • A value that is ascribed to the victim.

According to this definition, extreme force used by the police which clearly exceeds the amount of force necessary to subdue the assailant (victim), violates the integrity of the victim and is therefore unjustifiable. When the amount of force used violates the integrity of the victim we are dealing with violence and when the state or agents of the state initiate this action we are dealing with state-supported violence.

Thus, the Langa incident referred to above where the police responded to stones and other missiles with semi-automatic rifles, clearly constitutes state-supported violence.

As suggested above, violence is not limited to instances of physical harm. Structural violence is an example of a type of violence which may not be physical.

Structural violence is violence which is not exerted by an individual but by a structure, or structures, created or perpetuated by custom or by law (Degenaar, 1980). The structures thus created unfairly curb the freedom of subjects or discriminate unjustly against certain sections of the population.

According to the social scientist Johan Galtung, who is credited with coining the concept of structural violence, structural violence is "… present when human beings are being influenced so that their actual somatic and mental realisations are below their potential realisations" (Galtung, 1969: 168).

Structures are violent in that they prevent individuals from having equal access to society's life chances.7

This view is also supported by Kotze (1978) who goes on to argue that paternalistic and selective development are also examples of structural violence. In his definition of structural violence, Davies (1976: 131) emphasises the uneven distribution of power and resources:

Structural violence shows itself when resources and powers are unequally shared and are the property of a restricted number who use them not for the good of all but for their own profit and for the domination of the less favoured.

This definition introduces the economy as another sphere in which violence can be committed. One of the consequences of violence committed in the economic sphere is poverty.

Poverty is an important social problem and may in many cases be the result of violence committed in the economic sphere. Political, social, and economic structures can force poverty onto people. The violence lies in the fact that poverty forces people to live a sub-human life (Degenaar, 1980).

It is clear from the above that state-supported violence has many faces. What is obvious is that the definition of violence, and state-supported violence, should not be restricted to acts of physical violence.

What is of special interest here is how state-supported violence manifested itself in the context of civil unrest. Should the actions of the state, and its agents, which were devised to curb the incidence of civil unrest be viewed as actions with the intention to manage conflict, or do these actions make the state a party to the conflict. It should be recognised that the civil unrest the state attempts to curb may in the first place be a response to violent actions of the state, be it physical or structural.

It is generally accepted in the literature that the state has a monopoly over the use (of the tools) of violence. These tools include the police, the military and the legislative process.

Van der Vyver (1988) points out that violence committed by the state is almost invariably the result of a lack of legitimacy of a government, constitutional dispensation or legal system (p.71). It is when there is a significant loss of regime legitimacy that the state monopoly of violence will increasingly begin to be challenged (Zimmerman 1987: 337). The decline in regime legitimacy weakens the state and makes it easier for opponents to challenge the state by violent and/or non-violent means.

The wave of unrest which began in 1976 and reached unprecedented levels since 1984, is regarded by many to be the result of a decline in the legitimacy of the South African government. The question which needs to be posed here is, What is the role of the state in conflict situations such as this country has experienced since the mid-seventies? Should the state and its agents in these situations be seen as managers of conflict or are they party to the conflict?

The State as Conflict Manager or Party to Conflict?

Take the incident at Langa referred to at the beginning of the paper. According to one (it may be said to be the dominant view on the subject) theoretical perspective (as represented by the "Hegelian" theories, eg. Smelser 1962; Huntington 1968; Gurr 1970), the violence exhibited in the Langa incident was the throwing of rocks by the participants in the march. The actions of the police are viewed as conflict management aimed at reducing the occurrence of conflict situations. In this tradition, the use of force by authorities is typically viewed as social control, coercion, law enforcement, or the preservation of law and order. State-supported violence is considered to be an aspect of conflict management rather than violence. This perspective does not permit the view that the exercise of authority by the state can itself constitute political and physical violence. As Zimmerman (1980:191) points out: "Much of the violence involved in protest and rebellion is in fact initiated by governments for the purpose of controlling protest and rebellion."8

The other theoretical tradition views the state as a party to the conflict. This tradition has its origins in Marx's critique of the Hegelian conception of the role of the state in political conflict. The state, in Marx's view cannot escape a partisan attachment to certain interests at the expense of others. In fact, the state always represents the interests of the ruling class. According to this theoretical tradition, the violence is construed as residing in the actions of both the police and the protesters. An important characteristic of the theories in this tradition is that

… none of them is built upon the premise that the state necessarily must, or as a matter of contingent fact actually does, embody societal consensus, act in public interest, or play in conflicts the role of manager rather than partisan. (Nardin 1971:33)

Thus, the use of force by the state is seen to constitute not conflict management, but violence - the state is viewed as party to the conflict. Conflict management is conceptualised as the restraint of violence committed by any party in a conflict situation.

Whether we agree with Marx's conception of society or not, it makes intuitive sense to me that any party in a conflict situation can commit violence which is not justifiable. In view of this I do not support the Hegelian tradition with its view that the state and its agents are above the clash of private interest in civil society.

I therefore would like to argue for a view which allows for the possibility that both the actions of the state (or the police as agents of the state) and the actions of those opposed to the state (protestors in the case of an event of collective action) can be regarded as violent. Violence committed by any party in a conflict situation can be unjustifiable.

The wave of civil unrest which began in the middle of the seventies and continued spasmodically for more than a decade, resulted in the deaths of many individuals, some of whom were innocent bystanders. The police as agents of the state were frequently accused of aggravating the situation by their actions. The events which took place during 1990 once again highlighted this debate.

The SAP: Party to the conflict

A recent study (Olivier 1989b & 1990), which covered the period 1970-1984 in the PWV area, paid particular attention to the consequences of police action at events of collective action.

This study investigated, amongst other things, the effects of police action at 657 events of collective action on the subsequent rate of collective action. Police action was measured at five levels

  • No police presence
  • Police present only with no interaction with protesters
  • A minimal level of interaction which includes the putting up of roadblocks or asking participants to disperse
  • Physical contact between participants and the police with any number of arrests made, and
  • Police opened fire on protesters with rubber bullets, teargas and/or live ammunition and killed or wounded any number of participants.

The results suggest that the police actually contributed to the escalation of the unrest. The mere presence of police at events of collective action increased the subsequent rate of collective action by 68% over events where police were not present. When police opened fire on protesters, the subsequent rate of collective action increased by 107%.

What explains this dynamic? Few people will differ from me when I say that the South African Police has been, and to a large extent still are, political agents of a white minority government. They had to enforce unpopular pieces of legislation over a long period of time which brought them into constant contact with large proportions of the South African population, particularly members of the black communities.9 These laws include the pass laws, land laws, forced removals, colour bars, the migrant labour system, and measures under states of emergencies. These measures are, in Wilson and Ramphele's (1989) view, the major components of a systematic assault by the state on the poor.

Large numbers of people were detained under security legislation since the early sixties. During these contacts situations developed which were not favourable for the way in which the police are viewed by particularly members of the black communities. Different levels of physical violence developed frequently during these contacts which resulted in many deaths and injuries.

According to the Human Rights Committee (HRC) (1990) over 200 people have been killed and over 2 000 injured, either directly or indirectly, as a result of police action against gatherings for which permission was refused, or which occurred spontaneously, between 2 February and August 1990. Virtually all of these gatherings were for the purpose of expressing legitimate grievances around the issues of housing, education, health services and infrastructure.

In his study of detentions in South Africa, Foster (1987) refers to the physical and psychological process of detention.10 Such a process starts with the actual arrest and covers conditions of confinement and interrogation, and includes also the effects of such conditions. He argues that the package of laws under which people are detained is "… used as a form of political and psychological violence …" (Foster 1987:5). Detention under South Africa's security legislation is, according to Foster, above all a political act. He found that at least 70 000 people have been detained without trial since the 1960s.11 His results showed that of the 175 people in his sample (all held under security detention), 145 suffered from some form of torture. In fact, his results led him to conclude that torture, in terms of both physical and psychological abuse, is a relatively standard procedure in South African prisons.

At least 74 people died while in detention during the period 1963 to 1985, some under questionable circumstances.12 Official explanations of these deaths varied from natural causes, suicide, injuries, suffocation, falling from windows, to death due to torture. In some cases no reasons were disclosed (Meli 1988). Steve Biko, who died of massive head injuries on 12 September 1977, and Neil Aggett, who was found hanged in his cell on 5 February 1982, are well-known cases.

My research also investigated the effects of detentions on levels of collective action during the period 1970-1984 (Olivier 1989b). By removing the leaders from the community, the state would argue, levels of protest should decline. My results suggest otherwise.

As in the case of police action at events of collective action, I investigated the effects of detentions on levels of collective action subsequent to the detentions. As the number of people in detention increased from its lowest to its highest level during the 1970-84 period, the subsequent level of collective action increased seven times.

What explains this dynamic? One possible explanation is that those who were detained were replaced by less skilled leaders who had less control over their followers. This view is supported by the results of Jochelson's (1990) study of Alexandra.


These results show that actions by the state and its agents which were devised to curb the incidence of civil unrest did not have the desired effects. It was suggested that the high degree to which police activities had been politicised partially explains this dynamic.

Recent months saw a number of changes in legislation and policy which will contribute to the depoliticisation of the police force. A truly impartial police force, and one which is viewed by the people as such, is still an ideal. It should also be recognised that the legacy of apartheid will last for many years after its demise.

During his 1990 Christmas and New Year's address, President de Klerk had the following to say on violence:

No-one has the right to use violence against others to advance their political objectives. Those who have political differences should resolve them through negotiations.13

This, in my view, applies as much to the state as it applies to any other organisation or individual. Violence committed by any individual, structure or body is unacceptable and has no place in any democratic society. Who then has the responsibility for curbing violence in society? Allow me to answer this question in the context of protest activities.

Protest activity, within certain limits, only became an acceptable practice in South Africa during 1989. It can therefore be said that South Africans are, to a large extent, new to protest. This is true for protesters who air their grievances on any of a variety of social issues, and for the state. This new situation calls for a redefinition of the responsibilities of all parties involved.

In his evidence before the Goldstone Commission of Enquiry into the events at Sebokeng in March 1990, Colonel van der Merwe pointed out that the police are in a learning phase in respect of mass demonstrations (Goldstone 1990). The same is true for the organisers of mass demonstrations - a situation which places responsibilities on all parties.

Both the state and organisers of protest activities share the responsibility for preventing violence. However, as pointed out above, the state has a monopoly over the use of the tools of violence. This fact places a greater responsibility on the state to ensure that it, and its agents, act with restraint.

Any democratic government should realise that while it has a monopoly over the use of violence, it is in power only by the grace of its citizens. Accordingly the government should recognise that its task is not to govern the people, rather its task is to govern on behalf of the people.


1 See Lewis (1972); Adamek and Lewis (1973).

2 See Time, 19 June 1989.

3 See Time, 22 October 1990; The Economist, 13 October 1990.

4 According to the findings of a commission of enquiry into this incident, twenty people died as a result of police action (Kannemeyer, 1985).

5 See Harris (1988).

6 See Goldstone (1990)

7 See also Garver (1970).

8 See also Tilly et al (1975).

9 I use the term "black" in its broadest definition to include Africans, Indians and members of the so-called "coloured" community.

10 See also Riekert (1985).

11 The Human Rights Commission (1990) put this figure at nearly 80 000. Two-thirds of these detentions have occurred in the last 5 years.

12 Van der Vyver (1988) points out that since the inception of detention without trial, more political detainees have died while in detention than the total number of civilians who lost their lives as a result of urban terrorism.

13 Message by the State President, Mr F W de Klerk, 18 December 1990.


Adamek, R J and J M Lewis. 1973. Social Control Violence and Radicalization: The Kent State Case. Social Forces, 51(3):342-347.

Black Sash. 1988. Greenflies: Municipal police in the Eastern Cape. Mowbray: Author.

Cock, J. 1990. Political Violence. In: B McKendrick and Wilma Hoffman (eds.), People and Violence in South Africa. Cape Town: Oxford University Press.

Cole, Josette. 1987. Crossroads: The Politics of Reform and Repression 1976-1986. Johannesburg: Ravan Press.

Davies, J G. 1976. Christians, Politics and Violent Revolution. London: SCM.

Degenaar, J J. 1980. The Concept Violence. Politikon, 7(1):14-27.

Foster, D. 1987. Detention and Torture in South Africa. London: James Currey.

Galtung, J. 1969. Violence, Peace and Peace Research. Journal of Peace Research, 6:167-191.

Garver, N. 1970. What violence is. In: T Rose (ed.), Violence in America. New York: Vintage.

Goldstone, R J. 1990. Report of the Commission of Enquiry into the Incidents at Sebokeng, Boipatong, Lekoa, Sharpeville and Evaton. Johannesburg: Author.

Gurr, T R. 1970. Why Men rebel. Princeton: Princeton University Press.

Harris, Sarah. 1988. A Day that Made History: Sharpeville. London: Dryad Press.

Hibbs, D A. 1973. Mass Political Violence: A ross-national causal analysis. New York: John Wiley.

Human Rights Committee. 1990. State Violence: A study of repression. Johannesburg: Author (Report SR-8).

Huntington, S P. 1968. Political Order in Changing Societies. New Haven: Yale University Press.

Jochelson, K. 1990. Reform, repression and resistance in South Africa: A case study of Alexandra Township, 1979-1989. Journal of South African Studies, 16(1): 1-32.

Kannemeyer, D D V. 1985. Report of the Commission Appointed to Inquire into the Incident which occurred on 21 March 1985 at Uitenhage. Cape Town: Cape and Transvaal Printers (Repotr RP 74-1985).

Kotze, D A. 1978. Development and Structural Violence. Politikon, 5(1):30-41.

Lewis, J M. 1972. A Study of the Kent State Incident using Smelser's Theory of Collective Behavior. Sociological Inquiry, 42(2):87-96.

McKendrick, B and Wilma Hoffmann (eds.). 1990. People and Violence in South Africa. Cape Town: Oxford University Press.

Meli, F. 1988. A History of the ANC: South Africa belongs to us. London: James Curry.

Nardin, Terry. 1971. Violence and the State: A Critique of Empirical Political Theory. Beverly Hills, CA: Sage Publications.

Olivier, J L. 1989a. Collective Violence in South Africa: A Study of Ethnic Collective Action in the Pretoria-Witwatersrand-Vaal Triangle area, 1970-84. Ithaca, NY: Cornell University (Ph.D.Dissertation)

Olivier, J L. 1989b. State repression and collective action in South Africa, 1970 to 1984. Paper presented at the annual meeting of the American Sociological Association, San Francisco, August 1989.

Olivier, J L. 1990. Causes of Ethnic Collective Action in the Pretoria-Witwatersrand-Vaal Triangle, 1970-1984. South African Sociological Review, 2(2):89-108.

Riekert, J. 1985. The Silent Scream: Detention without Trial, Solitary confinement, and Evidence in South African "Security Law" Trials. South African Journal on Human Rights, 1:245-250.

Smelser, N J. 1962. Theory of collective behavior. New York: Free Press.

Tilly C, L Tilly and R Tilly. 1975. Rebellious Century, 1830-1930. Cambridge, Mass: Harvard University Press.

Van der Merwe, H W. 1989. Pursuing Justice and Peace in South Africa. New York: Routledge.

Van der Vyver, J D. 1988. State-Sponsored Terror Violence. South African Journal on Human Rights. 4(1):55-75.

Wilson, F & M Ramphele. 1989. Uprooting Poverty: The South African Challenge. Cape Town: David Philip.

Zimmerman, E. 1980. Macro-Comparative Research on Political Protest. In: T R Gurr (ed.), Handbook of Political Conflict. London: Free Press.

Zimmerman, E. 1987. Political Violence and Other Strategies of Opposition Movements. Journal of International Affairs, 40(2): 325-352.

© Centre for the Study of Violence and Reconciliation

Copyright © 2020 CSVR. All Rights Reserved.