The Human Face of Violence: Hostel dwellers speak.

The Human Face of Violence: Hostel dwellers speak.

Segal, L. (1991). The Human Face of Violence: Hostel dwellers speak. In The Journal of Southern African Studies, Vol. 18, No. 1, March.


In the Journal of Southern African Studies, Vol. 18, No. 1, pp. 190-231, March 1991.

Lauren Segal is a former Researcher at the Centre for the Study of Violence and Reconciliation.

This research would not have been possible without the assistance and cooperation of the managers and employees at the factory at which the interviews were conducted. I am grateful to Ronald, Richard, Jerry, Sihle and Steve who spent many hours interviewing and transcribing. I would also like to thank all members of the Centre for their valuable input and discussions on the subject, and to Graeme and Jonny for their editorial assistance and support. Discussions with Peter Delius were also most illuminating and Jean Leger helped with the interviewing schedule.

In the latter half of 1990 there was a rapid spiralling of violence in townships across the Witwatersrand. Hostel dwellers were a key group of protagonists in the bloody internecine clashes which still continue. Very few explanations, however, have penetrated beyond superficial explanations of this group's involvement in the violence. This paper, which contains excerpts from interviews with hostel dwellers, seeks to unveil some of the reasons behind their participation, and so presents the 'human face of violence'. It attempts to contextualise direct questions about the violence within, first, the broader historical conditions that have characterised the hostel community and, second, the broader political processes that have had an impact on the lives of hostel dwellers. While the small sample of interviews prevents any generalisations about the causes of the violence from being made, the interviews provide useful insights, as well as starting points, for a more extensive investigation of the subject.


Since July 1990 there has been a resurgence of violence in townships on the Reef, in particular in the East Rand townships of Thokoza and Kathlehong. The 'Transvaal war', as it has become known, has already claimed thousands of lives. Not surprisingly a wide range of explanations, accusations and counter-accusations concerning the major causes of, and protagonists in, the violence have been offered.

Hostel dwellers have been, and remain, an important group of protagonists in the violence that convulsed the East Rand. This is not the first time that hostel dwellers have been in the centre of township mayhem. The riots at Dube hostel in 1957 involving a vicious clash between a predominantly Zulu-speaking hostel population and the settled inhabitants of Meadowlands,1 and the involvement of the Mzimhlope hostel dwellers against students and township residents in the 1976 uprising, amongst other incidents, foreshadowed the clashes that we witness today.

Despite these historical precedents, hostel dwellers have remained either anonymous in the urban landscape, or have been assigned a semi-pariah status. Their present emergence as key actors has likewise been characterised by shallow attempts to understand their plight. In particular, for example, Zulu-speaking hostel dwellers have been described alternatively as 'warriors', 'blood-thirsty impis',2 as killers who 'proudly display the bodies of people they have battered to death'3 or who 'wave the parts of dead Xhosas in victory dances'.4

Thus, while political events may have reversed hostel dwellers' anonymous status, we remain no closer to an understanding of who these people are, or to the reasons for their violent actions. This paper attempts to venture beyond the ubiquitous and clich‚d image of hostel dwellers as 'Zulu impis moving through the streets of East Rand townships beating their axes and chanting war cries',5 and attempts to understand the role of hostel dwellers in the present violence through their own eyes. It will thus be investigating the hostel war from the perspective of the attitudes, beliefs and actions of individuals who have been involved.

Although this paper does not focus on structural aspects of hostel violence, it is important to stress that violence is endemic to the system of hostels. Structural elements such as the migrant labour system, single accommodation, and insalubrious living conditions, all contribute to and actively promote violence. It is another sign of the invisibility of the hostels that they are in the spotlight only when there is manifestly violent behaviour in the system. Suddenly the horrors of hostel life become a central point of concern in the media as well as in political circles. And yet the conditions in the hostels have, in reality, been health hazards long before FW de Klerk's trip into the township declaring them as such.

Explaining Violence An Ever Changing Kaleidoscope

There are always vested interests in explanations of social reality. This is especially true when one is dealing with the highly complex and sensitive issue of political violence. In almost all cases, the 'truth' that is presented about the violence is merely a story refracted through a prism of political interest groups and factions. As a correspondent of The New York Times working at the height of the civil war in Beirut said: 'There is no truth in Beirut, only versions'.

This is not to say that one cannot ascribe culpability to particular parties, or that some versions are not more accurate than others. Both The Weekly Mail and The New Nation newspapers have recently uncovered concrete evidence which shows that the state has aided Inkatha in organising anti-ANC activities and suggests that the SADF has been directly involved in the train massacres.6 This evidence supports some of the conspiracy theories accusing the government and Inkatha of colluding in the violence.

While subjective explanations are often motivated by self-serving political interests, this need not always be the case. In many cases such explanations emerge from the need to impose an official-sounding order on the overwhelming confusion and horror of the violence; the need to create light where in fact there is none. But the reports written either in the name of political ideology, or in the name of order, do not always reflect the wider sentiments – or even the expressed enunciations – of the people on the ground.

This disjuncture between the perceptions of those involved in political violence, and the coherent explanation of their actions advanced by political organisations, the press and others, is very clearly silhouetted through interviews with hostel dwellers who have participated in the recent violence.

Rather than portraying easy and defined 'truths' about 'what happened', the interviews conducted reveal a shifting assortment, a 'kaleidoscope' of explanations.7 The apparent incoherence of the explanations at times indicates that the view of violence as constructed from below, the human face of the violence, is far more diffuse and complex than most media or political accounts portray. It becomes clear that the actors in the violence are motivated by a host of factors, which at times intersect with, and at other times radically diverge from, popular explanations or expressed political party lines. These explanations of the violence from the people themselves are often contradictory, partly, as will be shown, because there are a range of traditions and ideologies (themselves often contradictory) which may operate simultaneously in an individual's life.

The most important lesson to be gained from talking to the actors themselves is, therefore, that explanations of violence are more complex than those articulated at the public level. The 'human face' of the violence forces us to move beyond the images described in the introduction of this paper.

There are, however, methodological problems of 'basing a theory of social action on the actors' own perceptions'.8 The process of talking only to the actors as a necessary corrective to public accounts and explanations, may result in the stick being bent too far in the opposite direction. The effect of this may be to overlook external political processes and the ways in which these impact on, and shape, individual responses. The interviews do nonetheless suggest routes for enquiry into these processes. Ultimately, it is the sometimes elusive intersection between collective experience and individual perception that allows one to arrive at an explanation of the violence.

The limited number of interviews, their regional specificity, and the fact that most of the interviewees were Zulu speakers, mitigates against generalising the findings of this study to the wider hostel-dweller population. This study does not, therefore, attempt to draw any watertight conclusions. Nor does it contend that the interviews necessarily reflect broader social processes. While this paper does not deliver complete historical truths, it does present the subjective accounts of a group of key actors in the current violence and, as such, is the embryo of an 'experiential mosaic'.9 It will doubtless be added to and radically altered through the addition of new, expanded and differing pieces of the wider picture.

The starting point of the next study might be to interview those people who are spoken about, but whose own voices are not heard in this study, namely, Xhosa-speaking hostel dwellers who have fled the hostels and the factories in the Thokoza area, the residents living in the townships, local political leadership and union officials. An entirely different version, a different truth, would almost certainly emerge from such interviews.

Literature Review

The system of migrant labour has been the subject of several historical and sociological studies.10 In particular, the compounds on the mines have been scrutinised as an example of 'total' or 'carceral' institutions and as pivotal forms of social control of the African workforce11 These studies emphasise functional aspects of the system, as well as the informal ways in which this control was contested.12

By contrast, inner city hostels, which were the logical outcome of the apartheid policy – designed as it was to prevent African people from settling in the urban areas on a permanent basis – have been sorely neglected.13 By an accident of research, the hostels in Cape Town have been under a far more intense sociological gaze than those on the Rand.14 Given the sudden prominence of the inner city hostels in the Transvaal war, and the wide spectrum of people who are grappling to understand their rapid transformation into military barracks, we have been forced to recognise and attempt to overcome this oversight.

Of the few published studies on the hostels in the Transvaal most focus on Johannesburg and are essentially descriptive.15 They point to the fact that while the inner city hostels were also designed as sites of control for the urban migrant population, they differ from the hostels on the mines in four important ways.

Firstly, government bureaucracies, rather than a single employer, are responsible for their management and upkeep. Secondly, they house workers who do not necessarily work for the same company or work the same hours or shifts. Thirdly, ethnicity is 'not the official organising principal of this world … and men from a variety of backgrounds [can] and [do] share rooms'.16 Fourthly, and most importantly, they do not exercise the same 'total control' or 'hegemony' over their inmates as do the mining compounds. Although both may be classified as 'carceral institutions', the urban hostel dwellers are a more diffuse grouping than their mining counterparts. They are embedded in an urban context which allows for a broader and more complex set of relations to develop with outside communities. This contrasts with the extreme isolation of migrants on the mines.

Sitas thus speaks of urban hostels as sites of 'diffusionary control' without 'the level of rationalisation achieved in the mining compounds'.17 Others speak of urban hostels as being 'more open' and 'less regimented'. This has led some writers to characterise hostel dwellers as 'men of four worlds'; the hostels, the black townships, their places of work as well as their rural homes.18

In the main, however, the political significance of this grouping and the deep-rooted cleavages that have historically divided hostel dwellers and township residents have either been overlooked or ignored by writers on South African history. With hindsight it is also arguable that those authors who have pin-pointed the importance of ethnic ties and political factors in hostel life have not fully assessed their significance in broader community struggles.19

There are some exceptions to this trend which should be mentioned. Delius's paper on Sebatakgomo, a rural migrant organisation operating in the 1940s, shows how migrants in the urban hostels provided an 'effective bridge between the ANC, SACP and rural politics'.20 In a more recent paper he explores how migrant hostel dwellers, who have gained trade union experience in the industrial setting, have been increasingly assertive in village affairs and have helped to cement an alliance with youth in the area.21 Mthetwa's work on hostel dwellers and their participation in community struggles in Alexandra township, sheds important light on the question of the mobilisation of the migrant populations during community struggles.22 Sitas's thesis only touches briefly on political issues; the thrust of his work is on unions and production.

The importance of hostels in the urban environment has not only been overlooked by academic writers. Political leaders have also overlooked their distinct identities and specific needs thereby turning them into invisible actors in the political landscape. The result of this neglect is that hostel dwellers and their struggles have commonly been subsumed under the general category of 'community struggles'. This trend will be examined in greater depth below.


The extent of the violence on the Rand precluded the possibility of interviewing in depth in the hostels of Thokoza and Kathlehong. We therefore had to find an alternative means of gaining access to this key group of participants in the violence. This was done by conducting interviews at a factory in the Thokoza area in which a large number of hostel dwellers are employed. After a lengthy process of negotiation with the hostel committee in the factory, it was agreed that we could conduct the interviews.

Interviewees were selected by the leader of the hostel committee in the factory and this may have introduced a bias within the sample selected. For example, almost all the people came from either Emhlabathini or Kwa Nongoma, both of which areas are key Inkatha strongholds in rural Natal where chiefly control appears to have remained intact. Had another group of nineteen individuals been interviewed, a different picture may well have emerged.

Six out of the nineteen interviewed were formerly resident in Khalanyoni hostel blocks (numbers four and five), which are situated next to Phola Park squatter camp and which were burnt to the ground during the fighting. These six subsequently moved to the hostel in Thokoza. This hostel is divided into three blocks: Mdala (number one), Khutuza (number two), and Unshayasafe (number three). Two interviewees were living at Kwesine hostel in Natalspruit, just a few kilometres from the Thokoza hostels. One further interviewee lived in the hostel in Vosloorus. Most of the hostel dwellers had lived and worked in Johannesburg for at least ten years (5 for over 30 years, and 4 between 15 and 20 years). Their ages ranged from 25 to 60, with most falling in the category of 40-60 years old.

The interviewing procedure itself posed several challenges. Despite the protracted negotiations with the hostel-dweller committee prior to the interviews, there was, without exception, a high degree of suspicion about out intentions. This is hardly surprising given the political context. Many of the hostel dwellers expressed the belief that we were trying to elicit information to give to the squatters or to the ANC. The presence of a tape recorder aroused further suspicions in this regard. Matters were mad worse by the fact that many of the hostel dwellers had seen volunteers from a student organisation at Wits university helping the Phola Park squatters rebuild their shacks. Many wanted to know why Wits wasn't helping to rebuild their hostel (Khalanyoni) which had also been destroyed. The more hostel dwellers in the factory got to hear of our work, the more suspicion and reluctance there was by individuals to be interviewed.

Generational issues were also an important consideration during the interviews. Most of the hostel dwellers we interviewed were elderly men, deeply rooted in rural traditions and ideology. The fact that the interviewers were young men meant that some of the questions were difficult to ask without breaching the boundaries of respectability. The presence of a white woman was another influencing variable in the process.

In general, the interviewing process presented several concerns for sociological research, in particular, ethical considerations of conducting research in beleaguered communities. There was a strong expectation amongst the hostel dwellers that the problems that they faced were at last going to be addressed and there were constant enquiries as to what practical assistance we were going to provide them.. Despite making our intentions very clear, this remained a central concern during the weeks at the factory.

Because most the interviews were conducted in Zulu and then transcribed into English, several further problems were encountered. Unspoken aspects of the interviewing process, namely the tone in which questions were answered, the silences and difficulties in answering some of the questions, are a vital source of information and a valuable reflection of a person's mindset. They are however, difficult elements to incorporate into a transcript. I tried to capture the unwritten aspects of the interviewing process by sitting in on the interviews and by asking the interviewees to describe the feelings that they encountered.

Who are the Hostel Dwellers and What are Their World Views?

In order to grasp the meaning of, and motivation for, the violence amongst the hostel dwellers, it became apparent that we needed to step outside of our own particular cultural prism. Within the urban environment, migrant hostel dwellers have developed their own peculiar set of beliefs and ways of seeing the world, their own 'unique cultural formations',23 which filter their perceptions of why and how they act. These cultural formations have been forged through their years of living as migrants, and through the specific historical exigencies with which they have had to contend. Part of this paper therefore focuses on the worldviews of the actors themselves. This provides a vital backdrop to the violence and, it is hoped, will allow us to see the present situation through a less jaundiced lens. Four main areas are explored, namely, community relations, hostel-dweller associations, rural connections and age determinants amongst hostel dwellers.

Relationship Between Township Residents and Hostel Dwellers

Hostels are geographically isolated. They are literally on 'the edge' of society and hostel dwellers are visibly located as outsiders. Historically, this geographical marginalisation mirrored their political marginalisation. In Sitas's study on hostel dwellers in Vosloorus township on the East Rand, the township emerged as an alien environment for hostel dwellers with most of those interviewed having no friendship ties in the township community. He spoke of the isolation of migrant workers not only being objectively defined 'by the structures of apartheid' but as being 'internalised and subjectively experienced by the migrants'.24 The extent of this marginalisation seems to have changed somewhat since than as hostel dwellers have been integrated to a greater degree with the wider community.

Thus, Mthetwa, in his recent study on Alexandra, remarked on a blurring of the clear distinction between hostel and the rest of the community:

there is a process of intermingling between some of the hostel residents and the broader outside environment of the township. To the extent that this intermingling occurs, it remains contrary to the thesis of 'total' exclusionist tendencies of total institutions.

He qualifies this by stating that the interaction of hostel residents is with the 'lowest stratum of urban residents', namely squatters and tenants, and that in general hostel dwellers remain marginal 'from the ensemble of social associations in the township'.25

The interviewees in this study reflected the fact that hostel dwellers are not totally isolated from township life, and almost all had several points of contact with township residents. Many even perceived themselves as members of the township rather than just the hostel: 'I see myself as part of the whole township. The hostel is just a place I am using for sleeping. I am in the location for the whole day on the weekend'. (8)

There are three identifiable levels at which contact with the township takes place. Friendship networks are the predominant reason for going into the township:

We have lived together with township residents for a long time There is no problem between us. I have many friends and relatives in the township. If my wife visits me I go to the township to ask for temporary accommodation whilst my wife is around.

Besides these friendship networks, the other forms of associations which featured prominently were the church and political meetings. Church communities were seen as straddling the hostel dweller/township resident, as well as ethnic, divides: 'Well my team is the churches. Whether Umsuthu or UmZulu we just join. I go to church in the location'. (4) The political meetings will be discussed in some detail later. Suffice here to say that they are another important point of intersection between the residents and hostel dwellers.

The hostel dwellers' view of themselves as members of the township community has developed in spite of the fact that these feelings are not perceived to be reciprocated by most of the township residents. Thus, when asked about how the township residents viewed them, hostel dwellers were unequivocal in pointing to the animosity and antipathy of township residents. In doing so, they did not appear to recognise the contradiction of regarding themselves as members of a community from which they were clearly ostracised. The contradictory thrust of their explanations are embodied in this description: 'I see myself as a member of the township in which I live. I have friends in the township. They visit me and I also visit them'.
And then he added:

When we are in township we have to hide from them that we are hostel dwellers. It is not safe at all. If you are unfortunate to be known that you are hostel dweller they kill you. (9)

They see us as inyamazane (animals). They say that we are the ones responsible for the violence that is taking place … . They don't even trust you even if you live alongside them. My brothers and relatives cannot come here and spend many days. (3)

A more nuanced account suggests that the residents who live nearer the hostels are friendlier and identify with the hostel dwellers. But these people 'who are better, they are such a small clique you can't really count them. With the majority its different'. (12)

These perceived antagonisms of township residents take on real and material forms when it comes to community struggles as will be explored below.

Forms of Association Amongst Hostel Dwellers

Within the hostels themselves, the predominant form of association was the burial societies. Although some of these were described as village-based structures under the control of a chief, membership appeared to be fluid and was expressed as being open to 'anyone who wanted to join', or 'open to all different tribes'. Some were members of two burial societies – one from home and another more general hostel-based society. Prior to the violence, burial societies did not therefore necessarily function as particularist forms of association or as a means of fostering ethnic divisions.

The hostel-dweller committees, responsible for the upkeep of the hostels, were likewise points of contact, rather than division, for hostel dwellers from diverse places. They were comprised of indunas from each room and were responsible for 'laying our problems to the municipality'. The indunas were elected according to how 'brave' and 'responsible', they were not according to political or ethnic allegiances.

Perpetuation of the Rural Social Network

The unchanging variable in studies amongst migrants is their rural orientation. It has been repeatedly demonstrated that migrants do not leave behind the countryside in their journey to the city. Their rural consciousness is deeply internalised, and their responses to happenings in the city are strongly determined by an ingrained set of values and expectations.

Why has the rural homestead remained so central to a migrant's consciousness? Sitas argued that in the 1970s, the strengthening of influx control and the homeland system underpinned the strong rural orientation of migrants. In the 1990s, formal influx control is no longer operational and the credibility of homeland structures has been radically undermined. And yet the importance of rural attachments persist.

In attempting to understand the continued importance and attraction of the rural world we asked hostel dwellers to tell us about how their lives in the country differed from the city and which they prefer. Most answers were uniform and included references to the expenses of city life, and the lack of freedom in the city. Only one person wanted to bring his family to the city and could see any advantages to living in the city. By contrast, the country was described in positive ways:

… No I think it is better at home. Here if you wake up in the morning, if you take one step you must pay. At home you don't do that. At home you don't pay rent. It's very expensive to live here. (8)

At home … there are things we can get from our gardens and we can slaughter a cow if we want meat and so on. But here, you buy everything. (9)

Here the houses are too close to one another which creates tensions … (10)

In Johannesburg, you are actually being enslaved … (14)

I don't want them to live in the township because I've got too much family. My father has two wives, I've got one wife and still trying to get another. At home, I've got cows and sheep, fields, I've got all the space I need. (3)

Despite the perception of some of the hostel dwellers that they were part of the township community, this perception is coupled with a more general disapproval of township lifestyle and culture. This is evidenced in the way in which many of the hostel dwellers describe township residents:

Township residents do not impress me from adults till to (sic) children. Respect is unknown amongst them. Sometimes you hear that comrades have assaulted the father of the house That is not human, it's very barbaric. It is an insult to humanity. Nature doesn't allow that a child can beat an elderly person, under no circumstances. (19)

And from another perspective:

The problem with urban people is that they want to take the law into their own hands, that's what I think is wrong Even if you try to explain to a township person that the law has to be respected he won't listen to you. (7)

I hate this place. It is rotten. The people who live here are uncultured. There is nothing which impresses me about this place. (19)

This study shows again that Johannesburg is not 'home'. On average, the hostel dwellers reported that they returned to their 'real' homes at least once a month. A third of the sample returned home at least twice a month.26 Rural ties were kept alive and channels of communication remained open.

This resilient rural connection amongst Zulu migrant hostel dwellers has been cited as an important factor in the leap-frog of the war from rural Natal into the urban centres of the Transvaal.27 It may well be responsible for the contact between the two 'war zones' and so acting as a vehicle for the transportation of many of the conflicts which have fed into the Transvaal war.

Generational Issues

The urban/rural divide was often mirrored by generational issues which featured prominently as a source of division and hostility for the hostel dwellers. The way in which age divisions have operated historically in hostels is important to understanding this issue. Over the years, it is the older workers in the hostels who have come to understand the complexities of urban life. This understanding functions as a 'survival kit' for the younger newcomers; it is the elders who initiate the younger migrants into the struggles and complexities of the city by sharing their knowledge of job opportunities, survival strategies, and so on. As a result, a gerontocratic form of control has been established in the hostels. This ensures that 'each migrant [is] constituted as a public person with duties, obligations and rights in the hostel … and that the past as transmitted by elders, acts as a judge on the present'.28

The role of elders in the hostels coincides with the gerontocratic forms of control in the rural areas. Inkatha has skilfully drawn on this tradition by blending an ethnic ideology with one based on age into its politics. This is manifest in the historically antagonistic relationship that Inkatha has had with urban youth, as witnessed, for example, in the stand it took against the children of 1976, as well as in their attack on the University of Zululand students of 1983.

In terms of township politics, hostel dwellers categorically place the youth at the centre of their alienation from the township:

The problem is that the comrades back up the children. They fully support them. Take for instance a meeting for parents; you find that whatever is being discussed, it can be rent, water, electricity, they will go back home and inform their children on decisions taken in the meetings If the meeting was held at 10 am, just in the afternoon around four you will see children rioting. Just from that programme we felt that no, this is not the way we should conduct things. So consequently we opted for division. (4)

Parents should stop their children from provoking hostel dwellers. (1)

What makes the people from the hostel not to like the youth from the location is because of that unnecessary stayaways, unnecessary boycotts. If there is a consumer boycott in the location, you cannot see an adult standing on the road, you only see the teenagers who are the forefronters. (5)

These tensions have led some authors on the current violence to conclude that the conflict has been between urban youth and rurally-based and traditional hostel inmates.29 The picture however is more complex than this neat division suggests. Age divisions within the hostel community continued to play a noticeable role during the violence, but not in the conventional sense outlined above. During the interviews, it was the elders who expressed more considered and conservative opinions. They were more ready to see themselves as mediators, voices of authority and rationality, and were less keen to rush onto the battlefield. The violence has thus to some extent upset the balance of gerontocratic rule as the younger hostel dwellers have taken the lead in organising the violence. 'The elderly people are the people who stay behind and guard the hostel'; (9) thus being assigned to a more passive role in hostel politics. This reflects a significant shift in the traditional locus of control within the hostel community itself a shift which, it will be argued, has been significant in entrenching the divide between hostel dwellers and township residents in the context of urban-based political conflict from mid-1990.

The Broader Context

We have attempted in broad strokes, to construct a picture of the individual consciousness of the hostel dweller. We are thus able to develop some sense of the hostel dwellers' perspective on their environment. Among the major perceptions are: a township community which, despite forms of social involvement is alienating and politically hostile for hostel dwellers: an expensive and corrupt urban world where money determines one's well being and which is juxtaposed with the 'freedom' of the countryside; an over-confident and arrogant group of young people who attempt to dictate the terms of one's existence; and finally, a rural life which is far away, but in nonetheless regarded as home.

It is now necessary to turn to some of the broader socio-political and economic processes that have shaped migrants' existence. These 'macro' developments afford us a more complex understanding of the specific position of hostel dwellers at the time of the outbreak of violence in 1990.

A Picture of Migrancy on the East Rand in the 1970s

The migrant labour system in the 1970s is described by Sitas30 as having entered its second major crisis (the first being in the 1940s). Migrants spoke of having a 'lean social existence' and of facing a series of new pressures.31 This crisis centred around the collapse of the rural economy and the snowball effect that this had for them in the cities. In particular, the new group of 'illegals' arriving at the hostels increased the number of raids of the 'blackjacks', or municipal police whose function it was to end this process of urbanisation.32 These battles, together with a growing criminal sub-culture, the greater pressure on already scarce resources, and deteriorating wage packets, all served to increase intra-hostel aggression and the spread of physical violence in the hostels.

Interwoven within these dynamics, there was also a new political mood developing. Sitas claims that, despite the absence of active political participation, the question of homeland independence directly and tangibly affected hostel dwellers and started to enter into their conversations. It was here that Inkatha was mentioned for the first time as starting to 'open branches on the East Rand and [as] calling for mass rallies'.33 The responses of the hostel dwellers to these changes in the 1970s provides an important comparative perspective to the 1980s and early 1990s.

According to Sitas, the cumulative effect of these changes was to 'transform working class life and interaction on the East Rand'.34 The development of union organisation in the factories was, Sitas's opinion, the ultimate catalyst of these changes. In the factories:

leadership arose that was not hostel or regionally based: this leadership by taking on issues at work, shifted the space association from the hostel back to the factory. It cut across regionalism and ethnicity by uniting producers at the point of production. With the arrival of unions, legitimation shifted from the gerontocracies to rank and file control, where elected representatives became the new leaders, and public discussion on every issue entrenched itself as the public form of decision making. Behind the survival of old relations, lurked a tremendous transformation. (my emphasis)35

The rapid unionisation amongst migrant metalworkers on the East Rand was said to have had massive consequences for all hostel dwellers. The new-found unity based on mass mobilisation on the factory purportedly penetrated hostel life.

The impact of this was four-fold: there was a growing disgruntlement with conditions in the hostels and secondly an 'us and them' mentality developed as those who were not part of the unions were referred to as impimpis (informers). A new militancy emerged in the hostels with the rise of younger, more determined union leaders with the result that new power relations were generated. Most importantly, issues of race and ethnicity were seen to have been subsumed under a new working-class perspective.

Sitas then generalises to the East Rand where he speaks of a 'moral revolution' having occurred amongst hostel dwellers, as well as a reorganisation of cultural formations based on union participation. All of this led migrant hostel dwellers into the forefront of worker militancy during the strike wave on the East Rand. The distinction between urban and migrant workers was apparently dissolving as was the relationship between migrant trade unions and the community.

The situation in the hostels ten years later bears very little resemblance to the above descriptions. The question of what transpired in the intervening years is a vital one. It may explain why the supposed 'breakdown of cultural formations' appears to have been short-lived and why at present, the East Rand hostels can in no way be characterised as having a 'working-class leadership which cuts across regional and ethnic ties'. It may also explain why internecine violence, rather than working-class militancy, was the response of hostel dwellers to the crisis which they faced in the late 1980s.

This paper tentatively suggests three parallel developments which may have caused a turnabout in the nature of hostels. Firstly, there were changes in the style and operation of the union movement which unwittingly undermined migrants' positions within the union. Secondly, new political developments likewise isolated hostel dwellers. Lastly, the Abolition of Influx Control Act of 1986, ushered in a new period of crisis and was a watershed in hostel life. Each of these will be examined in turn.

Changes in the Union Movement

Although every interviewee is a member of Numsa (the sole union at this particular factory), the relationship of migrant workers to unions appears to have shifted dramatically in comparison with Sitas's portrayal. Thirty per cent of the interviewees stated that they had joined the union because they expected the union 'to solve their money problems'. The vast majority, however, (60 per cent) joined because 'it solves problems between us and employers'. With varying degrees of emphasis, people echoed the following sentiment:

We joined because whites treated us as play things … . Now if you come across a quarrel with a white man your case will be attended to … you're no longer a kaffir that will be shown to the gate … . Since I joined the union I found myself great help because even if I may be wrong I can now sit down and solve the matter peacefully through negotiations. (4)

The union was unanimously regarded as a beneficial force in the workplace. The benefits of belonging to the union were, however, restricted to the factory: 'The union only connects me with the employer, that is it'.(4)

In fact, the interviewees felt that hostel issues were directly neglected by the union. Unlike claims that union organisation penetrated and transformed hostel life on the East Rand in the early eighties, the link between union organisation and hostel life in the early nineties was shown to be virtually non-existent. Not a single interviewee could recall a union organiser coming to their hostel. Some appeared complacent about the issue and simply stated that 'The union does not ever come and hold meetings at our hostels. They operate at our workplace only'.(8) Others spoke out against this in tones which suggested animosity:

They (the union) have never held any meetings in our hostel. They didn't fear anything, its just that they didn't see the need and yet most of its members are hostel dwellers. All hostel dwellers are undermined. (13)

No, they don't come to our hostel. We meet here in the workplace, that's all. It's not because they are afraid to hold meetings in our hostel, it's just because they are not interested. That's why we had a conflict with them because they always went to meetings in the township and make decisions that would affect us without consulting us. That is especially during stayaways so that is what makes us not recognise their calls for stayaways. (19)

The interviews suggest that since the mid-1980s, hostels have no longer been as central to union organisation as in the late 1970s and early 1980s, and 'migrants who are unionised have experienced a degree of alienation.'36 The Congress of South African Trade Unions' (COSATU) specific style of mobilising new members in and around the hostels, no longer appears to have been operational. Although there was no deliberate policy to oust migrant leadership, this trend accompanied the rise of younger, more urban-based shop stewards. An outside observer aptly commented that, 'as the unions become more established, they have remained committed to focusing on factory floor issues' but 'they have ceased to rely on hostels as their organisational base'.37 Delius comments that in the 1980s, even those unions which were increasingly sensitive to community issues rarely gave organisational attention to 'the rather different difficulties confronting migrants in their home villages'.38

Evidence from hostel dwellers in Alexandra corroborates these findings. The interviews conducted there likewise showed that by the 1990s the hostel population were alienated from the central concerns of union organisation. The reasons offered in this case were again the presence of younger shop stewards who were less in touch with rural issues.39

Moreover, in this sample of interviews, unions were not perceived as a conduit to the wider world of politics. This became clear when we posed the question of COSATU's alliance with the ANC and SACP. Many responded to the question by stating that they saw the activities of the union as being restricted to the workplace and thus they were unaffected by, or saw no significance to this alliance. 'I'm only interested in my rights as a worker, that is all'. 'I only joined Numsa for their protection at the workplace and not their political alliance'. (18)

Six out of the thirteen people who answered this question, however, saw the alliance as a problem and denied knowing about it when they joined the union:

That is a serious problem, which gives us a serious problem as well with this union, because our beliefs are different … (17)

I don't support these things. I did not know there'd be such an alliance between the union and these political organisations. I wouldn't have been a member if I'd known. (11)

I'm not for that alliance … I only discovered about it after I joined … (12)

You see the time I joined, it was not involved in politics. It was an orthodox union. It changed subsequently and that was late for us to withdraw. (15)

When we first joined the union we were not aware that we were indulging into politics. But now we know that we are members of an enemy union We are still going to establish ours because we don't want to be seen as members of the ANC and the SACP. (19)

One person pointed directly to the issues of dual membership:

All the people who stay in our hostel who also work with me are Numsa members but their membership is only centred here in the workplace. When they get to the hostel they become Inkatha members. (9)

Most of these people seemed to, 'forget about these difference though when we are at work …' (7) Mention was made by one interviewee of starting an UWUSA branch at the factory.

The interview material suggests that either the changes wrought in the hostels in the early 1980s were not 'fundamental breakdowns in old cultural formations' as argued by Sitas, or that they have been dramatically reversed in the intervening years. Either way, by the late 1980s, the injection of factory floor cultural formations into hostel life no longer appeared to be occurring. Moreover, ethnicity was able to raise itself as a powerful force obscuring the commonality of working class interests. The interviews suggest that gerontocratic rule persisted, home-based allegiances remained intact, and union-based associations were restricted to the factory floor.

Political Developments

The hostel dwellers' feelings of alienation from the unions was mirrored by their sense of alienation from urban politics. The peripheralisation of rural and migrant issues has been a persistent thread of the struggle in South Africa. The particular style of politics that emerged in the early 1980s, as well as in the agendas of political organisations reinforced this trend as they both dragged the political process into an urban-centric discourse: town councillors, black local authorities and other issues relation to 'permanent urban residents' became the primary issues of the day. This focus, whether intentionally or not, wrote hostel dwellers out of the political script as organisations hardly attempted to, let alone succeeded in, 'translating rural and migrant experiences in to the national political programme'.40

It was ultimately in the cut and thrust of political action, however, that the battle lines between hostel dwellers and township residents were forged. In particular, it was the popular strategies of stayaways and rent boycotts which reinforced the hostel dwellers' marginalised status. In this respect, the hostel dwellers appear to be operating according to a basic political truism, namely that boycotts are a perfectly acceptable way of protesting if, and only if, they are carried out with the consent of those participating. What was abhorrent to the hostel dwellers was the fact that they had already settled their own rent boycott by the time the township boycott was initiated and that the township boycott was, from their perspective, enforced. Several interviewees repeated the sentiment that they were being 'poisoned' not to pay rent. The following words expressed a typical attitude:

We told them [township activists] that we were the first people to start the rent boycott. We didn't ask for your help, so do not interfere into our affairs and we also won't interfere with yours. (13)

Descriptions of their own rent boycott suggest that it was not an organised political strategy but rather an 'obvious' form of protest. The collective response emerged directly from deteriorating conditions and was indicative of the lack of 'total control' exerted by municipalities over the hostels:

In early 1990 there was no electricity and the water stops at any time. We stopped paying because we didn't see what our money was doing. Everyone stopped paying until they told us that we'd be chased away. It wasn't well organised but something bad just spreads easily in a short space of time. No one had to call and tell us to stop paying. Even a child would see that we mustn't pay. (8)

On the other hand, the reasons for the rent boycott in Thokoza township were far less apparent to the hostel dwellers and the civic association quickly came to be regarded with hostility as a result of its organisation of the township's boycott:

The civic association was formed to serve the interests of the Thokoza community. Our hostel committee tried to seek information about the aims and objectives of the civic association. This was after we had our rent boycott which came before theirs. When they decide on their boycott they just came to order us to join as well.

They were asked if their association was affiliated to any political organisation … . They said they were under the ANC. We said we could not be part of it then. Our reason was that we wanted a neutral body accommodating all people … . After that we heard rumours that the hostel must be demolished. These rumours were passed around in the taxis and everywhere. We didn't understand why they didn't speak about demolishing squatter camps. (13)

The only time anyone from the township will come and talk to you is to suck information and then leave to inform the komblese (comrades) saying that the hostel dwellers would not like us to go ahead with the boycott. (4)

The other pivotal source of division with the community was around work stayaways:

Another thing we cannot tolerate is to be intimidated not to go to work during stayaways. No one will stop us from going to work. We go to work even on the 16th of June. Anyone who tries to stop us will be disappointed of what he will see. We come to Johannesburg to work, not for anything else … . The ANC must speak to their supporters not to intimidate us, they should stop their youths burning cars with Natal registration numbers … (1)

In summary:

There is no one who ever came to inform us of anything in the hostel. We were forgotten about. That's where they [the Civic Association) went wrong. They didn't inform us of boycotts or stayaways. We asked them once if they represented the whole of Thokoza or not. They did not invite us in their meetings. We just see youth attacking us during boycotts and stayaways. (10)

Township activists' neglect of the hostels was evidenced again and again in the interviewees' responses to whether anyone came to the hostels to ask about the conditions and grievances in the hostels during the turbulent years of 1984 to 1990. No one could recall an activist from the township approaching their hostel to investigate the situation. Part of the hostel dwellers' reasoning for this neglect was again explained as the fact that township residents did 'not consider us to be normal human beings', or 'they still look at us as objects without brains'.(13 & 19) 'There is no one who comes to see us because the ANC is a group of criminals who do not want to come up front where everybody can see them. They do all their things secretly'.(19) 'The only time they come is to tell us not to go to work, but hostel dwellers refused because we came here to work'. (17)

The situation in Alexandra hostels in the 1984-7 period appears to mirror the East Rand.41 During the height of 'popular' protest in Alexandra township, Mthetwa shows how political activists failed to mobilise hostel dwellers: all three hostels, for example, were paying rent 'while there was a popular rent political struggle going on around them'. He argues that the use of terms like 'popular protest' prevents one from disaggregating 'specific social groups like hostel residents which were marginal in the conceived "communal" struggles'.42

The violence itself has brought these deeply rooted antagonisms between the two communities to a head. Many interviewees speak of relations with the township as differing 'before' and 'after' the violence: 'Before the violence, we used to be part of the community. We used to sit down with friends. We had very cordial relations. The violence has stopped that.' (6) This strongly suggests that the limited interaction between hostel and township communities was jeopardised firstly by the hostile political processes and secondly by the indiscriminate process of violent struggle. In this context, the scope for ethnic prejudice and political intolerance was greatly enhanced, allowing all of these factors to become self-perpetuating through the violence itself.

Abolition of Influx Control

The period after 1986 may well be deemed the migrant labour system's 'third major crisis'. Although the themes of the struggles of the 1970s continued into the 1980s, this article suggests that they took on a new character after 1986.

In brief, the abolition of influx control unleashed processes of social, economic and political dislocation for urban and rural communities alike. Although on paper the Act did not necessarily guarantee freedom of movement for Africans in urban areas, its effect was to spawn rapid urbanisation and a burgeoning squatter movement across the Rand. The massive inflow of people into the cities, and the concomitant struggles over urban space and resources, reshaped the contours of city life. More and more people were jostling for already inadequate places in the cities. While old cleavages ossified, new cleavages cut through urban communities.

The issue of space has long been a contested feature of hostel life. In hostels, 'the concept is of a "bed-holder" as opposed to a "house-holder"; this immediately introduces the politics of space, where people are limited to a bed as the only space over which they have some measure of control.'43 The interviews reveal that the 'politics of space' was further affected by the abolition of influx in two central ways.

(i) Increased Unemployment in the Hostels

In the words of one hostel dweller: 'After influx [control] was abolished, the number of people wanting to stay in the hostels increased as people could migrate from rural areas to towns regardless of employment'. (4) This factor is restated throughout the interviews with almost everyone repeating the sentiment that, 'After 1986, people started pouring in the hostels in large numbers'. (9)

The hostels, to an even greater extent than before, became shelters for the unemployed, and were often the first port of call for people first entering urban environments. Although respondents were unable to identify the exact number of unemployed people in their rooms, most of them spoke of 'about half' the room being unemployed. This also depended on the particular time of year; 'From November it was more than that because another guys who finished schooling came to stay'. (8)

Today, yes, the hostels are full. But we are the ones responsible for that. What we would often do is harbour our unemployed brothers. So during the night, they go to the extent of sleeping on the floor. (2)

The 'unemployed brothers' were not simply strangers who arrived in their rooms but were people who often had fairly intimate relationships with the hostel dwellers:

There are a lot of unemployed people; they all come to their relatives and friends who have got beds in the hostel. Sometimes I'm working on night shift, my relative would sleep on my bed, but otherwise he can sleep on the floor. They all look for jobs during the day. (7)

These unemployed people, they used to be people we knew from home. My cousin. This guy's brother. So there was no tension. We helped them buy food, give them money for transport, or to look for a job. They come to our hostel because they knew us. (8)

Although the hostel dwellers seldom expressed resentment towards the unemployed people in their rooms (perhaps because they were often people whom they knew), there were definite undertones of the difficulties that resulted: 'This [the increase in the number of people] has also created some problems when it comes to cooking. There are people who steal our food and we sometimes fight over the stove.' (7) Many spoke of giving money to these people for transport to look for jobs as well as sharing their food with them. In the context of a shortage of space and a scarcity of resources this undoubtably created further strains in the daily existence of hostel dwellers. Moreover, the general picture seemed to be that 'they usually didn't get a job. They can stay for a year without a job. Sometimes they sell beer or cooldrinks in the hostel to get cents.' A host of informal businesses developed in the hostels as people attempted to eke out a living.44 The informal sector, however, is incapable of absorbing all the newcomers and there remains a large group of unemployed. The omnipresence of this group acts as a daily reminder of the shortage of jobs and probably feeds into the reluctance of hostel dwellers, as described above, to participate in stayaways which threaten their jobs.

It is likely that those hostel dwellers who were neither friends nor relatives of the unemployed found them more intrusive in the hostel. These hostel dwellers operated according to the principle that when 'there [are] too many [people] for the available means of subsistence and recreation … the "other" grouping [in this case the unemployed] becomes a … potential enemy.'45

Many writers on the current political situation suggest that the violence would not have occurred without a large pool of poverty-stricken and unemployed workers who were readily available as 'soldiers' in the conflict. It appears that material gains such as loot from houses that were attacked, acted as an added incentive, as is borne out below.

(ii) Slackening Control

The second major consequence of the abolition of influx control was the change in the form of control over hostel dwellers exercised by apartheid bureaucrats. Although it is difficult to ascertain the exact extent of the authority of the municipalities over hostels after 1986, the interviews reveal that their control undoubtedly slackened after influx control was abolished. For a start, municipal administrators no longer had the legislative muscle to assert the illegality of people in urban areas, and there was thus no longer the need to raid hostels in search of 'illegals'.46 One interviewee remembers how in 1986 or 1987 'it started not being as strict as it used to be'. He goes on to explain:

There was less security, less municipal police. This was especially after the comrades came and burnt down the pay office the municipal police ran away from that time. It was much less strict Maybe this new police were not well disciplined You found that if a person wanted to see you, he didn't have to ask security. Even it's a wrong thing that person can get in. No information is taken and so people moved in and out much easier. (8)

For many, this slackening of control was a source of concern:

There is no more security. In the past everyone who wanted to visit inside the hostel had to ask for permission at the gates: it was safe, we were secured, but now things have changed. We no longer get any security guards at the gate, so it's not safe.

The full impact of loosening of control over the hostels was felt most acutely with the growth of the squatter settlements which sprang up alongside hostels. The dwellers of Khalanyoni hostel next to Phola Park spoke of a new feel to hostel life:

When Phola Park started to build up, more women started popping in. They would come and get water … I didn't like it at all because a lot of women would stay and there would be a lot of corruption. They'd come for water but relationships started and then they'd come to see their boyfriends … . Where women are there's a lot of shit. Only young guys like such style … . Old guys, they didn't like such thing. We didn't do anything because we didn't have the power to stop them coming in. Before women used to stay at the gate and tell security who they wanted to see … . But the feeling of the hostel changed. Kids were popping in as well. (8)

Even in those hostels where squatter settlements did not directly feature, women appear to have become frequent visitors. At Kwesine hostel in Katlehong a hostel dweller spoke about the increased number of prostitutes over the last few years and the lackadaisical attitude of the police: 'Police used to arrest them, but now it seems as if they have given up because they can't stop them'. (9) While there appears to be a great increase in the number and frequency of visits by women to the hostel, only a few spoke of women as permanent residents.47 (It should also be added that many of the interviewees were reluctant to speak on this issue).

They just come in and out at any time. There are even full-time residents amongst them. Even the police have failed at moving them out because they say they want to live alongside their husbands. (4)

Yes, there are so many women who visit the hostel, there are even others who are now staying permanently in the hostel. They stay with their boyfriends. There are others who stay just like us. You find that maybe they come because of their friends and so on. Other men who don't have girlfriends try to get used to it, even though its difficult.

He added that after the violence: 'Most women who are staying inside the hostels are Zulus. They ran to the hostel because they were killed in the township. Others have settled down and are even paying rent.' (19) The general attitude towards those women who stayed in hostels was that they were 'not decent women'. At the same time hostel dwellers appeared resigned to their presence as they've 'got no choice in stopping them because they stay with their boyfriends'. (16)

The abolition of influx control lent a new flavour to hostel life. It was, however, a double-edged sword for the hostel dwellers. On the one hand, the lifting of influx control allowed for a greater 'freedom' in hostel life, as well as a restructuring of power relations. On the other hand, the customary tools of social control which had determined the boundaries of hostel life have been gradually blunted, and new pressures have been brought to bear on the lives of hostel dwellers.

The importance of these changes cannot be underestimated when considering the rapid spread of violence from one hostel to another. Although it would be impossible to draw any direct correlation between these changes and the outbreak of hostel violence, it is possible to make a correlation between the potential ease with which a community was mobilised and the increasing over-population and impoverishment of hostel life. The generalised difficulties and daily struggles of survival made this population particularly vulnerable and insecure. The impending threats to the hostel dwellers' existence and the resort to violence must be understood in the context of this already beleaguered community.

This observation coincides with the more general theory that it is the disintegration of apartheid which is the cause of the contemporary escalation in violence. The breakdown of apartheid laws in general, has meant that there are no longer the stringent and repressive measures which had, for so long, kept control over every aspect of black life. Chaos and violence has sprung out of the very struggle to create a new social order, especially as new pressures arise around material resources.48

Developments after the abolition of influx control also intersect with explanations which seek to place socio-economic conditions at the heart of the violence: that as racial cleavages begin to dissolve, conflicts converge around economic differentials. The current violence is therefore about the poorer and more marginalised sectors of black society struggling to make a space for themselves. The conflict is distinctive not because of the ethnic lines of division amongst the participants, but because the participants (hostel dwellers and squatters), 'are those who have been amongst the most severely exploited and disadvantaged by the apartheid system'.49 The above section bears out at least some part of this claim.

'Testing Fire with your Fingers: the Hostel War'

Individual experiences and accounts of the violence are diverse. For some hostel dwellers, participation in the violence has been driven by a harsh, survivalistic quality. For others, it has been motivated by a strong political philosophy. For all, it has had devastating consequences on their lives.

What becomes apparent is that the nature of the violence on the East Rand is dictated by a unique set of circumstances which are not necessarily replicated in other areas. Each area in which violence has erupted therefore requires an investigation of the specific circumstances of that particular community.

Inkatha Membership

Many fingers have been pointed at Inkatha for being the party which has both initiated and perpetrated the violence. For example, a study conducted by the Community Agency for Social Enquiry (CASE), covering the Reef violence from 22 July 1990 to 1 May 1991, found that Inkatha was responsible for initiating violent attacks in 66 per cent of the cases.50 Although the findings of this series of interviews might suggest a similar interpretation, it is important to note that the emphasis on Inkatha's (as opposed to the ANC's) role in the violence, arises from biases in the particular sample hostel dwellers interviewed for the study. This next section therefore sets out neither to prove nor disprove the complicity of Inkatha in aiding and abetting violence. The information gleaned allows us to do no more than tentatively reflect on some of the claims that have been made about the violence in the press and political circles. We start off by examining the nature of Inkatha membership.

Thirteen of the nineteen interviewees were members of Inkatha. One was a member of the ANC. The rest did not belong to any political organisation. The interview material allows us to identify three groupings of Inkatha membership. Of the four people who belonged to this category, three had joined when they were still at school: 'I will never at all join any other organisation except the IFP [Inkatha Freedom Party]. I joined it in 1973 when I was still at home'. From the perspective of one of the interviewees 'everyone who is inside the hostel is Inkatha, throughout the Transvaal'. (1)

The very strong allegiances displayed by this group towards Inkatha as an organisation did not appear to be matched by an in-depth understanding of Inkatha's programme. Instead, they espoused fairly vague and general descriptions of the aims and ideals of their organisation. People spoke of Inkatha as standing for 'equal rights and peace' (1), of having 'the desire to dismantle apartheid and promote alliance amongst South Africans whether you're white or black' (5), of promoting 'freedom of speech in negotiations for a new South Africa'. (7) One person stressed that Inkatha 'unites all people of the country … and strives for their liberation peacefully and not through fighting'. (10) The most concrete objective expressed was that the organisation was committed 'to return land back to its rightful owners who are the black people of South Africa'.(10) One person specifically joined because of Inkatha's attempts to 'fight for land and the right of existence'.(11) The interviewees showed no specific understanding of what differentiated Inkatha from any of the other liberation movements.

The second grouping is made up of those who claim that they were forced to join Inkatha 'in order to live' a decent life in KwaZulu. Four of the interviewees fell into this category. They stressed that without being a member of 'such an organisation [they] would not have made it to Johannesburg'. (11) They explained themselves as follows:

Politically I am forced to belong to Inkatha because I speak Zulu … . Back home, if you are not a member they encounter a problem whereby they think that you oppose their position. Take for instance myself. I wasn't a member, so my child passed away. Since I was not a member I couldn't bury my child … (4)

You see, at Natal, your children cannot get any place in local schools if you are not a member of Inkatha. There are many other things you cannot do if you are not a member. (15)

The last group, consisting of five interviewees, stated that they had joined Inkatha during the past year 'because of the Transvaal war.' Three of these claimed that they had joined because 'all Zulus are clubbed together as Inkatha members despite the fact that they may be non-members': (12)

I only started joining Inkatha last year because of the violence. Otherwise I wasn't interested. I joined because they said if you were Zulu, you were Inkatha … . All Zulus who live in the hostel were classified as Inkatha and were killed. We were left with no choice but to join Inkatha. I joined at Kwesine hostel in Natalspruit. There is headquarters there where people join. (3)

Another said that: 'Hostels must not be destroyed, that is why I joined. Inkatha wants that the hostels must be rebuilt and people must stay there'. (12) The varied reasons for joining Inkatha appear to influence responses to subsequent questions. This was particularly noticeable in questions relating to the presence of Inkatha and its mode of organisation in the hostels. Those in the first category stated that they found Inkatha 'existing since I arrived here in 1982'.(1) For these people 'Inkatha has always been there before the fighting'.(10) In a similar vein it was stated that 'Inkatha meetings were there long before the violence, the same place they are now being conducted today'.(11)

Within this grouping, responses to the question of how Inkatha had been organising over the past ten years were vague. From the scant information we received, it is, however, possible to infer that there are two levels at which Inkatha organises hostel dwellers. The first is at a hostel level, where meetings are arranged by an Inkatha member who lives in the hostel. This seems to be a largely informal and ad hoc method of organising:

We don't have special days for our meetings … . There are some meetings that are just organised by us for our own good sake. The organisers of the meetings are hostel dwellers. (7)

The second, and seemingly more consistent, mode of organisation is undertaken by Inkatha leaders who live in the township, and appears to involve hostel dwellers and members living in the township itself:

One of the organisers is from the hostel and one is from the location. The one from the location is one of the oldest members of Inkatha and is a well-known member of the organisation. He is responsible for our meetings. (11)

There is no-one who has a leadership role in the hostel. Our leader stays in the township … . He only informs us about outside rallies. There is no person who sponsors us to go to these rallies. No-one is compelled. We decide as individuals to go. (3)

According to these accounts, a liaison person, 'who is also a hostel dweller', is responsible for linking Inkatha organisation inside and outside the hostel.

When meetings are held in the hostel they are attended by township residents as well as hostel dwellers:

People from the location as well as the people from the hostel do attend these meetings. Before the violence there used to be very big members coming for the meeting, even from the location. (5) They come to the meetings without fear. (9)

One person ventured that 'some people come all the way from Natal' to the meetings. (12)

The reason advanced for township residents coming to Inkatha meetings at the hostel was because 'they can't hold any meetings in the township'. (9) The hostel dwellers are, however, also invited to township meetings and either seem to go 'when I have enough time', or else explained that 'we don't all of us go. It is our leaders who represent us, they are chosen by the hostel'. (19)

Those people attending Inkatha meetings pointed to some differences in the organisation of the meetings 'before and after' the violence:

Before the violence, no person would go around telling people about the meetings. They will only put a piece of paper informing us about the meetings, anywhere where everyone can read. Ever since the outbreak of violence the procedure has changed and they use the loudspeaker to inform us. (18)

The content of the meetings likewise changed with the onset of violence. Before the violence, the type of issues that were discussed at the meetings were described as 'unity among ourselves and issues about the land problem'. In recent meetings, however, they discuss 'safety, especially which places are labelled as dangerous'. (11)

These descriptions of Inkatha's activities in the hostels contrast with the opinions of those who stated that they had joined Inkatha because of the outbreak of violence conflict. This latter grouping believed that Inkatha started organising inside the hostel only once the violence had begun. 'The habit just started immediately with the outbreak of the violence' (4), 'after the Sebokeng violence'. As one interviewee put it:

Inkatha didn't have any meetings inside the hostel before the violence. All their meetings were held at the stadium. They started organising and campaigning for more members last year. (13)

Several others, some of whom had no political affiliation, thought that Inkatha had suddenly become prominent in the hostels, and spoke about its new and expanded role once the violence had begun:

I've seen IFP members starting to organise more strongly with the outbreak of violence. (17)

Inkatha was a very small organisation even last year. It gained more members with the outbreak of violence. (13) It started becoming strong last year. (9)

Ever since the outbreak of violence which isolated all Zulus as IFP members, IFP is much stronger than ever before. (19)

Before the violence there was no recruitment by Inkatha. During the violence, that's when they told people to come and join. (6)

These opinions lend credence to the oft-repeated claim that Inkatha capitalised on, rather than created, the conflict and that, 'Inkatha's toe-hold in the hostels under warlike conditions became a stronghold. Hostel dwellers resurfaced to haunt COSATU and the ANC as the warriors and cannon-fodder of Inkatha'.51 These interviewees intimated that the reason for Inkatha's newly strengthened position was the perception that Zulu were being attacked indiscriminately regardless of their political allegiances:

Now everyone is Inkatha because Xhosas attacked Zulus indiscriminately, so that everyone who was in the hostel became an Inkatha member. Even those who were in the location became Inkatha because as long as they were Zulus, they were also killed and their houses burnt.

Township dwellers do not differentiate. When you are a Zulu you are an Inkatha member. So under these circumstances, I joined Inkatha. (12)

There was also mention of those who had actively crossed the floor to join Inkatha because of the indiscriminate attacks on Zulus:

There were Zulus in the hostel who were in the ANC but not any longer. The Xhosas were only looking for a Zulu-speaking person and not for ANC or Inkatha so that's why the Zulus left the ANC. They were just looking for your tongue, which language you were. (8)

I remember when I was still staying in the hostel, there were many Zulus who were ANC members, they didn't want to hear anything about Inkatha. But now they have changed because ANC supporters kill anyone who is Zulu regardless of whether they are pro-ANC or not. (13)

Although the implication here is that some Zulu-speaking ANC supporters actually crossed the floor, it is also possible that many fled the hostels. The end result either way was that as the violence intensified, the hostel community became more homogeneous and uniform in its political identity.

With regards to Inkatha's role in the violence, the interviews do not offer us a definitive view. They do suggest, however, that intentionally or otherwise Inkatha has been greatly strengthened by the violence. The interviewees also give us insight into why this was the case. Part of the explanation seems to be the labelling and identification of Zulu with Inkatha irrespective of their genuine affiliations. Another, equally important, explanation is Inkatha's apparent readiness to mobilise along ethnic lines. Inkatha's major drawcard has always been its propagation of ethnic/cultural , and more specifically Zulu values. Once the violence erupted, the hostel dwellers' invitation card to join Inkatha was written in a language they well understood, namely their Zuluness and the preservation of life in the hostels. It is to this question of ethnicity that we now turn.

The Ethnic Factor

Ethnicity is the single most controversial aspect of the current violence. Three broad trends can be identified within ethnic explanations of the violence. These will briefly be explored here.

The most frequently espoused explanation in the South African media, and particularly noticeable in right-wing circles, focuses on the notion that there is 'deep-rooted tribal animosity' amongst black people in South Africa. Most of these writers are simply drawing on racist stereotypes and propagate a naive theory of ethnicity based on the 'inevitability' and 'naturalness' of ethnic conflict, based on the 'atavistic tendencies of blacks'. This tribal argument is often used interchangeably with that of 'black-on-black' violence. Many use the tribal argument to prove that non-racialism is untenable.

On the opposite side of the spectrum, some left-wing scholars and politicians have sought to deny the forces of ethnicity almost completely. They relegate ethnicity to being a mere product of manipulation, first by the colonial forces, and then by the South African state, and refer to ethnicity as 'an antiquated norm'.52 For these people, the conflict 'has a great deal to do with politics and very little with ethnicity'.53 Nelson Mandela has at times, reinforced this argument: 'This violence is not as alleged by the mass media, a conflict between Xhosas and Zulus. This is not a tribal conflict at all.'54

A third, and more sophisticated position, lying somewhere between these two, is also discernible. This position attempts to understand the role and tenacity of ethnic sentiments by taking account of their material basis, that is, the roots of such sentiments in a particular social formation. At the same time, these writers stress the ways in which the South African state has manipulated and entrenched these identities through structures of control such as the Bantustans, ethnically divided mine compounds, etc. On this view, it is this crucial combination of forces, a materially-rooted ethnicity and its manipulation by the state for its own ends, which has over the years fuelled the flames of ethnic conflict.

This article accepts the premises of the latter position. The interviews reinforce the argument that 'Zuluness' is not merely crafted from above by an organisation like Inkatha and therefore cannot, as such, be discarded as an explanatory tool. It is an easy mobilising agent, and is open to manipulation, precisely because it has a continued reality in popular consciousness, and because the intersection of that consciousness and the material conditions of people's lives means that 'Zulu consciousness' is continually being constructed and composed by the people themselves.55 This also means that ethnicity is not static as it has an internal dynamic of its own. This 'construction of Zuluness' from below is then greatly fortified by Inkatha's reliance on ethnic forms of mobilisation.

While ethnic identities have therefore always been a potent force available for mobilisation, they have, over that past decades, also appeared to exist side by side with a number of other distinct identities and have been part of the 'various layers of migrant consciousness'.56 As has already been shown, an ethnic identity has not meant exclusion from a variety of other non-particularist associations in the urban world. A migrant, for example, also derives his identity from being a church member in the township or a member of a local football team. In particular contexts, a migrant's ethnic identity is thus subsumed 'in a sense of belonging to a larger group'.57 More importantly, we have shown how many Zulu-speaking hostel dwellers have adopted potentially conflicting identities, for example being members of Inkatha and Numsa at one and the same time.

What we have witnessed in the current violence is thus the crystallisation, and coming to the fore, at a point of crisis, of one particular set of the coexisting identities. In this instance, it is the ethnic fault lines which have become manifest, and the hostel-dwellers' identities as 'Zulus' and 'Xhosas' which have been thrust to the forefront of their consciousness.

When we asked about the nature of the conflict, an ethnic component was central to the explanations:

It wasn't political violence to us. It was Zulu and Xhosa violence. If it was political it would have been Inkatha and the ANC. But because many Zulus are involved in Inkatha and many Xhosa in the ANC, those newspaper reporters then said it was political violence, it then spread like political violence.

The violence in Thokoza is not like other places. If you look at Vosloorus for example, the hostel dwellers were fighting with the ANC in the townships. Just like in Soweto as well. In Thokoza we are fighting with the Xhosas. (1)

The Xhosas and Zulus started fighting as soon as they heard about Phola Park. The relationship between the Xhosas and Zulus were cordial for a long time. Immediately when people in Phola Park who are Xhosa started fighting the Zulu hostel dwellers, they were then divided along those lines. (6)

I think it was a mistake of people to think that the fighting is between the ANC and Inkatha which is not true. If it were an organisations' war how can it happen that only the Zulus, or only the Xhosas are beaten up? The organisations don't consist of only the Zulus and Xhosas. (10)

For most, the ethnic dimension of the conflict was beyond doubt. The only attempt within the interviews to explore the factors underpinning the ethnicity focussed on the role of the media in feeding into these divisions:

I would like to blame the newspapers. Whenever there is fighting they won't explain it as it really is. They would say it is a fight between Zulus and Xhosas. When they reported the Sebokeng fighting they said Zulus killed Xhosas. So I think that this affected many people. They still had that resentment of Zulus. When the fighting started everyone who was in the hostel was fighting including Shangaans, Pedis, Sothos. Even at Phola Park there were Zulus who were fighting side by side with the squatters. … I want to blame the newspapers for all this distortion of facts. (13)

Some people believed that the violence had started off as a Zulu-Xhosa clash and then became a political struggle:

The fighting started last year between Zulu and Xhosa. Things have changed and now it's between ANC and Inkatha.

The reason the fighting is becoming political is the fact that in some areas Xhosas are backed by the ANC. Winnie Mandela is supporting them by giving them guns you see. We used to hear that in our hostel that she came to arm them. She knew she was supporting her Xhosa brothers. (1)58

The obvious question raised by these comments is why it is that ethnic, rather than other, identities, have crystallised amongst hostel dwellers. This question can be answered in a number of ways.

Firstly, after 2 February 1990 and the unbanning of the ANC, the political landscape has become increasingly polarised. For the first time since the 1960s, political parties can compete openly for support. As a result, political identities have been thrown into sharp relief and, although dual identities were able to coexist in the past, the new political climate has confronted people with more dramatic and stark choices as to their allegiances. In the case of these hostel dwellers, ethnicity has become a powerful reference point in an environment in which political allies and bases of support are being radically redefined.

Secondly, to the extent that the resort to ethnicity has come from within the hostel dwellers themselves (as opposed to labelling from the outside), it is the depth and endurance of ethnic identities, relative to other forms of allegiance, that brings them to the fore in times of crisis. The other forms of identity (described above), emerge from, and are related to existence in the urban areas. But the precariousness of that urban existence for the hostel dwellers, and the emphasis that they give to their ties to their rural home, give us a sense of the potentially ephemeral nature of these other identities. Under threat, then, hostel dwellers appear to turn to what is safe and what they are familiar with. It is their ethnic roots that provide this kind of refuge.

The coalescence of ethnic identities, as a response to the perceived senses of increasing social and political dislocation and increasing turbulence in urban communities, has historical precedents. Marks, for example, speaks of 'renewed demands for a return to a Zulu way of life … in the face of rapid social change … in the towns in Natal and the Rand in the 1940s'.59

These may be some of the ways to understand why hostel dwellers' own explanations hinge almost entirely on ethnic divisions and why their accounts of the fighting are peppered with ethnic terminology. People have resorted to their common identities and backgrounds in the face of a perceived onslaught, and ethnicity has become a defensive, survivalistic identity. But it is not as simple as stating that people have 'resorted' to ethnic identities in order to survive their situations. The deliberate manipulation of these identities by Inkatha, the state, and the media has been a vital ingredient in the complex web of events.

Stereotyping and Rumour

The conditions which combine to evoke strong sense of Zulu identity simultaneously generate an atmosphere in which stereotyping and rumours abound. This stereotyping and the spread of rumours in turn overlaps with, and strengthens the resort to, ethnic identities described above.

The hostel dwellers arrive in a world in which divisions are structurally preordained through the hostel and, more broadly, the apartheid system. This world is then further divided and categorised by the hostel dwellers through an elaborate system of labelling, ethnic and otherwise, much of which has persisted for a long time. Although these labels may not be systematically anchored in any empirical 'reality', they are nonetheless presented and expressed as 'reality'. It is therefore important to understand the stereotypes that were applied to Xhosa and Zulu people throughout the interviews and how these were transformed in a warlike environment:

Xhosas seem to be clever most of the time. They always like to rob claiming to be clever. Xhosas say 'these Zulus are very stupid.' Zulus are the type who don't like people to claim to be cleverer than them. (8)

If you look at the ANC Zulus, they have assimilated the Xhosas inheritance of lying and of being crooks. That's what I will never accept. Anyone who tries to criticize the Zulus is making a big mistake … (19)

From the perspective of a Pedi speaker:

I started in Johannesburg to live with different nations and discovered that the Xhosas believe the Zulus are not well educated while the Zulus believe you can't be taught by Xhosas. (8)

With the release of Mandela and the unbanning of the ANC, a whole new set of political perceptions and stereotypes appear to have been added to those that already existed. For the Zulu speakers, these stereotypes seem to hinge on a series of anti-Zulu sentiments:

All Zulus were reduced [by the Xhosas] to Chief Buthelezi's puppets, sellouts. This was because they wanted to uplift Mandela's name.

It will never happen in the history of the ANC that a Zulu person can be in power. You can do whatever for the ANC but forget about leadership role in the ANC if you are Zulu. Right now, Zuma who we were all looking into has failed. (19)

The same Pedi speaker had the following to say:

What makes it worse is since Mandela has been released. The Zulu people see it as a challenge, especially since he's been speaking to De Klerk and he might be the next ruler. Zulus find this undesirable because the Zulus say they are the ones who initiated the struggle against white people therefore they are entitled to rule. We all celebrated Mandela's release but soon after that were people who are against Mandela ruling. (8)

In another hostel dweller's opinion:

It started with silly remarks made by people against Zulus, especially township children. You see since these children favour the ANC, when these children see us they pass remarks and go to the extent of singing songs saying 'Yeh, Gatsha Buthelezi, yeh!' They thought the Zulus were against the release of Mandela. These things make us feel bad. So these are actually the things that have contributed towards the violence. It all started with these little things that built up until now we are fighting. (2)

The power of these political stereotypes is that they articulate with, and feed into, ethnic and political divisions. They are also the source of rumours and hearsay circulating in the township. It is these rumours which provide the spark by giving people the faith to undertake activities which common sense, or rational judgements, would preclude them from undertaking. Two interviewees proposed that rumours, rather than any concrete events, were directly responsible for causing the violence:

The fighting started because of rumours. People came to tell the Xhosas that Zulus were coming to attack. I think it was the tsotsis … . In fact the whole thing was organised by a bunch of criminals who went on spreading the rumour that the Zulus were going to attack the Xhosas. There's no one who knows the [cause of the] conflict. We don't seem to get any answers of what caused this violence. Its just the rumours. (3)

The spreading of rumours cannot simply be attributed to a 'bunch of criminals'. There is strong evidence to suggest that rumours were part of a deliberate campaign to sow divisions in the community. Evidence is provided by forged pamphlets which bore an out-of-date ANC crest, which the ANC believed to be the work of security forces, and which stated:

We want to destroy the Zulus. We want to drive them out of the hostels as we did in Sebokeng … . We have to end their Zuluness of which they are proud about it [sic] … . There will be no peace as long as the Zulus are still powerful. Let us destroy them all in South African townships, hostels, and in the working places. Let us burn down their houses in the townships and drive them out. Down with Zulus.60

That it was these types of rumours that fed into the tensions and started the conflict, is shown by this account:

All of a sudden, when this violence was about to start [on the Sunday night], they called a Zulu meeting only for Zulus and a Xhosa meeting only for Xhosas. We had never had different meetings like this before in the hostel although it had already started in other townships like Sebokeng. When we were at this meeting, there was hearsay that the Xhosas wanted to kill us now. When we asked why this was, they said it was because the Zulus had killed our people in Sebokeng. Zulu people started killing and now Xhosas wanted to retaliate, that's why we couldn't have joint meetings that night. The only issue was just hearsay … . There was an old man, he raised the question 'Why do the Xhosa want to kill us because we weren't even in Sebokeng?' There was no really solid answer for that until he said, 'No, this is just hearsay, those people won't do that'. The final word was that these people won't do this thing … it came to my mind that it was just hearsay.

The problem was with the other meeting with the different nation [Xhosas] because we didn't know their final word. They were meeting at the same time. I don't know what they said in that meeting. It came to my mind that their final word was a start this thing. It was my last meeting in Khalanyoni. The violence started the next day [Monday] at about 11 am. (8)61

With particular reference to the hostels, the rumour circulating was that the Xhosas and local residents were planning to abolish them:

We as hostel residents viewed that as being a stupid decision taken by the local residents. We actually went to the extent of deciding that in case our hostels were demolished, we will go into their houses and forcefully occupy them. (2)

The power of these rumours was again illustrated when we asked the hostel dwellers what started the violence in their own area. Many responded initially to the question by saying that 'there is no one who knows the conflict' or 'no one can tell you the foundation of this thing' or 'I really can't say'. They then proceeded to tell us what they had 'heard' or 'were told' was what they believed to be the cause:

I don't know. We were just told by people from Thokoza that Xhosas were killing them. That's how we got involved in the fighting . During that time we were still staying with Xhosas in our hostel. But the Xhosas we stayed with were beaten up and they ran away. This spread to other hostels in other places. We didn't exactly know what caused the fighting. (9)

I don't know exactly what started the fighting. I don't want to commit myself a lot from that because we just heard from hearsays what happened. (10)

I don't know exactly what the conflict was, we just heard that it started with a fight between a Zulu and a Xhosa was defeated. That's something we heard from rumours though.

Interestingly, there was a high degree of correlation between the different versions of 'hearsay' explanations as to why they believed they were now fighting. The story, repeated in a myriad ways, was allegedly about 'a woman dispute'. It took place between a Xhosa and Zulu man in Khalanyoni Hostel:

It started in Thokoza at Khalanyoni hostel. The people who stay there say that it all started in a shebeen. There was a fight between a Zulu and a Xhosa over a woman. She was staying in the hostel with a Xhosa guy but then ended up in a relationship with a Zulu guy. The other Xhosa guys influenced the Xhosa guy who was in love with the woman to demand her back. They asked him how they could allow his girlfriend to be taken by 'ichaka' [a Zulu]. The Xhosa guy then fought the Zulu guy but was unlucky as he died. His Xhosa brothers then went to Phola Park to tell them of the incident. They didn't ask, they just went to the hostel and attacked anyone who is Zulu indiscriminately. This all happened in a shebeen then to the whole hostel. That's how we were told and that's how the fighting started.

On Tuesday all Zulus were expelled from Phola Park62 and all Xhosas inside the hostel ran away to Phola Park. The fighting then involved our hostel [Mdala], because every person who survived the Khalanyoni attack came to our hostel for rescue. They came together as a group which divided Xhosas and us Zulus and we knew that Xhosas were in Phola Park and Zulus are in the hostel. (10)

In a shorter, more succinct description:

You see, you find two males unable to settle their dispute about a female involved. So the other would rush into his nation and find people who are willing to say 'Ag, kill him'. That is how it stared until it grew bigger to an extent whereby so many people have died. (4)

The obvious response here is to ask how a trivial love tale could generate the chaos and loss of life that followed. The point about this particular tale is that the different ingredients played perfectly into the underlying tensions, stereotypes and rumours which were already firmly in place; the man who started the carnage was seen to have 'invited other Xhosas from the squatter camp to wipe us Zulus out from the face of the earth'. Xhosas then came and attacked the Zulus, 'accusing them of being Inkatha, sellouts and Gatsha dogs'. As a result they were left 'no choice, we had to try and defend ourselves'. (19)

One Inkatha member aptly described how the 'love dispute' was simply the final precipitating factor:

There was tension and cold war long before the fighting started. The fight between the Xhosa guy and the Zulu exploded the ever-existing conflict in my opinion. (7)

And in another's opinion:

The fight over the women wasn't the cause of the violence. It wasn't the first time that had happened. It's what the people used to say because it happened when the violence broke out. If it was the cause of the violence it should have happened long ago because even in the hostel you find that if a Zulu did something wrong to a Xhosa … there was no violence. (8)

The fact that Khalanyone hostel was broken down brick by brick and most of the hostel dwellers property was stolen was seen as vindication of the beliefs of some that the violence was started by criminals or people who wanted to deprive Zulus of their property:

These criminals wanted to steal something. They looted our properties, all our possessions. At the new hostel we had to start afresh. People living in the squatter camps want the bricks of our hostel to build their shacks. (3)

Police Partiality

Partial policing is cited as one of the major contributing factors to the continuing violence. Several factors have borne out this claim: reports from lawyers and trade unions have alleged that in some instances the police were warned about the conflict in advance but took no pre-emptive measures.63 Photographic and video evidence of police involvement has likewise shown the police aiding a particular party in the conflict. More often than not, the party accused of receiving police aid is the Inkatha Freedom Party.

When we directly posed the question to the hostel dwellers whether Inkatha was supported by the SAP, everyone (with one exception) emphatically denied the validity of this claim:

That's an insult because the SAP has killed a lot of our people. (17)

There is no such thing. Hostel dwellers do not know anything like that because all the guns they had were confiscated by the police. In fact if they were supporting Inkatha they [Inkatha people] wouldn't be killed. So those are green lies. (13)

One hostel dweller pinned the blame for this accusation on the askaris:

We mustn't forget that there are ANC dissidents who have joined the police force and are called askaris. These people are never exposed for what they do, because definitely they want revenge against the ANC. So instead of trying to think about that, they just criticise the Zulus and Inkatha.

Another suggested that these allegations were false because the 'Xhosas don't want to accept that they have lost the battle'. (10)

These denials are made more complex, however, by the observations of several hostel dwellers, that ethnic allegiances are a divisive influence within the police force, and underpin police partiality in the violence. On this interpretation, Zulu police support Zulu factions while Xhosa police support their Xhosa counterparts. This ethnically-based support was, according to some interviewees, linked to shifts at the police station:

[Police complicity] actually depended on shifts of the station commanders. Take for instance if the station commander whose shift is at 8 am is a Zulu, he will use this powers to his advantage whereby he will command the troops to fight alongside Inkatha. The same situation does happen with the ANC. If the commander is a Xhosa he will use his powers to fight alongside the ANC. If it is the Xhosa's shift he will give the Zulus leave so that he can use the army against Zulus. That is why they government landed in so much trouble whereby people thought police were involved either way.

He went on to explain the rational behind the police bias:

I mean you cannot let your brother die … you will help by either way you can. They will actually help us with bullets but also the ANC was being helped. That is also the reason why most policemen quitted their jobs because they could not take it killing their people. (4)

This quotation not only introduces questions around ethnicity and policing but also around the power and autonomy of local police commanders in determining the political affiliations of the police deployed on the ground. Another implication seems to be that the army is being used at the discretion of the local commanders rather than because there is a genuine lack of police personnel.

The best illustration of police complicity along ethnic lines was a description of the fighting that broke out on the 13 December 1990 at Mdala hostel in Thokoza. The story, pieced together from several accounts, goes as follows:

We were told by one man that the police are coming to search us at our hostel. We tried to hide all our weapons. Instead of the police coming to search us, it's the police bringing the Xhosas to attack us. We were shocked when we saw Xhosas being delivered to our hostel by police to kill us, in fact police were also shooting us.

It was 2.30 am and we were all sleeping. We heard the noise of guns. We look through the window and saw a Caspir stopping at the gate of the hostel. We see that they [the Xhosa police] are protecting the many people outside who are wearing blankets. It's the Xhosas from Phola Park who are attacking. We got no time to go out. They shoot all over. One room they get inside, burn the people, shoot them. Seven people died there. Those that ran outside, they killed them.

The police is driving the Hippo. If we want to run away we don't have a chance. At something to five we try to get out but there are too many guns, there's bullets all over. We shout at the police, 'Why do you bring these people inside the hostel? Why don't you bring tear-gas inside? Why do stop in the middle of the gate? Why can't you get inside?'

When we saw our police, that is, Zulu police, we got courage with our spears. We went straight to where the Xhosas were and we killed 25 of them. We were stabbing them with spears and our police brought some of the guns. We beat them again. When we attacked them from behind, they ran towards Khutuza hostel. When we got to Khutuza hostel we found white policemen parked there not worrying about anything. They were just at the gates.

[Is that the reason why Zulu policemen came to your rescue, returning your weapons?] They didn't just return our weapons. They were helping us as Xhosa police were also helping their brothers. They took us out of our rooms to go and fight. Which was right. They couldn't let us just be butchered by Xhosa policemen and their brothers.

The episode continues the following day:

In the morning we go to report to the police station. We take the registration number of the caspir. I told the police, 'You bring Xhosas here to hit us'. But the police were Xhosas and they ran away from the police station because they were afraid. (1) The night before the Zulu police were instructed to be off-duty and only Xhosa police remained in charge and collected comrades to come and attack us. (12)

In the subsequent revenge attack by hostel dwellers from Mdala hostel on Phola Park residents, the police were describe as 'disappearing to the location'. This interviewee explained that 'I think that they were giving us a chance of revenging'. (19). The consequence of the attack was 'bodies strewn all over the place, making it difficult to walk'. (19)

These are indeed chilling descriptions of police complicity. But it is the casual reference to 'our police' and 'helping their brothers' which is perhaps the most potent illustration of the apparent lack of police neutrality and misconduct. Here however, it is revealed to be a more complicated equation than simply that of the police aiding Inkatha. Whereas previous debates have centred around racist policing, that is, differential policing of blacks and white, this material points to an ethnic dimension in partisan policing.

There seems little doubt, therefore, that the police have on occasions been implicated in the violence, both through active partisan involvement and through 'turning a blind eye' to the violence. The feeling that the South African Police were not to be trusted as agents of peace was clear from the interviews. In expressing these opinions, the hostel dwellers illuminated another interesting issue concerning policing. They baldly stated that it is police from outside South Africa who are the people who have contributed to stopping the violence. This resonates rather alarmingly with allegations made during this period, of Namibian and Mozambican involvement in the township violence as well as in the train massacres on the Rand:

The SAP has definitely played a role in the conflict but since they have been substituted by black troops from outside South Africa everything is fine. There have been no further attacks by the Xhosas. They used to attack us when the SAP was around. (17)

The people who curb this violence was the foreign troops from countries like Mozambique. They stand next to the hostels and defend them. (15)

The fighting has halted because of the presence of foreign soldiers who can't speak all the South African languages … . If they weren't there, fighting would still be on. (19)

We go to the municipalities and tell them to bring police to the hostel. They bring police in the hostel until today. They are black ones but not from South Africa. (1)

Who exactly are these foreign troops and why are they being used in this instance? Is it a sign of recognition that the SAP are incapable of stopping the violence precisely because of the ethnic affiliations described above? Why hasn't the public been informed of these troops? Are these troops being used to protect hostel dwellers, as the above quotations all imply? In the light of recent revelations about involvement of SADF units like Rekke 5 in township violence, these questions require urgent investigation.

Destruction of Hostels

Conditions in the hostels were uniformly expressed as disgusting:

If you can see our toilets, I doubt if you ever have appetite to eat ever again in your lifetime. There is no-one who bothers to clean them. We've got a few showers, we have to wait for each other even if you are in a hurry you just have to persevere and wait. We are lucky because we've got warm water. (9)

Any person who comes to the hostel for the first time can be shocked if he is told that there are people who live in it. (10)

Despite detailing these deplorable conditions, interviewees were adamant that hostels should not be destroyed. The appalling conditions appeared to be tolerable in the face of homelessness and the attendant implications that the destruction of the hostels signal. Hostels are the migrants 'entire urban reproductive base' and the retention of this 'urban umbilical cord' appears to be enough to overcome the degradation of the living conditions.

Sixteen out of the nineteen interviewed were thus vehemently against the destruction of the hostels. 'If they want to demolish hostels I will fight anyone'. (12) Of the three people who did not want to continue living in hostels, two were non-Zulu speakers. This again introduces the question of differential attachment to rural areas amongst different ethnic groupings.

The reasons given as to why hostels should not be demolished, echo the sentiments that pervaded hostel-dweller perceptions of themselves in the urban areas. As such, they hinge on the fact that they see themselves as temporary residents who were there only to work, and who had absolutely no wish to invest any money in a place which they do not consider home.

We are not here to settle, we are here to work, so homes are a waste of time. Whatever I get has to go home to develop my home at home, not anywhere else …

They should just leave us as we are. We are not interested in buying municipal houses … (1)

I don't want a house in the township especially because it takes only three minutes for a person to lose his house in the township if he is Zulu. The comrades can burn it down any time. (17)

I am here to work. Hostels are the cheapest place you can ever find. If you get a room in the township, the householder will charge you about R120.00 per month for a small room of 3m by 3m. Whereas in the hostel you pay R15 per month and we know that we are here temporarily because our homes are in Natal. (13)

And on a more practical note:

I think idom [useless] the idea that people should go to the locations and seek houses there. The residents in the location don't have houses themselves, so where do you think we will get houses from? We are having more than a thousand in number at the hostels so where do you think all the living space will come from?

The idea of bring women to the urban centres was equally unappealing:

Our problem is money. You can't bring your wife if you don't have money. Do you know how much I get paid a month? I am paid peanuts. It's much better that I'm alone because if I buy meat it lasts me for about three to four days … . I'm safe from overspending for a woman. Women are too demanding and too costly. The idea of staying with a woman is impossible … . We are going to kill each other if we stay with women … (1)

The strength of these assertions is an urgent warning of the need to reverse the perceptions held by hostel dwellers that both COSATU and the ANC plan to demolish the hostels. Although this is not the stated policy of either of these organisations,64 it has become part of the popular mythology circulating in the townships. As has been shown, this perceived threat is precisely the impetus that spurs the hostel dwellers to take up arms. As one hostel dweller warned, 'If there is still undermining other people for where they live, peace will never come. People must stop calling for the destruction of the hostels'.(13)

The question of the transformation of hostels, however, has become considerably more complex since the outbreak of the violence. Urban hostels are no longer home to the migrant population in general. The process of expulsion of Xhosa speakers during the fighting, and the subsequent fleeing of other ethnic groupings has essentially transformed the hostels into urban repositories for Zulu-speaking migrants:

… there is some space now ever since our 'brothers' left the hostel [his emphasis] … . We do not have other groups staying with us. The only people who do not stay with us are Xhosas. They all ran away. The majority are Zulus though, we don't have any problems with them, we live happily like a family now. (19)

The other tribal groups have all run away and right now we've got only Zulus in our hostel. (9)

And someone from Kwesine:

The hostels were mixed but now lately it is predominantly Zulus as other groups have all ran away. Most of the people didn't like the fighting among tribes, even some of the Zulus didn't like it. (9)

In an urban environment which has been characterised by a desperate shortage of space, the violence has 'reorganised the urban landscape' in favour of Zulu migrants. This complicates the trajectory of developments amongst the more general migrant population in urban areas. This is particularly true in a context where hostel dwellers feel that 'our lives have changed. There is tension among ourselves and as a result, we trust no-one'. (17)


Hostel dwellers' belief that they 'can never trust another human being' (9) goes along with their belief that the violence that has shattered their lives is not yet over. In fact, not a single one of the interviewees thought that the fighting had come to an end. Nearly three months after the interviews their prophecies proved to be well-founded. On the 8 September 1991, twenty-three hostel dwellers from the Thokoza hostels were shot dead. Only this time, it was by unknown assailants firing at a group of hostel dwellers on their way to a meeting to discuss the destruction of the hostels.

The press no longer speaks of 'dreaded Zulus' but now refers to the hostel dwellers as hapless victims of 'agents' who 'triggered the slaughter'.65 This article should have demonstrated the need to go beyond characterising hostel dwellers as either victims or vanquishers. The voices of the hostel dwellers raise questions and answers which afford us a more complex view of their actions.

Rather than answering all these questions, however, it presents a research agenda for those seeking both greater explanation and greater complexity. The more 'versions of the violence' which are made available for interpretation, the closer we will be to understanding the tragic destructiveness that has engulfed South African society.


1 S. Mokwena, 'The Dube Riots of 1957', unpublished Honours thesis, University of the Witwatersrand, 1990.

2 Windhoek Advertiser, 16 August 1990.

3 Business Day, 16 August 1990.

4 Windhoek Advertiser, 20 August 1990.

5 The Cape Argus, 12 August 1990.

6 The Weekly Mail, July 19-July 25 1991: The New Nation, July 19-July 25 1991.

7 Please note that figures in parenthesis refer to the interviews while superscripted figures refer to the footnotes.

8 M. Morris and D. Hindson, 'Political Violence and Urban Reconstruction in South Africa', unpublished paper, July 1991.

9 This term was used by Ari Sitas, 'African Workers' Responses on the East Rand to Changes in the Metal Industry', unpublished Ph.D thesis, University of the Witwatersrand, 1983. He uses the term to describe his project which he says 'both defines and is defined by the dramatis personae of the conflicts in industry'.

10 For example, F. Wilson, Migrant Labour in South Africa (Johannesburg, 1972), and C. Murray, Families Divided (Johannesburg 1981).

11 See C. Van Onselen, Chibaro (Johannesburg, 1980); P. Pearson, 'Authority and Control in a South African Goldmine Compound', African Studies Institute seminar paper, Johannesburg 1997.

12 D. Moodie, 'The Moral Economy of the Black Miners Strike of 1946', Journal of South African Studies, 12, 1986; 'Mine Culture and Miners' Identity on the South African Gold Mines', in B. Bozzoli, (ed.) Town and Countryside in the Transvaal (Braamfontein, 1985); D. Moodie, 'Migrancy and male Sexuality on the South African Gold Mines', Journal of South African Studies, 14 (January 1998).

13 This, despite the fact that in 1974, the West Rand Administration Board estimated that over 20 per cent of all male residents over 18 years old in Soweto, lived in urban hostels. The urban hostel population in the West Rand Administration area including Alexandria increased from 44 000 in 71 to 66 000 in 1974. This excluded company hostels which housed a further 20 per cent of Johannesburg's male working force in 1971. A. Peskin and M. Spiegel, 'Urban Hostel in the Johannesburg Area: A Preliminary Study', in P. Mayer (ed.), Migrant Labour: Some Perspectives From Anthropology vol.7 (Rhodes University Migrant Labour Project, Institute of Social and Economic Research, Rhodes University, 1976), p. 2, In 1982, the number of urban migrant workers was estimated to be 2 370 458, see A. Sitas, 'From Grassroots Control to Democracy. A Case Study of the Impact of Trade Unionism on Migrant Workers' Cultural Formations on the East Rand', Social Dynamics, 11, 1 (1985), pp. 32-43.

14 J. Segar, 'Living Conditions in the Hostels in Cape Town', unpublished manuscript, Rhodes University, 1988; M. Seloane, 'Nyanga East Men's Hostel: the Condition of Migrant Workers', SALDRU working paper, University of Cape Town, 1988; D. Selvan, 'Housing Conditions for Migrant Workers in Cape Town 1976', SALDRU Working Paper No. 10, University of Cape Town; M. Ramphele, 'A Quest for wholeness – Health Care Strategies in the hostels of the Western Cape', Department of Anthropology, University of Cape Town.

15 See Peskin and Spiegel, 'Urban Hostels'; G. Pirie and M. Da Silva, 'Hostels for African Migrants in Greater Johannesburg', GeoJournal, 12, 2 (1986), pp. 173-182.

16 P. Delius, 'Sebatakgoma: Migrant Organisation, the ANC and the Sekhukuneland Revolt', Journal of Southern African Studies, 15, 4 (1989), p. 14.

17 Sitas, 'African Workers Responses', p. 33.

18 Peskin and Speigel, 'Urban Hostels', p. 19.

19 Mayer in Migrant Labour, for example, pointed to the 'ethnic factor' in the hostel situation and how 'homeboy links … lend themselves to ad hoc rallying of ethnic support.' He goes on to say that 'since the (Zulus) are the largest group in some hostels … they tend to be perceived as the "most vociferous and ethnically aggressive" element'. The study however was conducted only a couple of months before the Mzimhlope riots in August 1976, during which Mzimhlope hostel dwellers attacked and killed over 20 Soweto residents. The potential role of hostels in political struggles was missed here.

20 Delius, 'Sebatakgomo'., p. 1.

21 P. Delius, 'Migrants, Comrades and Rural Revolt: Sekhukuneland 1950-1987', Transformation, 13, 1990.

22 T. Mthetwa, 'Community, Urban Migrants and Workers', Third-year Sociology essay, Sociology Department, University of the Witwatersrand, 1990.

23 Sitas, 'African Workers Responses', p. 33.

24 Ibid, p. 34.

25 Mthetwa, 'Community, Urban Workers and Migrants'.

26 The frequency of their visits is probably the result of the 'taxi revolution' in South Africa. A supposition that requires further investigation is that Zulu speakers are more tied to their rural base than their Pedi- and Sotho- or Xhosa-speaking counterparts. These other ethnic groupings appear to be more willing to consider living in, and bringing their families to, the urban areas. There have been several reports that there has been a general movement of Xhosa speakers, less attached to land in the Transkei than KwaZulu, out of hostels into shantytowns in the urban areas.

27 See Simpson, Mokwena and Segal, 'Political Violence in 1990: The Year in Review', in M. Robertson, (ed.), South African Human Rights and labour Law Yearbook 1991 (Cape Town, 1991).

28 Sitas, 'From Grassroots Control to Democracy', pp. 35, 36.

29 H. Vilakazi, 'Waging War On Our People', Tribute Magazine, September 1990.

30 Sitas's thesis deals specifically with metalworkers who worked in the foundries on the East Rand, and who lived in Vosloorus hostel.

31 Sitas, 'African Workers Responses'.

32 Ibid, p. 264.

33 Ibid, p. 271.

34 Ibid, p. 274.

35 Ibid, p. 36.

36 G. Ruiters, and R. Taylor, 'Organise or Die', Work in Progress, 70/71, 1991, p.22.

37 Morris and Hindson, 'Political Violence'.

38 Delius, 'Migrants, Comrades', p.11.

39 Mthetwa, 'Community, Urban Migrants and Workers', p.62.

40 W. Beinart, 'Worker Consciousness, Ethnic Particularism and Nationalism', in S. Marks and S. Trapido (eds), The Politics of Race, Class and nationalism in Twentieth Century South Africa (London, 1987), p. 298.

41 There have been dramatic developments in Alexandra township since 1988 and the situation may well have altered since then.

42 Mthetwa, 'Community, Urban Migrants and Workers'.

43 M. Ramphele, 'The Dynamics of Gender Politics in the Hostels of Cape Town: Another Legacy of the South African Migrant Labour System', Journal of Southern African Studies, 15, 3 (April 1989).

44 In Soweto, whole rooms in the hostels, have reportedly been converted into trading outlets, mechanic workshops etc.

45 Quoted in Sitas, 'African Workers' Responses', p. 248.

46 Certain regulations still enabled police to enter hostels in search of 'unauthorised persons' who could then be charged with trespassing. It has also been suggested that municipalities started insisting that money was given to local authorities for the extra people staying in the rooms.

47 The increase of women in the hostels in the Transvaal has not taken on the same proportions as in the Cape Town hostels. A penetrating study conducted in 1988 revealed that 'Cape Town hostels are neither migrant labourers nor single sex hostels but are in fact home to a large number of women and children endeavouring to live together as families', 36 162 people occupied 12 828 beds in the government-owned hostels of Langa, Nyanga and Guguletu, an average of 2.8 people per bed. Most of the extra people were women and children which essentially means that hostels have become overcrowded family slums. See J. Segar, 'Living in Anonymity: Conditions of Life in Hostels in Cape Town', paper presented at the Annual Conference of Anthropologists of Southern Africa. September 1988, p.10.

48 For a more in depth study of this, see Morris and Hindson, 'Political Violence'.

49 Ruiters and Taylor, 'Organise or Die', p. 20.

50 Press Release from the Community Agency for Social Enquiry (CASE), 20 May 1991.

51 Morris and Hindson, 'Political Violence', p.11.

52 K. Mhhize, 'Ways of Seeing', Indicator SA, 8, 1 (Summer 1990).

53 K. Shubane, 'The Violence in Some PWV Townships', unpublished manuscript from the Centre for Policy Studies, University of the Witwatersrand, 1991. This is not the opinion of the author in this paper.

54 Quoted in Ruiters and Taylor, 'Organise or Die', p.20.

55 For a highly complex and useful paper on the ways in which people's experiences give a different form to their 'Zuluness', see A. Sitas, 'Class, Nation, Ethnicity in Natal's Black Working Class', presented to the workshop on Regionalism and Restructuring in Natal, January 1988, paper no. 16.

56 This is an expression used by Beinart, W., 'Worker Consciousness, Ethnic Particularism and Nationalism', p. 306.

57 Ibid, p. 292.

58 When we asked how the nature of this conflict differed from Natal, interviewees were quick to point out that: 'At home, it is Inkatha and the UDF. But here, even though it has assumed that shape, it started as a feud between Xhosas and Zulus. In Natal it is the ANC and Inkatha'. (17)

59 S. Marks, 'Patriotism, Patriarchy and Purity. Natal and the Politics of Cultural Nationalism', in Leroy Vail and Lan White (eds), The Creation of Tribalism in Southern Africa (London and California, 1989).

60 Xerox copy of pamphlet obtained in Soweto, 17 August 1990. Quoted in an Africa Watch Report, 1991, p.59.

61 The squatters believed that after the meeting described here, the Zulu hostel residents attacked the Xhosa hostel residents, 'the motive being to avenge the attack of Zulu hostel dwellers in Sebokeng', p.6. For more accounts of these events from the perspective of the Phola Park squatting community, see affidavits published in a Lawyers for Human Rights report, 'Phola Park 10-13th September 1990 as witnessed by the residents'.

62 For a history of Phola Park, and its fight against the Thokoza Town Council's threats of removal, see 'Phola Park: A Struggle for Survival', a case study prepared by the Community Interface Project, Wits Business School, March 1991.

63 Reports of the Independent Board of Inquiry into Informal Repression, see in particular August and September 1990.

64 For a detailed discussion on COSATU's proposals on the question of hostels, see G. Schreiner, 'Transforming the Hostels', Indicator SA 8, 3 (Winter 1991).

65 The Star, 10 August 91.

© Centre for the Study of Violence and Reconciliation

+ posts

CSVR is a multi-disciplinary institute that seeks to understand and prevent violence, heal its effects and build sustainable peace at the community, national and regional levels.

Leave a Reply

Your email address will not be published. Required fields are marked *

Translate »