Mediated Paternalism and Violent Incorporation: Enforcing Farm Hierarchies on the Zimbabwean–South African Border

Paternalism and violence on South African farms have been famously intertwined. In a kinship idiom, fatherly white farmers confer ‘gifts’ on black workers, their ‘people’. This discretion maintains the conditions for racialised violence. But, on the Zimbabwean–South African border, mass-migration and globalised agriculture give paternalism and violence new significance. If white extra-legal violence was previously key to maintaining racialised, hierarchical order in rural areas, today a similar order is maintained through collusion between white farmers, senior black workers, and border guards. Such distributed coercion has precedents, with white and black patriarchs enforcing their positions and agendas through physical force. Today, however, white farmers are keen to perform a corporate style and appear removed from any vigilante violence. Senior black workers actively entrench themselves as powerful arbiters of farm order amid transience and widespread unemployment, drawing on the coercive power of border guards. But as the latter mediate paternalism down the chain of command, this remains at the pleasure of their employers. It all depends on having a job. Moreover, even as everyday influence is devolved down the hierarchy, senior workers are kept in their place by an absolute distinction between black and white.


Introduction
Just knock them around a bit more, and then hand them over to the soldiers.
Leaning out of his bakkie, 1 one of Grootplaas estate's white farmer-landowners quietly gave this command to Marula, 2 the top black foreman standing next to him. The two culprits waited nearby. They had been seasonal workers on the farm, employed to pick citrus. But, drunk one night, they had got into a fight with another worker, struck him repeatedly with a length of electrical cable, and fled. At month-end they had returned for their wages, and in the meantime had been sleeping in the bush and stealing workers' possessions from the labour compound. At dawn, as workers washed, ate and prepared for a day in the citrus orchards, the two men had been discovered. A security guard and 30 or so men had dragged them into the compound, past the rooms they would have previously occupied, to Marula's house. The foreman had already left for the farm workshop, so the crowd took the two men to the football field and beat them. And so now Marula presented the men, already bruised and hunched, to their former employer. Marula, in fact, would not follow the farmer's command to beat them further. They were already being described by workers as magumaguma -border gangsters -legitimising violence reminiscent of a lynch mob. And so, fearing for their safety, Marula simply handed them over to the soldiers for deportation -part punishment, but also part protection.
This article is about the role of violence in labour arrangements and hierarchies on South Africa's northern border farms. The precise setting is the southern bank of the Limpopo river, the boundary between Zimbabwe and South Africa, and the site of a string of large crop estates employing large numbers of settled and migrant Zimbabweans. The Limpopo river is a place of large-scale migration, transience, mass unemployment and short-term strategies of making do. In the often dry riverbed, gangs traffic, rob and rape. Police and soldiers act with virtual impunity. Resident workforces on farms along the southern bank become environments in which people strive for a provisional permanence. Following a pattern that is familiar in southern Africa, farm workers live on the estates in labour compounds: rows of brick or corrugated-metal cells if farmers provide them; collections of ramshackle mud huts if they do not. Workplace hierarchies spill over into compound life more generally, and shape vertical paternalist ties and what Blair Rutherford has called 'domestic government' on white farmers' land. 3 From one point of view, this is a case of the kind of incorporation that James Ferguson has recently depicted as an overdetermining pattern in recent southern African history -one in which people have attempted to find security in subjection. 4 By subordinating themselves to different authorities and thereby seeking protection -making 'declarations of dependence'people have opened up new options in life. This is a provocative starting point for thinking about contemporary centres of wage labour. In settings of transience and insecurity, it means examining how people gain a foothold, and with what consequences.
Uncertainty and the possibility of harm are the backdrop to people's search for incorporation. But any such protection is itself shot through with violence -a dimension for which Ferguson leaves room, but which he does not explore. What is the relationship between incorporation into -and the maintenance of -everyday arrangements, on the one hand, and violence, on the other?

Violence in the South African Countryside: Past versus Present
Social historical scholarship on rural South Africa in the first half of the 20th century has explored precisely the violence underlying a racialised status quo. The parameters of everyday arrangements in the countryside were profoundly unequal, and were backed by coercion -even a gratuitous 'surplus of violence', in André Brink's words. 5 Legal and extra-legal violence, 'market compulsion and brute force', combined to uphold white authority and black deference. 6 Higginson and Strobel extend George Fredrickson's account of 'unofficial violence' in the post-Civil War United States: 'bolstered by white authority and the disenfranchisement of blacks in the American South, [this] established "a rigid caste division between racial groups that were inextricably bound to the same culture, society, economy and legal system."' 7 Meanwhile, for migrants heading into South Africa in search of work, incorporation was literally coercive: recruitment for farm work meant 'not merely being assaulted, or robbed, or subjected to touts impersonating the police. Some were also being recruited at gunpoint, hunted like ostriches, or captured by dogs before being stuffed into lorries'. 8 On farms themselves, incorporation and violence were intimately related. In contrast to a culture of public lynching in the Unites States, Evans contends, rural violence in South Africa was often hidden away on white land. 9 Moreover, writing of sharecropping on Transvaal maize farms in the first half of the 20th century, Charles van Onselen argues that apparently opposite tendencies -paternalism as against violence; hegemonic and protective as against brutally coercive -were in fact two sides of the same coin. 10 An idiom of protective family relations depended on white farmers' authority not being contested, and senior black figures (sharecroppers and, increasingly, paid labour tenants) 11 continuing to supply labour from among their dependents. Both the paternalism and the violence here are best understood in terms of the relationship between white superordinate patriarch and black subordinate patriarch.
In the opening vignette, white and black patriarchs lie at the centre of both paternalism and violence. To this extent, they appear to resemble van Onselen's depiction. But what is striking about dramatic instances such as the one described above is their rarity. Indeed, contemporary South African agriculture seems worlds apart from earlier arrangements of overt violence. 12 On the Zimbabwean-South African border, farms are characterised by the apparent distance of white farmers from the paternalism or enforcement of workforce hierarchies. 13 Farmers rarely need to resort to overt coercion, or even present themselves as having the capacity to coerce. On the contrary, their denial of brute force is absolutely central to their self-presentation as export-oriented businessmen. Subject to regular state and supermarket inspections, farmers adopt a corporate style and retreat from their workers' everyday lives. Yet their authority remains conspicuously unchallenged. Unlike other parts of South Africa, there has been virtually no labour unrest on the Zimbabwean border, and farmers along the Limpopo are quick to underline how free the area has been from farm murders.
The transience that forms the backdrop to agriculture on South Africa's border with Zimbabwe may be especially acute. Nevertheless, broader trends in farming in that area appear to mirror those elsewhere in the country. Amid casualisation, neoliberalisation, and the intensified use of cross-border migrants as cheap and docile labour, 14 today's coercion of workers appears the result of exposure to vulnerability. Studies of farms on the Zimbabwean border and elsewhere in South Africa reveal insecurity and violence within workforces. 15 The context is a regional shift from a labour-scarce political economy, in which farmers and other white employers struggled to secure sufficient labour, to a labour surplus in which most people are excluded from secure employment altogether. This is precisely why Ferguson sees current 'declarations of dependence' -attempts to find some security in self-subordination -as so poignant. 16 This article examines how, amid instability and farmers' increasing detachment, established farm hierarchies remain firmly in place. It focuses specifically on the role of violence. Continuities in the maintenance of a racialised order co-exist with important changes in agriculture. They depend on a range of forms of coercion and vulnerability that are not simply extensions of past practices. In order to appreciate this, it is useful to extend definitions of violence to encompass quieter as well as more explicit forms of control. For workers, violence is not merely overt, but also structural and symbolic. Structural violence, in Farmer's formulation, refers to the established arrangements that render certain people vulnerable. 17 Symbolic violence, in Bourdieu's formulation, refers to the habits and predispositions that maintain and naturalise everyday domination. 18 Both are analytically important here because they form the foundations of more overt, conspicuous incidents, such as that described in this article's opening. Moreover, as Scheper-Hughes and Bourgois note, 'violence can never be understood solely in terms of its physicality -force, assault, or the infliction of pain -alone. Violence also includes assaults on the personhood, dignity, sense of worth or value of the victim'. 19 What this article investigates, however, is how quieter and more overt coercion are actually connected -how physical coercion is not merely enabled by a racialised hierarchical order, but also ultimately underpins it.
In what follows, I consider the relationship between a corporate style of management and the panoply of coercive conditions that continue to render a racialised order inviolable. As farmers today downplay their responsibilities in workers' personal lives and in noncontractual arrangements, senior black male workers are left as the central figures of a mediated paternalism. 20 They hold court, invoke vigilante justice, and call on the coercive power of stateemployed border guards. Workforce hierarchies incorporate transient people, in the process reconfiguring paternalist dependence and producing new forms of violence. Senior workers benefit from their roles in these hierarchies. Indeed, farmers tend to condone violent measures, rather than initiating them, as in the opening vignette.
The key role of black patriarchs has a long history, as van Onselen showed. So does the unequal distribution of vulnerability within black farm workforces, which has long been intertwined with the overt enforcement of racialised hierarchy. Du Toit's seminal analysis of agrarian paternalism presents vertical ties between workers and landowners as forged from a hegemonic discourse that shapes farm dwellers' possibilities. But he hints at the structural role of violence at the heart of the system, including out of sight -and beyond the direction -of white farmers. 'The emotional and physical battery workers experience in the work situation is reduplicated at home: it is through rape and domestic violence that male workers express their masculinity and authority'. 21 Vulnerability in the workforce is inseparable from the wider maintenance of hierarchical order. Yet today white farmers appear especially far removed from the maintenance of arrangements that, as I show below, leave them unambiguously superordinate -not merely as employers, but as white men. Nevertheless, behind the scenes, they do set the ultimate terms of hierarchy within the black workforce -and between black and white farm residents. Senior workers are put in a position where they have little choice but to engage in vigilante justice, if they are to sustain the arrangements that underpin farm life and labour. And, while coercion may be highly distributed, permanent workers still rely on farmers for the protection of their jobs and homes. At the same time, senior workers are quietly kept in their racially defined place, and reminded of the limits of their own power, such as through private acts and threats of physical harm. Even intimate relationships between employers and employees may be shot through with the possibility of racialised violence.
The article focuses on one border farm, which I call Grootplaas, and in whose labour compound I resided for 17 months between 2006 and 2008. A citrus estate, Grootplaas employs 140 permanent workers and a further 460 seasonal workers during the harvest. Political and economic crisis in Zimbabwe -including acute hyperinflation, supply shortages, and violent oppression of opposition supporters -was reflected in the large numbers of Zimbabweans climbing through the border fence in search of an income. Grootplaas and the other border farms became -and remain -islands of wage labour in a sea of informal arrangements.

Carving Out a Space for Corporate Agriculture
Farmers on the border today present themselves as businesspeople, whose factories in the fields are in principle no different from any other place of production. Prompted by market liberalisation and the threat of land restitution, they strive for flexibility by means of portfolios of estates and other agriculture-related enterprises (such as pack shed and transport services). 22 They respond to minimum wage and housing legislation by downplaying non-contractual obligations to workers. And they attempt to avoid depiction as apartheid-era anachronisms by foregrounding a version of agriculture that effaces the past.
Their approach is thrown into relief by the game and small-scale crop farmers behind the border, who preach and practise a more militarised lifestyle, which is in turn crystallised in the armed-response initiative called Farm Watch. This grew as a reaction to what they see as racially targeted attacks on farms, and involves a clear show of strength in the form of visible firearms and bakkies with green, flashing emergency lights. On one occasion, I was invited to a 'self-defence' camp run by Farm Watch in the Karoo. This combined training with a range of guns, and target practice in different farm-attack scenarios (including driving up to and away from imaginary ambushes), with evenings of group singing. When I returned to the border, I was told that I had been spending time with 'weekend warriors'farmers whose small-scale operations left them with too much occasion to imagine and enact apocalyptic scenarios.
By contrast, the border farmers claimed to be defined by their strict orientation to profitable enterprise. In some cases, resembling what Pilossof calls 'affirmative parochialism' among Zimbabwean farmers, 23 this is an attempt to avoid 'politics' in post-apartheid South Africa. Unlike their more conservative counterparts in Farm Watch, the border's commercial farmers present themselves primarily in 'economic' terms. Indeed, several are of Zimbabwean origin, and they claim precisely that this shapes their approach to agriculture. According to Jan, who farms next door to Grootplaas: We [Zimbabweans] weren't a politician or a businessman or what-have-you, or a teacher that owned a farm.… The whole country revolved around farming. So the best of the best were bred there, amongst this farming community. Business-wise and farming-wise and how to make success out of farming.
Despite Jan's claims to a nationally specific approach to agriculture, his views are echoed among the area's South African farmers. One, Thinus, put it like this: 'I am one of the difficult people in life. I'm a perfectionist. I will do a thing right or I will not do it. If I open a farm I will not open a bad farm'. Whatever their particular inflections, all of the border's farmers insist on their single-minded focus on successful enterprises.
Of course, this kind of profit-orientation does not by itself imply a lack of association with violence in labour arrangements. It was precisely the large, capitalist farmers of the Eastern Transvaal who were most conspicuously guilty of mistreating workers at the inception of apartheid. 24 But, in the mid-20th-century Eastern Transvaal, labour shortages were the backdrop to the state-sponsored exploitation of unfree deportee and prison populations. Farmers on the Limpopo today have few difficulties securing labour. A wider context of desperation and vulnerability means that large numbers of prospective pickers and packers appear at farm gates on recruitment days. Many who are employed are keen to retain their positions, however badly paid. 25 Moreover, on the border, this is combined with a corporate style, and even a veneer of de-racialised self-presentation. At Grootplaas, Willem -the farmer responsible for everyday operations and the brother-in-law of the farmer in the opening vignette -refers to an opposition between labour and management, not one between black and white. During one meeting in his office, he explained the Grootplaas labour hierarchy. At the top of the pyramid are the four 'shareholders' of the company -Willem, his father-in-law, Koos, and two of Koos's sons. A rung down are senior employees: the white workshop foreman and mechanic, and the white production manager, who is responsible for all operations around growing the fruit. They are joined in 'management' by Michael, the black personnel manager (sometimes known among workers in the area as the clerk); Marula, the black general manager (commonly referred to as the foreman); and the black pack shed manager (commonly, the pack shed foreman). Sitting on 'the management side of the table', these three are paid a monthly flat-rate salary.
It was no accident that, when I arrived at Grootplaas, I was assigned to Michael, the personnel manager. Although he lived in the compound like other workers, and was ultimately subject to the same hierarchies, he was also the only non-manual black worker, and his own investment in a corporate image was revealed by his pressed trousers, clean white shirt and Bluetooth earpiece as he sat in a distinct personnel office, used only by black workers but positioned just along a corridor from the farmer and his wife. The black receptionist of the neighbouring farm, who welcomes visitors at the offices in a shirt and tie, similarly disrupts established distinctions between white and black, in which the latter are ordinarily seen in blue or green overalls. A corporate style is evident even at the gates of different estates. Game farms to the south display warning signs replete with images of dogs, guns and electric shocks; the border's crop estates advertise their names, branded logos and pack sheds at their front gates, and even Coca Cola-sponsored store signs at turnings into compounds.
The audience for this apparent transformation is largely the export agents, private and government inspectors, and foreign supermarkets with which these family-run companies now negotiate. But the shift in style has been accompanied by a retreat from everyday workforce dynamics, which involves establishing distance from both strong-arm discipline and broad responsibilities to workers. Willem's nickname, Mpothe, is taken by workers to mean 'hit', memorialising his history of violence. Stories differ as to whether he hit a black butcher who was trying to sell meat in his compound after an earlier altercation, or whether he made a dent in an apartheid-era police van after a dispute about 'his people'. But he is now careful to assert clear parameters to his role. He keeps work to the office -indeed, his stipulation that our meetings take place there contrasted with older farmers who invited me to their houses. When disputes between workers are brought to him, he pleads ignorance, claiming a lack of sufficient understanding of 'the culture'. It is this attitude -not tales of a volatile temperament -that employees compare unfavourably with that of Koos, Willem's octogenarian father-in-law who previously ran the estate.
For border farmers, a corporate style of farming has been carved out from an environment that was itself characterised by violence. The white farmers came to the border in the first place as a buffer zone to enhance the state's territorial control, in the low-level 'border war' with Umkhonto we Sizwe (Spear of the Nation) 26 of the late 1970s and 1980s. They were offered low-interest loans for land, and provided with infrastructure -an electrical substation, telephone lines, and improved roads. In return, they were required to be resident, not absentee, landowners. And they were to undertake their compulsory military service as white adult males in the local commando, a model of conscription that drew on earlier Boer militias in farming districts. Willem explained: With the terrorists at that time that came from Zimbabwe, we were always on a standby, so they could call you anytime and say, 'ok well we've got a problem here' … We had a commander which would then phone us, or radio us, and say, '… Someone came through with landmines … We want everybody to get their kit and get together at a certain place and then follow their tracks and see whether we can pick them up'.
Mine-sweeping preceded farm work each morning. Of course, the risks fell on workers, not their employers. Indeed, while the buffer represented by the farmers, and the electrified razor-wire fence that followed it, were intended to block the movement of strangers, the farms needed workers. The devolution of risk down the hierarchy is key to how the farms have operated since their inception, and I return to this below. For now, it is important to recognise that, much like the pre-1948 cases mentioned above, the border farmers consolidated their place in the landscape through a show of physical strength -gun training, patrols, and armed response -that also parallels Farm Watch today.
It is precisely from this history that the border's big cultivators wish to distance themselves. Farmers like Jan are keen to underline that, despite insecurity in the past, the point of coming to the border was to make a success of their ventures, and that constituting an armed presence played at best a marginal role. They speak of their relative security in a place of strangers. Despite (or because of) a turbulent past, they feel little need to patrol. Few carry guns. Indeed, most who do so -Koos, for example -conceal the fact. Willem explicitly rejects the idea that he should (although his twenty-something son has begun to assert a particular masculinity by wearing a handgun in an underarm holster). It is important to recognise that farmers' attempts to distance themselves from violence do not mean that physical force has dropped off their radar. Koos was keen to impress on me that he is not a 'soft target', and that, if push comes to shove, he can look after himself. His scenarios for attack generally involved betrayal by a (hypothetical) trusted worker, after threats to his or her loved ones.
But, as I demonstrate below, farmers' safety in fact relies on a whole terrain of interlocking coercive mechanisms. Central to these are the workers who, in an earlier era, bore the brunt of the risk in clearing for mines. They leave farmers free to present themselves as savvy, globalised, business-oriented employers -neither obligated patrons to workers, nor landowners who look after themselves through their capacity for violence. Senior workers maintain and even promote paternalism within the workforce, which in turn enables them to control the environment around them. And, while permanent workers do the job of regulating compound life, complete with the threat of coercion, state institutions play a dual role. Soldiers are incorporated into the work of keeping the peace on the border estates, while the police produce the kind of environment that renders roads dangerous places for migrants and ensures docile workforces. The next section begins to unravel these arrangements by starting at the bottom of the farm hierarchy with migrant seasonal workers.

Mediating Transience: Provision and Vulnerability in the Seasonal Workforce
Focused on the capacity for harm inflicted by social arrangements, not simply by individuals, the concept of structural violence underlines how vulnerability is a feature of inequalities that are actively produced. In this regard, while earlier South African scholarship benefited from a comparison with extra-legal coercion in the American south, 27 recent ethnographic studies about migrant farm workers in the United States make for an equally useful reference point. For example, Benson, writing of tobacco estates in North Carolina, explores how farmers downplay their own roles in shaping their workers' living conditions. 28 The task becomes to trace these connections. Below I describe how seasonal workers' vulnerability is in fact an integral part of a wider system of coercion on the border farms.
When migrants cross through the fence, they are unable to plan with any certainty beyond the immediate present. Their vulnerability comes first from the possibility of robbery and rape at the hands of magumaguma -the border gangs -and, reportedly, South African soldiers. Once they have climbed through the fence, they risk arrest and deportation by patrols on the roads between the farms. Once employed, during the period of my fieldwork, seasonal workers were often left in an undocumented limbo -part of farms' labour quotas, but indefinitely awaiting ratification by the South African Department of Home Affairs. Police would often sweep through the compounds, rounding up workers for deportation. Home-made farm IDs would offer some degree of protection, as farmers negotiated with the police and army their prerogative as landowners and recruiters of labour. 29 At one level, all of this is an inconvenience that hinders farmers' enterprises. There is paperwork, and deported workers sometimes return for their jobs and pay. Farmers have been known to block police who attempt to enter their estates, in protection of 'their people'. Yet there are also local horror stories of farmers contacting the authorities at month-end, having workers deported, and thus enjoying a period of unpaid labour.
More generally, the insecurity experienced by seasonal workers is not merely accidental; it is a feature of power arrangements on the border, in which the farms' own hierarchies are implicated. In the 1990s, a 'special employment zone', running along the border and back 50 kilometres to the Soutpansberg, was instituted, which set the pattern for farmers' locally powerful positions and their control over mobility and settlement. 30 For Zimbabweans crossing the border, a system of special ('BI-17') permits placed a premium on connections with particular farmers and their black foremen. This highly personalised system of labour regularisation meant that networks developed between the South African farms and villages across the river in Zimbabwe, as people sought avenues for employment. Not only did this result in webs of kinship and other relationships emerging both across the border and between the border farm populations, it also led foremen to become key figures in the border area. Such dynamics were further consolidated by the opening of an 'informal' border post at Gate 17, a point on the border fence immediately adjacent to the farms. According to my informants -both farmers and farm workers -this was a somewhat ad hoc affair, in which a few South African officials would set up a mobile station on the southern bank of the Limpopo. Nevertheless, it was considered common practice for recruits to make their way to the farms without papers (in this case avoiding Gate 17), where they would in theory be issued with a permit. 31 The risk of arrest and deportation when undocumented was borne entirely by the workers themselves. Some farmers, indeed, would not document seasonal workers at all. One reason migrants crossed clandestinely was that the risk of deportation diminished once migrants reached the farms. The estates were considered 'protected ground', on which one would not be arrested. Moreover, post hoc legalisation was easy, and indeed remained the most convenient method of documentation. And many migrants, then as now, did not yet know their ultimate destinations. 32 Farmers welcomed this state of affairs, because it meant that recruits would turn up ready to work on the first day of the harvest; cumbersome documentation could be avoided, at least initially; and a 'grey' employment status made for a vulnerable, easily exploitable workforce. 33 Today, precisely these expectations about employment, documentation, and risk shape harvest-time recruits' lives on the farms. Recruits are expected to gain documentation once they commence work. During the period of fieldwork, they were taken to Beitbridge border post in groups for their paperwork to be processed, 34 with priority often given to those with histories of employment and connections among permanent workers. Those who were waiting for permits, or who never received them, slept in the bush for fear of police raids, despite the bitter winter nights, or asked friends to lock them in their rooms so that accommodation would appear empty. In short, the hazards of 'grey' employment status continue to be pushed down the hierarchy.
Today, the BI-17 permits and the 'informal' border post have disappeared, replaced since 2005 by procedures following standard South African immigration law. Corporate permits for fixed numbers of foreign workers are issued after employers demonstrate inability to find sufficient South African labour. In practice, however, this protects border farmers' existing autonomy: 'proof of the need to employ the requested number of foreigners [is] … evidently a formality'. 35 The result leaves intact what Rutherford has called 'farmer sovereignty' on the border, 36 in which farmers enjoy considerable autonomy or support from border authorities. Established workers often go through the border fence at the garrisons, with soldiers' permission, or simply use the farms' own gates, which lead to water pumps in the river. Border officials accept that recruits come straight through holes in the fence, bypassing conventional forms of state control. The degree to which workers are able to belong on the estates depends on their being seen to belong to white farmers. 37 Permanent workers live settled lives under these conditions; their seasonal counterparts experience far greater precarity. The effect is not too far from the logic of apartheid-era pass laws. Farmers stand as sponsors for migrants and govern the right to reside, while seasonal workers live a furtive existence that means avoiding roads and shuttling between orchards and compounds. The farm identity cards issued to workers, which mitigate the threat of arrest even if they are otherwise undocumented, underline the particular intertwining of incorporation and (structural) violence, provision and coercion, in this setting.
The structural violence of displacing vulnerability to harm down the hierarchy is allencompassing. As in Holmes's ethnography of migrant farm workers in the United States, it is assumed that the lower down the ladder one is, the more labour means physical damage. Permanent workers may be exposed to fumes when burning chemical fertiliser, for example, but seasonal workers complain of lacerations from thorns on citrus trees, exhaustion from fastpaced picking work while carrying heavy shoulder-bags, and even worn scalps from unsteadily balanced crates in the pack shed.
Yet the specific relationship between provision and structural violence goes further. It is inscribed in the very architecture of the labour compounds in which workers are housed once they are employed. Grootplaas's accommodation, much of it electrified, is some of the best on the border. The brick blocks are considered preferable to the corrugated metal cells or mud dwellings on neighbouring estates. At the same time, the long rows of outward-facing doors, the heat inside rooms from metal roofs, and overcrowding in seasonal workers' rooms leave new recruits virtually no privacy. The largest area of the compound, where most seasonal workers live, is built in a massive grid that leaves nowhere to hide from deportation patrols. The floodlight on one side replicates the ambivalence of provision and vulnerability. It is welcome in a place where robbery is always possible, but it also completes the classic picture of a regimented, surveyed living environment, and aids the task of night-time deportation patrols. All of this contrasts sharply with the living conditions of permanent workers. The latter have single-occupancy rooms, or two-room houses if particularly senior, located in more secluded parts of the compound. Despite an ultimate lack of security of tenure, permanent workers adapt their houses and plant gardens. Documented, and displaying confidence in their body language, they are rarely even checked by deportation patrols. Their accommodation and circumstances reflect their greater embeddedness, security and authority on the border estates.
Under the floodlight is the 'hostel', a row of large rooms that house 20-30 seasonal employees in unsanitary conditions. Here, as in the main grid, many residents spend as much time outside as possible. Workers' complaints about the risks indoors of infectious diseases, such as tuberculosis, were far from implausible; there had been a meningitis outbreak on the border farms shortly before my fieldwork, killing one person at Grootplaas. The hostel was itself an example of provision, albeit a limited one. It had been built after media coverage, the attention of municipal health services, and soon the scrutiny of other state departments around the meningitis outbreak focused on the housing of some seasonal workers in a large tent. Another measure that followed the scare was a worker-run housing committee, whose task it was to allocate housing according to due process. Yet the committee soon exemplified the power of permanent workers. Unsurprisingly, personal connections predominate and, for Marula, the main everyday role of the housing committee's two members is to call people to the court outside his house. Accommodation for seasonal workers illustrates their incorporation into farm arrangements, but also their exposure to far greater vulnerability than senior workers or their employers, in a setting where they risk arrest and the spread of disease.
On the border farms, vulnerability is mitigated but also ordered in new ways -it becomes part of how the farms operate. But there is more to this than structural violence in the sense of everyday indignity, risk and precarity. As David Graeber points out, structural violence is most usefully seen in terms of the threat of coercion built into institutionalised arrangements -in other words, the relationship between implicit and overt coercion. 38 Understanding how this works on the border requires a closer focus on Grootplaas. Crucial is the power of senior workers, as I show in the next section. But, as we shall see, farmers continue to shape workers' lives profoundly.

From Precarity to Punishment: The Power of Permanent Workers
The opening vignette offers a window into the regulation of the workforce and the compound in the context of the border. What comes into view is a black working population at Grootplaas built around the black foreman, the iconic figure who would formerly have been known as 'bossboy'. Marula is one of the farm's few South African workers. As emerged above in the discussion of seasonal workers, he and a small circle of top employees act as gatekeepers to employment and permits, allocate housing, and have the ear of the white farmers when it comes to endorsing or expelling residents. As in agriculture across the region, 39 senior black figures here govern their subordinates on behalf of white landowners. Their generalised authority is perhaps most sharply visible at the khoro, 40 where Marula's work supervisors from the orchards are elders. Marula himself has the final say as a kind of headman. These public tribunals are held outside his house at the centre of the compound, at a purpose-built semi-circle of benches. They draw on the shared origins of many senior workers in villages just over the border, their established positions within cross-border webs of kinship, and their history of bringing relatives to work on the farms. Marula himself offers personal protection to those who appeal to him for assistance. Other senior workers are also patrons in idioms other than that of labour: one is the manager of the seasonalworker football team; another is a bishop in the United African Apostolic Church. One way that established workers gain prestige and recognition is by safeguarding seasonal workers' earnings in their individually occupied, lockable rooms. Crucially, such acts of patronage, understood as manifestations of generalised authority and responsibility, and building on webs of kinship and shared origin, make for a more thoroughgoing paternalism within the workforce. 41 As the opening vignette suggests, the border setting brings its own inflection to the relationship between violence and paternalistic hierarchies. At the most obvious level, the role of senior workers in maintaining order is lent particular significance by the possibility of theft or violence in a place where people can simply disappear, whether southwards or through the border fence. The everyday risks of border life were impressed on me most strongly by children. On one occasion, I had driven to a game farm to speak to the black manager there. I was accompanied by Marula's daughters, who were close friends of the manager's own children. Only the latter were home, and we passed some time chatting and waiting for their father to return. Meanwhile, a man appeared through the bush, dressed in ragged clothes and asking for water. The response was immediate. With obvious concerns about the threat he represented, the children threw rocks at him and drove him away. They were adamant that asking questions first would only bring trouble. Transience, then, brings dangers for both border residents and for migrants, most of whom simply seek shelter, sustenance and employment. Part of the context is that the police are known for being overwhelmingly concerned with deportations, not people's access to justice. The border farms hold the possibility of violence even for the most established. Workforce hierarchies mean little away from the estates; one respected work supervisor has been robbed several times while crossing the border. Senior workers are affected by personal insecurity even on the farms themselves. Marula, for example, invited a ng'anga (healer) to stay with him, not just to play the part of patron, but explicitly to guard himself against invisible attack in a highly insecure setting. That the ng'anga later ran off with his wife only underlined his vulnerability, born of people's ability to disappear from the farm. His primary rival at the top of the workforce hierarchy -Michael, the personnel manager -has been assaulted several times by workers. After payday drinking, a bottle over the head around the back of a shebeen is always possible.
But senior workers appear secure in comparison to their seasonal counterparts. This is partly a matter of self-presentation: when attacked by other permanent workers, Michael avoids reporting these incidents because doing so would detract from his authority. Similarly, Marula's experience with the ng'anga was discussed in hushed tones. Beyond the maintenance of reputation, the odds are in fact stacked in senior workers' favour. If anything serious occurs, it is Marula and two other foremen who have the two-way radios connecting them to their employers. And, as discussed below, soldiers often rely on senior workers to understand what is happening around them. For senior workers -and especially Marula and his supervisor-elders -the transience around the farms appears as both danger and legitimation. The maintenance of order and personal safety confronts the insecurity of the border itself through the possibility of violent vigilantism.
More generally, permanent workers' discretion is only enhanced by their position at the centre of different patterns of movement and settlement on the border. The seasonal workforce is composed of migrants en route to Johannesburg, regular faces from across the border, the dependents of established workers, and 'semi-permanent' women, who sweep the compound and chase away baboons in the orchards out of the harvest season. Many people also come to the compound for shelter, or as a place for informal trade. For migrant women, male permanent workers represent access to housing and livelihoods, as well as a voice in a highly unequal setting, where -if disputes and bragging among men are anything to go by -sexual violence is widespread.
Yet this is no idyll of impartial and peaceful regulation. Representing a challenge to workforce hierarchies, the wider context of large-scale displacement from Zimbabwe has cast the authority of supervisors in a fresh light. Workplace camaraderie brings male recruits under the protection of influential figures, increasing the likelihood of receiving a work permit and offering the stable residential base to establish informal businesses. Yet some have experienced middle-class decline in the Zimbabwean crisis -at the most extreme, former teachers and nurses. Casting themselves as respectable in contrast to their team-mates and supervisors, and unwilling to accept their supervisors' superiority, they complain about kicks and punches in the orchards from those better embedded in workplace hierarchies. 42 In the compound, while male elders stake their claims to seniority and respect on notions of proper authority, they are arbiters of a highly gendered farm order, and there is little redress beyond them. This is especially the case for women. Cases of sexual violence may simply be argued between aggrieved men -the accused and the boyfriend -and even mediated by a soldier in private discussion with them. And, on one occasion, there was an attempt to evict a female employee and her taxi-driver husband after she rejected advances on her by permanent workers. Permanent workers maintain the conditions for a congenial domesticity by patching together relationships with transient women. But many of these relationships are short-lived, and the risk of pregnancy or sexually transmitted infection underlines women's vulnerability. Lying at a nexus of paternalist protection and violence, amid intersecting patterns of movement and settlement, permanent workers' roles both extend a kinship idiom into compound life and perpetuate forms of exposure to harm.
Farm hierarchies mesh with the border's military presence to create a peculiar enforcement of order. Manning a 10-kilometre-deep zone of jurisdiction, the army's formal role is to check southward migration and smuggling. During the period of fieldwork, I was told by one soldier, hundreds of Zimbabweans crossed the Limpopo daily when it was not in flood. Large numbers of cigarettes, as well as unprocessed organic products (such as grass mats and curios) and meat that risked carrying diseases such as Rift Valley fever and foot-and-mouth, made their way southwards. In both cases, despite the patrols of soldiers up and down the border in converted bakkies, and concealed outposts on farm roads tracking away from the border, the army's reach was limited at best. The fence was ragged with holes made by wire-cutters, and floods would leave whole stretches flat. Soldiers openly acknowledged that there was no way to apprehend even a fraction of migrants, and I heard complaints that there were no mechanisms to prosecute repeat offenders. The little control they did have depended on the co-operation of the border's settled residents -both black permanent farm workers and white farmers. The result was to give unofficial but unmistakable sanction and force to farm arrangements.
Soldiers spend a total of three months on the border and, although they are rotated every two weeks between the 'echo stations' distributed along the fence, they become embedded in networks of farm workers. New arrivals complain that, hailing from elsewhere in South Africa, they feel unable to speak to people. Many feel lonely in garrisons of around 10 troops, and they soon frequent shebeens in the farm compounds in search of beer and women. While some relationships I encountered were framed in explicitly transactional idioms -sex for army rations, for example -others emphasised enduring obligation. One soldier established a relationship with a farm resident, to which he devoted considerable energy even after he was circulated up the border to another echo garrison. More generally, soldiers would tell workers that they understood why Zimbabweans came through the fence, acknowledge the futility of trying to stop them, and appeal to those they spoke to in the compound not to cut new holes in the wire, since that would render their own failure visible. Soldiers, in turn, would be regarded as far more sensitive to local distinctions and dynamics in the compounds than the police, who would sweep in from Musina, 50 kilometres away, during deportation raids.
While white farmers appear removed from these networks, which set the terms of coercive arrangements, their own personal connections also inflect border regulation. Within the workforce, senior workers' power is possible only because of farmers' tacit support, regardless of any talk of distance. Farmers are extremely reluctant to dismiss key permanent personnel, even in the face of alleged criminal behaviour. Of course, the protection they offer only underlines their power as employers. Dismissal (which occurs, for example, when a worker is caught poaching) is dramatic. It means the abrupt loss of a job, a place of residence, and an anchor in established social networks -departure is ensured by the threat of compulsion by security guards. As for the military, during the period of fieldwork, the white commander of the border detail would regularly visit the crop farmers along the border, individually or in 'forums', to gather 'intelligence' on border dynamics and on the performance of his troops. This builds on relationships that precede the end of apartheid. Indeed, because of the continuity of senior military personnel on the border, relationships between civilians and the army extend back to the State of Emergency in the 1980s, when farmers did their military service in the border's commando.
Some bonds with the military go further. One game farmer runs the area's sole remaining proofplaas (test farm), measuring the land's carrying capacity. A close associate of the border's commander, he has a sideline training soldiers to track Zimbabweans through the bush. He showed me the variation in footprints, and told me that cheap Chinese soles indicate Zimbabweans who have not yet worked in South Africa; those returning south generally have the better shoes available there. For this farmer, tracking game and tracking people, and civilian and military engagement, had all converged: when I visited him, he showed me his collection of photographs of human corpses that he had found in the bush.
It is in this context, in which farms are central to the organisation of people and space on the border, that the opening vignette must be understood. Just as state officials rely on farm workers and farmers to navigate the area, so the latter also draw soldiers into their regulatory arrangements. Soldiers in the compounds are asked by senior workers to mediate disputes. As for punishment, the elders of the khoro have the sjamboks 43 of security guards and the mob violence of the workforce at their disposal, or the possibility of further retribution in villages just across the border in Zimbabwe. But these are rarely invoked. For farm dwellers, the soldiers represent ultimate coercive capacity.
Historical scholarship on early 20th-century South Africa has examined white extra-legal violence as a way to maintain racialised, hierarchical order in rural areas. A similar order is now maintained through collusion between white farmers, senior black workers, and border guards. This kind of distributed coercion has long existed to some degree, with white and black patriarchs enforcing their positions and agendas through physical force. Today, however, white farmers appear removed from any vigilante violence, while senior black workers actively entrench themselves as figures of substantial power. But if the latter mediate paternalism down the chain of command, this is at the pleasure of their employers -it all depends on having a job. Moreover, the final section examines how, even as everyday influence is devolved down the hierarchy, senior workers are kept in their place by an absolute distinction between black and white.

Naturalising the Hierarchy: Symbolic Violence and its Challenges
The sharp distinction drawn between white farmers and managers and black workers is sustained largely without commentary. Workers may speak of specific instances of racism, but it is broadly understood that they sit in the backs of bakkies, while farmers' dogs occupy the passenger seats. Even among Grootplaas's black managers, treatment different from white employees is assumed. Willem discusses plans for the orchards with the white manager in the air-conditioned tearoom. Marula, formally the lands manager, is told what to do in the workshop yard outside. Michael, the personnel manager, is spoken to in the same slow, hard monotone as is used for other workers. Conversely, workers who command the obedience of hundreds of people on the border avoid direct eye contact and drop their shoulders when speaking to white employers.
Yet, occasionally, assumptions about racialised hierarchy are made explicit. In this brief section, I focus on these moments to offer an indicative sketch. In interviews, farmers would reflect in the abstract on the differences between white and black people. In general, these revolved around the 'vision' of whites, and on their capacity to plan; in one case, this was given an environmental gloss by a contrast of the needs of 'ice man' (who learned to store for the winter) and 'tropical man' (who never had to). Conceptions of thoroughgoing difference translate into interpretations of divergent life conditions on the farm. Resembling the entrenchment of hierarchy among American farmers and Mexican migrant employees, farmers present workers as responsible for their living circumstances. 44 Hygiene problems, born of overcrowding, come to appear the result of a dirty 'cultural' space; visiting farmers expressed disgust that I was inhabiting such a setting. These depictions take different forms. Older farmers draw on an established, traditionalist rhetoric of cultural distancing. Koos, Grootplaas's retired founder, refers to Marula, the foreman, in language that suggests authority merely sanctioned by the farmers: 'we support Marula as an induna. That's why he drives the newest bakkie'. 45 During afternoon conversations in his living room, he would suggest to me that I speak to Marula and others about 'the tokoloshe' 46 while sitting 'around the fire' in the evening. The only fire in Marula's compound area is to heat a boiler, for the underground plumbing he has built to his shower unit; seasonal workers, on the other hand, have to burn wood for warmth and cooking because they are unable to purchase stoves. Yet, through a traditionalist lens, workers' fires evoke old, rural ways. Grootplaas's younger generations, who today run the estates, employ a 'modern', corporate register -they do not pride themselves on speaking workers' languages, or on 'understanding their people'. But the effect is the same. As discussed earlier, Willem sees life beyond work as outside his purview, and its complexities as beyond his expertise. Personal disputes in the workforce are put down to a different and impenetrable culture. Whether emphasising a traditionalist or a corporate register, he and other farmers rarely enter the compound.
Moments in which assumptions are made explicit are perhaps more revealing when they occur in the course of everyday interactions, not in private. Particular places of interaction between white and black become stages on which difference is performed. Early in my fieldwork, I was made acutely aware of the rigidity of the white/black distinction in farmers' presence, when I accompanied compound residents to a polocrosse 47 tournament organised by farmers on the next estate. During the day, farmers and farm workers stood around the sidelines of the pitch. This leisure-time proximity was an unusual experience for both. The result was that farmers would periodically move compound-dwellers away from them, so that they did not 'upset the women'. Meanwhile, one farmer, fuelled by a few beers, began shouting 'shamwari' 48 and 'ZANU PF' 49 aggressively in the direction of the gathered black spectators. Towards the end of the day, one worker (a companion with whom I had come to the tournament) approached me while I was talking to a white farmer, to inform me that he and others were returning to the compound. Unsure of how to negotiate the situation, he stood, head down and at some distance, until the farmer loudly asked him what he wanted. After he mumbled that he was waiting to speak to 'Mr Max', we stepped aside and made plans for the evening out of earshot. What was clear, however, was that there was no conceivable way to have an amicable conversation in front of the farmer, where the existing roles appeared inviolable.
Another stage is the farm workshop, where recurring interactions across the harvest cycle underline the white/black distinction. Once, when seasonal workers were gathered to take photographs for their work permits, the white secretary came to the personnel manager's office where the queue had formed. In a conspicuous manner, she complained that she would have to leave because of the smell. Later in the harvest, at payday, seasonal workers are asked to 'sign' for pay with a thumbprint, although most have some schooling, and many have O-and even A-levels. At various points, therefore, regardless of their diverse backgrounds and their own disgust at conditions in the compound, they appear simply as uneducated, unclean bodies.
One feature of this naturalisation of difference is what David Graeber calls the 'interpretive labour' involved in structural violence. 50 Those at the top of the hierarchy need to spend far less time and effort interpreting the perspectives and experiences of those at the bottom than the other way around. Farmers' lives are submitted to close scrutiny. The return of a son, or the approach of a sporting event, becomes the stuff of gossip in the compound. Farmers are compared in some detail in terms of their individual characteristics and histories. After all, making sense of them is key to workers' understanding and negotiating power. By contrast, with the occasional exception of senior employees, workers appear as a mass defined more by collective culture than by individual particularities. But, as Graeber argues, this very asymmetry relies on the threat of overt violence standing behind unequal arrangements. It is partly because force may be mobilised that superordinates need worry less about grasping subordinates' points of view.
All of this inflects senior workers' positions on the border estates. Once again, overt enforcement is rare but significant, and it can take very different forms. Michael, the black personnel manager, who casts himself in the same corporate idiom as his employer, Willem, was threatened with physical harm when he spoke English to Willem's father-in-law. He should have spoken an African language or Fanakolo, the hybrid language of labour. He was new to the job (and to farms in general), and such experiences must be understood in the context of workers' dependence on farmers for work (remember that being fired comes with its own threat of overt violence, in the form of the security guards overseeing expulsion).
In other cases, violence may be interwoven with intimacy. Benjamin, the store-man, would often attempt to explain the intricate relationship between himself and Willem's brother-inlaw. 51 Eventually he resorted to a story about the farmer's dog. Both were based in the farm's workshop, and so spent their days together. Occasionally, the farmer would order Benjamin to run to get something, knowing full well (as did Benjamin) that the Alsatian would attack him as soon as it looked like he might be running away. In doing so, the dog was responding to racial stereotypes, learned from its owners. And, for the farmer, these very stereotypes were easily manipulated in the service of a quick practical joke. For Benjamin, the whole process was rather less funny, and of course it was physically painful. He also read such moments as revealing a shared understanding between farmer and worker, reflecting trust (he would underline his employer's physical vulnerability after a serious motorbike injury) and a long, close working relationship. But it was acknowledged in a moment that also re-asserted racialised hierarchy. Permanent workers may call the shots on the border. But, behind the scenes, they are kept in their place by quiet assumptions of difference, and the occasional demonstration of power asymmetry.

Conclusion
The explicit farmer violence of van Onselen's description has disappeared from the border farms. Yet the farmer's joke with Benjamin bears its trace. As a farmer reaches out to his most trusted worker, his cultural resources have physical coercion at their very centre. Moreover, such relationships are between small numbers of white and black men, and their intimacy sets the terms for wider, highly gendered patterns of dependency and domination. The power of senior black workers leaves women and junior men extremely vulnerable, as farm hierarchies are reconfigured in the shadows of corporatisation and the language of business. Paternalist provision and disciplining violence remain central; the labour hierarchy is the safeguard against and the foundation of people's unequal exposure to harm. The order that results involves a complex, distributed network of coercion that leaves farmers appearing removed while also offering them the benefits of a kind of vigilante policing. This order is remarkably effective. It enables border farmers to feel relatively secure in a place of transient strangers. And, unlike 51 Not the one from the opening vignette. the Cape, 52 or even farms close to the former homeland of Venda to the east, 53 the border estates have seen virtually no organised contestation of farm hierarchies and arrangements.
Farmers continue selectively to promote elements of paternalism, while profiting from the structural violence that leads migrant seasonal workers to depend so heavily on them and their senior employees, and the symbolic violence that places them indisputably above their compound populations. The farms' mediated paternalism disguises their deep implication in the lives of their compound residents. What compound dwellers generally see is senior workers' apparent autonomy, with its coercive or protective potential. Underlying this power are threats that are only occasionally realised. Less visible are those quiet moments when a worker is reminded of his subordinate status, or his vulnerability, or when a farmer leans out of his bakkie, gently condoning vigilante justice in the workforce: just knock them around a bit more, and then hand them over to the soldiers.