‘We are not just a union, we are a family’ class, kinship and tribe in Zambia’s mining unions

Trade unions in Zambia and in several other developing countries have been understood to create ‘detribalising’ class consciousnesses. In contrast, we argue that Zambian understandings of unionism have developed through similar political economic processes to those that generated ‘tribes’. Values and structures that enable concepts of the good life more commonly found among Bemba speakers and Eastern Zambians have been naturalised into Zambia’s mining unions, guiding union policy and practice in a manner which limits North Western Zambians’ union participation. Utilising Lazar's (2018) understanding of unionism as kinship, we explore how Zambians of various tribes attempt to utilise unions to achieve what they see as human flourishing and social justice. We foreground that people’s understandings of the good life frequently incorporate gendered and gerontocratic hierarchies and we demonstrate that intra-national unionisms are co-created through (and influence) local cultural norms and political histories. This encourages anthropologists of trade unionism to ask what values and hierarchies are rendered invisible in other union ‘families’, and to explore intertwinements between unions and communities enabled through kinship, rather than through Civil Society Organisations.

The president's speech was met with roaring applause, yet it demonstrated two common misunderstandings of how tribe interacts with Zambian politics and with trade unionism. Most obviously, many members of the North Western Forum were not 'tribally' North Westerners. Rather, they were Copperbelt residents who had moved to the North West to work in the new mines. Their regionalist political claims responded to, and utilised, their Kaonde, Lunda and Luvale co-workers' frustrations at the region's perceived underdevelopment. While their engagement with chiefs and North West Zambian politicians often encouraged forms of ethnic entrepreneurship (see Kapesa et al. 2015), collective action in the North West was motivated by both local and migrant miners' inability to fulfil their personhoods as unionised workers and as family members.
More obtusely, the president's speech implied a categorical difference between North Western 'tribal' patterns of redistribution and hierarchy and those of the union 'family'. In contrast, we conceptualise both tribe and union through concepts of kinship, where actions are motivated by mutual inter-dependence, within self-defining hierarchies and identity structures (see Hickel 2015& Lazar 2018. The key claim of this article is that Zambian understandings of unionism evolved and continue to adapt through similar political economic process to those that shaped discourses of tribe. Bemba and other Northern and Eastern Zambians have been disproportionately influential in creating a wakopala-person of the Copperbelt-identity. This personhood guides Zambian unions, ensuring that these unions are primarily equipped to assist with the aspirations, hierarchies and concepts of the good life that connect the Copperbelt to North and East Zambia. The norms, values and kinship structures of being a unionist are therefore more compatible with those identities than of being a Kaonde, Lunda and Luvale. This, rather than tribalism, limits North Westerners' union participation.
This article writes into debates about the relationship between trade unions and the communities they operate within (Collins 2012;Lazar andSanchez 2019 &Mollona 2009). Unions' struggles to organise workers across identity cleavages have long been a concern of both academics and practitioners and in many African nations, these struggles manifest through accusations of tribalism (Chari 2004& Werbner 2018. Where some labour theorists, including those who write on Zambia, argue that unions should attempt to generate a detribalising class consciousness (Parpart 1987& Werbner 2018, we show that being a tribe member and unionists are not mutually exclusive forms of kinship. Both tribe and union inspire collective action and shape Zambians' obligations to their peers and community. Further, tribal identity is crucial to shaping the contours of union identity, and vice versa. In making these claims, we foreground that intra-national narratives of tribe and unionism are co-created through (and influence) patterns of capitalist exploitation, local cultural norms and political histories. This also allows us problematise community unionism's focus on alliances with CSO and NGOs (see Mollona 2009& Moody 1997. We explore how connections to community are entwined with locally constructed kinship norms and hierarchies and demonstrate that unionist political action is enabled or diminished through these kinships, as much as through class consciousness or CSO union alliances. Our paper combines insights from studies of the political economy of tribe in North West Zambia and of workplace-based experiences of Zambian unionism. Author 1 is a Zambianbased political scientist. Between 2015 and 2018, he conducted fifty interviews with North Western stakeholders, exploring how national political structures and international mining capital positioned union branch executives for ethnic entrepreneurship and how government interference has helped to naturalise Bemba and Easterner (Tumbuka and Nyanja) hegemony within the mining unions. Author 2 is a Europe-based anthropologist. Between 2017 and 2019, he spent 12 months on Zambia's Copperbelt and in its North Western province, using ethnography to explore how Zambia's mining unions reflect narratives of class and social justice tied to Copperbelt personhoods. The article was born out of our ongoing debates about the relationship between tribe as a political category and emic experience. It is structured through three claims found across both of our work. Zambians experience idealised ways of being and kinship obligations through notions of tribe and the obligations of being a unionist are often closely entwined with those of identifying as Bemba: tribal differences motivate labour militancy in North West Zambia, both through ethnic entrepreneurship and through workers' attempts to enact concepts social justice embedded in their differing kin obligations: calls for the unions to perform CSR in the North Western province reflect the unions' entwinement with Copperbelt aspirations, economies and networks, rather than signifying a politically engaged community unionism.

Tribe and union as kinship structures
The everyday politics of labour exploitation is deeply local and situated (Prentice 2012). Unions have therefore long struggled to represent the interests and aspirations of members, when these workers hold substantively disparate identities outside the workplace, including racial and, in several African nations, tribal inequities (Werbner 2018). Labour organisation has often been tied to specific ethnic identities. Brodkin Sacks (1988) describes how the black, female leadership of a hospital workers' union was contested by management, who favoured white workers to destroy solidarity (for a current parallel, see Durrenberger and Erem 2005:87-100). On the Panama Costa Rica Border, Bourgois (1989) recounted how racially categorised aspirations, patterns of migration and relationships to capital prevented pan-ethnic unionisation. Prentice (2012) observes how the racial ordering of Trinidadian workplaces dampens unionisation and how this is tied to a national order where the advancement of differing ethnic groups is perceived to be mutually exclusive. In contrast, Chari (2004) explores how Gounder management discourages the unionisation of its workforce through narratives linking ethnic solidarity between workers and employers to idealised masculinities and personal social advancement.
Responding to long patterns of racial segregation and tribal classification, several ethnographies of southern Africa explore how and whether trade unions can create a detribalising, nationalist working class (Hickel 2015& Parpart 1987. Werbner (2018) argues that unions have been able to affect real change in Africa only when they look beyond tribal interests and deploy a national political language. In a critique of this approach, Hickel (2015) argues that unions associated with the African National Conference failed to appeal to Zulu migrants, because of the migrants' aspirations for post-apartheid South Africa to incorporate values and hierarchies incompatible with urban unionists' presumptions of an anti-traditional 'modern' state. Union activists conceptualised themselves creating a new society, where individuals would be unfettered not only by apartheid but also by tribal and traditional hierarchies, while many Zulu workers desired to use collective action to restore the dignity of Zulu traditional life. Further problematizing cross-ethnic solidarity, Tanikella (2003) describes how attempts to celebrate ethnic differences can obfuscate the power differentials and structural inequalities that disillusion non-ethnic majority members of social movements. Unions (and to some extent, labour academics) that presume a powerful working-class identity, tied to a homogenous class-defined structural position, have therefore struggled with the complexity of identities based upon diversifying workplace experiences (Lazar and Sanchez 2019).
Understanding how intersecting identities shape labour regimes has been key to recent texts that explore political economies through sociality (Bear et al. 2015). Lazar (2018) argues for understanding unions as a kinship structure, where the intimacies of the workplace embody practices of kinning and support a 'mutuality of being' (Sahlins 2011). Lazar (2018) claims that unionism is motivated by commensality and sociability, rather than purely rational economic interests. Collective action (strikes, protests and the like) therefore represents a flow from kinship into politics, rather than vice versa, with political action inspired by personal relationships and obligations. Lazar's (2018) analysis fits nicely with Larmer's (2007) depiction of Copperbelt mineworkers' political consciousness, where miners consistently mobilised around perceptions of social justice embedded in their experiences and the demands of their kin, rather than the nationalist project of government or union leaders.
We find much inspiration in the insights of Lazar (2018) and Larmer (2007), while observing that almost all Zambians interact with non-union kinship obligations and aspirations, which they conceptualise through (emic) concepts of tribe. We explore how unionism has long been embedded within and evolved in conjunction with Bemba and Easterner 'mutualities of being', communal demands and visions of social justice-which are slightly different from those often held by the Kaonde, Lunda and Luvale. We also explore how union members' visions of social justice are convolved with political patronage, enabling ethnic entrepreneurship (Kapesa et al. 2015). Conceptualising unionism as a kin-based identity, and as one that interacts with other kinship structures, spurs insights into the varying relationships between unions and the communities they work within. Studies advocating community unionism and social movement unionism bemoan workers' failure to see beyond the workplace, and there is a tendency to perceive alliances with CSO as the premier evidence that a union is engaged in the politics of its encompassing community (Von Holdt 2002& Collins 2012. In response, we show how MUZ is embedded in the Copperbelt through economic structures, informal social networks and concepts of personhood and note that this is not the case in the North Western Province. In exploring how trade unions shape and respond to kinships embedded in hierarchies, we argue against seeing formal relationships between unions and CSOs the most significant evidence of union-community engagement. We demonstrate that demands for union-CSO alliances can, on occasion, represent a failure of the union to engage with the intra-community solidarities that generate kin-driven political action. Zambian tribes are imprecise categories, shaped through colonialism, the independence struggle and the modern political economy, yet almost every Zambian experiences themselves as having at least one tribal identity (Van Binsbergen 1985& Simpson 2009). Marten and Kula (2008) argue that Zambia's seventy-two official tribes reflect a codification of pre-colonial linguistic categories and migration patterns. This codification enabled chiefly indirect rule and encouraged temporary migration, through taxation and prohibitions on permanent urban settlement (Pesa 2019). Tribe pushed the cost of the social reproduction of copper extraction onto rural Zambians, while empowering those who guided the concept and resource claims made through it-the precursors of modern ethnic entrepreneurship (van Binsbergen 1994). Missionaries, journalists and anthropologists produced ethnohistories that legitimised tribe as a governance structure and as a component of Zambian identity (Macola 2003). As we show below, struggles against colonialism and then for democracy cemented tribe as an aspect of political and social personhood.
While literature and our own findings stress that tribal norms are constantly shifting, they influence the identity bricolage through which a Zambian shapes their expectations of kin and their concepts of the good life. Tribe has historically manifested through ritual or taboo. Richards (1956), Colson (1958) and Spring (1976) describe differences in the level of emphasis placed on sex, seclusion and domesticity in initiation ceremonies of Tonga, Bemba and Luvale girls. Colson's (2004) study of Tonga concepts of the Christian God finds a slight variation on Hinfelaar (1994) study of the Bemba understanding of the same. More significantly for this article, differing tribes explicate differing ideals surrounding relationships inside the household and between households and the community, as well as acceptable levels of inequality and sources of hierarchy. For instance, the Tonga are reported to be comparatively egalitarian and the Lozi deeply hierarchical (Taylor 2006). Tribal attributes and identities are never fully encompassing-a Zambian not only belongs to their tribe but also to their school, church, workplace and nation-and Zambians can belong to multiple tribes and move between them (Kapesa et al. 2015).
Our informants similarly described tribe through ritual and differences in demands from their kin and visions of the good life. While many reported fluidity or plurality in their tribal identities, this identity was almost always deeply held. Zambians continuously participated in the creation and negotiation of tribes, often through jest. A Facebook photo of an ugly man, beautiful woman or elaborate ritual would be captioned 'Which Tribe [is this]?' and miners would tag their friends of different tribes. When reading the online serial 'Love vs Tribe', union members described which tribes they would not want their family members marrying into. More significantly, a Tumbuka (Eastern province) unionist explained that Tumbuka masculinities were like those of the Bemba, but 'with more worries'. He claimed he faced additional burdens ensuring large bride prices for his paternal kin and that despite being happily married in Kitwe, he hoped to obtain a second wife in is fathers' village, to signify his success. Demonstrating this identity's importance, after bemoaning his responsibilities, this urban, educated unionist declared that he would 'rather cut of my own dick!' than fail in his tribal role. Other respondents explained that they had multiple tribal identities and responsibilities. A miner at Barrick Lumwana told us 'Mum is Kaonde, dad is Lunda, so I have two tribes…two villages'. A leader of Kansanshi's NUMAW branch explained: My mother is from Angola and my father is from Eastern province, but I live here …I have a farm… culturally I am Kaonde To be 'culturally Kaonde' referred to the kinship obligations and aspirations he experienced in his village just outside Solwezi township, which were different from those he would have felt had he stayed in Anglo or Eastern Zambia. This identity and kinship were intimately tied to political economic structures, including the mines and the unions.

Being Kaonde, Bemba and a unionist
Respect for village headmen, intra-family dependence and strong sanctions against individual wealth accumulation have frequently characterised studies of Kaonde and Lunda tribal identities (see Crehan 1983& Pesa 2019. Texts often attribute these traits to North West Zambia's isolation and population scarcity. The North Western province has consistently had a population density between one-half and one-quarter that of the rest of the nation, and tsetse fly discouraged any ongoing settler presence (Chileshe 2005& Crehan 1997). This created a situation where, until very recently, chiefs and headmen were powerful in their ability to interact with the state, but not to impose upon the lives of their constituents (Negi 2010). Headmen and more senior chiefs were codified through Rhodesia's colonisation, turning Kaonde-speaking clans into the Kaonde tribe (Chabatama 1999). However, the North Western Province's isolation allowed these headmen to impede forced Copperbelt migration and to facilitate paid labour for their constituents (Negi 2013). In contrast, inside their jurisdictions, chiefs had little power and were incentivised to be generous. Conflict and ambition regularly combined with land abundance to encourage the fissuring of villages. Smaller communities would emerge, creating new, non-hereditary headmen (Pesa 2019). Becoming a headman is a common Kaonde aspiration to this day (Jaeger 2015).
Geopolitical traits of the North West impeded the accumulation of individual wealth. While many Kaonde migrated to the Copperbelt in the Colonial era, Mitchell (1952) observed them being vastly less likely to aspire to stay there than Bemba or Easterners. (They were also more likely to hold low status mining jobs and therefore less prone to unionisation-see Bates 1972.) When miners, or other wage migrants, returned to the North West, most recommenced farming, rather than creating a merchant class. Monetization was uneven and partial until as late as the 1980s and land abundance discouraged the paid farm work required to maintain commercial farms (Crehan 1983). Today, government subsidized fertiliser for small holders similarly discourages farm consolidation (Jaeger 2015). Female ownership of fields, which is no longer common in North and East Zambia, has assisted Kaonde women in leaving unhappy marriages and Kaonde households have long been units of consumption and resource sharing (Crehan 1997). In describing modern life in the Lunda settlement of Mwinilunga, Pesa (2019:332) states: Long standing beliefs concerning the authority of chieftainship, prestige and the value of village life have remained significant, even as incorporate new elements and meanings.
When describing the moral authority of chiefs, sanctions against those considered to be greedy and the lack of emphasis on personal commodities, her work echoes Crehan's (1983Crehan's ( & 1997 earlier texts. Pesa's (2019) account also echoes the experiences of our North Western informants. Kaonde and Lunda received their jobs through their chiefs and village elders. While the North Western province is changing rapidly, many employees had only worked sporadically as labourers before mine employment and those who worked at Lumwana and Kalumbila mine frequently still lived in villages near the mine site. One former shop steward informed us that he had left the union after deciding that constantly attempting (and failing) to defend members was incompatible with his desire to become a headman. In contrast, a branch vice chairman who was married to the daughter of a more senior chief saw the combination of these roles as a rapid route to success. Often the idealised narrative of non-hereditary headmanship and equitable village life was in contrast with (an equally imagined) Bemba identity. We were told that wage negotiators focused on incremental pay-rises for all members, rather than increasing the lowest salaries (inevitably held by unskilled locals) because the Bemba were greedy and that branches in the North West had unstable leadership because 'the Bemba refuse to be ruled by anyone but themselves' and would therefore undermine locally born branch executives.
Bemba and Easterner kinship has historically been presented as evolving through proximity to and engagement with the Copperbelt, land scarcity and individuated patriarchal households. CiBemba-speaking Northern Zambians were forcefully relocated to the Copperbelt in the 1920s, creating ties between the regions that persist to this day (Money 2019). Richards' (1939) early work describes a matrilineal Bemba, with hereditary chiefs enforcing land laws and women controlling much of their household's resources. However, villagers consistently migrated to the Copperbelt for employment and to attend fee-charging private schools. This, combined with the shift from slash-and-burn to compost dependent agriculture, increased the importance of cash and of (primarily male) waged labour (Stromgaard 1985). By the 1940s, 60% of the Northern province's taxable population had migrated for waged labour, primarily to the Copperbelt, a portion that had likely increased by the 1980s (see Chileshe 2005). Moore and Vaughn (1994) describe land scarcity and labour migration as shifting the balance of power in Bemba areas from women to men and from chieftaincies to family patriarchs. In Luapula province, Poewe (1978) claimed that husbands and wives kept separate finances, including remittances from their consanguineous Copperbelt kin. Poewe (1979:80) also observed increasingly individualistic production, personal accumulation for reinvestment and respondents stating that their region was increasingly 'Not rural… It is the Copperbelt here now'. Many Zambians believe that even rural Bemba have lost touch with tradition (Taylor 2006). However, our informants described Bemba kinship as incorporating a scepticism of chieftaincy, an expectation of consanguineous remittances and an aspiration for personal wealth within patriarchal household units. The ability to assist with these aspirations has been key to being wakolapa and a unionist, with the unions therefore shaping the personhoods, both possible and logical, in the North and Eastern provinces, while demands from these provinces shaped unionism.
Copperbelt residents frequently describe themselves as wakopala, an identity that combines what its users see as 'modern' and 'African' traits, with a pan-Zambian narrative that nominally downplays the importance of tribe. Miners have consistently used militancy to enable the mutuality of being and fulfilment of personal ambitions of wakopala kinship and their unions have evolved through attempts to reconcile this imperative with nationalist political concerns (Larmer 2007). The prominence of ciBemba speakers on the Copperbelt has ensured that ciBemba is the lingua franca of the wakopala and the unions. Further, miners' early resistance to colonial exploitation was coordinated through the Bemba Mbeni Dance Society (Henderson 1970). Colonialists' response to this, and other labour disruptions, was to depict unionism as incompatible with 'traditional' life, creating tribal councils to oversee workers' complaints (Epstien 1958). Over time, the unions wrestled administrative and dispute resolution functions from these councils (Rotberg 1965). Workers voted to dissolve them and the union president was referred to as 'the Paramount Chief of the Bemba Trade Union' (Epstien 1958). While Rotberg (2002) claims that non-Bemba miners saw the union as 'Bemba serving' and dominated by Northern Province elites, many scholars perceived an anti-tribal class consciousness and an attempt at creating a national Zambian identity (Money 2019& Parpart 1987. Bates (1972:291) argued that on the Copperbelt: [Tribal] considerations govern many social patterns; hospitality, joking relationships and domestic and kinship affairs, but the structuring power of ethnicity is not allowed to carry over into the field of labour relations.
Even if the union's politics have rarely been overtly tribal, tribe continued to guide the kinship and domestic affairs that motivated day-to-day unionism and political action. For example, the wakopala are increasingly born on the Copperbelt (78% in 2000). They speak Bemba and practice initiation rites and bride payments drawn from Bemba norms (Potts 2005& Mususa 2014). Union leaders are commonly the guests of honour at life-stage ceremonies, entwining the union, wakopala and Bemba norms. More importantly, Larmer's (2007) study of the Mineworkers Union of Zambia describes the circulation and inter-dependence of urban and rural populations, with union wage demands reflecting these kin obligations. However, kin entitlements differ slightly between tribes and (as we will show) union recruitment and negotiating discourses invoked (and shaped) the responsibilities of Bemba kinship. A key wakopala aspiration is ensuring oneself and one's dependent kin are 'moving'; obtaining commodities that signify a successful household (Haynes 2017); placing children in feepaying private school; and financially supporting rural relatives. This concept of moving is vastly more coherent with Moore and Vaugh's (1994) account of individuated wealth among the Bemba, than Pesa's (2019) depiction of Lunda ideals (or earlier Crehan's (1997) study of the Kaonde).
Our respondents certainly saw a coherence between being Bemba, wakopala and a unionist. 'They [non-Bemba] just use ours [Bemba]' we were consistently and dismissively told in response to questions about non-Bemba rituals, initiations and bride-prices on the Copperbelt. When berating his shop stewards for not wearing full suits to a union meeting in a local scout hall, the chairman of MUZ's Chingolo branch (who was Tumbuka by name and ancestry) appealed to their pride, exclaiming; Gentlemen, we are Bembas, we are Africans! In the same meeting, while debating what wage increase to ask of management, one worker claimed: We miners, we are supposed to be the middle class, it is us who must send money to the villages, and place children in the universities.
Wakopala unionists often compared this classed (and covertly tribed and gendered) identity to the overt tribalism that they believe characterised North Western life. A branch secretary at Kalumbila explained: These people feel their employment is a gift, they cannot speak to a white man about wages, because it is biting the finger that feeds them. They were hired by their chiefs, who just said, 'you come here, you have this job' so they feel any money is a gift. But a man who is skilled, who has trained and worked before, who knows the value of his labour and knows that this value must be fought for, that is the kind of worker that has been brought here from the Copperbelt.
This urban unionist/chiefly subject dichotomy commonly appears in popular understandings of Zambian politics and academic texts on chieftaincy and tribe (see Kapesa et al. 2015& Negi 2010. We now show how tribal and unionist kinship obligations influence Zambia's political system. However, this is not the regressive tribalism that wakopala often claim, but rather is a combination of obligations, entitlements and solidarities, interacting with geographic and economic political claims and opportunities for entrepreneurship.

Tribe and the politics of mining in North Western Zambia
Accusations of tribalism between North Westerners and Copperbelt unionists took place within the context of sometimes violent ethnic tension in the North Western Province. Members of Parliament (MPs), chiefs and local non-governmental organisations (NGOs) demanded mine jobs be given to 'locals' and that the national mining tax revenues be spent in the North Western Province (Kapesa et al. 2015). Historically, North West Zambians have often perceived themselves to be excluded from Zambia's ruling coalitions (Pesa 2014), while Copperbelt unionists have had on occasion combative and at other times symbiotic relationships with national governments (Fraser 2017& Larmer 2007. Intra-union politics reflected these factors: firstly, through North Western political leaders presenting the union as a foreign interloper and then as branch chairmen depicted a disjuncture between 'locals' and 'migrants' in attempts to enhance their own political status. However, these factors did not fully explain workers' descriptions of their decisions to mobilise, which were embedded in their kinship obligations, both as a unionist and a (tribed) family member. This section explores how formal union roles enabled ethnic entrepreneurship and 'tribal' political categories. We agree with Negi (2013) that the legacy of tribe in North West Zambia facilitates national and international elites' resource extraction. However, we critique Negi (2013) for reducing tribal identities to easily manipulated political categories. Inspired by Larmer (2007) and Lazar (2018), we show how union action, including militancy, is linked to workers' attempts to fulfil their kinship obligations and to enable what they perceive to be social justice-perceptions influenced by their experience of tribe.
Tribe has long influenced and been legitimised through Zambia's post-colonial politics. During the anti-colonial struggle, Kaonde chiefs aligned with the African National Congress, who linked independence to 'African values'-(chieftaincy based governance) and to rural development (Pesa 2014). Copperbelt militancy, primarily in support of the United National Independence Party (UNIP), was crucial to decolonisation (Larmer 2007). Zambia's first president, Kenneth Kaunda, saw urban, Copperbelt-based development as key to the national project and he was widely (wrongly) presumed to be a Bemba (Kapesa 2019). The linkages and remittances that connected Copperbelt residents to Bemba communities in North Zambia (and Nyanja/Tumbuka communities in East Zambia) encouraged Bemba and Easterners to support Kaunda's UNIP, which many North Westerners saw as neglecting the rural economy (Fraser 2017). To this day, tribal identity and regional resource claims influence electoral support. In the 2016 election, almost all seats in Southern, North Western and Western Zambia were won by the opposition United Party for National Development (UPND), while all Copperbelt, Eastern and Northern seats were won by the incumbent Patriotic Front (PF) (Fraser 2017).
Most ethnic entrepreneurship takes place at the subnational level. Political operatives overtly link their position to their ethnic identity and directly tie tribal political support to government largess (Kapesa 2019). These entrepreneurs dominate many of the offices in which Zambians interact with the state and they control the community development funds that build bridges, roads and hospitals (Kapesa 2019). Union roles have been similarly invoked in Zambia's patronage-based politics. Before democratisation, the MUZ president and general secretary were given seats on the national mining company's board, in an attempt to placate Copperbelt miners (Larmer 2007). In 2008, the MMD allegedly financed the creation of the United Mineworkers Union of Zambia (UMUZ) in response to their perception that MUZ and NUMAW supported the PF (Uzar 2017). Linking this explicitly to tribe, we witnessed government employees pressure union leadership to 'resist tribal influences' by impeding the elevation of a Tonga unionist, assumed (by virtue of his tribe) to support the UPND. The most popular MUZ branch chairman in the North West, a Kaonde named Otis, was threatened with suspension if he ran in MUZ's quad-annual national election.
The long-term relationships between tribe and political power shaped opportunities for ethnic entrepreneurship when significant mining capital entered the regions. In Lumwana and Kalumbila, chiefs manoeuvred to ensure that the mines employed their 'local' subjects (Negi 2010). In 2008, the Kaonde chiefs surrounding Lumwana mine persuaded the company to restrict the recruitment of unskilled and semi-skilled labourers to 'the community'. In 2010, the Lunda Traditional Authority in Kalumbila allegedly exchanged a land concession to FQM for a guarantee that 'locals' would be employed for unskilled labour. 2 This 'local' was facilitated by the chiefs. Applicants required a National Registration Card (NRC) confirming they were from Solwezi and were interviewed in Kaonde for work at Lumwana and in Lunda at Kalumbila. This process depicted 'the local' as a tribal group, determined by linguistic competence and guided by the chieftaincy. Evidencing this, despite being merely 40 km away from Lumwana mine (much closer than Solwezi town), Senior Chief Musele's Lunda were initally excluded from mine employment, while semi-urban, Solwezi-based Kaonde received jobs.
In all three mine companies, HR and chiefs worked together to discourage unionisation. Union leaders claim that FQM paid mine workers to de-unionise, denied entry when unionists attempted to deliver legal documents of union incorporation and fired workers who attempted to start union branches. The construction of Lumwana was declared an 'economically essential activity', initially preventing mining unions from organising. The mine's recognition agreement with MUZ reads, 'The union is not the owner of people in the enterprise, it is an intruder'. One chief owned a labour hire company and his son explained that the union would simply siphon money to Kitwe, which should instead stay with the Royal Family, which all workers (about four hundred) were part of.
All three mines, and some subcontracting companies, eventually obtained union branches. As unionising co-workers in the North West could cost a person their employment, it disproportionately appealed to ambitious individuals with political connections. At FQM mines, these unionists became aligned with opposition politicians and at Lumwana, they attempted to parley their positions into government roles. At FQM Kansanshi, both the 2007 and 2008 negotiations led to illegal strikes, which were condemned by the government and by MUZ Head Office. However, Richard, Kansanshi's MUZ branch chairman, claimed that the union had neglected Kansanshi's workers by 'staying in Kitwe' when members were facing disciplinary charges. This had left miners afraid and enabled the company to offer a miniscule salary. Believing he would be suspended, Richard quit MUZ. He unsuccessfully attempted to become NUMAW's branch chairman. He then started his own union, the Consolidated Union of Mine and Allied Workers of Zambia (CUMAWZ) that would focus specifically on North Westerners. MUZ and NUMAW representatives alleged that Richard had been offered a role by the opposition party, then the PF, to incite the strike. Richard admitted this occurred, but claimed to have been acting on his conviction when he encouraged workers to rise against FQM and MUZ head office. Richard worked full-time as CUMAWZ President until 2014. He claims that, as he is 'famous for fighting for workers', he will be preselected for an opposition, now UPND, MP position in 2021.
In 2012, Gitemwa, the first MUZ chairman at Kalumbila mine, was similarly fired after accusations that he incited a strike to assist the opposition UPND. Gitemwa was a Bemba migrant from a wealthy family and had been a PF branch secretary at Copperbelt University. He became disillusioned with the party in 2011 and joined the UPND. When a miner died in a traffic accident, Gitemwa drove the corpse to Solwezi in his car and a strike occurred. Union representatives stated instead that Gitemwa had started the strike himself and that he used 'tribalist' discourses to mobilise the local Lunda. When the provincial minister arrived to mediate the dispute, Gitemwa publicly accused him and the PF of neglecting the region. He was then fired and quickly joined the UPND as an organiser. However, possibly responding to the PF Regional Secretary's dictation that no contracting work could go to 'anyone who is not PF', Gitemwa returned to the ruling party. He recently obtained international funds to train local civil society organisations (CSOs) on labour disputes.
Leading up to the 2016 Lumwana branch elections, branch executives worked with PF aligned businessmen to sell non-essential commodities (TVs and mobile phones) to miners. Many members went without income for several months, as their salaries were deducted to pay for the commodities they had bought on credit. The branch executives used their profit to campaign unsuccessfully to become PF candidates in the 2016 election. They then negotiated a 0% annual pay-rise and a year later a go-slow occurred during negotiations. This led to the firing of two popular shop stewards and a strike. When these branch executives were replaced at the next branch election, MUZ's head office came to Lumwana to give training that warned of the dangers of voting on 'tribal' lines.
In each strike, political pressure encouraged understanding all industrial action in the North West through a tribalist political lens. This enabled rapid promotion of unionists who engaged in ethnic entrepreneurship, which in turn solidified tribal discourse. We witnessed government and opposition representatives utilise flattery, bribes and threats to obtain branch official's assistance in either encouraging or discouraging strikes in their constituency. This pressure was often fierce, with employees of the governments' Presidents' office (the repressive secret police) arriving unannounced at unionists' homes. A recently promoted branch executive in Kalumbila thus attributed the removal of the previous three Branch Chairmen: There is a lot of political interference in the North Western Province, because it really is an opposition stronghold. So if you want to move up in the union it can be very easy to just tell the government [that your superior is encouraging tribal conflict] and they will fire that chairman.
Significant ethnic entrepreneurship occurred within the North Western mining workforce, yet miners did not risk termination and occasionally violence solely on the instruction of 'tribal' leaders.
Miners recalled their engagement in the strikes described above through violations of their kinship commitments to each other as miners and through their obligations to their families. Those involved in the Kalumbila strike foregrounded the indignity they experienced as miners and how this was linked to un-kept promises to the encompassing community. A North Western permanent resident explained that villagers were furious that the mine had not yet provided a hospital to the area, which they believed had been promised in exchange for their relocation and resource use; One fateful day a man was driving and he turned over and died… when the body was taken to Solwezi it was taken in an open van… not in a hearse or ambulance. People said 'No, why do we have to be treated like this, we are being treated like animals!' … Over three days [of striking] other things came up…issues of wages…issue of abuse.
A Copperbelt migrant electrician explained that he had participated in the protests, as had almost all his peers. They had been incensed by reductions to shift length from twelve to eight hours, with correspondingly longer working weeks and reduced penalty rates. These made it more difficult for them to maintain their Copperbelt-based social and familial lives and to send remittances to kin.
The strike at Lumwana was described by those who participated as caused by ongoing violations of unionist obligations by the branch executives and deliberate attempts by management and senior unionists to engineer tribal division. Workers perceived that the branch was too close to the management. They had received no pay-rise the year before, yet the union branch leaders had been invited to celebrate with HR in Lusaka (Zambia's capital). When the next year's wage negotiations were stalled for several months, members had no faith in the union branch leadership to negotiate a sufficient pay-rise. The shop steward who was fired after the slow-down explained: That led to a misunderstanding between management and the workforce where production went down… the [Kaonde dominated] mining guys were able to produce twentyfive thousand tons a day, but they started producing seventeen thousand, somewhere there.
At the meeting that management called to resolve the go-slow, this shop steward (who was Bemba and from the maintenance department) suggested reconvening with HR and was charging with inciting an illegal go slow. This had led to a strike that was again started in the mining section. He claimed: I was charged because I was in maintenance [which is primarily Copperbelt migrants], and mining is very strong so they charged me to keep maintenance quite … because they are sure that if the whole mine is brave they will be pointing out everything that is wrong.
Seemingly evidencing his claim, when the MUZ Head Office staff visited Barrick Lumwana to train the (almost entirely mining division and non-Bemba) branch executive, this executive and their shop stewards spent significant time demanding that the Head Office pay this miners' legal fees.
Ethnic entrepreneurship in the North Western mining unions rarefied tribalism as a political tool and enabled the ongoing exploitation of Zambia's resources by its former colonisers and political class (Negi 2013). However, the tendency among Zambian political commentators, and some academics, to understand industrial action and social protest through tribalism risks diminishing the legitimacy of miners' complaints. It also discourages seeing union action as a way to bring social change though simultaneous connections from the unions to many different communities (compared with a homogenised, detribalised working class). However, as the strikes at Kalumbila and Lumwana demonstrated, industrial action was inspired by workers' inability to fulfil their obligations to the myriad communities they came from and their belief that the union was not reflecting their expectations of kinship (see Hickel 2015& Lazar 2018. Understanding the norms, hierarchies and structures in which unions are already embedded and through which they are connected to communities is therefore crucial for exploring their ability to challenge political norms, which is often seen as primarily the purview of community unionism.

Unions, tribe and community engagement
Union solidarity did not mean that all workers' identities and aspirations were equally integrated into each union. Rather, the unions were shaped by, and better incorporated, concepts of personhood commonly expressed on the Copperbelt than those found in the North West. These Copperbelt values and hierarchies, which have evolved in conjunction with Bemba kinship structures and centred around a male head semi-nuclear family, were rendered invisible through the union's narrative of class consciousness. Further, through its presence in Copperbelt households and its ongoing engagement with Copperbelt social and political structures, the Mineworkers Union of Zambia was deeply embedded in the Copperbelt community. This is not the case in the North West, where the same norms and personal connections do not equally intertwine the unions and communities.
In observing how MUZ was embedded within and responded to Copperbelt identities, this section contributes to the literature on the relationship between unions and communities. In response to the increased financialization of daily life and growing employment precarity, many works call for workers to engage in political and social action outside the workplace. Models of connection between unions and their encompassing communities include social movement unionism and community unionism (Collins 2012). In both, trade unions act in concert with non-workplace community-based organisations and non-governmental organisations, utilising their resources for political change and community development (Von Holdt 2002). These tactics have rightly been praised for re-politicising the links between local labour conditions and intra-national capitalist regimes (Collins 2007). Mollona (2009) goes as far to describe community unionism and business unionism-based on the provision of services to members-as 'two sides of the same phoenix' due to the inevitability that a business union will engage with and draw support from the community in which it is based. Often however, calls for community unionism downplay existing links to a community and the apolitical nature of much NGO and CSO work (Collins 2012). Where unions have evolved in a specific location, like a company town, their members and representatives likely have deep ties to that community. More controversially, we show that unions' kinship norms and hierarchical structures can be intimately tied to those of these founding communities. Mollona (2009:680) claims that: One of the problems of implementing community unionism is that it requires a radical shift in workers' perspectives. They must move away from …their everyday workplace experiences and see their situation from the perspective of others.
Copperbelt-based unions did not need to shift anyone's perspective to engage with their community. Further, when North Western union members called for CSR projects and CSO engagement-reminiscent of community unionism-their claims reflected MUZ's failure to incorporate the kinship experiences that would motivate their collective action.
Despite little engagement with CSOs and a financial structure that combined membership fees with selling services at a profit, MUZ was deeply embedded in Copperbelt communities. Daily experiences in the workplace and social spaces ensured that fighting for the community and engaging in local politics did not require a radical shift in the perspectives of trade union members or leaders. Almost all Copperbelt union representatives, from branch chairmen to MUZ's general secretary, held communal leadership positions, commonly scout leaders, pastors or rotary members. During elections, they would stress their ties to the community by mentioning these and declaring if they came from a prestigious local school (Nkana Trust or Kitwe Boys). Some chairmen ran for election to local councils and the unions would support them, regardless of which political party that joined. Union executives would be reminded daily of their responsibilities to Kitwe. When nonminers' children were turned away from Mopani's mine hospital during a fever outbreak, a Mopani branch chairman called management and demanded their admission. During negotiations, another chairman's church prayed that god guide him as he helped their community. Almost every unionist from the NUMAW president to lowliest branch executives would be constantly helping members with requests for cash. They would also assist ex-miners, those who were looking for work and the relatives of each, essentially serving as an emergency provider for large portions of Kitwe.
MUZ assisted the mines with training and ran businesses in a manner that reflected and maintained links between the mines and Copperbelt communities. Miner training would often be facilitated by the unions, presenting the mine HR and union leaders as parents of the workers through topics like 'financial management', 'hygiene', or 'marital harmony'. Linking this explicitly to communal life, Mopani's branch executives were tasked with creating a quiz for miners' wives. The quiz was to comprise questions about mine safety that women could ask their husbands, encouraging these husbands to be diligent at work. Women were expected to band together to obtain as many completed quizzes as possible, with the winning group receiving funds for a church party. Similarly, the unions ran stores that sold essential commodities on credit and worked with a microfinance bank to facilitate payday loans. Almost every miner had some form of debt to the union or co-signed with it. This debt, combined with the cash gifts described above, connected the union to a vast network of kin, dependents and dependents of dependents, in much the same manner as many Copperbelt men perceived themselves to be connected to kin and dependents in North and East Zambia.
The union assisted Copperbelt workers, while embodying a (primarily masculine) wakopala identity and was embedded in a community that had shaped and been shaped by this identity. During wage negotiations, the unionists provided management itemised 'cost of living' documents based on the presumption that a miner was 'moving' and 'a provider' (descriptions of successful Copperbelt masculinity). These documents included the cost of renting a house with multiple bedrooms, paying schools fees for their children (differentiated to include both direct kin and those in villages) and funeral fees to cover the burial of parents and those of their spouse. As wage increases that matched inflation grew rarer, the unions increasingly negotiated for specific allowances that reflected wakopala understandings of the good life. For example, union negotiators frequently demanded that management provide a 'school fees allowance'. This was a tri-annual cash payment given to each worker to coincide with the beginning of the school term. It was expressly framed as a moral obligation, saving members from the humiliation of neglectfully placing their children, or those of rural relatives, in government schools.
Like trainings (and the claims of the NUMAW president in the opening vignette), wage negotiations were embedded in a discourse that presented the union as the household head in a family of miners, reminiscent of wakopala familial structures. This discourse invoked class consciousness, while creating a hierarchy of more rational parents and less rational children. During wage negotiations, company HR and head office staff would stressed that the company was 'like a family' and that the negotiators were the household-heads. Together, they needed to make members 'understand' the need for low wages. During training in Kitwe, MUZ branch executives were divided into small groups and asked the purpose of the union. One group claimed 'To bring ideas together to fight the capitalists' and another 'to work with management and guide our members for mutual benefit'. These two answers were placed on the whiteboard, seemingly without contradiction.
Not only were the unions not equally embedded in North Western communities, but they neither properly assisted with many North Westerners' aspirations nor embodied those aspirations' success. This was most obvious when discussing the role of chiefs. As mentioned above, Copperbelt miners saw chiefs as an impediment to a detribalized class consciousness, yet many North Westerners hoped to obtain promotion within the chieftaincy and perceived it to link mines to their encompassing communities. Several North Western miners assured us that it was not solely a chief, but a council of elders, that determined who the mines would hire. This council ensured fair representation of villages, while monitoring for things like good behaviour and participation in village affairs. These men also saw the chieftaincy as providing a rare opportunity for skill development. A former farmer who had been trained in utilising heavy mining equipment explained: The people from North Western, they had no skills, so they made a dictate from the chief who would say 'the investor who will come and mine here in North Western should employ our Lunda people and train them.' In contrast, educational allowances and cash for remittances had less resonance to North Westerners, who had frequently not completed school themselves and whose children were either in free government schools or in the schools that FQM provided to resettled communities. Many Kaonde miners were farmers and they desired that the mine, union and their mining jobs assist peers in the adjacent villages. They therefore wanted provisions relating to the environment placed into collective workplace agreements and for the company to provide cheap inputs to local farmers.
Similarly, without the personal connections and structural history, MUZ was vastly less embedded in North Western province economies than in the Copperbelt. While it directly ran branch stores and organised loans in the Copperbelt, it used Copperbelt-based suppliers and contractors in the North West, making goods more expensive and transferring wealth out of the North West. As many North Western-based union executives were Copperbelt migrants who lived in workers' townships created by the mines, they (and their wives and children if they migrated with them) were unlikely to go to church or join social clubs with North Westerners, impeding the semi-formal, semi-social transfers of cash and assistance that linked unions to Copperbelt solidarity. This was a constant source of complaint for miners who would argue that their labour was 'developing' Kitwe through offering paid employment to Kitwe residents, buying four-wheel drives and building buildings. When MUZ sent its trainers from Kitwe to the NW to train shop stewards, they were told: Miners should feel like the union is contributing to the place where they are living, but now we have problems as we are just seeing money go to Katalungu house [in Kitwe].
When they asked how the union could improve, these trainers were provided with a list of the mine CSR projects, which were primarily providing cash to CSOs and NGOs for microfinance and agricultural programmes. They were encouraged to partner with these CSOs and to undertake similar 'development'.
Not only did calls that the unions engage in CSR reflect the failures of union negotiating strategies to resonate with their lives but also the union's dissimilarity with North Western kinship norms. The 'parental', 'reasonable' Copperbelt patriarch that the union embodied was not an aspiration figure for North Western miners, who believed the union should be asserting its authority and defending its people, like a local traditional ruler. In early 2018, the unions were negotiating severance packages for laid off subcontractors in Mopani's Kitwe mine. While the union leaders promised to be tough with management, they were managing workers' expectations by informing them that the current arrangement was unsustainable and that the union would work hand in hand with the company. The Lunda and Kaonde executives laughed with scorn and recounted how the chief nearest Lumwana ensured that the mine continued subcontractors' employment: He goes to the HR and he says 'you, you have chased away all the other contractors, but how can you run a big mine like this all by yourself'.
Rather than discussing with management as two equal, rational adults, the union should be asserting its authority and defending its people, like a local traditional ruler. Similarly, rather than only giving financial assistance to miners, as a Copperbelt miner would to their extended family, North Westerners desired that the union works to improve the whole community, in line with the community-based resource redistribution practices common in the North West and with what several other academics describe as community unionism.
This mindset was 'tribal', in so far as it reflected an idealised Kaonde familial and patronage structures rather than the kinship idealised among the Bemba. It was a call for community unionism in that it expressly requested that the union align with (apolitical) CSOs engaged in community development. Yet, it foregrounded that the disconnection that made community unionism appealing was not a shift in workers' perspectives, but a disjuncture between the values, hierarchies and mutualities of being of unionist and north western kinships. As Lazar (2018) observes, kinship can be crucial for motivating unionists and for the political action that connects unions to the community. However, inspired by Hickel (2015), we note that kinship, both when identified as tribe and as unionism, manifests through historically created norms, political economies and concepts of the good. Attempts to involve unions in political change by encouraging workers to focus less on their personal responsibilities (tribally inspired or otherwise) and more on alliances with civil society therefore have the potential to diminish unions' links to the kinships that provide the strength for collective service.

Conclusion
This article has explored how mining unions shape (and are shaped) by Zambians' understandings of tribe. Other literatures present unions (in Zambia and elsewhere) as detribalisation agents. However, we show that the current structures of Zambia's mining unions codify tribe 'from above', with the state interfering to ensure Copperbelt dominance of senior union positions and with more junior unionists incentivised to mobilise tribal grievances in the North West (cf Fraser 2017& Negi 2013. Unions also shape and incorporate tribe 'from below' exemplifying the wakopala semi-nuclear family household and enabling the remittances which shape Northern and Eastern Zambian tribal identities (Pesa 2019& Haynes 2017. Rather than a culturalist understanding of tribe, we have simultaneously demonstrated that unionist kinships are similarly created from above and below. Ambitious (almost necessarily Copperbelt-based) leaders define a union family that mirrors their own patriarchal structures, while workers use militancy and solidarity to fulfil kinship obligations including co-worker and family member, shaping unions in the process. This paper therefore contributes to the literature linking trade unions to ethnic mobilisation and community unionism. It has foregrounded the potential of unions as places for ethnic entrepreneurship, which advantages foreign capital and former colonial rulers. However, we caution against reducing tribal identities to political categories. We instead show how labour mobilisation incorporates entrepreneurs' machinations and workers' failure to achieve what they see as the good life. We have also explored how Zambia's mining unions mimic and enable the familial structures and aspirations found on the Copperbelt and among North and East Zambians. The union 'family' is shaped like a semi-nuclear Copperbelt patriarchal home; it argues for wages and education allowances that ensure the dignity of Copperbelt 'providers'; offers small cash gifts when this is impossible; and enables the remittances and tax revenues that can develop the Zambian national family. The fact that this semi-autonomous, remittance providing Copperbelt family was different to the chieftaincy-embedded families found in North West Zambia was key to intra-union accusations of tribalism. Calls for engagement with CSOs by North Westerners therefore did not signify a desire for a politically engaged community unionism, but were instead a desire for apolitical development projects. These evidenced a lack of the intertwining between union, community and daily kinship, which had inspired political mobilisation on the Copperbelt.
Ethno-tension between North Westerners and Copperbelt residents were 'tribal' in that they reflected differing aspirations, personhoods and concepts of the good life, which Zambians often conceptualise through tribe. In this understanding, almost all union work is to an extent tribal-with people bundling together to attempt to recreate the socio-economic conditions they perceive to be necessary for social justice and human flourishing-concepts that are deeply embedded in kinship and culture (Hickel 2015). This article therefore opens space to question which cultural values are rendered invisible in the understandings of unionism and class consciousness naturalised by other unions and how these values shape mobilisation, calls for community unionism and inter-union tensions. Concept of 'family' and class consciousness would be an obvious place to start such investigations. Many ethnographies note that unionists claim their union is a family, often with primarily positive resonances, and more work could unpack the intra-union and intra-communal norm obligations and hierarchies this creates. More broadly, work could consider how class consciousness interacts with hierarchies like 'tribe' that seem unjust to many senior unionists, but are crucial to the personhoods and social conditions that some of their workers attempt to obtain through collective action.

Compliance with ethical standards
This research has been approved through the European Horizon 2020 ethical research process for European Research Commission Grants and has been approved by the University of Liege's Human Research ethics process.

Conflict of interest
The authors declare that they have no conflicts of interest.
Research involving human participants and/or animals The research involved human participants (see below).
Informed consent Informed consent was obtained for each research participant. Depending upon the participants' level of literacy and availability of informed consent forms, this was either taken in writing or orally after the project was explained to the participant and they were given the opportunity to ask for more details or clarification.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.