Introduction
Lungisile Ntsebeza’s empirical investigations of rule and resistance in the South African countryside are not value-free. As an engaged scholar, Ntsebeza does not just examine what has happened and why; he also explores whether systems of rule and means of resistance met defensible evaluative standards and, if they did not, how things might in future be better arranged. In his most thought-provoking writings, Ntsebeza focuses on an evaluative matter which, though extremely important, is conceptually contested and complex: he seeks to determine how well rural South Africans have been politically represented.
Representation is a topic of central importance for political thought in the liberal tradition (see e.g. Mill 2015), including studies which, like Ntsebeza’s, are alive to the structural constraints, whether from racial capitalism or from neoliberal globalisation, on both rule and resistance in postcolonial African societies. In South Africa, for most of the twentieth century, the ‘rural areas in the reserves were supposed to provide an economic base for the continued reproduction of rural people, at the same time justifying a cheap labour policy based on migrancy’ (Ntsebeza 2005, 22). In such circumstances, the class solidarity needed to challenge an exploitative economic system did not arise of its own accord: quite the contrary, since South Africa’s rulers encouraged ‘ethnic’ and ‘racial’ divisions among workers. Besides, as Vivek Chibber has recently emphasised, even in the absence of ‘false consciousness’, structural features of capitalism generally ‘make it alluring to opt for an individualized resistance strategy rather than a collective one’ – unless they are counteracted by the ‘arduous process’ of ‘class organization’, ‘the cultural work that cements the class project’, the creation of new institutions and ‘new avenues for representation’ (Chibber 2022, 74, 159–160, 164, 175). What is more, following the advent of universal adult franchise and a unified parliamentary democracy in 1994, neoliberal globalisation has constrained the ability of central governments, in South Africa as elsewhere, to act as true representatives of their citizenry (Ntsebeza 2005, 21–22).
Ntsebeza’s work on chiefs and headmen not only seeks to describe how chiefs make use of their power and to explain why the institutions of chiefship and headmanship have been so resilient in South Africa, but also raises the ‘question of whether an inherently undemocratic, hereditary institution can exist in a South African democracy purportedly modelled on the liberal tradition of representative government’ (Ntsebeza 2004, 72). He concludes: ‘In a sense, chiefs can only be “good” if they reject the very basis of chiefly authority and embrace democratic forms of representation’ (Hendricks and Ntsebeza 1999, 118). Equally, his work on farmworkers’ recent campaigns for better working conditions in the Western Cape does not just outline the material challenges faced by these workers, but also seeks to identify the failings of ‘non-governmental organisations representing rural people’ which have meant the ‘“invisibility” of rural communities’ persists post-Apartheid (Ntsebeza 2013, 143, 136).
I argue here that, if the evaluative insights about political representation in Ntsebeza’s work are to be fully appreciated and are to form the basis of valuable future research, the normative scaffolding which supports them must in certain respects be overhauled. Specifically, the conception of democracy which Ntsebeza uses as his evaluative yardstick is in one respect too narrow and in another respect too indeterminate. It is too narrow because it leaves no space for the crucial work of political representation carried out by unelected non-governmental organisations (NGOs) and local organisations, which Ntsebeza’s own writings often highlight. It is too indeterminate because it contains an unredeemed pledge. Ntsebeza insists that two conceptions of democracy which are often seen as incompatible – representative democracy and participatory democracy – can in fact be integrated into one. He insists that ‘democracy should … be both participatory and representative, rather than one or the other’ (Ntsebeza 2004, 86). But he does not tell us how these reputed opposites are to be reconciled.1
Democratic government seeks to realise the ideal of popular sovereignty: that the whole people should exercise ultimate control over how they are ruled. In a large, complex, modern-day country, this ideal could not be realised except through political representation – political actors speaking and acting on others’ behalf (see Mill 2015, 235). Thus, the whole people must be adequately represented, both by those who legislate and take executive decisions, and by those who influence decisions and legislation through persuasion, negotiation, the articulation of needs and demands, and the formation of political coalitions. But beyond stating this ideal, we must also be alert to the various ways in which the representation provided by both elected and unelected representatives can fall short and degenerate into misrepresentation. It is, in general, constructive for political philosophy, in cooperation with empirical disciplines, to articulate and diagnose social pathologies (see Honneth 2009, ch. 2) as well as formulating positive ideals.2 Ntsebeza’s work, I shall argue, provides us with important resources for the former task. By bringing his empirical studies into conversation with recent developments in the theory of political representation, I shall show that his work helps us identify several pathologies of representation, which any system of political representation needs to guard against.
I am glad to have the opportunity to write about the subterranean philosophical seam in Ntsebeza’s work. Almost a decade ago, Lungisile Ntsebeza and I worked together closely on a seminar series, ‘Philosophy in Africa, Africa in Philosophy’ (PIAAIP), featuring weekly invited speakers throughout the 2016 academic year, which we envisaged as one part of a broader cooperation between the Philosophy Department and the Centre for African Studies at the University of Cape Town.3 A topic which often came up in discussion at PIAAIP was the relationship between normative political and social philosophy and the empirical social sciences. Prof. Ntsebeza himself raised the question whether normative philosophy should not have more to say about actually existing conditions than that they do not live up to an ideal. My contention here is that, in conjunction with empirical research, it can.
Representation and election
One might think that the criterion for whether a political actor is providing their constituency with adequate representation is straightforward – namely, whether they have been fairly elected as a representative and can be removed by their constituency in a fair election.
In his analytical writings on South African Traditional Authorities, Ntsebeza tends to endorse this view. In Democracy Compromised, a book highly critical of the post-Apartheid continuation of Traditional Authorities, he writes:
a fundamental contradiction exists in … attempts to accommodate a role for the institution of traditional leadership and its incumbents in a liberal democracy based on multi-party principles and representative government. … In so far as the institution is made up of hereditary leaders, the possibility of people choosing or electing their representatives is automatically eliminated. (Ntsebeza 2005, 23–24)
Later in the same book, he posits an even tighter conceptual link between representation and election: ‘the institution of chieftainship is, in so far as its incumbents are hereditary leaders, inherently undemocratic precisely because the leaders are unrepresentative. Its subjects are not given the chance to choose their leaders’ (Ntsebeza 2005, 259).
Ntsebeza consistently holds to the line that it is always at least a necessary condition of headmen, chiefs, paramount chiefs and monarchs’ being adequate political representatives of their constituencies that they be elected and subject to removal by a vote; a necessary condition they fail to meet. However, as I shall now argue, the election criterion is neither always a necessary nor always a sufficient condition for adequacy of political representation. Moreover, Ntsebeza’s own empirical studies supply us with several illustrations of why it is not.
Even if a political actor has been elected and is subject to removal by a vote, they may still not provide their constituency with adequate representation. This could be due to individual character (idleness or incompetence, say); but it could also be due to more structural factors.
A territorial constituency might be too large and diverse for a representative to become sufficiently aware of their constituents’ needs, interests and opinions. Ntsebeza notes that this was true of elected councillors’ rural constituencies in the years following South Africa’s democratic transition – whereas the local unelected Traditional Authorities continued to be both ‘accessible and visible’ (Ntsebeza 2005, 292). But an elected representative might also be dependent on or answerable to parties other than their constituency, to such an extent that they sacrificed the latter’s needs and demands in order better to fulfil those of the former. The preponderant electoral system in South Africa since the coming of democracy has arguably had this effect. ‘In the proportional representation system,’ notes Ntsebeza, ‘officials are often prone to be primarily accountable to their political parties, or the leadership thereof rather than their constituency, the voters’ (Ntsebeza 2005, 279). Ntsebeza has also suggested that in the post-Cold War era of globalisation, a national government may be prevented from representing the needs, interests and opinions of its citizens in central areas of policy, because it is in effect held hostage by ‘the Bretton Wood[s] institutions (World Bank, International [Monetary] Fund and the World Trade Organi[z]ation)’ (Ntsebeza 2005, 21; see also Ntsebeza 2013, 142–143).
A principal finding of Democracy Compromised is that under both colonialism and Apartheid, South African ‘traditional authorities were upwardly accountable to the government, rather than to the rural residents’ (Ntsebeza 2005, 220), because it was government which paid their salary, could overturn the decisions of their courts, and ‘reserved the power of deposing traditional authorities, if they proved uncooperative or failing in their functions’ (Ntsebeza 2005, 34).4 But, as is clear from the above examples taken from Ntsebeza’s own writings, this problem of upwards accountability can also afflict elected politicians and even elected governments, detracting from the quality of the representation they provide to their constituents.
It also seems to be possible for a political actor to provide a constituency with adequate representation, despite not having been elected in a formal, open, constituency-wide election. Over the last few decades, NGOs, lobbying and pressure groups, and other unelected bodies which claim to represent the needs, interests and opinions of a public have proliferated. Political theorists have now begun to pay attention to the potential of such bodies to make up some of the shortfall in political representation experienced by numerous societal groups. Political representation provided by unelected and ‘self-appointed representatives’ (Montanaro 2017, 211) has been dubbed ‘informal’ (Brito Vieira and Runciman 2008, 146) or ‘virtual representation’ (Rehfeld 2017, 59).
Informal representation, though undoubtedly open to abuse (Montanaro 2018, 79), can in the right circumstances be useful – even necessary. For example, in a country with a formal electoral system based on territorial constituencies, there are likely to be societal groups whose members are a minority in all, or almost all, meaning there will be little incentive for the elected delegates of the territorial constituencies to represent the needs, interests and opinions of these groups – especially when they differ from, or conflict with, those of other groups (see Mill 2015, 264–274). Suppose that a nationwide NGO or pressure group were formed, representing one of these marginalised groups, and succeeded through its campaigning and lobbying in persuading the national government to change certain laws to the advantage of this group. Despite lacking a formal electoral mandate, this organisation would arguably have enhanced the quality of political representation in the country. It would have done so by rendering present to political decision-makers a set of important needs and interests which otherwise would have been invisible due to the structure of the electoral system (see Montanaro 2018, 49).
Again, this point is reinforced by Ntsebeza’s empirical writings. A major focus of his historical and sociological studies is the role informal representative bodies have played in making up some of the shortfall of political representation in rural areas – in mitigating ‘the “invisibility” of rural communities’ (Ntsebeza 2013, 136). In some cases, the informal representatives have been national or regional bodies with a clear organisational structure and a written constitution setting out rules for decision-making. Political organisations such as the Transkei Organised Bodies (TOB) and the All-African Convention (AAC) campaigned on behalf of disenfranchised rural South Africans against the government’s ‘Betterment Scheme’ and ‘Rehabilitation Scheme’ in the 1940s (Ntsebeza 2005, 196–197). Following the end of Apartheid, NGOs such as the Trust for Community Outreach and Education (TCOE) and Alliance of Land and Agrarian Reform Movements (ALARM) have pressured government to implement land reform in a way which benefits poor rural dwellers and South Africans with insecure tenure in the countryside (Hall and Ntsebeza 2007, 14–17). In other cases, local bodies with more improvised decision-making methods have taken on a representative role: Emnxe Residents’ Association (EMRA), for instance, which conveyed local people’s demands for residential land plots to Traditional Authorities perceived as corrupt and inefficient in the early 1990s (Ntsebeza 2005, 245–249); or the Landless People’s Movement (LPM), whose formation in 2001 Ntsebeza characterises as ‘by far the most significant event in the resurgence of resistance in South Africa’s land sector at the dawn of the 21st century’ (Ntsebeza 2013, 138).
Ntsebeza makes it explicit that he believes there is a role for such unelected bodies to play alongside formally elected representatives in post-1994 South Africa. He states that, given the ‘limited capacity’ of elected councillors in Xhalanga, ‘NGOs would be a good compl[e]ment’ (Ntsebeza 2005, 293). Based on his analysis of rural resistance in the Western Cape in the 2010s, he advocates for rural dwellers’ continued political representation by both NGOs and more improvised local social movements, alongside formally elected representatives (Ntsebeza 2013, 154). In an opinion piece (Ntsebeza 2018), Ntsebeza writes: ‘Given the power of global capital and a state that acts in its interests, organised social movements become a necessity.’
If the election criterion on its own is insufficient, we must seek alternative criteria for adequacy of political representation. Or, approaching the issue from the opposite direction, if the election criterion is not enough to establish inadequacy of political representation, we must find alternative specifications of the ways in which political representation can go wrong. I argue in the next section that, despite his stated allegiance to the election criterion, Ntsebeza’s empirical writings are a most helpful resource for the latter approach – the detection and articulation of pathologies of representation.
Constituency, audience, authorisation
Democracy Compromised addresses two principal questions about South Africa’s Traditional Authorities: (a) what explains their survival and resilience as an institution? (b) are they a legitimate component of a democratic system of government?
Ntsebeza, as we saw in the previous section, thinks the answer to (b) depends on how well Traditional Authorities can be said to represent their constituencies. If the answer to (b) were ‘Yes’, it could form part of the answer to (a) as well: Traditional Authorities might have survived because they were legitimate and recognised as such. But according to Ntsebeza, the answer to (b) is ‘No’, and his answer to (a) is that Traditional Authorities have survived despite their illegitimacy because, thanks to government policies, they have maintained a high degree of control over land allocation in rural areas. Consequently, most rural dwellers had and have little choice but to acknowledge and cooperate with them: ‘whether rural residents supported Tribal Authorities or not, they had no option but to use these structures’ (Ntsebeza 2005, 253).
If the question whether Traditional Authorities provide adequate political representation boiled down to the question whether their constituency has formally elected them, answering Ntsebeza’s question (b) would require hardly any empirical investigation. The rich empirical detail of Democracy Compromised would, in that case, be justified only to the extent that it formed the basis for an answer to Ntsebeza’s question (a). In the previous section, I argued that being formally elected by a constituency is neither a necessary nor a sufficient condition for adequacy of political representation. I shall now explain why detailed empirical investigation sometimes is necessary to determine whether a political actor provides a constituency with adequate political representation. Much of the empirical detail of Democracy Compromised and other studies by Ntsebeza is directly relevant to this evaluative question, even if he himself does not present it as being so. In other cases, considering Ntsebeza’s writings in the light of recent philosophical work on representation suggests further avenues for research relevant to Ntsebeza’s evaluative question ((b)). It will help us work through these matters systematically if we begin by examining more closely what political representation is.
Some types of representative (e.g. a lawyer or a letting agent) act on a detailed mandate, or else with discretion but within clearly defined parameters laid down by the represented party in advance. But political representation can almost never conform to this simple principal-agent model, on which initiative lies with the represented party and the representative simply obeys instructions. For one thing, political representation typically involves speaking and acting for people on such a wide and unpredictable range of matters that it would be impractical for them to lay down instructions for every eventuality in advance. For another, political representation usually consists in representing a large group of people which is (unlike a corporation or political party) not already internally organised with decision-making rules designed to enable the group to act or speak as a unit. Indeed, it is often not until it has a political representative that a group such as a territorial constituency – or even certain minority groups – has any capacity to act or speak as a group (see Hamilton 2014, 141). Moreover, a straightforward principal-agent model has no place for an important service which good real-life political representatives provide to the group they represent: namely, they ‘interpret the group’s interest’, giving the people they represent ‘an image of themselves to reflect on’ (Brito Vieira and Runciman 2008, 101, 139). Being represented politically can be a means by which members of a represented group come to articulate their needs, interests and opinions, or even form new ones. A political representative ‘creates a new version of the people and their interests’, with which the group they represent may or may not identify (Brito Vieira and Runciman 2008, 139).
On the other hand, a model of political representation which placed all the initiative with the representative would also be flawed. Groups represented by political representatives typically comprise mature adults capable of, and entitled to, a say in how they are represented. Political representation should therefore involve some degree of control on the part of group members – mechanisms by which they may endorse or reject the representation their representative provides, but which do not require a prior capacity on the part of the group to act as a unitary principal (see Hamilton 2014, 142).
One might think that ongoing interaction between representative and represented could be dispensed with, if only the political representative were similar enough in all relevant respects to the group represented. However, this conception of representation as representativeness rests on the implausible assumption that national populations and societal groups have fixed prior preferences, opinions and understandings of their needs and interests, which will not change during representation. It leaves out the bidirectional deliberative and mobilising aspects of political representation,5 whereby would-be representatives ‘give the people an image of themselves to reflect on’, do their best to ‘offer plausible reasons’ for why the people should identify with this image, and ‘attract a following’ or fail to do so (Brito Vieira and Runciman 2008, 139, 101, 103). Interest and opinion formation is, at least to some extent, endogenous to the process of representation.
If the above points are accepted, what emerges is a model of political representation which holds both the content of a representative’s portrayal of the group it represents to its audience and the relationship between representative and represented group to be key factors determining the quality of political representation a group enjoys. Though the representative should be free to propose a conception of the group’s interests, needs and opinions which no member of the group had thought of before, group members must also have some mechanism at their disposal for indicating endorsement or rejection of the representative’s portrayal.
Below, I examine in more detail four aspects of political representation so understood, with a particular focus on identifying ways in which groups may be misrepresented or otherwise inadequately represented. As I do so, I highlight several recurring findings in Ntsebeza’s empirical studies which are relevant to the assessment of rural dwellers’ political representation. Though in his discussions they are not explicitly linked to the evaluative question whether rural dwellers enjoy adequate political representation, I argue that these empirical findings form the basis for diagnosing several pathologies of representation.
Claimed constituency vs authorising constituency
An informal, self-appointed political representative will often have two types of constituency. The first is the group whose needs, interests and opinions it purports to represent: its claimed constituency. The second is another group of people – possibly, but not necessarily, a subset of or overlapping with the first – which exerts concrete control over the representative: its authorising constituency (Montanaro 2018). That an informal representative’s claimed and authorising constituencies do not precisely coincide is, in some cases, unavoidable, and will not necessarily prevent it from providing effective political representation to a group which otherwise would lack adequate representation (Montanaro 2018, 47). However, a mismatch between the needs, interests and opinions of its claimed constituency and those of its authorising constituency is clearly liable to create a situation in which an informal representative misrepresents its claimed constituency to its audience – for example, government legislators (Montanaro 2018, 129–131).
In his studies of rural dwellers’ political representation, many of Ntsebeza’s assessments in effect highlight precisely this problem of constituency mismatch. For example, he discusses Tshintsha Amakhaya, an NGO network established in 2010 with a claimed constituency of ‘farm workers, farm dwellers, communities claiming land restitution, and farmers and land users on municipal commonage and in the rural areas of the former bantustans’ (Ntsebeza 2013, 144). Ntsebeza observes that Tshintsha Amakhaya’s ‘dependence on external funding exposes it to the same pressures that most non-governmental organisations labour under’ (Ntsebeza 2013, 149). Prominent among these is that, following the end of Apartheid, most NGO donors – an indispensable authorising constituency – have been very positively disposed towards the African National Congress (ANC)-led government, meaning ‘NGOs were under pressure from their donors to collaborate with the government’, with some donors even ‘channel[ling] their funding through government departments and expect[ing] NGOs to do commissioned work for the government’ (Ntsebeza 2013, 138). Such a strong pro-government orientation on the part of its authorising constituency would tend to disable a NGO from properly representing its claimed constituency’s needs and interests, whenever the obstacle in the way of their fulfilment was government policy – as, argues Ntsebeza (2013, 137–138), it frequently is.
But the most flagrant case of constituency mismatch diagnosed in Ntsebeza’s work is that of many, perhaps most, of the Traditional Authorities during colonialism and Apartheid. Though notionally intended to represent the views of local people to the central government,6 Ntsebeza explains that local headmen’s power to appoint members of District Councils, combined with the government’s power to appoint and replace local headmen, meant that District Councils and headmen were answerable not to their claimed constituency (local people), but to an authorising constituency (the central government) with a very different set of interests. In 1903, a Transkei resident, Mr E. Mamba, told a select committee of the Cape Parliament: ‘The Government take it that [the councils in the Transkei] represent the people, whereas they only represent the headmen’ (Ntsebeza 2005, 73); and the headmen were answerable to the government.
What is so striking about this case is that upwards accountability had created a situation in which the audience (the central government) of these political representatives doubled as their authorising constituency. Such a configuration was liable to – indeed, arguably it was designed to – create a frictionless form of representation permitting the audience to hear only what it wanted to hear. Like the colonial governments exercising indirect rule before it (Mamdani 1996, 82), the Union of South Africa government was more interested in itself being represented in word and deed to rural subjects via Traditional Authorities, than in those subjects being represented to it. The advent of Apartheid meant, at least in the Transkei, that much of the power previously vested in headmen was now transferred to chiefs by government decree. However, the latter had very limited discretion in how they wielded this power. Following passage of the 1951 Bantu Authorities Act, ‘Recalcitrant chiefs – those who refused to be party to such an evidently oppressive system – were simply sidelined to make way for their more compliant counterparts’ (Hendricks and Ntsebeza 1999, 105).
Ntsebeza argues that, under the ANC-led government, the powers of Traditional Authorities in South Africa, at least as regards land allocation, have come to fit Mahmood Mamdani’s model of ‘decentralized despotism’ (Mamdani 1996, 18) more closely than before. Without magistrates or district commissioners looking over their shoulder, Ntsebeza contends, Traditional Authorities in many cases now have more autonomy to be despotic (Ntsebeza 2005, 287).7 But a missed opportunity in Ntsebeza’s work on the transition from Apartheid to democracy is a comparative assessment of the forms of political representation rural dwellers have experienced. Have Traditional Authorities become better or worse political representatives of rural dwellers to government, now that their upwards accountability has been removed or reduced? Did Traditional Authorities or the local informal representative bodies competing with them in the 1990s provide better representation to rural dwellers? Many chiefs were evidently personally corrupt (Ntsebeza 2005, 219–221, 244), but, as Leslie Bank and Roger Southall point out, ‘the evidence is that civics no less than chiefs can easily become corrupted’ (Bank and Southall 1996, 426). Equally, while Traditional Authorities resorted to banishment orders and corporal punishment to compel obedience, EMRA threatened locals that it would burn their houses unless they complied (Ntsebeza 2005, 234, 216, 248).
It would also be worth investigating whether Traditional Authorities have in effect acquired a new authorising constituency, taking the place of the central government, and whether there has been constituency mismatch either in the case of Traditional Authorities post-Apartheid or in that of the local civic representative bodies. Given that, as we saw in the previous section, Ntsebeza believes some formally elected representatives exhibit constituency mismatch in post-1994 South Africa, it would be important to assess the severity of these cases in comparison with the constituency mismatch of Traditional Authorities as well as informal civic representative bodies.
Constituency of what?
Misrepresentation can occur when a political actor depicts a group’s needs, interests and opinions as being other than they are. It can also occur when a political actor misrepresents the group’s scope or extension – when individuals to whom the category which forms the basis of the group applies are discounted as though it did not apply to them, or when individuals to whom this category does not apply are counted as though it did.
Ntsebeza’s work on Traditional Authorities in South Africa shows that there is yet a further way in which a group can be misrepresented: what I shall call type misrepresentation. This occurs when a group is spoken for and acted for as though it were a different type of group from the type of group it is.
The type misrepresentation to which Ntsebeza’s writings on South African Traditional Authorities draw attention is the portrayal as ‘tribes’ of groups which are not ‘tribes’ and, at least to a large extent, do not view themselves as ‘tribes’. As Ntsebeza writes, ‘tribalism is inherent in the recognition of separate chieftaincies’ (Ntsebeza 2005, 293). This is because, in South Africa as in many other African countries, the constituency of a chief or paramount chief has not been conceived of purely territorially. Rather, it has been conceived of as a group bound by culture, language and blood relations which, however widely dispersed it may currently be, aspires to ‘territoriality, primitive government through elders and chiefs, and a primitive subsistence economy’ – as Archie Mafeje puts it in a critical overview of the ‘ideology’ of tribalism (Mafeje 1971, 257–258). Mafeje notes that conceiving of Africans as divided into ‘tribes’ not only ‘over-simplifies, mystifies, and obscures the real nature of economic and power relations between Africans’,8 but has done so in a way which facilitated African societies’ being ‘drawn into a complex of extractive economic and political relations’ associated with ‘indirect rule’ (Mafeje 1971, 254, 261).
Mamdani (1996, 51, 115–117, 148–154, 279) has subsequently described this ‘complex’ in detail. Having stipulated that ‘The bearer of custom’ is ‘the tribe’ and having reserved the right to modify ‘custom’ where necessary to suit their purposes, colonial governments used the ideology of tribalism to justify subjecting even workers in large industrial cities to tribal authorities, and to justify maintaining a compulsory labour regime, supposedly decreed by ‘custom’, in rural areas alongside the market economy. During Apartheid, the National Party government not only drew on the rural reserves administered by tribal authorities as reservoirs of labour for South Africa’s mines and factories, but also, invoking the purported authority of ‘tribal law’, ordered rural ‘Conservation schemes’ to be ‘largely implemented by forced unpaid labor’ (Mamdani 1996, 183, 191).
Ntsebeza follows Mafeje and Mamdani in rejecting the tribalist portrayal of Africans as a false framing which has served the interests of a governing elite. What is more, he demonstrates that of the many people in Xhalanga District who objected to being politically represented by Traditional Authorities, some specifically objected to being represented as a constituency comprising a ‘tribe’. For example, at a meeting with the Magistrate in January 1956, when the Bantu Authorities Act was being implemented, Mr E. Qamata, as Ntsebeza recounts:
accused the government of introducing ‘the Jungle Law’ in insisting on dividing ‘the Bantu according to their tribes …’. In his words: ‘With the Europeans such a thing does not exist. We do not know the difference between different European tribes as they are all one white race.’ (Ntsebeza 2005, 141)
Of course, Ntsebeza does not claim that everyone in Xhalanga objected to being politically represented on a ‘tribal’ basis: some people did not (Ntsebeza 2005, 127). Nor does he claim that Xhalanga is an absolutely typical rural district of South Africa. But his work supplies strong grounds for insisting that, at the very least, rural dwellers should be able to opt out of being represented as a member of a ‘tribe’ or – in the post-1994 nomenclature – ‘traditional community’ (see Hendricks and Ntsebeza 1999, 109).9
An image of themselves to reflect on
If one function of political representatives of a group is to give group members ‘an image of themselves to reflect on’ (Brito Vieira and Runciman 2008, 139) – to present to the group itself a potentially novel account of its various needs, interests and opinions, and how these do or should relate to those of other groups (see Montanaro 2018, 40) – how are we to assess whether this function is being adequately fulfilled?
A tempting answer would be that it is fulfilled if, and only if, the representative’s portrayal of the group is accurate, if the portrayal depicts the group as it actually is. While it cannot be wholly wrong, there are some problems with operationalising this criterion as it stands. For one thing, an observer assessing the accuracy of a political representative’s portrayal of a group must recognise that they themselves are not infallible. Their own understanding of the group in question might be incorrect. Moreover, being confronted with a political representative’s very different portrayal of the group is one conceivable way in which an observer might be jolted into the realisation that their own understanding is incorrect.
For another thing, in some circumstances at least, it would seem that a political representative ought to represent the needs and interests of a group as the members of that group themselves construe them, even though the members’ construal genuinely is erroneous. Once the representative had offered the group their (the representative’s) construal of their needs and interests and the group members had persisted, throughout extensive engagement, in nigh-on unanimously rejecting the representative’s construal, it would be hard to justify the representative’s bringing to an audience their own construal of the group’s needs and interests, rather than the group’s construal of their own needs and interests. If this is right, then in some circumstances an inaccurate portrayal of a group’s needs and interests would constitute more adequate political representation of the group than an accurate portrayal.
In his work on political representation of rural South Africans, Ntsebeza stresses the importance of the ‘political education’ provided to groups of rural dwellers by informal representatives (Ntsebeza 2005, 118–20, 204–6). He tends to be in favour of those representatives – like Cala University Students Association (CALUSA), once a certain Bambo Qongqo had weaned it off ‘Black Consciousness philosophy, in favour of socialism’ (Ntsebeza 2005, 236) – who offer political education which portrays the salient dividing lines within and between rural groups as being lines of economic class, as opposed to lines of race or ethnicity. K. D. Matanzima’s ‘campaign to insinuate the notion that tensions at Emnxe, in particular, were “ethnic”, between amaMfengu and abaThembu’, on the other hand, Ntsebeza dismisses as a ‘strateg[y] … of divide and rule’ (Ntsebeza 2005, 187).
Though Ntsebeza gives strong reasons for favouring the portrayals of rural groups that he does, in so doing he has thrown his own hat into the ring as a would-be representative – at least in the sense that he is offering, and justifying, his own portrayal of the groups in question. Assessments of the portrayals offered to groups by their political representatives must evaluate not only whether their content is accurate, but also whether group members are exposed to a sufficient variety of portrayals and would-be representatives are not prevented by law or otherwise from offering a group any particular image of their needs, interests and opinions to reflect on.
Discourse and exit
Ntsebeza voices support for an overhauled form of democratic government which would involve both representation and participation: a ‘democracy’ that ‘incorporates both representative and participatory elements’ (Ntsebeza 2005, 23). Despite his criticisms of Traditional Authorities, he concedes that some precolonial governance forms involving chiefs and headmen operated according to ‘participatory principles identified with traditional democracy’; principles, he suggests, on which we in the present can fruitfully draw to create ‘a new form of democracy for South Africa and elsewhere’. This would be a form of democracy that ‘combine[s] the participatory elements of pre-colonial indigenous institutions and the representative aspect of liberal democracy’ (Ntsebeza 2005, 33, 299).
Participatory democracy and representative democracy are frequently viewed as incompatible (see Brito Vieira 2017, 1). It could even seem that combining the two is ruled out at the conceptual level. After all, if everyone participates in decision-making, what need is there for anyone to act or speak for anyone else – that is, for representation? At the least, one might think, those matters of public policy which are determined through the people’s participation in decision-making could not also be determined through political representation of the people. A ‘hybrid form of democracy’ (Ntsebeza 2005, 33) involving both would, then, in practice have to be a chequerboard of some decisions arrived at through participation and other decisions arrived at through representation.
However, putting together two points I argued for earlier reveals strong grounds for thinking that a process of political representation can itself involve participation of the represented, thus suggesting one potential way of spelling out more concretely Ntsebeza’s idea for a ‘hybrid’ conception of popular sovereignty. If, first, adequate political representation cannot be unilateral, but must involve some constituency check whereby members of the represented group indicate whether they endorse or reject their representative’s portrayal of them, and, second, formal constituency-wide elections are neither always necessary nor always sufficient for adequate political representation, then it may be that the constituency check – the mechanism by which members of a group exercise some control over how they are represented – could consist of something more substantial than simply casting a vote, even something approaching Ntsebeza’s conception of participation.
Elections are one form of constituency check, but they are not the only form, and they are unlikely to suffice on their own. When assessing systems of political representation we must, as Laura Montanaro has urged, also pay attention to ‘non-electoral forms of authorization and accountability’ (Montanaro 2018, 79). Some of these are discursive: members of a constituency communicate their agreement or disagreement with their representative’s portrayal of them and actions on the constituency’s behalf. If feasible, this communication can take the form of extended consultations, in which both representative and as many constituents as can take part seek to articulate and defend their respective understandings of the constituency’s needs and interests and the actions on its behalf which they consider necessary. Such consultative conversations involving giving, weighing, rebutting and re-articulating of reasons, and aiming at – even if they do not always arrive at – consensus, are presumably part of what Ntsebeza has in mind when he advocates for ‘the involvement of rural residents in decision-making processes’, which he claims ‘was … the hallmark of governance in most southern African societies at the advent of colonialism’ (Ntsebeza 2004, 85). That the constituency check can take this discursive form shows that there is, in principle, no conflict between the deliberative conception and the representative conception of democracy. 10
A discursive constituency check can also operate via its effect on a representative’s reputation. If enough members of a self-appointed representative’s claimed constituency state publicly that the representative’s portrayal of their needs and interests is badly wrong, then the representative is likely to lose credibility in the eyes of their audience, which may even cease to recognise them as a representative. Further, members of the claimed constituency may have the option of exercising or threatening to exercise a power of ‘exit’ – ending their subscription or membership, or ceasing to be a donor to the representative body. Those who cannot vote with their feet in one of these ways may still be able to exercise ‘discursive exit’ (Montanaro 2018, 89–91) – stating publicly, ‘So-and-so does not speak for me!’
At the beginning of this section, we saw that Ntsebeza believes Traditional Authorities have survived in twentieth- and twenty-first-century South Africa principally due to their control over land allocation. ‘[C]hiefs in general’, he observes (Ntsebeza 2005, 253), ‘through Tribal Authorities, had real power in the sense that rural residents could not gain access to government resources, including land, without their endorsement.’ We are now able to see that this finding is relevant not only to Ntsebeza’s empirical question about Traditional Authorities’ resilience (his question (a)), but also to his evaluative question about Traditional Authorities’ place in a system of representative government (question (b)).
If rural dwellers depend upon the Traditional Authorities’ administration and upon the goodwill of chiefs for access to vital resources, including land, then they are almost certain to be inhibited in making use of the non-electoral forms of constituency check described above. ‘Exit’ may be impossible, and the discursive forms of constituency check, including ‘discursive exit’, could result in the destruction of their livelihood. Though being unelected does not on its own disqualify Traditional Authorities from providing adequate political representation to rural constituencies, if it is also the case that the powers they wield render non-electoral forms of constituency check too dangerous and costly for constituents, then the representation they provide is likely to be unacceptably unilateral.
Combining the philosophical and the empirical
Ntsebeza’s sociological and historical studies of South African rural life are of particular interest because they involve not only description and causal explanation, but also reasoned evaluation of the forms of political representation which have been available to rural dwellers. His normative question about representation, important in its own right, is of particular concern to those who see rural dwellers as especially exposed to exploitation owing to structural features of the political economy. A sine qua non for efforts to overcome such exploitative circumstances is adequate organisation and representation (Chibber 2022).
Here I have sought to show that empirical work on political representation, such as Ntsebeza’s, can constructively be brought into conversation with recent developments in the normative theory of representation. I have argued that Ntsebeza often presents evaluation of the forms of political representation he considers as being a more straightforward matter than it actually is. He does this when he assumes that formal constituency-wide elections are always at least a necessary condition for adequate political representation. While it is probably true that a large country’s system of political representation as a whole would be inadequate unless it involved some formal elections, I have argued that the election criterion is neither always a necessary nor always a sufficient condition for a political actor to count as providing a group with adequate representation. Ntsebeza’s own writings on the political functions performed by rural NGOs and other informal representatives supply ample support for this view.
My principal aim, however, has been to show that much of the empirical detail in Ntsebeza’s studies can be harnessed to plug the justificatory gaps in his own evaluative assessments of instances of political representation. In pursuing this aim, I have not sought to list all the criteria which would be met in a country whose overall system of political representation was irreproachable – a fully representative democracy in which the people were truly sovereign. Such a system would ensure that all salient interest and opinion groups were properly represented by political actors with comparable advocacy skills; and, as I have granted, formal elections at various levels would almost certainly be necessary to achieve this. Rather than stating high-level democratic desiderata on the national plane, my approach has been more fine-grained and local. I have argued that the rich empirical detail in Ntsebeza’s studies of rural political representation enables us to articulate and diagnose several pathologies of representation. These are flaws either in the content of a representative’s portrayal of the local group they represent, or in the structure of the relationship between representative, the group they represent, and the audience or audiences to which they represent it, which mean a group is misrepresented or otherwise inadequately represented. Detecting and correcting such flaws in local representation is as important and necessary as theorising about ideal systems of national democracy; it is an important function which engaged scholarship such as that of Ntsebeza can discharge.
In addition, I have suggested that a promising way to develop Ntsebeza’s pregnant comments about democracy is to acknowledge that the process of political representation can itself involve popular participation. Whether a representative is formally elected or not, they should regularly undergo a discursive constituency check, involving deliberation, contestation and persuasion, in conversation with as many constituents as possible. If circumstances permit, this could amount to the joint, even consensual, decision-making which Ntsebeza associates with participatory democracy.