Locating the ‘customary’ in post-colonial Tanzania’s politics: the shifting modus operandi of the rural state

ABSTRACT This paper examines how both rhetoric about custom and practices drawing on elements of deep-rooted political culture remain relevant in post-colonial Tanzania. This is the case despite the Tanzanian government’s aggressively modernising stance and the erasure of colonial-era ‘traditional’ chiefs after independence. The paper identifies three patterns. Firstly, witchcraft cleansing remains a rare flashpoint over which rural people are willing to defy officials, amid legislation that has barely moved on from the colonial period. Secondly, for defenders of certain practices, describing them as customary is a way to try to place them beyond criticism, while for officials it becomes a way to wash their hands of the attendant problems. Lastly, a performative political practice can be discerned in the interactions between rural populations, officials and development experts that resonate with descriptions of pre-colonial political encounters. By looking for local legitimacy in interactions with so-called elders, development experts have become arbiters of (pseudo-)traditional authority despite their modernising identity. These observations show that discourses about and practices drawing on the customary have become deeply imbricated with the political practices of the rural state.

This paper is concerned with the role of customary authority in rural Tanzania, and particularly in the southern regions of Lindi and Mtwara. For discussions of 'customary authority' since independence, Tanzania presents a challenge. For the colonial period, Iliffe's treatment of 'Indirect Rule' in this territory is a classic on the topic of 'invention of tribes' that recognises the diversity of outcomes while casting it as fairly benign. 1 For the post-colonial period, though, 'customary authority' has garnered very little attention here. The Nyerere government abolished chiefship shortly after independence and positioned itself as one of the most assertively modernising and bureaucratic on the continent. 2 In keeping with the observations by Verweijen and van Bockhaven in the introduction to this special issue on the tendency to identify 'customary authority' with chiefship, the erasure of chiefship also reduced attention to customary authority. Meanwhile, ethnic mobilisation, elsewhere a useful lens for tracing loyalties with roots in custom, is conspicuous mostly by its absence. The literature on political relations since independence has consequently focused on the authoritarianism of the developmentalist state and its rural emissaries. 3 The question arises, then, of where, if anywhere, to locate 'customary authority' in post-colonial Tanzania.
This question draws the observer into the sprawling debate on the roles of 'custom' or 'tradition' in African politics since the colonial period. The definitions and characterisations of customary authority, and more widely of custom or tradition, at play here vary greatly. So do those of their antinomies, described as modern or formal institutions or simply the state. This paper takes leads from the introduction to this special issue, which makes a number of helpful observations. Firstly, Verweijen and van Bockhaven make clear that there is no point in trying to pick the practices of authority in African states and societies neatly apart into 'customary' elements on one hand and 'formal' or 'modern' ones on the other. In keeping with Spear's now-classic review of the topic, what passes as political tradition in Africa now is neither primordial nor constructed, but unstably drawn together from sources that vary from long-standing and deep-rooted beliefs or dispositions to context-bound, situational improvisations. 4 Accordingly, the discussion below finds practices rooted in local political tradition at work in self-consciously modern development interventions, without acknowledgment of their presence and notwithstanding the absence of recognised customary authority figures. Conversely, sometimes notions of long-standing tradition are invoked to bolster stances that are very much contemporary. This paper, then, accepts that there are popular political imaginaries at work in Tanzania today that are not instigated by formal authorities and draw on long-standing, arguably pre-colonial dispositions and claims. Further, these imaginaries can underpin forms of authority not derived from state endorsement. But rather than trying to extricate these claims, dispositions and authority positions from their contemporary environment, the aim is to trace their imbrication.
Secondly, in keeping with the simultaneously inherited and constructed, improvised and place-specific character of political authority in post-colonial settings, there is no generalisable account of how colonial rule affected customary authority. Nor is there one correct description of the interrelation between customary and formal authority in post-colonial Africa. Balances differ over place and time, as well as with observers' definitions and points of view. As will become clear below, the area examined here presents a deviation even from the rough likelihoods that the special issue's editors identify in that the lack of centralised authority here did not provoke attempts by colonial officials to create one. As a corollary of the salience of context-bound variation and constant improvisation, the observations below clearly chime with Verweijen and van Bockhaven's emphasis on the state as process as well as structure. In the interactions between rural officials, experts and villagers described below, it appears as constantly being put together.
Thirdly, Verweijen and van Bockhaven emphasise that elements of customary authority are not found only in connection with chiefship, to which colonial rulers often sought to confine it. They highlight that besides the discursive and socioeconomic sources of power that recent literature has emphasised in connection with chiefship, customary authority has ritual, symbolic, and arguably cognitive dimensions that dynastic-political definitions centred on chiefship marginalise. 5 This observation is very important for the case discussed here. As elsewhere in Tanzania and more widely in Swahili-speaking regions, the people I observed here had a developed vocabulary for phenomena to do with customary authority. Typically, they used mila to refer to customs inspired by Islam, and jadi to refer to customary practices not derived from any religion. They also knew utamaduni, roughly 'culture', as the purview of one of the officers of local governments (the afisa utamaduni), and its evil twin, imani potofu or ushirikina, roughly 'superstition'. 6 Nevertheless, while references to jadi and its kindred formed part of a widespread rhetoric, this rhetoric was not primarily about politics and the state. It was part of conversations that could be about the state, but also about family relations, religious life, or broad ideas about how individuals and communities related to social change.
Conversely, in relations between state envoys and villagers, customary practices were at work that were not explicitly identified as jadi. In other words, jadi and its kindred operated both as a means and object of rhetoric, and as deep-rooted practices, structures or habits of thought not explicated as jadi. Some of the arguably most powerful aspects of custom are ones not explicated as such, for instance, the performative practices outlined below. The paper by Komunji and Buescher in this issue suggests that this observation applies more widely than Lindi and Mtwara, and it could be very revealing to map these parallels.
In this paper, though, my interest is rather in exploring how the operations of such practices can help bridge the gaps between different modus operandi of the Tanzanian state: its 'authoritarian-technocratic-modernising' modus on one hand and its 'populistclientelist' one on the other. Paying attention to rhetoric about and uses of customary practice helps understand how the Tanzanian state continues beyond the defeat of the modernist agenda that initially underpinned its assertiveness, and how it balances its 'modernist tradition' of developmental authoritarianism with the accommodating practices necessitated by its straitened means and the turn towards consultative routines in development. 7 'Custom' and the developmentalist state To set this aim in context: the diverse appearances of customary authority in post-colonial African states are linked to another sprawling debate, this one on the functioning, dynamics and character of post-colonial states. Tanzania is particularly relevant to one aspect of this debate: the tension between the perceived strength of African states, if positioned as authoritarian, and their perceived weakness if considering their lack of resources and dependence on clientelism. 8 Consider, for example, the conflicting conclusions drawn by Bayart on one hand and Chabal and Daloz on the other, based on the observation of similar processes in countries with similar post-colonial trajectories. 9 Bayart characterises African states as strong enough to face down discontented urban masses and exploit rural producers. Chabal and Daloz describe it as overrun and overgrown by personal networks and barely there. Similarly, Herbst and Englebert have highlighted the difficulty of statemaking in Africa and the resulting limits on the powers of its states, while a recent political economy of Tanzania describes its state as very stable and firmly in the grip of an entrenched metropolitan elite, even though Tanzania clearly shares in the structural problems identified by Herbst and Englebert. 10 In part, this contrast arises from differences in definitions, terms and case studies. But it also reflects the complexities of understanding 'state survival' in post-colonial Africa, as states described by a variety of metrics as failing or barely there continue to make themselves felt strongly, if intermittently, in the lives of their citizens. Tanzania makes a useful site for trying to explain how these different descriptions make sense, partly because its stability is often taken as a given, in contrast to many other African states. Factors cited to explain it include first president Nyerere's exceptional political skills, the absence of large ethnic factions, the political quiescence of little-stratified peasant societies, and the state's ability to maintain the flow of aid by exploiting geopolitical competition. 11 But political stability in Tanzania is also constantly seen as under threat, given the ethnic and religious diversity of its population and the sheer grinding poverty, and thus practical limitations, of the state itself. 12 There has to be a process, then, that maintains the stability, and customary political and social practice forms part of it.
Put differently: tracing uses of the notion of custom, and elements of current political practice that can plausibly be described as customary, helps understand how nonelite Tanzanians contribute to the perceived stability of the Tanzanian state, rather than merely accepting it. Unlike the notion of 'development' and its kindred, which stand at the centre of official legitimating discourse, 'custom' cannot be monopolised by factions close to the state (or, likewise, business). It therefore directs attention to the involvement of ordinary citizens in the political process, and to the effort involved in keeping these apparently quiescent constituencies on the state's side. In this way, it provides space for the insights, provided for example by Giblin, on villagers' active efforts to manage their relations with state power. This step, in turn, enables an understanding of Tanzanian politics that relies less on implicit assumptions about the apparent gullibility and quiescence of rural populations than is common in much of the literature. 13 Thus the view taken here is that if people seeking control over others, or over resources, invoke 'custom' at certain moments, this highlights both the strength of the state and its limits. It shows the state's strength because state agents set many of the terms for the invocation of custom, and at times, use the designation 'customary' to delegitimate certain practices as 'backward'. 14 It shows its limits in that agents of the state, as well as ordinary citizens seeking to negotiate with it, find it necessary to draw on 'custom', even though this notion has no place in official understandings of the Tanzanian state. Moreover, at times state representatives designate certain realms of human activity as 'customary', thus beyond their remit, rather than try and fail to control them. 15 In this case, the invocation of 'custom' becomes a way for state agencies to limit their obligations and retreat from intractable problems.
When tracing these invocations of the customary, the rural state appears less as a stable set of institutions dominating a quiescent rural population, and more as a process, driven by the constant effort of state representatives to project power and reproduce the political order. 16 This processual nature of the state, in turn, suggests an explanation for the difficulty of deciding whether the Tanzanian state is authoritarian, thus strong, or weak, in consequence of its shattering poverty. As rural officials muddle through, their degrees of influence and legitimacy are constantly in flux; the power of different actors waxes and wanes in ways that static descriptions strain to accommodate. Rather than custom surviving from the stateless past, the spaces of the customary in public debate and the practice of governance are co-constituted by the state and by place-specific social practice. 17 Arguably, this view on the 'customary' also softens the contrast between Tanzania and those countries where chiefdom remains part of the political structure, such as Congo, the subject of several other papers in this collection. The conclusion returns to this observation.

The region discussed here
My empirical observations derive predominantly from the Southeast of Tanzania, the regions of Lindi and Mtwara. Different parts of Tanzania exhibited significant variation in the character of relations between citizens and state authorities, and officials in this part of the country were at times more and at times less intrusive than they were elsewhere. These distinctions arise from the legacy of variations in the organisation of government during the colonial period, in connectivity between different regions and the political centres, and from the ramifications of successive policy turns for these varying structures. 18 Thus my observations are bound to be idiosyncratic, shaped by the peculiarities of this region. It is all the more striking, then, that they chime at times with observations from historically, geographically and culturally quite different places.
The distinctiveness of the southeast lies particularly in its decentralised, notoriously (among colonial officials) ineffective indirect rule regime. Early British administrators here took the region's ethnic diversity as a sign that German rule had destroyed 'tribal' organisation. But rather than seek to reconstruct it as prelude to imposing paramount chiefs, they appointed unusually small-scale 'native authorities'; essentially village headmen grouped into regional councils. 19 Their ineffectiveness fed into the reputation of the area for 'backwardness', well established by the time of independence. In the 1970s, the perceived need to act decisively against this lack of progress, and the tenuousness of patronage links to the capital, fed into an aggressive villagisation campaign. 20 Yet here as elsewhere, villagisation quickly faltered. Its nation-wide economic repercussions were reinforced by the impossibility of moving the cashew trees that were the Southeast's main cash crop. 21 Subsequently, the donor agencies chosen for this region emphasised consultation and 'basic needs', rather than technocratic, approaches to development. 22 Parts of this paper examine the ensuing forms of accommodation between state representatives and those of communities targeted by development, tracing spaces for political practices informed by custom. Some of these forms are very place-specific, but others work on a national, or perhaps even 'Swahili-phone', regional level, particularly those involving explicit claims about Tanzanian or 'African' culture. 23 Among these, claims about and practical attempts to control uchawi, witchcraft, form one of the widespread and pervasive objects of disagreement and accommodation, and one of interest not only to development specialists, but to all officials concerned with order.
Customary challenges to state authority? Healing, witch finding and the law Uneasy negotiations surround forms of custom focused on healing and, especially, countering witchcraft, which are both impossible to discount and continually resistant to incorporation into the state's legal system. 24 Praise for traditional medicine as a popular resource and its condemnation as superstition lie close together in official discourse, leaving the broad problem of addressing illness and misfortune, potentially a source of substantial authority, in an ambiguous position. 25 Like biomedicine, traditional healing is regulated in Tanzania, and both the biomedical establishment and the state seek to harness it to medical aims. 26 While such attempts offer healers a modicum of status, they also make them vulnerable to scrutiny, regulation and potentially prosecution. Concomitantly, healers typically keep a low profile; they advertise their practice, often recommending contact by phone, but do not draw excessive attention to themselves.
Witchcraft eradication, though, is a different matter. As elsewhere in British Africa, law in colonial Tanganyika Territory did not recognise witchcraft as a punishable offense, but punished witchcraft accusations. A version of these rules remains in place. 27 Consequently, the urgent problem, to many people's minds, of addressing witchcraft remains removed from the purview of the state. In the late colonial and early independence period, some witchcraft eradicators acquired fame and maintained cooperative relations with representatives of the independence movement and the state. 28 Today, too, as Maia Green has shown, there are healers who position their witchcraft cleansing as a social service and a contribution to societal progress; in a way, an effort towards the ubiquitous aim of development. 29 Yet witch finding and cleansing also constitute the clearest example of 'insurgent knowledge' in the history of colonial and post-colonial Tanzania. In colonial Tanganyika, witch finding movements never quite caused the same degree of concern among administrators as Karen Fields has demonstrated for Central Africa. 30 Nevertheless, official unease about them ran deep, as the 'witchcraft ordinance' signals with its penalties for claims to witchcraft cleansing abilities as well as witchcraft accusations. In the post-colonial period, witchcraft cleansing or prosecution have been a rare cause over which rural people in Lindi and Mtwara have been willing to defy the state.
Thus for the 1990s, Stacey Langwick reports that villagers in Mtwara protested violently against official attempts to curtail witch finding, while in Central Tanzania the killing of suspected witches has been an intractable problem. 31 In the early 2000s, I observed the anxieties and negotiations around this practice when, on the heels of rising rates of illness and death from HIV/AIDS, witch finding experienced a minor renaissance. 32 Official responses varied. In one town, the village government refused to let witch finders operate, but did not prosecute anyone. In another village, elders invited a witch finder, and after he had done his job, left the village executive officer to be prosecuted for having facilitated his activities, which was construed as a public order offence. 33 One particular witch finder created an ephemeral youth movement, observed mistrustfully by elders and authorities alike, that clashed with the young Muslim reformists who were active at the same time. Like good modernists, the Muslim reformists agreed with the administration that witchcraft fears were mere superstition. 34 Evidently, officials picked their battles where witch finders were concerned; they improvised and sought paths of least resistance.
The most striking failure I have encountered to prosecute a witch finder who had been involved in unambiguously illegal activity concerned the murder of a man who was accused of doubling as a human-eating lion, on the accusation of a well-reputed healer. 35 These events occurred around 2003 in Mbekenyera, a remote town in Lindi region. After several people had been killed by the lion, the healer, a woman who lived in a remote location reachable only by foot, called a public meeting. During the meeting, young men interpreted her ambiguous gestures to incriminate a man who was anyway considered an outsider. He tried to hide, but they hunted him down and beat him to death. According to the reports that reached me, local authorities paid some of the young men to bury the victim, who had no local relatives, in a shallow grave, the vandalisation of which by scavenging animals was taken as further evidence of his guilt. But neither the healer nor the members of the lynch mob faced charges.
Evidently, the healer in question held enough authority to immunise herself from prosecution. Despite her role in these violent events, she was typically described along the lines of the pacific witchcraft-shaving expert that Green observed in the 1990s in a neighbouring region: remote from towns and their modern ways, supported by ancient powers closely tied to nature, a purifying counter-force to witches that normally calmed conflicts, rather than raising them. 36 Her safety from prosecution highlights paradoxical features of her position. She was supposed to hold moral authority precisely because she did not interact with this-worldly, modernising pursuits of power. 37 In this sense, her influence was self-limiting from the point of view of officials. While her reputation depended on her being seen as distant from contemporary ways, she was not likely to challenge officials, the quintessential purveyors of state-endorsed modernity, directly. Nevertheless, the failure to prosecute her suggests that officials assessed her influence as strong enough to mobilise rural people in her protection if need be.
Witch finding, and healing more broadly, here constitute an unstable realm whose relationship with the state oscillates between regulation, endorsement, toleration, repression and ostracism. However variable the relationships, though, the practitioners of witch finding always have to be ready to negotiate with state representatives. 38 Thus while the healer's authority is clearly rooted in a realm of claims and practices that predate the state, it is inextricably intertwined with political practices and discourses that cannot pass as customary. The politics around witch finding, then, make clear that Tanzania's self-consciously modern, formal and bureaucratic institutions do not have a monopoly on structuring politics and governance in the country, notwithstanding the government's decisive move against the customary authorities inherited from colonial rule. But the wielders of authority drawn from customary practices still have to reckon with officialdom, even if by evasion. Conversely, state representatives, too, need to pick their battles with figures like this healer. Their failure to prosecute her forms part of a broader pattern, whereby the designation of certain issues as matters of 'custom' or, more broadly, 'culture' allows officials to corral them into a space beyond their own responsibilities.
Invocation of 'custom' as a means to abdicate responsibility: families, domesticity and the politics of culture While the authority derived from witch finding hovers uneasily round the edges of the rural state, custom and culture are routinely invoked in today's Tanzania to assert and justify authority structures in families and the domestic sphere. The kind of practices described as cultural and the values ascribed to them on these occasions vary enormously. Thus an aversion to violence and conflict is often described as a trait of Tanzanian political culture, while homoerotic sexual practices are described as against 'African' culture, implying that 'African' culture prescribes heterosexuality. More specific arrangements pertaining to marriage, childbirth and domesticity, to things such as birth spacing or commensality, may be associated with specific ethnic groups. 39 For example, defenders of girls' initiation in Lindi and Mtwara position it as a part of Mwera or Makonde culture, thus part of a heritage deserving protection. 40 This kind of claim draws on several elements of official political rhetoric. The 'village museum' outside Dar es Salaam, with its display of huts and implements 'typical' of discrete ethnic groups, is a concrete expression of the official stance that all ethnic identities form part of a national cultural heritage, and as such deserve protection. In this line of reasoning, the designation of ethnic identity as a question of culture serves to de-politicise it: ethnic traditions can live on in the realm of culture, while the political sphere can remain designated as thoroughly modern. 41 Thus a 1998 'holiday of the Makua and Yao' at the village museum featured the heir to one of the long-standing big-man titles among Yao-speakers, Mataka, seated at a table while women passed him on their knees, ululating. What would once have been a demonstration of political power was repositioned as spectacle. The utility of this approach for a state as multi-ethnic as Tanzania is obvious. Nevertheless, it also has the effect of moving 'ethnic' culture beyond debate: if a practice is designated as part of a group's culture, then the default position is that it is protected alongside all others. Hence the invocation of culture in defence of girls' initiation.
Indirectly, though, this way of reasoning confers its own kind of political power on the designation of certain practices as part of, or inimical to, particular 'cultures'. This is most immediately evident in the way the condemnation of certain sexual preferences or practices as 'un-African' has served the socially-conservative agendas of religious activists in Tanzania. 42 Similarly, various versions of 'culture' (African, Tanzanian, ethnic or religious) are invoked with striking frequency to justify patriarchal domestic relationships. 'Here in Africa, we believe in culture', is a sentence that I have heard uttered in defence of asymmetrical gender roles, and there are countless permutations of this trope. Thus it is 'tradition' for wives to take off their husband's shoes when they arrive home, to bring them water, to provide them with the choicest piece of meat at dinner, to not contradict them, and so on. 43 Clearly, these are schematic assertions that only very partially reflect the practice and power structure in individual households. But they are often invoked to counter official developmentalist rhetoric on gender equality and women's rights. 44 That said, challenges to such assertions of gendered entitlement protected by custom are possible. They involve drawing on a different register of public discourse on culture; one that is aligned with its modernising, developmentalist strand. For 'custom' can be discarded if it is designated as Imani potofu or ushirikina, roughly, superstition or false belief. 45 It is hardly coincidental that while men often wield the rhetoric of tradition, these terms critical of tradition are heard particularly from women. They occur for instance in connection with opposition to female circumcision, and Hodgson shows their role in connection with choice of marital partner. 46 Obviously, there is no consistent standard by which to sort protected 'custom' from obsolete 'superstition'. The dividing is drawn in the constant interaction of policy agendas, official rhetoric, popular claims and private negotiations.
Ultimately, the effects of this mock-antiquarian manner of depoliticising and quarantining particular social practices and claims by declaring them cultural are ambiguous. It fences off questions that are clearly of enormous relevance to social policy and to the kind of social change often glossed as development, such as the negotiation of reproductive choices. It corrals them into a realm that formal, modernising state authority washes its hands of. Yet while this move threatens to deprive the losers of these 'ringfenced' cultural practices of recourse, challenges to this kind of rhetoric remain possible by appealing to the state's modern identity. The language in which they have to be put, though, is of an assertively progressive kind, the kind that dismisses beliefs as superstitions. Such modernising rhetoric has begun to sound almost old-fashioned; it is part of what could be called the state's 'modern tradition'. 47 The ambiguous role of government, either endorsing or censuring presumed custom, indicates a process of mutual accommodation between defenders of custom and purveyors of modernity. Another form of accommodation is evident in public political performances, where developmental rhetoric becomes intertwined with unnamed customary practices so as to help rural officialdom be seen to be in charge.

Bureaucratic performance and the limits of quiescence
In tracing ways of exerting authority that draw on custom without saying so, this paper proceeds from the scepticism towards the assumptions of rural political quiescence stated in the introduction. 48 This scepticism is based on the observation of villagers' perceptive conversations about and strategic interactions with state representatives, as well as reports of rural unrest from newspapers and secondary literature. 49 They indicate that, while rural officials can rely on material support from the centre if rural discontent becomes an emergency, such emergencies are expensive and undesirable for a peripheral state that remains, in spite of the declared rural focus of official development policy, pathetically under-resourced. 50 Moreover, the 'buying of consent' through patronage is not a straightforward process, as patronage resources are limited, the processes whereby they are procured complicated, and rural audiences are opinionated on their needs and preferences. 51 The above discussion of their refusal to abstain from witchcraft cleansing and, in the case of the murdered 'lion-man', witchcraft prosecution, has made this refractoriness evident.
Under these circumstances, there is a genuine element of debate and persuasion to rural politics in Lindi and Mtwara. Officials have to make their cases, and if they fail to do so, risk quiet sabotage or at times open confrontation. Close up, the developmental authoritarianism of the rural state devolves, in this region, into encounters between officials and their audiences in which both sides are under some obligation to demonstrate and affirm their commitment to shared goals, and to produce contributions towards them. 52 Typically, these take the form of access to NGO inputs that officials provide to villagers, and of voluntary labour and other efforts that villagers provide to the interventions proposed by these NGOs. Development NGOs, then, have effectively become intermediaries between the state and its citizens. 53 Some observers, most prominently Maia Green, have characterised this symbiosis between development NGOs, the state and an aspiring Tanzanian middle class of development facilitators as a mechanism that keeps rural people in their marginalised, largely receptive, condescended-to place. 54 But the occasions when experts descend on villages in all their splendour are part of a broader process in which, I think, provincial citizens have somewhat more leverage than she allows. While provincial people need the inputs and services that rural officials and experts control, the former also need the latter. In particular, while NGO representatives themselves can be characterised as intermediaries between donors and target audiences, or donors and state authorities, they, in their turn, need intermediaries from among the populations they target, and more broadly, need cooperation in their projects from both local authorities and target audiences. 55 The result is a messy process of mutual cajoling and criticism. Take, for example, a 2003 attempt to get people to show up for an elephantiasis vaccination campaign in the rural town of Rwangwa. It involved loudspeakers mounted on pick-up trucks incessantly circling the town, calling on people to join the queues in a tone that appeared borrowed from radio advertising. They did not, however, directly address the rumours that were making the rounds about fatal allergic reactions to the vaccine or its being derived from pigs, thus haram for the town's Muslim majority. The debates around witch finding in Rwangwa further illustrate the mutual exasperation and mistrust. Supporters of the practice surmised that both Muslim reformists and the town government were involved in zombie-keeping, because these otherwise inimical factions were united in their opposition to witchcraft cleansing. 56 There is, then, an element of public debate and performance to rural governance: NGOs and local government deployed speakers, podiums, loudspeakers, shiny cars and so on to make their case, but failed to sway public opinion.
In particular, mikutano, meetings, have become a mainstay of developmental intervention: they are where vaccination campaigns, new buildings, visits from 'on high', agricultural interventions, bore holes and so on are announced; a constant part of the consultative processes around the kind of targeted input provision that rural NGOs have come to focus on post-villagisation. 57 Inevitably, they are highly choreographed, amounting to public performances in which ranks, priorities and common purposes are confirmed. Observing them, it is hard not to remember accounts of meetings by early colonial district officers with some of the still fairly independent local power-brokers they encountered. 58 The choreographed, disciplined and decorous nature of these public performances merits further examination. Inevitably, they were in Swahili, and speakers would draw extensively on the language's rich registers for the expression of respect and obligation. This included both the inclusion of Arabicate, 'coastal' overtones, with greetings such as shikamoo, 'I hold your feet', and forms of address such as waheshimiwa or waungwana (the honourable or the patrician/freeborn). It also included call-and-response interactions associated with mainland rather than coastal culture, often with the interjection 'oye' used as an emphatic greeting: Warwangwa oye! 'citizens of Rwangwa, I greet you!' To which the citizens addressed responded with a collective oye. Ululating, clapping and crowds running with cars were further forms of popular participation that conferred status on the people thus welcomed. The sense of protocol and decorum was highly developed also in the attention given to the order of speakers, the invitations and honours extended to 'official guests', and the way participants were seated. 59 The content of these meetings revolved around the reaffirmation of shared goals between the emissaries of the state, NGOs and citizens. The inevitable and often lengthy invocations of maendeleo, which can mean both development and progress, traced out common ground for the different parties. 60 This explicit, ideological 'developmentalism' is a widely recognised feature of the Tanzanian state. Yet the self-conscious modernism is asserted through forms of conduct that resonate clearly with descriptions from the pre-colonial nineteenth century, of the way missionaries, travellers and early representatives of the fledgling colonial state were received by the 'big men' who then dominated the area.
These pre-colonial 'big men' were self-made leaders who attracted and controlled followers through a mixture of patronage and violence, both enabled by their access to longdistance trade. Their power lay in control over people and trade goods, and when they met visitors, they brought their followers out in the open to demonstrate their clout. 61 In this way, the ability to literally mobilise villagers, to get them moving as they throng visitors in the village square, mattered to the display of authority in the 1890s as in the 2000s. The efflorescence of NGOs pursuing broad-based participation and 'local ownership' of their projects has reinforced the similarity between present-day demonstrations of authority and their pre-colonial precedents, as such participation is best demonstrated performatively.
It is perhaps worth emphasising that the failure of bureaucratic structures to monopolise the exertion of authority is not an exclusively 'African' thing. Performative, at times carnivalesque aspects of politics (especially electoral politics) are clearly in evidence also outside Africa. 62 What is distinctive, though, is the juxtaposition of a sense of propriety closely tied to the Swahili language and long-standing forms of public performance with the in-origin modernising language of development. A similar invocation of disparate sources of legitimacy is observable also in NGOs' efforts to find 'community representatives' to interact with.

NGOs and their traditional or anti-traditional community leaders
Since the rural state has struggled to recover from the over-stretch that accompanied villagisation, NGOs in Lindi and Mtwara have pursued models of 'integrated programme support' that involve targeting interventions partly according to stated or assumed 'community' preferences. 63 But who speaks for the 'community'? The most routine answer to this question again draws on widespread notions of custom or culture as patriarchal. Thus one group who are routinely afforded a part in this interaction consists of those loosely referred to as wazee, elders. They may be elders in a number of senses: party veterans, mosque committee chairs, organic intellectuals, relatively successful cash croppers, even rural traders. 64 Almost inevitably, they are married men, and they are bound to be relatively well-placed in social networks. The assumptions around who passes as an 'elder' in this context are arguably as unexamined now as they were when colonial officials were out looking for 'chiefs'. 65 Really, what qualifies somebody as an elder or, more widely, as a 'community representative' or 'volunteer' is their usefulness in the context of a particular official intervention.
At the same time, though, the development NGOs active in Lindi and Mtwara routinely insisted and prided themselves on the inclusion of constituencies cast as traditionally (customarily?) marginalised; above all, women and youth. These groups, too, are therefore routinely included among the 'stakeholders'. This means that some 'community intermediaries' will be drawing on discourses of kinship and of hierarchies cast as traditional, while others will derive their relevance precisely from representing constituencies of whom it is assumed that they were 'traditionally' ignored. 66 Either way, the actual social relations that make a person's claim to either kind of status plausible are bound to be personalised and idiosyncratic.
Clearly, then, this kind of intermediary status is acquired in a process of constant construction and reconstruction of social relationships, in keeping with so many observations on the nature of 'customary' authority in colonial and post-colonial Africa. 67 Both appeals to 'custom' and tradition and appeals to progress, development and modernity can be used to claim such authority. Perhaps, the aggressive way in which the early post-colonial state moved against chiefly authority is one of the reasons why developmentalist ideology has come to take up so much space: the appeal to development had to do work that elsewhere was taken on partly by the notion of custom. This perspective offers an alternative to accounts of villagisation's developmental authoritarianism that place the global intellectual and political currents of 'high modernism' as the prime mover. 68 Rather than modernist expectations 'running away' with Tanzanian officials, from this point of view their aggressive invocation of modern aims and methods was partly the function of a lack of alternatives; their inability to fall back on invocations of customary authority as part of the structure of governance.

Conclusion
It is evident from the above that despite the abolition of 'formal' customary authorities in post-colonial Tanzania, 'the customary' remains a productive category of political discourse, in ways that are very much concerned with authority. But perhaps the most revealing part of the preceding sentence is the bracketing of 'formal' with 'customary'. That it is necessary to distinguish between formally-constituted customary authority and popular, informal versions indicates that the notions of customary authority at stake here have themselves long been drawn into the orbit of formal state power, notwithstanding the fact that customary authority was (and remains) supposed to be alien to formally constituted, bureaucratic, modern state power.
In fact, this distinction between customary and formal realms is highly unstable. As we have seen, there are development NGOs, quintessential harbingers of modernity by their self-definition and in the eyes of their target populations, who consult with and thereby endorse so-called traditional elders. Conversely, attempts to challenge practices that their defenders seek to protect by terming them customary have to call on notions of modernity that have themselves become elusive. Appeals to development continue to evoke an automatism of progress and supposedly-unchallengeable forms of rationality that have turned out to be mirages; rather than describing facts they function as part of a 'discursive tradition' of modernisation.
By contrast, when it comes to the place of customary forms of authority in popular, dissident politics, remarkably little appears to have changed since the colonial period, at least when considering notions of witchcraft. Now as then, the phenomenon of witchcraft is always a potential and often an actual embarrassment to state power. Open engagement with it is impossible while official authority refuses to countenance the reality of witchcraft to the vast majority of its citizens. Instead, officials oscillate between dismissing concerns over witchcraft, trying to 'translate' it into categories they can accommodate, opposing witchcraft cleansing, tolerating it or turning a blind eye. For people to whom witchcraft is an indisputable problem, the urgency of dealing with it can provide a rare cause for defiance of state authorities. In this sense, witchcraft experts can acquire an (albeit very context-bound) authority that clearly draws on long-standing custom and stands apart from, at times even counters, state authority.
The significance of witchcraft arguably extends beyond its ability to crystallise dissent and defiance. Witchcraft discourses have been shown to be both deeply moral and often concerned with fairness in the distribution of scarce goods, and in this way, they provide signposts to a wider discursive field. 69 They signal the vitality of political imaginations that occur also in many forms that have not been discussed here, from the politics of Christian and Muslim congregations to the political commentary in newspaper poetry, journalistic commentary, popular music and rumour. In conjunction with the careful strategising by rural officials, seeking to deploy their connections with NGOs as well as with the political centre in the carefully choreographed performances outlined above, these imaginations are a reminder that the absence of more visible and vigorous political dissent in Tanzania is not simply a non-event. Rather, the reproduction of peripheral state authority in Tanzania takes constant work.
These performances, moreover, highlight the importance of looking beyond rhetoric and explicit, verbalised discourse when tracing the spaces of the customary. The distinctive way of doing things that I encountered, strongly reminiscent of pre-colonial sources, points towards continuities in the form and practice of politics. One likely implication is that then as now, language, verbal statements, cannot be taken at face value; they are one element in a much richer spectacle. Statements that appear disingenuous to an observer expecting pragmatic, technical reasoning may be less so if speaker and audience know they are participating in a kind of theatre.
In this manner, the recurrent reliance on performance points to the influence of a longstanding political culture that explains aspects of political practice otherwise easily dismissed as post-colonial dysfunction, such as apparently empty rhetoric. Yet the continuities should not be over-emphasised, either: the same or similar sense of decorum and spectacle clearly ends up doing very different work whether deployed in a slave-trading big man's diplomatic mission or in the interactions between development experts and their target populations. As the example shows, present uses of customary practice are not by default worse than the past.
The recurrence of performative elements nevertheless reflects the persistence of the challenges confronting those seeking control over people. In part, the challenges arise from this political culture, as when witchcraft discourse is deployed to express mistrust towards the powerful. But they also arise from the practical difficulty of maintaining territorial control, with the straitened means available to a state as impoverished as the Tanzanian. In this context, the public performances form a way of making maximum use of those times when citizens are reachable for the representatives of the state, which matters precisely because a lot of the time the state's presence remains fairly ephemeral.
The salience of references to development on these occasions has in itself become part of a post-colonial political tradition that has formed in Tanzania over the last half-century. As a contributor to the effort to keep a struggling administration going, the search for 'community representatives' in contemporary development resembles indirect rule, highlighting the gap left by the erasure of the formal customary authorities that served similar purposes in the colonial period. At the same time, the role of political performance in connection with development makes clear that customary authority is not summed up in chiefship. It also shows the active involvement, rather than mere compliance, of villagers in the reproduction of the rural state, and the instability of the distinction between 'custom' on one hand and 'modernity' on the other.