Jump to ContentJump to Main Navigation
Becoming Muslim in Mainland Tanzania, 1890-2000$

Felicitas Becker

Print publication date: 2008

Print ISBN-13: 9780197264270

Published to British Academy Scholarship Online: January 2012

DOI: 10.5871/bacad/9780197264270.001.0001

Show Summary Details
Page of

PRINTED FROM BRITISH ACADEMY SCHOLARSHIP ONLINE (www.britishacademy.universitypressscholarship.com). (c) Copyright British Academy, 2022. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in BASO for personal use. Subscriber: null; date: 27 June 2022



(p.1) Introduction
Becoming Muslim in Mainland Tanzania, 1890-2000

Felicitas Becker

British Academy

Abstract and Keywords

Muslims and Muslim polities have been present for hundreds of years on the coast of East Africa. Some 80 percent of the population in Southeast Tanzania is estimated to be Muslim. How and why this came to be the case, and how this process has shaped both the ritual practice of these Muslims and the way they understand their place within their country, is the subject of this book. It concentrates on the role of proselytizers rather than the motives of converts, emphasizing the former's personal commitment and piety. The conversion to Islam among non-Muslims in the countryside, and the spread of Sufi orders in the towns where many people already were Muslim, are addressed. The religious practice and everyday life, and the role of the state, travel, and local society in the production of oral records are described. Moreover, the chapter discusses the historicity of local views of history and religion, kin networks, villages and religious affiliations, and the polyvalence of religious change, struggle, and negotiation.

Keywords:   Muslims, Southeast Tanzania, kin networks, religious affiliations, Sufi orders, religious practice

The topic

On the coast of East Africa, Muslims and Muslim polities have been present for hundreds of years. Kilwa, one of the Swahili towns on the stretch of coast studied in this monograph, had Muslim citizens, ruled by a sultan, when Ibn Battuta visited the town in the fourteenth century.1 Beyond the coastal towns, though, European colonizers in the last decade of the nineteenth century found that almost all the people on the mainland practised their own religious traditions, distinct from Islam. Anglican and Catholic missionaries who worked in Southeast Tanzania from 1876 and 1896 respectively, confirm this observation. Sixty years later, however, the situation was very different. In spite of the missionaries’ presence, by the time of independence in 1961 Muslims outnumbered Christians, while those who exclusively followed local religious practice had become a minority.2 Today, some 80 per cent of the population in this region are estimated to be Muslim. How and why this came to be the case, and how this process has shaped both the ritual practice of these Muslims and the way they understand their place within their country, is the subject of this book.

That many people in Africa converted to Islam during the colonial period, between c. 1890 and 1960, is a well-known fact, and is one of a number of unintended apparent outcomes of colonialism that observers (p.2) began to comment on in the late colonial period.3 ‘Islamization’ figured on this list of outcomes alongside ‘detribalization’, ‘urbanization’ and ‘industrialization’. Ever since it was noted, though, it has also been said that the process is poorly documented and only tentatively understood. To counter the first problem, the present study relies heavily on oral information obtained through interviews. Interviews have provided concrete answers to many questions about the ‘how’ of conversion: about individuals who sponsored mosques or taught others how to be Muslim; about the sites of mosques and Quran schools and about ritual changes. Given the importance of multiple individual actions and subjective evaluations in the process of religious change, though, it is inevitable that motives can be recovered only indirectly. To understand why these changes occurred at the time and at the scale they did, we depend on interpretation.

The present study therefore has to pay close attention to what people in Southeast Tanzania understood by Islam, or by being Muslim, and what they sought to attain by becoming Muslim. This question may seem contrived: the personal reasons why a set of people, most of whom are now dead, changed their religious allegiance are unrecoverable. Still, the story to be told here depends on recovering the significance of becoming and being Muslim for Southeast Tanzanians, and the way this significance arose from changing contexts. Rather than the codified, explicit normative statements through which non-Muslim observers often conceptualize Islam, we have to follow often unstated significance bound up with social interaction.4 Rather than the explicit, or at least explicable, subjective commitment to a particular interpretation of personal experience and choices suggested by the term ‘meaning’, we can take ‘significance’ to be a quality that attaches to personal trajectories from their wider context. In the present case, this context was provided by combination of colonial subjection with its attendant social and economic constraints, and the memory of Zanzibari hegemony.

Based on an understanding of major themes in East African history, the reader may yet imagine a fairly straightforward account of the spread of Islam. It would run as follows: the Swahili coast had long been Muslim. During the nineteenth century, interaction between the coast and the (p.3) interior of East Africa increased greatly in the course of the expansion of Zanzibar’s commercial empire.5 With increased contact, people in the interior became cognizant of the religion of the coast. Their subsequent acceptance of Islam can then be attributed to a number of motives, including networking, acquiring a new and less parochial identity, or the pursuance of patrician wealth and status through the culture associated with it. A polemical version of this line of reasoning can be found in the writing of missionaries and early colonial observers, who used it to imply that Africans converted to Islam for materialist rather than religious reasons and that these conversions were only ‘skin deep’.6 Still, the basic factors of increased contact and the attractiveness of coastal culture are beyond dispute. This line of reasoning is in keeping also with a good deal of literature on other countries that became Muslin gradually and without conquest by a Muslim power. Accounts of Islamization that focus on the role of traders, of immigrant Muslims and their contact with local elites have been written, for instance, about Java, Sahelian West Africa and Central Asia.7

Yet the literature on East African history also indicates the problems with this line of reasoning.8 The first lies with the ambiguous attitudes to the coast among up-country people. Impressed by and covetous of its (p.4) wealth, people in the interior also knew the coast as the origin of the constant threat of enslavement and, with country people everywhere, derided their inhabitants’ reliance on comfort and niceties. In some locations, rural people actively opposed Muslim proselytizers.9 Moreover, even within one location, not all people might feel the same about Islam. From a different angle, it has been suggested that Islam served as a means for chiefs to centralize ritual authority, and counter the dispersal of political authority among the people of south-eastern Africa. Yet, as we will see, it is far from clear that for aspiring leaders becoming Muslim was an effective way of pursuing this aim.10 Even on the coast itself, people outside the dominant patrician milieu looked upon coastal culture with a good deal of ambiguity, vacillating between imitation and mockery. Muslim patricians, in turn, were not averse to treating Islam as an exclusionary ideology, rather than a missionary religion, and a marker of their own exclusive status.11 This mixture of attitudes can hardly be taken as an obvious precursor to the widespread adoption of Islam.

Similarly, the nature of the evidence for villagers’ contacts with coastal Islam confronts us with the aggregate nature of Muslim identity. Much of the evidence, especially the precious scraps for the early years, before 1930, does not concern persons saying of themselves ‘I am Muslim’, but the occurrence of practices, dress styles or names associated (at first, in the observer’s mind) with Islam. It is impossible to interpret this kind of evidence without discussing what the people involved understood by ‘Muslim’ and ‘Islam’.

Arguably the most fundamental problem lies with the timing of the expansion of Islam. The apparently straightforward explanation based on increased contact and the attractiveness of Islamic culture refers to events (p.5) during the nineteenth century. Most of the conversions in the Tanzanian interior, though, occurred during the colonial period between c. 1910 and 1950. At this time, South Asian traders had displaced coastal patricians and Zanzibari/Omani immigrants in commercial relations, and in politics Zanzibari hegemony had given way to that of the colonial powers.12 That exchanges between coast and interior remained as intensive as they had been before colonialism, and coastal culture as attractive, cannot be taken for granted under these circumstances. How, then, can we explain the continuing attraction of Islam? If the power and wealth of the coast invited conversion, what are we to make of the fact that so many people became Muslim after the end of long-distance trade? If the dominance of Muslim milieus favoured conversion to Islam during the nineteenth century, should villagers not, by the same token, have preferred to become Christian in the twentieth? Unless we want to assume that up-country converts ignored the changes going on around them, we need to explain how the significance of being Muslim shifted and remained positive under changing circumstances.

One way of addressing the question is by focusing on the role of proselytizers rather than the motives of converts, emphasizing the former’s personal commitment and piety.13 From the late nineteenth century onwards, so this argument runs, saintly pioneers from the Muslim coast implanted Islam in the countryside. It is an account largely in accordance with interviews, which tend to focus on highly moral motivations and personal piety among both converts and proselytizers, and it has some undeniable truth. Two Sufi orders, the Qadiriyya and the Shadhiliyya, rapidly gained adherents during the first half of the twentieth century, and informants associate their growth with a handful of major saintly leaders. Actually, though, the focus of Sufi expansion in Southeast Tanzania lay on the coast, not in the countryside, where most new conversions to Islam took place.14 We are effectively looking at two parallel processes: conversion to Islam among non-Muslims in the countryside, and the spread of Sufi orders (p.6) in the towns where many people already were Muslim.15 Ultimately, the appeal to the role of Sufism shifts rather than solves the explanatory problem: why did these Sufi saints emerge now? If interaction with Muslims and Sufis was enough to lead to gradual conversion, then Islam and Sufi orders should have spread slowly inwards from the coast for centuries.16 The injunction upon Muslims to convert others, if such it was, was not new in the twentieth century, so why was it practised now?

A further problem arises from the mosaic of Christian and Muslim areas that developed in the former Zanzibari sphere of influence. Within the region considered here, Masasi district, though neither culturally nor structurally exceptional, became mostly Christian.17 In other districts, there were strong Christian minorities, especially near mission stations. Muslims and Christians may be found within the same families or lineages. Clearly, the acceptance of Islam was not predestined by the character of local society. It was in effect part of a process of religious change that had three strands: Islamization was intertwined with Christianization and with change in indigenous religious practice.

It is also difficult to associate any of these religions systematically with a particular attitude to colonialism.18 At first sight, there would seem to be a clear contrast: Africans’ encounter with Christianity under colonial rule inevitably drew its meanings partly from the fact that Christianity was the religion of the colonizers, a part of their culture and knowledge that could (p.7) be appropriated, tried, changed, and challenged.19 Christian communities were not therefore ‘pro-colonial’, but they performed a kind of cultural accommodation to colonial influences.20 Obviously, the relationship between colonialism and the acceptance of Islam cannot be construed in quite the same way. Conversely, it would be tempting to ascribe an oppositional tendency to the shift towards Islam, but that too is difficult: areas where resistance against colonial rule had been fierce later became mostly Christian; Muslim areas were not systematically more restive than Christian ones; and some Muslim leaders endorsed colonial rule.21

We can salvage some insights from this confusion: if becoming Muslim remained a valid option under colonialism, then the need to engage with the colonial presence had not monopolized the interests of religious innovators. Likewise, the attempt of European observers to explain Islamization as Africans’ ambitious imitation of socially ‘superior’ Arabs is disqualified by the persistence of conversion beyond the marginalization of the coast. The explanation given below instead turns on two quite different claims: firstly, not ‘Islam’ as such, but certain tenets of Muslim allegiance had become part of local idioms of power by 1900. Concomitantly, when a new generation of local proselytizers drew upon them in later decades, they were not addressing their relationship with powerful outsiders. Rather, becoming Muslim mattered in the villagers’ own terms, within the field of social relations constituted by the village. Secondly, Muslim converts after 1920 effectively revaluated Islam, giving it an inclusive and egalitarian rather than exclusive and hierarchical character. Becoming Muslim was a way of overcoming a heritage of exploitative social relations and reconstituting the field for the negotiation of dependency and control. It was a way of (p.8) addressing shifts in power relations partly created by colonial rule that went almost unnoticed by colonial officials.

Religious practice and everyday life

As far as the present study succeeds in tracing out the less obvious implications of changing religious allegiance, this was made possible largely by African converts’ tendency to relate views about religious practice and change to reference points in many other areas of life. Ideas that European observers identified as religious provided ways to express and address concerns that the same observers might not have though of as religious, such as healing. Diverse social, cultural and intellectual influences played themselves out in the religious sphere, among others, because religious practice was considered relevant to them.

Religious change, then, was not only a matter of worldviews, but also of practical options. With people in many parts of Africa, Southeast Tanzanians considered relationships with spirits and ancestors crucial for human well-being. As these forces were deeply implicated in human affairs, changing one’s relationship to the spirit world was a valid way of acting in the human world. The way individuals interpreted relationships with ancestors and spirits, and increasingly with god, mattered because it defined their options in interacting with the spirit world, hence in enhancing and safeguarding their lives.22 Nevertheless, as long as being Muslim and interacting with the spirit world were considered compatible, becoming Muslim could be thought of as a way of improving, not renouncing, previously known religious practices. Religious change, in this sense, can be described as part of a search for ways to improve life. It need not be motivated by disenchantment with existing ways, but it implies an admission that things could be better.

It is telling that, as far as we know, in our region of study the spirit world was never thought of as separate from the human sphere but as deeply enmeshed with it. Spirits were seen as ever-present in the wilderness between human dwellings, and also among them.23 Moreover, spirits (p.9) continually act upon the human world: they can cause illness and remove it, they disrupt and placate, possess, ignore or support humans. Concomitantly, a person’s views and actions on one sphere are likely to have consequences also for his or her relations to the other. Two points follow that are important for our discussion. Firstly, the distinction between ‘nominal’,‘pragmatic’ or ‘skin-deep’ conversion on one hand and ‘genuine’ conversion on the other misses the point.24 There are bound to be differences in the degree of personal commitment among converts, but the ‘spiritual’ and pragmatic aspects of conversion cannot be separated. Secondly, thinking about religious questions can well support, possibly even attract a variety of this-worldly subtexts. Again, notions informing religious practice also were conceptual tools for thinking about many other things.

The present study does not dismiss the factors specified in the existing literature on the popularization of Islam in twentieth-century East Africa, but seeks to link them both to the social implications of the colonial encounter and to fine-grained local context. Thus the acceptance of Islam in the countryside was indeed an engagement with Swahili Muslim identity, part of a continuing negotiation over the boundaries of Swahili polities and the meaning of belonging to them. To understand the terms of this engagement, though, we have to look at relationships between the coast and the countryside from the rural end. Sufi leaders and other Muslim ‘missionaries’ were an important feature of the spread of Islam, but we have to look at their interaction with the people who now accepted their guidance, and to give due place to the low-profile local teachers who spread their word. Similarly, while Islam stood for ‘civilization’ of a kind, we will find that the meaning of being civilized by virtue of being Muslim was defined in a way that was inseparable from villagers’ own historical experience and local interpretations of social relations.

Clearly, then, the widespread acceptance of Islam in the twentieth century was not merely a delayed effect of the commercial and cultural contacts of the nineteenth century; rather, it was possible only through active reinterpretation of the connotations of being Muslim, which both added to and in some respects erased its pre-colonial significance. The saintly figures who are now remembered as the main figureheads of coastal (p.10) Islam in the mid-twentieth century achieved this prominence because they contributed to a new interpretation of their religion, overcoming the heritage of brutal stratification in which it had earlier been implicated. In the towns, this process was bound up with the popularization of Sufi Islam. In the countryside, it was bound up with conversion to Islam itself. In both locations, converts followed visions of Muslim polities as less hierarchical and more inclusive than they had been in the nineteenth century. Muslim identity effectively passed from being a marker of social stratification to being a marker of equality.

State, travel and local society in the production of oral records

The process of gathering oral records is bound to influence the understanding of such records. Thus, the way in which I have come to understand society in Southeast Tanzania depends partly on inexact observations made in the present time.25 In particular, interviewing confronted me with the diffuse, malleable, polycentric nature of power relationships, a phenomenon well known from the anthropology of rural Africa.26 In this context, the state was one aggregation of agencies among others, if the most powerful and invasive one. In every one of my interview sites, respondents reckoned not only with the influence of the central state (which they normally assumed to be on my side), but equally with that of local officials, fellow villagers, international aid agencies and businesspeople from town. Arguably, villagers similarly reckoned with overlapping networks of different spatial and social reach when cultivating their religious affiliations.

The conditions imposed on my work immediately placed me within the orbit of the state. Due to the exigencies of Tanzanian bureaucracy, I had to explain my purposes to local government officials wherever I went, ‘proving’ my probity with documents issued by offices in Dar es Salaam.27 (p.11) They would then, with more or less enthusiasm, set out to find me interview partners, while I went my own way to find others. The standing of the local official determined to whom he (sometimes she) would introduce me. An unpopular official might dragoon gullible individuals, ironically enabling me to speak to people who would not have volunteered as interviewees, especially elderly women. Better-connected officials would find less reluctant informants. Either way, it was impossible for me not to come across as a somewhat privileged client of the government. Generally, the government was important not only as an agency of control, but also the most important patron in the region.

Villagers met the taint of state power attaching to me with subtle yet active efforts to channel and control it within their frames of reference.28 This became evident in the fact that, concurrently through word of mouth and official channels, I would always be introduced to a specific set of local notables whom I was expected to interview. No less revealing, though, were the efforts taken by some of my interlocutors to control who spoke to me, and about what. They might try to create an opportunity for themselves or people close to them to speak publicly, to find a particularly impressive setting for their interview or to create a paper trail for it. Most evident, though, was the effort to exclude some people from speaking. Even though society in Southeast Tanzania is technically not patriarchal (as it is matrilineal), this was above all a question of gender. In the villages particularly, it was difficult to speak to women. In the presence of sons or husbands, women often declared themselves incompetent to speak or were simply sidelined. Some urban women were considerably more articulate, but these were women who had an unusual measure of social and economic independence. These gradations in women’s independence highlighted the pervasiveness of hierarchy and interdependence within each research site. ‘Local’, then, does not translate into undifferentiated ‘communal’ living.

It was, however, not simply a matter of women being silenced, but also of what they considered appropriate. Their reticence alerted me to the sensitivity of my interlocutors to their social situation. Women spoke much more freely and directly in domestic situations that were female by definition—in the back yard, while cooking. Here, they had a sphere that (p.12) was their own and where they acted with considerable assurance.29 Conversely, public gatherings (sometimes prompted by my arrival) were very much male affairs. On those occasions, it was important to recognize status and express respect (heshima), both verbally and procedurally (for instance, by offering the right chair to the right person). In the villages, against the background of material poverty, the importance of procedure and rhetoric as means to mark public occasions was particularly palpable. Together, these low-key gestures and utterances amounted to a strong code for appropriate public behaviour that recognized differences of status and authority based on gender, age, descent, relationship to the government, education, and personal accomplishment.

A major reason why one’s fellows were important lay with the mutual, but unequal dependency that permeated social relationships. Poverty is rife in rural Tanzania, and in the Southeast in particular. Yet while most people, except for a handful of state employees and entrepreneurs, are poor, there are innumerable gradations, and social relationships functioned partly as a means of mitigating one’s vulnerability. I had to get used to the relative poverty of plots and dwellings before I could appreciate the nevertheless marked differences in degree—to observe whether there was a mud or cement floor; whether there was one chicken or five chickens in the yard; whether there were holes in the thatched roof—but such things could spell the difference between dried cassava root for dinner or a cooked meal. The vulnerable faced shortages not only of staple foodstuffs, but also of water (which during dry spells in the villages had to be bought or fetched from far away) and fuel. They had to choose between washing and eating, between scrounging for food and sleeping on an empty stomach.

The resilience of the poorest was impressive, but it was clear that heshima mattered not only to the ego, but as a foundation for claims to material and practical support from others, and to the ear of administrators. Among the less marginal, too, disputes often arose over access to land, especially after somebody had put in effort to clear or plant it, and over inheritances. In these situations, the conflicting parties’ standing with neighbours and officials influenced the outcome. Here, one’s heshima was, in effect, an insurance for potential conflicts. Villagers were attentive not (p.13) only to what they owed or claimed from one another, but also to what they might be able to demand or might be asked for; not only to who had bewitched whom, but also who might bewitch whom. Such a speculative approach to power relationships was nevertheless realistic, given the uncertainties that villagers had to reckon with. Economic fortunes and the rules of state patronage changed unpredictably year after year.30

Geography added to stratification. In the outward signs of poverty, the difference between towns and villages, especially isolated villages, was great. The former contained some well-furnished homes; the latter might not count even a corrugated-iron roof. Children in town were rarely seen in the dust-coloured rags that most village children wore. Although there were at least as many signs of former as of current wealth in the coastal towns, where boarded-up buildings and gutted cars were common, the reasons why these towns used to be considered the antipode to the rough countryside were still recognizable. Meanwhile, as with poverty, there were infinite gradations in isolation. A distance of twenty minutes on foot between two villages could mean different prices for basic goods if one of the villages was on a graded road and the other was not. Yet while the poorest people in town were probably less likely to starve than those in the countryside (as the appearance of paupers from isolated villages in town during a hungry season suggested), poverty could be as grinding in towns as in the countryside and, heightened by contrast, even more demeaning.

Agriculture was precarious for most villagers. Of the four years I travelled in the region, three had poor food harvests, and two had poor cash crop harvests. This was not the outcome of proper drought, just badly timed rains and rat infestation, and did not invite particular comment. It was a function of middling soils, unpredictable rain, and tsetse and other diseases that hampered animal husbandry.31 As many, especially poorer, townspeople farmed food crops, and wealthier ones had money invested in cashew plots, these difficulties were felt in the towns as well as the countryside. The contrast between town and countryside was again tempered by parallels between them.

(p.14) Yet while heshima had material implications, this was not the only reason it was sought after. People, especially men, wanted to have authority. The term ‘networks of authority’, with authority vested in personal relationships, although coined with pre-colonial Africa in mind, fits present-day Southeast Tanzania.32 The extent of authority was malleable, its sources multiple and changing. They included connections to the state, commercial success and land ownership as well as persuasiveness or specialized (including religious) knowledge. Still, it was always the individual’s responsibility to accumulate. The multiplicity of sources of authority and the separateness of social spheres (such as, for instance, gendered ones) implied that authority was not easily solidified into all-purpose social control. On the other hand, it meant that the quest for authority was intensely personal. Their attention to public decorum notwithstanding, my interlocutors did not treat public and private situations as opposites, but as parts of a continuum. The hierarchies that came into play in public situations were based in personal relationships.33 Linkages to the state were embedded in this web of many strands.

That religious roles counted in such networks of authority was evident. Interviewees demanded and commanded respect in their capacity as shehe (religious scholar) or imamu (prayer leader).34 Challenges to established religious authority, such as have been launched since 1980 by young supporters of Islamic reform, were taken very seriously. These recent conflicts have only become possible because of conditions that have arisen since the late 1970s.35 Still, the relevance of religious practice at large and Islam in particular to these constant negotiations of dependency and control is not a recent phenomenon. The relationship has, however, always been an unpredictable and changeable one, and its permutations are an important strand in the narrative below. Initially, though, we have to ask how information derived from such a loaded context can be used to trace these permutations.

(p.15) The historicity of local views of history and religion

Given these contextual observations, the reader may doubt the usefulness of interviews as sources of factual information. The politicking around interviews implies that speaking to the researcher was seen, in the first instance, as a way of asserting one’s heshima: it was important to perform the interview process, independently of what was being said. Moreover, all informants were partisan. For instance, women discussed questions of girls’ inheritance rights under Muslim law differently from men, and local notables were liable to emphasize the history of influential elders and claim it as their own. But not every aspect of history had an immediate present-day use. Unflattering as this is for my profession, the importance of questions about local history (for instance, the succession and actions of colonial intermediaries) was not always obvious to respondents. Still, this is reassuring, as it means that the answers of respondents could not be driven entirely by present-day interest.

Moreover, even factually doubtful information could be helpful in understanding the intellectual currents and interests that have shaped interviews. Over time, interviews provided recurrent tropes and themes, and these were shaped by past debates. My understanding of the historicity of the claims of present-day informants is indebted to Steven Feierman’s account of ‘peasant intellectuals’ among the Shambaa, even though I have arrived at a very different terminology.36 Particularly important was his assertion that terms whose importance to a discourse appears constant, such as the notion of the chiefs’ ability to ‘heal the land’ and bring rain in Shambaai, can nevertheless change meaning and status, can fade, lie dormant, and recur. On the other hand, new claims arise in new situations. Even notions that respondents themselves look upon, or at any rate present, as authentic, indigenous, and timeless, are historical and context-bound.37 I take this insight as a licence to attribute certain tropes—recurring claims, particularly evocative notions—found in my interviews to particular historical junctures, using the scant written record to corroborate their appearance at certain points in time.

Current-day informants, in other words, rely on a store of manners of speaking and habits of thought that contains much of the usage of earlier (p.16) generations. Not everything in this inventory needs to be of immediate or unambiguous use; parts of it are swept along by conversation, to be redeployed one day or eventually forgotten. The most revealing tropes need not be the most elaborate and most current ones; they may be stated tentatively and may be misleading within a present context. Yet their significance comes to light by relating them to independent information on earlier periods. Ways of doing and understanding things that are now considered obvious then become discernible as having arisen from the negotiation of past historical conjunctures.38

It is also true that present-day informants can be credited with a degree of intuitive grasp of the issues faced by earlier generations. Southeast Tanzania saw enormous changes over the twentieth century, but many of the practical exigencies of life changed remarkably little. Cultivators today, including urban ones, still eat what they produce, hoe in hand. Women still cook in earthen pots; houses are still built from mud and wattle. Occasionally, outspoken villagers remarked to me on the self-sufficiency of these techniques and the limited relevance of the innovations proposed by outside advisers. Much of the knowledge handed down from the past still works, and past generations are given credit for it.

But the ‘local’ in ‘local knowledge’ is a political rather than a spatial category. Spatial locations are constituted by their relationship with a gamut of outside agencies, from district administration to central government, markets and international donors. These agencies construe rural society as ignorant and helpless, and villagers feel their dependence on them keenly. It is an irresolvable contradiction in people’s minds, but an adequate reflection of the precariousness of life in this part of the world, where self-reliance is essential but does not remedy dependency. This precariousness has been a constant over the period studied here, but was constituted differently at different points in time. The following paragraphs identify the most fundamental of the shifting points of reference within which it was negotiated.

(p.17) Changing structures: kin networks, villages and religious affiliations

Following Martin Klein, society in Southeast Tanzania can be described as ‘decentralized’: since the pre-colonial period and through to the present, outside observers have commented on the absence of ‘paramount chiefs’ or similar stable large-scale rulers.39 Yet if we take a long-term view, from c. 1880 to the present day, it is the mutability of centralized authority that is most striking, rather than its absence. During the late pre-colonial era, several leaders within the region exercised control over fairly large areas.40 Colonial warfare and the administrative structures imposed by the colonial rulers destroyed their fiefdoms, but shrewd intermediaries in later generations continued their quest for the territorial extension of authority, with occasional success.

Still, the use of terms of kinship, family and household as metaphors for social connectedness and control, runs through our sources. As elsewhere in Africa, kinship was involved in negotiations of a kind European observers would have looked for in political institutions; they were ‘a court of claims rather than a family tree’.41 The basic, if much varied, pattern of kinship relationships in this region was that of the ‘matrilineal belt’ of East Central Africa: individuals were members of their mothers’ lineages and clans; their senior male elders were maternal uncles, rather than fathers. Some other features recur in discussions of kinship relationships, without being universal: early in our period of study, husbands were less pivotal than in patriarchal families, especially when married couples resided with the wife’s kin. ‘Sorority groups’, the daughters of one mother, often kept close ties and relied on their brothers for male support.42 Although coastal dwellers recognized a patrilineal pattern, the contrast with the interior was sharp only (p.18) with Arab immigrants, as Swahili kinship was malleable and could be in effect bilineal.

Southeast Tanzania could serve as a case study of the claim that matrilinearity is most likely to be found in land-rich environments with few storable assets, where persons are the most valuable resource.43 On the face of it, matrilineality also helps explain the prevalence of decentralization. Ambitious men, when seeking to consolidate control over kin, faced a more complicated situation than in patriarchal families. Their own offspring lived with wives on whom they had relatively weak claims, rather than forming the obvious core of a kinship network within or near their own household. Yet the most important feature of matrilineal kinship relations that emerges from the literature is their compatibility with a large variety of actual political situations and the ease with which ties waxed and waned.44 In fact, matrilineages lost importance over the twentieth century, a decline related to religious change. As we will see, both occurred within a quiet process of negotiation over roles and obligations within households and villages.

Yet kinship, whether actual or metaphorical, never worked in isolation; actual social control was the outcome of kinship position combined with whatever other resources were available to an ambitious person at the time. The persons who held authority, then, could either set out from or eventually claim pivotal positions within lineages. As a discoursive legitimization for authority, the importance of kinship increased under colonialism, because colonial authorities recognized it, while other sources of power, including slavery and arms, waned. At the same time, it became more difficult to exploit. Administrators functioned as an unpredictable court of appeal in disputes over positions in kin networks, and used the malleability of these networks to impose self-interested choices of personnel. Against this background, Muslim identity provided an idiom in which social relationships could be examined and to some extent reformulated without outside interference.

(p.19) Shared settlements, varying in size from isolated extended-family settlements to large multi-ethnic clusters and towns, have continually provided frames of reference for power relations and their expression. Even if territorial control was weak and settlement impermanent, towns and villages always had their intricate, if contested, hierarchies.45 They provided the most enduring setting for structured, if not uncontroversial, social life. We can treat settlements of all sizes, villages and towns, as parts of one continuum, rather than as sites of opposed rural and urban ways of doing things. Obviously, towns, better connected and more populous than villages, were more diverse, offered more ways of expressing difference, and were arguably more complex. Still, villages were neither self-contained nor simple, and many townspeople were former villagers.

We will see that rural converts to Islam invoked their being Muslim to claim rights to participate, ritually and socially, in village life, and to limit the ambitions of local notables. Within the villages, in other words, Muslim allegiance helped ‘flatten’ social hierarchies rather than sharpen them. In other parts of British colonial Africa, religious authority has been described as an important source of social control for colonial intermediaries, as well as the prime source of challenges to it.46 Two factors limited its usefulness for the ambitious in the present case. Firstly, it was widely dispersed: the performance of an initiation ritual, the largest-scale collective ritual, required contributions from numerous different experts, including the ‘owner of the grounds’ (where the ritual was held) and the ‘owner of the bag’ (of medicines needed to protect the ritual), a circumciser for boys, instructors of both sexes, and others. Sacrifices could be performed on different scales by different representatives of a lineage. Some ritual roles were so bound to specific contexts that it is questionable whether they afforded the bearer any effective authority in everyday social relationships.

Moreover, religious expertise could be socially and spatially marginal. Healers, who were among the most powerful experts, often lived in relative isolation, acquiring their knowledge in a secret process from fellow healers and spirits. This neutral status made their judgements more persuasive in contentious cases. It also made sense because the wilderness was the source of many of their medicines and, through spirits, of their knowledge of (p.20) them. At the same time, the preference for having healers on the margins of society suggests that villagers doubted their capacity to contain their powers.47 Above all, religious authority thus conceived could also become the source of a challenge to other holders of authority, for instance when healers, who held essential but unfathomable and potentially anti-social powers, claimed leadership. Arguably, this is what occurred during the Maji Maji War at the beginning of colonialism.48

Religious authority, then, was as negotiable as kin-based authority. It did not offer an obvious means to compensate for the limitations and unpredictability of endorsement by colonial rulers. Both were involved in the renegotiation of power relationships under the conditions created by colonial rule, and both were part of an ongoing process of negotiation with outside influences.

The polyvalence of religious change, struggle, and negotiation

The preceding account of the pervasive attention given to mutual dependency and control by Southeast Tanzanians is in marked contrast to the content of informants’ accounts of the acceptance of Islam. The latter are narratives of moral deliberation, of searching for knowledge, of positive cultural and ritual change. Moreover, tension exists between the observers’ interest in identifying human, societal dynamics to change and the respondents’ readiness to accept change as something that just happened.

This contrast is a reminder of the polyvalence of Muslim allegiance and of religious change more broadly. It has already been said that ritual change could be a means of acting about one’s relationships with spirits, personal well-being, and relations with others. By extension, changing religious (p.21) observance could be a means to implicitly address social affiliations and rights. Once the colonial state had put a damper on political competition by arrogating the conferral of publicly displayed authority to itself, power relations were rarely openly adversarial. But the necessity to negotiate dependency continued under colonialism. Religious experimentation formed part of the range of options that came to hand. The actors involved typically did not explicate demands for inclusion or for specific rights; rather, shared experience brought individuals together around practices that spoke for themselves.

Jonathon Glassman has provided a compelling argument for the intrinsic connection between debates on public ritual and social struggles on the late pre-colonial Swahili coast. He shows that here (as in many pre-industrial societies), the marginal and excluded (those whom he calls ‘plebeians’) tended to try to challenge dominant definitions of rights and obligations within existing institutions, rather than these institutions themselves.49 As public ritual was the pre-eminent way of affirming existing institutions and hierarchies, marginal peoples’ quest for inclusion took the form of struggles over ritual. Glassman spoke of pre-industrial society as the site of this type of struggle, and his argument clearly cannot be taken forward into the colonial period without modification. But moral discourses, and those focused on religious allegiance, identity, and practice, continued to have implications for personal rights and entitlements.

The narrative below will, in fact, show that the people studied here continue to argue over ritual and religious practice partly in pursuit of inclusion. The rituals in question have changed over time, as have the fora where the arguments take place, and the terms of debate, which now include such formerly unheard-of categories as ‘Tanzanian’. But performative, rhetorical elements have, in different guises, continued to shape the contestations both at the local level and in interaction with state authorities. Muslim allegiance has been shaped and reshaped by interaction with these processes.

Glassman leaves no doubt that the debates he described derived their dynamism from the ‘ongoing struggles of ordinary men and women’, which continued unabated in the colonial and post-colonial periods.50 While nearly every individual life in this region was in its way a struggle—against hardship, inequity, and the untoward consequences of one’s fellows’ attempts to counter the hardships in their lives—the term is more suited (p.22) to the relatively open social arenas and partly transient participants of the towns during a period of intense exchange than the colonial period with its much narrower economic and political options. The term ‘negotiation’ better reflects the degree of practical, personal, social, and emotional interdependence among the protagonists of change in the colonial and post-colonial period, and the equal reality of cooperation between them. It should not, however, be taken to imply that the stakes were any lower among those negotiating in 1930 or 1990 than they had been among struggling urban plebeians in 1880. The gradations of contentment and bitterness found among present-day informants make this very clear.

It is also important to keep in mind the experimental, speculative elements in the process of religious (and by extension societal) innovation. Such innovation formed part of a search for ways to live with the constantly changing, but always harsh conditions in this region. It also is a subject of tentative discussion and intellectual exploration. Listening to people with very little formal education, or practice of interaction with foreigners, narrate and interpret it in sometimes very incisive and original ways has been an invaluable experience.


The preceding discussion makes clear that certain challenges and processes recurred over time in different guises. They can be briefly summarized as follows:

The parallelism of religious change in different regions and towards different denominations

The acceptance of Islam in Southeast Tanzania was part of the continent-wide advance of monotheist religion, both Christian and Muslim, in the mid-twentieth century. Christianization within Southeast Tanzania and in neighbouring regions, and the handful of other cases of rural Islamization in East Africa that have been studied, present a range of parallels and divergences. This geographic context helps situate Southeast Tanzania within the dynamics of East African history, identify changes that worked across regional and religious boundaries and, perhaps most importantly, avoid teleological interpretations.

(p.23) Changing configurations of knowledge and education

Both townspeople and villagers consistently valued knowledge, the latter well before it could be codified in writing. They ascribed far-reaching practical uses to it and accorded status to the bearer. Concomitantly, knowledge was a means for exerting control over others, but also for challenging that authority. The range of knowledge and the way in which it was canonized changed greatly over the course of the nineteenth and twentieth centuries. This was due to schooling, to the increasing use of both Arabic and Latin writing, but also to the confrontation with the technical capabilities of the colonizers. In order to understand how new Muslims adopted Islamic knowledge, we need to pay attention to this broader process of the reformulation of knowledge, and to its connection with ongoing struggles over status and influence.

Overcoming slavery, changing social identity and the reformulation of personal dependency

African converts to Islam reshaped the religion that had been instrumentalized by nineteenth-century slave owners as a means of social exclusion into a creed of free commoners. Still, social, economic and cultural differentiation did not disappear. Within the idiom of shared religious commitment that Muslims found in Islam, the differences reappeared and mutated constantly. In order to understand the adoption of Muslim identities by villagers and the arguments among townspeople about how to be Muslim, we need to pay attention to their connection to the ongoing negotiation of actual and remembered dependency.

Changing trans-regional interaction and changing local politics

Villagers and townspeople alike reasoned and still reason about power relations as a local concern, but nevertheless one that is framed by far-flung networks. Few people were unambiguously ‘local’. Their spatial and social frames of reference ranged between the Indian Ocean region, the colonial, later national territory, and commercial or kinship networks spanning some towns, villages or houses. The reformulation of criteria for control over people under colonialism, and the combination of popular and bureaucratic politics after independence, brought constantly changing conditions for aspiring power-brokers. Islam was implicated in the negotiation of these conditions because it provided terms for the ongoing debate over the (p.24) meaning of ritual as well as over geographic and social allegiances. With the recent importation of Islamist ideas, Muslims face new divisions, but they have some experience in debating politics and religion to fall back on.

The present study charts the permutations of these issues in roughly chronological fashion.Chapter 1 deals with relationships between town and countryside, traders and villagers, at the onset of colonialism, demonstrating the depth of interaction, but also the limits of the attraction of Islam for rural people at this time. Chapter 2 examines the beginnings of rural Muslim congregations during the early colonial wars, arguing that their founders were rural people working to change, rather than perpetuate, pre-colonial hierarchies.Chapter 3 examines in greater detail the oral evidence for the origins of a number of rural mosques. It argues that the congregations of the new mosques developed an understanding of ‘Muslim’ social relations that implicitly restrained the actions of the ambitious among them. Chapter 4 focuses on the role of rural Quran schools in extending and affirming Muslim allegiances, and on the way their trajectories intersected and sometimes clashed with those of mission schools.Chapter 5 examines changes in both family relationships and ritual practice, arguing that, though halting and inconclusive, change on both these planes contributed to the development of a new kind of public sphere.Chapter 6 recounts the history of the eminent coastal shehe of the mid-twentieth century, and their role in addressing both the aftermath of slavery and the changing social hierarchies in the shadow of colonial domination.Chapter 7 examines the changing significance of Muslim allegiance in the context of nationalist politics, arguing that it reinforced the nationalist loyalties of many rural Muslims, but left urban notables in an uncomfortably parochial position.Chapter 8 traces the origins and course of conflicts between ‘mainstream’ and reformist Muslims since the 1980s, in relation to national politics and international doctrinal exchanges.


(1) Ibn Battuta, The Rihla: Travels in Asia and Africa, tr. H. A. R. Gibb, vol. 2 (London: Routledge, 1956), 380–81.

(2) On the history of the Mission Benedictines in this region, see CdK 1927–76, and Godfrey Sieber, The Benedictine Congregation of St Ottilien. A Short History of the Monasteries, General Chapters and Constitutions, Biographies of its Superiors General (St Ottilien: Eos Verlag, 1992). On the Universities Mission to Central Africa, A. E. M. Anderson-Morshead, The History of the Universities Mission to Central Africa (London: The Universities Mission to Central Africa, 1909), and Terence Ranger, ‘Missionary adaptation of African religious institutions: the Masasi case’, in Terence Ranger and Isaria N. Kimambo (eds), The Historical Study of African Religion (London and Nairobi: Heinemann, 1971), 221–51. Figures on the relative strength of the different religions are scattered through this literature; see also Lindi and Mtwara, Diocesesan Offices, ‘Takwimu za jimbo la Mtwara na Lindi’ (Statistics of Lindi and Mtwara diocese).

(3) J. Spencer Trimingham, The Influence of Islam upon Africa (London: Longmans, 1968); Michael Crowder, West Africa under Colonial Rule (Evanston: Northwestern University Press, 1968), 356–62.

(4) The need to observe Islam rather than rely on textual statements is, of course, familiar to anthropologists of Islam. See Clifford Geertz, Islam Observed: Religious Development in Morocco and Indonesia (Chicago: University of Chicago Press, 1968), and the long debate he sparked off, especially the criticism by Talal Asad, in The Idea of an Anthropology of Islam (Washington: Georgetown University Center for Contemporary Arab Studies, 1986).

(5) John Iliffe, A Modern History of Tanganyika (Cambridge: Cambridge University Press, 1979); Abdul Sheriff, Slaves, Spices and Ivory in Zanzibar: Integration of an East African Commerical Empire into the World Economy, 1770–1873 (Oxford: James Currey, 1987).

(6) Martin Klamroth, ‘Ostafrikanischer Islam’, Allgemeine Missionszeitschrift 37 (1910), 477–93; Carl H. Becker, ‘Materials for the understanding of Islam in German East Africa’, tr. B. G. Martin, TNR 68 (1968), 31–61.

(7) See e. g. Merle Ricklefs, Mystic Synthesis in Java: A History of Islamization from the Fourteenth to the Early Nineteenth Centuries (Norwalk, CT: EastBridge Books, 2006); Nehemia Levtzion and Jay Spaulding, Medieval West Africa: Views from Arab Scholars and Merchants (Princeton: Markus Wiener, 2003), 29–30, 115–18. For general statements, Nehemia Levtzion ‘Toward a comparative history of Islamization’, in Nehemia Levtzion (ed.), Conversion to Islam (New York: Holmes and Meier, 1979), 1–23, and Marshall Hodgson, The Venture of Islam: Conscience and History in a World Civilization (Chicago: University of Chicago Press, 1975), vol. 2, 532–74. Ira M. Lapidus, A History of Islamic Societies, 2nd edn (Cambridge: Cambridge University Press, 2002), puts the emphasis on Islam as a force for political centralization.

(8) Many studies relevant to East African Islam deal largely with the population of the coast, often summed up under the term Swahili. For an overview see Randall L. Pouwels, Horn and Crescent: Cultural Change and Traditional Islam on the East African Coast, 800–1900 (Cambridge: Cambridge University Press, 1987). Jonathon Glassman’s Feasts and Riot: Revelry, Rebellion and Popular Consciousness on the Swahili Coast, 1856–1888 (Oxford: James Currey, 1995) is the best study of the social dynamics of which Islam formed part in the nineteenth century. For the twentieth century, August Nimtz, Islam and Politics in East Africa: The Sufi Orders in Tanzania (Minneapolis: University of Minnesota Press, 1980). Anthropological studies of East African Islam include David Parkin and Stephen Headley (eds), Islamic Prayer across the Indian Ocean: Inside and Outside the Mosque (Richmond: Curzon, 2000); Pat Caplan, African Voices, African Lives: Personal Narratives from a Swahili Village (London and New York: Routledge, 1997).

(9) The ethnographer Karl Weule gives a vivid description of his up-country Nyamwezi porters asserting their separate identity during a sojourn on the southern coast. Karl Weule, Native Life in East Africa, tr. Alice Werner (London: Sir Isaac Pitman & Sons, 1909), 28–38; a German official in a rural location near Kilwa reported that a local jumbe physically threw three ‘Arab’ travellers out of his hut when they tried to convert him (TzNA, G 9/48, 90: Kibata Sub-Station to Kilwa, 22 November 1912). There are also a handful of unpublished studies of rural Islamization that grapple with these questions: Alan Thorold, ‘The Yao Muslims. Religion and social change in Southern Malawi’ (University of Cambridge: PhD thesis, 1995); Joseph T. Gallagher, ‘Islam and the emergence of the Ndendeuli’ (Boston University: PhD thesis, 1976); Robert Louis Bunger Jr., Islamisation among the Upper Pokomo (Syracuse: Maxwell School of Citizenship and Public Affairs, 1979), and David Sperling, ‘The growth of Islam among the Mijikenda of the Kenya coast, 1826–1933’ (University of London: PhD thesis, 1988).

(10) This is Edward Alpers’s argument, in his ‘Towards a history of the expansion of Islam in East Africa: the matrilineal peoples of the southern interior’, in Terence Ranger and Isaria Kimambo (eds), The Historical Study of African Religion (Nairobi and London: Heinemann, 1971), 172–201.

(11) Glassman, Feasts and Riot, 133–45.

(12) For changes attendant upon the imposition of colonial rule in Tanzania, see Iliffe, Modern History; for the politics of the colonial state in rural Kenya, Bruce Berman and John Lonsdale, Unhappy Valley: Conflict in Kenya and Africa (Oxford: James Currey, 1992).

(13) Nimtz, Islam and Politics; the same narrative strategy is employed in Edward A. Alpers, ‘East Central Africa’, in Nehemia Levtzion and Randall L. Pouwels, The History of Islam in Africa (Athens: Ohio University Press, 2001), 303–26, and David C. Sperling and Jose J. Kagabo, ‘The coastal hinterland and interior of East Africa’, in ibid., 273–302.

(14) See below, Chapters 2 and 3.

(15) According to Nimtz’s account, it seems likely that the Sufi orders were more deeply involved in converting rural people to Islam in Central and Western Tanzania, where Nimtz’s Sufi informants in Bagamoyo, Dar es Salaam and Zanzibar had their contacts. It is also possible, though, that field work in rural sites there would turn up as many local Muslim teachers with no tangible Sufi connections, as in Southeast Tanzania. Nimtz’s research was undertaken at a time when rural Sufism had probably just reached its greatest extent, so that it was relatively easy for Sufi leaders to claim an instrumental role in the spread of Islam.

(16) We do not know very much about the interaction between the coastal towns and their environs for the many centuries before nineteenth-century commercial expansion, but it is clear that there was interaction. Justin Willis, Mombasa, the Swahili, and the Making of the Mijikenda (Oxford: Clarendon Press, 1993), Ch 1., presents information on the well-established patterns of interaction between Mombasa and its hinterland in the early nineteenth century that had not led to villagers becoming Muslim.

(17) For the growth of Christianity in Tanzania and East Africa, see Adrian Hastings, The Church in Africa 14501950 (Oxford: Oxford University Press, 1994), 519–28 and 540–567; Iliffe, Modern History, 216–36 and 543–47 and Isariah Kimambo and Thomas Spear (eds), East African Expressions of Christianity (Oxford: James Currey, 1999).

(18) A fact recognized in early studies of its history, which described pioneers of both Islam and Christianity in the region as ‘modernisers’. See Terence Ranger, ‘The Apostle: Kolumba Msigala’, in John Iliffe (ed.): Modern Tanzanians (Nairobi: East African Literature Bureau, 1972), 221–51.

(19) For a general portrayal of Indirect Rule in Tanzania, see Iliffe, Modern History, 318–341; for the late colonial period in the Southeast, see J. Gus Liebenow, Colonial Rule and Political Development in Tanzania: The Case of the Makonde (Evanston: Northwestern University Press, 1971).

(20) Jan-Bart Gewalt, Herero Heroes. A Socio-Political History of the Herero of Namibia 1890–1923 (Oxford: James Currey, 1999); and J. D. Y. Peel, Religious Encounter and the Making of the Yoruba (Bloomington: Indiana University Press, 2000), describe the incorporation of Christianity into reconstructed ethnic identities. Iliffe, Modern History, 216–37, describes the mixed effects for Christianity of its association with colonial rule in Tanganyika. The argument for the acceptance of Christianity as a step towards challenging colonialism is presented by John and Jean Comaroff in their study, Of Revelation and Revolution: Christianity, Colonialism and Consciousness in South Africa (Chicago: Chicago University Press, 1991).

(21) Some well-remembered rural proselytizers in southern Tanzania worked for the colonial government: see Gallagher, ‘Islam and the emergence of the Ndendeuli’, 122–130; interview with Shehe Ali Saidi Mkumba, Ndanda, 12 August 2004.

(22) This is not a new observation: Ranger and Iliffe have portrayed pioneers of both religions in Southeast Tanzania as ‘modernizers’. See Ranger, ‘The Apostle: Kolumba Msigala’.

(23) This is not to say that the proximity of spirit and human worlds was the reason why the former was considered to have such a direct influence on the latter. Seventeenth-century Calvinists imagined God and the Heavens as remote from humans and their affairs, but still saw a causal relation between their religious demeanour and the way they fared in the world.

(24) The wide currency of this distinction may be due to the influence of missionary notions of conversion as a conscious and deeply emotional act of submission to God. An engaging, if somewhat cynical, version of the distinction between ‘wild-eyed’ and ‘stodgily conservative’ converts is found in Richard Bulliett, Conversion to Islam in the Medieval Period: An Essay in Quantitative History (Cambridge, MA: Harvard University Press, 1979), ch. 4.

(25) I choose ‘society’ over ‘societies’ even though it was made up of sub-units that could be spoken of as societies in their own right, because the processes I will discuss depended on relations between these units; especially between Swahili towns and rural dwellers.

(26) For a classical account of a nearby region, see J. Clyde Mitchell, The Yao Village: A Study in the Social Structure of a Nyasaland Tribe (Manchester: Manchester University Press, 1956).

(27) The lore of expatriates in the Southeast has it that their numbers are small partly because local officials here were for long particularly mistrustful of foreigners. The reason given is the vicinity of Mozambique and its liberation—and subsequent civil war. The suggestion is fair, especially as Portuguese troops once planted mines on the northern, Tanzanian, bank of the Ruvuma river (interview with Mzee Hamisi Kidume, Mikindani 28 June 2004). I was occasionally confronted by local notables who felt a need to make sure that I had signed the village guest book, introduced myself to the village chairman, or otherwise deferred to authority.

(28) For a case study of rural interlocutors trying to involve a Western researcher in local power struggles, see Mitzi Goheen,‘Chiefs, sub-chiefs and local control’, Africa 62 (1992), 389–411.

(29) For the importance of the domestic sphere to women, and its vulnerability, see Megan Vaughan and Henrietta Moore, Cutting down Trees: Gender, Nutrition and Agricultural Change in the Northern Province of Zambia 1890–1990 (London and Lusaka: Heinemann, 1994), and Megan Vaughan, The Story of an African Famine: Gender and Famine in Twentieth-Century Malawi (Oxford: Oxford University Press, 1987).

(30) For a study of a recent experience of boom and bust, see Seithy L. Chachage and Joyce Nyoni, Economic Restructuring and the Cashewnut Iindustry in Tanzania: A Research Report (Dar es Salaam: Tanzania Agriculture Situation Analysis, 2001).

(31) The most comprehensive, and fairly unenthusiastic, assessment of the agricultural environment in Southeast Tanzania is found in J. G. Bennett, L. C. Brown et al., Mtwara/Lindi Regional Integrated Development Programme: The Farming Systems and Development Prospects. Report of the Zonal Survey Team in Phase 2, vol. 2 (Surbiton: Project report 67 by the Land Resources Development Centre, 1979).

(32) Lonsdale, ‘The conquest state of Kenya’, 13–44, especially 19–21.

(33) See Shula Marks and Richard Rathbone, ‘Introduction: the history of the family in Africa’, JAH 24 (1983), 145–61, for a discussion of the relation of public and private spheres in descent-based political structures.

(34) Generally speaking, a shehe is a religious scholar, that is a person knowledgeable enough to pronounce on religious issues, while an imamu is a person who leads the prayer in the mosque. Both titles, though, could also be used as honorifics for people who did not actually fulfil these social functions.

(35) Felicitas Becker, ‘Rural Islamism during the “war on terror”: a Tanzanian case study’, African Affairs 105 (2006), 583–603.

(36) Steven Feierman, Peasant Intellectuals: Anthropology and History in Tanzania (Madison: University of Wisconsin Press, 1990).

(37) Ibid., 1–12; James Giblin, The Politics of Environmental Control in Northeastern Tanzania, 1840–1940 (Philadelphia: University of Pennsylvania Press, 1992), introduction.

(38) Notwithstanding their greater reliance on orally transmitted information, it appears that people in Southeast Tanzania participate in the habit of European historians to essentialize supposedly unique traits out of the historical experience of their particular group. It is very convenient to assume that people behave in a certain way because of the way they are.

(39) The term ‘stateless’ occurs quite early in the ethnographic literature on the region; see Weule, Native Life, passim. For the advantages of the term ‘decentralized’, see Martin Klein, ‘The slave trade and decentralised societies’, JAH 42 (2001), 49–65.

(40) Rochus Schmidt, Aus kolonialer Frühzeit (Berlin: Safari Verlag, 1922), on German dealings with the Yao ‘sultan’ Machemba on the Makonde plateau; Terence Ranger, ‘European attitudes and African realities: the rise and fall of the Matola chiefs of Southeast Tanzania’, JAH 20 (1979), 63–82, for the Matola dynasty in Masasi; Weule, Native Life, on the reduced status of these rulers after the Maji Maji War.

(41) Lonsdale, ‘The conquest state’, 21. James Giblin and Gregory Maddox, ‘Introduction’, in Giblin and Maddox, Custodians of the Land: Ecology and Culture in the History of Tanzania (Oxford: James Currey, 1996), 1–14, point out the tendency to construe kinship in pre-colonial Tanzanian history as the opposite of politics.

(42) Vaughan, African Famine; for further discussions of matrilinearity in this region see Mitchell, The Yao Village; Elisabeth Colson, ‘The resilience of matrilineality. Gwembe and Plateau Tonga adaptations’, in Linda Cordell and Stephen Beckerman (eds), The Versatility of Kinship: Essays Presented to Harry W Basehart (New York: Academic Press, 1980), 359–74; Ladislav Holy, Strategies and Norms in a Changing Matrilineal Society: Descent, Succession and Inheritancy among the Toka of Zambia (Cambridge: Cambridge University Press, 1986).

(43) As argued by Mary Douglas, in ‘Is matriliny doomed in Africa?’, in Mary Douglas and Phyllis Kaberry, Man in Africa (London: Tavistock, 1969), 121–36.

(44) Douglas, ‘Is matriliny doomed’, posited the association of matrilinearity with certain social and ecological conditions. See Mitchell, Yao Village, for the elusiveness of stable authority; and Marks and Rathbone,‘Introduction: the history of the family in Africa’, 145–61, for the ability of historical actors to manipulate and change kinship ties.

(45) Many large present-day villages have only been in existence for a generation, since the ‘villagization’ programme of the 1970s. Leander Gunther Schneider, ‘Developmentalism and its failings: why rural development went wrong in 1960s and 1970s Tanzania’ (Columbia University: PhD thesis, 2003), is the most perceptive account of villagization.

(46) Karen E. Fields, Revival and Rebellion in Colonial Central Africa (Princeton: Princeton University Press, 1985), passim.

(47) Rain-making, a crucial source of chiefly authority among Feierman’s Shambaa, in Southeast Tanzania is a collective undertaking, often dominated by otherwise undistinguished women. See Peter Lienhardt, ‘A controversy over Islamic custom in Kilwa Kivinje, Tanzania’, in I. M. Lewis (ed.), Islam in Tropical Africa (London: International African Institute, 1966), 374–86. Different people often shared ritual-competence for different aspects of a ritual, but this did not make any one of them a ruler. Joachim Amman, ‘Sitten und Gebraeuche der Wamwera’ (Ndanda Mission Library: undated typescript); Bantu M Munga, ‘Unyago wa wavulana wa Kimakonde wa Tanzania [originally Tanganyika]’, (University of Dar es Salaam Library, typescript, c. 1955).

(48) For oral accounts of Kinjikitile, the ‘prophet’ of the maji medicine, see the records of the Maji Maji research project (MMRP), for healers and diviners as outsiders, Joachim Amman,‘Sitten und Gebräuche’; Alkuin Bundschuh, Ostafrika: Land und Leute. Aufzeichnungen aus der Sicht eines langjährigen Missionars (St Ottilien: Eos Verlag, 1976), 233–67.

(49) Glassman, Feasts and Riots, introduction.

(50) Ibid., 270.