THE MUMIANI SCARE OF 1945: THEORETICAL APPROACHES

In his annual report for 1945, Digo District Commissioner (DC) Noel Frederick Kennaway describes a period of acute public panic at the end of that year:Footnote ¹

Kennaway's brief report is the only archival trace of the 1945 panic, and the only hint of a local scare ‘some seven years’ earlier, although Peter Pels notes that one did occur in Morogoro District, Tanganyika around that time.Footnote ³ Neither episode stands out in oral histories of Mumiani activity in the region. How to approach such a fragmentary account of such an unusual event?

Fortunately, shadowy figures like Watu wa Mumiani (and the occasional episodes of panic about them) have been important sites for innovative theoretical and methodological work in African history in recent decades. These ‘vampire’ figures are, as Luise White has persuasively argued, an ‘epistemological category, with which Africans described their world’.Footnote ⁴ The stories told about them, however, ‘are not constructed on a moment-to-moment basis’, but rather ‘drawn from a store of historical allusions … kept alive and given new and renewed meanings by the gossip and arguments of diverse social groups’.Footnote ⁵ By treating formulaic elements of Mumiani stories not as distortions or distractions but rather as the key ingredients of ‘a good and thus credible story’, White directs our attention to something like what Nancy Rose Hunt calls ‘the lexicon of debate’.Footnote ⁶ In this approach, identifying the key elements ‘kept alive’ and — importantly — transformed by rumor helps reveal ‘wider terrains of belief and theory’.Footnote ⁷

James Brennan has made a complementary argument with reference to Mumiani phenomena, about the selection and prioritization of ‘explanatory contexts’ for historical events with no clear cause or effects.Footnote ⁸ Analyzing an anti-Mumiani riot in Dar es Salaam, Tanganyika in 1959, Brennan notes that the interpretation of such events often ‘raises demands for speculation on evidence that is generally evocative but specifically elusive’.Footnote ⁹ Any plausible explanation of events like these demands inference and interpretation based on indirect, contextual clues, when ‘no conclusive cause or causes are evident, but some are more likely than others’.Footnote ¹⁰

Illuminating as these two approaches are, there are potential risks associated with each.Footnote ¹¹ On the one hand, White's innovative historical method risks re-inscribing the very opposition between reality and belief that it seeks to overcome. White ‘takes these stories at face value’ to see what they reveal about colonial experience, but ultimately understands herself to be writing ‘the history of things that never happened’.Footnote ¹² To use these stories as evidence in this way, one may treat them ‘as if they were true, or accurate’, but keep in mind that ‘the stories are false’ and that ‘their very falseness is what gives them meaning’.Footnote ¹³ Brennan's method, on the other hand, risks imposing an external and purely analytic distinction — one perhaps not shared by the historical subjects themselves — between ‘event’ and ‘context’ in order to posit an explanatory, causal relationship between the two. If White is interested in showing how and why these ‘vampire stories’ are so widespread ‘even though they do not depict actual events, conversations, or things that really happened’, Brennan is interested in explaining how ‘a rumour that is ubiquitous relate[s] to a violent event that is rare’ — but at the risk of losing sight of why those rumors were ubiquitous in the first place.Footnote ¹⁴

I want to argue that in the case of the Digo District Mumiani scare of 1945, a more dialectical approach — one capable of reconciling White's methodological emphasis on the content of rumors with Brennan's focus on context of events, while avoiding the risks associated with each — will help us arrive at a better understanding, first, of why these Mumiani stories re-emerged when and where they did, and second, what their renewed salience then and there tells us about how that moment was understood by the people involved. Later, I will suggest that Walter Benjamin's concept of a historical ‘constellation’ is useful in this regard. But such an analytic cannot adequately be employed until all of the relevant empirical material — the elements of the constellation — have been presented. Let me begin, then, by tracing the metamorphoses of ‘Mumiani’ and its referents over the nineteenth and early twentieth centuries. This will show how the term was ‘kept alive and given new and renewed meanings’, in White's words, ‘by the gossip and arguments of diverse social groups’, rather than positing an explanatory link between its Arabic-language origins and its significance in World War II-era East Africa.Footnote ¹⁵ Following this, I return to DC Kennaway's speculative explanation for the episode's origins in the ‘two mysterious murders’ in Kilifi District to the north. Finally, I explore the historical association of famine with the pawning of kin for food loans in the nineteenth century, and argue for the importance of a contemporary drought, labor conscription campaign, and provision of ‘famine relief’ as more salient ‘explanatory contexts’ for the return of Watu wa Mumiani to Digo District in 1945. This is not the ‘background’ to the event. Rather, I am arguing that the panic was as much ‘about’ these earlier forms as it was about more proximate contextual ‘triggers’.

MUMIANI IN WORLD HISTORY

As Kennaway points out, the history of Mumiani is, in the first instance, the history of a substance referred to as mumia, momiyai, mumie, and mummy. The etymology is variously given as the Arabic or Persian mum (‘wax’), amomum (a perfume), or mumia (an embalmed body).Footnote ¹⁶ In the course of translating Pliny, Dioscorides, and Galen from Latin and Greek into Arabic, and later Rhazes, Serapion, and Avicenna from Arabic into Latin and Greek, uncertainty emerged over whether ‘mumia’ referred to preserved and perfumed corpses or a waxy (or bituminous) ingredient (or exudate) of them.Footnote ¹⁷ The medical communities of early modern Europe and the Indian Ocean remained divided over whether mumia referred to bodies or bitumen, whether one was an acceptable substitute for the other, and whether either had any medicinal value at all.

By the mid-nineteenth century, however, mumia (and cognates across the Indian Ocean world) had come to refer not to a substance derived from Egyptian mummies, but to one created from the bodies (or bodily substances) of people recently captured and killed for its manufacture. In 1854, for instance, Alexander Cunningham wrote that ‘the common momiai of Indian medicine is of course a manufactured article; although not made, as generally asserted, of the melted fat of Abyssinian boys, who have been roasted for the purpose’.Footnote ¹⁸ A footnote in Raja Sivaprasad's History of Hindustan from 1874 includes ‘among the nursery tales we heard in our childhood that in Habash (Abyssinia) men make Momiyái, a sort of medicine, from the blood of the slain’.Footnote ¹⁹ Another colonial Indian source (from 1885) describes ‘mimiyai’ as ‘a kind of bitumen … said to be extracted from the heads of coolies who emigrate to the colonies, by hanging them head downward and roasting them over a slow fire’.Footnote ²⁰ And in East Africa, the Mombasa-area missionary Johann Ludwig Krapf's 1882 Dictionary of the Suaheli Language defines Mumiani as ‘a fabulous medicine which the Europeans prepare, in the opinion of the natives, from the blood of a man’.Footnote ²¹

This newer conception, however, had not fully replaced the older one of a substance derived from preserved corpses. In his 1870 Handbook of the Swahili Language as Spoken at Zanzibar, Edward Steere defined ‘Mumyani’ as ‘A mummy. Mummy is still esteemed as a medicament’, and Charles Sacleux's 1891 Dictionnaire Français-Swahili gives ‘mumiani’ as the Swahili translation of the French ‘Momie’ (‘Mummy’) with no further elaboration.Footnote ²² His 1939 Swahili-to-French dictionary, however, seems to assimilate these earlier definitions of ‘Mumiani’ to what would have been, for him, a more familiar genre of European blood-stealing rumor, namely blood-libel: ‘Mummy. Used in the medicine of the indigenous, who say that it is made of dried human blood: Jews would be responsible for obtaining it by bleeding subjects’.Footnote ²³ In that same year (possibly drawing upon material appearing in the journal Man between 1929 and 1934), Frederick Johnson published a definition of ‘Mumiani’ as ‘a dark-coloured gum-like substance used by some Arabs, Indians and Swahili as a medicine’, detailing its indications (‘cramp, ague, broken bones, &c.’) and modes of use (either ‘as an outward application’ or ‘melted in ghee for drinking’).Footnote ²⁴ Johnson adds, ‘many natives firmly believe that it is dried or coagulated human blood taken from victims murdered for the purpose, and when a rumour is started that mumiani is being sought for, the natives of a town are filled with terror and seldom go out of their houses after sunset’.Footnote ²⁵ The Indian Ocean world of the late-nineteenth and early-twentieth centuries was thus the scene of protracted disagreement, debate, and reimagination of the nature, origins, and merit of medical techniques and technologies, including Mumiani. And among the ‘diverse social groups’ re-vivifying and re-imagining this concept were not only the Arab, Indian, and African inhabitants of the East African coast, but a host of European missionaries, administrators, and social scientists as well.

There is good reason, then, to question Pels's ‘neo-diffusionist’ (but really just ‘historical’, in a linear sense) argument that this was an ‘Indian rumor’ brought to East Africa by Indian soldiers and police in the 1890s.Footnote ²⁶ Pels's article has the virtue of attempting to account for the timing of an East African Mumiani panic in 1904, but the claim hinges on the rough contemporaneity of rumors about ‘Mumiani’ extraction by Indians in East Africa (in 1904–6) with rumors about ‘momiai’ extraction from Indians in British hospitals in India during the plague epidemic of 1896.Footnote ²⁷ Earlier Indian accounts cited above already identified momiai extraction as an ‘Abyssinian’ phenomenon, however. Stories about Mumiani killings, it seems, were not an ‘Indian rumor’ at all, but a richly imaginative, highly adaptable body of images and ideas — ‘historical allusions’ — shared by a broad range of social actors and capable of expressing international concerns about the sources, availability, and effectiveness of foreign medical technologies on the one hand, and the fate of individuals swept up into global circuits of unfree labor on the other.Footnote ²⁸

‘MYSTERIOUS MURDERS’ AND RITUAL KILLING ON THE KENYAN COAST

Like Pels, Kennaway also explains Mumiani rumor in terms of a roughly contemporary but slightly prior event: the ‘two mysterious murders in Kilifi district’. Although The Mombasa Times reported that the ‘naked body’ of ‘a murdered native … was found in the bush on the mainland at Mikindani’ in late September, and that ‘the man's throat had been cut and he was otherwise brutally mutilated’, I have found no reference to a body ‘found hanging by its feet in a tree’ in Kilifi District records.Footnote ²⁹ This would seem to raise the question of whether Kennaway was reporting as the ‘origin’ of the rumor information related to him in an instance of its telling.Footnote ³⁰ On the other hand, there is evidence of ‘an inquest … into the death of a Giriama whose mutilated body was found in a pond at Kaloleni’ in September.Footnote ³¹ Determined to be a case of ‘murder by a person or persons unknown’, this is presumably the body Kennaway describes as found ‘face down in a puddle’.Footnote ³² The only detail common to both bodies, according to Kennaway, was their postmortem mutilation ‘in the Boran and Abyssinian fashion’ (a polite way of indicating that the victims’ genitals had been cut off). Such mutilation is not described in any available sources as characteristic of rumors about Mumiani killings from this period, however — unless we take the ‘body found hanging from a tree’ to be an instance of just such a rumor.Footnote ³³ Yet it is on the basis of this ‘curious’ detail that Kennaway links the incidents in Kilifi District to the ‘curious belief’ in Digo District. While offered as an ‘explanatory context’ for the timing of the 1945 Mumiani scare, reports of these deaths may have been more significant for an ‘historical allusion’ that would have called up a rich set of images and associations for Digo District residents, but which would have been lost on Kennaway.

Although descriptions of unnamed East African peoples emasculating the victims of initiatory killings can be found in twelfth-century Chinese encyclopedias (and in early seventeenth-century accounts, in which they are identified as ‘Mosseguejoes’), nineteenth-century missionary texts attribute the practice to ‘the Galla’, Oromo-speaking peoples (which include the Borana) of the northeast Kenyan borderlands.Footnote ³⁴ Church Missionary Society catechist George David reported in 1877, for instance, that ‘the Gallas once had a law that no man should have a wife until he had killed a man, or at least produced his privy parts, and for this barbarous purpose they used to catch the Wanika, which caused them to leave that part of the country’.Footnote ³⁵ It is unclear how far in the past David understands this ‘law’ to have been in effect, or when the practice had stopped, but there is another important nineteenth-century association on the Kenyan coast to ritual killing and mutilation: Mung'aro, a ritual of Mijikenda male initiation.

In 1845, Krapf describes Mung'aro as ‘a horrible sport, practiced from time to time by the young people when they have reached a certain age. … [T]hey remain in the woods until they have killed a man … [S]olitary travellers, especially slaves, are their favorite prey.Footnote ³⁶ Krapf advised travelers ‘to ask after the time of the Wagnaro [Mung'aro] of a village, as he might risk his life, if alone on the road during its observance’, because the ritual ‘cannot terminate unless they have slain somebody in the fields, or bought, by common contributions, a slave, whom they may kill’.Footnote ³⁷ In his opinion ‘the Wadigo … appear to be especially attached to this cruel practice’.Footnote ³⁸

Although absent from nineteenth-century texts, mutilation of the victim's body became a standard element of twentieth-century descriptions of Mung'aro (at which point it was no longer being performed). Take, for example, the description of Methodist missionary J. B. Griffiths: ‘They formed themselves into parties and went in quest of the initiation victim, who had to be a man of an enemy tribe, and whose right hand and sexual organs they had to take to the elders’.Footnote ³⁹ By the time Kennaway was reporting the ‘curious’ deaths in Kilifi District as the origin of Mumiani rumors in Digo, in other words, mutilation ‘in the Boran and Abyssinian fashion’ — collection of genital trophies in the course of an initiatory ritual killing — seems to have been fully incorporated into Mijikenda understandings of their own ritual history.

What these twentieth-century accounts of killing and ‘mutilation’ during Mung'aro share — and share with twentieth-century Mumiani rumor — is a clear sense of the vulnerability of individual travelers along roads and paths. Mung'aro and Mumiani also relate this anxiety to political power: Mung'aro involves a ritual killing (and in later accounts, mutilation) at the hands of senior men as part of their initiation into precolonial political office. Mumiani beliefs highlight such periodic killings as an element of the colonial ‘Government’ power that organizes it. And as the political ritual of Mung'aro faded from practice at the end of the nineteenth century and began to include the collection of body parts for the ritual use of secretive elders, Mumiani imaginaries changed as well, from a medicine made from African blood by Indians and Europeans, to the gangs of Africans said to be employed by them.Footnote ⁴⁰

Nineteenth-century accounts of ritual killing and mutilation share another important theme: the violation of broader norms governing patronage and dependent (or unfree) labor. Most sources mention a slave as an acceptable alternative to ‘a stranger’ in Mung'aro.Footnote ⁴¹ But David's report includes an additional reference to the vulnerability of enslaved peoples and dependents. After the Mijikenda abandoned their northern territory, according to David, escaped slaves from the coast took their place, putting themselves ‘under Galla protection and render[ing] tribute in return’.Footnote ⁴² But David suggests that by the late nineteenth century, this arrangement was breaking down. Speculating about the consequences of growing British presence in the region, the Oromo had begun to kidnap and sell back into slavery those who had once sought refuge in their patronage.Footnote ⁴³

The themes of slavery, subordination, and betrayal of the terms of clientage or dependency provide important clues to the timing of the 1945 Mumiani panic. Without undertaking a detailed analysis of the nineteenth-century regional slave economy (which has been carefully analyzed by others), a brief discussion of certain key dimensions is thus in order before returning to the 1945 Mumiani episode.Footnote ⁴⁴ In what follows, I focus on ‘pawning’ relationships, the importance of kinship as an idiom or ideology of power and patronage, and the role of regional famines in producing a mobile and vulnerable pool of refugee labor through contractual forms like pawning.

SLAVERY, PAWNSHIP, AND AUTHORITY

In the second half of the nineteenth century, a series of treaties between Britain and Zanzibar slowly restricted the exportation of slaves from East Africa into the broader Indian Ocean world. Limitations on slave exports meant that the number of enslaved people available locally for purchase increased until the overland trade in slaves (primarily from present-day Tanzania and Malawi) was abolished in 1876.Footnote ⁴⁵ During this period, slave raiding in the coastal hinterland also increased, and the paths connecting fortified settlements became increasingly threatening spaces of potential abduction and enslavement.Footnote ⁴⁶

Following a brief decline in the wake of the 1876 treaty, ‘a substantial revival of the slave trade took place’ in 1884, coinciding with a major famine remembered in southern Kenya as mwakisenge.Footnote ⁴⁷ The mwakisenge famine affected much of the coastal hinterland and interior, and lasted for several years. ‘Faced with starvation’, many southern Mijikenda ‘pawned themselves or their children to the coastal plantation owners’.Footnote ⁴⁸ This was, however, only the first in a series of disastrous famines and food shortages to coincide with the imposition and first decades of European rule.Footnote ⁴⁹ Importantly, several of these early twentieth-century famines and food shortages — kodi kwanza (‘first taxation’), rupia mbirimbiri (‘two rupees each’), rupia tahutahu (‘three rupees each’), faini (‘fine’), pishi mwenga (‘[one] pishi [of grain for] one [rupee]’) — are named after contemporaneous extractive arrangements (in taxation, livestock confiscation, or ‘famine relief’) imposed by the new colonial administration. This expresses a clear understanding that the causes of endemic famine were as much social and political as they were ecological.

Persistent and recurring food shortages during this period put already marginal individuals in a precarious situation: At risk of capture while out in search of food, their situation within the homestead grew increasingly tenuous as potential pawns. ‘Pawning’ (kuweka rahani) refers to an arrangement in which kinship-based rights in people could be transferred temporarily in exchange for a loan of food or cash in times of hardship. The pawn's labor during the time of their pawnship served as interest on the loan, and they could (ideally) be ‘redeemed’ at a later date by repayment of the original loan of food or cash. Children were especially at risk of being kidnapped or ‘pawned’ by senior male kin to settle a debt, serve as security for a loan, or in exchange for grain to sustain the rest of their homestead.Footnote ⁵⁰ And as Justin Willis and Suzanne Miers have shown, ‘the increased availability of bought slaves in the latter part of the nineteenth century enabled “new men” to reduce the autonomy of all their dependents’.Footnote ⁵¹ Kinship, however, remained the shared ‘idiom’ of excision and incorporation in such circumstances.Footnote ⁵²

Non-kinship-based frameworks of authority in precolonial Mijikenda society included a male age-grade system.Footnote ⁵³ Mung'aro, described earlier, was the ritual of initiation from nyere, ‘youth’, into ngambi, the ‘adult’ status entitling men to sit in councils (also called ‘ngambi’) with the power to allocate land, adjudicate disputes, and impose fines for legal and ritual violations.Footnote ⁵⁴ But by the early twentieth century, partly as a result of the dramatic transformations of coastal society over the calamitous nineteenth century and partly due to the administrative measures of the Imperial British East Africa Company and British Government, the system of age-sets and ngambi councils eroded and eventually lapsed entirely.Footnote ⁵⁵ Power had by then become highly localized: ‘Homestead heads developed new ritual forms which emphasized their autonomy and their own authority’ and ‘local councils of homestead heads usurped the prerogatives of … elders in the dispensation of justice’.Footnote ⁵⁶ When the Kenyan coastal strip was reconstituted as a British Protectorate in 1895, the ‘new men’ who had risen to prominence in the late nineteenth century were quickly recognized as the administration's best hope for a supply of labor in a reorganized coastal economy, and it took steps to strengthen their position with respect to the labor question.Footnote ⁵⁷

By creating new, state-backed forms of power and corresponding incentives for junior men to dodge them (typically by abandoning the Reserves to become either casual laborers in Mombasa or squatters on defunct coastal plantations), these measures aggravated existing inter-generational tensions and introduced a new source of intra-generational conflict, as Willis points out.Footnote ⁵⁸ At the same time, there is evidence that headmen and elders would actively avoid the administration's labor demands to prevent the flight of women and junior men from their homesteads.Footnote ⁵⁹ These conflicts over labor — who had the right to dispose of the labor power of others, who the right to compensation, and what the nature of the relationship between parties was — all have roots in precolonial relations of patronage, dependency, and indebtedness that included the pawning, abduction, and sale of individuals in times of scarcity, transacted through the idiom of kinship. The durability of these precolonial concerns and their transformation in the early colonial period are key to understanding the nature and timing of the Mumiani panic that unfolded in Digo District in 1945, coinciding as it did with the failure of the rains on the one hand, and a large-scale wartime civil labor conscription scheme on the other.

‘THE FAMINE OF WHEAT’ AND ‘THE YEAR OF TAVETA’

In Digo District (now Kwale County), 1944–5 is remembered as the Mwaka wa Taveta (the ‘Year of Taveta’).Footnote ⁶⁰ Between June 1944 and October 1945, the colonial administration conscripted over 1,500 Digo men (with hundreds more summoned, who either refused to appear or were rejected on various grounds) for labor on sisal farms and an irrigation project in Taveta District.Footnote ⁶¹ Although conscription was formally conducted through the relatively new administrative apparatus of chiefs and headmen — the ‘native authorities’ empowered by the administration to supply conscript labor — in practice it was effected, like pawning, through kinship-based relations of authority and subordination within and between homesteads. While monthly intelligence reports detail the numbers of conscripts, summonses, rejections, and Tribal Police required for enforcement, what emerges most clearly in oral histories of the Mwaka wa Taveta is pervasive conflict between men and their brothers-in-law over the right to ‘pawn’ their respective sons and sisters’ sons to the colonial government.Footnote ⁶² What was from the government's perspective a bureaucratic administrative exercise was, from the perspective of Digo District residents, the imposition by the government of a form of debt-bondage that it had forbidden them to contract amongst themselves.Footnote ⁶³ Importantly, the Mwaka wa Taveta coincided with the Nzala ya Ngano, ‘The Famine [literally, “Hunger”] of Wheat’, the famine taking its name from the edible form in which ‘famine relief’ rations were provided. In local understanding, ‘Government’ was using the position it had spent the last fifty years establishing at the apex of coastal patron-client networks to demand labor in exchange for the food aid it supplied to the drought-affected regions in 1944–5.Footnote ⁶⁴

Informants disagreed over whether fathers or maternal uncles had the authority to force ‘their’ young men into the labor contracts for work in Taveta.Footnote ⁶⁵ Most asserted that the recruitment was undertaken on a homestead by homestead basis — that the District Commissioner or District Officer would hold a public meeting announcing the labor requirement for that month, then send Chiefs or Tribal Police to collect a young man from each homestead, for whom the homestead head was responsible.Footnote ⁶⁶ Who a homestead head might be able to send — a son or a sister's son — thus depended on the composition of the homestead in each instance, and his ‘strength’ and influence relative to that of his brothers-in-law. Two informants claimed that the demand for labor during the Mwaka wa Taveta put such strain on the structural conflict between fathers and mother's brothers over women and children — what Audrey Richards famously called ‘the matrilineal puzzle’ — that it ‘broke’ the precolonial system of inheritance from one's mother's brother (ideologically grounded as it was in his ‘right’ to pawn his sister's son).Footnote ⁶⁷ Another informant also made the connection between conscription and pawning explicit, adding that, like pawning, conscription was not expected to be permanent. Young men and their families hoped that they would return home when the contract was finished (though as with nineteenth-century pawns, not all did).Footnote ⁶⁸ Some compared the work to slavery, claiming that the workers were not paid, just fed and clothed.Footnote ⁶⁹ Others denied this, on the grounds that the government had ended slavery and could not simply take people for work without pay.Footnote ⁷⁰

Much of the conscript labor was allocated to private estates. This situation ‘was accepted only with reluctance by His Majesty's Government’, and the question of its necessity was ‘raised from time to time in the House of Commons’.Footnote ⁷¹ There were concerns in the metropole about ‘the exploitation of compulsory labour for private profit’, including the retention and letting out of conscripts for work in the vicinity of the camps on projects not considered to be ‘essential undertakings’ for the war effort.Footnote ⁷² Wages notwithstanding, the ‘compulsory labour scheme’ was a disconcertingly unfree form of labor not only to Kenyan conscripts, but to the British Government as well.

This, then, was the socio-historical situation in which the Mumiani scare of 1945 unfolded. But what is the link between these forms of unfree labor and labor recruitment and local ideas about the Watu wa Mumiani? To return to the earlier discussion of White's ‘store of historical allusions’ and Brennan's ‘explanatory contexts’ — and to demonstrate the utility of Benjamin's notion of ‘constellation’ for grasping the historical consciousness of this moment — I want to argue that the famine and conscription campaign of 1944–5 were both. To be clear, I am arguing against an interpretation that would see famine and conscription as being what the Mumiani rumors were ‘really about’. Watu wa Mumiani were not an allegory for (or reflection of) other forms of extractive social relations. But they were molded by the experience of those other forms (which is the very reason the rumors were plausible to begin with) and were understood to be part and parcel of them. The claim of a link between certain extractive relations and Mumiani is not solely mine, in other words. Rather, as will be shown below, the connection is posited by southern Mijikenda themselves in a variety of ways. This aspect of the Mumiani panic gives us the clearest picture of the historical consciousness of that moment — how it was understood as one episode of more general social vulnerability and stress in a series of such episodes — and it is to these explicit local links between Mumiani and the history of pawning and famines that I now turn.

PREDATORY PATRONS AND WATU WA MUMIANI

The association of Mumiani with asymmetrical relations (like pawning) contracted with wealthy coastal patrons was not new. Recall that in the later nineteenth century, when the region was plagued by a series of famines that resulted in the pawning of thousands from the hinterland, Mumiani was the name of a powerful medicine made by and for the use of coastal Arabs, Indians, and Europeans out of human blood. This was not a metaphor for the unequal relationships established with representatives of these stereotypically landowning, moneylending, and political groups, but an assertion about these figures that expressed popular concerns about the perils of contracting such desperate arrangements with them.

There is direct evidence of patrons-as-Mumiani from the decades leading up to the 1945 panic, but with a twist: ‘The chief mumiani of the 1930s’, according to Willis, ‘is said to have been Muses Muhammad’, also known as ‘Chuba Muses’.Footnote ⁷³ Willis's interlocutors describe Muses as a Mombasa-born and Arab-dressing — but ethnically Kauma (one of the Mijikenda peoples) — businessman and predatory lender ‘whose incorporative networks of credit stretched into the hinterland’.Footnote ⁷⁴ Muses was also, according to one interviewee, a founder of the Mijikenda Union, launching the political organization to protect his ethnic compatriots from being robbed of their land by Arabs.Footnote ⁷⁵ As the interview progressed, however, the picture began to change. Before his death, the man's father had mortgaged a coconut plantation to Muses, who was bullish and dismissive when his heirs tried to redeem the property at the agreed-upon price.Footnote ⁷⁶ Contracting these sorts of usurious relationships with cash-strapped but landholding Mijikenda was, according to Willis's informant, Muses's ‘primary occupation’.Footnote ⁷⁷ But he was also said to ‘take other people and make them into slaves’.Footnote ⁷⁸ ‘Back then’, Willis's interlocutor continued, ‘it was like he was Mumiani. … Mumiani is a person who catches people and cuts their throat. … Mr. Muses Muhammad, people were saying he was the real boss of them’.Footnote ⁷⁹

Another of Willis's informants steered a discussion of Mombasa workers’ housing in his father's time to the issue of personal security, and from there to the threat of capture by gangs employed by ‘Chuba Musis’.Footnote ⁸⁰ His explanation for the occasional disappearance of workers in early twentieth-century Mombasa shifted subtly from gangs of slave catchers to gangs of Watu wa Mumiani, but ‘Chuba’ Muses Muhammad — wealthy coastal patron, Muslim but Mijikenda, insider and outsider — was said to have organized them both: ‘if a person got lost here, they had been captured by the forces of Chuba Musis’.Footnote ⁸¹ Initially claiming that Muses ‘sold people at Zanzibar’ and would send out a ‘force’ to obtain them, the alleged purpose for which they were captured changed.Footnote ⁸² From 1936 onward, he claimed, the government would announce when there would be Mumiani, and ‘if you're caught out on the road by Mumiani, that's your affair. … When that time arrived, it was the troubles of Chuba Musis’.Footnote ⁸³ In short, the man many believed to have founded the Mijikenda Union to promote the interests of his co-ethnics also betrayed them, capturing them at night either for sale abroad as slaves or to kill them and drain their blood. On the one hand the link between these two forms of predatory patronage is diffuse and associational: Thinking about work in interwar Mombasa led Willis's informant to think about enslavement and sale in Zanzibar, which led him to think about how ‘Government’ allowed Watu wa Mumiani to operate freely. On the other hand, the link is direct and concrete: The same historical figure was said to be engaged in both.

White's work on colonial ‘vampires’ emphasizes their qualities of novelty and alterity as a way colonized Africans would ‘speak about the demands of their rulers’.Footnote ⁸⁴ But what the Mumiani panic of 1945 and the example of Chuba Muses articulate is a concern less about colonial ‘rulers’ directly than about intermediary figures who mediate between foreign powers and vulnerable communities. In other words, while the fear that seized Digo District in 1945 did mark the alterity of a source of power and its demand for blood (locating it in ‘Government’), it articulated much more acutely an anxiety about the Africans who worked for or with those outsiders, performing the abductions and killings on their behalf. It not only parallels but seems strongly shaped by contemporary concerns about the pawning of young men by senior male kin and ‘native authorities’ in service to the foreign power that backed them. One old man, for instance, claimed that the Mumiani rumors of 1945 had been started by escaped conscripts.Footnote ⁸⁵ Though skeptical that Watu wa Mumiani had been abducting people at the time of the Taveta conscription, he was certain that such people did (and indeed, still do) exist. To him, these stories were a way for men like his father's brother to explain being sent from home by old men to do strange work in a strange place for an uncertain length of time at an uncertain rate of pay.Footnote ⁸⁶ It was precisely due to the historical association of Mumiani with the perils of patronage that these stories were plausible. In White's words, such stories ‘were not hysterical accusations but concerns and interests grounded in local historical experience. They do not “explain” the rumor, but they explain how it was locally credible’.Footnote ⁸⁷

Finally, Mumiani is explicitly associated with drought in the region, not as a metaphor for this more general desiccative process, but as a claim about its social and physiological effects. An analytic claim about a relationship between the two would not itself be novel. Erin Pettigrew, for instance, has recently drawn attention to the importance of forms of unfree labor and drought-related social stress for understanding bloodsucking accusations in colonial Mauritania.Footnote ⁸⁸ What stands out in the coastal Kenyan case is that the relationship between drought and blood-stealing is itself an explicit dimension of the local theory of why Watu wa Mumiani operate when and where they do. According to Digo and Duruma, the environmental condition of drought thins the blood and accelerates its movement through the body. The resulting condition is experienced as one of both anxiety and lethargy. Not only is blood thinner (and thus easier to drain), but food prices go up and money becomes scarce. And as one woman explained, Watu wa Mumiani are aware of these effects and readily take advantage during ‘times of intense sun (wakati wa jua kali), when there is no money’.Footnote ⁸⁹ Under such circumstances people are willing — reluctantly, as they were in the past — to engage in various unsavory practices to survive. In the nineteenth century, these included kidnapping, pawning, or sale of children to outsiders for grain or cash. In the twentieth century, these strategies are understood to include cooperation with Watu wa Mumiani. Periods of drought have thus become, to use a phrase I encountered repeatedly during a 2013 panic in the same area, ‘Mumiani season’ (musimu wa Mumiani). In the local imagination, Watu wa Mumiani are pragmatic, rational actors maximizing profits and efficiency, operating — like predatory patrons of the nineteenth century — within a regionally shared understanding of how environmental conditions affect human physiology and moral psychology. And in this way, the episodic nature of Mumiani predations — like the threat of pawning — is conceptually linked in local thinking to the periodicity of drought and the desperation it can engender.

The claim of a connection between Mumiani, drought, and the patrons to whom one is vulnerable under such conditions, is thus both an analytic and an ethnographic one. If the contemporaneity of the Mwaka wa Taveta conscription campaign and the Nzala ya Ngano famine are, in Brennan's terms, the most likely ‘explanatory contexts’ for the Mumiani scare of 1945, it is only because, in White's terms, they ‘suggest genealogies of local concerns and historical fixations that would not otherwise be apparent’.Footnote ⁹⁰ In this sense, the 1945 Mumiani panic can be seen as an historical ‘constellation’; a moment in which elements of the past and the present — between which there may be no necessary or causal links — are seen to come together suddenly in a meaningful new configuration.Footnote ⁹¹ The particular ‘constellation’ of this moment was made up of elements as diverse as blood medicine and labor conscription, mutilated bodies and matrilineal descent, roads and rainfall. Approaching it in this way allows one to interpret the entire episode without trying to make such a ‘rich text of half truths and local knowledges’ be ‘about [just] one thing’, as White cautions.Footnote ⁹²

Focusing on both the direct and more diffuse associational links that the Mijikenda themselves draw between Mumiani, drought, and predatory patronage practices involving the transfer of rights in people also allows us to relativize our definitions of event and context in this case. Were the contemporaneous Ngano famine and Taveta conscription the ‘contexts’ for the Mumiani panic of 1945, or do Watu wa Mumiani form part of the context of these other events, informing popular understandings and responses, as a durable (if also malleable) feature of famine and debt-bondage in southern coastal Kenya since the nineteenth century? The answer, of course, is both, meaning that the empirical material itself calls, in this case, for a dialectical approach like the one I propose. Such an approach interprets the complex, non-linear, ‘constellation’ of meanings at play in Digo District in the dry, waning months of 1945 without flattening them out into metaphorical beliefs for something more ‘real’. Neither does it impose an outside analytic distinction between event and context in order to then explain the former in terms of the latter. Instead, it allows one to study such moments as the conjuncture of multiple forms of historical knowledge, practices, and images with different temporalities and associations, and to show how such arrangements draw the past implicitly and explicitly into the present, shaping the very concerns they articulate.Footnote ⁹³