This article traces a series of transformations in the historical imagination of a ritual at the heart of the once vigorously-debated “Singwaya tradition” of Mijikenda origins.Footnote 1 Called “mung'aro” (“[the] shining”), this ritual of elder male initiation has figured prominently in descriptions of the Mijikenda-speaking peoples of the Kenyan coast since the middle of the nineteenth century. It was not until the twentieth century, however, that it appeared in Mijikenda traditions of origin (in which, moreover, it would become centrally important). Rather than attempt a detailed reconstruction and interpretation of the ritual itself (for reasons that will become clear), I will instead trace the shifting arrangements of a cluster of ritual motifs as mung'aro became an object of inquiry for generations of missionaries, administrators, anthropologists, and historians. The significance of these motifs and the relations between them have changed dramatically over this period as they were reimagined and reinterpreted in Asian, Arab, European, African, and American descriptions. Indeed, between the 1870s and the 1970s, the image of ritual violence in these descriptions is reversed. In what follows, I aim to show how these motifs (but not their meanings) have endured despite these changes, and to highlight some of the implications for the study of traditions of origin like the Singwaya narrative.Footnote 2
The motifs in question are violent and unsettling. They include: the ritual killing of a “slave” or “stranger,” the collection of an anatomical trophy from the victim's body, intercommunal and intergenerational conflict, forced migration, seclusion, masquerade, and the dangerous transfer of power and authority. The basic arc of their rearticulation in a series of historical constellations is as follows. In the Western Indian Ocean between the sixteenth and eighteenth centuries, an image of East Africa develops as a place where Oromo-speaking pastoralists from the north kill their neighbors to the south in order to collect genital trophies as prerequisites for marriage. In the nineteenth century, this image persists alongside another in which the Mijikenda-speaking peoples of southern coastal Kenya effect the generational transfer of power through a ritual called “mung'aro” that was also said to involve the killing of an ethnolinguistic outsider. In the early twentieth century, the motif of postmortem emasculation — absent from nineteenth-century descriptions of mung'aro — is incorporated into Mijikenda representations of their own recently-abandoned ritual practice. This synthetic image is further transformed in the second half of the twentieth century, finally, through its incorporation into the comparatively recent tradition of Singwaya origins. In these narratives, rather than fleeing the ritual violence of Oromo pastoralists (as they had in traditions of origin recorded in the nineteenth century), the Mijikenda came to cite their own ritual violence — mung'aro, now understood to include the collection of a genital trophy from the ritual victim — as precipitating their expulsion from Singwaya.
In tracing this cluster of motifs over an almost 500-year period, I draw inspiration from Steven Feierman's examination of “long-term continuities in political language,” David Schoenbrun's reconstructions of “durable bundles of meaning and practice,” and Luise White's work on the “vocabulary” of rumor.Footnote 3 These conceptual frameworks are well-suited to linguistic phenomena (“streams of discourse” for Feierman, “semantic histories” for Schoenbrun, “vampire stories” for White), given language's unique capacity for pure reference (think, “Words and Things”) and metasemantic glossing.Footnote 4 One can trace, for instance, continuities in the “associated propositions” of terms like kuzifya shi (“healing the land”) and kubana shi (“harming the land”), as Feierman has done, or retentions and shifts of semantic meaning in lexical reconstructions like *-bándwa and *-sámbwa, as Schoenbrun has done. Here, however, I am tracing the history of acts and images which may evoke or express things other than themselves, but do so iconically (by resemblance) or indexically (by contiguity or “pointing-to”), rather than semantico-referentially.Footnote 5
For the case at hand, then, I propose a modification of White's method for the study of rumor. As is well known, White focuses on “the formulaic elements with which a good and thus credible story is told” — a story worth telling which, when told, its audience finds compelling.Footnote 6 White calls these formulaic elements the “vocabulary” of rumor. Instead of “vocabulary,” I draw a concept from aesthetics — namely, “motifs” — to characterize the persistent imagistic and associational elements of practices that become, in the second half of the twentieth century, key features of a transformed historical imagination of Mijikenda origins. In what follows I suggest that these motifs can be conceptualized — to mix my metaphors — as something like centers of mnemonic gravity in coastal Kenyan historical consciousness. What I mean is that the simultaneous density of associations surrounding the motifs and the looseness of semantic fit (if any) facilitates their movement across communicative genres, semiotic modalities, and domains of knowledge. Their significance remains flexible and open to further elaboration as they are incorporated into new repertoires of ritual practice and narrative discourse, but these “formulaic elements” — this cluster of motifs — are, I suggest, an important part of what come to make the Singwaya narrative of Mijikenda origins “a good and thus credible story” to the Mijikenda themselves. Tracing the recombination of a cluster of enduring motifs through the history of mung'aro highlights the extent to which such adaptation is constrained and molded by the weight of the past, opening up the question of “the limits of invention” in new ways.Footnote 7
Organization and evidence
Tracing the persistence of these motifs across ritual and narrative genres over five centuries means drawing on a wide range of different kinds of historical evidence. In what follows I examine the (re)creation and (re)interpretation of that evidence to show how, over an extended period of time and from a range of different perspectives, something called “mung'aro” has been construed as a variety of objects of analysis while at the same time being treated as if it were a stable historical entity existing outside those analyses. For heuristic purposes, I divide this history into four parts that chart a chronological progression and parse source materials roughly by evidentiary type, but which are essentially moments in the transformation of an image of ritual violence.
The first section explores travelers' accounts of the East African coast from the sixteenth to the eighteenth century. These sources report the postmortem emasculation of victims of ritual killings as a step toward marriageable male adulthood in the Northeast African coastal hinterland among peoples referred to as “Gallas” and “Mosseguejoes.”Footnote 8 They include a twelfth-century Chinese encyclopedia but consist primarily of sixteenth- and seventeenth-century texts by Portuguese priests recording information reported to them by Arab and African intermediaries. They do not describe the practice of collecting genital trophies as part of a ritual called “mung'aro,” but they are the earliest articulation of the cluster of motifs that becomes central to twentieth-century understandings of it.
The second section introduces nineteenth-century missionary accounts of two apparently distinct sets of contemporaneous ritual practices. First, they describe a persistent fear among coastal Kenyan “Wanyika” of violent attack by the “Galla” — attacks (still) said to include the collection of genital trophies as a prerequisite for marriageability.Footnote 9 Second, they describe a “Wanyika” ritual practice called “mung'aro” (or cognate terms) which does not include the collection of genital trophies, but which does — like the alleged “Galla” practice — involve the killing of a stranger or “slave.” These accounts of apparently separate and distinct “Galla” and “Wanyika” practices were either composed by or collected from Church Missionary Society (CMS) and United Methodist Free Church (UMFC) missionaries and catechists in the southern coastal Kenyan hinterland.
The third section examines early twentieth-century accounts recorded by colonial administrators as part of an effort to identify local political systems through which to administer the peoples of the coast interior. In these texts, mung'aro appears as a recently abandoned — but potentially revivable — ritual integral to nineteenth-century structures of political and jural authority in the coastal Kenyan hinterland. Importantly, it is in this moment that the collection of genital trophies is first incorporated into local understandings of mung'aro and the ritual killing it was said to have entailed.
The fourth section, finally, considers the incorporation of this new image of mung'aro into Mijikenda traditions of origin collected by local intellectuals and foreign social scientists over the twentieth century. In these traditions, the Mijikenda killing and mutilation of a “Galla” during mung'aro leads to a war that drives the Mijikenda south from Singwaya (in southern Somalia) to their present locations along the Kenyan coast. This last constellation, then, consolidates the twentieth century reversal of a nineteenth-century image: no longer the victims of ritual killing and emasculation at the hands of the Oromo, the Mijikenda had, by the second half of the twentieth century, come to represent themselves as having been the perpetrators of the same form of ritual violence — against Oromo victims.
The emergence of an image of ritual violence
The earliest reference of which I am aware to the practice of ritualized postmortem emasculation in East Africa is in the Shilin Guangji, a late twelfth- or early thirteenth-century encyclopedia compiled by Chen Yuanjing:
When a marriage is to be arranged the bride's family announces the agreement by cutting off the tail of a cow in calf as (a gesture of) good faith. … The groom's family must respond … by bringing a severed “human tail” to the house of the bride. The “human tail” which serves as a betrothal gift is the male organ. … Each marriage (consequently) deprives a man of his life.Footnote 10
Such “fifth-hand sailor' yarns,” Justin Willis points out, “probably tell us more about Chinese images of the other than about African society,” but for the purposes of my argument, that is the point.Footnote 11 What is important about this early Chinese text is the fact that by the thirteenth century an image of the ritualized collection of genital trophies on the East African coast was circulating throughout the Indian Ocean world — one that would prove to be a remarkably durable feature, whatever its epistemological status, of how Northeast Africa and the East African coastal hinterland have been imagined and understood since.
Although the peoples supposed to have engaged in the collection of “human tails” for marriage go unnamed, the image of postmortem mutilation reappears in sixteenth-century Portuguese accounts where they are identified as either “Gallas” or “Moceguejos” (or “Mosseguejoes”).Footnote 12 The first such text, written by Joao Bermudez, purports to describe his residence in Northeast Africa between 1541 and 1556. Published in 1565, it is the first to attribute the collection of genital trophies to a named people: “the Gallas.”
These Gallas live in the lands neighboring Magadoxo [Mogadishu]; they are a wild and cruel people … In the lands they conquer, they kill all the men, with the young men they cut off the genital members [os membros genitaes], the old women they kill, the young women they keep for their use and service.Footnote 13
Although the association of ritual killing and emasculation with Oromo speakers continues, subsequent Portuguese accounts also describe the practice among another population farther south, in what is now Kenya. Francisco Monclaro, a Jesuit priest, describes the “Moceguejos” populating the territory surrounding the city of Malindi in 1570 as follows:
They live in the fields and forests, they wear their heads covered with very foul-smelling clay. … They are very warlike and they say that it is their custom in fights to cut off the foreskins and swallow them, and later when they appear before the King render them up by casting them from the mouth so that the King may make them knights [os arme Cavalleyros].Footnote 14
Joao dos Santos, a Dominican Friar who had visited the East African coast between 1586 and 1589, also describes “Mosseguejo” men dressing their hair with clay that cannot be removed until they “bring before their leader [capitão] an obvious sign of the man that they killed.”Footnote 15 Dos Santos adds that “the Abyssinians … and the Gentile Galla of Ethiopia: all have this same custom.”Footnote 16
Another early seventeenth-century text compounds the confusion about the identity of the peoples described. In an account of his travels in the 1620s, Jerónimo Lobo describes the “Galla” of Abyssinia in terms almost identical to those in which Monclaro described the “Moceguejos” of Malindi in 1571, strongly suggesting that Lobo was familiar with Monclaro's text and in fact borrowed from it.Footnote 17 Alternatively, the similarities raise another possibility — not mutually exclusive with the first — that the peoples referred to as “Gallas” and “Mosseguejoes” were more culturally similar than has generally been acknowledged.Footnote 18 As will be seen, however, the discontinuities among constellated images of mung'aro — especially the fact that the collection of genital trophies is described as a feature of mung'aro only after the ritual was no longer performed — undermines any straightforward understanding of the transmission and retention of such cultural forms between peoples historically.
In 1728, fifty years after Lobo's death, Joachim Le Grand translated a manuscript version of Lobo's Itinerário into French and published it in Amsterdam and Paris as Voyage Historique d'Abissinie. In a “Dissertation on the Coast of East Africa” accompanying his translation, Le Grand provides another description of the “coast of Melinde.” But although he cites Lobo, the true source of his description of the coastal “Mossegueios” is immediately recognizable as dos Santos, who is not cited.Footnote 19 Then, in 1735, Samuel Johnson condensed and translated the Voyage Historique — Le Grand's French translation of Lobo's Portuguese Itinerário — into English, publishing it in London as A Voyage to Abyssinia by Father Jerome Lobo.Footnote 20 The Voyage Historique was further translated into German and published in Zürich decades later as Reise nach Habessinien.Footnote 21
From the sixteenth to the eighteenth century, then, this image of “Gallas” and “Mosseguejos” circulated in Europe through the translation, repetition, and free elaboration of earlier texts. This early modern image of East and Northeast Africans as killers and collectors of anatomical trophies would go on to shape understandings in subsequent encounters between East African peoples and European explorers, missionaries, and colonial administrators. It was current among the Arab and Swahili populations of the western Indian Ocean, from whom it had made its way into early Portuguese accounts (and who were also the most likely source of the “fifth-hand sailor' yarns” of Chen Yuanjing's encyclopedia). Finally, its repetition along this intertextual chain tended to include the addition of some new detail at each new link — like the foreskins of Monclaro's account, or heads-as-trophies with Le Grand — in ways that point to the plasticity of the image even as its core motifs persist.
In the following section, I show how this image of Oromo-speakers' ritual violence continued to circulate alongside descriptions of a different ritual among the Mijikenda-speaking peoples of the southern Kenyan coast in two different sets of nineteenth-century texts. The first are early nineteenth-century accounts that continue to describe the “Galla” as ritual killers and collectors of genital trophies. The second are descriptions of an apparently unrelated ritual practice among Mijikenda-speakers called “mung'aro,” said to involve the killing of a “stranger” or “slave” but not the taking of a genital trophy from the victim. The two images will eventually combine in the early twentieth century into a new constellation, but until the end of the nineteenth century the collection of such trophies never appeared in descriptions of the Mijikenda mung'aro. All the more striking, then, that it becomes such an important feature of twentieth-century constellations of the lapsed Mijikenda ritual.
A nineteenth-century missionary diptych
On 15 March 1844 in Mombasa, CMS missionary Johann Ludwig Krapf met with the Kadhi of Mombasa and Rashid bin Salim, who had been “the chief of Mombasa” under the short-lived British Protectorate of 1822. “At this opportunity,” he writes, “I heard some account of the customs of the Wonica [Wanyika] pagans:”
In the present month is the Wangnāro [Mung'aro] of the Wonicas i.e. the time when the young people assume the mastery of the aged ones. They whiten their faces with lime in order to make a more ghost-like appearance. If any spectator should laugh at this comic parade, they would beat strip and send him off empty-handed. Therefore the Sooahelees [Swahilis] do not like to travel amongst them at the time when their annual pranks take place.Footnote 22
Subsequent experience led Krapf to revise this understanding. It was, he was later told (although it is unclear by whom), more sinister than he had first been given to understand:
I did not know at that time, that the Wagnāro [Mung'aro] … cannot terminate, unless they have slain somebody in the fields or bought (by common contributions) a slave, whom they will kill. When this has been done, the festivity terminates with eating and drinking, and with the washing of their bodies which they cover with mud during the Wagnaro, in order that they may remain unknowable, when they slay anybody on the road.Footnote 23
A few weeks later, as Krapf was evangelizing on the mainland south of Mombasa, he was advised by a local leader “not to go to Bumbo and its vicinity, as the Wagnāro … had just commenced, and it were not advisable, that I should stroll about the plantations in that quarter.”Footnote 24 Although he does not witness it himself, he sees signs of others' participation: among the people assembled in a market, Krapf noticed, were
a few men and women, who were come from Bumbo, and who had bedaubed their faces with mud, to give themselves the appearance of evil spirits. Besides they make their faces unknowable, in order that they cannot be discerned, when they slay a lonely traveler in the fields or forests.Footnote 25
In 1846 Krapf, together with Johannes Rebmann, established a CMS mission station at Rabai Mpya on the mainland northwest of Mombasa. Rebmann later shared his notes on the peoples of the area with Richard Burton, who makes the following brief mention of mung'aro based on Rebmann's material:
Once about every twenty years comes the great festival “Unyaro” [Mung'aro], at which the middle-aged degree is conferred. … Candidates retire to the woods for a fortnight, and clay themselves for the first half with white, and during the second with red earth; a slave is sacrificed, and the slaughter is accompanied by sundry mysteries, of which my informants could learn nothing.Footnote 26
Burton's account is the earliest periodization of mung'aro, framing it as an initiation into a gendered gerontocratic hierarchy that, to him, resembled “masonic degrees.” He repeats Krapf's earlier claims about the “mud” (now “clay”) decoration, elaborating this detail to include its division into sequential white and red phases. Burton omits the ritual killing of a stranger, describing it instead as the “sacrifice” of a “slave.”Footnote 27
In 1863, Charles New arrived in Mombasa to support Thomas Wakefield in the expansion of UMFC mission operations. Wakefield had, together with Krapf, completed construction of a new mission station in Ribe. Although the year is not given, New describes a brief encounter with a ritual there that, although unnamed, resembles Krapf's “Wagnaro” and Burton/Rebmann's “Unyaro.” New's initiate wore “a covering of soft mud, an inch thick, looking like a close-fitting cap” resembling Portuguese descriptions of the clay headdresses of “Mosseguejo” youths (which, recall, could only be removed after the youth had slain an enemy and collected a genital trophy).Footnote 28 With the adornment complete, according to New, “the man is turned into the wood, and is allowed to do as he pleases.”Footnote 29 New states that the initiate was formerly “expected to kill someone before the ceremony is over” (citing Krapf for this claim), but adds that he “believe[s] it is not so now,” without explaining why.Footnote 30
The most detailed nineteenth-century account is also the only one from an individual who claimed to have once undergone the ritual as an initiate. Rabai elder Abe Mjeni Mwasunga's description of “Ugnaro” was recorded in 1879 by CMS catechist George David.Footnote 31 “Being one of the number among those who were showed the Wanyika Customs of special office by their grandfathers, the rest having all died excepting he and another,” Mwasunga (David's interlocutor) had been asked to oversee the initiation of a new cohort in 1879.Footnote 32 Rather than participate, however, he offered David a description of the ritual from memory:
On the day on which the Ugnaro [mung’aro] begins, the elders order their young men to spend the night dancing. … Both they and the elders go out on the open field, away from their Kayas (Forts) to fetch clay for their bodies. … Then 16 particular elders, go first to their sacred place to remove the charms. … [T]hey are not to wash off the clay from their body till they murder a man, i.e. a stranger or slave seen passing alone anywhere on their Country.Footnote 33
This is the first mention of kaya “charms” in a description of the ritual, although these will become an important element of twentieth-century accounts, as will be seen.
Two years before taking down Mwasunga's account, David related a “Wanika” origin story to Frere Town Mission Secretary James A. Lamb, which cites the alleged “Galla” practice of killing for genital trophies as the cause of “Wanika” migration from the north. This nineteenth-century tradition of origin thus belongs to the second set of texts examined in this section which, like earlier Portuguese texts, focus on the collection of genital trophies as a feature of “Galla” violence against Mijikenda-speaking peoples:
The history of Jilori [Jilore, inland from Malindi, on the Sabaki River] as George David gives it is that it was originally Wanika territory—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, & for this barbarous purpose they used to catch the Wanika which caused them to leave that part of the country, whereupon it was resorted to by runaway slaves who put themselves under Galla protection & render tribute in return; & now that the Gallas fear a white man is coming they are beginning to sell the slaves.Footnote 34
It is unclear how far in the past David understood this migration from Jilore to have taken place, or how long ago the Oromo had been subject to this “law.” But almost a quarter of a century earlier — in 1853 — Krapf had published an account of the very practice David describes:
Throughout the Galla Nation the abominable custom prevails to emasculate a prisoner in war either when he is alive or slain in the battle. … Without this exhibition a Galla cannot get a wife. He is consequently compelled to go to war or waylay innocent travellers of other nations, until he gets this requisite for the marriage-contract. But as he cannot always quickly succeed, he has found out the horrid expedient of buying a slave from the coast, in order to cut his privity and carry it to his bride.Footnote 35
This was, according to Krapf, an ongoing practice: “I have some years ago seen myself some slaves brought from Mombas to the Galla market at Mberria, who were sold for this wicked purpose to the Galla who sell their ivory to Mombas partly on this count alone.”Footnote 36
David's and Krapf's texts each link the collection of genital trophies to the East African slave trade, albeit in different terms. Rather than abducting fugitive slaves to sell back into slavery at the coast (as in David's account), the Oromo were, according to Krapf, buying slaves at the coast to kill them for body parts. What is important for my argument here, however, is not the exact nature of the relationship that may have existed between slavery, ritual killing, and genital trophies, but rather the fact that in these accounts these three motifs are understood to be related in some way. The loose, associational quality of the conceptual links between these practices is, I argue, an important aspect of their durability as motifs. The potential killing of a “slave” — although the specific local category of persons to whom this term referred remains unclear — is also an element that these descriptions of Oromo ritual share with the contemporaneous accounts of mung'aro with which this section began. As another point of conceptual contact between the two sets of practices, then, transformations in the regional slave economy over the nineteenth century may have shaped the ongoing transformation of mung'aro imaginaries by facilitating the absorption of details or motifs drawn from understandings of Oromo ritual.Footnote 37 Krapf remarks in 1853, for example, that as a result of a recent British-Omani treaty banning their export outside the Sultan's territory, “slaves have got cheaper on the African coast so that a person which formerly could not afford the prize, can now buy a slave with little expense,” and “the Galla-tribes to the North of the Wanika-country … will not be slack in gratifying their horrid propensity and practise.”Footnote 38 When David reported on the history of Jilore in 1877, meanwhile, overland transportation of slaves had just been banned by a new treaty, and the capture and sale of escapees seems to have been at least partly driven by speculation about the eventual banning of slavery itself by the British (“now that the Galla fear a white man is coming they are beginning to sell the slaves”).Footnote 39
Before proceeding to the transformed image of mung'aro in the early twentieth century, let me reiterate the motifs of this associational nexus in the nineteenth century. First, there is the killing of a stranger or “slave,” especially one abducted from a path while traveling alone. Second, the adornment and disguise of the ritual participants with clay. Third, the transition between social statuses, power, or authority by initiates. Fourth and finally, there is the seclusion of initiates in the kaya ritual enclosure before and after the killing of a stranger. By the end of the nineteenth century in southern coastal Kenya, then, the (1) killing of a stranger or slave by (2) disguised initiates is associated with (3) the transfer of political power anchored in (4) the ritual space of the kaya. In the first half of the twentieth century, the constellation of these motifs in mung'aro imaginaries comes to include the collection of a genital trophy — a detail derived not from accounts of nineteenth-century Mijikenda ritual, but from stories of Oromo predation in which the Mijikenda figured as victims.
Mung'aro in the anthro-administrative imagination
The 1879 initiation in which Abe Mjeni Mwasunga declined to participate is the last for which there is any evidence. Descriptions of mung'aro from the early twentieth century, then, are elder male recollections of a recently abandoned ritual practice still within living memory.Footnote 40 Colonial administrators collected these descriptions in an effort to identify political institutions free of Arab and Swahili influence, and believed they had identified a “Wanyika” system of age-grades and ritual authority as just the kind of political formation amenable to indirect rule as “Native Authorities.” Those institutions, however, seemed to be disappearing before their eyes despite their efforts to shore up the authority of “elders” in office.Footnote 41
These “anthro-administrative” descriptions of elders' councils, age-grades, and “secret societies,” and of the rituals regulating advancement within (or admission to) them, involved translating local political concepts and categories into more familiar administrative ones.Footnote 42 For example, in what is probably the earliest twentieth-century account (undated, but circa 1913, collected by Assistant District Commissioner [ADC] Sydney La Fontaine in Kilifi District, north of Mombasa), Giriama initiation was said to involve the selection from a “newly elected” kambi (or senior age-grade) of two “headmen;” one “pre-eminent,” the other “subordinate.”Footnote 43 Together these two would have “jurisdiction over the whole tribe of the Wa-Giriama” — that is, if such an “election” could be held.Footnote 44 If the Portuguese had conceptualized these male rituals as related to warfare, and nineteenth-century missionaries had viewed them as “pagan” religious practices, colonial administrators understood them as fundamentally political in nature.
Arthur Champion, who succeeded La Fontaine as ADC of Kilifi District, also believed that Native Authority figures needed such a ritual backing for local legitimacy.Footnote 45 To that end, he obtained (in 1914, and does not say from whom) the following description of a Giriama mung'aro performed in the 1870s:
The nyeri [junior generation], their bodies smeared in red mud and castor oil (mbono) and wearing a garment known as marinda wa makindu (a kilt made of leaves), were grouped into marika and each rika [age set] was given its name. The feasting continued for some days in the kaya [ritual center] and then the young men went forth into the bush, still in the same guise, and so they had to remain till a foreigner could be found and killed. They then scraped off the mud and oil and threw their kilts on to the body of the dead man.Footnote 46
In contrast to nineteenth-century accounts, Champion describes mung'aro as involving a kind of ritual cross-dressing. “Marinda,” which he translates as “kilts,” are women's pleated skirts, though in this case they are made from palm fronds — disposable ritual replicas of young, female attire, cast off at the culmination of a ritual that turns initiates into senior men. The marinda are removed along with the clay and castor oil and placed on the body of the victim whose death effects their ritual transformation. The initiates' new status as powerful senior men figurates not only their prior existence as having weak, junior, female qualities, but perhaps also the nature of “outsiders” like the ritual victim relative to themselves as well.
In 1917, administrators recorded similar accounts from representatives of the Duruma.Footnote 47 Unlike Champion's Giriama account they are not presented as memories of a specific historical performance of the ritual, but differ from it only in that one account suggests that “a python may be substituted for the human victim,” and that the ritual was “also performed on sons of Chiefs, members of the Kambi, and for this purpose, persons who have already been through the ceremony visit the former in their villages.”Footnote 48 It is unclear whether “chiefs” refers here to government-appointed chiefs (a recent addition to the South Coast political landscape), to regionally important “big men” within a field of dispersed and relatively autonomous homestead settlements, or to the heads of individual settlements or settlement clusters. Mung'aro, in this account, seem less like a rite of “elderhood” than one capable of shoring up power along some lines of descent and not others.
The last early twentieth-century account of mung'aro as a lapsed historical practice still within living memory is drawn not from a colonial administrator but a Methodist missionary. Published posthumously in 1935, J. B. Griffiths's description includes many familiar images — clay disguises, dancing, palm leaf skirts, abduction from a path, killing, charms — but introduces a number of variations and new details.Footnote 49 Most important among these is that the death of the ritual victim now involves the removal or their right hand and their genitals — the earliest mention of the collection of genital trophies as a feature of mung'aro.Footnote 50 These “relics,” according to Griffiths, were brought to “the elders” and turned into a chirumbi “war charm” placed either at the gate to a fortified kaya settlement, or kept in the elders' “house of secrets” at the center of the kaya.Footnote 51 In Griffiths's description, then, the ritual not only effected a transition between gendered and generational social statuses, but renewed and revitalized the objects that anchored elder male authority in kaya ritual centers.
Griffiths explicitly refutes Mood's informant's claim that a python could be substituted for a human victim. The Duruma, he thought, were “too much afraid of the shades of their forefathers to make a change.”Footnote 52 But despite Griffiths's use of the present tense to deny Duruma claims that “they have now substituted a python or a leopard for a human being,” there is no evidence that a mung'aro of any configuration was performed in the twentieth century.Footnote 53 Griffiths did, however, claim to have been “initiated” as a Duruma “elder” himself — although it is unclear what “elder” means in this context, and what his “initiation” entailed (it seems unlikely to have included killing another person, for instance).
In the early twentieth century, then, we see not only the first claims that mung'aro involved the collection of anatomical trophies, but also the earliest claims by Mijikenda-speakers themselves that mung'aro might not necessarily involve killing a human being (an assertion refuted by a missionary interlocutor). It is possible that the latter claim was an innovation in response to colonial administrative efforts to identify, reform, and perform a new round of initiations (but then, recall Krapf's 1845 assertion that killing was not required among the northern “Wanika,” and Charles New's claim that by the mid-1860s killing was “no longer” an initiation requirement). In any case, administrative interest in kambi ritual did have a practical dimension that may have had consequences for how mung'aro was remembered (or at least for how it was presented to colonial authorities). As David Bresnahan has shown, administrators came to believe that “to establish a legitimate kambi they needed to hold genuine kambi initiations,” and so sought to induce regional elders to initiate their successors.Footnote 54
For various reasons these efforts failed or were only partly realized. With the Duruma, for instance, the government appointed two uninitiated men as headmen over the handful of surviving mung'aro initiates, and then tried to oblige the latter to initiate the former to cement their authority.Footnote 55 Perceiving that the administration had “removed their prerogative by appointing non-initiates as Headmen over them,” the older cohort simply refused to initiate their juniors (or indeed to assist the colonial administration in any way) until, in November 1923, the government-recognized headmen were forced publicly to resign, and two initiated elders — Mwaiona wa Munga and Kidanga wa Mwaruwa — installed in their place.Footnote 56 District Commissioner H. B. Sharpe claimed to see an immediate improvement, but Mwaiona would be prosecuted for “extortion” within a year, prompting Sharpe to remark that “if it were not that the Government needs to squeeze out of him the Duruma Initiations … it might have been better to remove him.”Footnote 57 Mwaiona would resign in 1925 without having overseen the initiation of a new kambi.Footnote 58
In addition to elders' manipulation of the administration's desire to perform what it viewed as a necessary rite of legitimation, a series of droughts, local rebellions, and the disruptions of the First World War also interfered with the ritual transfer of generational authority.Footnote 59 And as the number of living representatives of senior age-grades dwindled, officials worried that the proprietary esoteric knowledge they claimed to possess threatened to disappear with them. In 1920, DC Thompson described the question of Duruma kambi initiations as an urgent one, given that “a bare half dozen elders of the original kambi remain … and if efforts are not made soon to re-establish the council, there will be no elders left to initiate intending candidates.”Footnote 60 By 1934, the administration believed that there was “only one living person who has ever been initiated into the practices and mysteries connected with the tribal Kayas”: Kidanga wa Mwaruwa, the man installed as headman alongside Mwaiona wa Munga eleven years earlier.Footnote 61 No such initiations were held, and by the second half of the twentieth century, mung'aro as an element of nineteenth-century political and legal structures had faded with those structures from living memory. It was, however, preserved and transformed through its incorporation into a new historical imagination of Mijikenda origins: the Singwaya narrative.
Mung'aro, Singwaya, and Mijikenda origins
Although nineteenth-century sources attest multiple and varied traditions of origin for different Mijikenda-speaking peoples, by the early twentieth century a consensus — more or less — had developed about their shared origins in a place called Singwaya in or near what is now southern Somalia. Willis, however, has argued that the apparent consistency of these narratives is “illusory” — an effect achieved only by “disregarding differences.”Footnote 62 What the Singwaya narratives have in common, for Willis, is that they mention Singwaya; other narrative elements are consistent only in their “extreme negotiability.”Footnote 63 But what is negotiable, I argue, may be less the details themselves — at least, not the motifs I have traced up to this point — and more how those details are combined into larger narrative structures (which themselves seem to fall into two or three general patterns). The larger claim, however, is that as the basis of elder authority was transformed over the twentieth century, the historical understanding of mung'aro was transformed through its incorporation into Singwaya narratives which in turn changed over the course of the century.
The first mention of mung'aro in a Singwaya narrative is in an undated text from the interwar period, compiled by H. M. T. Kayamba from the statements of sixteen Digo elders in northeastern Tanganyika.Footnote 64 Kayamba's informants related that “some ten generations ago there was a tribe called the Wambokomu [Pokomo] which inhabited the country known as Chungwaya [Singwaya], north of Lamu.”Footnote 65 After siding with the “Wasegeju” in a war against the “Wagalla,” the “Wambokomu” migrated south, fragmenting into present-day Mijikenda groups along the way. In this narrative, mung'aro — a “grand tribal meeting” at which men “smeared themselves with mud” and elected representatives “to fill up three grades of a council” — founds the post-exodus political order of the Mijikenda in their new home.Footnote 66 It is not cited as the cause of the war with the “Wagalla,” nor is any claim made that the ritual was performed in Singwaya at all.
This changed in the second half of the twentieth century as mung'aro emerged as a new explanation (alongside others) for exodus from Singwaya. The earliest iteration of this new understanding of which I am aware is that of Giriama Chief Kalu Birya, recorded in 1955.Footnote 67 Having relocated from “Muthothana [Tana River] … near Abyssinia” to “Pokomo” and establishing there “a Kaya which they called Singwaya,” the Giriama “held a big ngoma known as Mungaro, … and during that ngoma it was decided that a man must be killed from the Wagala tribe,” who in response “decided to fight the Giriama.”Footnote 68
This new image of mung'aro as the specific precipitating cause of a war with the “Galla” is repeated and elaborated in the corpus of oral narratives collected separately by Thomas Spear and Cynthia Brantley in 1970–71. There is however considerable variation across these traditions as to the exact cause of migration from Singwaya. Most cite the killing of an Oromo boy, but not all give mung'aro as the context of that killing. Of the twenty-eight traditions Spear collected that give a reason for the departure from Singwaya, twelve cite mung'aro as the cause.Footnote 69 But almost as many describe a killing resulting from conflicts among men of different ranks and anachronistic “tribal” identities over sexual access to women.Footnote 70 Of the ten Giriama historical traditions Brantley collected which cite conflict with the “Galla,” three give mung'aro as its cause.Footnote 71 Four, however, describe the killing as the response of a jealous Giriama husband to an Oromo man's claim to a “traditional” right to sexual intercourse with a new Mijikenda bride.Footnote 72 It is as if the centuries-older constellation of ritual killing and emasculation by Oromo men in order to marry had, by the twentieth century, fragmented into its constituent motifs and been reassembled into two twinned versions of the relatively new Singwaya narrative: a hated Oromo marriage practice in which Mijikenda are the perpetual victims, and a ritual killing involving the collection of genital trophies — only now as part of the Mijikenda mung'aro.
In two of Spear's narratives the victim was an Oromo woman either married or impregnated by a Mijikenda “youth” and then killed by his people.Footnote 73 In four others (three from Digo informants), the victim was a handsome Oromo man killed by Mijikenda men out of jealousy — a cause cited in a much earlier colonial-era version of the story, also from a Digo informant (in which, however, the killer was a Segeju).Footnote 74 Brantley and Spear also recorded one tradition of origin each that, like George David's from 1877, attribute the initiatory killing of outsiders to the Oromo and that cite this as the cause of a Mijikenda migration — only now, from Singwaya.Footnote 75 Aspects of nineteenth- and early twentieth-century migration narratives persisted, in other words, alongside more recent versions that attributed this same form of violence to the Mijikenda during mung'aro. The motifs are the same, only the resulting constellated images differ.
By the time mung'aro was incorporated into the repertoire of Singwaya narratives as a cause of the war with the Oromo, the purpose of the ritual — what it was understood to do — was open to a wide range of local interpretations. Explanations by Spear's informants include: to bring rain in times of drought, to “differentiate between kambi and nyere,” to “cut a rika [age-grade].. during the vuri [short rains],” a dance “performed during mourning ceremonies or when people were going to fight,” a circumcision ritual, a dance performed during the execution of “wrong-doers condemned to death by the kaya elders,” to “select leaders,” and to “cleanse the sick.”Footnote 76 But this ambiguity only emerges when descriptions of the ritual as a historical practice are elicited by the interviewer, not when informants spontaneously insert mung'aro into the Singwaya narrative of origins. In other words, the ambiguity emerges only in elicitations of the ritual's function in the recent historical past, not in spontaneous descriptions of its consequences in the deep mythic past. In the latter, what is important for informants is not the social function of the ritual, but the fact that its violence led to the expulsion of the Mijikenda from Singwaya.
Despite the variation in these descriptions of the ritual process, significance, and function, Spear argues that “the initiation ceremony of the age-sets, mung'aro, was a detailed reenactment of the migration from Singwaya.”Footnote 77 But in these accounts, mung'aro is cited as the cause of a migration that had not yet occurred, not a reenactment of it. An alternative hypothesis might be that the periodic performance of mung'aro was a ritual repetition not of the migration itself but of the “primal crime” that, in these versions of the Singwaya narrative, led inexorably to it. But whatever its virtues, such an interpretation would be only a tempting optical illusion made possible by the mid-late twentieth-century incorporation of mung'aro into the Singwaya story roughly half a century after the consolidation of that new idea of origins around the turn of the century. One might argue instead that the cluster of motifs condensed in this image of mung'aro was, in some sense, more compelling to Spear and Brantley's informants than the idea of Singwaya origins on its own. Having a much longer history among coastal peoples than the Singwaya narrative itself, these motifs may have been unconsciously incorporated into it for that very reason — to shore up, as it were, a widely shared but historically shallow origin story with an image composed, in part, from more deeply ingrained cultural motifs.
The specific image of mung'aro that became a staple of Singwaya narratives was, recall, consolidated during the colonial period in relation to administrators' quest for rituals of elder legitimation. In the second half of the twentieth century, this is the image of mung'aro projected backwards as one of a few plausible explanations for conflict with the “Galla” (the notion of which is clearly far older than claims that it took place in “Singwaya” or that it was the result of mung'aro). The incorporation of this image into the emergent repertoire of Singwaya narratives took place during a period of Kenyan history — the 1950s and 1960s — in which coastal elders' claims of authority increasingly drew on the “discursive contrast” of tradition and modernity that had also haunted the failed administrative effort to ritually transubstantiate “modern” Local Native Tribunals into “traditional” kambis.Footnote 78 In the absence of any further mung'aro initiations, claims to proprietary esoteric knowledge of there having once been such a ritual, and of the motifs understood to have been its distinctive features, came to underwrite assertions of “elder” authority instead.Footnote 79
As proprietary historical knowledge of Mijikenda “tradition” displaced lived ritual experience as the putative ground of “elder” authority, the incorporation of mung'aro into the Singwaya story in the second half of the twentieth century transformed both. Narratively locating mung'aro at the origins of the Mijikenda was an assertion of the ritual's antiquity, and so also of its “traditional” character and of the authenticity of those who knew about it qua “traditional” elders. At the same time, the “traditional” quality of this image of mung'aro — including the cluster of motifs that I have described — serves to authenticate the Singwaya narrative as a “tradition” of origin — a traditional tradition, so to speak. Foregrounding mung'aro in these Singwaya narratives demonstrates a command of esoteric historical knowledge about Mijikenda ritual and origins, and is thus also part of a larger performance of “expertise” as a “traditional” kaya elder.Footnote 80 It also logically entails “owning,” in some sense, the violence that led to their ancestors’ expulsion from Singwaya, but the question of “who started it” does not actually seem to be what is at issue. What is important is the violence itself, and that the image of that ritual violence include the right combination of certain of key motifs.
For some, a new version of an older regicide narrative did just that. In a Segeju origin story collected by E. C. Baker in 1919, a jealous Segeju husband kills the “Sultan” of Singwaya for attempting to exercise his “customary” right of initiatory sex with new brides.Footnote 81 It is this killing that sparks the war leading to expulsion. Although the killing is not in the context of mung'aro, the killer does don women's clothes as a disguise, and collects an anatomical trophy (though not a genital one) from the victim as proof of the killing.Footnote 82 When Spear, Brantley, and Morton's Giriama informants told the same story in 1970–71, the killer was Mijikenda, not Segeju.Footnote 83
For others, the image of ritual violence in mung'aro now provided the right combination of these essential motifs. In a 1918 Digo origin story, for example, a “Gala youth” is clubbed to death by a group Segeju men and his body buried in a cattle kraal.Footnote 84 The Segeju blame the killing on the Digo, against whom the “Wagala” then wage war. By the 1970s, however, Mijikenda elders had begun to assert that their ancestors did the killing, not the Segeju, and that the killing was in the context of their mung'aro ritual rather than an unnamed Segeju one.Footnote 85 These versions of the Singwaya story, collected over half a century and hundreds of kilometers apart are identical except for this detail: whether in the context of a successful mung'aro ritual or a sabotaged marriage ritual, by the second half of the twentieth century, the killers are no longer Oromo or Segeju, but Mijikenda, reversing a centuries-older image of ritual violence while reinscribing it in a new tradition of northern origins.
Conclusion
Through a careful review of the available evidence, I have shown how an image of ritual violence slowly transformed as it became the object of missionary, administrative, and historical inquiry, becoming a key element of Mijikenda oral traditions in inverted form. What first appears as a form of “Galla” or “Mosseguejo” violence directed against the “Wanyika” was, by the second half of the twentieth century, incorporated into traditions of origin as a ritualized form of Mijikenda violence against the Oromo. This process of reversal began in the second half of the nineteenth century when missionaries in the coast hinterland began to report both a “Wanyika” ritual involving the killing of a slave or a stranger, and an unnamed “Galla” practice of collecting genital trophies from slain enemies or “slaves.” The latter image had already appeared in sixteenth- and seventeenth-century Portuguese texts, but it was only after both sets of practices went from being the objects of quasi-ethnographic inquiry to historical ones in the early twentieth century that the motifs began to converge and merge in the twentieth-century image of the nineteenth-century “mung'aro.”
In George David's 1877 version of the Mijikenda origin story, recall, the killing of an outsider and the removal of his genitals is described as a necessary step in the transition between social statuses — but for the “Galla,” not the “Wanyika.” It is also cited as the cause of a migration from the north — but from Jilore, not Singwaya. In this version, the Oromo are the perpetrators and the Mijikenda the victims. In Singwaya narratives collected almost a century later, however, the Mijikenda figure as the perpetrators of the same form of ritual violence, and the Oromo as their victims. Between the 1870s and the 1970s the geographical point of Mijikenda origin is displaced northwards and the roles of killer and victim are reversed, but the motifs — capture, killing, and dismemberment of or by outsiders, a ritual transition between social statuses, and elder male power — are unchanged. Rather than a “durable bundle of meanings and practices,” then, these motifs seem to serve instead as mnemonic anchor-points — non-negotiables of a sort, through which a variety of historical narratives may (and seemingly must) pass, out of which a range of constellated images may be generated. The reconfigurations, reversals, displacements, and transpositions that weave in and around these centers of mnemonic gravity are, I suggest, facilitated by the associational qualities of the violent motifs themselves, rather than by strict and consistent understandings of durable “meanings and practices” — or indeed of any empirical historical relationship between them.
In tracking these transformations, I have drawn inspiration, in particular, from White's methodological focus on the “vocabularies” of stories about the past — formulaic narrative elements once dismissed as distractions from historical truth, but which, she shows, are precisely the details around which compelling stories are spun. Recasting this methodological move in terms of aesthetics (“motifs”) rather than linguistics (“vocabulary”) highlights the multimodal durability of these details across genres of ritual practice and historical discourse as they wind their way into a vision of origins that, over the course of the twentieth century, became “a good and thus credible story” for the Mijikenda to tell themselves about their own history.Footnote 86
Tracing the articulation of this cluster of motifs into a series of constellated images also reopens the question of “the limits of invention” in new directions. While their rearticulation is not the result of manipulation or conscious “invention of tradition,” this is not — given the variability of the resulting constellations — because this cluster of motifs constitutes either a “durable bundle of meaning and practice” or “long-term continuities in political language.”Footnote 87 Focus on the constituent motifs of this series of constellated images allows us to explore how and why these motifs might endure despite the instability of their significance, the abandonment of the practices in and through which they were once expressed, and dramatic transformations of social, political, and ritual landscapes in which they have been embedded over the last five hundred years of coastal East African history.
Acknowledgements
Research was supported by a Fulbright-Hays Fellowship and by grants from the Wenner-Gren Foundation and the Nicholson Center for British Studies. A version of this argument was presented at the African Studies Association annual meeting in November 2021 on a panel organized by Raevin Jimenez. I thank her for the invitation, and Adam Ashforth and Yaari Felber-Seligman for their questions there. I would like to thank Ralph Austen, Robert Blunt, Richard Bodek, Hannah Chazin, John Cropper, Luther Gerlach, Colin Halverson, Britta Ingebretson, Paul Ocobock, Emily Osborn, Anna Weichselbraun, and especially David Bresnahan, for their valuable comments on different stages of the manuscript's development. I also wish to thank the editors and anonymous reviewers for their thoughtful readings and helpful suggestions. For Ralph Austen. Author's email: [email protected].