Essence is expressed by grammar.
———Ludwig WittgensteinIt is now an anthropological truism that ethnic identity is “other”-oriented, such that who we are rests on who we are not. If there is anything primordial about ethnicity, it is not the blood-based body of affective affiliations that plagued an earlier generation of modernization theorists, but rather, as Comaroff (Reference Comaroff2010: 531), revisiting Barth (Reference Barth and Barth1969), maintains, “the act of drawing boundaries among populations.” The contexts in which such demarcations occur are of course crucial to their social and historical significance, whether they are motivated by politics, resource competition, class formation, marketing, immigration, religious encounter, or the apocalyptic violence of state-directed genocide. Such contexts also illuminate the cultural “stuff”—such as language, ritual, kinship, or costume—that is selected and produced as the content of ethnicity. But whatever its significance and cultural material, ethnicity rests on prior conditions of differentiation and othering, forming “a dialectic of identification and contrast” (Comaroff op. cit.) that accounts for its fluid and dynamic characteristics. In its widespread Hegelian variations, this dialectic surfaces in self-designating ethnonyms originally generated by ethnic outsiders, as external coinages.
The development of Yoruba identity in the late nineteenth century falls well within this view from without. According to the standard narrative, “Yoruba” was a Hausa or Fulani term designating Old Oyo, later extended to its southern vassals and neighbors who were otherwise organized into subcultural groups such as Egba, Egbado, Ijesha, Ijebu, Ondo, Ekiti, and Akoko but lacked any overarching identity as such. Law (Reference Law1997: 206, 216 n. 11) ventures the earliest appearance of the ethnonym in a West African Arabic language source from 1615 (a Muslim treatise on enslavable infidels), and cites Bowdich (Reference Bowdich1819: 208–9) to document the use of the term among Muslim residents of Kumasi in 1817.Footnote 1 In his monumental History of the Yorubas, Samuel Johnson (Reference Johnson1921: 5) quotes a “copious extract” from Denham and Clapperton (Reference Denham and Clapperton1826, app. 12, p. 22), who translated an Arabic manuscript by Sultan Mohammed Bello of Sokoto that describes “Yarba” as “an extensive province” whose people bought slaves from the north and “resold them to the Christians” who transshipped them from the coast. The appropriation of “Yoruba” as a conscious term of self-identification also occurred from without, among the Aku community of liberated slaves in Freetown, where the CMS with protégé Samuel Ajayi Crowther standardized Yoruba language and orthography for what became the elite consolidation of a Pan-Yoruba identity (Ajayi Reference Ajayi1960). This occurred slowly and unevenly, beginning with Oyo as “Yoruba proper” and gradually expanding into a wider regional identity through the ideological framework of a Christian nation (Peel Reference Peel, Tonkin, McDonald and Chapman1989; Reference Peel2000). More recently, Matory (Reference Matory1999; Reference Matory2005) has judiciously demonstrated that a crucial contribution to modern Yoruba identity came from late nineteenth-century African-Brazilian travelers to Lagos whose manifold impact on the Lagosian Renaissance left a lasting legacy of cultural nationalism.Footnote 2
It is not my intention to challenge the specifics of these externalist perspectives, but rather to complicate the very distinction between “inside” and “outside” that they presuppose.Footnote 3 If any prior demarcation of boundaries involves “a dialectic of identification and contrast,” then how are these identifications made and contrasts drawn? By what cultural logics and semantic principles are objects, entities, particulars, and collectivities constituted in the first place, in a sense “from within”? What I am suggesting in this exploratory essay is, first, that the “primitive” function of Yoruba negation is not in fact unconditioned, but is already embedded within a cultural semantics of quantification; and second, that such demarcating and objectifying modalities played an important if largely overlooked role in the mediation and incorporation of those “external” factors and diasporic influences that eventually gave rise to Yoruba ethnicity. My aim is both to complement and destabilize externalist perspectives by turning Yoruba lineage theory on its head—relating descent, residence, kinship, and kingship in Nigeria to their reconstituted ritual frameworks in Cuba and Brazil. Following Barber (Reference Barber1991) on praise-poetry and Verran (Reference Verran2001) on Yoruba quantification, I reexamine the semantics of the category ilé in the emergence of Lucumí and Nagô houses in order to provide an internal counter-perspective on Yoruba ethnogenesis from within.
If such an exercise helps explain the rise of Yoruba ethnicity, it does so circuitously, by focusing on two specific trajectories in Cuba and Brazil that shaped the transnational field of Yoruba social and religious capital—an important episode in the broader consolidation of “pan-Yoruba” identity in Nigeria (Matory Reference Matory2005). Whereas the first trajectory charts the emergence of ritual lineages out of Catholic brotherhoods, the second relates colonial ideologies of racial stratification to growing concerns with ritual purity. My reanalysis of these trajectories through a Yoruba lens reveals historical continuities between racial whitening and ritual purification in Santería and Candomblé.
THE HOME IN THE TOWN
The basic Yoruba concepts of “home” (ilé) and “town” (ìlú) are marked by considerable socio-semantic ambiguity, embracing diasporic idioms of “homeland,” ritually reconstituted centers, kingdoms, and quarters, as well as houses in which people live. “Home” or “house” can shift between narrow conceptions of a localized residential family group, a collection or compound of such groups (agbo ilé, lit. “flock” of houses), or in a more genealogical idiom, between a core of agnatic lineal descendents (ọmọ baba kan) and the broader range of cognates, affines, strangers, and, in the past, pawns and domestic slaves that would become attached and in some ways “absorbed” within the lineage (ìdílé) or compound (ilé). Such “absorption” is rarely absolute, in that “strangers” may be barred from inheriting lineage-vested titles; they may commemorate distinctive orisha (deities) that they “brought” to the compound generations before; or, as Barber (Reference Barber1991) so cogently demonstrates, their praises (oríkì) can set them apart from the “core” by resurrecting associations with alternative origins. But even the agnatic core is open to negotiation, since sons seeking titles through complementary filiation—that is, through their mother's connection to the title-holding patrilineage—can “opt” into their mother's agnatic line.Footnote 4 I knew a family in Ayede-Ekiti that seemed to have two cognomens—they were ilé Ilétogun (see how ilé is already nested in a prior “Ogun” line) and ilé Balógun, holders of the Balógun war-title of Ayede. A woman in the family explained that her paternal uncle received the title from his mother's Balógun line, with which his own patrikin became genealogically associated.Footnote 5
It was precisely such slippage and ambiguity that gave rise to a debate about Yoruba descent that has never been fully resolved (Barber Reference Barber1991: 156–58). Peter Lloyd (Reference Lloyd1966) noted that genealogical claims through the mother's line for access to corporately held titles and land were more common in southeastern Yoruba areas such as Ondo, and he concluded that Yoruba descent itself took different forms—agnatic in some regions, cognatic in others. This conception was based on certain assumptions about the character of Yoruba lineage structure and descent, not only that it was primarily patrilineal, but also that localized corporate groups of families and compounds formed the primary building blocks of Yoruba social and political organization. Given the Fortesian dogma of the day, Lloyd saw domestic groups as reproductive cells of the social order, clustered within lineages that served as dominant segments of Yoruba kingdoms. If the king (ọba) in council mediated the political interests of his ìwàrèfà chiefs, the chiefs promoted the interests of the “houses” or “lineages” through which their corporately held titles devolved. There is already significant slippage within the model itself, because the jurisdictions of civil chiefs typically extend beyond the descent group as such to clusters of associated lineages and houses. In the Ekiti region where I worked, lineages were grouped into “wards” or “quarters” (àdúgbò), and these were the recognized domains of chiefly jurisdiction. But here too, the semantic slippage was extreme. In the relatively decentralized kingdom of Ishan, for example, each quarter was named after the dominant lineage or ilé (“house”) in which its chieftaincy was vested (Apter Reference Apter1995: 376–78). But more inclusively, these quarters could also be considered “towns” in their own right, marked by distinctive “dialects” and orisha (deities) within the kingdom at large. I wondered why Lloyd (Reference Lloyd1954; Reference Lloyd and Lewis1968) had overlooked these broader political sodalities, which were so central to the political organization of towns like Ado-Ekiti where he had worked. In a sense, Lloyd abstracted the simplest, descent-framed paradigm as the primary political unit, and saw more inclusive co-residential arrangements as corporate, and in some cases cognatic, manifestations of the lineage principle itself.Footnote 6
In what might seem as a minor footnote in Yoruba kinship studies, Bender (Reference Bender1970) challenged Lloyd's agnatic and cognatic variations on methodological criteria. Whereas for Lloyd, and much British social anthropology of the time, the corporate lineage was structurally primary, for Bender, with an American culturalist emphasis, the patrilineal idiom was logically prior to the actual social groupings that it generated on the ground. He argued that whatever the localized corporate groupings we may encounter empirically—predominately agnatic in some areas, cognatically mixed in others—we should not conflate descent as a symbolic idiom of genealogical reckoning with the residential, economic, and demographic characteristics of the actual groups. The debate is significant because it approaches socio-semantic slippage between meaning and reference as a problem to be resolved one way or the other, rather than—as I shall suggest in due course—as central to the very deployment of the category “ilé.” Despite their different approaches to the shifting meanings of descent, however, both Lloyd and Bender saw it as a building block of Yoruba political and social organization, generating those patterns of opposition and inclusion which characterized the kingdom-in-council. The dominant conception that households and lineages exist in towns, and build higher-level political relations between their representative chiefs, remained unchallenged.
In a break from this foundationalist vision of descent, Eades (Reference Eades1980), following Fadipe (Reference Fadipe1970: 97–118), turned the genealogical meaning of ilé on its side, treating it as a residential group or compound first and foremost, with lineage groupings within its walls, rather than as the residential correlate of a “prior” corporate lineage. Such a view not only matched Yoruba conceptions of the “compound” as the primary unit of affiliation and corporate organization within towns, Eades argued, but also directly resolved the agnatic-cognatic debate by identifying a bilateral kinship ideology that is effectively “pushed” in a patrilineal direction due to the primary factors of virilocal residence and economic cooperation between fathers and sons (1980: 49–54).Footnote 7 The agnatic descent group as household “core” emerges, in this view, as a secondary consequence of primary residence patterns. Households as primary building blocks “are best seen not as localised families, but as groups of co-residents, some of whom are related” (ibid.: 51). The “secondary” character of agnatic descent is further evidenced by the “Hawaiian” terminology of Yoruba kinship, which makes no distinctions between lineal and collateral relations.
The admixture of relational types within the compound—residential and lineal, consanguineal and affinal, core and stranger, freeborn and slave—is further complicated by patterns of internal segmentation around “half-brothers” (ọmọ baba) of the same father but different mothers, reinforced by inheritance rules. If groups of full siblings (ọmọ ìyá) form minimal units of segmentary opposition vis-à-vis domestic resources, over generations the same principle of differentiation extends throughout the more inclusive compound (ilé) of smaller households consolidated within larger political units. Not only are the primary factors “governing” composition of the ilé ambiguous—agnatic for Bender, agnatic and cognatic for Lloyd, and residential for Eades—but the referential scope of ilé remains fluid, opening up to frame larger political groupings of related compounds, and narrowing all the way down to distinguish sibling groups within polygynous households. Indeed, the core unit of Yoruba social and political organization has defied clear demarcation and definition.
THE TOWN IN THE HOME
In her monumental study of oríkì orílẹ̀, the attributive praise-poetry that invokes “origins,” Karin Barber clarifies the fluid and dynamic parameters of the Yoruba ilé through the very discourse genre that governs its composition, since oríkì “play a part in the actual definition and constitution of groups” (1991: 138). Barber's subtle and nuanced exegesis of shifting origins and boundaries within the category ilé in Okuku reveals a complex, labile, and negotiable social field in which descent, residence, patrilineal and matrilateral ties, as well as cooperative arrangements between “attached” lineage segments and fictive kin are variably invoked and adjusted to include or exclude according to context. From this more actor-oriented perspective grounded in social and discursive practice, the “debate” over the primacy of descent versus residence becomes moot. After reviewing the arguments of both “camps,” Barber (ibid.: 158) concludes:
In Okuku, it was not possible, in the end, to propose either the ‘compound’ or the ‘lineage’ as the fundamental social unit. Rather, the principle of descent and the principle of residence were entwined and interpenetrated at every level, down to the foundations of social identity. And this identity was continually redefined according to the circumstances, giving rise to different ‘groups,’ differently recruited in different situations, so that no single definition of a primary social unit was in the end possible. One term—ilé—was used for almost all significant groups: but it turned out to refer to different kinds of units in different circumstances.
From this more situated perspective, no objective “inventory” of ilé within any town is possible because the units themselves are flexible and fluid. As Barber (Reference Barber1991: 159) notes, her table of twenty-nine ilé in Okuku should not be seen in terms of “solid and permanent social units” but rather “as an indication of the range of possibilities open to social groupings as they adjusted their boundaries according to context.” Indeed, to treat the table as an objective inventory would succumb to what Bourdieu calls the synoptic illusion, when practical schemes of social classification that only “make sense” when strategically deployed are abstracted into fixed hierarchies that violate the actual “economy of logic” on the ground (Reference Bourdieu and Nice1977: 97–109).
It is not Barber's move toward a theory of practice, however, that I wish to highlight, important as it is, but an even more radical paradigm shift that her rich ethnography suggests. As we shall see, the fluidity of the category ilé is not a function simply of its strategic deployment and shifting boundaries on the ground, but also of a deeper mode of semantic configuration. We begin with Barber's paradoxical insight that the very oríkì orílẹ̀ with which “related” members of an ilé are praised, and which shape the flexible parameters of the compound unit, invoke not common ancestors on a genealogical tree but the towns of origin from which their forbears originally migrated:
Oriki orile are one of the principal means by which groups of people who regard themselves as kin recognise each other and assert their unity. But they do so in terms of a common town of origin, and not, in the first instance, in terms of ancestry. The key emblems in oriki orile are always associated with the names of places. Oriki orile do include allusions to illustrious men and women among the ancestors of the group, but these allusions are attached to the notion of town of origin. Oriki orile do not trace genealogies, nor do they revolve around the notion of a lineage founder. The members of a group assert that they are ‘one’ because they all came from the same place of origin, and distinguish themselves from people coming from other places (Barber Reference Barber1991: 145).
Barber reflects on how this association between kin groups and emblematic towns of origin might have developed, beginning with “primordial towns of origin” in which all members shared the same oríkì orílẹ̀ vis-à-vis outsiders from other towns, and distinguished themselves internally by lineage oríkì corresponding to “notional patrilineal kin groups” (ibid.: 148). According to this primordial starting point, the ilé must have originally designated agnatic lineages and lineage segments, and would only later become associated with distinctive towns or origin through the subsequent population movements associated with political fission, competition for resources, slave-raiding campaigns, and the nineteenth-century wars. As some towns expanded and others were sacked, regrouping under warriors and strong men who offered protection and war-booty, a resettlement pattern emerged of reconstituted “houses” defined primarily by their towns of origin.
There is no question that the turbulent warfare of the nineteenth century following the collapse of Old Oyo circa 1836 generated much population movement and resettlement, giving rise to new sociopolitical amalgamations. My own case study of the kingdom of Ayede provides an example of just this process, of stranger lineages and quarters settling around an original “core” to defend against Ibadan and Nupe predators (Apter Reference Apter1992). But I also would argue that such resettlement patterns are built into the cultural frameworks of Yoruba homes and towns in the first place. Barber appears to acknowledge this possibility when she reflects on the earlier forms of ilé:
It is quite possible that before the nineteenth century, there was less diversity in the ile and in the relationships between them, fewer ‘stranger’ groups and weaker bonds between and across compounds. Perhaps ile were more unitary and more strongly bounded. On the other hand … [b]earing in mind the probable high degree of population mobility before the nineteenth century, one may suspect that flexibility of group boundaries, and the possibility of invoking a variety of principles of recruitment, was already present in the social system, to be drawn on in different ways and with increasing intensity as the need increased (1991: 164).
It is just this latter consideration that I would like to push further. Rather than speculate about primordial towns in a somewhat mythic past, I would argue that the primary “place” of the town (ìlú) within the home or compound (ilé), as invoked by oríkì orílẹ̀, is built into the cultural category of ilé—that the very location of the “town” in the “home” is always already primordial. If such a radical reconceptualization appears counterintuitive at first, it more accurately reflects Yoruba concepts of group membership based on principles of semantic designation and logical quantification that differ significantly from their English language counterparts (Verran Reference Verran2001).
One of the first implications of this paradigm shift—placing the town within the home—concerns notions of socio-logical priority. As long as we see lineages-cum-residential ilé as primary units of sociopolitical organization, as building blocks of the segmented groups that vie against each other for political power and control, then it makes no sense to say that they come from elsewhere, from “outside” the system itself. If towns are made up of houses, then how are houses made up of towns? Yet from a Yoruba perspective this conundrum can be solved. Not only does the town in the home make perfect sense, but once grasped, it resolves the agnatic versus cognatic debate, the “lineage” versus “residence” debate, and obviates appeals to primordial towns. Getting there, however, requires a quick detour through Yoruba number theory.
THE WHOLE IN THE PART
“The different practices in classification that underlie the generation of predicating terms in English and Yoruba create different types of referring categories” writes Helen Verran (Reference Verran2001: 186) in her brilliant study of the logic of Yoruba quantification. There is no way I can do justice to the depth and sophistication of the argument she develops, which itself instantiates the very recursive tallying embedded in Yoruba enumeration. Rather I shall grossly simplify one of her major insights in order to illuminate an important difference between the English and Yoruba semantics of number. In English, and “Western” epistemology more generally, things, objects and numbers in the world are conceived as “spatiotemporal particulars,” individual entities that form collections of specific kinds and types—in more formal terms, as members of an abstract set. The number six is a collection of six objects forming a group or set, or more abstractly, six units of “one” in an extended series on a number line; a family of four is comprised of four individually related persons; a collection of residents living in houses within a spatiotemporal area combine to form a neighborhood. In Yoruba language and culture, Verran shows, things, objects and numbers in the world are conceived as “sortal particulars,” qualitative sorts of “thinghood” that infuse the universe and which manifest themselves in different modes at particular times and places. Sortal particulars can manifest themselves within a plurality of objects that form what “we” would see as members of a set, but the “objects” themselves are secondary to the sortal particular which they instantiate. “Number, in Yoruba language talk, is a degree of dividedness” (ibid.: 198). Things, objects, and numbers in the world are modes of manifesting sortal particulars in a given situation, time, or place. Five oranges are not five individual oranges forming a group, but “orangeness” divided into a plurality of five. Set membership is not additive; rather it is differentiating or decompositional—it starts with the whole and breaks it up into parts. A family of four is the sortal particular of familyhood broken into four related persons; a neighborhood is a sortal particular of neighborhoodness broken down into residents and their homes, according to its mode of manifestation at a particular time and place.
A few well known Yoruba motifs and practices illustrate this logic of objectification and quantification very clearly. There is the myth of Orishanla as the primordial orisha who was sabotaged by his slave when the latter rolled a large boulder onto his master and smashed him into hundreds of pieces. When Orunmila tried to collect the pieces and put “Orisha” back together again, those fragments left behind became the different orisha in Yorubaland (Idowu Reference Idowu1962: 59–60; Beier Reference Beier1980: 6–7). There are many interesting interpretations of this classic myth, especially as a master-slave dialectic, but the point I want to highlight here is how the very multiplicity of the orisha themselves represents a mode and degree of dividedness relative to a primordial whole. Naming ceremonies also illustrate individuation from a prior whole. Much is made about how Yoruba naming ceremonies held seven days after the birth of a child weigh against infant mortality before assigning a social identity, but I would underscore how the ceremony itself is a ritual of nominalization, individuating the child from the house or lineage from which he or she emerges. In a sense, the child is not “born into” a house or lineage, but is born out of it, coming to manifest the lineage through the ritual mode of naming itself, as in the proverb: “Ilé ni à ń wò k'á tó s'ọmọ l'órúkọ” (“It is to the house that we look before we name the child”).Footnote 8Oríkì themselves, as “modes of attribution,” are also central mechanisms of manifesting sortal particulars—the very shifting boundaries of inclusion and exclusion invoked by paternal lines, maternal lines, towns of origin, and significant landmarks instantiate “homeness” or “compoundness” in its plurality of forms at particular times and places. The meaning of ilé becomes less a question of what it “is” than of how—in what mode—it manifests.
When we return to the “place” of the Yoruba town in the home, with Verran's manifesting modes in mind, we can see that the “building block” approach to household and lineage, in which individual ilé combine to form increasingly inclusive “compounds” and quarters, is doomed from the start. The search for the primary principles of these primary units, be they descent (agnatic or cognatic), virilocal residence, or some basic entanglement of the two, approaches the problem from the wrong direction. The very “origin” of ilé as a prior “town”—the town of origin indicated by oríkì orílẹ̀—implies that the home or compound is a manifestation of the town; it is a mode of dividing the town into ilé, or a mode of dividedness of the town itself. We thus start from the town as it subdivides into ilé, not from the ilé as they combine to form towns. From the standpoint of Yoruba quantification, whereby “objects” emerge from prior wholes, every ilé is a manifestation of ìlú, a microcosm, as it were, of the kingdom at large.Footnote 9 And just as every kingdom has a royal lineage with exclusive claims to the kingship, so every ilé has an agnatic core with exclusive claims to a lesser political or ritual title. The primordial ilé must have non-agnatic “extras” in the form of maternal kin, attached strangers, or fictive kin because if every ilé is potentially the ìlú which it manifests, it must contain within itself the basis of differentiating royal from non-royal lines. In this model, the core ilé must contain both agnates and non-agnates, or outsiders living at the very center of the compound.
Historically, the dynamics of political fission illustrate how this actually occurred. Disaffected princes, ambitious civil chiefs, or entrepreneurial big men who accumulated clients and resources frequently broke away to found their own kingdoms, establishing ruling dynasties with beaded crowns that eventually invoked the authority of Ile-Ife. Or within a kingdom, a rival house could usurp the kingship to inaugurate a new ruling line. And if a powerful “house” could expand into the very kingdom that it “manifested,” so a “kingdom” could “shrink” into a quarter or house within a town. In Ayede, when the Eshubiyi “line” rose up in the nineteenth century to usurp the kingship from the Olua house, the latter dwindled to a shadow of its former glory although its members retained ritual icons of sovereignty associated with their orisha, and still refer to themselves as a “town” (ìlú) (Figure 1; Apter Reference Apter1992: 45–54). Local histories often recall how subordinate towns that seceded from the center eventually rose up to displace their former rulers, reconstituting, as it were, the town in every home.
Figure 1 Priest of Orisha Olua, representing the former ruling house in Ayede. From Apter (Reference Apter1992: plate 15). Author's photo.
The same cosmological whole in the part governs the manifesting modes of orisha worship as it transposes shifting boundaries of ilé and ìlú between social, political, and ritual domains. As I have argued elsewhere (Apter Reference Apter1992: 149–61), the orisha are simultaneously one and many, allowing followers to form around a lineage core, lineages to combine within the jurisdiction of a quarter, or emerging royal lines to incorporate the gods of their displaced predecessors. But there is one ritual transposition I would like to highlight because it explains how a “town of origin” associated with an ilé becomes, or comes to manifest, the town where it resides. For during the celebration of its orisha, the “house” qua town shrine (ilé òrìṣà) invokes its forbears and literally occupies the town, bringing its power from the outside bush into the town shrine, dominating the streets and crossroads, entering into other shrines and houses, stopping traffic, imposing fines, and taking possession of public space. During such ritual takeovers, the “house” of the orisha becomes the palace and kingdom that it manifests, and assumes control over the town.Footnote 10 Its priests and priestesses represent the kings as well as the military and civil chiefs of the former kingdoms and towns from which they migrated. Such ritual maneuvers are powerful because they manifest the potential of a real political takeover as well.
Town shrines as ritual “houses” are internally organized both as palaces (ààfin) and kingdoms (ìlú), mirroring the actual palace of the kingdom with its expansive courtyard, verandah, parlor, and inner chambers, while protecting within its walls the ritual crowns, calabashes, and deities of the orisha's town of origin. In principle, all kings, crowns, and orisha come from Ile-Ife as the sacred locus of original kingship, and in this sense all shrines manifest Ile-Ife within their cores. Historically, however, they represent the towns and quarters from which they migrated, as strangers joining a group of indigenes, disaffected political factions founding new kingdoms, or persecuted communities that were forced to relocate during the turbulent nineteenth-century “inter-tribal” wars. In Ayede-Ekiti, the three dominant orisha cults consolidated in the 1850s represent the composite character of the political community as it was centralized under the warlord Eshubiyi. The royal Yemoja cult, with its associated orisha, represents Ayede's link with Ibadan, from which the Àtá Eshubiyi received his crown, and reassembles the prior constellation of Oyo-centric deities that the Ibadan warlord Oluyole reconstituted, according Oyo's Shango a central place. As the Yemoja and Shango priestesses lead the ritual procession throughout Ayede, they carry the beaded crowns and calabashes of kingship, which they incarnate through spirit possession. The priestesses receive salutations of “Kábíyèsí!” (“Your highness!”), to which they respond “Ẹṣẹun!” in a hypercorrected accent that explicitly marks Ibadan provenance. The cult of Orisha Ojuna represents immigrants from Ikole, with an altar dedicated to the Elékọlẹ́, recalling the Àtá Eshubiyi's town of origin before he married into Iye and founded Ayede. The non-royal cult of Orisha Iyagba represents immigrants from the Yagba kingdom of Alu and its associated towns, glorifying the military chieftaincy of the Balógun-Ààfin with incipient icons of rival kingship. If Orisha Iyagba is a non-royal cult within Ayede, it nonetheless manifests Yagba kingship in its festival praises and ritual iconography.
I have discussed the political dynamics of ritual mobilization in greater depth elsewhere (Apter Reference Apter1992), but here I wish to highlight the ritual manifestation of the whole in the part as town shrines double as kingdoms unto themselves. Indeed, town shrines are internally organized as miniature governments. Fadipe explains: “The priesthood of every òrìsà is organized on the model of the political system. There is not only a hierarchy of officials, but these officials also bear titles that have been adopted from the secular government” (1970: 284–85). This model of the orisha cult as ritual monarchy or kingdom is not an aberration of an “original” family or lineage-based cult that evolves over time into complex forms, as one eminent scholar of Yoruba religion has maintained (Idowu Reference Idowu1962: 130–32), but rather constitutes the primordial form of the orisha cult house as it manifests the whole in the part, from which the genealogically defined lineage or family emerges. Nowhere is this mode of particularization more clearly evidenced than in Cuba and Brazil, where the Lucumí and Nagô “nations” developed into genealogically defined ritual “houses” and “families,” replete with initiatory blood-lines and branches.
ATLANTIC TRANSFORMATIONS: LUCUMÍ AND NAGÔ
There is no question that the lexicon of African “nations” that emerged in the era of Atlantic slavery played an important if complex role in the development of ethnic identities in the Americas as well as in West and Central Africa. We know that much was invented on both sides of the Atlantic (Palmié Reference Palmié, Apter and Derby2010), as well as during the Middle Passage (Mintz and Price Reference Mintz and Price1992), making the identification of New World “nations” with prior source populations in Africa problematic at best. Curtin (Reference Curtin1969) gives three reasons for confusion, stemming from the interests of slavers and planters: (1) the ethnic lumping of different groups in Africa under an umbrella ethnonym, such as “Mandingo”; (2) the identification of Africans with their ports of embarkation (e.g., Mina) rather than the hinterlands from which they may have arrived; and (3) the association of slaves with behavioral stereotypes in the colonies, such that “rebelliousness,” “docility,” physical strength, or entrepreneurial propensities became ascribed “national” characteristics by which they were classified. Yet rather than occlude ethnic origins as such, these factors help us understand how African “national” identities emerged out of Afro-European encounters, subsuming specific populations within broader named collectivities. Much also has been said, following Bastide (Reference Bastide1971; Reference Bastide and Sebba1978), of “the new social frameworks” that reshaped neo-African communities in the Americas, including patterns of labor segmentation, color stratification, and the underlying dynamics of class formation that motivated re-Africanization among free blacks. Throughout the Iberian imperial landscape, the religious brotherhoods of the Catholic Church stand out as especially formative institutional loci in the making of New World African identities, serving as “generative bases” (Brown Reference Brown2003: 34) or “incubating cells” (Palmié Reference Palmié and Binder1993: 341) of ethnically denominated socio-religious groups such as Congo, Hausa, Ibo, Carabali, Jeje, Mina, Fanti, and Ganga, among which the Lucumí of Cuba and the Nagô of Brazil emerged as preeminent ethnic “nations.”
The historical sociology of New World Catholic brotherhoods—cofradías and cabildos in Cuba, irmandades in Brazil—involves a complex web of economic, sociopolitical, and religious connections that begin with sixteenth-century colonization, when religious fraternities of patron-saints incorporated enslaved Africans into religious sodalities that doubled as social clubs and mutual-aid societies. Based primarily in cities and provincial towns, the cabildos de nación organized African-born slaves according to their professed “nations” of origin, establishing crucibles of “syncretism” and “creolization” as African-derived religious and cultural practices coalesced within a Catholic frame. If, as administrative arms of Church and state, these brotherhoods recognized the authority of the governor, they also pushed toward political autonomy as self-governing corporations that financed manumission for their members, and in extreme cases, fomented rebellion.Footnote 11 It is not the politics of the brotherhoods as such that I wish to emphasize here, however, but their development from miniature kingdoms into ritual houses during the nineteenth and twentieth centuries.
Lucumí Counterpoint
Like the orisha cults of Yorubaland, the cabildos de nación in Cuba were modeled on the political system in which they were embedded, incorporating monarchic, military, and religious orders within their iconography and institutional organization. As David Brown explains in his formidable study of Cuban Santería, all of the cabildos and their associated ethnic “nations,” whether Lucumí, Congo, or Caribali, initially emerged within the same organizational framework, even if they differed in influence and prestige: “The cabildo de nación borrowed the title of official state, church, and civic institutions, especially from the monarchy and the military, to describe their offices. Cabildo kings…were supported by a ‘court,’ consisting of a ‘queen’ or two (second in rank), and lower-ranking officers, such as the abanderado (standard bearer) and the mayor de plaza (chief of ceremonies)” (Brown Reference Brown2003: 35).
Cabildos were, in effect, “miniature neo-African monarchies” (ibid.), referred to as reinados (“kingdoms”) in popular discourse. Represented by flags that linked African “nations” to Catholic saints, their members paraded under the umbrella of the Church, taking to the streets during Carnival, saint's days, and the Day of Kings, while demanding small payments (aguinaldo) from onlookers as a form of ritual tribute and protection. Entering Havana from beyond the city walls (extramuros), the “African” kings and queens in procession constituted something of an occupying force: “The group literally took over the street from curb to curb in a public space, where its collective identity and complex hierarchy were on display” (ibid., 36–37). Fusing European codes of royalism and militarism with African drum rhythms and associated choreographies, the cabildo processions established performative arenas of symbolic appropriation and sociopolitical empowerment, where enslaved blacks could flex some muscle within limits uneasily tolerated by the authorities.Footnote 12
There is no question that European markers of hierarchy, power, and socioeconomic value fed into cabildo iconography and processional conventions, producing an Atlantic Creole assemblage that circuitously fed back not only to Freetown and Lagos, but also into the Nigerian hinterland. If the vectors of transmission are not always clear, the circum-Atlantic parameters are expansive. Nunley (Reference Nunley1987) shows how the Ode-Lay masquerades of Freetown were embellished with a “fancy” and “fierce” aesthetic that carried with it New World finery; Roach (Reference Roach1996) traces performative chains of ritual surrogation which link Mardi Gras pageants to the Elizabethan stage. Indeed, Brown (Reference Brown2003: 39) cites a contemporaneous account of an 1856 Day of Kings festival in which a Havana cabildo ruler sported “a genuine costume of a king of the Middle Ages, a very proper red, close coat, velvet vest and a magnificent gilt paper crown.” Yet such carnivalesque phantasms of feudal monarchy are not so easily reduced to plantocratic kitsch, for they reference a broader royal lexicon imbued with deadly authority. In the orisha festivals of Ayede-Ekiti, for example, the Àtá (ọba, king) wears not only a European-styled beaded crown, but also royal robes of embroidered àrán cloth made from an imported European red velvet brocade reserved exclusively for kings and their ritual representatives (Apter Reference Apter1992: 108–9). My point is not to reduce sacred kingship to African replicas of European replicas “all the way down,” but rather to emphasize the African grammars of sovereignty motivating the appropriation of European emblems and signs, a coterminous process of royalist codification that occurred on both sides of the Atlantic.Footnote 13
With this “deeper” cultural logic in mind, we can revise the standard depiction of cabildos as miniature Bourbon states—providing safe-haven for African gods and “kings”—by emphasizing a much greater continuity with orisha cult performance and organization than has generally been acknowledged. If much has been written on the dynamic correspondences between Yoruba deities and Catholic saints, less has been said of their isomorphic performance genres; between saint festivals and the Day of Kings in the Catholic calendar, and the public annual festivals culminating in the Day of Carrying Water (ijọ ipọnmi) for the orisha. In formal terms, when Yoruba priestesses mobilize their cults to take over the town, they take possession of space and time. As they process from bush shrine to the market and palace, they honor political patrons and historic locations with propitiations and oríkì, demanding tribute from those praised, fining cars and lorries that interrupt them on the roads, and reinscribing public space with the history of the cult. “Today does not own itself,” they sing, “today belongs to X,” naming the principal deity of their orisha cult. In Ayede, the Yemoja and Orisha Iyagba processions intersect, waging muted ritual battle as both groups take over the town. Iyagba warrior priestesses brandish cutlasses and spears, mobilized by military rhythms as they confront Yemoja's avatars from Ibadan and Oyo. Each cult empowers the king and the kingdom in its own name, activating its distinctive past to bear upon the present. Each cult also manifests “hot” and “cool” gods with characteristic choreographies and rhythms—measured and staid versus rapid and staccato—emphasizing reproduction and transformation, authority and power. During the orisha's festival, the kingdom that it reproduces and transforms becomes the cult's kingdom of origin.
The processions of Catholic virgins and saints mirrored the outings of the orisha with striking symmetry, converging within the cabildos de nación in both formal and substantive terms. Both traditions carry icons invested with power through public streets and pathways before delivering them to their sacred destinations. As in the cabildo processions, the orisha priestesses take over public space, exacting economic tribute for cult support and protection just as aguinaldos were demanded in Havana. Like the miniaturized court societies of the cabildos, the orisha cult priestesses-in-procession formed ritualized kings-in-council, with associated chiefs and warrior priestesses among the inner entourage. Furthermore, the “fancy” and “savage” aesthetic opposition that characterized European versus “African” codes (Brown Reference Brown2003: 47–51) in Cuba, contrasting formal decorum and comportment against those raffia fringes and animal horns that dominated periodic break-out sessions of animated drumming and dancing (corros), were already embedded within orisha cult registers that indexed the “cool” hierarchy of the kingdom against the “hot” powers of transformation and subversion. Even specific ritual syntagms carried over. Describing a “royal” cabildo procession in 1844, the American physician J. G. F. Wurdemann noted, “The whole gang was under the command of a negro marshall, who, with a drawn sword, having a small piece of sugar-cane stuck on its point, was continually on the move to preserve order in the ranks” (cited in Brown Reference Brown2003: 37). For David Brown, this telling motif suggests a potential assault on slavery itself: “What emblem could better embody the bitter-sweetness of black carnival in the period's premier sugar-producing slave society than this precious detail? The entire Cuban ‘sugar palace,’ at its height between 1820 and 1860 … revolved around that piece of cane, now sword pierced and held aloft by the formally dressed and officially empowered black officeholder, who was also a slave. Did his performance slyly comment on ‘cane’ as the ‘signifying’ slave songs of the antebellum United States ‘sang’ upon ‘corn’…?” (ibid., 47–48). Without denying the plausibility of this interpretation, given the surplus of meanings generated by carnival, I would only underscore that this “precious detail” is also found in Ayede's orisha festivals, where, among the category of ológun “field marshals,” priestesses waging ritual battle carry spears and cutlasses tipped with kola nut pieces—instruments of death capped by gifts of life.Footnote 14 There, the kola nut serves as a minimalist protective cover, highlighting the fragile containment of militant dismemberment so easily unleashed.
My intention is not simply to pinpoint the Yoruba “origins” of specific ritual patterns and practices, but to show that the core ritual systems within orisha cults and at least some of the cabildos de nación were virtually the same in the nineteenth century—they replicated within their restricted codes and corporate organization the larger monarchies in which they were embedded, while also repossessing the public sphere during processions of collective renewal and empowerment. Moreover, the quantification of the whole in the part informing the òrìṣà-ocha system helps account for two trends that have inspired considerable debate: The first concerns the scope and character of Lucumí influence in the colonial period vis-à-vis other ethnic nations and religious societies that developed coterminously. The second addresses the historic transformation of Lucumí cabildos into “house-temples” (casa templos).
The question of Lucumí influence is all the more puzzling given the lower number of their cabildos compared with those of the Congo and Carabalí nations (Brown Reference Brown2003: 63; Lopéz Valdés Reference Lopéz Valdés1994). How did a minority ethnic nation emerge as such a dominant Afro-Cuban identity? Part of the answer seems to lie with its religious culture's “absorptive” capacity. Characteristically open to “stranger” deities, the Yoruba-Fon-related religions of Cuban Santería and Brazilian Candomblé have incorporated European, Congolese, Indian, and Creole gods and spirits within their “houses” and altars, thereby gaining ground in the field of Afro-Caribbean religious production while setting the stage for subsequent contests over purity and authenticity. The same principle of ecumenicalism seems to have applied to Lucumí organizations that not only consolidated those subgroups (Oyo, Ketu, Ijesha, etc.) speaking varieties of what we now call the Yoruba language, but also came to claim as its ritual “descendants” non-Yoruba speaking nations as well. In what may well have been a remodeling of Spanish hyphenated naming conventions, we find Lucumí-Eyó (Oyo), -Egua (Egba), -Agguado (Egbado), -Iyecha (Ijesha) also extending to Lucumí-Achanti (Ashanti), -Fanti, -Popos (Dahomey), -Araras, -Benin, -Magi (Mahi), -Bariba, -Jausa (Hausa), -Ibo, and -Tapa (Nupe).Footnote 15 In addition to boosting Lucumí recruits and affiliates, this expansive Lucumí association with other nations earlier on may have further influenced the organization of their distinctive cabildos, by transposing the modular organization of the “miniature kingdom” into black cabildos across the board.Footnote 16
Such a broadening of Lucumí socio-religious jurisdiction might be cast as strategic consolidation at a time when the population of African-born blacks was declining and government assaults against black cabildos and people of color were on the rise, triggered by the 1886 emancipation of slaves. But the cultural conditions of such broad consolidation were built into the concept of a Lucumí totality, which came to manifest its segmented particulars both laterally, as sub-ethnic affiliations, and lineally, as members of initiatory lineages. As social kinship gave way to ritual kinship from the 1880s to the 1920s, the kingdom-cum-cabildo was reconstituted within the house or home.
Brown (Reference Brown2003: 62–66) identifies three historical trajectories by which the cabildo de nación gave way to the turn of the century casa-templo: one of continuous transformation; one of covert derivation; and one of more independent inauguration by ex-slaves and creoles who established ritual family lines. But in many ways these different trajectories represent manifestations of an underlying process of ritual segmentation and differentiation, giving rise to a “transitional” cabildo—an “umbrella organization connecting numerous emergent houses and ritual family lines”—that culminated in fully fledged family fission. Thus emerged the modern twentieth-century “‘house’ of Ocha … composed of a single extended ‘ritual family’ (familia de santo) directed by a single priestly elder who practices within a private domicile in more or less discrete or underground fashion” (ibid.: 67). Critical to Brown's analysis is how ritual kinship and kingship are mutually sustained. If ritual kinship is traced through the padrino (godfather) or madrina (godmother) as house owner and elder, actual initiation is performed by the ọba (lit. “king,” “master of ceremonies”) in the preparation and consecration of the neophyte's head.Footnote 17 Moreover, icons of kingship are intimately associated with the ritual prerogatives of the ocha themselves, whether channeling past kings of Oyo with Chango, or by symbolic and iconographic association with royal crowns, cloths, scepters, and tributes.
Revising Ortiz's theory of cabildo origins for the Cuban Lucumí-Santería religion, Brown places greater emphasis on eponymous ancestors as “root founders” (raíces) of houses:
A deeper ‘starting point’ of the Lucumí religion lies not so much in the cabildos as in revered personages who founded and led the religion's Ifá and Ocha ‘houses’ (casas or ilés) and gave birth to their corresponding ramas (‘branches’). A casa (ilé in Lucumí) refers not merely to the physical domicile of a priest, but to a ritual ‘family’ of priests initiated by an elder of Ocha.… Rama refers to the genealogical lineage or lineage tributary from which priests (and their houses) descend. Casas are the nodes, as it were, of the ramas, and these terms together constitute the sacred genealogical organizing principles of the Lucumí tradition.… Upon consecrating the ‘head’ of an initiate, a ritual elder—the ‘owner’ (dueño) of a ‘house’—becomes the ‘godparent’ (padrino or madrina) of a new ‘godchild’ (ahijado or ahijada). The new member of the house comes to have ‘brothers’ and ‘sisters’ (hermanos and hermanas) among the house's other initiated priests, both alive and deceased (eguns). The term ‘house’ can also subsume collegial or ‘working’ relationships with priests of other houses or ramas (ibid.: 74).
I have quoted this passage at length because it so clearly identifies the genealogical matrix within which Lucumí houses and their ritual lineages were produced, mirroring the characteristic principles of fission and fusion in segmentary “lineage” systems more generally. Although house lineages and their segments appear to continuously divide and break free over time, such fissiparous pressures are counterbalanced by consolidating trends of lineage fusion, when, as Brown points out, the house subsumes “‘working’ relationships with priests of other houses or ramas.” My point in highlighting this countervailing trend is not merely to complement fission with fusion, but to reestablish the immanent frame of the Lucumí casa (ilé) as whole-in-the-part. What Brown sees as a basic break between the old cabildos and the founding of new “houses” can be recast in much more continuous terms, as a mode of sortal—in this case “genealogical”—particularization, privileging an initiatory principle of lineal descent that came to dominate the “town” in the “home.” If the miniature kingdoms of the nineteenth century no longer shaped corporate architecture and organization, they nonetheless persisted within the domestic iconography of the casa templos, manifesting through the ọbá's rites of consecration, as well as in royal and warrior ochas, and their emblems, altars, and shrines.Footnote 18
Nagô Themes and Variations
The development of the Nagô nation within Brazilian Candomblé is so embedded within twentieth-century ideologies of Yoruba purity and authenticity that it is difficult to excavate its genesis from within the lay Catholic brotherhoods (irmandades) of the colonial period. In his foundational study of the African religions of Brazil, Roger Bastide (Reference Bastide and Sebba1978) explored the multifarious affinities between the Fon and Yoruba divinities of the Jeje and Quêto/Nagô nations and their saintly counterparts in Catholicism, locating them within a broad range of class contexts and regional variations. Salvador de Bahia represented the genuine or “pure” Candomblé for Bastide, with “corrupted” and “degraded” forms spreading inland and southward, reflecting the “anomic” consequences of industrialization and labor migration. Areas like Minas Gerais, where gold mining in the eighteenth century supplanted the declining sugar economy in the northeast, were hardly worthy candidates for the true Candomblé, where instead one found the so-called Bantu “sects” of “congadas,” festival groups organized by such black brotherhoods as Our Lady of the Rosary, St. Balthazar, Saint Benedict, Saint Iphigenia, the Girdle of Saint Francis, and so on. But it is here, in the mining communities of the colonial period that Bastide developed his analysis of Brazil's “two Catholicisms,” one white and European, the other black and Africanized, ritually reconstituting black kingship and court society.
Indeed, Bastide's discussion of these brotherhoods parallels the Cuban cabildo in crucial ways. Both served as mutual aid societies that rotated credit, guaranteed funerals and burials, and financed the manumission of a certain number of slaves each year (Bastide Reference Bastide and Sebba1978: 116). Both fomented a sense of shared ethnic kinship. Both took over the streets in public processions during holy days. But the most striking congruity concerns the royal offices and titles, and the protocols that these latter commanded:
…the still vivid memories of African kingship emerge even more plainly in the congadas. These festivals accepted the continuance of a monarchic regime for Brazilian Negroes—in an adulterated form, of course, and incorporated into the worship of Our Lady of the Rosary. The earliest mention of a congada is in the town of Iguarassu in Pernambuco in 1700, but it already existed, at least in fragmentary form, in the middle of the seventeenth century, and its origin can be traced back to Portugal. Pereira da Costa tells us that each parish had its king, queen, secretary of state, marshal, herald of arms, ladies-in-waiting, etc., who were addressed as ‘Your Majesty,’ ‘Your Excellency’ or ‘Madam.’ The election was held on the feast day of Our Lady of the Rosary and was the occasion for dancing, which varied in type according to the ethnic origin of the king (ibid.: 120).
Here we confront the miniature monarchy in motion, replete with entourage and royal salutations within a “Bantu” rather than Mina or Nagô ethnic nation. But such designations had a fluid content, indicated by the variable ethnicity of the king and associated choreographic styles. The festivals culminated in ritual coronation, processing to the chapel, where “the priest consecrated the man whom the brotherhood had chosen by placing a cardboard crown on his head” (ibid.: 121).
There is no question that Congolese slaves—especially those from the Kingdom of Kongo—would have recognized the marriage of Catholicism and monarchy in the brotherhoods from their own dynastic conversions, beginning with Nzinga a Nkuwu's baptism in 1491, and consolidated under his son Alfonso, who developed Christianity into a royal Kongo cult (Thornton Reference Thornton1984: 148). Though many of the African coastal societies were touched by missionaries during the centuries of the Atlantic slave trade, none appropriated the theocratic models of Catholicism as fully as did Kongo in the sixteenth century and seventeenth centuries. Yet it would be a mistake to attribute the Minas Gerais ritual monarchies to Congolese influences alone, since the majority of enslaved Africans during the formative years of the brotherhoods, from 1690–1750, came from the Mina coast, and were primarily Fon and Yoruba (Kiddy Reference Kiddy2005: 39–45). It was only after 1750 that Congolese slaves came to dominate demographically, putting their “Bantu” ethnic stamp on the black brotherhoods of Minas Gerais.Footnote 19 As Kiddy explains:
The documentation demonstrates that Brazilians had started importing Mina slaves into the port of Rio de Janeiro at the beginning of the eighteenth century, and many of those slaves were destined for the mines. These slaves became the earliest members of the rosary brotherhoods in Minas Gerais and participated in shaping the heterogeneous communities that the brotherhoods would become. Within the brotherhoods, West African slaves could begin to reconstruct their worlds around common cultural elements such as kinship, kingship, expressions of hierarchy, and a link with the ancestors (ibid.: 49).
In other words, the formative framework of black brotherhood royalism was primarily Fon and Yoruba in origin, establishing a West African politico-ritual “grammar” that was subsequently “filled” with Central African content, as in songs, rhythms, names, and even “nations.” I would argue, however, that beneath the “Bantu” exteriors that built up around it, a more fundamental Yoruba-Fon substrate endured, motivating the very formal transposition of whole into part, of kingdom into brotherhood.Footnote 20 I am aware that this line of interpretation can be pushed too far, succumbing to a cultural chauvinism not unknown among Yoruba specialists.Footnote 21 However, it is precisely the more abstract and formal features of generative cultural models that are most structurally determinative yet most difficult to perceive.
The complementary trajectory is easier to discern—how the Yoruba framework of gods and kings directly associated with the Nagô nation came to take over Bantu and other non-Nagô Candomblé houses, whether by absorption, imposition, or conscious appropriation. In Bahia, Bastide explains, despite the multiplicity of Congo and Angola houses, a general Yoruba paradigm prevailed: “The Yoruba imposed their divinities and the structure of their ceremonies on the other ‘nations.’ The result is that today everyone worships the same gods in his own language, with his own music, in structurally similar ceremonies.… Throughout this region the prestige of the Nagô was so high that rivalry compelled the other ‘nations’ to borrow the organization system of their cult, along with their orixás, which they identified with their own voduns or spirits. They borrowed not only the essential features of the Nagô rites but even their priestly hierarchy” (1978: 194–95, my emphasis).
From this Yoruba-centered perspective, all gods are orisha (orixa), all rituals, sacrifices, and invocations follow Nagô conventions, and all Candomblé nations replicate the core features of the ritual Nagô nation form. Even the cult of Indian spirits adopted Nagô norms, giving rise to the caboclo Candomblé houses. Whether rightly or wrongly, Bastide attributes such mimetic appropriation to Nagô's ruling status within the hierarchy of Bahian Candomblé houses, a position earned by its imputed fidelity to an authentic African tradition: “Thus the prestige of the ‘Nagô’ finally won out everywhere. Their prestige derived from their having upheld the ancestral religion most faithfully in the original form in which it had been brought to America by Ketu priests captured by the Dahomans and sold into slavery Bahia. Thanks to the initiation of generation after generation of new filhas de santo, vestals of the sacred fire, the tradition has been maintained without any adulteration or falsification” (ibid.: 197).
Bastide's line of argument reveals not only the modular reproduction of the Nagô ritual nation throughout non-Nagô Candomblé houses, but also the complicity of the anthropologist in ratifying claims of ritual authenticity in what would become a religious purification movement.Footnote 22 Central to Bastide's perspective, however, is a model of ritual reproduction through initiatory lineages that consolidates the “blood-lines” of Nagô purity within a competitive field of Candomblé houses, designated by ilé (“house”) in Yoruba and either casa (“house”) or teirrero (“yard”) in Portuguese.
We are now in position to trace the trajectory of Candomblé's privileged Nagô line from Afro-Catholic irmandades to the genealogically constituted houses and families that—in Yoruba terms—they came to manifest. The emergence of Candomblé houses within Afro-Bahian society is difficult to pinpoint, but Parés (Reference Parés, Falola and Childs2004: 189) argues that by the early nineteenth century a number of ethnic “nations,” particularly Angola, Jeje, and Nagô, were coalescing around work crews (cantos), Catholic brotherhoods, secular dance-drum gatherings (batuques), and what were already identified as Africa-derived Candomblé congregations. Parés (ibid.: 190) further suggests that a predominantly Jeje (Fon-Dahomean) religious template had already formed when the relatively late and concentrated influx of Nagô slaves in the 1820s began in Bahia, revaluing the Jeje-based “vudum” deities with the Nagô-Yoruba “orixá.” By the 1830s, Nagô ritual dominance was portended with the founding of the Casa Branca do Engenho Velho, also known as Ile Iyá Nassô after the mother of its putative founding priestess, which became known as the original Candomblé house from which the “pure” Nagô branches have descended. The dual-naming is significant because it points to two important originary frames: the first, as “the White House of the Old Sugar Mill,” invokes nothing less than the Great House of the sugar plantocracy, relating colorations of whiteness and purity that, I will argue, took on racial overtones. To this day, this “house” and those branches founded by its ritual descendants are known as the “great houses” of Candomblé (Matory Reference Matory2005: 125). The second designation as the house of “Iyá Nassô” frames the temple with reference to the its charismatic founder, Marcelina-Obatossi, whose mother held the title of Iyá Nassô in the Nagô-Yoruba Xango (Shango) cult.Footnote 23 Variant accounts of how together they returned to the kingdom of Ketu in Yorubaland, and thereafter brought back the “true” tradition to Bahia, establish an important return-to-the-homeland precedent for re-Africanizing Nagô-Candomblé lines according to authoritative standards of ritual purity.Footnote 24 But it is the logic of eponymous nominalization and the lineages which founding ancestors inaugurate that—as with the Lucumí casas de templos—I wish to emphasize at this juncture. The very plurality of names—the one institutionally framed, the other focused on personages—captures a movement of particularization in which lineal modes of genealogical reckoning emerge from the “house” and rise to the fore.
If cosmologically the Candomblé houses manifested Oyo and Ketu kingship within their walls as condensed “mystical geographies,” sociologically the two dominant “wholes” were Catholic brotherhoods and sugar plantations, representing the urban and rural “niches” in colonial Brazil where African religions reemerged (Bastide Reference Bastide and Sebba1978). Like the congadas of Minas Gerais further south, the Candomblé houses developed from brotherhoods. According to Johnson, “As a community … the terreiro Engenho Velho had its roots in the Catholic brother- and sisterhoods (irmandades), the men with Our Lord of Martyrs, the women with the order of Our Lady of the Good Death (Nossa Senhora da Boa Morte)” (2002: 75). It was out of this sisterhood that Marcelina-Obatossi emerged, founding the terreiro named after her “mother,” Iyá Nassô (ibid.) to inaugurate a lineage model of what Matory (Reference Matory2005: 125) calls “initiatic families,” thus setting the stage for segmentation and fission within the ritual lineage at large. As Johnson explains:
With the death of the priestess and leader, Marcelina-Obatossi, succession disputes led to fission and the splintering off of two new terreiros: one in the neighborhood of Gantois, Iyá Omi Ase Iyámase, usually referred to simply as Gantois; the other called Axé Opô Afonjá (the force of the staff of Afonjá). These three houses, along with Alaketu, comprise the traditional houses of the Nagô-Ketu nation, the trunk of the tree from which thousands of descendant houses would branch and flower. They serve both as the genealogical progenitors of many terreiros in Brazil and as the authoritative model of tradition and correct liturgy for many more (2002: 76).
Here we see in full evidence how the blood-based logic of ritual initiation within the “royal” line of Marcelina-Obatossi took precedence and continued to shape relations within and between Candomblé houses; that is, how blood became the dominant idiom of the “kingdom” within the cult. The connection between ritual kinship and the collective power of the terreiro is nicely captured by Matory's discussion of the Yoruba concept of àshe, a concept in Yorubaland that relates the animating force of sacrificial blood and associated medicines to personal power, agency, authority, and kingly command, but which in Brazil has come to further specify the very shared genealogical substance of initiatic families:
In Brazil, axé can be a countable thing. An axé is the membership of a temple (or a family of temples), united by the same continuous ritual transmission of axé from a founding priest or priestess to the ‘children-in-saint’ whom they have initiated and so forth. Thus, one might ask a worshipper which axé he or she belongs to. In reply, he or she might identify the initiatic lineage by the name of its founder, its temple of origin, or its currently ranking temple. An axé (temple community or family of temples) is held together by shared axé (ritually constituted life-force) and, ideally, by a shared set of ritual conventions that are supposed to have remained unchanged since the founding of that axé (family of temples) (2005: 124).
Through the lens of axé, we can see how the great White House of the Old Sugar Mill, sometimes referred to as just Casa Branca (White House), came to manifest that of its founding priestess, Marcelina-Obatossie, disseminating through lineages of her initiated god-children. The very names of those houses that broke away through fission are at once “linked to the axé of Iyá Nassô” (ibid.: 125), yet emphasize their own distinctive identities as crucibles of their own axé. The “Iya Omi Ase” of Gantois means “mother of the water of àshe” in Yoruba, at least initially referencing the ritually loaded term “water” which in orisha worship refers to the reproductive “blood” of mothers as a potent manifestation of collective àshe (see Apter Reference Apter1992: 97–116), whereas the “Axé Opô Afonja” references the àshe of the historic Are-Ona Kakanfo of Ilorin who, as head of the army of the Oyo Empire, led a rebellion against its king (Johnson Reference Johnson1921: 191–92), thereby serving as an effective charter for a Candomblé house that was born of fission.
But as Matory has also emphasized in his quotation above, “axé can be a countable thing,” a statement that accurately captures, from a Yoruba perspective, the emergence of quantity out of quality or thinghood; that is, from an encapsulating category or totality of axé rather than additive quanta forming aggregate sets. That the sense or reference of this totality may change according to spatial location and historical context does not detract from its “countability” through particularization, but only underscores the variable “wholes” through which its parts—as cult initiates—are manifestly made. The members of a Candomblé “house,” like those of an orisha cult or Lucumí casa, are not recruited or assembled as a subset of a given population; rather they are generated by axé, ritually crowned and reborn through initiation. One vector of transmission, the town in the home, ritually transposed as the kingdom in the cult, was manifested through the historical emergence of the Candomblé house from the Catholic brotherhood, in which the various kingdoms of God, the colonial state, and the Dahomean-Yoruba homelands were reinscribed. Another powerful vector of transmission motivating discourses of ritual purification, during what Parés (Reference Parés, Falola and Childs2004: 191–98) identifies as the first and second phases of Candomblé's Nagôization (1870s–1930s) appears to derive from the earlier sugar plantations of the nineteenth century and their caste-like ideologies of racial purity and miscegenation.
Like the Great Houses of Nagô Candomblé, the casa grande of the sugar plantocracy was obsessed with hierarchy, rank, and purity. Although much has been written both for and against the myth of racial democracy in Brazil, and its antecedent ideologies of color stratification under slavery, there is no question that ideas about honor, status and blood naturalized productive relations on the sugar plantations. Bastide describes the casa grande as the hub of the plantation system at large:
The master's family was endogamous; it wanted no black blood in its veins. A wife was chosen with an eye to her racial purity and her fitness to bear her husband's children and propagate his line.… The intermediate class consisted of poor whites who could survive only by integrating themselves as dependents within the only stable units the colony, the big landowning families, and of mulattoes or free Negroes almost totally assimilated into the Portuguese civilization. The house slaves were selected for their beauty, intelligence, health, and cleanliness from the Creole blacks or the Mina or Nagô Africans, i.e. almost exclusively from the West African group. The field hands were unusually Bantu or semi-Bantu. In short, social status increased with proximity to European values as represented by the master and his wife (1978: 68).
Even if we allow for a certain degree of stereotyping in this portrayal of the casa grande, the language of racial purity was clearly central to an ethnically stratified productive regime ruled by white “blood,” followed by descending orders of status and labor correlated with increasing degrees of “polluting” blackness.Footnote 25 Upward mobility was extremely limited for blacks within this hegemonic system, resulting in strategic miscegenation from below, which, again following Bastide, was “characterized by the expression ‘limpar o sangue,’ the purging of the blood by sleeping with whites and producing children with lighter skin, whose white fathers would help them along and who might in this way be freed from the yoke of slavery and enjoy an advantage when it came to economic competition” (ibid.). From the standpoint of black bondage, if such “cleansing” led toward freedom, for white planters it threatened the categorical separations that supported their power and privilege. Thus the fear and threat of racial mixing encouraged the policing of bloodlines from above, excluding mulattos from the inner circles of elite marriage, education, and society while criminalizing blackness at the margins. The Brazilian Creole was paradoxically a sign of “progress” and a creeping threat to white overrule.Footnote 26
When seen against the historical backdrop of the casa grande and its logics of cleansing, the cultivated purity in the “great houses” of Bahian Candomblé resonates with striking symmetry. They maintain the “pure” Nagô-Ketu-Oyo tradition where other houses are “syncretistic,” mixing categories and practices from dubious sources among mestizo-caboclo spirits and their devotees. Nagô orthodoxy and ritual purity not only claim fidelity to African precedents, but also refer to the quality of a House's axé, that ritual blood of the initiatic family that is channeled by lineages and produces new “children.” Like the white plantation overlords, the “matriarchs” and “patriarchs” of the Nagô houses regulate the propagation of “pure” lines of descent, uncorrupted by the dubious ancestors (egum) and mixed spirits (caboclos) which are kept at bay (Matory Reference Matory2005: 129). There is a systematic inversion and displacement at work that in some ways reverses the casa grande, since the telos of ritual purification is embedded in a project of re-Africanization. Yet to say that blackness triumphs over whiteness in this context misconstrues the colorations of race and nation in Candomblé, for ritual whitenesss sustains the language of purity in the cooling white (funfun) deities of Oxala and his congeners. To be sure, the mulatto mixing of black and white is displaced onto the mestizo mixing of Indian and white in the unstable category of caboclo, which in Nagô temples embodies the essence of ritual pollution, and includes Bantu admixtures as well. But such shameful taints can be overcome through healing rites of “cleansing” (limpeza) that send meddlesome egums and caboclos away (ibid.: 130–31), or by seeking connections with a “purer” house. As Matory explains: “Any given priest and his or her temple might abandon their original axé and seek affiliation with an older and more prestigious axé,” engaging in a form of ritual upward mobility that further benefits one's initiate-offspring (ibid.: 125).
What I am proposing as an extended hypothesis is that the logic of ritual purity in the great houses of Nagô Candomblé reworked the racial ideologies of blood and stratification in the casa grande of the colonial sugar estates, transposing principles of kinship and descent from social to ritual domains.Footnote 27 To be sure, the diasporic returns and commercial interests of free black Bahians in the nineteenth century, and the codifications and ratifications of anthropologists in the 1930s, were crucial to the Nagôization of Candomblé, as has been demonstrated by Matory (Reference Matory1999; Reference Matory2005), Johnson (Reference Johnson2002), Parés (Reference Parés, Falola and Childs2004), and Capone (Reference Capone and Grant2010), and this helps explain how Nagô temples converted their ritual resources into social and political capital. But the idioms of royalism and genealogy which mediated and motivated these broader exchanges were generated by Yoruba scheme-transpositions “from within.” It is not the cultural models of kingdom (ìlú) and house (ilé) as such that I have emphasized, but the modes of manifesting the former within the latter, a process that has brought the parallel developments of Lucumí and Nagô houses into clearer focus.
***
Throughout this essay I have argued that an important strand of Yoruba ethnogenesis was fundamentally grounded in a cultural mode of quantification and boundary construction in West Africa and the Americas.Footnote 28 For all of the interest in how Yoruba identity developed from “without,” beginning in the mid-nineteenth century with Christian missionaries and diasporic returnees, the “making” of the Yoruba was deeply embedded in the generative semantics of sortal particularization (Verran Reference Verran2001) and the transpositions of “house” and “home” through which idioms of genealogical descent emerged. In Yorubaland proper, I reversed received approaches to the building blocks of the social order by placing the town or kingdom as “prior to” and thus structurally immanent within the residential compound or home. Such a counterintuitive approach to the town in the home—at least from the conventions of Western social science—was inspired by Barber's exposition of oríkì orílẹ̀, that form of praise-poetry celebrating lineage identities through towns of origin rather than eponymous ancestors (1991: 135–53, et passim). Moreover, once grasped as kingdoms in microcosm, Yoruba homes and their lineage dynamics can be seen as instantiating a range of organizational modes, including those made manifest through the very performances of oríkì themselves, rather than as primary principles of a specific form and type—residential versus genealogical, agnatic versus cognatic, corporate versus symbolic, which the lineage debates could never pin down.Footnote 29 When recast as emergent sortal determinations rather than primary structural types, Yoruba descent and its residential dimensions become manifestations of wholes in parts.
If in Yorubaland we saw how orisha worship effects such transpositions through collective rites of renewal, the same operational logic was also discerned at the core of Cuban Santería and Brazilian Candomblé. In each of these contexts, cult “houses” manifested prior kingdoms of origin that repossessed towns during public festivals, as cabildos de nación and irmandades that were ritually remade into neo-African monarchies. Within the emerging fields of socio-religious production in colonial Cuba and Brazil, the Lucumí and Nagô ethnonyms came to dominate along two axes of ritual accommodation and resistance: one lateral, assimilating non-“Yoruba” ethnic nations within their expansive brotherhoods; the other vertical, filtering the iconography of European royalism through Yoruba grammars of kingship and sovereignty. As the Lucumí and Nagô nations expanded, largely through religious avenues after the 1870s and 1880s, their social capital increasingly accrued through ties of ritual rather than ethnic kinship.
The capitalization of ethnic entrepreneurship that followed in Cuba and Brazil, and the consolidation of a Yoruba ethnicity in Nigeria that it energized, belong to the longer durée of Yoruba ethnogenesis in trans-Atlantic perspective (Matory Reference Matory2005: 38–148). My goal has been to identify a critical cultural modality—a distinctive semantics of quantification—that rendered these developments both possible and intelligible, while also bringing novel historical interpretations to light.Footnote 30 In Cuba, the transition from cabildo de nación to casa de ocha was a continuous development of town into home—largely in response to state persecution—as ritual ramages and lines of consecration remade neo-African monarchies in genealogical terms. In Brazil, the emergence of Candomblé houses from black brotherhoods and sisterhoods resonated with the historic casa grande, transposing plantocratic fears of racial mixing into obsessive concerns with ritual purity. Thus when Matory (Reference Matory2005: 115) asks, “Why is it that Brazilian Nagô Candomblé and the Cuban Lucumí Regla de Ocha pursue ritual objectives of ‘purity’ and ‘cleansing’ that are virtually absent from the cognate Nigerian òrìṣà religions that are typically regarded as their origins?” we can point to the ritual reworking of color stratification within the New World cults. That the considerable literature on Nagô purification has overlooked this specific transposition of racial and ritual bloodlines attests to the insights afforded by Yoruba modes of demarcation and boundary formation in the plantation societies where neo-African identities initially emerged. If such motivating logics are difficult to perceive from without, they provide new perspectives on the diasporic trajectories of Yoruba ethnogenesis from within.
