Indigenato and Interpellation

If naming others was part and parcel of colonialism, interpellation by the colonial state, which sought to make legal categories equivalent to social places, was both juxtaposed with and displaced by vernacular forms of designating oneself and others. The Estatuto Político, Civil e Criminal dos Indígenas de Angola e Moçambique (1957) [Political, civil, and criminal native statute of Angola and Mozambique], passed in 1926, created legal categories of difference in these Portuguese colonies.Footnote ³ It set forth who was to be considered “civilized,” and therefore a cidadão (“citizen”), and who was to be considered “non-civilized,” and therefore an indígena (“native”). It thereby transformed the social, economic, and political places of people who previously inhabited life on the basis of social recognition. The statute’s division between “citizen” and “native” resembles the distinctions between “citizen” and “subject” outlined by Mamdani (Reference Mamdani1996). However, Portuguese “native law” was peculiar in that it allowed for someone who was born an indígena to become a citizen, to become assimilado, or “assimilated,”Footnote ⁴ through the display of the attributes that “civilized” people were expected to possess. It was never easy for someone to become a citizen in the Portuguese colonies, but the possibility shaped people’s desires. This mechanism of interpellation was in force until 1961, the year in which the first anticolonial uprisings broke out in Angola and all inhabitants of Portuguese colonies were granted citizenship and thus considered “Angolans.”

The beginning of decolonization in other African colonies, international criticism, and condemnation of labor conditions in the Portuguese Empire made it clear to Portugal that it was time to loosen its grip on the colonies. This loosening was accompanied by the dissemination of lusotropicalist ideology, propagated by the Portuguese state in the 1960s. According to lusotropicalism, Portuguese colonialism differed from that of other colonial powers because it was not utilitarian, but moral, and done for the benefit of Portugal’s subjects. Lusotropicalism presented miscegenation as evidence that the Portuguese were affable and prone to conviviality with “other races” (see Bender Reference Bender1978; Lusotopie 1997; Messiant Reference Messiant2006; and Castelo Reference Castelo, Anderson, Roque and Santos2019). The eventual end of the indigenato regime was certainly related to external circumstances. But it is also the case that the colonial attempt to homogenize subjects through their racialized interpellation as indígenas was never smoothly translated into vernacular forms of handling difference (Scott Reference Scott1998). I will focus here on the iteration of forms of designating oneself and others in Portuguese and Umbundu and relate such names to the broader context of colonial interpellation.

Although Umbundu was the most widely spoken language in the Central Highlands, some knowledge of Portuguese was available to most Umbundu speakers through the incorporation of loan words that went hand in hand with the caravan trade, which dominated the region until the military subjugation of local political formations in the early 1900s. This is to say that Umbundu and Portuguese should not be regarded as self-contained entities, but should be considered structurally different, even if with blurred borders, both due to linguistic difference and to their indexation of different social places: that of the colonizer and that of the colonized, along with the unequal power positions that characterized them (Mignolo and Schiwy Reference Mignolo, Schiwy, Maranhao and Streck2003). It is the creation of borders through translation that gives one the impression that discourse circulates between languages (Gal Reference Gal2015). Yet, Portuguese and Umbundu have been constituted as separate precisely through the translational process that had its necessary counterpart in the reduction of Umbundu to grammatical and dictionarized form by Catholic and Protestant missionaries (Rafael Reference Rafael1993; Gilmour Reference Gilmour2006). In this process, translation emerges as a site of subject constitution within the asymmetrical relations of power in colonialism to the extent that it “bring[s] into being hegemonic versions of the non-Western other” (Niranjana Reference Niranjana1992: 4).

There are few historical sources in Umbundu language (cf. Dulley Reference Dulley2021), since the majority of its speakers were and remained illiterate throughout the colonial period. Yet, vernacular forms of designating oneself and others surface now and then in the available sources,Footnote ⁵ which were the product of a translational effort involving Portuguese, English, and Umbundu speakers.Footnote ⁶ Vernacular forms might appear in Umbundu if a term is considered to be untranslatable into Portuguese, as frequently occurs with ocimbali, or in translation, as with ocindele ^{Footnote 7} (translated as branco, “white”) and ocimbundu (more or less derogatorily translated as preto or negro, “negro” or “black”). This paper focuses on the equivalences and non-equivalences established between these names in their translations and iterations. I argue that colonial translation, or its absence, is frequently motivated by the desire to equate a name in Umbundu that suggests disposition and prestige (or its lack) with a name in Portuguese that fixes the inevitability of race. Still, translational displacement can challenge such equivalences.

Table 1. Umbundu Terms in Translation

The attempted equivalence between ocindele and branco, on the one hand, and ocimbundu and preto or negro, on the other hand, begs the question of why the category ocimbali was perceived as untranslatable. Why could it not, for instance, have been translated as assimilado,Footnote ⁸ the Portuguese equivalent of the French évolué? This refusal of the possibility of translation suggests different conceptions on the part of Portuguese and Umbundu speakers of what it meant to be “civilized” and that “these divisions were not as easily (or permanently) drawn as one might imagine” (Stoler Reference Stoler1989: 155). In considering the reasons for the non-translation of ocimbali as assimilado, although both are terms that “neither color nor race could readily or reliably delimit or contain” (Stoler Reference Stoler2009: 8) and frequently referred to the same kind of people, one realizes that what these names made clear—that the racial and color component of the description is never straightforward—also applies to the other equivalences mentioned earlier. That is, although there is an attempt to equate colonial racial categories with Umbundu terms, the iteration of the materiality of the signifier—repetition that necessarily contains difference and similarity—reveals displacement in translation (Derrida Reference Derrida1981).

The Native Statute [Estatuto do Indígena] of 1957 Footnote ⁹ set forth the “moral duty” of indígenas to work while also discriminating against the subjects of its interpellation: “For the effects of the present Statute, the individuals or descendants of black race who, in their education and customs, do not distinguish themselves from what is common in that race shall be considered natives [indígenas]; and the individuals of any race who do not fulfill these conditions shall be considered non-natives [não indígenas].” If the criterion to measure the degree of “civilization” for someone to be considered assimilado and become a citizen was the person’s capacity to assimilate Portuguese culture, the notion of “education and customs” was inevitably attached to race: “race” and “stage of development” were, as a rule, equivalent. Even though the status of indígena did not totally preclude a person thus interpellated from becoming assimilado, and therefore a “citizen,” the requirements for attaining such a privilege became increasingly restrictive after 1926. Messiant (Reference Messiant2006: 75–76) argues that it was necessary for Africans and Europeans to have different statuses so that the former could be made available as labor to the latter; the small layer of assimilados served as intermediaries between the vast indígena population and Portuguese colonizers. The whole indigenato system was based on the racialization of indígenas, race being a form of hierarchization in which a population is socially and physically made inferior (Chatterjee Reference Chatterjee1993; Andersen Reference Andersen2014).

Becoming part of the privileged class of assimilados required relinquishing “uses and customs” considered to be “non-civilized” and embracing so-called “civilized” ones. That is, becoming assimilados meant living and dressing according to European standards; speaking and writing Portuguese fluently in a context in which most Portuguese colonists were illiterate; not living on subsistence agriculture and earning a minimum wage; being monogamous; and being a Christian. The interpellations addressed to colonial subjects supposed the connection between a name and the incorporation of a series of bodily dispositions that allowed symbolic power to draw distinctions among the non-white population. Racialization became part of the colonial habitus (Bourdieu Reference Bourdieu1985). Assimilado inhabitants of the colonies remained second-rank citizens in relation to Portuguese citizens, but they were not compelled to undertake forced labor and could have low-paid jobs in the colonial administration and private sector. They had to pay higher taxes than indígenas but earned better wages and enjoyed freedom of movement. However, assimilados had to comply with the criteria based on which they had gained citizenship. They could lose their citizenship if they “relapsed” into the customs attributed to the so-called gentios (“gentiles”), such as polygamy, following non-Christian cults, not sleeping on a bed, or not speaking Portuguese.

The indigenato regime guaranteed that whiteness remained synonymous with the privilege of citizenship, paving the way for an influx of poor white settlers throughout the twentieth century (Castelo Reference Castelo2007). And, while the law established the criteria for the admission of non-whites into this privileged circle, admittance depended on good relations with those who were in the position to assign privilege. Assimilados, representatives of the sole exception to the equivalence between race and alleged degree of civilization established in the indigenato regime, acquired their status either because their fathers were citizens or through education and good relations with local officials. Assimilation served the purpose of defining, by contrast, those who could be recruited as labor: men between the ages fifteen and sixty who were neither in the army nor employed as rural policemen (cipaios), by a European family, or as permanent workers for a private company (Cahen Reference Cahen2012). Women were not subject to forced labor but, along with children, were expected to feed the workers and perform unpaid “public works” such as building bridges and roads. Assimilation was only possible for an indígena woman via her white or assimilado father or husband. A clear hierarchy assigned cultural and racial superiority to whites, who were followed by mestiços (“mestizos”), assimilados, and indígenas. Among the latter, those considered to be “pagan” ranked lowest.

Colonial law contained within itself the potential for its own abuse (Messiant Reference Messiant2006: 86), for it not only restricted indígenas’ social mobility and choices, but also assigned decision-making power regarding assimilation to the very people that were least likely to grant it. More assimilados meant local officials would collect less taxes and recruit less people as labor. However, this dependence on local officials was also a point of rupture in the colonial system. Since corruption was widespread, it was possible to gain an official’s favor through status or advantage. This ambiguity in the actualization of colonial law led to the privileging of those Christians who were closely related to missionary organizations and thus in a better position to support their claims to assimilation with the colonial administration. This fostered continuities and ruptures in the patterns of hierarchization and privilege prevalent in the Central Highlands prior to the establishment of the colonial administration in the twentieth century. At a national level, the constitution of a class of educated Christians led to the formation of the competing regional elites that fought for the power of the state after Angolan independence in 1975 (Messiant Reference Messiant2006; Péclard Reference Péclard2015). In Umbundu expression, were one to describe the social positionalities of these elite subjects, they would most likely be designated as ocindele or ocimbali, although few of them would be classed as “white” in Portuguese.

Colonialism as synonymous with “civilization” is an endeavor that promises to deliver the colonized from what is seen as their inferiority by changing the dispositions through which they are identifiable as inferior. Yet, the colonized are simultaneously precluded from achieving the standards set as desirable since they are not provided with the necessary means to reproduce that which the name “civilization” supposedly entails. And the very exceptions that confirm the rule—the less than 1 percent of assimilados around at the time of Angolan independence—function as the confirmation of a possibility that remains unfulfilled due to the supposed incapacity of the colonized. Nowhere is the extralegal violence of law more evident than in the act that inaugurates the interpellation that will try to conceal its arbitrary origin through repetition (Derrida Reference Derrida, Cornell, Rosenfeld and Gray1992). As “the one hailed always recognizes that it is really him who is being hailed” (Althusser Reference Althusser2001: 118), the state does not need to address its subjects verbally and individually—the historicity of names is contained in their iteration. Rather, it is the iteration of the equivalence between a name and the position it designates that produces subjects. Thus, interpellation is not a description, for it is neither true nor false (Butler Reference Butler1997); it is a performative statement that attributes a place to a subject by quoting an existing convention (Austin Reference Austin1962) that constitutes subjects before they come to exist—the son of an indígena is, by law, an indígena. Interpellation sets the terms by which a subject might have a place in society and largely determines that place, for the iteration of someone’s position depends on the (always imperfect) equivalence between a name and a place.

Colonial legislation, although attempting to stabilize a convention that determines social places, is haunted by the disjunctions between the forms of hierarchy that precede it and colonial attempts to fixate a different order. This disjunction appears in language, in which the equivalence between signifier and signified is challenged by the counterviolence of displacement (Derrida Reference Derrida1980). In a structuralist system—such as Lévi-Strauss’ “classification” (Reference Lévi-Strauss1966) and Bourdieu’s (Reference Bourdieu1991) “rites of institution”—a name is meaningful insofar as it differs from other names. This clear-cut opposition depends on a somewhat stable, though arbitrary, coupling of signifier and signified that creates meaning through contrast. For this to be the case, “white” would have to be equivalent to ocindele, and the opposition between “white” and “black” should be homologous to the opposition between ocindele and ocimbundu. However, the rupturedness of the connection between signifier and signified appears in translation. As argued below, ocimbundu is not the same as “black”; preto does not equate with “black”; and ocimbundu might even differ from ocimbundu. This is the case because iteration depends on the historicity of names, and so does translation. Names, as they are cited, also imply the social positionalities of the previous political formation indexed by Umbundu designations,Footnote ¹⁰ for citation is both imitative and novel (Gal Reference Gal2015: 227).

If every attempt at stabilization contains contradiction, and therefore poses the possibility of subversion, the degree to which nomination is successful depends on the power of the convention it quotes. The performativity of naming is necessary because the arbitrariness of classification is filled with instability. Naming makes evident the need for constant re-enaction, and therefore also the possibility of iteration’s failure to reiterate, that is, its failure to iterate the same to the prejudice of difference. And if naming is grounded in convention, and the stability of the linguistic convention that guarantees reproduction depends on social stability (Siegel Reference Siegel2006), a destabilization of forms of interpellation is to be expected whenever social instability affects convention. If “[t]ranslation … leads to the emergence of hierarchy,” this entails the “possibility that haunts every communicative act: that at some point translation may fail and the social order then may crumble” (Rafael Reference Rafael1993: 211). Competing forms of designating oneself and others in Umbundu can be viewed as traces that rip apart the seemingly placid surface of colonial law. It is my contention that the possibility of assimilation, a brisure (Derrida Reference Derrida1976) opened in the structural racialization promoted by colonial law, found fruitful soil in the vernacular expression of privilege and status. If Umbundu names index the social positionalities of caravan-trade Central Highlands, they also contain a potential to disrupt colonial order and hierarchy.

Vernacular Names

Although the indigenato regime determined the form and structure of interpellation, there were other, unofficial ways of designating oneself and others. Vernacular names bore a relation of displacement to official forms of interpellation. Yet, Portuguese and Umbundu did not fully correspond to two different systems of expression; rather, the linguistically diverse but interdependent names uttered in both languages were distributed along the lines of an overarching hierarchical relation. The iteration of names in Umbundu reveals both the social places available in Angolan colonial society and how they were negotiated, disputed, and transformed. Whereas names such as indígena encompassed a wide range of social positions and dispositions, which they sought to reduce and fixate (Scott Reference Scott1998), names in Umbundu were not marked by the constraints of legal consistency. Yet, despite the hierarchization in which Portuguese ranks higher and has more fixating power than Umbundu, the attribution of “resistance” to a vernacular “hidden transcript” (Scott Reference Scott1990) can be challenged by an analysis that takes translation into account. To consider the work that translational displacement can do, let us begin with one iteration of the juxtaposition of Portuguese and Umbundu names.

In an ethnographic study conducted in the Central Highlands, Edwards (Reference Edwards1962: 136–37) tells us the story of Justino, “a man of fifty or more, who lived in Portuguese style making money from coffee and other crops” in the 1950s and “claimed to be assimilado” although he was not legally recognized as such (a rather common phenomenon). He used to mediate conflicts between villagers and employed workers on his farm, but as someone whose position in colonial society was not irrevocably that of a citizen, his status was subject to challenge. A labor recruiter was said to have called him um preto qualquer, which was translated into Umbundu as ocimbundu cango, “a mere black.” Ocimbundu, the singular of Ovimbundu, the ethnonym attributed to the inhabitants of the Central Highlands during the colonial process, was thus juxtaposed with the social position of the non-assimilated indígena, “black” by default.Footnote ¹¹ In response to the recruiter’s remarks, Justino made a public speech in which he affirmed that he was not an ocimbundu. In the performance of this ritualized act, he challenged the accusation leveled against him by enacting his capacity to summon people to his house, including the labor recruiter. His main statement that he was not an ocimbundu, Ame sitjimbunduko, might have been translated in different ways: “I am not black,” “I am not indígena,” “I am assimilated,” “I am a Portuguese citizen,” or even “I am white.” Although the term’s translation into Portuguese is disputable, the derogatory character of the term ocimbundu was taken for granted in its citation. Not many such situations are described in colonial records, but many such occasions composed the thread of citations through which people were interpellated.

In colonial sources, ocindele was equated with branco, and ocimbundu was juxtaposed with preto or negro (respectively, in a more or less derogatory way). Yet, ocimbali was not translated as assimilado, even though Portuguese and Umbundu speakers understood the term to indicate a position and disposition involving closeness to Europeans. In order to understand why such a translation could not be established, one needs to focus on two interrelated aspects: the logic that structured the designation of oneself and others in Umbundu, which I argue was related not to origin, race, or descent, but to disposition, status, and prestige; and the treatment given to the vernacular term ocimbali by colonial agents, who ethnicized it. Let us follow the juxtaposition of two Umbundu/Portuguese terms: ocindele/branco and ocimbundu/preto. We shall then compare this to the non-juxtaposition of ocimbali/assimilado and inquire into the ethnicization of ocimbali as Mbali. These two cases—that of a racialized translation and that of an ethnicized non-translation—point to possible displacements in the relationship between interpellation and social positioning in Angolan colonial society.

Translation, Race, Irony

The unpublished Umbundu grammar put together in the 1930s by Merlin Ennis, a Congregationalist missionary in the Central Highlands, states, “The more common method for deriving one noun from another is through employing either the augmentative prefix oci or the diminutive prefix oka. Take for instance the word ondele, infant (still pink) and ocindele, whiteman [sic] (they say ‘red man’). A mulatto is a okacindele, a diluted whiteman. In scorn they may apply okacindele to an African who apes the ways of the whiteman” (ABCFM archives, Houghton Library, box ABC 76). The association between ocindele and white skin color, a common slip in colonial translations, ignores the fact that ocindele, as employed in Umbundu, made no such reference. While Ennis speaks of the “mulatto” as a “diluted whiteman,” okacindele hints at the fact that such people sat awkwardly in the hierarchy that had “white” as the highest position and “black” as the lowest. The metaphor of the ape that intrudes into Ennis’s narrative unwittingly reveals that “the menace of mimicry is its double vision which in disclosing the ambivalence of colonial discourse also disrupts its authority” (Bhabha Reference Bhabha1994: 126). Translation makes apparent how the incorporation by the colonized of traits conventionally attributed to the colonizer unsettles the latter’s positioning in the colonial situation: If the colonized can act like the colonizer, the former’s mimicry of “civilized” behavior challenges the juxtaposition of the privilege of whiteness with “civilization.” Thus, translation implicitly challenges and unsettles the constitution of the colonizer as privileged subject. Moreover, colonial dictionaries contradict their own racialized rendering of ocindele.

In his Umbundu-Portuguese dictionary, Spiritan bishop Alves separates the root ndele into three different entries. The first, “ndele (o; olo),” is translated as “evil spirit; demon” and gives as synonyms otjilũlu, ehamba, and osande, the names of worshipped ancestors (Alves Reference Alves1951: 820). As examples of the word’s use in context, he presents the phrase okutuñuha olondele, glossed as “having diabolic convulsions,” in a rather common juxtaposition of local rituals with the diabolic (see Meyer Reference Meyer1999; Behrend Reference Behrend2011). Alves’s attempt to keep separate what he presents as the three meanings of the root ndele is reflected in his attribution of a derogatory character to the first and a complimentary meaning to the second entry: “ndele (otji; ovi, i),” both “that which shines” and “white, person of white race [branco, pessoa de raça branca].” The second entry is followed by an example: Omunu otava-tava etji tjipopya vakwavo, okatjita ovindele; u pwãyi opatãla-patãla, okatjita ovikwamanga, translated as “The person who follows orders gives birth to whites [brancos]; but the person who contradicts other people gives birth to crows (pitch black people [pretos retintos])” (Reference Alves1951: 823). It remains unclear whether the sentence in Umbundu is Alves’s creation or belonged to the common pool of vernacular proverbs. However, its racialization of what is perceived to be the inferiority of others is certainly the author’s doing, since ovikwamanga, in Umbundu, refers merely to crows, while pretos retintos, a derogatory Portuguese expression used to mean that a person’s blackness is the result of being painted black many times, was added by the author in the guise of an explanation.

The third entry calls into question Ennis’s attempt, posed elsewhere (ABCFM archives, Houghton Library, box ABC 76), to separate whites from spirits, as “ndele (u; ova)” is said to be the “custom, manner, way of being of the white,” but is obviously also that of the spirits if “u” stands for “ways, habits, speech, customs and life.” And if the ways of the white are related to those of the spirits, they might well share their spectral quality. The conflation of ndele with race is ruptured in the very act of trying to establish it, for if undele is the “way of the white,” it is possible to become white through mimicry. The example given by Alves seems to confirm it: “Walikapa undele = Took to himself the ways of the white [Tomou para ele a maneira dos brancos]” (Reference Alves1951: 823). It is thus clear that ocindele and its chain of substitutions designated not only people the colonizer never considered to be “white” but also pointed to the realm of ancestors.

In his collection of fables in Umbundu, Spiritan missionary José Francisco Valente affirms that the Umbundu terms used to refer to white people, ochimbali ^{Footnote 12} or ochindele, might be used interchangeably. In Valente’s words, “For the Bundo [Portuguese colonial term for Ovimbundu], the term ‘Chindele,’ White [Branco, the capitals are his], is not limited to color. ‘White’ is everyone who has the habits of a civilized person, even if this person is pitch black [preto retinto]. White is, therefore, the detribalized [destribalizado]” (Reference Valente1973: 74). In attempting to understand where the term ocindele comes from, Valente explains the association between “white” and the spirit world as one in which the power of spirits is attributed to the white man: “The white man carries a ‘charm’ [feitiço, also translatable as “fetish”Footnote ¹³], he has ondele, and this is perhaps why our spirits are inclined towards him” (Reference Valente1974: 111). The white man’s power is, thus, clearly associated with what missionaries contributed to describing as the universe of sorcery: the possibility of interfering with the expected course of events by drawing on forces not available to everyone else (see Geschiere Reference Geschiere1997; Comaroff Reference Comaroff and Comaroff1999). It is also related to power as coercive force, as in the following excerpts: Kwende ñgo, ame ndakutumila, ame ndichindele! (Valente Reference Valente1973: 71), translatable as “Just go! I order you! I am white,” and Twatundamo, tuvindele, which Valente renders as “We obeyed you and became people” (Reference Valente1973: 297) but is literally translatable as “In obeying you we became white.”

In vernacular narratives, ocindele is associated with wealth, prestige, and power, as in the phrase walinga ochindele, which Valente translates as fazendo dela uma senhora (“making a lady of her”) but which could also be translated as “making her white.” However, I do not wish to suggest that branco was equated with ocindele because “whites” were thought to be powerful, as colonial ethnographic records seem to suggest. There seems to have been more to it. Goodsell mentions that in Huambo in the late 1880s, “a young man, after hearing about God as Creator, asked [Sanders, a Congregationalist missionary]: ‘Why did God allow the whiteman meat and chicken to eat and cloth to wear, while black men carry loads, eat corn mush and wear rags?’” (ABCFM archives, Houghton Library, microfilm, reel 154). This question, which is unfortunately only rendered in English, directs our attention to the fact that the equivalence between whiteness and power was an imperfect one, for while white people bore a similarity to ancestors, they were also thought of as people whose privileges were not inherent. Bishop Alves’s instruction to Catholic catechists makes this clear. While he exhorts catechists to respect “all whites, even if they practice a false religion or have no religion at all” (Alves Reference Alves1954), he also attributes to these same whites—and we should not forget that being a literate Christian was considered a sine qua non condition for assimilation—the inherent possession of civilization, translated into Umbundu as elunguko lyovindele (ibid.), roughly “the liveliness of the white,” lunguko being rendered as esperteza (“cleverness”) in Alves’s dictionary (Reference Alves1951: 593).

Displacement is also revealed in irony. Ennis (ABCFM archives, Houghton Library, box ABC 76) and Valente (e.g., Reference Valente1973: 269) translate ondele, the word from which ocindele derives, as “infant.” While Ennis points to the fact that the skin of newborns is red and this contributed to their association with the reddish skin of the “white,” it also seems to be the case that both infants and ancestors are insatiable. This is how Goodsell, drawing on Ennis’s description of ancestor worship, depicts ancestors: “Interestingly enough, ancestors who have died, called ilulu (spirits out of the body), are reckoned as a part of the village community. They may get angry and are thought to be creatures of other emotions which must be respected by the living. One must be careful to keep in their good favor. This is done by sharing with them any good fortune one has and by talking with them (prayer)” (ABCFM archives, Houghton Library, microfilm, reel 154). Like the white, ancestors are to be appeased. And like ancestors, “whites” used to behave in a demanding paternalistic way. This is how Goodsell describes the missionaries’ arrival at Olimbinda, a future mission site: “The place was with due ceremony declared ours. I then arranged to have a hut built in my absence, and a lot around it fenced in, so as to have the natives accustomed to the idea of a white man living as a friend and helper among them, and also to have them take a little responsibility upon themselves” (ibid.). Those who responded to the demands of an ocindele, ancestor or “white,” expected to be rewarded with gifts. However, this was not always the case. It depended on the ocindele’s goodwill.

Ocimbali, the Untranslatable In-Between

It was the supposed universality of the colonial endeavor, which “misrecogniz[es] the historical conditions by which comparison becomes materially possible for an abstract omniscience” (Morris Reference Morris2007b: 172), which created the conditions of possibility for the hierarchization of colonial society to ignore other ways in which colonial relations were translated. Misrecognition of the irony contained in the translation of branco as ocindele points to this fact. So does the assumption that ocimbali intransigently refuses translation because it evades the (impossible) colonial project of equivalence. Untranslatability points to the impossibility of the subjects designated by the name ocimbali “acced[ing] to the universal” (ibid.: 191). In the excerpts below, the term ocimbali and its chain of substitutions are iterated as if detached from discourse, as something that cannot be fully rendered in the language of a nostalgic colonial endeavor that desperately seeks to enact the imperialism it projects into its “glorious past.”

As far as the fractures of colonialism are concerned, Barbeitos (Reference Barbeitos2008) relates racialization in Portuguese colonies to Portugal’s peripheral place in world imperialism, whereas Messiant (Reference Messiant2006) associates racism with the settlement of Angola by dispossessed Portuguese citizens for whom racial privilege was the only possible means of distinction. In trying to equate the terms in Umbundu with a fixed social position, colonial attempts to define them struggle with the eerie evocation of the times that preceded military subjugation, when Europeans were encompassed by African social formations (Cândido Reference Cândido2013). One might even argue that it was the “extraversion” of the Central Highland societies (Péclard Reference Péclard2015), their very openness to strangers—who were named on the basis of disposition, wealth, and social status—that eventually led these societies to fall prey to a legal framework based on fixed social status. Yet, the times when Europeans could only inhabit Angola in vernacular terms haunted the colonial imagination, and traces of this phantasmatic social order disrupt colonial discourse in the form of untranslatable names such as ocimbali.

In its regional variations, ocimbali was transliterated as mbali, kimbari, mambari, and quimbar (the most common Portuguese variant). It was also ethnicized as Mbali, but never translated. However, from a social perspective, people thus designated were in-between the two terms available to interpellate Africans: indígenas and assimilados. The Mbali on the southwest coast of Angola, for example, were former slaves or descendants of slaves whose process of distinction from common indígenas started in the nineteenth century and culminated in the formation of a “stable proletariat” to which the designation of assimilado was applied (Clarence-Smith Reference Clarence-Smith1979: 8, 42). In the wake of the abolition of slavery, they were considered by colonial agents to be distinct from “what was common in that race” due to their incorporation of some of the ways of their masters, which led to their qualification as “detribalized.”

Spiritan missionary and ethnographer Estermann suggests that the ethnic term Mbali was used in southwest Angola as a self-designation by former slaves who understood themselves to be different from ordinary Africans (Reference Estermann1956: 49). This seems to be confirmed by the entry for mbali in Alves’s dictionary: “a slave marked on the face and sold to the white, whose customs he partly acquired; civilized (now a common meaning)” (Reference Alves1951: 653). Neto describes the ovimbali in the Central Highlands as “broadly those who adopted some of the ways of the European, or had close trading relationships with them,” and affirms that, “Long-distance trade and association with foreign goods and European habits gave some Umbundu-speaking groups the name of Vimbali (sing. Ocimbali) among many eastern neighbours” (Reference Neto2012: 39–40). The Catholic missionary Lecomte’s 1887 report entitled Chez les Ganguelas [Among the Ganguela] seems to confirm this. Lecomte affirms that the population of Ngalangi [Galangué], one of the Umbundu-speaking polities in the Central Highlands in the late nineteenth century, is considerable and

Thus, one learns that the term ocimbali and its variations, sometimes equated with ocindele and the social position associated with the prestige and status of whiteness, was not confined to the Umbundu-speaking area. It was disseminated throughout the regions connected to the Central Highlands through trade, raiding, and migration.

In the period of the caravan trade, the term olofumbelo—plural of ofumbelo, the Umbundu version of pombeiro, or caravan trader—was employed to designate wealthy merchants of every skin color. In Magyar’s (Reference Magyar1859) travel accounts in the mid-nineteenth century, kimbáló, his transliteration of the term ocimbali into Hungarian (his native language) (see Fodor Reference Fodor1983: 116), is equated with trader.Footnote ¹⁴ This points to a certain overlapping of ocimbali and ofumbelo, a hypothesis that Alves’s rendering of mbali as “black merchant” (Reference Alves1951: 654) seems to confirm. After the Central Highlands were militarily subjugated and incorporated into the colony of Angola in the early twentieth century, the olosoma [plural of osoma, usually translated as “king”] and olofumbelo, who formerly concentrated political and economic power through the monopoly of trade routes, became indígenas just like their former subjects, who became the subjects of colonial rule through the mediation of the olosoma and olofumbelo. Nuanced forms of interpellation resulted from the ambiguous overlapping of the social configuration of the caravan trade and the legal colonial differentiation between citizens and indígenas. Before indigenato, an ocimbali might have been someone who possessed Western goods and enjoyed prestige but was polygamous and neither spoke Portuguese fluently nor had any formal education. In the colonial context, an ocimbali could be someone who lacked the necessary instruction to pass the exams required for assimilation but shared the material conditions of Europeans and could speak some Portuguese—like Justino, the farmer who said he was not an ocimbundu.

Although ofumbelo gradually lost its referent with the weakening of the caravan trade, the term ocimbali persisted. Clarence-Smith (Reference Clarence-Smith1979: 47) differentiates between the Mbali of the Caconda highlands, who participated in the caravan trade and became small farmers after the rubber trade declined in the 1910s, and the coastal Kimbari. To a certain extent, ocimbali and its variations overlapped with assimilado due to the implied proximity to the colonizer and his goods, and thus to political power and prestige. Yet, from the perspective of colonial agents, even if the Mbali were viewed as “acculturated,” they were still referred to as indígenas, and most were interpellated as such within the indigenato regime. In his article on “the Mbali of Southern Angola,” Estermann affirms that it was difficult to know how many Mbali inhabited a given place, for the statistics did not distinguish them from “other indígenas,” although they had “assimilated a great number of cultural elements from their former masters” (Reference Estermann1983: 135). The ambiguous place of the people thus designated might explain why the term was ethnicized in some of its iterations, such as the 1960 census, but not in others.

In Southern Angola, the connection between slavery and those designated as Mbali led to an ethnicization of the term that qualified these people as more “civilized” than their neighbors. In the Central Highlands, however, where farmers and/or traders who did not live close to white settlement areas assimilated fewer elements from the universe of the colonizer, the ethnicized signifier was not Mbali, but Ovimbundu, or Vimbundu, the plural of ocimbundu, which was made equivalent to “black” in colonial translations. The term ocimbali, the most common iteration of Mbali in the Central Highlands, remained in use but was not equated with a fixed position in colonial society. Its chain of substitutions interpellated subjects in different ways in the various regions of Angola before the establishment of colonial rule. Edwards, for instance, affirms, “The Ngangela call the Ovimbundu ‘Ovimbali,’ a name used in Angola for Africans who live with or imitate the white. I have heard an Otjimbundu [variation of ocimbundu in the ethnicized singular form] use it for the Kimbundu of Northern Angola” (Reference Edwards1962: 8). As the inhabitants of Umbundu- and Kimbundu-speaking areas were seen by their neighbors as traders and raiders, with those speaking Kimbundu being seen as closer to the Portuguese than those speaking Umbundu, ocimbali and its variations seem to have been used to designate those who, from the perspective of the speaker, shared the social position of Europeans as well as their goods.

This hypothesis leads Messiant (Reference Messiant2006: 45–46) to advance the structuralist argument that ocimbali and its variations imply closeness to and imitation of the white. Thus, such terms are applied by one ethnic group to another that it sees as more civilized than itself: the Ngangela call the Ovimbundu ovimbali; the Ovimbundu call the Mbundu ovimbali; and the Mbundu, who are closer to the colonizers, call no other ethnic group kimbari but reserve this term for the ones among themselves who are closest to Europeans, since the Mbundu, as an ethnic group, are considered a sort of prototypical occupier of the intermediary position signaled by this term. Although Messiant’s understanding of ocimbali and its chain of substitutions as a relative ethnic category illuminates the historical relationship between regions in Angola as well as the rural-urban divide that is an important marker of social distinction (Cahen and Messiant Reference Cahen and Messiant1989), her systemic approach does not allow one to make sense of the competing ways in which prestige, wealth, and status were iterated in relation to the signifier ocimbali. If it is undeniable that colonialism structurally reconfigured social relations in Angola, it is also the case that translation and iteration, as displacement, unsettle the fixity of classification. And ethnicity, a historical marker of distinction, is also displaced in translation.

For instance, the affirmation that “the mestiços who formed the upper and most numerous layer of the non-white group of the pombeiros [were] simply, for the Whites as well as for the Blacks, mestiços” (Messiant Reference Messiant2006: 46) is hard to sustain if one considers that in Umbundu there was no perfect equivalent for mestiço, which in this specific case could be translated in many different ways: ofumbelo, in view of their occupation; ocimbali, to signal greater closeness to Europeans on their part than on the part of the speaker; ocindele, implying the position of status and prestige associated with being white or having a white father; or, ironically, okacindele (little white), to hint at the fact that their whiteness might be incomplete. Racial hierarchy remains operative, yet names are not equivalent to fixed social positions.

According to Estermann, “the Mbali themselves” proudly affirmed in Portuguese, the language associated with civilization, that Mbali means a “civilized black,” as opposed to “non-civilized blacks” or gentios (Reference Estermann1983: 135). Valente, who equates ochimbali with preto evoluído (“evolved black”), rapaz evoluído (“evolved young man”), and civilizado (“civilized”) (Reference Valente1973: 112, 193–94, 46), seems to agree with “the Mbali.” Valente (ibid.: 46) translates Ame ndichimbali, sisentiyuko as Eu sou civilizado, não sou supersticioso como um gentio qualquer, which is translatable into English as “I am civilized, I am not superstitious like a mere gentile,” although the Umbundu statement would be closer to “I am an ocimbali, not a gentile,” sentiyu being the transliteration of the Portuguese word gentio into Umbundu. In one of Valente’s iterations of the term ocimbali, where he recalls how one of the whites in Caluquembe “offer[ed] his mulatto daughter [filha mulata] to a chimbali as a concubine” (ibid.: 112), he in fact censors the “kaffir-like life” (vida cafrealizada) of whites. Although Estermann mentions the existence of “affective and sexual relations between white and mbali” (Reference Estermann1983: 145), it is implicit that he refers to relations between white men and Mbali women, something that was not considered abnormal. What shocked Valente was that the Caluquembe white man’s offer called into question the hierarchical intersections of gender and race in the relation between a mulata, the daughter of a white man and a black woman, and a black man, even one considered to be more “civilized” than other black men.

A similar slip in translation is to be found in Valente’s work, for example, when the word civilizado (“civilized”) is rendered in Umbundu alternately as ochimbali or ochindele. Ennis affirms that ocimbali can also refer to the “white”: “The word ocimbali would appear to be derived from the verb viala, to rule in the following way, the whitemen [sic] are called olombiali, rulers. As a derogatory the augmentative prefix replaces the o prefix of ombiali and the i in mbia is omitted for euphony sake” (ABCFM archives, Houghton Library, box ABC 76). In what he presents as the word’s etymology, ocimbali means “ruler,” but in an ironic way, as is usually the case when the augmentative is employed as a prefix to a noun designating a certain kind of person in Umbundu. Bell (Reference Bell1922: 145–47) translates ocimbali as “native servant,” and Edwards (Reference Edwards1962: 7) renders it as “a domesticated black.”

For Estermann, it is as if the Mbali’s self-definition as “civilized” hides a fundamental element: their origin. He affirms that the language of the Mbali was a mixture of Kimbundu and “Bantuized Portuguese,” although most of them could speak “the language of the white” (Reference Estermann1983: 135). Childs, however, argues that the language spoken by the Mbali was Umbundu, for “the term ‘Mbali’ (vaMbali, Mambar, Ovimbali) is an Umbundu term used for nearly one hundred years to designate those who imitate the Europeans or who live at or near the European towns which are collectively designated by the cognate term—Lupali” (Reference Childs1949: 19). Once again, the overlapping between ocimbali and ocindele is hinted at, as the first “l” in lupali is a locative that indicates “with the mbali” or “close to the mbali,” in this case, the “white.” What is striking is that the main issue to which colonial authors paid attention was the origin of the language spoken by “the mbali,” not its structure. No structural description of it has been registered apart from anecdotal remarks such as those by Ennis above, possibly because it would not be easy to attribute unity to the ways of speaking of people who shared a designation but inhabited such different regions and social positions. Cardoso’s (Reference Cardoso1966) attempt to determine the origins of the language spoken by the Olumbali in Moçâmedes denies the hybridity that characterizes it, although his own account seems to challenge this in presenting almost every word in its variations according to region and speakers.

Transforming those called Mbali into an ethnicity was not an easy task. Yet, colonial authors insisted on assigning an unequivocal primordial origin to this hybrid language, which could only make its phantasmatic character apparent (Morris Reference Morris2010; Ivy Reference Ivy1995). As a result, as in Ennis’s narrative below, transformations are taken for corruptions:

Despite Ennis’s disdain for Umbali, his account seems to confirm the association between ovimbali and prestige, as those he calls “genuine Ovimbundu” would also switch to this more prestigious enregistrement, replacing the ls in Umbundu with Portuguese rs in order to transfer to the former the status associated with the latter, in a clear association between a language variant, its name, and a social positionality (Gal Reference Gal2018). However, there were instances in which the equivalence between ocimbali and prestige was called into question, and the translation of ocimbali as “ruler,” “civilized,” “merchant,” or “servant” points to the ambiguity and fracture that characterized this place in society. It is my contention that the vernacular ocimbali is not fully translatable because it opposes colonial racialization while simultaneously mocking the dream of assimilation. Maybe the reason why ocimbali could never have been smoothly translated as assimilado is because irony, in its challenge to fixation, has no place in (colonial) law. The indigenato regime can thus be understood as a device aimed at halting the potential for unsettling the colonizer’s subjective constitution that is contained in the mimicry of his embodied dispositions by the colonized.

Final Remarks: On Iterative Translational Subject Constitution

Very few of the inhabitants of the Central Highlands of Angola were aware of the letter of the law by which they were interpellated. However, their possible places in colonial society were related to the iteration of names such as ocimbundu, ocindele, ocimbali, branco, negro, and assimilado. Umbundu names and Portuguese legal categories were not equivalents, but translations. Colonial translation, in its misrecognition of the impossibility of perfect juxtaposition between signifier and signified, forces the equivalence of branco and ocindele as well as that of negro and ocimbundu while precluding the translation of ocimbali as assimilado. The equivalences accepted are the ones that fit colonial hierarchy and thus become plausible. That is why the various authors discussed above, although somewhat differently positioned in colonial society, translate in similar ways. Yet, translation can never escape the work of différance, that is, the necessary disjunction and displacement between signifier and signified that leads to the possibility, for instance, that those indígenas classed as “black” (negro) in Portuguese might challenge colonial classification and speak of themselves as “white” (branco) in Umbundu—ocindele or ocimbali. Thus, the effects of displacement are both linguistic and material: The possibility to subvert racialization was, in this case, contained in the very form of the law that violently inaugurated the pattern of exclusion to be iterated (Derrida Reference Derrida, Cornell, Rosenfeld and Gray1992); for colonial law ignored the local social formations that preceded it as well as the vernaculars in which they were expressed. Translation threatens hierarchy in that it reveals the alien character of authority (Siegel Reference Siegel1997).

An overarching hierarchy structures the Umbundu-Portuguese chain of substitutions. The translations of ocindele as branco and ocimbundu as negro present these terms as timeless equivalents. Fabian’s (Reference Fabian1983) “denial of coevalness” is thus echoed in colonial records in a ghostly way, for these translations erase the ambiguous relations between Portuguese and Umbundu speakers that led to the production of these equivalences and their “implicit meanings” (Asad Reference Asad, Clifford and Marcus1986). Yet, traces of the relations on the basis of which ocindele was equated with “white” and ocimbundu was translated as “black” can be found in the iteration of the fragments in which this structural equivalence is posed. The same is true for the non-translation of ocimbali and its situational correspondence with the privilege of whiteness, the history of slavery, and assimilation. The mosaic of conjunctions and disjunctions that arises from the iteration of colonial interpellations and their translations reveals fractures that are recognized in ironic vernacular expression as the possibility or impossibility of certain subjects inhabiting certain places (Likaka Reference Likaka2009). Umbundu forms of designation bear a relation of displacement to colonial ones. Transformation is, therefore, to be sought in the shifting juxtapositions of designation and social positioning. To this extent, some of the transformations that happened in post-independence Angola, when both indígenas and assimilados became “Angolans,” were already contained as a possibility in Umbundu forms of interpellation.

On one hand, colonial law largely fixated the assignment of social positions; on the other hand, translational displacement was potentialized by the possibility of assimilation foreseen in the indigenato regime, even if it was rarely practiced. For instance, one could be officially classed as indígena and still call oneself not ocimbundu, but assimilado, ocimbali, or ocindele. Designations index a subject’s social position. But since the relation between names and social positioning is not fixed, different—and sometimes contradictory—names can designate the same individual in different contexts. My analysis has argued that imperfect equivalence between Portuguese and Umbundu designations enabled subjects to occupy the different positions associated with names applicable to them in both languages. Moreover, since names and positions are also imperfectly equivalent, subjects could occupy different positions while bearing the same name and change their positioning by using a different name. For instance, ocindele, a term used to name the position of privilege and high status of merchants during the caravan trade, could designate both the colonizer and those officially or socially perceived as assimilated during the indigenato regime. The relationship between names and social positioning is transformed as names are iterated.

It is in the space created by such disjunctures that agency can take place. For instance, were one to describe in Umbundu the main political leaders associated with the MPLA (Popular Movement for the Liberation of Angola) and UNITA (National Union for the Total Independence of Angola)—the political organizations that fought each other during the civil war (1975–2002)—they could certainly be designated as ocindele. For they were precisely those who occupied the ambiguous social positionalities that in Umbundu were connected to whiteness, colonization, Christianity, and the power of what is foreign (Dulley and Sampaio Reference Dulley and Sampaio2020). Moreover, the associations between the MPLA and Luanda, and between UNITA and the Central Highlands, points to the instability in the use of the signifier ocimbali and its chain of substitutions across different regions. That is because if most MPLA and UNITA leaders originated from the Christian assimilado elite (Péclard Reference Péclard2015), the alternative national projects advocated by UNITA and the MPLA index the different sociopolitical positionings of most of their leaders: the former associated with the more rural and ethnically “Ovimbundu” Central Highlands and the latter with urban and ethnically “Kimbundu” Luanda (Pearce Reference Pearce2015). This positioning was relational within and across regions. The term ocimbali could be applied to a member of the MPLA elite by rural Central Highlanders given the implicit association of this party with Luanda and urbanness; yet, whether or not an Umbundu speaker would apply this term to another Umbundu speaker depended on their social positioning vis-à-vis each other.

Iteration contains the possibility of transformation because the signifier is always dependent upon the context in which it is uttered, and context is never simply linguistic and sociocultural, but also historical and political (Siegel Reference Siegel2006). The indigenato regime contained in itself, and in its relation to vernacular forms of interpellation, the very fractures that eventually led to its extinction. The beginning of armed resistance against colonialism in 1961, at a time when most African colonies had already become independent, signaled to the Portuguese state the need to revise the possibilities provided for its colonial subjects. However, this system had always been challenged from within, for it is not possible to make a name the perfect equivalent of a person, let alone of a group of people. Transformations in postcolonial Angola are related both to external circumstances (i.e., changes in the sociopolitical context in which speech is uttered and rituals are practiced) and the disjunctions in social positionality that were already manifested in attempts at colonial translation. Thus, although one needs to acknowledge the violence and arbitrariness of the colonial system, one must also recognize its (partial) infelicity, in Austin’s (Reference Austin1962) terms, to the extent that the enactment of the indigenato regime could never fully reproduce the law. No strict distinction can be drawn between linguistic and ritual iterability to the extent that both draw on conventions that are simultaneously linguistic and social. As a consequence, transformation occurs both inside and outside the speech act because the latter draws its ritual force from the convention that sustains it. And convention is always already social, political, cultural, and linguistic.

Without denying the oppressive character of colonialism in Angola and elsewhere, one should explore its inherent contradictions. Drawing on différance to do so allows us to think of transformation as that which inhabits the brisures of structure, ripped apart in iterational displacement. For agency does not usually occur in the heroic and straightforward mode attributed to sovereign subjects; rather, it is both unpredictable and dependent on the language in which it is performed (Inoue Reference Inoue2006: 73). As iteration contains both sameness and difference, it is unlikely that anything will repeat itself in a merely reproductive mode. Thus, an institution is capable of enduring to the extent that the iteration of sameness can conceal the manifestation of difference. A position is maintained through the iterability of its relationship to the names and subjects it designates (Hansen Reference Hansen2001), but iteration always entails uncertainty. For a name to interpellate, it must appear as something other than a name: it must be taken for a proposition that simultaneously describes and transcends reality by quoting from established convention (Butler Reference Butler1997). Colonial legislation supposes that one should be able to judge whether a person corresponds to a name and the qualifications associated with it. The fantasy that structures the interpellation of subjects by the colonial state is the fantasy of absolute identity between the name and that which it names. Yet, as that which is can only remain what it is if it can add to itself the possibility of being repeated as such, the attribution of identity implies a continuous struggle against the possibility of subversion contained in the very names on which identity depends. Consequently, the possibility of transformation lies not in the intentional action of subjects,Footnote ¹⁶ but in their capacity to operate within the fractures of the relationship between language and society by drawing on the disjunctions that inhere in the unstable relation between signifier and signified, names and social positioning. Translation renders apparent the subject’s lack of control over language (Rafael Reference Rafael1993).

If ethnographic theorization is to be concerned with both historicity and difference within difference in order to avoid totalizing description (Dulley Reference Dulley2019), special attention is to be paid to the ways in which the units of analysis that stand for social groups are named in accounts of colonial and postcolonial contexts. In the case discussed here, this means not taking signifiers such as “Ovimbundu” for clear-cut units of analysis that supposedly bear a stable relation with a group of people “on the ground.” As this piece has showcased, such names contain in themselves not only the historicity of their iterations and ambiguities, but also a range of social positionalities that close attention to translational practices of subject constitution can reveal. This can only be observed if one considers the language in which these subjects have named themselves and others—a language that is only meaningful when placed in sociocultural, historical, and political context. The disjunctions between Portuguese and Umbundu forms of interpellation reveal the ambiguities and uncertainties that made it possible for colonial subjects to inhabit alternative social positionalities in post-independence Angola. This paper, grounded in the singularity of its ethnographic theorization in its engagement with the effects of colonial interpellation on historical and ethnographic rhetoric, invites a consideration of the consequences of translational displacement for interpellative subject constitution in colonial and postcolonial contexts.