Navigation – Sitemap

Relationships in motion

Oracular recruitment and ontological definition in Cuban Ifá cults
Rapports en mouvement : élection oraculaire et définition ontologique dans le culte d’Ifá à Cuba
Martin Holbraad
p. 219-264


Based on ethnographic fieldwork in Havana among initiates of the male diviner-cult of Ifá, this paper addresses the classical anthropological conundrum of how to make sense of practitioners’ contention that initiation transforms them into «new people». Taking into account the role of divination in «calling» neophytes, it is shown that Ifá initiation involves an accumulation of relationships with deities and fellow initiates that are «internal» in the sense that they pertain to the very definition of the neophyte as a person. It is argued that for such a notion not to appear logically absurd («apparently irrational») it is necessary to revise radically a series of analytical assumptions regarding the nature of change, obligation and personhood.



Dedicado a Javier Alfonso Isasi, Ifá alá Olorun,
Oggunda Teturá (1921-2004)
Padrino, ibaé elese Olodumare.


My Ph.D. fieldwork in Cuba from 1998 to 2000 was funded by the Economic and Social Research Council of Great Britain and hosted in Havana by the Centro de Antropología. Among the many religious practitioners who helped me in my research, I am particularly grateful to the late Javier Alfonso Isasi, to whom this article is dedicated. Without his guidance, as well as that of his son Javielito, I would be unable to write about Ifá. Some of the ideas developed in this paper took form in e-mail exchanges with Eduardo Viveiros de Castro to whom I am much in debt. Andrew Moutu’s ideas on internal relations were inspirational. Discussions in the Magic Circle at the Scott Polar Research Institute in Cambridge were similarly catalytic. Murray Last and Will Rea were generous with their advice on the West African literature. For insightful comments, I thank the members of seminars at the London School of Economics, the Museu Nacional at Rio de Janeiro, University College London and the University of St. Andrews, where I presented versions of this paper. Finally I am grateful to Michael Houseman for allowing me relationally (and so expansively) to reinvent my earlier ideas about motility for the present issue of Systèmes de pensée en Afrique noire, as well as the journal’s two anonymous reviewers for their apposite commentary.

  • 1 Houseman & Severi, 1998 : 278; Houseman, 2004.

1The most enduring anthropological puzzle regarding initiation ceremonies is how to account for the notion that a mere social occasion – however ritualised and exotic – may be deemed to transform people in such a thoroughgoing manner that they may become « new persons ». How might we account for the putative power of these ceremonies to transform those who undergo them? Residing in the tension between the way the analyst might expect things to be and the way they are elsewhere or for others, the question begs a further one, as to how to account for alterity more generally. The idea of initiatory transformation, after all, is no more or less « surprising » than other classics of anthropological debate, such as the appeal of witchcraft, oracles and magic, the obligation to return gifts, or the advisability of cross-cousin marriage. Faced with cleavages of this kind, between our own default assumptions (that initiations don’t really metamorphose neophytes in the way participants suppose, that witchcraft doesn’t really work, etc.) and the phenomena as we find them ethnographically, there are two available strategies. One is to predicate our sense of bizarreness on to the phenomena themselves, and then deem it our job to account causally for how such surprising phenomena may come about. Thus for Houseman & Severi, for example, the task of making sense of initiation ceremonies is spelt out as that of accounting « for the particular mechanisms whereby the various participants become committed to the irreversible metamorphosis these ritual experiences are held to enact1 ».

  • 2 Turner, 1969.
  • 3 Bloch, 1974, 1992.
  • 4 Whitehouse, 1996, 2000.

2This approach is common to most influential accounts of initiation. We may take just three examples. Turner’s famous idea that Van Gennep’s « liminal » stage of ritual is characterised by the negation of social differentiation (anti-structure) in favour of a more holistic « communitas » that « revitalises » social bonds, is most comfortably taken as an account of the efficacy of initiation rituals in terms of their social function2. Bloch’s early idea that the formalised character of ritual expression in Merina circumcision ceremonies serves to establish the authority of the elders, is essentially an argument about the causal conditions for the birth of ideology and its false constructs3. And, to take but one more recent example, Whitehouse’s suggestion that ceremonial violence during initiation is a means of encoding the symbolic content of the ritual in the neophytes » episodic memory, amounts to an argument regarding the cognitive conditions for the transmission of ritual cosmologies4.

  • 5 Latour, 2002.

3Perhaps there is nothing disreputable about causal explanations in anthropology. But there are perils when they are used to account for « alterity ». For the key assumption that allows causal accounts of « strange » phenomena to get off the ground at all is that the phenomena in question are well understood. Having branded the task of analysis as that of explaining their occurrence as effects of this or other process, the task of understanding phenomena like initiation is assumed to be a relatively straightforward matter of ethnographic description. Alterity, then, comes down to no more than a disagreement (e.g. they think initiation makes people new, we don’t). But liberal as it may sound, the danger with this view is that alterity is pasted over by way of misunderstanding5. For there is always the possibility – a strong one – that our surprise at things like initiation has something to do with our difficulty in understanding them in the first place. And, assuming that description is not equivalent to understanding (you could probably describe the mutterings of a schizophrenic, but that wouldn’t mean you’d understood them), this would imply that there is something more to the anthropological analysis of surprising data than a causal account of their occurrence.

  • 6 Viveiros de Castro, 2003; Holbraad, 2003.

4A second anthropological strategy in the face of alterity, then, would be to theorise misunderstandings rather than differences of opinion. On this view causal questions become irrelevant – even dangerous – inasmuch as they seek to provide antecedents to misunderstood effects. Rather, the job of analysis becomes that of clearing up the misunderstandings, and this project must be modelled not on science (which causalists worry about endlessly) but on philosophy, or at least that part of philosophy that takes conceptual analysis as its method6. After all, if alterity is an index of the inadequacy of the analyst’s assumptions, then the way to deal with it anthropologically must be to explore how those assumptions might be changed. Anthropological analysis, then, becomes a search for concepts that may be adequate to make sense of alien and therefore surprising data.

5This paper accounts for initiation from within this analytic space. The task is to make sense of initiatory transformation with reference to Ifá, the Afro-Cuban male cult that I have been studying in inner city Havana since 1998. The central thesis will be that the transformations that Ifá initiation effects need to be understood as ontological ones, in the sense that through them neophytes are redefined as different persons. This difference is articulated as an organised transformation of the neophyte’s relationships with deities as well as with fellow practitioners. This constitutes an ontological redefinition because the relationships effected are, as philosophers put it, « internal » rather than « external », i.e. they do not merely reposition the neophyte vis-à-vis others, but rather they constitute part of his definition as a person.

6In order to establish this analytically an indirect strategy is adopted which, instead of examining the transformed relations directly, focuses on the role of divination in bringing them about. This is both ethnographically feasible and analytically advantageous. It is feasible because divination has pride of place at the very core of Cuban Ifá initiation ceremonies, and is the key mechanism that determines neophytes’ initiatory course, and hence the relational transformations that this course implies. The focus on divination is also advantageous because, unlike the cult relations themselves, the diviners’ pronouncements offer a degree of transparency when it comes to deciding analytically on the matter of internal versus external relations. As it will be argued, these pronouncements are best understood as speech-acts that « redefine » the neophyte by relating him to meaningful entities in a way that reconstitutes him as a person (and hence the relations established are internal par excellence). As we shall see, the key notion of « inventive definition » has a number of surprising corollaries regarding the analysis of transformation itself. The brunt of the argument will be that these corollaries, which collectively comprise what I call a « motile » logic, provide an analytical bridge, so to speak, across which internal relations transmute from divination to the initiation ceremonies themselves.

  • 7 Holbraad, 2007.
  • 8 Bascom, 1991 : 84-85; Akinnaso, 1992. Nevertheless, I suspect that oracles may have been (or still (...)

7Before proceeding, and in view of this journal’s African remit, we may comment on questions of comparison. As an attempt to elucidate the logic of initiation in Cuban Ifá, my account is unapologetically parochial. Unlike causal explanations, which draw strength from universality (or perhaps « cross-cultural application »), the value of the philo-philosophical analysis here on offer is gauged only in relation to the material it purports to illuminate – a matter of one’s ability to transform concepts to fit an ethnographic description, rather than extending them to cover many7. Thus, for example, although in some ways Ifá initiation in Cuba is similar to Yoruba practices, ethnographers of Ifá and other related cults in West Africa have not emphasised the role of divination in cult recruitment, having tended to present cult initiation largely as a matter of family tradition or even personal choice8 Since the relationship between initiation and divination is key to my analysis, the present argument about the ontological character of initiation in Cuban Ifá cannot pretend to « apply » also to Yoruba cases.

8Still, an orientation toward concept production makes virtue of this. The possibility that concepts produced to account for the Cuban case might not « fit » in West Africa presents an opportunity to transform concepts further. So, for example, to the extent that initiatory transformation presents as much of an analytical problem in West Africa as it does in Cuba, one may test the limits of a Cuban- derived « motile logic », which renders initiatory relationships « internal » and therefore effective in an ontological sense, against the West African ethnography. If, say, in Yoruba initiation heredity sometimes plays a role equivalent to divination in Cuba, then what equivalent or alternative to motility (derived from Cuban divination) might one derive from the field of West African heredity? While such questions are not pursued here, they may be suggested by the ethnographic juxtapositions of this volume.

Ifá initiation as accumulation of relationships

  • 9 Henceforth when I write of « Ifá » I mean Cuban Ifá. In spelling ritual terms of Yoruba origin I fo (...)

9Ifá in Cuba is very closely related to Santería, the most well known Afro-Cuban religious tradition. Both have evolved on the basis of elements brought to Cuba primarily by Yoruba speaking slaves from West Africa, mostly during the 19th century9.The relationship between the two is most obvious in the fact that they share an extremely rich mythical and devotional universe, but they are also related ritually, since babalawos (i.e. full initiates of Ifá) are often required to officiate as diviners in Santería rituals. The prestige of the babalawos as diviners stems from the fact that, unlike santeros, they are initiated into the cult of Orula, short for Orunmila, the Yoruba god of divination, whom they have the privilege to adore. Indeed, the prestige of the babalawos is rather enhanced in the Cuban context by the macho credentials conferred upon them by the fact that only heterosexual men are admitted into the cult. Nevertheless, throughout its history in Cuba and up until today, Ifá has largely been practised by « marginal » groups, as Cuban intellectuals often say, in predominantly non-white urban neighbourhoods.

  • 10 Menéndes, 1995 : 51; Akinnaso, 1992 : 92.

10The starting point for understanding the structure of Ifá worship is that participation depends on initiation. While non-initiates (aleyos) may participate in some of the rituals that comprise worship, they are barred from dispensing them for two main reasons. Firstly, Ifá worship is deemed effective only insofar as it involves the use of a number of consecrated entities, and these are only bestowed on initiates. Secondly, the knowledge of how to dispense the various rituals is strictly esoteric. Initiates’ commitment to secrecy is captured by the very term babalawo, translated as « father of secrets » : baba = father, awo = secret10.

11A simple initiate/non-initiate distinction, however, does not capture the complexity of relations between practitioners since, on the one hand, Ifá initiation itself admits of various grades, and, on the other, practitioners who wish to enter Ifá must previously be initiated to Santería, and may also be initiated into other Afro-Cuban cults (for example the Bantu-speaking based Palo Monte complex). These two features complicate matters considerably, so it is worth outlining the possible stages in the career of an Ifá initiate in some detail.

12Table 1 depicts initiation as a gradual process that unfolds as a series of ceremonies over time. Ceremonies that are considered absolute pre-requisites for Ifá initiation appear in the central two columns of the table. Other ceremonies, which may emerge in any individual case as additional prerequisites if the oracle so ordains, are listed in the far left and far right columns.

Table 1 - Ceremonies undertaken on initiatory « path » of Ifá

Table 1 - Ceremonies undertaken on initiatory « path » of Ifá
  • 11 For a detailed discussion of the historical evolution of this ceremonial sequence in Cuba cf. Brown (...)
  • 12 The « warriors » include Elegguá (messenger god and « owner of the paths »), Ogún (patron of smithe (...)

13Not all the ceremonies listed here are properly considered as initiations. « Initiation » (iniciación), a term used by practitioners themselves, refers to the more ceremonious rites – such as the receipt of « mano de Orula », as well as full Santería or Ifá initiation11 –, which involve lengthy divinations that regulate various aspects of the neophyte’s further progress within the cult, as we shall see. Other ceremonies are best described as consecration rites, whereby the neophyte « receives » various consecrated paraphernalia for his or her personal protection (e.g. receipt of the so-called « guerreros » – the popular « warrior » deities12). Indeed, while each step of initiation or consecration opens up for practitioners a correspondingly wider range of ritual activities, the crucial leap from participation to dispensation is only achieved by those « practitioners who ‘‘become’’ or ‘‘make’’ Ifá » (hacerse Ifá), i.e. those who achieve the status of babalawos by undergoing the weeklong initiation ceremony. Only babalawos are entitled to dispense Ifá rituals, be they of divination, magic, or consecration.

14Nonetheless, seen from the viewpoint of their social implications, all stages of consecration have an initiatory character. From the first step of consecration practitioners are incorporated in networks of religious kinship, referred to as « families » (familia en la religión, familia de Ifá), which from then on provide a primary context for worship. As shown on figure 1, at the heart of Ifá kinship lies the relationship between a padrino (godfather) and his ahijado/a (godson/daughter). This relationship arises because consecration always involves « receiving » (recibir) the crucial consecrated items from an initiator who « has » (tener) them and is able to act as godfather by « giving » (dar) them, usually in return for money. Once completed, this ritual transaction engenders a lasting relationship that is cultivated in hierarchical terms : the godchild owes his or her padrino respect and is expected to visit him regularly, bringing gifts when possible. The padrino, in turn, is obliged to help and advise his godchildren on matters religious, both by coaching them through mythical knowledge and ritual practice, and by making his ritual services (divination, magic, etc.) available to them according to their circumstances.

Figure 1 - Ritual kinship relations in « religious family »

Figure 1 - Ritual kinship relations in « religious family »
  • 13 Holbraad, 2005.

15While this vertical kinship relation (godfather to godchild) is as simple as it is fundamental, it can engender a large degree of complexity due to the fact that godfathers can have many godchildren and vice versa. Babalawos have a free hand at attracting godchildren at quantity, and indeed tend to welcome as many as possible, since this is felt to enhance their personal prestige among practitioners, including considerable material accrual (money, gifts, etc.13). Conversely, a single practitioner may have different godfathers corresponding to separate ceremonies. Furthermore, the ceremony of full Ifá initiation requires the ritual assistance of a « second padrino » (usually referred to as ayibbón) who subsequently may take a primary role if the godfather-godchild relationship is disrupted for some reason (the godfather’s death or immigration, personal conflict, etc.). Finally, complexity accrues also on the horizontal dimension, since godchildren who share a godfather are viewed as siblings, and refer to each other as abbure, or, in Spanish, as « brothers/sisters in the religion » (hermanos en la religión).

  • 14 Brown, 2003.

16Still, of all ego’s ritual kin-relations by far the most important is with the godfather, not only because of its substantive ritual significance, but also because of its structural implications. Unlike other relations within the ritual « family », godfather-godchild relations can be iterated indefinitely through the generations, as godchildren who become fully initiated babalawos gain the right to acquire godchildren of their own, and so on. Referred to as ramas (branches), lineages of babalawos may in principle extend indefinitely into the past, although, since Ifá has a relatively short history in Cuba, babalawos will typically trace their genealogical line from three to maybe six or seven generations back, often to an « African » of the 19th century14.

  • 15 Menéndez Vázquez, 1995.

17The genealogical reckoning involved in lineage affiliations is extremely important to babalawos both ritually and politically. Ritually speaking, the lineage is considered sacred in its own right, and its efficacy is invoked at the outset of all rites by means of a ceremonious incantation of ancestors’ names – the so-called « moyubba15 ». Babalawos explain this sacred quality of lineages in terms of the ritual of initiation itself : the Orula idol-deities that godfathers « give » during initiation are said to be « born » (nacen) out of their own Orula. Moreover, the consecration of the neophytes themselves as babalawos is also spoken of as a birth, and specifically as « the birth of a king », as babalawos like to emphasise. Lineages of babalawos, then, are coupled with isomorphic lineages of Orula idol-deities, as illustrated in figure 2.

Figure 2 - Ritual lineages and the « births » of Orula idol-deities

Figure 2 - Ritual lineages and the « births » of Orula idol-deities
  • 16 Bascom, 1941; Holbraad, 2007.

18In summary, Ifá initiation is articulated as a gradual transformation of relationships on two planes. On the one hand neophytes enter increasingly wide circles of divine influence by receiving various kinds of consecrated items, including the Orulas themselves. My unwieldy term « idol-deity » is significant here inasmuch as practitioners make clear that, once consecrated, these objects are deities – in other words they are not to be glossed as « representations » of deities16.

19In fact what distinguishes the initiated from the uninitiated in this context is precisely the fact that while both groups are in principle subject to divine influence, the former are better equipped to enter into reciprocal relationships with the deities, by virtue of the care and attention they bestow on the idol-deities they have received during their initiation. Each deity has likes and dislikes in terms of the position they should occupy in the initiate’s home, offerings they should be given and, above all, the animals’ blood with which he or she ought to be fed. Over and above the regular care that they owe their deities, initiates may also appeal to them on specific matters by making occasional requests. These are accompanied by appropriate offerings, which may be decided through divination, and are often spoken or whispered directly at them.

20As we saw, however, such human-deity relationships are connected to a second plane of relationships, namely ritual kin-ties between humans – be they immediate, in the « family », or mediated along the lineage. Significantly for the present argument, such human-human relations are a logical consequence of human deity relations. So, for example, in the « coupled » constitution of lineages outlined above (figure 2), it is by virtue of « having » their own Orula that godfathers are able to preside over the « birth » of a new Orula, which subsequently the neophyte « receives » as a prerequisite for his own « birth » as a babalawo (see figure 3).

Figure 3 - The logical priority of initiate : deity relationships over initiate : initiate relationships

Figure 3 - The logical priority of initiate : deity relationships over initiate : initiate relationships
  • 17 Van Gennep, 1960.

21The same priority holds for ritual « family » ties (as opposed to lineages). For example, you might be my sibling (abbure) because we both received our Warrior deities, say, from the same babalawo, who in turn was only in position to give the Warriors to us because he had his own. And note that the logical implication here is not mutual. With respect to this (and each) ceremony, the babalawo does not give us the deities by virtue of being our godfather, but rather becomes our godfather by virtue of giving us the deities. Renovating Van Gennep’s famous thesis about the sacred character of social transitions, we may say that in Ifá relationships between people are both mediated and engendered by relationships with deities17.

Internal versus external relations

  • 18 Houseman & Severi, 1998 : 223-285; Houseman, 2005.

22We have so far characterised Ifá initiation as an unfolding accumulation of neophytes’ relationships with deities on the one hand and fellow participants on the other. In adopting « relationship » as a guiding concept to describe the transformations neophytes undergo I follow Severi and Houseman, who place relational transformation at the core of the very definition of all ritual18. In line with my introductory comments, however, I shall not follow Severi and Houseman in using relations to explain practitioners’ commitment to initiatory transformation. Rather the present argument is limited to using the concept of relationship to characterise this transformation. The question is this : insofar as practitioners are committed to the transformation initiation is supposed to engender; and if this transformation can be described as an organised accumulation of relationships; how, then, is the practitioners’ commitment to transformation best conceptualised? What could the idea that initiation induces relational alterations mean to us as analysts?

  • 19 Strathern, 1992, 2005; Latour, 2005; Law, 2004; Viveiros de Castro, 2002.
  • 20 Moore, 1922; Ollman, 1975.

23The question is reasonable when one considers that in a very ordinary sense relational transformations just are the fabric of social living. One would be hard pressed to think of any social occasion that did not involve a transformation of relationships (just think of the way your day has panned out so far today, and the multitude of relationships that have altered during its course, including ours as you read these words…). In fact one could probably argue that to call an occasion « social » merely indicates that it pertains to « relationships ». The present paper follows a recent trend of self-conscious attempts to provide this truism with a content, setting up the notion of the « relation » as an object rather than just a premise of analytical scrutiny19. In particular, I argue, in order properly to characterise initiatory transformation in terms of altered relationships in a way that does not just conflate it, bluntly, with all other social occasions (catching the bus, reading an academic paper, etc.), one would need an analytic distinction between different kinds of relational change.This is the kind of topic that has preoccupied philosophers at least since Aristotle, who parsed change itself as an alteration of the relations that hold between a subject and its predicates, a problematic pair of ideas to which I shall return. For now, it is expedient to draw attention to one particular philosophical distinction that arguably goes to the heart of the matter of initiation, between so-called « internal » and « external » relations20. An entity is « internally » related to another if this relationship is part of its definition. If the entity can be defined independently of its relationship to the other, the two entities are related « externally ». Series, systems, part-whole relations, and essential properties are often cited as instances of internal relation. E.g. the members of the natural number series (1, 2, 3…) are defined in terms of their relations to the rest; kinship terms constitute systems of mutual definition (« mother » in terms of « children », etc.); sentences are defined by their having words as their parts; and Socrates was defined by his humanity, just as humans were defined by their mortality. By contrast, my relationship, say, to my shirt is an example of an external relation : my definition does not, presumably, depend on wearing this (or any) shirt.

24Might the distinction between internal and external relations serve to sharpen our characterisation of initiation, so as to understand its difference from catching the bus and such? The criterion would be this : are the relationships that neophytes amass as a result of successive grades of initia-tion constitutive of their definition as persons? In such a case we could qualify these relationships philosophically as « internal », thus distinguishing them from relationships engendered in a bus queue, say, which are clearly external.

  • 21 Cf. Wedel, 2004 : 100-105 for santería initiation, and Goldman, 1985 for a remarkable analysis of t (...)
  • 22 The number 16 reigns supreme in Ifá divination, since the system of binary configurations on which (...)
  • 23 Holbraad, 2007; Cabrera, 2000 : 99, 103.

25Of course, my intention to argue that initiatory relationships are indeed internal, that they define initiates as persons, is transparent. However, ascertaining this analytically is not as easy as it may seem. No doubt, initiates do seem to indicate in a number of ways that they consider that initiation into Ifá has made them different people, as they also would express it21. A very common conception, for example, is that Ifá initiation « gives you health and development » and, more specifically, that it prolongs one’s life by sixteen years22. It is also said that every grade of initiation « gives » the neophyte aché, a mana-type concept-cum- substance, which in this context could be translated as potency, ritual as well as personal23. Most obviously perhaps, we have seen that babalawos do talk of full initiation to Ifá as the « birth of a King ». Nevertheless, babalawos tend to make such comments in a loose way. At any rate, one may be sceptical about using informants’ statements as replies to analytical questions, particularly when these questions cannot be addressed to them in a way that is not leading (« Do you think Ifá has changed your essence as a person? » is not a particularly good interview question!).

26I want to argue that the case for viewing initiatory relationships as philosophically internal can be made if one turns away from the relationships themselves, to look instead at the way they are comprehensively regulated by means of divination. For this it pays to examine the crucial role of divination in the process of Ifá initiation.

Initiation and oracles that define

27Divination is at the core of Ifá initiation in two important ways. Firstly, lengthy divinatory seances (called itá) are arguably the centrepiece of all the major ceremonies on the ladder of Ifá initiation, including the ceremony of full initiation as a babalawo. Secondly, each ceremony on the ladder of Ifá initiation can only be carried out following a prior oracular prescription, so that initiation is essentially premised on divination. In this section the first point is considered, arguing that the divinations carried out during initia-tion ceremonies transform neophytes by redefining them as new persons. In the following section the consequences of this are explored in view of the second point, regarding the divinatory premise of initiation. The overall aim will be to show that since the cult relationships that initiation implies are premised on divination, and divination is a process of redefinition, the cult relationships are also best thought of as defining the neophytes involved, which is just another way of saying that these relationships are internal. Before getting to these analytical tasks, however, we may give an idea of how Ifá divination works.

28In Ifá the consecrated paraphernalia associated with Orula are used to make the deity « speak », as it is said. Orula consists of a clay or porcelain pot containing a number of consecrated palm nuts. Practitioners emphasise that, along with a consecrated stone (otá) also held in the pot, these nuts « are » Orula. Divination involves a series of techniques that yield one out of 256 possible configurations of the palm nuts. The babalawo achieves this by casting 16 nuts 8 consecutive times in a manner that is equivalent to tossing a coin 8 times (hence the total of 256 possibilities). The resulting configurations are referred to either in Yoruba as oddu, or in Spanish as signos (signs) or letras (letters or verses). While practitioners explain that the oddu are a means by which Orula speaks, they also emphasise that each oddu is a divine being in its own right, sometimes thought of as a guise or avatar of Orula – what practitioners refer to as the deity’s « paths » (caminos). Furthermore, each oddu has its own name, as well as its own graphic sign (hence the Spanish term signo).

  • 24 See Holbraad, 2003. To compare with Yoruba cases see Abimbola, 1997; Akinnaso, 1992; Barber, 1991.

29Each oddu is associated with an open – though particular – set of characteristics, much like star-signs in Western astrology. These are described in an inordinate number of myths associated with each oddu, which are often referred to as « paths » (caminos del oddu), or « stories » (historias) – the ritual term is patakin (pl. patakines). Each oddu also corresponds to a number of specific aphorisms, character traits, advice, warnings about specific types of events, prohibitions, and other rules of conduct, as well as assorted ritual and cosmological data. Dedicated babalawos spend a lifetime memorising and contemplating these data, and bring them to bear on the circumstances of consultants during séances24.

  • 25 The permanence of one’s « sign of itá » (signo de itá) is most clearly manifested by the fact that, (...)

30The most ceremonial séances that babalawos are called on to perform are those of itá, which are performed for the benefit of neophytes during initiation ceremonies. The main reason why itá ceremonies are considered so important – more so than the more everyday consultations (consultas) that babalawos perform for clients – is that the information imparted in them is understood to be relevant to the whole life (and death) of the neophyte25. In itá, following the initial designation of the neophyte’s oddu, which is marked early on in the seance on a divining tray covered with consecrated powder (iyefá or aché), the neophyte is coached on the various features associated with his oddu by the babalawos presiding over his initiation. Unlike ordinary consultations, more than one babalawo is needed to perform itá (usually a minimum of four) because, as initiates explain, this ensures that as many aspects of the oddu as possible are covered.

31The initiators’ coaching typically takes the following form. First, with magisterial and sometimes competitive machismo, the presiding babalawos give a rather fragmented account of what they take to be the salient features of the oddu. « Ifá says you are a leader, your sign is the highest ranking of all », « this sign is about problems in the home », or « this is a strong sign, it means death is after you » are statements one might typically hear in this initial phase of coaching. These are usually explained or elaborated upon for the neophyte’s benefit with reference to the corpus of myths associated with the oddu. This process can take considerable time, as babalawos show off their rhetorical talents, and involves narratives such as the following, which was recounted during the itá of a friend of mine, in order to explain the aforementioned warning about domestic problems :

Once tiger visited the goat, asking him if he could sleep on his patio. Goat agreed, so the next day tiger asked him if he could sleep in the living room. Again the goat let him, so the next day tiger asked if it would be all right to sleep in the bedroom. Goat went to consult Orula to find out what to do. Orula told the goat that tiger was planning to eat him, and advised : « put some specs of corn in one hand and stones in the other ». So the goat took tiger for a walk and, on the way, started eating the corn. The tiger asked, « what are you eating? », and the goat showed him the stones. Tiger got scared and fled.

32Following the lengthy process of mythical story telling and discussion, each of the babalawos takes turns to give the neophyte instructions of various types, with reference to the characteristics of his signo. Significantly, these pronouncements result from what we might call a process of customisation. Since the various elements that enter divinatory pronouncements (regarding death, domestic affairs, and so on) are initially associated generically with each oddu, the diviner’s job is to render this data personally relevant the neophyte who « drew » the oddu in question (sacar el oddu). That this can require considerable skill is shown by the above example of mythical interpretation. Nothing in the myth of the obtrusive tiger speaks directly to the consultant’s circumstances. In this case, it is the babalawo’s awareness that in Cuba housing scarcity has put domestic problems at the forefront of people’s minds that allows him to « customise » the myth so as to render it relevant to the consultant.

33This raises the question : is the diviners’ ability to relate the data of the oddu to the neophyte during the itá ceremony to be understood in terms of internal or external relations? If it could be shown that internal relations are at stake here, then a big step would be made towards showing that other relational transformations that Ifá initiation effects are also internal (viz the cult relationships between neophytes and deities, and neophytes and ritual kin). This is because these relationships are also instigated by divination, as we shall see. However, the advantage of focusing attention on the relationships that diviners announce during the itá (e.g. « death is after you »), rather than the cult relationships that itá engenders, is that the former take the form of verbal affirmations. This is useful for purposes of our argument, considering that a definition – the operative concept that allows one to distinguish internal from external relations – is also a form of affirmation. Hence to broach the matter of internal relations we may consider the logical characteristics that may distinguish definitions from other types of affirmation, to see whether the oracular pronouncements of itá fit the bill.

  • 26 For a synthesis of the classical debates, cf. Robinson, 1950.

34Although a survey of philosophical discussions of definition is beyond the scope of this paper as well as my expertise, it may be fair to say that the debate among analytical philosophers of language in particular is concerned with questions of the adequacy of definitions26. Is it possible to capture the meaning of a word, a sentence, a concept, an entity, etc. (positions differ and proliferate in these debates!) in terms of other meanings, and what are the logical entailments of such a possibility or of its absence? Such a concern with adequacy, implicated in an assortment of intertwining debates in contemporary philosophy of language, is hardly surprising considering that the requirement for good definitions (or adequate at least) has been deemed central to philosophical inquiry, from Socratic « what is… » questions onwards.

  • 27 I adopt the standard distinction in the philosophical literature between definiendum and definiens (...)

35Here, however, we may foreground a notion of definition that analytical philosophers tend to pass over, and which arguably captures what is at stake in Ifá divination. One could broach the problem by pointing to the conservative, so to speak, character of analytical philosophers’ attention to adequacy. The assumption there seems to be that, when it comes to the problem of definition, the most interesting question is how meanings that already exist as a stock (albeit perhaps an uncharted one) may or may not relate to each other. Here is a horse, say, and the problem is to decide whether and how a sentence like « quadruped with a flowing main and tail » may serve to define it. Appeals to truth-conditions are prominent in these discussions. Since we have a sense of what a horse is, defining it must be a matter of getting this sense right, using terms that hold true of horses, such as « quadruped ». On this conservative view, then, definition must at the very minimum be a species of truth-claim, understood in the ordinary sense : a predicative assertion of important facts about the definiendum27. Aristotelians would call these facts « essential ».

  • 28 Heidegger, 1968; Deleuze, 1994; Nietzsche, 1997.
  • 29 Wagner, 1981; Ardener, 1989; Strathern, 1999; Latour, 1999; Viveiros de Castro, 2003.

36But what of inventive definitions? It is remarkable that philosophers, who do nothing if not muster powers of clear thinking and creativity in order to change what and how we think should presume a stock of meanings that is given. Maybe defining a horse, or even virtue, is a matter of articulating a sense one already has, but who could even begin to have a sense of a Platonic Form, a monad, or sense data before philosophers ventured to define them? Indeed, those philosophers who have followed Nietzsche in thinking of philosophy as an « untimely » enterprise, have sought reflectively – inventively – to theorise invention28. And so have anthropologists who see the creation of new meanings not just as a philosophical prerogative, but as an irreducible aspect of social living29.

  • 30 Wagner, 1977.
  • 31 It may be noted that in order to make this argument (identifying the predication of novel meanings (...)

37Here we may venture one distinction between inventive definitions and the ones analytical philosophy is mainly concerned with – let us call them « conventional », following Wagner30. The distinction pertains to the question of truth. As already noted, conventional definitions are a species of truth-claims : if « quadruped » serves as part of the definition of a horse, this is because the statement « horses are quadrupeds » is true. Now, as an example of what I take to be an inventive definition, consider a definition of inventive definition – as pudding to my proof, so to speak. Let us define inventive definition as a speech-act that inaugurates a new meaning by combining two or more previously unrelated meanings. I did it just now! Taking the meanings « speech-act », « inauguration », « novelty », and « meaning », I combined them to inaugurate a new meaning – new relative to convention –, which I call « inventive definition ». Now consider whether this kind of definition is a species of ordinary truth-claim. If it were, « speech-act », « inauguration », « novelty », and « meaning » would all have to be construed as properties that predicate « inventive definition », like « quadruped » and « flowing main » predicate « horse ». Our definition of « inventive definition » would be deemed adequate insofar as it picked out properties that hold true of such definitions (viz that they are speech-acts, they inaugurate, etc.). But this would be contradictory. The notion that definiens must pick out properties of their definiendum implies that the definiendum is not new : it must already be there to be picked at, acting as a reference by which its definition may be checked for truth. Posited as a condition for its own definition, the definiendum takes logical precedence over its definiens, and hence the latter cannot be said to inaugurate it. So, since inventive definitions are defined inventively as inaugurations – as inventions of (new) meanings –, it follows that, unlike conventional definitions, inventive definitions are not predicative truth-claims31.

  • 32 Holbraad, 2003.
  • 33 Evans-Pritchard, 1976.
  • 34 This would be obvious if one took the statements that in Ifá there are no mistakes or lies as attem (...)

38I wish to argue that divinatory pronouncements are inventive definitions in just this sense. Consider two statements that babalawos frequently repeat, adage-like, when discussing the wonder of divination. « Ifá doesn’t make mistakes » (Ifá no se equivoca), they say, and « in Ifá there are no lies » (en Ifá no hay mentiras). Babalawos themselves may certainly lie or make mistakes, since they are « imperfect humans », as one practitioner put it, but not the oracle of Ifá. As I have argued elsewhere, these statements effectively bar the logical possibility that genuine oracular pronouncements might be false at all32. Contrary to Evans-Pritchard’s famous suggestion that practitioners of divination shield themselves from oracular errors by means of « secondary elaborations » – further « mystical beliefs » that may account for the error33 –, for babalawos the very notion of « oracular error » is an oxymoron. If oracles don’t lie and don’t make mistakes then oracles that are wrong are not oracles34.

  • 35 Holbraad, 2004.

39The most telling conversation I had during fieldwork on this matter was with a young babalawo about, as he called them, « exploitative » babalawos35. One of his main complaints was that some babalawos seek to impress their clients by attaching Orula’s verdicts to specific dates or people’s names (e.g. « your daughter will fall ill next Tuesday », or « the witch is your neighbour, Maria », etc.) :

I don’t give people dates : Ifá speaks past, present, and future, and gives advice, but [consultants] should know for themselves their own situation and act on it as they see fit. […] Some [babalawos] do give them, but that’s just showing off, and clients complain when things don’t turn out that way. How can they know these things? [i.e. the babalawos] Ifá doesn’t work like that.

  • 36 Holbraad, 2003 : 52-53; Werbner, 1973.

40How are we to understand the claim that « Ifá doesn’t work like that »? Why can Ifá « give advice » but not dates and names? When I asked my informant to elaborate, he reverted to a point babalawos tend to repeat like a mantra in such conversations, namely that « Ifá is interpretation », implying that interpretation of itself could not yield specific names and dates. Admittedly, the implication is moot since Ifá is used, in certain contexts, to give unambiguous « yes » or « no » answers to specific questions posed to it by the diviner36. I would suggest rather that the import of babalawos’ normative insistence on such terms as « advice » (consejo) and « interpretation » (interpretación) has more to do with the question of falsifiability than with ambiguity. By adding dates and names, « exploitative » babalawos present Orula’s verdicts as statements of fact that may be verified or falsified according to how things actually turn out. « Advice », by contrast, may be « interpreted », or even « acted upon », but not falsified as such. Similarly, the idea that « Ifá speaks past, present and future » (Ifá habla pasado, presente y futuro), which babalawos often emphasise during divination, is another normative mitigation of falsifiability. When, in the itá for my own mano de Orula ceremony, I queried – in falsifying mood! – the oracle’s contention that I am prone to impotence, one of the babalawos reminded me with macho gusto that I have no children after all. « And don’t forget, » he added with emphasis, « Ifá speaks past, present, and future ».

  • 37 See note 33 above.
  • 38 Viveiros de Castro, 2002 : 134.

41Normativity is important here. For, like the statement « your daughter will fall ill on Tuesday », an oracular verdict like « you are prone to impotence » could perfectly well be interpreted as truth-functional, and thus rendered falsifiable. But what if oracular pronouncements were to be interpreted in non-truth-functional terms as inventive definitions? Babalawos’ insistent muddying of truth-functional waters through their normative commentaries suggests that such an alternative may be analytically appropriate. For my informant’s dispute with « exploitative » babalawos who give names and dates can be cast precisely as a dispute over whether oracular pronouncements ought to be taken as falsifiable truth-claims or not. Just like Evans-Pritchard, and other analysts of divination who have taken for granted that oracles deliver predicative truth-claims37, practitioners themselves are liable to make the same assumption – and this pitfall may be readily exploited by babalawos who wish to « show off ». No wonder, then, that babalawos should find it necessary to stress normatively that the oracles’ pronouncements are not affirmations that might, least of all in principle, be falsified. And if they, as practitioners, find such a category mistake perilous, then so should we as analysts38.

42Let us then spell out the implications of treating divinatory pronouncements as inventive definitions. We noted earlier that diviners arrive at verdicts during the itá by establishing relations between neophytes and various features associated with the oddu (myths, injunctions, names, etc.). So, verdicts like « death is after you » or « you are prone to impotence » established a relation between myself and death or impotence. Treating the verdicts as inventive definitions would imply that « death », « impotence » and the like are not predicates that « hold true » of me. Rather, they are meanings that are related to me so as to redefine me. By telling me that death is after me the oracle renders a specific relationship to death as part of my definition as a person, which is to say that my relationship to death is rendered internal.

43In sum, rather than ascribing the consultant with a set of properties that may be falsified in light of experience, divination defines the consultant as a new kind of person. From being a person who stands in no particular relation to death, the oracle transforms me into a person death is after. To ask whether such a shift is « true » or « false » is fundamentally to misunderstand the ontological character of the transformation, by confusing it with the epistemological question of how the shift may be ascertained. The fact that practitioners of Ifá might be as liable to fall into this trap as analysts have been explains why babalawos put such normative emphasis on the requirement that the oracles’ pronouncements be interpreted as non-falsifiable.

44Now, this way of thinking of divination may seem awkward. Maybe a concept can be defined afresh, but can the notion of invention really be stretched to include people as well? For, more than just a meaning (e.g. « inventive definition »), I am a flesh-and-blood person, and it is unclear how as such I can be « brought together » with meanings such as « death » (and if I were literally brought together with death presumably I’d be dead?), just as it is unclear how I, flesh and blood, can be a « new meaning », or new at all. In fact, isn’t this talk of oracles’ ability to transform people philosophically suspicious? For it would seem that what is being propounded here is a version of social constructivism (« oracular constrictivism » if you will), based on the preposterously idealist notion that entities of the world can be brought in and out of existence by mere human fancy – divinatory or otherwise.

  • 39 Latour, 1999.

45Such appeals to common sense are cheap in their professed transparency. As Latour has pointed out, from a properly anthropological perspective the dilemma between realism and idealism is false39. Constructivism is indeed preposterous inasmuch as it comes as a remedy for a bias of its own premise, namely that « our ideas » and « the world » cannot but constitute distinct ontological camps. For only on this premise does it even become possible – let alone necessary – to say that what « appears » to be real and mind-independent is actually « only » a human construct.

46But all this means is that to understand Ifá divination and initiation it is necessary to deny that the distinction between concepts (or meanings) and things (or people) is axiomatic. Instead of reducing to the absurd our claim that oracular pronouncements have ontological effects, appeals to common sense just alert us to the fact that the sense of divination, with its seemingly bizarre entailment that assertions may be unfalsifiable at the level of logical principle, is quite uncommon.

47And it is uncommon in just the way Latour has envisaged. For what differentiates predicative (conventional) from non-predicative (inventive) definition, and what makes the latter rather than the former appropriate to the analysis of divination, is precisely that predication presupposes the « common » distinction between word and world, while invention does not. As we saw, the truth-functional character of conventional definitions is premised on the logical priority of their definienda. And, as we also saw above, although this logical priority does not imply that the definienda in question exist (e.g. defining unicorns as « horses with horns » does not mean that they exist), it does imply the existence of a domain of entities – the « world » – from which conventional definitions may draw their truth-values. By contrast, inventive definitions do not presuppose the existence of a world of entities : such a world is their conventionalised outcome. When the oracle of Ifá defines me as a person death is after, it is not speaking of an entity (viz me) existing out there in the world, of whom certain properties may be said to hold (viz that death is after me). Such a construal would imply that the definition in question is conventional and open to falsification – a notion babalawos are keen to deny. But nor does the inevitable consequence, namely that in defining me the oracle is bringing me about as a new person, imply the idealist notion that the world’s constituents exist insofar as people (such as diviners or anthropological analysts) think or speak of them. For, just like the notion of conventional definition, idealism is premised on a logical distinction between word and world, concept and thing, etc. – in other words, precisely that distinction which the logic of invention obliterates.

48It follows that in committing us to a notion of invention, divination leads us a fortiori to posit an ontological parity between what common sense distinguishes as « concepts » and « things ». Thus our analytical imagination is further stretched to include a monistically construed plane of concepts-cum-things that incorporates elements like me, death, impotence, inaugurations, novelties, meanings, and horses with horns too. Each of these may be brought together with others so as to engender new elements, such as me whom death is after, inventive definitions, or even fairytales with unicorns in them. Before closing I’ll make explicit some important corollaries of positing such a plane of concepts-cum-things. But first we may note that the claim that divinatory pronouncements redefine neophytes by relating them internally to data associated with the oddu puts us in good position to settle the matter regarding cult relationships also. This becomes apparent when one considers the second point mentioned earlier, namely that it befalls the oracle to regulate neophytes’ course through successive ceremonies of consecration and initiation.

Divinatory obligation and internal relations

49As already mentioned, it is the oracle that determines whether a given practitioner « has the path of Ifá » (tiene camino de Ifá), i.e. whether he has to become a babalawo. The sense of journey that the notion of a « path » conveys is reflected along the vertical axis of the matrix of consecration and initiation presented in table 1. Within this matrix of possible ceremonies the oracle’s regulative role is pervasive, since every stage of consecration or initiation is subject to divinatory confirmation, and cannot proceed without it. Hence, under this arrangement, in order to become babalawos potential neophytes are required to undergo a whole series of divinations, the most crucial being the itá séance that is carried out in the ceremony of mano de Orula. Whether the neophyte has the path of Ifá depends on which of the 256 possible signos he draws in this itá. Since only a proportion of combinations indicate that Orula is « calling » the neophyte to initiation, as practitioners sometimes say (la llamada de Orula), the matter is uncertain. To illustrate how oracles regulate practitioners’ initiatory course I quote from a letter written in response to my request by my friend Javielito Alfonso, a babalawo in his 40s from a well-established lineage in Havana :

  • 40 The Cuban authorities’ tolerance of the cults in a socialist context is a relatively recent phenome (...)
  • 41 There is no need to reveal this personal information.
  • 42 « Washing » a deity (lavar el santo) is a cheap alternative to full Santería initiation (see table (...)

In relation to my steps of initiation, I can tell you that at a very early age I received the Warriors from the hands of Isaias Mejias, my father’s godfather. I suppose I must have been 3 or 4 years old, I do not remember the moment. At 10 years of age at 1969 I received mano de Orula […] Only from that moment onwards did I begin to have notions of Ifá, though still in a rather indirect manner […] I say this because Ifá was not my interest as such, since at that time what really interested me were music groups of rock, pop, and soul of the 60s, and furthermore I was aware of the negative attitudes regarding the religions at the time40. In my mano de Orula [viz the ceremony] I drew the signo ****41 and it was determined that my guardian angel is Ogún. To the question whether I have to seat the santo [viz undergo the full initiation to Santería], Orula said no, so it was determined that I have to wash42 Ogún and become Ifá. My disinterest continued until finally I decided to become Ifá, influenced by the fascination of the stories [viz the myths], listening to the babalawos speak, and seeing that many people with less background or none at all were beginning to enter the religion. Also, a principal aspect of the signo that I had drawn in my mano de Orula is difficulty at advancing in life, or bringing one’s aspirations to fruition. This signo speaks of the crab of whom there is the proverb that says that he walks backwards and has no head, so the babalawos tell you that you need to become Ifá in order to change and to be able to prosper in life, etc. Thus in 1988 I washed Ogún and in August of that year I became Ifá. In Ifá I drew **** as my principal signo.

50The first point to note is the clarity with which Javielito lays out the regulative role of the oracle, which dictated his course from ceremony to ceremony. The itá of his mano de Orula ceremony « determined » that he should wash Ogún and then become Ifá, and so he does, notwithstanding his initial « disinterest » – which will be addressed presently. This suggests an important step towards establishing that the relationships that initiation generates are « internal ». If, as we have argued, oracular pronouncements are best construed as redefining the neophyte, then this must hold also for the oracle’s « call to Ifá », as well as all its other regulative prescriptions regarding previous and subsequent ceremonies. On this view, the oracle’s determination that Javielito should wash Ogún and become Ifá redefines him as a person by relating him « internally » to these deities, just like, by saying that death is after me, the oracle redefined me as a person by relating me internally to death.

51However, this is only a first step towards characterising the relationships that are generated by the ceremony of initiation itself. As Javielito’s delays demonstrate, the neophyte-to-deity relationships that oracles establish are not identical to those established during initiation. In fact, all the oracles can do is to redefine potential neophytes as persons who have to receive certain deities. In other words, the internal relationships established with the deities in itá divination imply only an obligation, projected into the future. What needs to be shown is that the internal character of these obligations carries over to the relationships that consummate them, when neophytes actually receive the deities during initiation. If this were to be established, the final step of showing that ritual kin relations between initiates are also internal would be a small one, since logically speaking ritual kinship comes part and parcel with receiving the deities, as seen in figure 3.

52The key to understanding how the internal character of relations carries over from divinatory prescriptions to initiation lies, I want to argue, in the notion of obligation itself, which is the second element that Javielito emphasised in his cogent account. Indeed, Javielito’s references to « difficulties » and the need to heed the oracle’s call to become Ifá in order to bring aspirations to fruition are atypical only in their relative mildness. As a matter of fact, the obligation to heed the oracle’s prescriptions in all matters, including sacrifices and other injunctions and prohibitions as well as the all-important matter of initiation, is often a matter of considerable anxiety for practitioners. To disregard oracular imperatives – or, as practitioners say, to « fall into disobedience » (caer en la desobedencia) – is considered an affront to Orula himself. And, as the myths of the oddu continually reiterate, the deities’ wrath can be awesome in such cases, as indeed can be their favour when the oracle’s strictures are followed.

  • 43 Holbraad, 2004.
  • 44 Hagendorn, 2001.

53Initiation is particularly problematic in this context. With prices currently ranging from $1 000 to more than $5 000, Ifá is one of the most expensive of what is in any case an inordinately pricey set of ceremonies, so most practitioners have mixed feelings about the prospect of having the oracle mark them a path of Ifá43. Notwithstanding its great prestige, only a small proportion of potential neophytes can actually afford to go through with the ceremony. For most, having a path of Ifá implies a long and frustrating period of waiting for a financial opportunity to come along, and this is by no means an easy matter given the present economic outlook in Cuba44. Indeed, the frustration that potential neophytes experience for as long as initiation remains beyond their financial reach is compounded by an element of fear. For it is understood that a pending initiation can anger the deities concerned, particularly if the chosen neophyte seems to be ignoring the oracle’s call by procrastinating his initiation, failing to save up, or to take advantage of the right circumstances when they come his way. In such cases all sorts of incidental misfortunes that might afflict the potential neophyte may be explained in terms of his failure to heed Orula’s call, and this provides babalawos with a certain amount of emotional leverage when it comes to impressing the urgency of initiation upon potential god-children – a theme to which I shall return.

54So obligation, with its emotional sanctions, is the link that connects the oracle’s initial call with the ceremony of initiation itself. I want to argue that obligation acquires certain peculiar logical characteristics in the divinatory context of Ifá, and that these characteristics explain why obligation furnishes, so to speak, a logical bridge across which internal relations are carried from divination to initiation. Note that such an argument is necessary precisely because, construed in an ordinary sense, the notion of obligation would appear not to provide such a bridge. On the commonsense view, Orula’s « call » would be deemed obligatory, by means of the divine sticks and carrots Javielito refers to, just to compensate for the absence of any ineluctable logical link between what one is told to do (divination) and what one actually does (initiation). Divination does not imply initiation, on this view, so it may as well compel it.

  • 45 For a well-rounded though technical exposition of the logic of obligation, cf. Horty, 2001.

55Here it is worth making explicit some of the premises that underlie this commonsense view of obligation45. For a start, obligation is taken as a means of regulating choice : it only makes sense to oblige someone to act in a certain way (e.g. to get initiated) if he may have acted otherwise. In other words, obligation is posited against the background of a set of distinct possible courses of action, including the prescribed course and its assumed alternatives. In the case of divinatory « calling » the alternative that seems to motivate obligation in the first place is a state of inaction, namely the fact that potential neophytes have not yet been initiated, and this state of affairs may be taken as logically distinct from the prescribed action, i.e. initiation itself. Thus the common sense take on divinatory obligation implies a particular view of temporal distribution, whereby actions and/or inactions succeed each other as logically discrete units. What unifies these logical snapshots is, of course, the fact that they are ascribed to a single agent. So if obligation presupposes a series of discrete actions distributed over time, it also presupposes a single perduring subject who either performs or fails to do so at different times. In other words, the commonsense notion of obligation is premised on a straightforward Aristotelian analysis of change, based on digital alterations of a subject’s predicates at different times. In our case this would be spelled out as follows :

At the time of the oracle’s « call » the person (subject) is not initiated (predicate).

At the time of initiation the person (subject) is initiated (predicate).

56Straightforward as it is, talk of perduring subjects and altering predicates should alert us to the fact that this view of obligation may be quite inappropriate to the analysis of the oracle’s call to initiation. For, as we have seen, Ifá divination turns on a logic that precludes subject/predicate distinctions. In fact, by inaugurating new meanings, divinatory pronouncements turn the logic of Aristotelian change inside out or, better, back to front. Rather than imagining that definition must start with a subject (definiendum) upon which predicates (definiens) may be applied, the inventive definitions that oracles effect take the definiens as raw material out of which a definiendum may be crafted afresh by a process of synthetic combination.

57Now, it may be objected that this logical discrepancy is irrelevant to the question of divinatory obligation. On this view obligation, with its subject-predicate baggage, pertains to the product of divination, i.e. the « new person » that oracular definition has invented, whereas the reverse logic of divination pertains only to the process of invention itself, which is both temporally and logically anterior. The problem with this suggestion, however, is that is presents the oracle’s power to compel as somehow separable from the business of divinatory definition, as if obligation were a contingent or « extra » feature that may or may not accompany the oracle’s call. This hardly tallies with practitioners’ own view, according to which the oracle’s pronouncements are taken as obligatory by virtue of being oracular (much like they are taken to be true for the same reason). For practitioners of Ifá it would be nonsense to say of a pronouncement that it is divinatory and then to wonder ex post facto whether it is also obligatory. As we saw, Orula’s word is the kind of word that demands obedience, and that’s that.

58It follows that obligation needs to be understood as a function of the logic of divination itself, and this requires a wholesale departure from subject-predicate assumptions. The question is this : what are the logical contours of obligation when it is seen as a corollary of inventive definition? In particular, if subject-predicate assumptions stand in the way of properly analysing divinatory obligation, then what logical image are we left with if we discard such assumptions? What would obligation look like if it were not parsed as a digital alteration of a subject’s predicates?

59As we saw, subject-predicate analyses imply a specific image of change itself : a self-identical subject perdures through a serial succession of different discrete predicates. A crucial point to note here is that this analysis renders change as an essentially derivative phenomenon : changes are broken down in terms of a temporal succession of putatively more basic logical units, which are not in themselves subject to change. On the one hand, the substrate that guarantees the unity of the phenomenon of change (viz the subject) is posited as a perduring and self-identical entity. On the other, the predicates that hold of it do not in themselves change either : they succeed each other as discrete self-identical states which, as such, can be placed between quotation marks, or designated by a single symbol (property « U » for uninitiated, « I » for initiated, etc.), just as subjects can (« p » for person, etc.).

60What if one denied the Aristotelian model by rendering change non-derivative? What if one took the basic logical unit for the analysis of change to be, not unchanging and discrete data such as subjects and their predicates, but rather changes themselves? According to such a view the problem would not be how to account for the fact that things change (to which the logical armoury of subject and predicate is one solution), but rather that of showing how changes can become things, so to speak. But this problem we have already begun to solve in our analysis of inventive definition. As we saw, for the idea of « inventive definition » to make sense one has to deny the (otherwise) axiomatic distinction between a subject and its predicates (e.g. me and death) in favour of a monistic plane of concepts-cum-things that can be combined with each other to define new concepts-cum-things (e.g. me whom death is after). So we already have a rudimentary account of how (concepts-cum) things come about, and this begs our present question nicely : is this account compatible with the analytical reversal that divinatory obligation seems to force upon us, namely that the basic units that make up things like « me whom death is after » (viz I, death, etc.) are changes rather than stable entities?

Figure 4a - « Motile » transformation

Figure 4a - « Motile » transformation

Figure 4b - Aristotelian predicative change

Figure 4b - Aristotelian predicative change
  • 46 The fact that for purposes of this minimal example the fusion has had to be characterised verbally (...)
  • 47 Bateson, 2002 : 103.
  • 48 Viveiros de Castro, 1998 : 15-17.
  • 49 Geertz, 1993 : 28-29.

61The non-predicative or, as I’ll call it, « motile » image of change that follows from the inventive logic of divination is depicted in figure 4a, and can be contrasted with the image of predication in figure 4b. Three features should be emphasised. First, as the symmetry along the lateral axis of 4a would indicate, this monistic picture presents myself and death (which in 4b are presented as subject and predicate) as the same kind of thing, namely as trajectories of change. And note that the same goes for the altered trajectory that results from their fusion, myself-whom-death-is-after46. A second point follows, namely that according to this view change occurs seamlessly : rather than a digital disjunction of a property and its negation before and after the change as in 4b, what we have here is an « analogue » fusion of concepts-cum-things that produces more of the same : another concept-cum-thing – viz me-whom-death-is-after47. So the novelty that the change produces should here be thought of as the kind of novelty transformations bring about, the creation that ensues when things differentiate themselves by combining with others48. After change, it’s changes all the way down, to recast the anthropological saying about turtles and analysis49.

  • 50 For a classic argument on the analogy between temporality and modality (i. e. between different tim (...)
  • 51 Holbraad, 2003 : 65-66.

62From this idea of seamless transformation follows the third and, to our purposes, most significant point, which has to do with choice. As explained earlier, the idea of choice turns on subject-predicate logic, since choices are premised on logical distinctions between possible courses of action that can be distributed discretely over time50. So choice is incongruent with our motile image. Indeed, to imagine change as a seamless transformation of further changes is to imagine what could be called a hyper-determinist world. If the trajectories that make up a transformation are seamlessly connected, and no data other than trajectories influence their course (…change all the way down), it follows that their courses could not but be as they are. For to imagine an alternative course (as one must if one is properly to speak of choosing) is conceptually to interrupt a change in order to intercalate its putative alternative. And this would be to render the change discrete – effectively, to « conventionalise » it as a subject. In other words, each transformation of concepts-cum-things is as it is because the concepts-cum things that it combines are as they are because of yet other concepts-cum-things, and, since this regress is infinite, each concept cannot but be as it is. Elsewhere I have called this characteristic of motility « necessity », but perhaps « pure actuality » would be more accurate, since ordinary philosophical analyses of necessity show that it can be imagined from within a predicative frame, whereas what we have in mind here cannot51.

  • 52 In contrast to well discussed cases in West African ethnography, notions of « destiny » or « fate » (...)

63These considerations allow us finally to answer the question regarding the motile view of obligation. True, this view leaves no space for a concept of obligation that is bound up with the option of choice. But this is its virtue for our purposes. For the notion that having a « path of Ifá » marked by the oracle implies that the potential neophyte has to undergo the initiation can be rendered in purely motile terms. On such an analysis the divinatory verdict that says « you have to become Ifá » does not « prescribe » a course of action in the face of possible alternatives, but rather describes (or predicts) a course that cannot but be as it is, since it is trajectile. On the oracle’s say-so, becoming Ifá is something one « has to » do in the sense of having the ceremony ahead of one, as a necessary constitutive element of who one is (rather than in the normative sense, where « have to » is synonym to « ought to »). Were the term not too loaded, we would say that on a motile view what seems like obligation is constituted as destiny in the strict sense : what one « has to » do is one’s destination, a destination that is given by the trajectile concepts-cum-things that oracles bring together so as to redefine the people who consult them52.

64At this point it should begin to become apparent why obligation, in its motile guise, allows us to extend the point about internal relations from divination to initiation – the crux of the present argument. What originally prevented such a step was the idea that the divinatory « calling » to initiation and the ceremony of initiation itself are logically separate. However, a motile analysis of obligation dissolves this logical separation. The temporal lag from « calling » to ceremony is rendered a logical feature of initiatory transformation, since, on this view, the constituents that are transformed (the neophyte and the deities) are themselves transformations that unfold over time. Time, in other words, unites divination and initiation as seamlessly connected moments on a trajectory, rather than separating them as discrete actions. By the same token the logical difference between obligation and its consummation is effaced. Here obligation is not an extraneous normative add-on, over and above the actions that it « modifies », as the famous is/ought distinction would have it. It is a function of the motility of the action of divination itself : the transformation of transformations that the oracle effects is one of « directing », seamlessly, the neophyte towards his initiation, by redefining him as someone who « has to » do it.

65So when practitioners speak of the initiatory course of Ifá as a « path » (camino) they are being more telling than metaphorical : practitioners move from one ceremony to the next as paths whose course is set by divinatory stricture. The tota simul view of table 1, therefore, which presents practitioners’ initiatory ascent as a sequence of distinct possibilities, is strictly speaking misleading. Defined by the oracles, each practitioner’s path (or better, each practitioner-as-path) unfolds amidst the available steps of initiation, as a unique, continuous and therefore purely actual transformation. Mixing a metaphor, we may say that oracles regulate initiation by channelling neophytes’ paths like dams that harness a river. And the paths are as they are « obligatorily » in the uncommon sense of having their course marked by oracles that preclude the possibility of choice itself – the possibility of possibility if you like.

66It follows that if divination relates people to deities internally, then so do the initiation ceremonies in which divinatory « calls » are supposed to be, as it were, consummated. Obligation subsumes its own consummation as an inevitable result inasmuch as the consummation is as motile as the obligation. After all, imagine the opposite : say we construed initiation in predicative terms. A neophyte’s receiving the deities would then have to be thought of as the process by which he comes to « have » them like a subject « has » properties – an apparently tenable position. This would be inconsistent with saying that the ceremony (viz the receipt of the consecrated deities) was obligated by the oracle’s call. For treating the « receipt » in predicative terms amounts to treating the neophyte as a logical subject (i.e. conventionalising him as a discrete entity), and this is effectively to sever the motile trajectory of obligation at its most crucial point, namely that of consummation. If deities were related to neophytes as predicate to subject then, from a logical point of view, their relation would be necessary at most, i.e. the deities would at most be what Aristotelians would call « essential » to the neophyte. We have seen that oracular obligation requires the stronger relation of pure actuality – a relation that bars the possibility of possibility, which predicative concepts of necessity and essence do not.

67Thus, by virtue of its own motility, the model of oracular obligation presented in figure 4a reduplicates itself in the initiation ceremony that it precipitates, as seen in figure 5. It should be clear that the relations involved here, between deities and humans, are indeed internal. In their receipt, the deities transform the neophyte as someone who « has » them. This logical analysis provides a rigorous sense to our earlier ethnographic hints at ontological transformation, such as the idea that initiation is a regal « rebirth » that gives health, development and potency (see above). Inasmuch as these characteristics are gained by virtue of receiving the deities, they can only be thought of as characteristics that define the neophyte, and definition is the hallmark of internal relation.

Figure 5 - Initiatory transformation in Ifá

Figure 5 - Initiatory transformation in Ifá

68By the same token, the logic can be extended to the second plane of relationships initiation precipitates, namely the networks of ritual kinship associated with cult lineages and « ritual families ». As we saw in figure 3, gaining a godfather (and with him the whole network of ritual kin relations) is a matter of receiving deities from him. It follows that ritual kin relations cannot be external. If receiving a deity (i.e. becoming initiated) just is to gain a godfather, and receiving a deity is an internal relation, then gaining a godfather is also an internal relation, constitutive of the redefinition that neophytes undergo. To wonder whether the kin relation might be external is to wonder whether it might be possible to be initiated without the auspices of a godfather, and this possibility is barred by the logic of Ifá initiation.


69In one sense the idea that initiation transforms neophytes by redefining them as different people is rather banal, at least inasmuch as it reflects an intuition that may seem obvious. And the central distinction used to establish this here, namely between external and internal relations, is itself hardly novel. Nevertheless, the logical framework that has had to be developed to account for these conclusions – with notions like inventive definition, concepts-cum-things, pure actuality, motile obligation, etc. – shows that banal intuitions may sometimes rest on counter-intuitive assumptions. As discussed in the introduction, the counter-intuitive character of these assumptions is a function of the bizarreness of the ethnography they claim to illuminate – in this case babalawos’ conviction that their initiatory « birth » is rendered obligatory by divine « calling », expressed indubitably through the casting of certain consecrated nuts in divination.

70By way of conclusion, I want to defend the counter-intuitive character of our analysis from a strong objection, namely that it results from an overly exotic account of Ifá initiation. There is sense in this complaint. As set out, the whole argument about internal relations depends on a detailed elucidation of the « motile » logic of divinatory invention. Insofar as initiation ceremonies are indeed precipitated by the oracles and the motile obligations they induce, this may indeed make sense. But isn’t this an idealised image of what actually goes on when a practitioner decides to undergo initiation? « Decision » is used advisedly here because from the practitioner’s perspective – as opposed to that of the oracle, on which we have focused – that is exactly what initiation represents. While the oracle may present initiation as the inevitable outcome of Orula’s « call », practitioners’ anxiety on the matter shows that there is more to it than that. The mixed feelings with which potential neophytes view the prospect of hugely expensive ceremonies indicate that from their point of view the progression from « calling » to consummation is hardly ineluctable. Indeed, if practitioners had no choice in the matter, why would babalawos raise the spectre of divine wrath as a sanction against « falling into disobedience »? Fear and anxiety, it seems, are methods for keeping potential neophytes on the straight and narrow (« path »), and this would indicate that, contrary to our analysis of motile obligation, things could have taken a different course.

  • 53 Holbraad, 2004.

71Furthermore, as we saw, they often do. Notwithstanding the obligations divinatory « callings » impose, there are plenty of men in Havana who at one time or other were told by diviners that they have a path of Ifá but nevertheless remain uninitiated, often because they can’t afford to pay for the ceremony, or even because they ignored the oracle’s prescription, as Javielito admitted doing for a while. Conversely, as I have described elsewhere, in Havana one sometimes meets babalawos whose initiation seems to have been owed more to personal choice than to divinatory obligation53. For example, in the late 1990s, when Ifá was enjoying its first period of effervescence following years of suppression by the state authorities, some men treated Ifá initiation as a life-style choice – an element of their « street cred », so to speak, along with owning a motor-bike or designer clothes. While they would hardly admit this in the presence of more senior babalawos, who would criticise such trends as a « lack of respect » (falta de respeto), these new initiates were rarely inclined to invest much time in learning how to practice as babalawos, so that for many of them initiation hardly marked a transformation at all – let alone an ontological one.

  • 54 Note that these normative sanctions are incorporated in the logic of Ifá in an entirely circular wa (...)

72Nevertheless, while such trends do form an important part of the picture of Ifá practice in Havana, the argument of this article is that they illustrate the logic of Ifá only by contradicting it. This is ethnographically clear, insofar as individuals’ putative « disobediences » and « disrespects » are subject to criticism, are expected to attract divine sanctions, and are therefore a source of embarrassment and fear for those who commit them. But there is also a principled reason for resisting the charge that our analysis relies on an overly exotic account of initiation, focusing as it does on the logical absolutes of the oracle at the expense of practitioners’ more ambivalent perspectives. For note that this contrast, between the logic of divinatory obligation and people’s capacity to contradict it, is strictly analogous to a contrast examined earlier in relation to the oracles’ claim to truth. As we saw, the fact that diviners’ truth-claims are stipulated as indubitable does not imply that they are never actually doubted. Like anthropologists who write on divination, practitioners too may wonder whether what oracles say is « really true ». However, for babalawos who take such pains normatively to emphasise that the oracle is unfalsifiable this is just a misunderstanding : to doubt whether an oracular pronouncement is true is to doubt whether it is an oracular pronouncement. The same, I would argue, holds for the obligations that the motile logic of the oracles induces. It is perfectly possible to view initiation as a matter of choice – a discrete possibility of action to be weighed against alternatives, or even just ignored. However, such reversals of the priority of obligation over choice in Ifá amount to a form of logical « denial ». To weigh what one « has » to do against possible alternatives is to deny the special sense in which one has to do it (as one’s motile destiny), conventionalising obligation as a predicative « ought ». Or, putting it in the earlier formula, to treat initiation as a matter of choice is to treat it as if it were not an initiation. Again, the only way to bar this logical avenue is normative : enter notions of disobedience, disrespect, and the fear of Orula’s wrath54. Indeed, in view of how serious the matter of initiation is to the life of the cult, it is hardly surprising that the normative stakes, and the emotive responses they induce for practitioners, should rise in such a fashion.

73This brings us full circle, to the critical point that motivated the strategy of the present argument. We noted that sceptical practitioners who question the oracles’ truth are equivalent to anthropologists who wonder why the natives « believe in them ». An analogous equivalence, then, would bind those anthropologists who seek to explain why natives « commit » to initiatory transformation (making it sound rather as if the natives commit an error), with practitioners who disobediently compare initiation to alternative courses of action. Of course, one would hardly want to demand that anthropologists obey Orula. Still, babalawos may have a point in treating similar instances of disobedience and lack of respect as one big misunderstanding of what initiation is.



1 Houseman & Severi, 1998 : 278; Houseman, 2004.

2 Turner, 1969.

3 Bloch, 1974, 1992.

4 Whitehouse, 1996, 2000.

5 Latour, 2002.

6 Viveiros de Castro, 2003; Holbraad, 2003.

7 Holbraad, 2007.

8 Bascom, 1991 : 84-85; Akinnaso, 1992. Nevertheless, I suspect that oracles may have been (or still are) more important in regulating matters of initiation among the Yoruba than has hitherto been discussed to my knowledge. Such hints do at least exist in the literature (e.g. for a discussion of oracles’ role in individuals’ devotion to particular deities, cf. Bascom, 1969 : 77-78), and, in view of Abimbola’s account of the pervasive use of divination in Yoruba Ifa (1997), it would be surprising if this did not extend also to matters of cult recruitment.

9 Henceforth when I write of « Ifá » I mean Cuban Ifá. In spelling ritual terms of Yoruba origin I follow Cuban rather than contemporary Yoruba conventions; or better, I follow a Cuban convention, since among practitioners spellings have not been formulated consistently. For example, Cubans often distinguish « heavy » consonants of Yoruba-derived terms by spelling them either quasi-phonetically (e.g.Yoruba-derived od as in ordu) or by double consonants (e.g.oddu). I follow the latter convention, which is simpler.

10 Menéndes, 1995 : 51; Akinnaso, 1992 : 92.

11 For a detailed discussion of the historical evolution of this ceremonial sequence in Cuba cf. Brown, 2003.

12 The « warriors » include Elegguá (messenger god and « owner of the paths »), Ogún (patron of smithery and war), and Ochosi (god of hunting), who are given together with Osun, considered a symbol of the neophyte’s personal well being.

13 Holbraad, 2005.

14 Brown, 2003.

15 Menéndez Vázquez, 1995.

16 Bascom, 1941; Holbraad, 2007.

17 Van Gennep, 1960.

18 Houseman & Severi, 1998 : 223-285; Houseman, 2005.

19 Strathern, 1992, 2005; Latour, 2005; Law, 2004; Viveiros de Castro, 2002.

20 Moore, 1922; Ollman, 1975.

21 Cf. Wedel, 2004 : 100-105 for santería initiation, and Goldman, 1985 for a remarkable analysis of the ritual constitution of personhood in Brazilian candomblé.

22 The number 16 reigns supreme in Ifá divination, since the system of binary configurations on which it is based yields sixteen principal combinations, the so-called signos or oddu (see below). As illustrated by the majestic numerological speculations of the Cuban mystical scholar Heres Hevia, 16 does indeed have metaphysical potential (Heres Hevia, 1962). Considered as the first iteration of squaring on the nutural number series (2, 4, 16…), one might say that 16 is the mystical number par excellence in that it tends most economically to render the discrete Many as a Continuum One – 16 exemplifies circularity in number, so to speak, and is in this sense « magical ».

23 Holbraad, 2007; Cabrera, 2000 : 99, 103.

24 See Holbraad, 2003. To compare with Yoruba cases see Abimbola, 1997; Akinnaso, 1992; Barber, 1991.

25 The permanence of one’s « sign of itá » (signo de itá) is most clearly manifested by the fact that, after having been determined in the itá ceremony, this sign becomes the name by which the initiate is referred to in subsequent rituals. This is particularly so in full Ifá initiation, since, both in formal ritual and in more informal contexts, babalawos will refer to each other by their signo, often dispensing with Christian names altogether.

26 For a synthesis of the classical debates, cf. Robinson, 1950.

27 I adopt the standard distinction in the philosophical literature between definiendum and definiens (e.g. Robinson, 1950). The former is the object of definition, while the latter is its means.

28 Heidegger, 1968; Deleuze, 1994; Nietzsche, 1997.

29 Wagner, 1981; Ardener, 1989; Strathern, 1999; Latour, 1999; Viveiros de Castro, 2003.

30 Wagner, 1977.

31 It may be noted that in order to make this argument (identifying the predication of novel meanings as a « contradiction ») I have effectively conventionalised the definition of inventive definition. After all, a contradiction is just a joint statement of truth and falsehood : if it is true that inaugurations of meanings (viz inventive definitions) are predicative truth-claims then it is false that they are inaugurations. « Inauguration » is thus treated as a predicate of inventive definitions, in order to be treated truth-functionally for the sake of the argument. This, however, does not show that inventive definitions are, willy-nilly, predicative truth-claims. Rather it demonstrates that, for purposes of logical argument – with its reliance on truth-functional criteria such as contradiction –, the meanings that are brought together in inventive definition can be disaggregated as predicates and treated as conventions – moving, as Wagner would put it, from figure to ground (Wagner, 1972 : 6-7, 1987).

32 Holbraad, 2003.

33 Evans-Pritchard, 1976.

34 This would be obvious if one took the statements that in Ifá there are no mistakes or lies as attempts to inventively define what kind of thing verdicts are, rather than as wide-eyed attempts to characterise a miraculous feature of the verdicts by assuming an analogy with ordinary (and fallible) truth-claims. But since, following Evans-Pritchard, existing accounts of divination do assume that divinatory verdicts are just a special case of ordinary truth-claims, it is necessary to press the case home (Bascom, 1941; Park, 1963; Fortes, 1966; Turner, 1975; Sperber, 1985; Zeitlyn, 1990, 1995; Boyer, 1990, 1994).

35 Holbraad, 2004.

36 Holbraad, 2003 : 52-53; Werbner, 1973.

37 See note 33 above.

38 Viveiros de Castro, 2002 : 134.

39 Latour, 1999.

40 The Cuban authorities’ tolerance of the cults in a socialist context is a relatively recent phenomenon (see Argüelles & Hodge, 1991 : 141-171; Bolívar, 1997 : 165; Pedraza, 1998 : 24-30; Hagedorn, 2001, Holbraad, 2004)

41 There is no need to reveal this personal information.

42 « Washing » a deity (lavar el santo) is a cheap alternative to full Santería initiation (see table 1).

43 Holbraad, 2004.

44 Hagendorn, 2001.

45 For a well-rounded though technical exposition of the logic of obligation, cf. Horty, 2001.

46 The fact that for purposes of this minimal example the fusion has had to be characterised verbally in terms of the relation « …is after » should not be taken as evidence that relations are ontologically different from the trajectories they relate. Rather, the verbal characterisation is shorthand for the fact that many more trajectories are involved in the fusion, including the one of the concept-cum-thing « after » – which, of course, in this context is a synonym of « persecution ».

47 Bateson, 2002 : 103.

48 Viveiros de Castro, 1998 : 15-17.

49 Geertz, 1993 : 28-29.

50 For a classic argument on the analogy between temporality and modality (i. e. between different times and different possible worlds) see Lewis, 1986 : 202-204 and passim.

51 Holbraad, 2003 : 65-66.

52 In contrast to well discussed cases in West African ethnography, notions of « destiny » or « fate » did not show up as a salient part of cosmology or practice in my research on Cuban Ifá. Not only are references to such ideas largely absent in babalawos’ discourse, but also in Cuban Ifá and Santería there is no adoration of orí (the head) as such, which the Yoruba consider the locus of individual destiny (referred to as iwa, cf. Bascom, 1969 : 71-74; Lawal, 1985; Fortes 1981).

53 Holbraad, 2004.

54 Note that these normative sanctions are incorporated in the logic of Ifá in an entirely circular way, since both disobedience and the anger of the deities that it provokes are diagnosed primarily by means of divinatory consultations.



Titel Table 1 - Ceremonies undertaken on initiatory « path » of Ifá
Datei image/png, 193k
Titel Figure 1 - Ritual kinship relations in « religious family »
Datei image/png, 45k
Titel Figure 2 - Ritual lineages and the « births » of Orula idol-deities
Datei image/png, 43k
Titel Figure 3 - The logical priority of initiate : deity relationships over initiate : initiate relationships
Datei image/png, 30k
Titel Figure 4a - « Motile » transformation
Datei image/png, 35k
Titel Figure 4b - Aristotelian predicative change
Datei image/png, 25k
Titel Figure 5 - Initiatory transformation in Ifá
Datei image/png, 42k



Martin Holbraad, « Relationships in motion », Systèmes de pensée en Afrique noire, 18 | 2008, 219-264.


Martin Holbraad, « Relationships in motion », Systèmes de pensée en Afrique noire [Online], 18 | 2008, Online erschienen am: 05 Juni 2013, abgerufen am 26 Mai 2017. URL : ; DOI : 10.4000/span.817



Martin Holbraad

University College London



© École pratique des hautes études

  • Logo EPHE
  • Logo CEMAF