Holbraad: Truth beyond doubt

This work is licensed under the Creative Commons | © Martin Holbraad. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 (Online)

Truth beyond doubt:

Ifá oracles in Havana

Martin Holbraad, University College London

This article analyzes the concept of truth on which the practice of Ifá divination in Cuba turns. Motivated ethnographically by Ifá practitioners’ claims that the truths their oracles issue are indubitable, I argue that from the viewpoint of commonplace conceptions of truth such an assumption can only be interpreted as absurd. To avoid such an imputation, the article is devoted to reconceptualizing what might count as truth in such an ethnographic instance. In particular, it is argued that in order to credit the assumption of divinatory indubitability, representational notions of truth must be discarded in favor of what I call a “motile” conceptualization, which posits truth as an event in which trajectories of divinatory meanings (called “paths” by diviners) collide. In advancing such an analysis, the article exemplifies what I call an “ontographic” approach, dedicated to mapping the ontological premises of native discourse through the production of concepts which, while not the native concepts themselves, comprise their close equivalents. Elaborated in greater detail elsewhere (Holbraad 2012), this is put forward as my take on what the editors of HAU call “ethnographic theory.”

Keywords: divination, truth, Cuba, Ifá, ontography, motility

When I tell people in Europe that I went to Havana to work with certain cult practitioners who think that oracles tell them the “truth” about things, I am invariably called to answer for myself: “do you think that oracles work?”1 I both love and hate this question. One reason that I appreciate it, particularly when it comes from the kind of self-avowed rationalists that one often meets in academic life (even in anthropology departments!) is that, with its mixture of intrigue and incredulity, the question reminds me that anthropology does have something to say, even to rationalists. For a testing moment, I—as a metonym for what I study-become as fascinating to my rationalist friends as what I study is to me—the anthropologist. And, as an anthropologist, I am in good company, since Frazer and Tylor, by trying to explain why savages might be credulous enough to think that things like oracles might work, were quelling the same kind of disquiet in the Victorian psyche as I am expected to when talking to my rationalist colleagues. Of course, such rationalists today don’t bat an eyelid when told that people in the Caribbean trust oracles—and for this dubious merit they may as well thank the anthropologists. But the enduring question—yes, but do you trust oracles?—shows that the grounds for their disquiet are still there. Now, it would be great if I could respond to this question by lending people a book—if not by Frazer or Tylor, one by another, perhaps more recent, anthropologist—but as it happens, I can’t. That is to say, in my view anthropologists have not yet arrived at an adequately reassuring analysis of oracular truth. This claim, however (which I shall only defend with reference to one example2), does not explain why I find the challenge as to my beliefs uncomfortable. What I hate about it is that it tends to put me in a double bind. If I say that I do not believe in oracles, I provide a quick fix for my interlocutor’s disquiet with what is really a lie. For there is an important sense in which I do believe in oracles—but if I admit this, I create the conditions for misunderstanding, since the sense in which I trust oracles is very different from the more sensational one in which they are more than likely interested. This paper aims to provide the kind of clarification required to avoid misunderstanding on this score.

The main reason I frame discussion of Ifá oracles in Havana with thoughts drawn from the familiar is not rhetorical. The point can be made with reference to an argument presented by Eduardo Viveiros de Castro, as part of a critique of what he calls the “classical anthropological solution” to the problem of how to take seriously such astonishing claims as “peccaries are human”—his favorite Amerindian example (2002). With Durkheim, Lévi-Strauss and Sperber in mind, among others, he argues that varieties of the “classical solution” turn on a common assumption—namely, that if natives are to be taken seriously when they say or do things that the anthropologist finds unreasonable, this must be in spite of what they say or do. Unable to admit that peccaries might be human, anthropologists feel they are left with no alternative but to enunciate the conditions under which the natives could entertain such outlandish notions. The crucial assumption here, Viveiros de Castro notes, is that when anthropologists think (to themselves), “peccaries, clearly, are not human” they think in terms of the same concepts as natives do when they say that peccaries are human. Otherwise anthropologists would be in no position to judge the truth or error of native statements. Or, to put the point in philosophical terms, the assumption is that the statements “peccaries are not human” (as anthropologists know) and “peccaries are human” (as natives claim) are cast in terms that have identical “intensions” (or thereabouts), and the opposition between the two is truth-functional (viz. true v. untrue3 respectively) because the role of each statement is semantically to fix the “extension” of the terms involved4 (cf. Viveiros de Castro 2002: 134). And indeed, under such an interpretation, the natives’ “error” would be palpable. While the concepts of “peccary” and “human” are as distinct for them as they are for us, natives insist on misapplying those concepts (i.e. getting their extension wrong) by bizarrely conflating the two classes in statements like “peccaries are human,” statements which here are understood as stating empirical as opposed to conceptual claims. The job of anthropological analysis, then, is to explain why natives might get their own concepts wrong, as it were; which is to say that the anthropological “problem” is epistemological through and through.

As this attempt at interpretation would indicate, the “classical” approach is not untenable in its own terms. It is, however, implausible. Note the double miracle implicit in this position. One would certainly need a reason to expect people, who are in so many ways as different from us as Amerindians, to nevertheless share our concepts—concepts as socially important as “human,” or indeed as peculiarly theirs as “peccary” (and by “us”—an offensive word to some!—I mean simply us anthropological analysts). Equally, one would need a reason to explain why they seem to get the empirical implications of these concepts wrong so systematically. After all, as Viveiros de Castro reminds us, peccaries are not just human; “they walk in gangs . . . have a chief . . . are noisy and aggressive . . . and so on” (2002: 136). So, far from incidental, Amerindian “errors” must be serially—and seriously—compounded. Now, we know that reasons which would explain both miracles away are forthcoming—cognitive anthropologists’ attempts to define a species-wide conceptual baseline, for example, or the old Popperian argument about the “closed” character of mystical “belief systems” (cf. Horton 1967). But the point is that, regardless of their merits, in the present context such theoretical moves have a semblance of beating the sledgehammer with the nail, considering how unlikely their explananda really are.

Admittedly, if there were no alternative to the “classical solution” one would have to settle for miracles—or at least the disenchanting theories that they engender. But, as Viveiros de Castro shows, a suitably radical alternative does exist, and it runs as follows. What if we stipulated that the analyst’s puzzlement in the face of surprising native statements is not caused by an epistemological disagreement over the correct empirical application of shared concepts (in other words, a difference of opinion), but rather, by the radical alterity of the concepts involved? So if, in line with the present interpretation, the “classical” position stems from the idea that native terms like “peccary” and “human” have the same intension for the natives as they do for the analyst, then Viveiros de Castro’s proposed alternative is a straight negation: the terms involved have different intensions for the analyst—and this is precisely why native statements are genuinely bizarre to him.

Viveiros de Castro offers a number of arguments in favor of this conceptual reversal, but here I want to focus on two main advantages, before going on to discuss critically some implications for my own strategy in this paper. First, in denying the first “miracle” of the classical approach—that somehow the natives’ concepts must be the same as ours—the move to intensional alterity also dispels the second one—that natives consistently misapply their own concepts. For, once the possibility that native concepts might be different from our own is opened up, then native statements like “peccaries are human” need no longer be viewed as attempts to “apply”—that is, determine the extension of—predefined terms (“peccaries” and “human”). Rather, they can be recognized as attempts on the part of the natives to express the meaning of their own concepts—in other words, to define them intensionally (cf. Wagner 1972: 5-8). Extensional “applications,” then, are not the issue here, and hence the possibility of native “error” does not even come into play. Indeed, since statements like “peccaries are human” are meant to define what counts as “peccary” (and, by this “bizarre” definition, also “human”), they ought to be understood as ontological statements, on a par with, say, Descartes’ definition of “I” as res cogtans.

The second advantage of Viveiros de Castro’s reversal is that it promises a much more fecund analytical agenda than its “classical” mirror image. In taking the sense of native statements out of the brackets, as it were, and instituting it as a prime object of anthropological analysis, Viveiros de Castro is effectively proposing a conceptual field that is fresh by definition. After all, consider what the job of analysis must amount to on this view. Rather than enunciating the conditions of native error (be they epistemic, cognitive, sociological, political, or whatever), the analytical task now becomes one of elucidating new concepts—again, “new” by definition. Indeed, note that this project is necessarily quite different from the notion, familiar from so-called “relativist” approaches, of “cultural translation” or “emic description” (see also Holbraad 2012: 34-53). For ideas of translation or description turn on the assumption that unfamiliar concepts must, to some satisfactory degree, have familiar equivalents, as if “their” conceptual repertoire must ultimately be revealed as isomorphic to “ours”—a third miracle: verstehen. And this, of course, is precisely the opposite of what Viveiros de Castro is arguing.

Perhaps the best way to think of the kind of analysis Viveiros de Castro proposes is as an ethnographically guided version of what some philosophers often see as their own trade, namely “conceptual analysis.” Certainly this is the direction that he himself seems to be pointing at when he defines anthropology as a species of thought-experimental “fiction.” He writes:

What does such a fiction consist in? It consists in taking indigenous ideas as concepts, and following the consequences of such a decision: to determine the preconceptual ground or plane of immanence that such concepts presuppose, the conceptual personae that they deploy, and the material realities that they create (2002: 123, my translation).

The prospect sounds enticing, but I wonder if here Viveiros de Castro comes close to invoking his own miracle ex machina. For, in view of the guiding premise of his approach, what he is effectively suggesting is a process of arriving at ontological presuppositions and/or consequences of concepts that are still in want of definition. Indeed we should treat indigenous ideas as concepts, but, as already seen, the point of doing so is to render explicit—for us analysts—their intensional alterity. But the kind of ontological inferences that Viveiros de Castro seems to be envisaging cannot be made from a position of such aporia. Indeed, this is the point at which the analogy between anthropological analysis and philosophy breaks down. As a purely autochthonous intellectual exercise, philosophy has the luxury of creating unfamiliar concepts by testing the limits of familiar ones. The “new,” in this case, may rest on the shoulders of the “old.” The anthropological challenge, by contrast, seems hyper-philosophical. What we are called upon to do is to create (our own) new concepts out of concepts, which to us are just as new (the natives’), which would seem tantamount to creation ex nihilo.

A fourth miracle perhaps? The argument of the present paper proceeds on the assumption that it is not. I would argue that the methodological tools for performing the kinds of thought-experiment that Viveiros de Castro has in mind can, in fact, be extrapolated from the terms of the contrast between native statements and our default assumptions. Consider the following strategy. Following Viveiros de Castro’s “rules of the game,” we must accept that, as anthropologists, we are in the dark: we begin with no sense of native concepts. We do, however, know two things. Firstly, we know the sense of our own default concepts (e.g., that peccaries are pig-like animals that live in South America). And secondly, we know that a symptom of the difference between our concepts and the natives’ is that, in a number of contexts (specifically, those where intensions differ), our own translations—or misunderstandings—of native comments appear as statements of falsehood.

Arguably, we have here the makings of a method that may allow us to approximate an understanding of native concepts and the strange statements that define them. For, like philosophers, one thing that we can do is transform the sense of our own concepts. So what if, through conceptual analysis, we were to alter the premises of our concepts (“peccary,” “human,” etc.) so as to transform them to such an extent that, when used to gloss native statements, they would yield statements of truth? The anthropological thought experiment would then proceed from the question: how do we need to change the intension of our own terms (by defining them differently than we ordinarily do) in order to make them behave, in truth-functional terms, like the natives’ concepts appear to? How far do we have to change our assumptions about what counts as a “peccary” before we could say that peccaries are human? To be sure, the promise here is not that of appropriating the native concepts themselves, but rather one of arriving at our own approximate equivalents—a truth-functional imitation of sorts. But this doesn’t make the project of analysis any less fecund. After all, the kind of Copernican revolution of concepts proposed here is by definition aimed at arriving at new concepts. For ease of reference, I call this method “ontographic,” as it offers a way of charting out the ontological premises of native discourse (Holbraad 2009, 2012).

So in this limited methodological sense, I would disagree with Viveiros de Castro’s unqualified contention that anthropologists’ truth-judgments are altogether irrelevant to analysis. “I am an anthropologist,” he writes, “not a swineologist. . . . An awareness of the falsehood of native statements . . . is only interesting in having awakened the anthropologist’s interest” (2002: 134-5). From then on, he argues, anthropologists should leave their own truth-judgments to one side, and focus on unearthing the ontological presuppositions that underlie the natives’ statements. In light of the above considerations, however, I would maintain that, logically speaking, there is no way of gauging the premises of native statements except in light of our own, and that such comparisons must thus ultimately be guided by truth-functional considerations (see Holbraad 2012).

This paper is designed to make explicit the analytical fecundity of this method. The task here will be to do an ontography of the concept of truth, as deployed in Cuban Ifá divination. In light of the comments just made on the role of truth-concepts in ontographic method itself, this may seem like a peculiarly recursive choice of topic. And indeed, an analysis of divinatory concepts of truth does present the prospect of pertinent comparisons with the kinds of truth-concepts anthropologists themselves might rely on in their own analytical strategies,5 including that of “ontography” perhaps. Nevertheless, while emphasizing that, by definition, an ontography of divinatory truth can indeed only proceed from a critique of truth-concepts that may be deemed ordinary in a general or “common” sense, I should make clear that the extra step of comparing divinatory truth to those truth-concepts that are at stake specifically in the context of anthropological analysis (ontographic or otherwise) cannot be made here (though see Holbraad 2012).

Starting off with an attempt to set the parameters of my approach with reference to a short critique of Pascal Boyer’s recent discussions of divination, my argument is set on track by presenting salient ethnographic facts regarding Ifá divination as I witnessed it in Cuba. Emphasizing practitioners’ contention that the oracle of Ifá is infallible, I make the analytical point that the oracle’s verdicts need to be understood as being the kinds of truths that are immune to doubt—in philosophical parlance, truths that are “indubitable.” A short and rather abstract second section is devoted to establishing that from the standpoint of ordinary understandings of what counts as truth, the indubitability of divinatory verdicts can only appear as dogmatic nonsense. For purposes of ontographic analysis, these are grounds for doing away with ordinary understandings of truth, and for constructing an alternative conceptualization that accords (extensively) with my informants’ convictions. In the final section this task is carried out by working back from the ethnography, using a nexus of concepts and practices associated with the oracle, in order to gauge the premises that guarantee truth and its emergence in Ifá divination. After all, what distinguishes ontography from arbitrary ontological speculation, is the method of extrapolating analytic abstractions from the ethnographic material, rather than, say, heaping Western philosophical concepts upon it.

The central idea, then, will be that divinatory truth-claims are beyond doubt because their truth conditions are not specified with reference to facts. Rather, divination exhibits what I call a nonrepresentational “motile” logic, which presupposes a notion of truth understood not as correspondence across an ontological divide (representation v. fact), but as proximal motion on a single ontological plane, that allows for revelatory events. If this sounds esoteric at this stage, that is because we have not yet seen the ethnography, to which I shall turn presently. Before doing so, however, it may be helpful to set the parameters of my own “nonrepresentationist” approach by contrasting it with a relatively recent attempt to account for divinatory truth offered by Boyer. For Boyer’s argument also turns on a negation of the role of “representation” in divinatory contexts, though in a way that is quite different from the one I shall be proposing—and the difference may be instructive.

Boyer on divinatory truth

Like his arguments on religious phenomena more generally (Boyer 1990, 1994, 2000, 2001), Boyer’s argument on divination turns on a cognitive premise, namely that explaining why people think what they do—in this case, why they think that oracles give truth—must ultimately be a matter of showing how their minds are able and likely to entertain the notions in question. This is because the ideas that anthropologists usually summarily describe as “culture” (such as ideas about divination) in fact boil down to highly complex aggregates of mental representations, distributed across human populations in accordance with constraints imposed by individual human brains—the instruments for mental representation par excellence (cf. Sperber 1996).

In line with this anti-culturalist premise, Boyer eschews the general (or, as he calls it, “epistemic”) question of why people believe in divination, and sets out to analyze the cognitive processes involved when a given individual represents an oracular pronouncement (henceforth “verdict”) as being true (Boyer 1994: 49-52). These processes, he argues, can be seen as a peculiar variant of the cognitive processes involved in ascribing truth to any ordinary representation, say in the course of conversation. The first point to note about the cognitive structure of truth-ascription is that it is “meta-representational,”—that is, it turns on the mind/brain’s ability to represent representations (ibid.: 243-245). Take the representations that are uttered when people communicate with each other—representing such ordinary utterances as “true” involves spontaneously representing what cognitive psychologists call an “evidential account.” This specifies two things: (1) that the representation that the speaker is expressing was caused by the events or states of affairs that his utterance describes (i.e., that his mental representation emanates from the way things really are); (2) that the utterance in question is expressing that representation, and not some other. So if, for example, you said to me “Boyer is a cognitivist,” my default presumption—that what you said is true—would be built on the spontaneous assumption that: (1) your mental representation BOYER IS A COGNITIVIST was, somehow, caused by Boyer actually being a cognitivist; and (2) that your utterance actually conveyed that mental representation. In other words, I believe what you say because I believe that you know what you are talking about, and that you are not lying about what you know. The evidential sequence, then, takes the following form:

[fact] causes → [MENTAL REPRÉSENTATION] expressed by → [utterance]

So much for everyday communication. What is interesting about divinatory proceedings, argues Boyer, is that they explicitly preclude the possibility of constructing the representational stage of the above sequence. What is so important about trance, “randomizing” elements (cowries, cracked scapulae, etc.), references to supernatural agencies, etc. in divinatory methods is that they are all means through which the diviner is himself effectively divested of responsibility for the verdict (Boyer 1994: 246). Hence, to use the famous example, when a Zande man consults an oracle, it is clear to him that it is the poison which is fed to the fowls that determines the verdict, not the oracle operator (Evans-Pritchard 1976: 146- 149). It follows that the truth of the verdict cannot be evaluated with reference to a correspondence between the verdict uttered and the mental representation of the diviner. The diviner cannot lie because, properly speaking, he does not speak at all. If suspicion does arise that the diviner’s mental representations are in fact deflecting the causal series that leads up to a given verdict, then the action simply does not count as divination (cf. Boyer 1994: 207). Bypassing mental representations, then, evidential accounts for divinatory verdicts correspond to immediate sequences:

[fact] causes → [verdict]

In Peircian terms, Boyer argues, genuine divination is assumed by practitioners to deal in indexical utterances—that is, utterances that are assumed to be caused by the states of affairs they express, as a smile is assumed to express goodwill (Boyer 1990: 72-75; cf. Rappaport 1979). This point is of central importance to Boyer since for him the indexical character of divinatory verdicts lies at the heart of the answer as to why practitioners tend to deem such verdicts to be true. The idea is that the causal nature of the connection between indices and the facts they describe increases the probability that practitioners will suppose that the verdict is true. This, claims Boyer, is because from a very early stage of human cognitive development causal relationships are represented as stable connections, so that a given effect tends to be conjoined spontaneously in the mind/brain of the observer with its supposed cause. Consequently, to the extent that divinatory technologies coerce practitioners into assuming that their outcomes are indexical, they also tend to force the assumption that they are true. Boyer’s logic on this last point is as pivotal to his argument as it is abstruse, so it is worth quoting his own words:

If a causal connection is assumed between two events or states “c” and “e,” a subsequent occurrence of “e” will lead the subject to assume that “c.” . . . Representing a connection as causal leads to conjecture that it may correspond to a stable pattern. . . . In metaphorical terms, the utterances (e.g., divinatory verdicts) are supposed to be true because they are construed as the stable symptoms or indices of the situations they describe. (Boyer 1994: 251)

But the point hardly follows. All that anti-representational divinatory techniques can do is force upon practitioners the assumption that, were the verdict to be true, it would be because it was caused immediately by the facts that it describes (i.e., it would be their index). In other words, Boyer’s pivotal appeal to the stability of causation is question-begging. Spelling out the correspondence between indices and stable causal connections motivates not the proposition that divinatory verdicts need to be taken as indices, but only the tautology that, were the verdicts to be so taken, they would be assumed to be true. To return to the Azande, the fact that when verdicts are taken as true they are assumed to be caused by—say—witchcraft, in no way explains why verdicts are taken as true in the first place. Indeed, in light of Evans-Pritchard’s famous point about the coexistence of divinatory and “common sense” explanations (see below), the question remains: why do Azande presume that the poison kills fowls because of witchcraft rather than because of its toxicity?

It may be objected that, rather than a weakness, this issue of under-determination is more of a virtue, since it allows Boyer’s cognitive argument to accommodate the fact that verdicts are often doubted, not only by exceptious analysts, but also by skeptical practitioners. Certainly in Cuba (and the point may well hold for most places) there are plenty of people who do not believe in the oracles at all (ideological Communists and converted Christians are most vehement in this respect). More intriguingly perhaps, a significant proportion of practitioners attend divinations in what may best be described as an agnostic or half-hearted spirit, explaining, for instance, that although they are interested in what the diviners have to say they are not “really” sure whether to believe them (cf. Bascom 1941). So, in light of these possible attitudes, the aim of analysis cannot be to render the truth of verdicts entirely foolproof, because, as skeptics know, it is not.

This is a point worth taking, but only because, in Viveiros de Castro’s terms, it makes explicit the “classical” tendencies of Boyer’s cognitive approach. For, while the divergence between the skeptic-cum-analyst’s and the (committed) practitioner’s views on divination are undeniable, it is not an ipso facto requirement that this divergence be interpreted as a disagreement over the truth-value of divinatory verdicts (“false” versus “true” respectively). In line with the considerations outlined above, an alternative would be to interpret the divergence as a difference in the intension of the concepts that each side—skeptic and practitioner—relies on. Much of the rest of this paper is devoted to showing that, with regard to divinatory truth, a preference for this alternative is dictated by the ethnography of divinatory practice. In this regard, I shall be defending the idea that the difference between the two conceptions of divinatory truth pertains to the question of doubt: while skeptics assume that oracles’ claims to truth are at least open to doubt (indeed they may often turn out to be false), practitioners make clear that genuine divinatory verdicts are so special precisely because they are in principle indubitable. So what is at issue here is really a botched conversation. Pointing ostensibly at the same referent (viz. divinatory verdicts), the skeptic imagines that practitioners simply attach a different truth-value to verdicts (true to his putative false), while the practitioner thinks that, in even entertaining the possibility that verdicts might be false (let alone asserting it), skeptics misunderstand the nature of divinatory claims to truth.

Boyer’s argument cannot encompass such a scenario. Indeed, it is probably accurate to say that, in terms of the contrast between doubt and indubitability, his cognitive approach is prejudiced in favor of the former—the skeptical position. On his cognitive analysis, practitioners’ evidential accounts allow for divinatory verdicts to be represented as true because they posit a direct causal link back to the state of affairs that the verdicts describe. As we saw, this leaves open the cognitive possibility of positing an alternative causal account—one that would connect the verdict not to the state of affairs it describes, but to a more mundane cause like the toxicity of poison. But this is tantamount to saying that, in principle, practitioners of divination are able to represent verdicts as being false, which is another way of saying that even for them verdicts are only doubtfully true. As we shall see, there is no way out of this problem as long it is assumed that the same concept of truth is at issue for practitioners and skeptics (and analysts) alike. In order to begin to approximate a suitably new concept of truth, we now turn to the ethnography of Ifá divination in Cuba.

The oracle of Ifá: a thumbnail sketch

Ifá is closely related to Santería, the most well-known Afro-Cuban religious tradition. Both have evolved primarily on the basis of elements brought to Cuba by Yoruba-speaking slaves from West Africa, primarily during the nineteenth century. 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 (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, 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 (Holbraad 2004). Nevertheless, Ifá has largely been practiced by “marginal” groups, as Cuban intellectuals often say, in predominantly non-white inner city neighborhoods of Havana, Matanzas, and Cardenas (my material comes mainly from Havana). Rights to worship in Ifá are distributed according to strictures associated with initiation (Holbraad 2008). Part of the reason why Ifá is so prestigious is that in order to be fully initiated—becoming a babalawo, with rights to participate in and dispense all aspects of worship—one has to undergo a series of initiatory ceremonies, part of the point of which is to ascertain whether Orula—the god of divination—will “call” the neophyte to the next step of initiation. Orula’s will on this matter—as in all—is expressed through the oracle of Ifá, so each “grade initiation” ceremony involves a long divinatory ceremony (called itá). Being called to “make oneself Ifá” by the oracle, as initiation is referred to (hacerse Ifá), involves agreeing with someone who is already initiated to preside over the ceremony as “godfather” (padrino). In this way neophytes are recruited to ritual “lineages,” which, from then on provide the primary context for worship and tutelage in the secret mythical and ritual knowledge that babalawos spend their lives “studying” (that is the word they use).

By far the most important element of initiation is the bestowal of the consecrated idol of Orula upon the neophyte by his godfather. In fact, it is probably more correct to refer to this as an “idol-deity” rather than just “idol,” since the consecrated paraphernalia that babalawos receive are not understood to “represent” the deity but rather, as practitioners emphasize, they are the deity (Bascom 1950; Holbraad 2007; cf. Palmié 2002: 166). Orula, then, basically consists of a clay pot, which, among other consecrated items, contains a bunch of twenty-one palm nuts (mano de Orula). These are the prime and most ceremonious means of divination, and the young babalawo will, from his initiation onwards, be able to use them to conduct divinations.

Although babalawos perform a range of ceremonial and magical services, the lynchpin of Ifá worship is divination, and it is chiefly in their capacity as diviners that babalawos are consulted by clients (see Holbraad 2005). In their general thrust, myths about the origins of Ifá divination recount how Orula was given the gift of interpreting Ifá as a way of bringing order into a then-chaotic universe. In myth Orula is presented as an arbitrator of both divine and human affairs, putting his divinatory powers to work for the benefit of all those who approach him for help, by revealing the will or “word” of Ifá. This archetypal role is one that babalawos are expected to fulfill both in regulating matters of worship within the cult (such as questions of initiation), and for the benefit of clients who, in exchange for a fee, visit the babalawos in order to clarify issues of personal concern regarding health, financial matters, love and sex or problems with the police.

The oracle of Ifá is built on a series of techniques which are designed to yield—in an apparently random manner—one out of the 256 possible configurations that the system allows. When it comes to the more ceremonious types of séance, in which the consecrated palm nuts are used (and on which I shall be focusing here), the babalawo achieves this by casting sixteen nuts, eight consecutive times, in a manner that is equivalent to tossing a coin eight times (hence the total of 256 possibilities is equal to two in the power of eight). 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 “through” the oracle, they also emphasize that each of the oddu is a divine being in its own right, (sometimes thought of as guises of Orula) which practitioners also sometimes refer to as his “paths” (caminos, see Holbraad 2007). Furthermore, each oddu has its own name, as well as its own sign (hence the Spanish term).

Each divination involves the casting of a large number of different oddu, according to a fixed order of questioning. The first cast is the most significant, however, since it determines what is referred to as the “principal” oddu of the session, which is taken as the basis for characterizing the personal circumstances of the consultant—be it a client, a neophyte, or otherwise, depending on the occasion. But before revealing the significance of the principal oddu, the babalawo casts a long series of further oddu which are designed to yield “yes” or “no” answers to specific questions through a complex algorithm. The first and most crucial of these is whether the consultant “is,” on this occasion, iré or osobbo (loosely, whether his or her circumstances are favorable or not). The Cuban folklorist Lydia Cabrera translates these Yoruba terms as “for good path” and “for bad path” respectively (Cabrera 1996: 192). Once the consultant’s state is ascertained, the babalawo proceeds to ask a fixed series of more specific questions, which determine the nature and causes of their state of iré or osobbo, as well as the appropriate ritual remedies and precautions. Once all questions are duly dealt with, the babalawo initiates the last and most artful phase of the divination, whereby the principal oddu, which up to now has passed without comment, is—as it is said—”spoken” (hablar el oddu). This idea of “speaking” the oddu stems from the fact that each of the 256 configurations is associated with a large number of myths which babalawos spend a lifetime memorizing. Each of these myths is colloquially referred to as a “path” of the oddu (camino del oddu). Depending on the extent of his own knowledge, the babalawo begins to recount one or more “paths” of the principal oddu, in order then to interpret it for the benefit of the consultant.

To give you a flavor of what this involves, let me present an extract from the transcript of a divination at which I was present. The occasion is an ordinary client consultation conducted for the benefit of a single mother in her mid-thirties by my godfather, Javier, who was seventy-seven years old at the time, having been initiated to Ifá in 1968. The principal oddu for this session was marked as being Obbeyono (one of the 256 divinatory “signs”), and through questioning it was determined that the woman was osobbo, at risk of illness due to sorcery. The remedy prescribed was a consecrated necklace dedicated to Babalú Ayé, the deity of disease, who is often identified “syncretically,” as they say, with St. Lazarus. In “speaking” the oddu, Javier recounted four “paths” of Obbeyono. This extract presents just the second one:

J: “Now, let me tell you, never mind your osobbo—San Lázaro will take care of that as long as you thank him. People like it when this ‘sign’ comes up, and it has been coming up a lot in these times. It speaks of a trip.”
C: [Laughter] “That’s what everyone wants!”
J: [Lugging on his cigarette] “Ifá says that on Lucumí land, in Africa, there was a territory that was owned by Ogún [the fearsome deity of smithery] and with his machete he’d cut down the people when they tried to enter. It so happened once that he felt someone was penetrating the border, so he took his machete and went to meet the intruder. But when he arrived he saw San Lázaro struggling with his crutches and took pity, so instead of attacking he got to work opening the cripple’s path across the land with his machete. . . . When people get this signo in their itá[the long divinatory séance conducted for neophytes as part of their initiation], we usually say that they are travelers. But in this case, Orula is just telling you of the possibility of a trip.”
C: “Yes, if only. But every time that things look up, something comes along and spoils my luck.”
J: “Of course, you are osobbo. Let’s see if Ogún wants something to open your path [casts etc.]. No. He says he doesn’t want anything. In any case, when you go home you should attend to your Ogún [referring to an idol-deity she had received years earlier], give him rum to drink, but not too much in case he gets drunk and then he can’t help you. Everyone wants to travel, no?”

It will be evident from this extract that interpretation is a crucial part of “speaking the oddu.” After all, the myth about Ogún and San Lázaro does not of itself speak to the woman’s travelling prospects. It is only because the babalawo knows that travel may well be among her concerns (since in Cuba, these days, travel is on everybody’s mind, as she herself also openly confirms), that the relevant verdict is located in this area. Indeed, for babalawos themselves, the yardstick of a good “orator of Ifá,” as they say, is above all his skill at bringing the myths to bear with precision on their clients’ personal circumstances (cf. Matibag 1997: 151-152). Javier illustrated this to me with a vivid account that merits full quotation:

[To consult] you need to know how to speak—to be an orator of Ifá—to manage the “metamorphosis,” as we call it. . . . You might come to me and from one story I can tell you three things. But you go to someone else and they might tell you ten, knowing how to get the most out of the oddu (sacarleprovecho). There was one guy ... who was famous when I was young. Once I was in a [séance] with him; he was arrogant but with good reason since he knew more than anyone else. . . . The other babalawos were speaking the oddu—I did too—but at some point he just stood up and said: ‘now listen to me!’ and, turning to the neophyte [curtly], “the fridge in your house is broken!” [The neophyte], bewildered, goes “yes, it is.” The babalawo turns to the rest: “Did you hear that?”—that was his way of teaching. We wondered how Ifá can speak of the guy’s fridge. . . . So he explained himself—I think the oddu was Obara Meyi: “Ifá says that there was an island where fishermen lived, but all their fish kept rotting. Close by there was another island which always had snow, so the fishermen brought snow from there to put their fish in.” And so with metamorphosis he says that in the house there must be a fridge, and since the neophyte had turned out osobbo, that it must be broken. Do you see how it works?

One could list here a variety of ways in which babalawos guard the prerogative to interpretation as a good and proper constituent of a successful divination. One could also elaborate on how clients too fully expect the diviners to bring interpretative skills to bear on their case, often availing the diviners of relevant information in order to help them “get to the point,” as one client put it to me. Though we cannot enter into this here, there is one point that pertains to anthropologists’ take on this common phenomenon, which we may call “interpretative openness.” In a classic version of what Viveiros de Castro calls the “classical solution” to the problem of belief, a number of anthropologists have sought to explain practitioners’ conviction that divinatory means deliver true verdicts by pointing to the subtle inter-subjective negotiations of meaning that divinatory interpretation so often involves (e.g., Bascom 1941; Bohannan 1975; Lévi-Strauss 1963; Sperber 1982; Parkin 1991; Zeitlyn 1990, 1995). The idea is that verdicts are best seen as blank slates upon which practitioners are able to draw up interpretations that they can reasonably represent as true. The diviner’s skill in achieving this air of plausibility (in good or bad faith) is therefore taken as pivotal in sustaining people’s trust in the putatively mysterious ability of divinatory means to deliver the truth.

Good manners towards one’s informants are not the only reason for resisting this “smarter than thou” stance vis-à-vis the practice of divination. To my mind, what the ethnography shows is that the practice of divination turns on a peculiarly quirky reversal of the premise of these approaches; the premise being that truth-ascription must come after interpretation since, logically speaking, if diviners and their clients are to decide whether their oracle tells them the truth, they must first understand what it is that the oracle is telling them. Quite to the contrary, I would maintain that what makes divinatory truth so special is the fact that practitioners put the cart before the horse in just this respect. From the practitioners’ point of view, what makes divinatory verdicts worth interpreting in the first place is the fact that they must be true.

I cannot establish this point with reference to all of the ethnography of divination, but the case of Ifá, at least, is a clear one. Discussing this point with babalawos and uninitiated clients alike, again and again I was given the following two maxim-like statements with an air of self-evidence: “Orula doesn’t make mistakes” (Orula no se equivoca), and “in Ifá there are no lies” (en Ifá no hay mentiras). As for the babalawos, well, “they are human beings,” as one of Javier’s godchildren put it to me, “and that means they’re imperfect.” Such comments do not only show the distortion involved in claiming that the specter of divine truth is constructed out of crafty interpretative projections. They also point to a conclusion that places the analysis of divinatory truth on a different footing altogether. Namely that, insofar as they are construed as genuine, Orula’s verdicts are not just taken to be true by practitioners, but rather are taken as statements that are true beyond doubt. For, by saying effectively that oracular statements which (for whatever reason) turn out to be mistaken or deceitful are not genuine verdicts, practitioners bar the logical possibility that verdicts might be false at all, which is just another way of stipulating that genuine oracular verdicts cannot but be true. This point about divination has been made before, not least by Evans-Pritchard, who famously branded as “secondary elaborations” the logical safety-nets with which the Azande were able to turn seemingly falsifying circumstances into confirmations of the infallibility of their poison oracle. The reason why the point bears restating is that, more like Viveiros de Castro than like Evans-Pritchard, I see practitioners’ conviction that their oracles are infallible, not as a consequence of the closed character of their epistemological presuppositions, but rather as a mark of ontological alterity regarding what kind of thing truth itself might be.

Indubitability and the discarded premise of representation

This possibility becomes clear when one considers how absurd practitioners’ views on infallibility seem when judged from the perspective of ordinary notions of truth. The problem is as follows. From the point of view of familiar ways of thinking about truth, there are indeed certain types of truth-claims to which one can accord the status of indubitability. For example, philosophers generally agree that truth-claims that are true “by definition” (i.e., the kinds of statements philosophers call “analytic”) are strong candidates for indubitability. Hence, to take the usual illustrations, it is impossible to doubt statements such as “all bachelors are unmarried” because their truth is entailed by the very meaning of the terms involved. Following the landmark argument by philosopher Saul Kripke (1980) about “a posteriori necessity,” one might also want to extend the category of indubitable truths to include statements such as “water is H2O,” that is, empirical statements that designate the “essential” properties of things—the properties without which things would not be what they are (water that is not H2O is not water). Similarly, depending on how far one would want to enter the exegetical debates among philosophers about the validity of Descartes’ infamous cogito argument, one might extend indubitability to include statements like “I think, therefore I am,” the truth of which is supposed to be entailed by their very utterance. Or, further, one could also mine indubitability in theological arguments for the existence of God (see for example Plantinga 1964, 1974).

It is easy, however, to see that divinatory truths cannot be construed as being indubitable in any of these familiar senses. To take our earlier example of the divinatory statement “your fridge is broken”—this hardly looks like an “analytical” statement, since there is nothing in the definition of “your fridge” as a concept that entails its being broken. Apart from anything else, if the state of repair of a refrigerator could per impossible be ascertained merely by virtue of the meaning of terms (by analogy to “bachelors are unmarried” and so on) then going to diviners to find such matters out would be quite redundant. Nor does “your fridge is broken” look anything like a Kripkean designation of essential properties. Unlike water and its chemical formula H20, presumably the fridge in question would still be itself even if it were not broken. And as for the Cartesian cogito and theological concerns with the nature of God, it is hard to see their relevance in this context at all—whatever it might be, a statement such as “your fridge is broken” is not an example of the cogito argument, much less a statement about God’s existence.

Indeed, if it is feasible to bypass philosophically-minded deliberations of this kind, that is because it seems obvious that divinatory statements are of a very familiar type, namely statements of fact—as “a posteriori,” “synthetic” and “contingent” as any ordinary statement such as “you are tired” or “the sun is shining.” Certainly in terms of its form, a claim like “your fridge is broken” is indistinguishable from such statements of fact. After all, just as with such ordinary statements, it would appear that the truth of the claim—divinatory or not—would depend on the facts of the matter, that is, it would depend on whether the fridge in question was indeed broken, as Javier’s own account also appears to illustrate. The problem, however, is that such matters are inherently doubtful. To render the divinatory verdict “your fridge is broken” dependent on whether the refrigerator is indeed broken is just to render its truth doubtful, since there is always the possibility that the fridge is fine. Treating such verdicts as statements of fact, in other words, takes us straight back to the skeptical position. From an “ontographic” point of view, then, there is only one thing for it: we need to identify and then discard those underlying assumptions that render ordinary conceptions of truth incompatible with the practitioners’ position on divination.

Let me keep this unavoidable digression into the philosophy of truth tolerably short by being very clear as to what it is we are looking for. We have just identified the possibility of doubt as the prime sticking point when it comes to appreciating the difference between divinatory notions of truth and ordinary ones. If we can find out, then, what it is about our ordinary concept of truth, which leads inevitably to the conclusion that oracular statements are open to doubt, we will have made an important step towards characterizing an alternative take on truth, which might accord with divinatory practice. In fact, a narrow focus on the concept of doubt leads us directly to a distinction that lies at the heart of ordinary assumptions about truth, namely, that between representations and facts. Indeed, one might say that the possibility of doubt can only arise in terms of this commonplace distinction. For one way to express the difference between representations and facts is to say that while facts are just actual, representations can equally well be about things that are not as they can be about things that are. Just because of this difference, it is natural to assume that representations rather than facts are the proper bearers of truth and falsehood. But if notions of truth and falsehood presuppose representations as bearers, then so does the concept of doubt, since it in turn stands or falls by the distinction between truth and falsehood. As already explained, something is doubtfully the case if it might have not been the case. But this negative possibility is a pure function of representation, since negations of facts can by definition only feature as representational contents—as what representations are about. In other words, the possibility that something may be false (upon which the possibility of doubt depends) can only arise representationally; there are no false facts. Analytically speaking, then, without a concept of representation we cannot have a concept of doubt.

But this conclusion alerts us to an intriguing—if counterintuitive—analytical possibility. Since in the effort to make sense of practitioners’ understanding of divinatory truth we find the possibility of doubt standing as a hurdle—and since the possibility of doubt depends on the idea that truth is a property of representations—might it not then be reasonable to question whether this latter assumption is appropriate when it comes to the analysis of divinatory truth? In fact, might it not be worth wondering whether an alternative conceptualization of truth—one which does not rely on the idea of representation at all—might serve as a better analytical frame for Ifá divination? I propose to pursue this possibility ethnographically in the final section of this paper.

Motion and divinatory truth

Given space constraints, my appeal to the ethnography will be more summary than it perhaps ought to. Nevertheless, it is expedient to take as starting point the notion of “paths,” which, as we have noted, features in a number of surprising ways in the way practitioners talk about the procedure of Ifá divination. As mentioned already, there are two ways in which the notion of “paths” features in what practitioners say about the oracle. First, as we saw, the 256 configurations that the palm nuts yield (the oddu are themselves sometimes referred to as the idol-deity Orula’s “paths.” In fact, we may note that this is just a specific case of a more general logic in Ifá and Santería, whereby each deity of the pantheon (Orula being just one of them) is understood to “have” a multitude of “paths,” each of which is known to have peculiar mythical and ritual characteristics. The second sense in which “paths” feature in practitioners’ discourse has to do with what Orula says during the divination, rather than how he appears. As we noted, both the options of good or bad fortune that get ascertained to start off with (on which I shall not be focusing here), and the individual myths that get spoken towards the end of the consultation (on which I focus), are thought of as “paths.”

When one asks babalawos to explain why deities and myths should be thought of as “paths,” the response one gets is more or less speculative, or indifferent in any case, which seems odd, since, with their knack for magisterial wisdom, babalawos can be rather prone to the idea that they have an answer for everything. This could be considered grounds for concluding that the concept of a “path” does not, in itself, have cosmological significance. Its import, I would argue, is “ontographic.” The fact that practitioners find the concept self-evidently appropriate for referring to such diverse data tells us less about what they think and more about how we should think about what they say. This would probably be to put too fine a point on my informants’ quaint manner of expression, were it not for the fact that these references to “paths” correspond neatly to two crucial ways in which motion is integral to the process of divination itself.

The first of these is the most obvious, and has to do with the mechanics of this form of divination. As noted, the business of casting oddu (Orula’s “paths”) involves a seemingly chaotic event through which a single configuration of nuts is determined. From a technological point of view, this poses the problem of how to infuse “chaos” into the set of 256 discrete possible configurations that the sixteen divining-nuts allow for. And the solution is found, of course, in motion. The boundaries that render the 256 configurations discrete dissolve indeterminately in a swift continuous movement as the babalawo shifts the divining-nuts from one hand to the other.

The second way in which motion enters the divinatory process might seem less straightforward, though my argument relies on the idea that ultimately it is not. What I have in mind here is the process of interpretation, through which babalawos”metamorphose” the mythical “paths” of Ifá, so as to provide a verdict that is relevant to the consultant’s personal circumstances. As we saw in the examples from my field-notes, divinatory interpretation involves a dialogical process, whereby myths that initially appear as rather general and opaque “stories” are gradually brought to bear on the immediate circumstances of the consultant. This process comes down to the babalawo’s skill at transforming (or “metamorphosing”) the myth so as to render it specific enough to be taken as a message that “gets to the point,” as my informant put it. In other words, the ability interpretatively to arrive at a true verdict (“your fridge is broken,” or whatever) is premised on the capacity that meaningful data—such as myths—have to change themselves, or—if you like—to move. If this sounds vaguely metaphorical it is only because the entrenched habit of representationism involves a predisposition toward imagining that (in their pristine state at least) meanings must correspond to something discrete and stable—what analytical philosophers who like to get their heads around things call “propositions.” But even that phrase—getting one’s head “around” a concept—betrays artifice here. Consider what is happening now, as you read. A stream of meaningful data is gushing forth out of this paper—just like a continuous current of sound would come out of my mouth were I to read this to you—and is apprehended by you as some kind of morphing beast, which can only be tamed—or, as we say, “pinned down”—through a certain amount of exertion. Only once you “get the point” (and only if there is one to be gotten) does meaning begin to acquire the semblance of stasis. At the first instance, then, meaning moves—quite literally so. And note also that such a “motile” view of meaning is thoroughly at odds with the ordinary idea that meanings are “representations” (which might “match” or “reflect” facts of the world), since such a match would presuppose that the meanings in question are already constituted as “propositions about the world,” which is just another way of imagining meaning in a state of rest.

I want to argue that the central role of the notion of transformation through motion in Ifá can be seen as the prime constituent of what I call a “motile” logic, which in turn has profound implications for our conceptualization of divinatory truth and indubitability. To give you an idea of what I mean by “motile logic,” it is useful to return once again to the work of Evans-Pritchard, this time to his famous distinction between “how” and “why” questions, which may be taken as an exemplary—if unassuming—case of “ontographic” analysis. The Zande brewer’s hut burns down. He consults the oracle to find out what happened, and is told that witchcraft is at play. This explanation, says Evans-Pritchard, is not meant to replace or even compete with a commonsense account in terms of the causal sequence of events that led to the fire, which Azande are as capable of producing as anyone else. For such an account can only explain how the fire came about, whereas what interests the brewer when he goes to the oracle is why this misfortune should affect him in particular and on this particular occasion. In other words, while commonsense explanations tell causal stories, divinatory verdicts relate events to personal histories. One may say, then, that divination picks up where “commonsense” leaves off. After all is said and done (the fire in the hut having been accounted for in the most exhaustive and hard-nosed manner possible), the owner is still left with the question: “Why me? Why now?” The only commonsense answer is a non-answer: “coincidence.”

Now, the substantive distinction between “how” and “why” is not that important here, not least because Ifá oracles are called upon to answer all sorts of questions that have nothing to do with misfortune. In what sense, for instance, should questions regarding the timing of a ceremony be thought of as “why” rather than “how” questions? Much more significant for the argument at hand are two key insights that underlie Evans-Pritchard’s distinction. First, there is mileage to be had out of the notion that divinatory truth-claims relate things, for example, events to personal histories. For the difference between “why” and “how” ultimately turns on a distinction between two orders of relation. “How” as we saw, is cast in causal terms, by linking events linearly in logical sequences, consequent to antecedent: “this happened because that happened . . . ,” etc. Such links may be called relations ofconjunction (see Figure 1). “Why” questions, on the other hand, seem to pertain to something like a hidden dimension, squeezed laterally “in between” linear conjunctions: when all causal chains are said and done, as tightly as can be, there is still space enough to ask “why” as an extra question. This “extra” quality is just the product of the logical shift involved in relating causal chains to data that lie outside of them—outside by definition, since positing further causal links would keep analysis at the level of “how.”

Figure 1
Figure 1: conjunction (causal links).

But what kind of relation could that be? A clue, I think, lies in Evans-Pritchard’s second insight, namely that common sense tends to brush off diviners’ “why-questions” as matters of “coincidence.” Notwithstanding its normative vacuity, the concept is graphically dynamic: we call “coincidences” those events that constitute a singular outcome of two or more unrelated causal series. (I walk into a bar and find you there “by coincidence,” if the events that brought me there were causally independent of the events that brought you.) This may seem like a negative way of characterizing “why-relations,” since the obvious distinguishing mark of coincidence (as opposed to conjunction) is that it is non-causal. But a more abstract analysis reveals the positive, dynamic facets of coincidence. First, coincidence involves interaction: coincidental relations do not in themselves pan out as ordered series, but are rather constituted at the intersection of two or more causal chains (or their members), as illustrated in Figure 2. Second, the points of intersection that constitute coincidental relations correspond to dynamic events, since they represent meeting-points of series that are in motion. This follows from the fact that causal chains themselves comprise events (that is, alterations over time) so that their meetings properly constitute temporary collisions of trajectories. One may say, then, that coincidences are best glossed oxymoronically as non-causal interactions.

Figure 2
Figure 2: coincidence (non-causal interactions)

It will be clear that these abstract considerations allow for an analysis that goes beyond the distinction between “how” and “why.” For the difference between conjunction and coincidence is not one of meaning or content (cast in terms of distinct categories of questions), but rather a purely formal antithesis. If “common sense” works on identifying the conjunctions that link events to their causes, divination works laterally at establishing collision points between causally independent trajectories of events. Indeed, note here that the distinction can also be put in terms of an opposition between rest and motion. As we have defined it, the difference between causal conjunctions and non-causal interactions comes down to the difference between giving logical priority to series of isolated—or at least distinct, and in that sense “stable”—events, and starting with continuous trajectories of motion. From this point of view common sense and divination are diametrically opposed: while the former is given “events” as determinable points and then must work at linking these points in an implicitly temporal order to form causal “chains,” the latter is given motion as the raw material so that its job becomes one of arriving at “events,” which in this case are constituted as temporary definitions at the vertices of motion. The concept of “motile logic” refers precisely to this ontological reversal, which posits motion as primordial, and stable entities as derivative outcomes.

Given the ethnography already presented, perhaps the point of the present analysis begins to clarify itself. For in Ifá both the process of “metamorphosis” in “speaking the oddu” by which verdicts are arrived at, and the technical procedure by which the oddu itself are cast, lend themselves to an analysis in terms of coincidental relations between trajectories of motion, or “paths”—to use the indigenous term. Let me start with the case of mythical interpretation. Two relevant points have already been made. First, that interpretation is a dialogical process whereby myths are brought to bear on the consultant’s circumstances. Second, that this process is transformative, and that therefore the meaning of the myths is best construed as being in motion. To this we should add that, inasmuch as the consultant’s personal circumstances also feature in the process of interpretation as meaningful data that are made to interact with the meaning of the myth, their meaning must also be thought of in motile terms. So, to return to the earlier extract, the mythical “path” that tells the story of San Lázaro’s meeting with Ogún need not be construed as fundamentally different to, say, the consultant’s personal frustrations in her attempts to travel. Both data refer to events or states of affairs that are meaningful, and both can be thought about, narrated and transformed in motion. Effectively what we have, then, are two strands of meaning which initially seem unrelated, and the diviner’s job is to make the two “meet” so as to produce a verdict that “goes to the point.” So it should be clear that what we have here is a relation of coincidence (see Figure 3).

Figure 3
Figure 3: coincidence in divinatory interpretation

In view of Figure 3, we may make a couple of points regarding the motile emergence of the oracle’s verdicts, which will wrap up the argument on truth. The first point is that, according to this model, verdicts emerge as singular events. Now, this may sound like a point about epistemology rather than ontology—since it is cast as an answer to the question of how divinatory truth-claims emerge, rather than what kind of things they are. But this would be a misunderstanding. As we have emphasized all along, the process of divination itself, as well as what practitioners have to say about it, leaves no doubt as to the fact that, in Ifá divination, interpretation is constitutive to the definition of divinatory truth. Hence, if the process of interpretation is premised on the motility of meaning, then motion is also the ontological ground of divinatory truth as such. So, and this is really the crux of the present argument, truth must in this case be defined precisely as the event that results out of the meeting of causally independent trajectories of meaning, which is just the kind of meeting that diviners are able to induce through interpretative metamorphosis.

The second point to note is that this definition of truth has consequences with regard to the question of indubitability, from which this whole discussion took off. For, having defined truth as a kind of event, we may then ask how such a definition fares with respect to the notion of doubt. On the ordinary definition of truth, we saw, doubt pertains to the possibility that a representation may fail to match a fact. What, then, would be the equivalent possibility on a motile definition of truth? Looking at Figure 3 one might be tempted to offer a reply. If, on this motile image, the truth of divinatory verdicts is defined as a meeting (or collision) of causally independent trajectories of meaning, then to cast “doubt” on such truths would be to raise the possibility that the trajectories in question might not have met as they did. The trajectories might have taken a different course, and hence could have intersected at a different point, or not at all.

However, I would argue that there is a perspective from which this is a slide “from the model of reality to the reality of the model,” to paraphrase Bourdieu (1990: 39). And that is the perspective of motion itself. Consider what it is about motion that gets reduced or taken for granted for the purposes of graphic representation. In order to indicate movement on paper all that is needed is a line to show its trajectory, with a little arrow on its end indicating its direction: since only movements have directions, this is enough. But why is a line appropriate for representing a trajectory? The answer, clearly, is that trajectories are necessarily continuous, and this is because movements have a momentum, an “intraneous” power that “keeps them going.” Now, if you think about it, the continuity of plotted trajectories is only a very faint way of expressing momentum. This is hardly surprising since tota simul representations on paper have to be “economical”: they do not move in themselves, and hence they cannot really have a momentum. But economy comes at a price. For the point about momentum is not only that it renders motion both continuous and directional, but also that it does so as a matter of necessity: momentum describes the inner compulsion of motion. The best way to understand this, I think, is cinematic: imagine panning away from the bird’s-eye perspective of diagrams, and placing the “camera” at the helm of a moving trajectory, cockpit-style. What you see now is hardly a contingent matter, since your immanent purview is dictated at every moment by the propulsion of the trajectory itself. What previously seemed like one possible course among many now seems like the only possible course, for the momentum of motion—its propulsion in one direction—carries you with it. With momentum, one might say, motion entails its own necessity. So viewed from the point of view of motion (which, as we have seen, is constitutive to the definition divinatory verdicts), truth-claims emerge as events that cannot but be what they are. To doubt them—or in other words to posit the possibility that they could have turned out differently—is just to deny their motility.

It needs to be emphasized here that the foregoing constitutes a radical departure from the “representationist” take on truth, for the truth that it defines is not the one we are used to. I venture to call the truth in question “revelatory.” For at issue here is not the veracity of the way things are thought about or represented, but rather the capability that things—moving things—have to reveal themselves to each other, when they come into relation through mutual proximity. Once again, this should not be read as a metaphor, since “things” in this context does not refer to “objects” or “entities,” but to meaningful data that register (and can interact) in motion and as motion. Taken in this sense “revelation” is far from mysterious. Consider conversation: your ideas reveal themselves to me as they collide with—and thus transform—my own “in exchange”; just like my analysis of Ifá divination reveals itself to you when you “put your mind to it.” Taking the mystery out of it, one may cast the truth-claims of divination as “revelatory” in terms of the modification that results when two initially independent strands of meaning are brought together. No accident, that Newton’s eureka moment came about as a meaningful collision with an apple: it’s just these kinds of eureka moments, writ small on the pages of personal diaries, that divination induces in its motility.

Now, this kind of definition of truth may have an air of vacuous mysticism about it. But I would argue that this impression is owed to the fact that thinking of truth in this way brings to light those aspects of truth-ascription that notions based on representation take for granted (this goes to “vacuity”) and thus obscure ( . . . and this to “mysticism”). Consider the “representationist” account of truth for a moment. Truth, common sense tells us, is an attribute of representations that reflect facts. Hence truth-ascription involves a comparison between representations and facts in order to establish a match of “correspondence” or “coherence,” depending on one’s philosophical preferences.6 However, the notion of comparison indexes a deep circularity here. Logically speaking, comparison presupposes data that are already given “to it” as comparable, for comparison is not the kind of thing that takes place indeterminately: to compare is always to select to compare something with something. Hence, in the case of truth-ascription, the comparison between representation “p” and fact “p” (i.e., the truth-giving match) already presupposes that “p” is selected as the right datum with which to compare “p” (of course it may turn out to be the wrong one, but the point is precisely that match-making must always start somewhere). Now clearly this assertion of comparability is itself implicitly comparative: in supposing that “p” and “p” are worthy of comparison, one is already comparing them—indeed one is establishing an initial match between them. But, from the “representationist” definition of what truth is, it follows that establishing such a match between a representation and a fact (albeit initial) is just to take a tacit stance with regard to the truth of that representation7. The circularity of the definition is obvious: truth-matches presuppose a comparison, which presupposes a truth-match which presupposes comparison, etc.

It is important to stress here that this circularity arises in connection to a deeper conceptual conflict. On the one hand the “representationist” account posits truth as a relational property, inasmuch as it attributes it on to representations that stand at a certain relation to facts (a matching-relation). On the other hand, the relata involved are taken as belonging to distinct ontological camps (representations v. facts). The problem that arises then is one that is typical to dualist ontologies in general, namely the problem of “interaction”: how can tokens of distinct ontological types be brought into relation with one another (as truth-matches are supposed to)? And it is in light of this problem that the “representationist” account both takes for granted and obscures the “initial” truth-matches that we have been pointing to. Initial liaisons between representations and facts need to be assumed in order to render the matching-relations of truth-ascription possible. But they need to stay un-theorized since their ontological anomaly as “half” representation “half” fact, if you like, would become apparent in the glare of analysis.

These kinds of problems, of course, often have solutions, and my argument regarding divinatory truth certainly has no bearing on the philosophical question as to whether or not the “representationist” quandary is soluble. I only claim that Ifá divination turns on an alternative account of truth, and that this alternative can be defined conceptually in terms of its freedom from that particular quandary. This is because what fails to register in the “representationist” account—the “initial” position of truth—is now foregrounded as the basis for the very different conceptualization of truth that we have been exploring. If truth turns on “meetings” between moving trajectories of meaning there is no ontological anomaly to contend with at all and hence no circularity either: unlike “matches” the meetings in question are constituted as relations between tokens of the same ontological type.

One could sum up the point about divinatory truth by way of a reply to the kind of level-headed objection one gets from rationalists. Even if all these ponderous points were valid, is it not nevertheless obvious that diviners and their clients are as interested in ascertaining facts of the world as anyone else? When a babalawo pronounces that a witch is harming his client, or that his client’s fridge is broken, is he not making statements about how things actually stand in the world—about statements of fact? And isn’t this the simple reason why practitioners are interested in what oracles tell them? The answer is no. But “no” not because practitioners are not interested in discovering things about the world, but because these discoveries are not captured by a notion of “ascertaining facts,” or not insofar as that notion implies a process of matching ideas “about” the world with the way the world “actually” is. Verdicts are rather temporary truth-claims that emerge as and when the world reveals itself to itself, if you like. These revelations are “discoveries” in a full sense since they allow practitioners not only to gauge matters that concern them, but also to understand their significance. So when, for example, the arrogant babalawo of Javier’s story exclaimed that the consultant’s fridge must be broken, he was not demonstrating the predictive power of the oracle. The fact that the fridge was indeed broken makes for a good story, but, after all, the consultant did not need the oracle to find this out. Nor would the babalawo consider it a failure if the consultant had answered that the fridge was alright—he would simply project the problem in terms of past or future difficulties, just like Javier did with regard to his consultant’s trip. What the babalawo was demonstrating was the oracle’s ability to unearth even seemingly insignificant data and instate them as constituents of the consultant’s circumstances, intrinsic to the very course of his life.

By way of conclusion, let me defend this hypothesis against a very serious objection. For as things stand, it would appear that all meaningful data must a fortiori also be construed as “true” whenever they are brought in relation to each other, and such a consequence should surely render this notion of truth vacuous. Indeed, if this analysis gives sense to the idea that divinatory verdicts are indubitably true, then why does it not render all truth-claims indubitable? And if it does, then why go to diviners at all?

On this point we can just bite the bullet and admit that as long as they are viewed on a motile premise, all collisions of meaning-trajectories are ipso facto true (indubitably so). But crucially, the reason why truth-claims are not in general posited automatically as true in this sense is that only very few of them carry their motile credentials on their sleeve. So, to take a previous example, in full conversation your ideas may reveal themselves as they collide with my own; they may constitute truth-events on the motile account. But this does not stop me from abstracting away from the trajectories that lead up to these “meetings,” and representing your ideas as statements of distinct “propositions” about things. Indeed, it is probably fair to say that these kinds of representational judgments dominate my thinking during conversation, and that motile considerations remain dormant as a background condition. This is arguably because the representational mould dominates thinking generally, and there are good reasons why this should be so, including, if one is that way inclined, good evolutionary reasons: the ability to align our thoughts with our environment (i.e., to judge stable representations for truth) is an indispensable condition for acting effectively. If this holds for anyone, then it holds for Cubans too, as, Evans-Pritchard assured us, it did for the Azande.

But our analysis suggests that this kind of truth-reckoning is not only different from the motile one, but also logically incompatible with it. From this incompatibility follows an “either/or” clause, whereby representational truth-ascription unavoidably eclipses colliding trajectories and vice versa, since the act of isolating a motile truth-event as a representational proposition is just a way of divorcing it from the trajectories of motion that brought it about, and rendering it as a discrete abstraction. Hence the dominion of representational thinking lodges itself at the expense of motile truth-events, by obviating them as hidden premises. Therefore, the motile account of truth is rendered non-vacuous not because its general applicability is dented as such, but because it is habitually obscured.

Now, I would argue that what distinguishes the interpretative dialectics of Ifá from ordinary conversations is just the fact that divinatory proceedings meticulously maximize the scope for treating verdicts as motile truth-events, and thus resist the dominion of representation, if you like. And this, I would argue, is the crucial role of that chaotic technique of casting palm nuts to arrive at the oddu in the first place. For coincidence is pivotal to this process and the diviners’ verdicts turn on that pivot, as prescribed by the divinatory procedure and its technologies of motion. Casts, then, are truth-events par excellence since they just are temporary equilibriums that emerge from non-causal interactions between salient movements, as we saw.

Or is this too psychedelic? Even if we accept that the motile model of truth might be tenable in some circumstances, surely merely physical events like casts of palm nuts cannot fit the bill. For, unlike the processes of interpretation that follow them, casts do not bring together trajectories that are meaningful as such (what meaning could there be in a mere hand-movement, or a bunch of nuts). Indeed onthis view, the fact that practitioners are prepared to attach such significance to “merely” coincidental outcomes (by virtue of the elaborately meaningful “paths” of each oddu) could be taken just as an indication of the dogmatic and arbitrary character of divinatory belief. However, this objection just amounts to a flat refusal to take the motile premise of Ifá seriously. Oddu-myths appear as arbitrary semiotic appendages on “mere” physical movements only if one takes for granted that meaning is separable from its material “manifestations” (see Holbraad 2007).8 But such an assumption iterates the “representationist” ontology by insisting that meaning can be thought of only as an abstraction. Our motile analysis denies this. Since on a motile premise we can accept that meanings just are parts of the world, we can also accept that parts of the world (like moving hands and palm nut configurations) can be meanings—not as signs that “have” meaning, but as instantiations of meaning pure and simple. The problem then becomes one of revealing what meanings the given movements instantiate, and this, as we have seen already, is a matter of bringing relevant meaning-trajectories together “by coincidence,” to give a truth-event. Divinatory casts do just that, and they do so indubitably.


Thanks are due to Maurice Bloch, Michael Houseman, Caroline Humphrey, Peter Lipton, Morten Pedersen, Rafael Robaina, David-Hillel Ruben, Alan Strathern and Marilyn Strathern for comments on written versions at various stages. The argument’s meta-anthropological frame was inspired by a series of email communications with Eduardo Viveiros de Castro, and I am hugely in his debt for his commentary and encouragement in developing this argument over the years. I also record my gratitude to the Economic and Social Research Council for funding my original Ph.D. fieldwork in Cuba, as well as the Centro (now Instituto) de Antropología for hosting me in Havana.


Bascom, William R. 1941. “The sanctions of Ifá divination.” Man IXXI (1/2): 43–53.

———. 1950. “The focus of Cuban Santería.” Southwestern Journal of Anthropology (6): 64–68.

———. 1991 (1969). Ifá divination: Communication between Gods and Men in West Africa. Bloomington, IN: Indiana University Press.

Bohannan, Paul. 1975. “Tiv divination.” In Studies in social anthropology, edited by John Beattie and Geoffrey Lienhardt, 149–66. Oxford: Clarendon Press.

Bourdieu, Pierre. 1990. The logic of practice, translated by Richard Nice. Cambridge: Polity Press.

Boyer, Pascal. 1990. Tradition as truth and communication: a cognitive description of traditional discourse. Cambridge: Cambridge University Press.

———. 1994. The naturalness of religious ideas: A cognitive theory of religion. Berkeley: University of California Press.

———. 2000. “Functional origins of religious concepts.” Journal of the Royal Anthropological Institute 6 (2): 195–214.

———. 2001. Religion explained: The evolutionary origins of religious thought. London: Random House; New York: Basic Books.

Cabrera, Lydia. 1996 (1980). Yemayá y Ochún. Miami, FL: Ediciones Universal.

Chalmers, David J. 2002. “On sense and intension.” In Philosophical Perspectives 16: Language and Mind, edited by J. Tomberlin, 135–82. Oxford: Blackwell.

Cicero, Marcus Tullius. 1997. Nature of the Gods and On divination, translated by C. D. Yonge. Essex: Prometheus Books Publishers.

Detienne, Marcel. 1996. The masters of truth in archaic Greece, translated by J. Lloyd. London: Zone Books.

Evans-Pritchard, E E. 1976 [abridged from 1937]. Witchcraft, oracles, and magic among the Azande, edited by E. Gillies. Oxford: Clarendon Press.

Fortes, Meyer. 1987. “Religious premises and logical technique in divinatory ritual.” Philosophical Transactions of the Royal Society of London, Series B (251): 409–22.

Holbraad, Martin. 2004. “Religious ‘speculation’: the rise of Ifá cults and consumption in Post-Soviet Havana.” Journal of Latin American Studies 36 (4): 1–21.

———. 2005. “Expending multiplicity: Money in Cuban Ifá cults.” Journal of the Royal Anthropological Institute 11 (2): 231–54.

———. 2007. “The power of powder: multiplicity and motion in the divinatory cosmology of Cuban Ifá (or Mana, again).” In Thinking through things: theorising artefacts ethnographically, edited by A. Henare, M. Holbraad and S. Wastell, 189–225. London: Routledge.

———. 2008. “Relationships in motion: oracular recruitment in Cuban Ifá cults.” In Systèmes de Pensée en Afrique Noire 18: 219–64.

———. 2009. “Ontography and alterity: defining anthropological truth.” Social Analysis 53 (2): 80–93.

Holbraad, Martin. 2012. Truth in motion: The recursive anthropology of Cuban divination. Chicago and London: The University of Chicago Press.

Horton, Robin. 1967. “African traditional thought and Western science.” Africa 37 (1; 2): 50–71; 155–87.

Horwich, Paul. 1998. Truth, 2nd ed. Oxford: Clarendon.

Jackson, Michael. 1989. Paths toward a clearing: Radical empiricism and ethnographic inquiry. Bloomington and Indianapolis: Indiana University Press.

Jules-Rosette, Bennetta. 1978. “The veil of objectivity: prophesy, divination, and social inquiry.” American Anthropologist 80 (3): 549–70.

Jung, Carl G. 1989 (1951). The I Ching or Book of changes. London: Arkana.

Kripke, Saul. 1980. Naming and necessity. Oxford: Blackwell.

Lévi-Strauss, Claude. 1963. Structural anthropology, Vol. 1, translated by C. Jacobson and B. Grundfest Schoepf. London: Allen Lane the Penguin Press.

Matibag, Eugenio. 1997. “Ifá and interpretation: An Afro-Caribbean literary practice.” In Sacred possessions: Vodou, Santería, Obeah, and the Caribbean, edited by M. Fernández Olmos and L. Paravisisni-Gebert, 151–70. New Brunswick, NJ: Rutgers University Press.

Needham, Rodney. 1972. Belief, language and experience. Chicago, IL: University of Chicago Press.

Palmié, Stephan.. 2002. Wizards and scientists: Explorations in Afro-Cuban modernity and tradition. Durham and London: Duke University Press.

Park, George K. 1963. “Divination and its social contexts.” Journal of the Royal Anthropological Institute 93: 195–209.

Parkin, David. 1991. “Simultaneity and sequencing in the oracular speech of Kenyan diviners.” In African divination systems: Ways of knowing, edited by Philip Peek, 173–90. Bloomington and Indianapolis: Indiana University Press.

Plantinga, Alvin. 1964. Faith and philosophy: philosophical studies in religion and ethics. Michigan: William B. Eerdmans Publishing Co.

———. 1974. The nature of necessity. Oxford: Clarendon Press.

Putnam, Hilary. 1975. “The meaning of ‘meaning.’” In Mind, language and reality: philosophical papers, Vol. 2. Cambridge: Cambridge University Press.

Rappaport, Roy A. 1979. “The obvious aspects of ritual.” Ecology, meaning, and religion Richmond, CA: North Atlantic Books.

Sperber, Dan. 1985. On anthropological knowledge. Cambridge: Cambridge University Press.

———. 1996. Explaining culture: A naturalistic approach. Oxford: Blackwell Publishers.

Turner, Victor. 1975. Revelation and divination in Ndembu ritual. Ithaca, NY: Cornell University Press.

Viveiros de Castro, Eduardo. 1998a. “Cosmological deixis and Amerindian perspectivism.” Journal of the Royal Anthropological Institute 4 (3): 469–88.

———. 2002. “O nativo relativo.” Mana 8 (1): 113–48.

———. 2012. Cosmological perspectivism in Amazonia and elsewhere: Four lectures given in the department of social anthropology, Cambridge University, February-March 1998. Hau Masterclass Series, Vol 1.

Wagner, Roy. 1972. Habu: The innovation of meaning in Daribi Religion. Chicago, IL: Chicago University Press.

———. 1981. The invention of culture, revised ed. Chicago, IL: University of Chicago Press.

———. 1986. Symbols that stand for themselves. Chicago, IL: Chicago University Press.

Zeitlyn, David. 1990. “Prof. Garfinkel visits the soothsayers: ethnomethodology and Mambila divination.” Man (N.S.) 25: 654–66.

———. 1995. “Divination as dialogue: Negotiation of meaning with random responses.” Social intelligence and interaction, edited by Esther N. Goody, 189–205. Cambridge: Cambridge University Press.

———. 2001. “Finding meaning in the text: The process of interpretation in text-based divination.” Journal of the Royal Anthropological Institute (N.S.) 7: 225–40.

La vérité au-delà du doute : oracles Ifá à La Havane

Résumé : Cet article analyse le concept de vérité autour de laquelle tourne la pratique de la divination Ifá à Cuba. Partant du constat ethnographique que les praticiens Ifá considèrent les vérités énoncées par leurs oracles comme indubitables, je soutiens que du point de vue des conceptions communes de la vérité une telle hypothèse doit être considérée comme absurde. Pour dépasser une telle imputation, cet article s’attache à reconceptualiser de qui pourrait compter comme vérité dans un tel cas ethnographique. Il est proposé notamment que pour accorder crédit à cette hypothèse d’indubitabilité divinatoire, les notions représentationnelles de la vérité doivent être abandonnées au profit de ce que j’appelle une conceptualisation “mobile”, qui fait de la vérité un événement au sein duquel entrent en collision les trajectoires de significations divinatoires (appelés les « chemins » par les devins). Par une telle analyse, l’article illustre ce que j’appelle une approche « ontographique », dédiée à la cartographie des prémisses ontologiques du discours indigène par la production de concepts qui, bien que n’étant pas des concepts indigènes eux-mêmes, comprennent leurs équivalents proches. Élaboré plus en détail ailleurs (Holbraad 2012), ceci est présenté comme mon point de vue sur ce que les éditeurs de Hau appellent la « la théorie ethnographique ».

Martin Holbraad teaches social anthropology at University College London. Having conducted fieldwork on Afro-Cuban religion and socialist politics in Havana since 1998, he is author of Truth in motion: the recursive anthropology of Cuban divination (Chicago, 2012) and co-editor of Thinking through things: Theorising artefacts ethnographically (Routledge, 2007) and Technologies of the imagination (Special Issue of Ethnos, Spring 2009).


1. This article is a much worked-over and revised version of one that appeared in Portuguese in the Brazilian Journal Mana (Holbraad 2003—the first journal publication to come out of my doctoral research). It presents some of the core arguments explored in my Ph.D. thesis, which are further developed in a monograph on the notion of truth in Cuban divination and in anthropology (Holbraad 2012). Whereas the present argument focuses on establishing what I refer to as the “motility” of divinatory truth, the book advances the claim that this “motility” allows us to conceive of divinatory truth as a peculiarly inventive (sensu Wagner 1981) species of definition.

2. The debate about divinatory truth transcends the disciplinary limits of anthropology (see, for example, Cicero 1997; Jung 1989; Detienne 1996). Within anthropology, and apart from recent arguments by Boyer which are addressed below, the debate has a long history (see for example Evans-Pritchard 1937; Park 1963; Fortes 1966; Bascom 1969; Turner 1975; Jackson 1989; Zeitlyn 1990; 1995, 2001). On the whole, these discussions may be described as species to the genus of anthropological debates about “apparently irrational beliefs” (Sperber 1985). Although I probably join a worthy bandwagon in finding the concept of rationality pernicious as an analytical tool, in this context I retain Sperber’s phrase for heuristic reasons. The phrase is useful inasmuch as it identifies “the problem” not with the beliefs themselves, but with the way they appear to us—not least problematically as “beliefs” (cf. Needham 1972; Boyer 1994: 229).

3. “Untrue” is preferable to “false” here, since it allows us to include as a variation of the “classical solution” Sperber’s influential idea that, like all “symbolic” expressions, “apparently irrational beliefs” are not so much false as empty, in the sense that they do not correspond to determinable propositions that can be judged for truth or falsehood (Sperber 1985).

4. The extension of an expression is its reference. Intension is harder to define, but for present purposes it can be understood as a description of sufficient and/or necessary criteria for determining the extension of a given expression (see Chalmers 2002; cf. Putnam 1975). So, for example, if I ask you what a peccary is and you point one out to me (“there’s one!”), you are giving me the meaning of “peccary” in terms of its extension. But if you explain to me that a peccary is a pig-like animal that lives in South America, you will be giving me the intension of the term. Loosely, we may say that a term’s extension depends on empirical considerations, while its intension depends on conceptual ones.

5. Note that similar comparisons have been made before, notably by Jules-Rosette (1978), Jackson (1989), and Boyer (1990).

6. Note that this is not just an epistemological question about how truth-claims might be arrived at, but rather pertains first to the definition of what truth consists in as such.

7. In relation to the philosophical literature on truth and intentionality, this way of putting the point will sound strange: the “initial match” that I describe is what most people call “reference.” The ordinary intuition on this matter is that representations are true/false insofar as: (1) they contain some type of referring expression (a name, a description, a token-reflexive, etc.); and (2) they combine that referring expression with a property, a relation, etc For example, the truth-claim “you are ire” contains a referring expression “you” and combines it with the property of being iré. Now, on the representationist account, “you are iré” is true if and only if you are iré. But the reason why your being iré is deemed as the truth-giving fact is because the representation “you are iré” has the semantic property of picking you out as a referent. Hence the semantic power or representation—the ability to refer to things without making truth-claims about them—renders comparison a noncircular premise in the definition of truth.
But this escape is superficial because the circularity of the representational account of truth can be recast in terms of reference as well. The only difference is that when it comes to reference, “matching” pertains not to “facts” but to “objects” (broadly construed to include things, people, concepts or whatever one would want to include in the class of referents). For example, “you” has a reference (relative to its context etc.) insofar as you match it as a thing of the world, just like “you are iré” is true insofar as it is matched by the fact of your being iré (for a formal exposition of this parallel, see Horwich 1998: 108). But the reference-match is as much a comparison as the truth-match, and is hence subject to the same circularity: an “initial” reference-match has to be posited, etc.

8. The alleged ontological distinction here is analogous to the one encountered in our discussion of divinatory interpretation. If there representationism amounted to assuming an ontological gap between representations and the world (i.e., as an issue regarding the metaphysics of semantics), here the distinction is drawn between representations and the worldly vehicles through which they get expressed (i.e., at the semiotic level: signified v. signifier). However, since our motile analysis denies a premise that these variants of representationism have in common (viz. that meanings are abstract), it serves as a tenable alternative to “semiotic” representationism also. On such a view, oddu are not arbitrary signifiers of abstract (signified) meanings, and nor does the system of 256 oddu constitute a semiotic “code.” Rather, the relationship between the material manifestation of oddu during the séance and its meaning as expressed in mythical “paths” can be thought of as analogous to the relationship between a person and his personality: there is no arbitrariness because the oddu just is its meaning, for those who are acquainted with it (i.e., the babalawos who “study”). Hence the fact that each oddu is properly considered a deity in its own right (see above). So, inasmuch as oddu in Cuba are commonly referred to as “signs of Ifá” (signos), then these signs, quite literally for practitioners, speak for themselves, to echo Roy Wagner’s famous phrase (Wagner 1986).