The French verb croire [“to believe”]¹ is paradoxical in that it expresses doubt as well as assurance. To believe [croire] is to state a conviction; it is also to add a nuance to that conviction: “I believe” [je crois] often signifies “I’m not sure.” This ambiguity involves the subjective side of belief [croyance]. As regards its object, the situation is no less equivocal, since the complement of the verb can be produced in two ways: direct or indirect. What is more, the indirect construction itself has two forms: croire á . . . [“to believe in,” “to think of ”] is not the same thing as croire en. . . [“to believe in,” in the sense of being willing to rely on], which both differ from croire + direct object and croire que . . . [“to believe that . . .”]. Finally, the meaning of the verb and the construction of the complement can vary depending on the nature of the object: man, god, fact, value, statement. . . .

This suggests two questions (at least): is it possible to order this diversity of usages? If so, is this order universal or does it characterize only a certain type of culture, and in this case what is the basis for the word’s unity? In other words: how is it that multiple meanings do not require diverse expressions?² But since this is apparently the case, is a translation of the verb in all its senses possible in other languages, using a single term?

Croire á . . . is to state that something exists; croire en . . . is to have confidence; croire que . . . is for something to be represented in a certain way. Although the difference between the two indirect constructions may appear slight, it is undeniable, as the following example shows: a person believes in [croire en, “trusts in”] God, while one believes in [croire á] the Devil, that is, one recognizes that he exists without, by definition, putting one’s faith in him: one cannot croire en the Devil. Croyance en [“belief in, trust in”] God does imply croyance á [“belief in”] his existence, but implication is not identity. On the other hand, this implication seems so obvious that it often goes unformulated: a believer believes in [croire en, “trusts in”] God, he feels no need to say that he believes in [croire á] it, one would say, implicitly. But is this certain? In fact, the believer not only need not say that he believes in [croire á] the existence of God, but he did not even believe in [croire á] it; precisely because in his eyes there can be no doubt about it: the existence of God is not believed in [crue], but perceived. On the contrary, to make God’s existence an object of belief, to state this belief, is to open up the possibility of doubt—which begins to clarify the ambiguity with which we started. So one could say that it is the unbeliever who believes that the believer believes in [croire á] the existence of God. One could call this a play upon words; but these words do lend themselves to it, and it is precisely this possibility that must be explored, if not elucidated, in trying to organize the field of their usages. Besides, what I have just said will appear much simpler if we leave the religious domain. If I have confidence in a friend, if I believe in [croire en, “have faith in”] him, will I say that I believe in [croire á] his existence? Certainly not; that existence is, simply, undeniable. It is only if it were not unquestionable that I would have to believe in [croire á] it, and believe in it explicitly. Again, it will probably be said that this is playing on words, this time on the word “existence,” for the man’s existence, by definition, is not the same level as that of the deity. By definition, yes, but by cultural definition: the distinction between a cultural world and a natural world, or between a “this world” and a “beyond,” is widespread, but it is not universal. It is this distinction between two modes of existence that leads to a knowledge on one side, belief [croyance] on the other. From this kind of perspective, the existence of supernatural beings can only be an object of belief, and this is why wherever this distinction is made the phenomenon of belief as the affirmation of existence takes on this ambiguous aspect, between the certain and the questionable.

This is not the only reason. Let us now consider the relations between croire á . . . [“to believe in (a fact)] and croire que . . . [“to believe that . . .”]. To believe in [croire á] the existence of X—“god, table, or washbasin”—can be expressed in a direct construction: to believe that [croire que] X exists. But this is a statement of a peculiar type—the existence of a god or of a hundred thalers is not an attribute—and is different from the statement that endows X with certain characteristics which permit X to be represented to oneself. The representation, the content of belief, is accompanied by an affirmation of existence but is separable from this affirmation; the affirmation can be bracketed—the Husserlian epoché—and this is what makes possible studies of beliefs as such: one need not believe in [croire á] what one believes in order to analyze it. The “I believe” [je crois] which precedes so many statements of the most diverse kinds, is the mark of a distancing and not of an adhesion.

These two movements, which a single verb is able to express, appear radically opposed, or else completely unrelated. Belief as representation, as statement, pertains to what is also called ideology; there is no isolated belief, every representation is part of a global system which is more or less clearly, more or less consciously articulated, a system which may be religious but may also be philosophical, political . . . . Believe as faith [confiance] is the conviction that he to whom one has given something will reciprocate in the form of support or protection; it calls forth a relationship of exchange, of which the relationship between the believe and his god is only a particular case, even if a frequently privileged one. One puts one’s faith [confiance] to the same end, whether in an individual, in a party, in an institution. It is significant in this regard that Benveniste, in his Indo-European Language and Society (1969; English edition, London, 1973), discusses belief [la croyance] not in the section on “religion” but in that on “economic obligations.” For he sees the original meaning of belief in this credit which has been accorded and should be returned. Must we then see belief-as-representation as a derivative meaning? Or else as a meaning that has been added on, and which would turn the verb “to believe” into a conglomerate without unity?

This derivation is certainly a possible one: to beliefe in [croire en] someone, to give him credit, is, among other things, to believe [croire] what he says, and in this way one goes from the trust to the statement that it allows one to take as established fact. This is especially evident when the belief appears in the form of religious faith: trust in [croyance en] a god is usually the basis of what we call a credo, a group of statements which become the direct object of belief. The same is true in many other domains. For political examples, there is an overabundance of choices. But it is also possible (more often than is usually . . . believed!) to accept a proposition that is said to be scientific just as one accepts a dogma or even the possibly fantastic assertion of a man who is judged worthy of trust; I believe it not because I am able to prove it, but because I have faith in those who say they have proven it, for example, in Einstein when, following him, I write E=mc². But we would miss the essential part of belief-as-representation if we reduced it to this sole case in which it is based on the argument of authority. The specific trait of a representation is to appear obvious, to be self-evident, and the fact that it is always possible to bracket the judgment or the feeling of obviousness changes nothing: the obvious is replaced by the arbitrary, but both simply mean that this form of belief is based on nothing but itself or the cultural system within which it finds its meaning.

So it seems impossible to overcome the polysemy of the word. Its religious usage does allow us to unify the verb’s three constructions, but it does not eliminate the other usages; over and above this, it only unites the three constructions in religions of a certain type. This observation leaves us to question its anthropological usage, now well established and apparently unproblematic.³ What anthropologist would deny that he seeks to uncover the beliefs of those of whom he studies, to compare them with our own beliefs or those of other peoples, as if this object of study and its designation presented no a priori problem, as if it were obvious that every human being “believes” [croire]—this being one of our beliefs—in the same way, if not, of course, in the same things? The danger in this situation is not simply the well-known if not always foreseen one of inappropriately applying a category that may have meaning only in our own culture; it has to do with the fact that this category may not be a single, unified one at all, even for us, or at the least that is a shattered category, whose fragmentation is, precisely, a singular cultural phenomenon. What is more, anthropological usage reduplicates the paradox I emphasized above when I said that it is the unbeliever who believes that the believer believes. If for example I say that the Dangaleat⁴ believe in [croire á] it in the same way that I imagine I could, if I did. But how can one tell whether they believe [croire] and in what way? What question can one ask them, using what word of their language, in what context? Or, inversely, how is it possible to translate into French the word or words they use to talk about what is to our eyes an object of belief?

In J. Fédry’s Dictionnaire dangaleat,⁵ we find the verb àbidé “to perform the rites faithfully.” It comes from the local Arabic abada, “to adore God,” adoration being understood as a ritualized activity. It is a matter of worship [du culte], of faith in action, and not of the representation of a being whose existence must also be affirmed. This verb is used with a direct-object complement: this being, God for converts to Christianity or Islam, or the margaï for others. The best way to translate it is thus “to serve,” in the biblical sense of the trm: to worship [rendere un culte à]. No abday margaï, “I serve the margaï.” Another verb, amniye, signifies “to bestow one’s trust on,” “to rest on,” “to believe in” [croire en]. It is constructed with an indirect-object complement, introduced by the preposition ku: no amnay ku marigo, “I have faith in the margaï,” “I give my faith to the margaï”; this is the verb that Christian use to say “I believe in [croire en] God,” no amnay ku bungir. In contrast to the foregoing, this verb is not used exclusively in religious contexts: one can evidently, as in French, put one’s faith in another person. The first sense given by the dictionary, besides this, is “to be used to, familiar with . . . ,” and one will say, for example: no amniy-iy-g pisò, “I am used to horses.” This too is a word of Arabic origin whose Semitic root has given us the “amen” of Christian liturgy which, as Fédry points out, marks adhesion to a person more than to a conceptual “truth.” As this author notes, “one may wonder about the fact that both of these verbs come from Arabic, whose linguistic influence is very strong Dangaleat have taken what fits their own way of “believing” [croire]: the terms that designate a specific behavior and mental attitude—worshipping [rendre un culte] and trusting in the addressee of the worship—and not terms that are based on definite representations or propositions.

One may thus translate our “believe in” [croire en] into Dangaleat, and the fact that the Hadjeraï took the word from Arabic suggests that for them it expresses the essential aspect of belief (and of religious faith in general, says Fédry, who belongs to the Society of Jesus and should know whereof he speaks): faith [la confiance]. But in this case, how can we translate “to believe [that]” [croire que]? To find out, to know, to know about something, is ibiné; pakkine serves to render: think, suppose, figure out, foresee. These two verbs are pure Dangaleat. The first will be used to mark certainty and so translates “to believe” [croire] in cases where the French verb is more or less equivalent with “to know,” when for instance Don Juan says to Sganarelle, who is questioning him about belief, “I believe that two and two make four.” The second verb covers the doubtful usages of our verb, all those in which the speaker takes a certain distance with regard to what he is talking about.

In sum, we can translate all aspects of the verb “to believe” . . . except the verb itself. What we have been able to translate has been the French equivalent of “to believe” in each of its particular usages, but in Dangaleat there is no single term that serves at the basis of their unification. In other words, we can translate everything except the ambiguity. We must therefore return to the reasons for this ambiguity. Ambiguity is not simply polysemy, the fact that a verb sometimes has one meaning and sometimes another, each of them unequivocal; it is, rather, that all of these meanings, even the contradictory ones, are intrinsically liked; that, above all, there is always doubt at the heart of the conviction, and that the affirmation itself indicated that it could always be suspended. But why condense this paradoxical liason into a single word instead of sorting out its elements, as the Hadjeraï do? The answer, “I believe,” lies in the comparison between a religion like Christianity and a religion like that of the Dangaleat.

It is not so much the believer, I would say, who affirms his belief as such, it is rather the unbeliever who reduces to mere believing what, for the believer, is more like knowing. Nevertheless, the Christian cannot avoid expressing his faith not only as trust in God [confiance en], but also as belief in [croyance à] his existence and belief that [croyance que] God possesses such and such attributes, that the world was created, and so forth. He states this as a belief, even though he knows it—but also because he knows that by this very fact it is contestable and contested. Above all he knows that there are other beliefs, on the one hand, because his religion has a history and was constituted against the “false” gods, on the other because this history is not over yet and there are still idols to be eliminated; and there can be other beliefs only because his own belief is one among others. Next, he knows—it is even an essential point in his credo—that the object of his belief is in a “reality” of a different order than the realities of the world of creation, which are the object of a permanently revisable scientific knowledge, or of calculations, of predictions that can be proven wrong; and he also knows that this possibility of revision lies in the demonstrable or verifiable character of the knowledge or the hypothesis, a character whose legitimacy he challenges in the case of his belief, but which, inversely, challenges the legitimacy of his belief. Thus he must simultaneously assume both his affirmation and the challenge to it, a challenge that belief is, nonetheless, supposed to make impossible on its own level. In other words, the contradiction is inside his faith, and that is what it is “to believe.”⁶

This situation is the result of the distinction made between two worlds: the Kingdom of God and this world. In our culture such a distinction seems so characteristic of religion, to those who reject it as much as to its adherents, that religion in general and so-called “primitive” religions in particular are usually defined by a belief in supernatural powers and by their worship. There is even a tendency to think that the extent and importance of the supernatural world are much greater for “primitives” than for “moderns,” that super-nature is not only the domain of gods and spirits but also, for example, the domain in which the power of the magician and the sorcerer operates. I certainly do not mean to deny that at any latitudes one can find people who believe in [croire à] the supernatural, but one equally finds people for whom such an affirmation is completely meaningless—without them being, for all that, areligious—far from it. For here we have a major misunderstanding: because we have constructed the concept of natural law, we are ready to admit the supernatural (whether as illusion or as other reality hardly matters) as a place to put whatever contravenes, or seems to contravene, natural law; but this is our own notion, whether we judge it well grounded or not, and not that of the people to whom we abusively attribute it. As Evans-Pritchard remarks, “many peoples are convinced that deaths are caused by witchcraft. To speak of witchcraft being for these peoples a supernatural agency hardly reflects their own view of the matter, since from their point of view nothing could be more natural.”⁷ For his part, Claude Lévi-Strauss has stressed the realist, materialist character of magic, its monistic, not dualistic, conception of the world.⁸

The margaï, the spirits who have such an important place in the individual and social lives of the Hadjeraï, are invisible, nonhuman powers; they act unpredictably, and are the cause of whatever disturbs the natural course of things. But they are no less a part of the same world as human beings. The latter believe in [croire à] the existence of the margaï like they believe in their own existence, in that of animals, things, atmospheric phenomena . . . . Or rather they do not believe in [croire à] it: this existence is simply a fact of experience:⁹ there is no more need to believe in [croire à] the margaï than to believe that if you throw a stone it will fall. One fears and/or trusts in them, one gets to know them, one gets used to them, one performs the special sacrifices that please each margaï, and one is careful to make no mistakes, for fear of getting sick or being affected in some unpleasant way. If we can speak of a Dangaleat religion—another untranslatable expression—it is not in the sense in which the faithful share a single elaborated body of beliefs about supernatural beings, but rather in the etymological sense, according to Benveniste, of the Latin religiō:¹⁰ that of a meticulous concern for the proper carrying out of the cult, without, however, being able to define the necessary correctives in advance; at every occasion, one takes aim within uncertainty. One can only estimate what each margaï desires. The four verbs mentioned above define these behaviors equivocally and without contradictions: one serves the margaï, one trusts in them (that is, in the mutually fruitful nature of the exchange inaugurated by the sacrifice), one knows from experience that they exist, and one tries to guess their intentions. All this does presuppose a particular representation of the world, but one which excludes the possibility of its explanation in the form of “belief,” of an assertion that in spite of itself is doubtful, relativized. The Dangaleat certainly know that others think differently, and it happens that many of them convert to Islam or to Christianity. But this situation cannot surprise them: one does not believe in [croire à] the margaï; one experiences them, and this experience is first of all a local one; such spirits do not necessarily exist everywhere. While the encounter with otherness relativizes Christian belief in an otherworldly absolute, it confirms the Dangaleat experience of the world, which is relative from the beginning and so cannot be disturbed by diversity. This is why religions of the Dangaleat type are without the proselytizing inherent in religions founded on beliefs whose vulnerability impels their formidable dynamism.¹¹

If the Dangaleat have no need of the verb “to believe” this is not solely because of their monism, as opposed to Christian dualism. Equally in play is another opposition, one between the historicism of the Christian religion and the empiricism of Dangaleat religion. This empiricism assures everyone of the presence of the margaï, and has no need of an intercessor. Every man performs his own sacrifices and will have recourse to the diviner only to know what animal, or what sex and what color, he should kill and on what day. A religion like Christianity or Islam is based, on contrary, on a revelation, testimony, a transmission whose fidelity is guaranteed by a church or specialized experts. This revelation is, precisely, that another world exists; the revelation is a unique historical event, its content is constituted by the words of its protagonist, God incarnate or prophet. So everything rests on a faith, which is simultaneously a trust and a specific credo. All the meanings of the verb “to believe” should then come together, but his necessity is nothing more or less than a cultural necessity. It is only in this perspective, in my opinion, that we can speak of “religious belief,” and it is only when it is understood that this notion does not have universal value that we can appreciate how difficult the problem of a general definition of religion really is; but this may also be the point from which we can try to resolve the problem.