The relative native

This work is licensed under the Creative Commons | © Martin Holbraad. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau3.3.033


Turning a corner

Preamble for “The relative native” by Eduardo Viveiros de Castro

Martin HOLBRAAD, University College London


In his own preamble to the original Portuguese version of “The relative native” (2002), Eduardo Viveiros de Castro explains that the article’s aim is to elaborate “metatheoretically” upon his earlier arguments regarding the cosmology of Amazonian animism, published in Portuguese in 1996 and in English in 1998—the now classic JRAI article “Cosmological deixis and Amerindian perspectivism” expanded upon in his Cambridge lecture-series of the same year, titled “Cos-mological perspectivism in Amazonia and elsewhere,” which is now available as Volume 1 of HAUs Masterclass Series (2012). Viveiros de Castro’s move to extract metatheoretical implications from his theory of perspectivism in Amazonia recalls Roy Wagner’s claim that The invention of culture (1981) sets forth the “epistem-ology” (1981: xv) that corresponds to his earlier ethnographic presentation of the role of invention in Daribi social life in Habu (1973; see also Strathern 1980—the mediating term in a Wagner-Strathern-Viveiros metatheoretical triptych). Certainly, each of these “moves to meta” can be seen as exercises in anthropological “recursion” (Corsin Jimenez, in press; Holbraad 2013). For example, where Wagner Melanesianizes anthropology by taking Daribi manners of invention as an ethnographic template from which to reinvent the activity of anthropology itself, namely as a manner of invention also (and note the virtuously double circularity of the recursion), Viveiros de Castro’s equivalent attempt to Amazonianize anthropology in the present article consists in adopting for anthropology the core tenet of Amazonian perspectivism, namely, the idea that differences between “perspectives” are to be seen in ontological rather than epistemological terms.

In Amerindian cosmologies, Viveiros de Castro writes in the earlier ethnological article, different species do not see the same thing (or “world”) in different ways, but rather “see in the same way … different things [or ‘worlds’]” (Viveiros de Castro 1998: 478). Similarly in anthropology, runs the metatheoretical argument elaborated here, the difference between anthropological analyst and ethnographic subject lies not in the different perspectives each may take upon the world (their respective “world-views” or even “cultures”) but rather in the ways in which either of them may come to define what may count as a world, along with its various constituents, in the first place. For Amazonians and anthropologists alike, then, difference pertains to the ontological question of what things are or indeed could be, rather than how they might be differentially “represented,” “known” (or at least “believed”), or for that matter “constructed.” So in this sense, anthropologists, again like Amazonians, may best be conceived as multinaturalists. If nature is meant to designate what there is and cultures are the different ways in which this can be seen, then Viveiros de Castro’s anthropology is one that multiplies natures rather than cultures (hence also the deep affinity of this vision of anthropology with Bruno Latour’s parallel argument about the ontological pluralization of nature in the practice of science—e.g., Latour 1993).

So Viveiros de Castro brings Wagner’s assault on the nature/culture matrix of anthropological thinking full circle. If Wagner recasts the idea of culture as the manner in which the world invents itself (culture does as nature is, one might say), Viveiros de Castro recasts the idea of nature as the manner in which people conceive it (nature is as culture does it), provided “conception” here is understood in sharp contrast to representation, as an irreducibly ontological operation—establishing this point, and drawing out its consequences for the practice of anthropology, is one of the prime tasks of the article. If there is any asymmetry at all in these complementary operations it lies in the fact that, by working at the reversible “nature = culture” equation from its “nature” end—the end of what things are and what they could be—Viveiros de Castro’s argument brings the question of ontology to the very center of anthropologists’ metatheoretical deliberation. Indeed, in view of the current furor about the so-called “ontological turn” in anthropology,1 one way to read the article that follows is as just that: the moment when anthropology turned the ontological corner.


Corsin Jimenez, Alberto. In press. “The right to infrastructure: A prototype for open source urbanism.” Environment and Planning D: Society and Space.

Holbraad, Martin 2013. “Scoping recursivity: A comment on Franklin and Napier.” Cambridge Anthropology 31 (2): 123–27.

Latour, Bruno. 1993. We have never been modern. Translated by Catherine Porter. London: Prentice-Hall.

Scott, Michael W. 2013. “The anthropology of ontology (religious science?).” Journal of the Royal Anthropological Institute 19 (4): 859–72.

Strathern, Marilyn. 1980. “No nature, no culture: The Hagen case.” In Nature, culture and gender, edited by Carol MacCormack and Marilyn Strathern, 174222. Cambridge: Cambridge University Press.

Viveiros de Castro, Eduardo. 1998. “Cosmological deixis and Amerindian perspectivism.” Journal of the Royal Anthropological Institute (N.S.) 4 (3): 46988.

———. 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, University of Cambridge, February-March 1998.” HAU: Masterclass Series 1: 45–168.

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

———. 1981. The invention of culture. Revised and expanded edition. Chicago: University of Chicago Press.


Martin HOLBRAAD teaches Social Anthropology at University College London where he codirects the Cosmology, Religion, Ontology and Culture Research Group (CROC). He conducts ethnographic fieldwork on religion and socialism in Cuba and is the author of Truth in motion: The recursive anthropology of Cuban divination (Chicago, 2012).


1. See Scott (2013) for a recent review of the literature on the anthropology of ontology.


This work is licensed under the Creative Commons | © Eduardo Viveiros de Castro. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau3.3.032


The relative native

Eduardo VIVEIROS DE CASTRO, Museu Nacional, Universidade Federai do Rio de Janeiro

Translated from the Portuguese by Julia Sauma and Martin Holbraad

The human being, such as we imagine him, does not exist.

—Nelson Rodrigues

Ground rules1

The “anthropologist” is a person whose discourse concerns the discourse of a “native.” The native need not be overly savage, traditionalist nor, indeed, native to the place where the anthropologist finds him. The anthropologist, on his part, need not be excessively civilized, modernist, or even foreign to the people his discourse concerns.2 The discourses in question (and particularly that of the native) are not necessarily texts, but rather may include all types of meaning practice.3 What is essential, however, is that the discourse of the anthropologist (or the “observer”) establishes a certain relation with that of the native (or the “observed”). This relation is one of meaning or, when the anthropologist’s discourse aspires to be Scientific, a relation of knowledge. By this token, anthropological knowledge is also a social relation, since it is the effect of the relationships that reciprocally constitute the knowing subject, on the one hand, and the subject he comes to know, on the other. As with all relations, this form of knowledge brings about a transformation in the relational constitution of anthropologist and native alike.4

The (meta)relation between anthropologist and native is not one of identity: the anthropologist always says and, therefore, does something different than what the native says or does, even when he intends to do nothing more than repeat the native’s discourse in a “textual” form, or when he tries to establish a dialogue—a dubious notion—with the native. This difference is nothing other than the knowledge effect created by the anthropologist’s discourse, which is produced by the relation between the meaning of this discourse and the meaning of that of the native.5

Clearly, this kind of discursive alterity is grounded in an assumption of similarity. The anthropologist and the native are of the same species and share in its condition: they are both human, and each of them is positioned in their respective culture, which could (even) be the same. But this is where the game starts to get interesting or, better, strange. For even when the anthropologist and the native share the same culture, the relationship of meaning between their respective discourses serves to differentiate them: the anthropologist’s and the native’s relationship with their respective cultures are not exactly the same. What makes the native a native is the presumption, on the part of the anthropologist, that the native’s relationship with his culture is natural, which is to say, intrinsic, spontaneous, and, if possible, nonreflexive or, even better, unconscious. Thus, the native gives expression to her culture in his discourse. The anthropologist does so too, but if he hopes to be something other than a native, he must also be able to express his culture culturally, which is to say, reflexively, conditionally, and consciously. The anthropologist’s culture is contained (in both senses of the word) in the relationship of meaning that his discourse establishes with that of the native. The native’s discourse, by contrast, is merely penned in by his own culture. The anthropologist’s deployment of his own culture is a necessary condition of his humanity, one might say, while for the native being deployed by his is a sufficient one.

Obviously, these differences are not in the so-called nature of things. They are a feature of the language game that we are describing here, and serve to define the very characters we have been designating as “the anthropologist” and “the native.” So let us turn to some other ground rules.

The anthropological idea of culture places the anthropologist and the native on an equal footing, inasmuch as it implies that the anthropologist’s knowledge of other cultures is itself culturally mediated. in the first instance, this sense of equality is simply empirical or de facto, since it refers to the common (or generic) cultural condition of the anthropologist and the native. However, their differently constituted relationships with their respective cultures, and therefore also with each other’s, are such that this de facto sense of equality does not imply an equality de jure—that is, an equality with regard to their respective claims to knowledge. The anthropologist tends to have an epistemological advantage over the native. Their respective discourses are situated on different planes. While the anthropologist’s capacity to produce meaning does depend on the meanings produced by the native, the prerogative to determine what those native meanings mean remains with the anthropologist—explaining and interpreting, translating and introducing, textual-izing and contextualizing, justifying and signifying native meanings are all jobs of the anthropologist. The anthropological discourse’s relational matrix is hylomor-phic: the anthropologist’s meaning is form; the native’s is matter. The native’s discourse is not the master of its own meaning. De facto, as Geertz might say, we are all natives; but de jure, some are always more native than others.

This article proposes the following questions: What if we refuse to give this kind of strategic advantage to the anthropologist’s discourse over that of the native? What would happen if the native’s discourse were to operate within the discourse of the anthropologist in a way that produced reciprocal knowledge effects upon it? What might occur if the form intrinsic to the matter of native discourse were to be allowed to modify the matter implicit in the form of anthropological knowledge? it is said that to translate is to betray. But what happens when the translator decides to betray his own tongue? What happens if, unsatisfied with a mere passive or de facto equality between discursive subjects, we claim an active or de jure equality between their respective discourses? What if, rather than being neutralized by this equivalence, the disparity between the meanings produced on either side, by anthropologists and natives, is introduced into both discourses, thus releasing its full potential? What if instead of complacently admitting that we are all native, we take the opposite wager as far as it can go, namely, that we are all “anthropologists” (Wagner 1981: 36)—and, to boot, not some a little more than others, but just each in their own way, which is to say, very differently? in short, what changes when anthropology is taken to be a meaning practice that is epistemically continuous with the practices that it discusses, and equivalent to them? What changes, in other words, when we apply the notion of “symmetrical anthropology” (Latour 1991) to anthropology itself, not to condemn it as colonialist, exorcise its exoticism, or landmine its intellectual field, but rather to turn it into something else? Something different not only to the native’s discourse (for that is a difference that is constitutive of anthropology), but different also to the discourse that anthropologists habitually enunciate about themselves, often in hushed tones, when commenting on native discourses.6

If we do all of this, I would say that we would be doing what has always been called “anthropology,” properly speaking, rather than (for example) “sociology” or “psychology.” My hesitation here is due to the fact that much of what goes, or has gone, by the name of anthropology turns on the contrary assumption that the anthropologist has a privileged grasp of the reasons for the native’s reasons-reasons to which the native’s reasonings are oblivious. The anthropologist, according to this view, is able to provide a full account of how universal or how particular any given native might be, as well as of the illusions that the latter may have about himself-at times providing an example of his native culture while imagining that he manifests human nature in general (the native as unselfconscious ideologue), while at other times manifesting his human nature while thinking that he is displaying his own particular culture (the native as unwitting general cognizer).7 Here, the knowledge relation is conceived as unilateral, such that the alterity between the anthropologist’s and native’s respective discourses dissipates as the former encompasses the latter. The anthropologist knows the native de jure, even as he may not know him de facto. Or we could go the other way around: even though the native may know the anthropologist de facto (often better than the anthropologist knows him in turn), he does not know him de jure, precisely because the native is not an anthropologist, which is what the anthropologist, well, is. Needs must, the anthropologist’s science/knowledge is of a different order to the native’s: the condition of possibility for the former includes the denial of the latter’s claim to legitimacy—an act of “epistemocide,” to use Bob Scholte’s acute expression (1984: 964). The subject’s knowledge requires the object’s ignorance.

But we need not be overly dramatic about all this. As the discipline’s history attests, this discursive game and its biased rules provided lots of instructive information about the natives. The experiment proposed in the present article, however, consists precisely in refusing to play it. This is not because this game results in objectively false results, say in representing the native’s nature erroneously; the concept of objective truth (along with the notions of representation and nature) is part of the rules of that game, not of the one proposed here. In any case, once the aims of that classic game are set, its results are frequently convincing, or at least, “plausible” as adepts of the game like to say.8 To refuse to play the game amounts simply to giving oneself a different set of goals, appropriate to different rules, as outlined above.

What I am suggesting, in short, is that there are two incompatible ways of conceiving anthropology, and that one needs to choose between them. On one side, anthropological knowledge is presented as the result of applying concepts that are extrinsic to their object: we know what social relations, cognition, kinship, religion, politics, etc. are in advance and the task is to see how these play out in this or that ethnographic context—how they play out, of course, without the knowledge of the people involved. On the other side (and this is the game proposed here), we have an idea of anthropological knowledge that is founded on the basic premise that the procedures involved in anthropological investigation9 are of the same conceptual order as the procedures being investigated. It should be emphasized that this particular equivalence of procedure at once presupposes and produces the radical nonequivalence of everything else. For, if the first conception of anthropology imagines each culture or society as embodying a specific solution to a generic problem—or as filling a universal form (the anthropological concept) with specific contents—the second, in contrast, raises the prospect of the problems themselves being radically diverse. Above all, such an approach takes off from the principle that the anthropologist may not know in advance what these problems might be. In such a case, anthropology poses relationships between different problems, rather than placing a single (“natural”) problem in relation to its different (“cultural”) solutions. The “art of anthropology” (Gell 1999), I suggest, is the art of determining the problems posed by each culture, not of finding solutions for the problems posed by our own. It is just for this reason that positing a continuity between the procedures of the anthropologist and the native is such an epistemological imperative.10

It bears repeating that this pertains to the procedures, not to those that carry them out. After all, none of this is about condemning the classic game for producing faulty results that fail to recognize the native’s own condition as Subject-observing him with a distant gaze, devoid of empathy, which constructs him as an exotic object, diminishes him as primitive rather than the observer’s coeval, denying him the human right of interlocution-we are familiar with the litany. The problem is rather the opposite. It is precisely because the anthropologist very easily takes the native to be an other subject that he cannot see him as an other subject, as an Other figure that, more than subject or object, is the expression of a possible world. It is by failing to accept the native’s condition of “nonsubject” (i.e., his being other than a subject) that the anthropologist introduces his sneaky advantage de jure, under the guise of a proclamation of de facto equality. Before the game even starts, he knows too much about the native: he predefines and circumscribes the possible worlds expressed by this other, radically separating the other’s alterity from his capacity to induce difference. The authentic animist is the anthropologist, and participant observation is the true (meaning, false) primitive participation.

It is therefore neither a matter of advocating a kind of intersubjective idealism, nor of standing up for some form of “communicative reason” or “dialogic consensus.” My touchstone here is the concept evoked above, namely the Other as an a priori structure. This concept is proposed in Gilles Deleuze’s well-known commentary of Michel Tour-nier’s Vendredi.11 Reading Tournier’s book as a fictional description of a metaphysical experiment—what is a world without Others?—Deleuze proceeds to gauge the effects of the Other’s presence through the effects of its absence. The Other thus appears as a condition of the field of perception: the existential possibility of those parts of the world that lie beyond actual perception is guaranteed by the virtual presence of an Other that perceives them; what is invisible to me subsists as real by being visible to an other.12 Without an Other the category of possibility disappears; the world collapses, reduced to the pure surface of the immediate, and the subject dissolves, turning into a thing-in-itself (while things-in-themselves, in turn, unravel into phantom doubles). An Other is thus no one (neither subject nor object) but rather a structure or relation—the absolute relation that provides concrete actants with their relative positions as subjects or objects, as well as their alternation between the two positions: the Other refers (to) me to the other I and the other I to me. The Other is not an element within the field of perception; it is the principle that constitutes such a field, along with its content. The Other is thus not a specific point of view to be defined in relation to the subject (the “point of view of the other” in relation to my point of view or vice-versa), but rather it is the possibility that there may be a point of view at all—that is, it constitutes the concept of a point of view. It is the point of view that allows the I and the Other to adopt a point of view.13

On this point, Deleuze is critically extending Sartre’s famous analysis of the “gaze,” by providing a prior structure for the reciprocity of perspectives associated with the Sartrian regard. What is this structure? It is the structure of the possible: the Other is the expression of a possible world. A possible world that exists, really but not actually—or not beyond its expression in the form of an Other. This express possibility is implicated in, and constitutive of, the perspective from which it is expressed (which nevertheless remains heterogeneous), and is effectuated in language, or the sign, which provides the reality of the possible as such—meaning. Thus, the I renders explicit this implication, actualizing its possibility by taking its rightful place in the language game. The subject is therefore an effect, not a cause, inasmuch as it interiorizes a relation that is initially exterior to it—or rather, a relation to which it is initially interior: relations are originally exterior to the terms, because the terms are interior to the relations. “There are many subjects because there are others, and not the contrary.” (Deleuze and Guattari 1991: 22)

The problem is thus not that of seeing the native as an object, and the solution is not to render him a subject. There is no question that the native is a subject; but what the native forces the anthropologist to do is, precisely, to put into question what a subject can be. This is the cogitation peculiar to anthropology: one that allows anthropology to take on the virtual presence of an Other who is also its condition—the condition for passage from one possible world to another—and that is only as a consequence able to determine the derivative and vicarious positions of subject and object.

The physicist questions the neutrino, and cannot disagree with it; the anthropologist answers for the native, who can thus only (de jure and, frequently, de facto) agree with him. The physicist must associate himself with the neutrino—he must think with his recalcitrant object; the anthropologist associates the native with himself, thinking that his object makes the same associations as he does—that is, that the native thinks like him. The problem is that, like the anthropologist, the native certainly thinks, but, most probably, he does not think like the anthropologist. The native is certainly a special object: a thinking object, or, a subject. But if he is objectively a subject, then his thinking also takes objective form—just as the anthropologist’s thinking does—as the expression of a possible world. Thus, the Malinowskian distinction between what the native thinks (or does) and what he thinks that he thinks (or does), is spurious. It is precisely this cleavage, this bifurcation of the nature of the other, that the anthropologist (who would have himself do exactly as he thinks)14 hopes to exploit. However, a better distinction— the difference that really makes a difference—is between what the native thinks (or does) and what the anthropologist thinks the native thinks (and acts accordingly). The true confrontation is between these two manners of thinking (and acting). Such a confrontation, note, need not be reduced to similar forms of equivocation in each case—the misunderstandings are never the same on either side, just as the sides are not the same in the first place. In any case, who could venture to define what may count as mutual understanding here? But nor is it necessary to content ourselves by imagining this manner of confrontation as some kind of edifying dialogue. The confrontation should implicate the two sides mutually, altering the discourses it brings into play in equal measure, since the aim of the procedure is not to arrive at a consensual optimum, but a conceptual maximum.

I evoked earlier the critical distinction between quid facti and quid juris. This seemed a useful distinction since the first problem consists in evaluating the claim to knowledge that is implicit in the anthropologist’s discourse. The problem is not a cognitive or a psychological one, since it is not about whether knowing another culture is empirically possible.15 The problem is rather epistemological, which is to say, political. It speaks to the properly transcendental question of how to decide on the legitimacy of discourses that enter into a relationship of knowledge. In particular, it speaks to the manner in which relations of order are established between these discourses—these relations are in no way innate, after all, and nor are their enunciative poles. No one is born an anthropologist and, as curious as it may seem, even less a native.

At the limit

In recent years, we anthropologists have worried greatly about the identity, as well as the destiny, of our discipline: what it is, if it continues at all, what it should be, if it has the right to exist, what its proper object might be, its method, its mission, and so on (see, for example, Moore 1999). Let us focus on the question of our discipline’s object, since the rest of these questions turn on it. Is it culture, as in the North American tradition of anthropology? Or is it social organization, as it was for the British? Or maybe human nature, as per the French approach? I think that the appropriate response is: all and none of the above. Culture, society, and nature— much of a muchness: such notions do not so much designate anthropology’s object or topic, but rather point to its basic problem, namely that it cannot adopt (Latour 1991: 109–10, 130) any such themes as its own, if in doing so it neglects to take into account the one “anthropological tradition” that counts most, namely, that of the native.

If we must start somewhere, let us be British and acknowledge from the outset that the anthropologist’s privileged domain of concern is human sociality, that is, what we are happy to call “social relations.” We could then also suggest that “culture,” for example, cannot exist beyond its actualization in such relations.16 And we could add, importantly, that these relations vary in time and space. So, if culture does not exist beyond its relational expression, then relational variation is also cultural variation. Or, to put it another way, “culture” is the word anthropologists use to talk about relational variation.

Thinking, then, about relational variation: would such a notion not willy-nilly imply some kind of subject—an invariable substrate to which relational variations would stand as predicates? This seems to be the ever-latent question, insistent always on some sort of immediate evidence. But this question is, above all, badly formed. For what varies, most crucially, is not the content of relations, but rather the very idea of a relation, i.e., what counts as a relation in this or that culture. It is not the relations that vary, but rather the variations that are related. And if this is so, then what is imagined as the substrate of variation, namely “human nature”—to turn to the darling concept of the third great anthropological tradition—would completely change its function, or better, it would stop being a substance and would become a true function. Nature would stop being a type of highest common denominator of cultures (the minimum high, so to speak, of a humanitas minima), or a sort of backdrop of similarity generated by cancelling differences so as to arrive at a constant subject—a stable referent capable of emitting variable cultural meanings (as if differences were not just as natural!). Instead, human nature could be conceived as something like a minimum common multiple of difference—bigger than cultures, rather than smaller—or something like the partial integer of the different relational configurations we call “cultures.”17 The “minimum,” in this case, is the multiplicity that is common to humans—humanitas multiplex. Thus conceived, nature would no longer be a self-same substance situated within some naturally privileged place (such as the brain, for example). Instead, nature itself would be accorded the status of a differential relation, best placed between the terms that it “naturalizes.” It would consist in the set of transformations that are necessary in order to describe variations between different known relational configurations. Or, to use yet another image, nature would become a pure limit—but not in the geometric sense of limitation, understood as a perimeter or a term that constrains and defines some substantive form (recalling the idea of “mental enclosures” [enceintes mentales], which is ever present in the anthropological vocabulary), but rather in the mathematical sense, as the point to which a series or relation converges: a tension-limit, as opposed to a contour-limit.18 In such a case, human nature would be the theoretical operation of a “passage to the limit,” indicating what human beings are capable of virtually, rather than a limitation that consigns them actually to being nothing other than themselves.19 If culture is a system of differences, as the structuralists liked to say, then so is nature: differences of differences.

The theme of the contour-limit (so characteristically Kantian, and ever-present in the discipline’s imaginary) is at its most conspicuous when it provides a limiting horizon in the guise of so-called human nature, as is the case with naturalist-universalist approaches such as sociobiology, evolutionary psychology and, to a large degree, in structuralism itself. But it is also present in discourses about human cultures, where it renders clear the limitations—if I may call them thus—of the classic cultural-relativist position. This recalls the notion enshrined in Evans-Pritchard’s phrase about Zande witchcraft—a Zande man “cannot think that his thought is wrong” (1976: 109)—or the current anthropological image of culture as a prosthesis of the eye (or classificatory sieve) that only permits one to “see things” in a certain way (or which hides certain aspects of reality); or even, to cite a more recent example, the fishbowl metaphor, which encloses each historical period (Veyne 1983).20 Whether in regard to nature or cultures, the theme appears equally “limited.” If we were to be perverse, we could say that its strategic neutrality, its copresence in the otherwise opposed camps of universalism and relativism, is a good indication that the notion of a “mental enclosure” is one of the mental enclosures that most characterize our common historical “fishbowl.” In any case, it demonstrates that the supposed opposition between naturalist universalism and culturalist relativism is, at least, very relative (and perfectly cultural), for it can be summed up as a matter of choosing the dimensions of the bowl, or of the size of the cell where we are imprisoned: should it include all of human kind ecumenically, or should it be made to order for each culture? Or perhaps we might have one great “natural” prison with different cultural cell-blocks, some of them with slightly more spacious cells than others?21

Thus understood, anthropology’s object would be the variation of social relations. Not of social relations as a distinct ontological province, but of all possible phenomena taken as social relations, or as implying social relations: of all relations, in short, as social. This, however, would require adopting a perspective that is not completely dominated by the Western doctrine of social relations—a perspective that would be ready to accept that handling all relations as social could lead to a radical reconceptualization of what “the social” might be. Let us say, then, that anthropology distinguishes itself from other discourses on human sociality, not by holding any firm doctrine about the nature of social relations, but, on the contrary, by maintaining only a vaggue initial idea of what a relation might be. For its characteristic problem consists less in determining which social relations constitute its object, and more in asking itself what its object constitutes as a social relation—what a social relation is in that object’s terms, or better, in terms that can be formulated through the (social, naturally, and constitutive) relationship between the “anthropologist” and the “native.”

From conception to concept

Would all of this not simply suggest that the point of view defended here, and exemplified in my work on Amerindian perspectivism (Viveiros de Castro 1998), is “the native’s point of view,” which anthropologists have professed to be grasping for some time now? To be sure, there is certainly nothing particularly original in the point of view that I am adopting here. The only rightful claim to originality belongs to the indigenous point of view itself, and not to my commentary on it. Still, when it comes to the question of whether the object of anthropology ought to be the native’s point of view, the response must be both “yes” and “no.” “Yes” (certainly!), because my problem in the above-cited article was to discover what a “point of view” is for the native. In other words, what concept of a point of view do Amazonian cultures enunciate—what is the native point of view on the point of view? The answer is “no,” on the other hand, because the native concept of a point of view does not coincide with the concept of “the native’s of point of view.” After all, my point of view cannot be the native’s own, but only that of my relation with it. This involves an essentially fictional dimension, since it implies making two entirely heterogeneous points of view resonate with each other.

My article on perspectivism, then, was at once a thought experiment and an exercise in anthropological fiction. Here, however, the expression “thought experiment” should not be understood in the usual way, as an attempt to think oneself into another form of experience but rather as a manner of experiencing for oneself an other’s form of thought. It is not a matter of imagining a form of experience, if you like, but of experiencing a form of imagination.22 The experience, in this case, is my own—as ethnographer, as well as reader of the ethnological literature about indigenous Amazonia—and the experiment is a fiction that is controlled by that experience. In other words, the fiction that is involved is anthropological, but the anthropology that it produces is not fictional!

What does such a fiction consist in? It consists in taking indigenous ideas as concepts, and following through on 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. And note that treating these ideas as concepts does not mean that, objectively or actually speaking, they are something else. Individual cognitions, collective representations, propositional attitudes, cosmological beliefs, unconscious schemata, embodied dispositions and so forth: these are the kinds of theoretical fictions I choose not to heed here.

Thus, the type of work for which I am advocating is neither a study of “primitive mentality” (supposing such a notion might still make sense at all), nor an analysis of the natives’ “cognitive processes” (supposing these were accessible, given the current state of psychological and ethnographic knowledge). My object is less the indigenous manner of thinking than its objects, the possible world that its concepts project. Nor is it a matter of reducing anthropology to a series of ethno-sociological essays about worldviews. This is because, in the first place, no world that is ready to be viewed exists—no world that would precede one’s view of it, or precede even the distinction between the visible (or thinkable) and the invisible (or presumed), which provides the coordinates for this manner of thinking. Second, because treating ideas as concepts involves refusing attempts to explain them in terms of some transcendent notion of context (ecological, economic, political, etc.), opting rather to treat them immanently as problems, i.e., placing them in the field of problems in which ideas are implicated. And nor is it, finally, a matter of proposing an interpretation of Amerindian thought, but rather one of carrying out an experiment with it, and thus also with our own. In Roy Wagner’s words: “every understanding of another culture is an experiment with one’s own” (1981: 12).

To treat indigenous ideas as concepts is to take an antipsychologizing stance, since what is at stake here is a de jure image of thought, irreducible to empirical cognition, or at least to the empirical analysis of cognition psychologists provide. The domain of concepts does not coincide with subjects’ cognitive faculties or internal states: concepts are intellectual objects or events, not mental states or attributes. They certainly “cross the mind,” as the English expression has it, but they do not stay there and, above all, they are not to be found there readymade. They are invented. To be clear: I am not suggesting that Amerindians “cognize” differently to us, or that their “mental” categories are different to those of any other human being. Certainly, it is not a matter of imagining them as instantiating some peculiar form of neurophysiology that processes difference in a different way. For my own part, I am inclined to think that Amerindians think exactly “like us.” But I also think that what they think, that is, the concepts that they deploy, the “descriptions” that they produce, are very different to our own—and thus that the world described by these concepts is very different to ours.23 And as far as the Amerindians are concerned (if my analyses concerning perspectivism are correct), I think that they think that all humans, and aside from them many other nonhuman subjects, think exactly “like them”—this being precisely the reason for subjects’ divergences of perspective; that is, the very opposite of a universal convergence of reference.

The notion of a concept implies an image of thought as something other than cognition or a system of representations. What interests me in Amerindian thought, then, is neither local knowledge and its more or less accurate representations of reality—the so-called “indigenous knowledge” that is currently the focus of so much attention in the global market of representations—nor indigenous cognition, its mental categories, and how representative they are of the species’ capaci-ties—this being the main concern of human psychology as a “natural science.” Neither representations, whether individual or collective, rational or (“apparently”) irrational, which might partially express states of affairs prior and exterior to themselves; nor categories and cognitive processes, whether universal or particular, innate or acquired, which manifest the properties of some thing of the world, whether it be the mind or society. My objects are indigenous concepts, the worlds they constitute (worlds that thus express them), the virtual background from which they emerge and which they presuppose. In short, my objects are the concepts, which is to say the ideas and problems of indigenous “reason,” rather than indigenous categories of “understanding.”

It should be clear by now that the notion of concept has a very specific meaning here. Treating indigenous ideas as concepts means taking them as containing a properly philosophical significance, or as being potentially capable of philosophical use.

It might be said that this is an irresponsible decision, since neither the Amerindians nor even (and this must be stressed) the present author are philosophers. One might wonder, for example, how to apply the notion of a concept to a form of thought that has, apparently, never found it necessary to dwell on itself, and which would rather evoke the fluent and variegated schematism of symbols, images, and collective representations than the rigorous architecture of conceptual reason. Is there not, after all, any sign of the well-known historical and psychological abyss, or “decisive rupture,” between a panhuman mythical imagination and the universe of Hellenic-occidental rationalism (Vernant [1966] 1996: 229); between the sign’s bricolage and the concept’s engineering (Lévi-Strauss 1962); between the paradigmatic transcendence of the Figure and the syntagmatic immanence of the Concept (Deleuze and Guattari 1991); between an imagistic intellectual economy and a doctrinal one (Whitehouse 2000)? On all of this, which is more or less a direct legacy from Hegel, I have my doubts. I insist instead on talking about concepts, and this for a number of reasons. And the first among them, on which I shall comment here, stems from the decision to place native ideas on the same footing as anthropological ones.

As stated above, the experiment I am proposing posits an equivalence de jure between the anthropologist’s and the native’s discourses, taking them as mutually constitutive of each other, since they emerge as such when they enter into a knowledge relation with one another. Anthropological concepts actualize this relation and therefore can only be construed as being completely relational, both in the manner of their expression and in their content. They are to be construed neither as truthful reflections of the native’s culture (the positivist dream), nor as illusory projections of the anthropologist’s culture (the constructionist nightmare). They reflect, rather, a certain relation of intelligibility between two cultures; a relation that produces the two cultures in question by back projection, so to speak, as the “motivation” of the anthropological concepts. As such, anthropological concepts perform a double dislocation: they are vectors that always point in the other direction, transcontextual interfaces that function to represent, in the diplomatic sense of the term, the other in one’s own terms (that is, in the other’s other’s own terms)—both ways.

In short, anthropological concepts are relative because they are relational, and they are relational because their role is to relate. Indeed, their relational origin and function is marked by the habit of designating them with alien-sounding words: mana, totem, kula, potlatch, tabu, gumsa/gumlao…. Other concepts, no less authentic, carry an etymological signature that evokes analogies between the cultural tradition from which they emerged and the traditions that are their object: gift, sacrifice, kinship, person…. Yet other (and just as legitimate) ones constitute terminological inventions the role of which is to generalize the conceptual mechanisms of the people being studied—animism, segmentary opposition, restricted exchange, schismogenesis …—or, inversely, and more problematically, terms that are deployed in order to inject notions that are already diffuse in our own tradition into the interior of a specific theoretical economy—incest taboo, gender, symbol, culture—so as to universalize them.24

We can thus see that numerous concepts, problems, entities, and agents that are to be found in anthropological theories emerge through the imaginative efforts of societies on which the discipline hopes to shed light. Might one not say, then, that anthropology’s originality lies in just this synergy, between conceptions and practices pertaining to two worlds—the “subject’s” and the “object’s” respectively? Recognizing this might help, among other things, to mitigate our complex of inferiority in relation to the “natural sciences.” As observed by Latour:

The description of kula is on a par with that of black holes. The complex systems of alliances are as imaginative as the complex scenarios conceived for selfish genes. Understanding the theology of Australian Aborigines is as important as charting the great undersea rifts. The Trobriand land tenure system is as interesting a scientific objective as polar icecap drilling. If we talk about what matters in a definition of science—innovation in the agencies that furnish our world—anthropology might well be close to the top of the disciplinary pecking order. (1996a: 5)25

In this passage an analogy is made between indigenous concepts and the objects of the so-called natural sciences. This is one possible, and even necessary, perspective: it should be possible to produce a scientific description of indigenous ideas and practices, as if they were things of the world, or better, so that they can become things of the world. (One must not forget that for Latour the objects of science are anything but “objective” and indifferent entities, patiently awaiting description.) Another strategy would be to compare indigenous conceptions with scientific theories, as suggested by Horton in his “similarity thesis” (1993: 348–54), which anticipates some aspects of Latour’s symmetrical anthropology. And yet another is the strategy advocated here. In this connection it is worth noting that anthropology has always been overly obsessed with “Science,” not only in relation to itself (is it a science? can it be? should it be?), but above all—and this is the real problem—in relation to the conceptions of the peoples it studies. The question then becomes whether to disqualify such conceptions as errors, dreams, or illusions, in order then scientifically to explain how and why the “others” cannot explain them(selves) scientifically; or to promote native conceptions as more or less continuous with science, fruits of the same desire to know, which unites all humans. Horton’s similarity thesis and Lévi-Strauss’ science of the concrete are two examples (Latour 1991: 133–34). And indeed, the image of science may well be considered a kind of gold standard of thought, at least as far as our own intellectual tradition is concerned. It is not, however, the only or necessarily the best terrain on which to establish fruitful epistemo-political relations with the intellectual activity of peoples who have no truck with our much-cherished cult(ure) of Reason.

So, we might imagine a form of analogy different to the one suggested by Latour, or a manner of similarity other than Horton’s. A form of analogy in which, instead of taking indigenous conceptions as entities similar to black holes or tectonic faults, we took them as being of a kind with the cogito or the monad. We could thus say, to paraphrase the previous citation, that the Melanesian concept of the “dividual” person (Strathern 1988) is as imaginative as Locke’s possessive individualism; that understanding the “Amerindian philosophy of chieftainship” (Clastres [1962] 1974) is as important as commenting on Hegel’s doctrine of the State; that Maori cosmology is equivalent to the Eleatic paradoxes and Kantian antinomies (Schrempp 1992); that Amazonian perspectivism presents a philosophical challenge of the same order as Leibniz’s system…. And when it comes to what matters most in a given philosophical elaboration, namely its capacity to create new concepts, then without any desire to take the place of philosophy, anthropology can be recognized as a formidable philosophical instrument in its own right, capable of broadening a little the otherwise rather ethnocentric horizons of our philosophy—and, in passing, ridding us of so-called “philosophical” anthropology too. In Tim Ingold’s (1992: 696) pithy phrase: “anthropology is philosophy with the people in.” By “people,” Ingold means “ordinary people” (ibid.), to be sure. He is also playing, however, with the term’s connotation of “the people” or, more likely yet, “peoples.” A philosophy, then, with other peoples in it: the possibility of a philosophical endeavor that places itself in relation to the nonphilo-sophy—simply, the life—of other peoples on the planet, beyond our own.26 Not only the common people, but above all with uncommon people, those that are beyond our sphere of “communication.” If in “real” philosophy imaginary savages are altogether abundant, the geo-philosophy proposed by anthropology conducts an “imaginary” philosophy with real “savages.” “Real toads in imaginary gardens,” as the poet Marianne Moore has it.

Note the significant displacement involved in the above paraphrase. It is no longer (only) a question of the kula’s anthropological description (as a form of Melanesian sociality), but (also) of the kula as a peculiarly Melanesian description (of “sociality” as a form of anthropology). Similarly, it would still be necessary to understand “Australian theology,” but now as constituting a form of understanding in its own right, just as, to give another example, complex alliance or land tenure systems can be seen as exercises of an indigenous sociological imagination. Clearly, it will always be necessary to describe the kula as a description, to understand Aboriginal religion as an understanding, and to form images of the indigenous imagination. Doing so is a matter of transforming conceptions into concepts, extracting the latter and returning them to the former. And a concept is to be understood here as a complex relation between conceptions—a manner of activating preconceptual intuitions. In the case of anthropology, the conceptions that enter into this kind of relation include, before all else, the anthropologist’s and the native’s—a relation of relations. Native concepts are the anthropologist’s concepts. or so we may suppose.

Neither explain, nor interpret: Multiply and experiment!

In The invention of culture, Roy Wagner was one of the first anthropologists to draw out the radical consequences of the idea that anthropologist and native can be treated on an equal footing due to their common cultural condition. From the fact that the anthropologist’s attempt to approach another culture can only be conducted through terms taken from his own, Wagner concludes that anthropological knowledge is defined by its “relative objectivity” (1981: 2). At issue here is not a deficient objectivity, that is, subjective or partial, but an intrinsically relational objectivity, as can be gathered from what follows:

The idea of culture …places the researcher in a position of equality with his subjects: each “belongs to a culture.” Because every culture can be understood as a specific manifestation …of the phenomenon of man, and because no infallible method has ever been discovered for “grading” different cultures and sorting them into their natural types, we assume that every culture, as such, is equivalent to every other one. This assumption is called “cultural relativity.” …The combination of these two implications of the idea of culture, the fact that we ourselves belong to a culture (relative objectivity) and that we must assume all cultures to be equivalent (cultural relativity), leads to a general proposition concerning the study of culture. As the repetition of the stem “relative” suggests, the understanding of another culture involves the relationship between two varieties of the human phenomenon; it aims at the creation of an intellectual relation between them, as understanding that includes both of them. The idea of “relationship” is important here because it is more appropriate to the bringing together of two equivalent entities, or viewpoints, than notions like “analysis” or “examination,” with their pretensions of absolute objectivity. (Wagner 1981: 2–3)

Or as Deleuze might say, it is not a matter of affirming the relativity of the true, but rather of affirming the truth of the relative. It is worth observing that Wagner associates the notion of a relation to that of a point of view (the terms that are related are points of view), and that the idea of the truth of the relative defines what Deleuze calls “perspectivism.” Whether it be Leibniz’s or Nietzsche’s, or, equally, Tukanoan or Jurunoan, perspectivism is not relativism, that is, the affirmation of the relativity of truth, but relationalism, through which one can affirm that the truth of the relative is the relation.

I asked what would happen if we refuse the epistemological advantage of the anthropologist’s discourse over that of the native: what if we took knowledge relations as modifying, reciprocally, the terms they relate or, rather, actualize? This is the same as asking: what happens when native thought is taken seriously? What happens when the anthropologist’s objective ceases to be that of explaining, interpreting, contextualizing, or rationalizing native thought, but instead begins to deploy it, drawing out its consequences, and verifying the effects that it can produce on our own thinking? What is it to think native thought? I say “think,” here, without worrying whether what we think (namely, others’ thoughts) is “apparently irrational,”27 or, even worse, rational by nature.28 At issue is a manner of thinking that does not think itself from within the coordinates provided by these alternatives—a form of thinking entirely alien to this kind of game.

For a start, taking native thought seriously is to refuse to neutralize it. For example, one ought to put in parentheses all questions of whether and how native thinking illustrates universal processes of human cognition; whether it can be explained as a result of particular modes of the social transmission of knowledge; or as the expression of a particular cultural world; or whether its functional role is to validate a particular distribution of political power. All such forms of neutralizing foreign thought should be resisted. suspending such questions or, at least, refusing to enclose anthropology within them, one might opt rather, say, to think other thought simply (so to speak) as an actualization of as yet unsuspected virtualities of thinking.

Would taking the Amerindians seriously mean “believing” in what they say, taking their thought as an expression of certain truths about the world? Absolutely not; here is yet another of those questions that are famously “badly put.” Believing or not believing in native thought implies first imagining it as a system of beliefs. But problems that are properly anthropological should never be put either in the psychologistic terms of belief, or in the logicist terms of truth-value. It is not a matter of taking native thought as an expression of opinion (the only possible object for belief and disbelief) or as a set of propositions (the only possible objects for truth judgments). We know the mess anthropology made when it decided to define natives’ relationship to their own discourse in terms of belief: culture instantly becomes a kind of dogmatic theology. And it is just as bad to shift from “propositional attitudes” to their objects, treating native discourse as a repository of opinions or a set of propositions: culture turns into an epistemic teratology—error, illusion, madness, ideology.29 As Latour observes (1996b: 15), “belief is not a mental state, but an effect of the relation between peoples”—and it is precisely that effect that I do not mean to produce.

Take animism, for example—about which I have written previously (Viveiros de Castro 1998). Lalande’s Vocabulary, which is hardly incompatible with more recent psycho-anthropological studies on this topic, defines “animism” in just these terms: as a “mental state.” But Amerindian animism is anything but: it is an image of thought that separates de facto from de jure, that which pertains to thought by right from what contingently refers to a state of affairs; it is, more specifically, an interpretive convention (Strathern 1999a: 239) that, formally speaking, involves personifying objects of knowledge, thus turning thought into an activity—an effect of the (“social”) relation between the thinker and what she or he thinks. After all, would it be appropriate to imagine, say, legal positivism and jus-naturalism as mental states? The same is (not) the case for Amazonian animism: it is not a mental state of individual subjects, but rather a transindividual intellectual disposition that, if anything, deploys the “mental states” of different beings in the world as one of its objects. It is not the native’s mental condition, but a “theory of the mind” applied by the native. Indeed, it is a manner of resolving—or better, dissolving—the eminently philosophical problem of “other minds.”

If it is not a matter of describing American indigenous thought as a set of beliefs, then nor is it a question of relating to it via some prior notion of belief that might lend it its credibility—either by benevolently pointing to its allegorical “grain of truth” (a social allegory, for the Durkheimians, or a natural one, for the cultural materialists) or, even worse, by imagining that it provides access to the intimate and final essence of things, acting as a portal into some kind of immanent esoteric science. “An anthropology that …reduces meaning to belief, dogma and certainty, is forced into the trap of having to believe either the native meanings or our own” (Wagner 1981: 30). But the plane of meaning is not populated by psychological beliefs or logical propositions, and the “truth” of Amerindian thought is not, well, granular! Neither a form of doxa, nor of logic—neither opinion, nor proposition—native thought is taken here as an activity of symbolization or meaning practice: a self-referential or tautegorical mechanism for the production of concepts, that is, “symbols that stand for themselves” (Wagner 1986).

The refusal to pose the question in terms of belief seems to me a critical anthropological decision. To emphasize this, we might reinvoke the Deleuzian Other: the Other is an expression of a possible world; but in the course of social interaction, this world must always be actualized by a Self: the implication of the possible in an Other is explicated by me. This means that the possible goes through a process of verification that entropically dissipates its structure. When I develop the world expressed by an Other, it is so as to validate it as real and enter into it, or to falsify it as unreal: the “explication” thus introduces the element of belief. To describe this process, Deleuze indicated the limiting condition that allowed him to determine the concept of the Other:

These relations of development, which form our commonalities as well as our disagreements with the other, dissolve their structure and reduce it either to the status of an object or to the status of a subject. That is why, in order to grasp the other as such, we felt right to insist upon special conditions of experience, however artificial: the moment at which the expressed still has no existence (for us) beyond that which expresses it— the Other as the expression of a possible world. (1969: 335) [Emphasis removed in author’s translation. —Trans.]

And he concludes by recalling a maxim that is fundamental to his thinking: “The rule invoked earlier—not to explain too much—meant, above all, not to explain oneself too much with the other, not to explain the other too much, but to maintain one’s implicit values and multiply one’s world by populating it with all those things expressed that do not exist outside of their expressions” (ibid.). Anthropology can make good use of this: maintaining an Other’s values implicit does not mean celebrating some numinous mystery that they might hide but rather amounts to refusing to actualize the possibilities expressed by indigenous thought— opting to sustain them as possible indefinitely, neither dismissing them as the fantasies of others, nor by fantasizing ourselves that they may gain their reality for us. The anthropological experiment, then, involves formally interiorizing the “special and artificial conditions” that Deleuze discusses: the moment in which the world of the Other does not exist beyond its expression transforms itself into an abiding condition, that is, a condition internal to the anthropological relation, which renders this possibility virtual.30 Anthropology’s constitutive role (its task de jure), then, is not that of explaining the world of the other, but rather of multiplying our world, “populating it with all those things expressed that do not exist outside of their expressions.”

Of pigs and bodies

Rendering native possibilities as virtualities is the same as treating native ideas as concepts. Two examples.

Amerindians’ pigs

In American ethnography one often comes across the idea that, for Amerindians, animals are human. This formulation condenses a nebula of subtly varied conceptions, which we shall not elaborate here: that not all animals are humans, and they’re not the only ones (plants etc. may also be human); that animals are not humans at all times; that they were human but no longer are; that they become human when they’re out of view; that they only think that they’re human; that they see themselves as human; that they have a human soul beneath an animal body; that they are people like humans are, but are not exactly human like people are; and so on. Aside from all that, “animal” and “human” are equivocal translations of certain indigenous words—lest it be forgotten, we are faced with hundreds of different languages, and in most of them the copula is not commonly marked by a verb. But no matter, for present purposes. Let us suppose that statements such as “animals are humans” or “certain animals are people” make sense for a certain indigenous group, and that their meaning is not merely “metaphorical”—as much sense, let us say (though not exactly the same kind of sense), as the apparently inverse (and no longer scandalous) affirmation—”humans are animals”—makes to us. Let us suppose, then, that the first statement makes sense to, for example, the Ese Eja of the Bolivian Amazon: “The affirmation, that I frequently heard, that ‘all the animals are Ese Eja’” (Alexiades 1999: 179).31

Right then. Isabella Lepri, an anthropology student who, coincidentally, at the time was working with the same Ese Eja, asked me whether I believed that the peccaries are humans, like the Amerindians say they are. I answered that I did not—doing so because I suspected (without any particular reason) that she believed that, if the Amerindians say such a thing, then it must be true. I added, perversely and rather untruthfully, that I only “believed” in atoms and genes, the theory of relativity and the evolution of the species, class war, and the logic of capital, in short, in that type of thing; but that, as an anthropologist, I took the idea that peccaries are humans perfectly seriously. She challenged me: “How can you maintain that you take what the Amerindians say seriously? Isn’t that just a way of being polite with your informants? How can you take them seriously if you only pretend to believe in what they say?”

To be sure, this intimation of hypocrisy obliged me to reflect. I am convinced that Isabella’s question is absolutely crucial; that all anthropology deserving of the name must answer it; and that it is not at all easy to do so very well.

Naturally, one possible response is that contained in Lévi-Strauss’ cutting remark on Ricoeur’s mythical (and mystical) hermeneutics: “It is necessary to choose which side you are on. Myths do not say anything capable of instructing us on the order of things, on the nature of reality, the origin of man or his destiny” (1971: 571). Instead, the author continues, myths do teach us much about the societies from which they originate and, above all, about certain fundamental (and universal) operative modes belonging to the human mind (Lévi-Strauss 1971: 571). One can thus oppose the referential vacuity of myths to their diagnostic richness: to say that peccaries are human does not “say” anything to us about the peccaries, but is highly telling about the humans who say it.

The solution is not specific to Lévi-Strauss—ever since Durkheim or the Victorian intellectualists it has been a standard anthropological posture. In our days, for example, much of so-called cognitive anthropology can be seen as a systematic elaboration of this attitude, which consists in reducing indigenous discourse to a set of propositions, selecting those that are false (or alternatively, “empty”) and producing an explanation of why humans believe in them, given that they are false or empty. One such explanation, to continue with the examples, would be to conclude that such propositions are really forms of citation—statements to be placed between implicit quotation marks (Sperber 1974, 1982)—and therefore do not refer to the world, but rather to the relation between the natives and their discourse. This relation is, once again, the core theme for so-called “symbolic” anthropologies, of the semantic or pragmatic type: statements such as the one about peccaries, “in reality,” say something about society (or do something to it), not about what they are about. They teach us nothing about the order of things and the nature of reality, however, neither for us nor for the Amerindians. To take an affirmation such as “peccaries are humans” seriously, in this case, would consist in showing how certain humans can take it seriously and even believe in it, without showing themselves to be irrational—and, naturally, without the peccaries showing themselves to be human. The world is saved: the peccaries are saved, the natives are saved and, above all, so is the anthropologist.

This solution does not satisfy me. In fact, it profoundly bothers me. It seems to imply that to take Amerindians seriously, when they affirm things such as “peccaries are humans,” is precisely not to believe in what they say, since if we did we would not be taking ourselves seriously. Another way out is needed. As I do not have either the space or, above all (and evidently), the ability to go over the vast philosophical literature that exists on the grammar of belief, certainty, proposition-al attitudes, et cetera, in what follows I will simply present certain considerations that have emerged intuitively, more than reflexively, through my experience as an ethnographer.

I am an anthropologist, not a swinologist. Peccaries (or, as another anthropologist once said about the Nuer, cows) are of no special interest to me, humans are. But peccaries are of enormous interest to those humans who say that peccaries are human. As a result, the idea that the peccaries are human interests me also, because it “says” something about the humans that say this. But not because it says something that these humans are not capable of saying by themselves, and rather because in it the humans in question are saying not only something about the peccaries, but also about what it is to be “human.” (Why should the Nuer, for example, not say in their turn that cattle are human?) If the statement on the peccaries’ humanity definitely reveals something about the human mind (to the anthropologist), it also does more than that (for the Amerindians): it affirms something about the concept of humanity. It affirms, among other things, that the notion of “human mind” and the indigenous concept of sociality include the peccaries in their extensions—and this radically modifies these concepts’ intension in relation to our own.

The native’s belief or the anthropologist’s disbelief has nothing to do with this. To ask (oneself) whether the anthropologist ought to believe the native is a category mistake equivalent to wondering whether the number two is tall or green. These are the first elements of my response to Isabella. When an anthropologist hears from his indigenous interlocutor (or reads in an ethnography) such things as “peccaries are human,” the affirmation interests him, no doubt, because he “knows” that peccaries are not human. But this knowledge (which is essentially arbitrary, not to say smugly tautological) ought to stop there: it is only interesting in having awoken the interest of the anthropologist. No more should be asked of it. Above all, it should not be incorporated implicitly in the economy of anthropological commentary, as if it were necessary (or essential) to explain why the Indians believe that peccaries are human whereas in fact they are not. What is the point of asking oneself whether the Indians are right in this respect—do we not already “know” this? What is indeed worth knowing is that to which we do not know the answer, namely what the Indians are saying when they say that peccaries are human.

Such an idea is far from evident. The problem that it creates does not reside in the proposition’s copula, as if “peccary” and “human” were common notions, shared by anthropologist and native, the only difference residing in the bizarre equation between the two terms. We should say in passing that it is perfectly possible for the lexical meaning or semantic interpretation of “peccary” and “human” to be more or less the same for both interlocutors; it is not a translational problem, or a matter of deciding whether we and the Amerindians share the same “natural kinds” (perhaps we do…). The problem is that the idea that peccaries are human is part of the meaning of the “concepts” of peccary and human in that culture, or better, it is just this idea that constitutes the conceptual potency of the statement, providing the concept that determines the manner in which the ideas of peccary and human are to be related. For it is not “first” the peccaries and the humans each in their own place, and “then” the idea that the peccaries are humans: on the contrary, peccaries, humans and their relation are all given together.32

The intellectual narrowness that afflicts anthropology, in such cases, consists in reducing the notions of peccary and human merely to a proposition’s independent variables, when they should be seen—if we want to take Amerindians seriously—as inseparable variations of a single concept. To say that peccaries are humans, as I have already observed, is not simply to say something about peccaries, as if “human” were a passive and inert predicate (for example, the genus that includes the species of peccaries). Nor is it simply a matter of giving a verbal definition of “peccary,” much as a statement of the type “‘bass’ is (the name of) a fish.” To say that peccaries are human is to say something about peccaries and about humans, something about what the human can be: if peccaries have humanity as a potential, then might humans not have a peccary-potential? In effect, if peccaries can be conceived as humans, then it should be possible to conceive of humans as peccaries: what is it to be human if one is also “peccary,” and what is it to be peccary if one is “human”? What are the consequences of this? What concept can be extracted from a statement like “peccaries are human”? How can we transform the conception expressed in a proposition like this into a concept? That is the true question.

Hence, when told by his indigenous interlocutors (under conditions that must always be specified) that peccaries are human, the anthropologist should ask herself or himself, not whether or not “he believes” that they are, but rather what such an idea could show him about indigenous notions of humanity and “peccar-ity.” What an idea such as this, note, teaches him about these notions and about other things: about relations between him and his interlocutor, the situations in which this statement is produced “spontaneously,” the speech genres and language games in which it fits, et cetera. These other things, however—and I would like to insist on this point—hardly exhaust the statement’s meaning. To reduce the statement into a discourse that only “speaks” of its enunciator is to negate the latter’s intentionality, obliging him to exchange his peccary for our human—a bad deal for a peccary hunter.

Thus understood, it is obvious that the ethnographer has to believe (in the sense of trusting) his interlocutor: the native is not giving the ethnographer an opinion, he is effectively teaching him what peccaries and humans are, explaining how the human is implied in the peccary. Once more, the question should be: what does this idea do? What assemblages can it help constitute? What are its consequences? For example: what is eaten when one eats a peccary, if peccaries are human?

Furthermore: we still need to see if the concept that can be built by way of such statements can be expressed adequately in the “X is Y” form. For it is not so much a matter of predication or attribution but of defining a virtual set of events and series into which the wild pigs of our example can enter: peccaries travel in a pack …they have a leader …they are noisy and aggressive …they appear suddenly and unpredictably …they are bad brothers-in-law …they eat palm fruit …there are myths that say they live in huge underground villages …they are incarnations of the dead …and so forth. It is not a matter of establishing correspondences between peccaries’ and human’s respective attributes—far from it. The peccaries are peccaries and humans, they are humans inasmuch as humans are not peccaries; peccaries imply humans, as an idea, in their very distance from them. Thus, to state that peccaries are human is not to identify them with humans, but rather to differentiate them from themselves—and therefore us from ourselves also.

Previously I stated that the idea of peccaries being human is far from evident: to be sure, no interesting idea is ever evident. This particular idea is not nonevident because it is false or unverifiable (Amerindians have many different ways to verify it), but because it says something nonevident about the world. Peccaries are not evidently humans, they are so nonevidently. Could this mean that the idea is “symbolic,” in the sense given to this adjective by Sperber? I think not. Sperber conceives of indigenous concepts as propositions, and worse, as second-class propositions, “semi-propositional representations” that extend “encyclopedic knowledge” in a nonreferential manner: he seems to identify the self-positive with the referentially void, the virtual with the fictional, immanence with closure…. But one can see “symbolism” differently from Sperber, who takes it as something logical and chronologically posterior to the mind’s encyclopedia or to the semantic capacities it informs: something that marks the limits of true or verifiable knowledge, as well as the point at which this knowledge becomes transformed into an illusion. Indigenous concepts can be called symbolic, but in a very different sense; they are not subpropositional, but superpropositional, as they suppose encyclopedic propositions but define their most vital significance, their meaning or value. It is the encyclopedic propositions that are semiconceptual or subsymbolic, not the other way round. The symbolic is not semi-true, but pre-true, that is, important or relevant: it speaks not to what “is the case,” but to what matters in what is the case, to what is interesting in its being the case. What is a peccary worth? This, literally, is the interesting question.33

Sperber (1982: 173) once wrote, ironically, “profound: another semi-propositional word.” But then it is worth replicating—banal: another word for propositional. In effect, indigenous concepts certainly are profound, as they project a background, a plane of immanence filled with intensities, or, if the reader prefers a Wittgensteinian vocabulary, a Weltbild composed of foundational “pseudo-propositions” that ignore and precede the distinction between true and false, “weaving a net that, once thrown over chaos, can provide it with some type of consistence” (Prado Jr. 1998: 317). This background is a “foundationless base” that is neither rational/reasonable nor irrational/unreasonable, but which “simply is there—much like our own lives” (Prado Jr. 1998: 319).34

Amerindian bodies

My colleague Peter Gow once narrated the following scene to me, which he witnessed during one of his stays among the Piro of the Peruvian Amazon:

A mission teacher in [the village of] Santa Clara was trying to convince a Piro woman to prepare food for her infant child with boiled water. The woman replied: “If we drink boiled water, we catch diarrhea.” The teacher, laughing in mockery at this response, explained that common infant diarrhea is caused precisely by the ingestion of unboiled water. Without being flustered the Piro woman answered: “Perhaps that is true for the people from Lima. But for us, people native to this place, boiled water gives diarrhea. our bodies are different from your bodies” (Gow, personal comm., October 12, 2000).

What can the anthropologist do with the Amerindian woman’s response? Many things. Gow, for example, wove a shrewd commentary on this anecdote:

This simple statement [“our bodies are different”] elegantly captures what Viveiros de Castro (1996) called cosmological perspectivism, or multinaturalism: what distinguishes the different types of people are their bodies, not their cultures. However, it should be noted that this example of cosmological perspectivism was not obtained in the course of an esoteric discussion about the occult world of spirits, but during a conversation about eminently practical concerns: what causes diarrhea in children? It would be tempting to see the positions of the teacher and of the Piro woman as representing two distinct cosmologies, multicul-turalism and multinaturalism, and imagining the conversation as a clash of cosmologies or cultures. I believe that this would be a mistake. Both cosmologies/cultures have been in contact for some time, and their imbrication precedes the ontogenetic processes through which the teacher and the Piro woman came to formulate them as being self-evident. But above all such an interpretation would translate the dialogue in the general terms of one of the parts involved, namely, multiculturalism. The coordinates for the Piro woman’s position would be systematically violated by the analysis. of course, this does not mean that I believe that children should drink unboiled water. But it does mean that the ethnographic analysis cannot go forward if the general meaning of such a meeting has been decided from the word go.35

I concur with much of this. The anecdote told by Gow is certainly a splendid illustration of the irreducible divergence between what I have called “multicul-turalism” and “multinaturalism,” particularly as it stems from a banal everyday incident. But Gow’s analysis does not seem to be the only possible one. Thus, on the question of the conversation’s translation into the general terms of one party— in this case, the teacher’s—would it not be equally possible, and above all necessary, to translate it into the general terms of the other? For there is no third position, no absolute vantage point, from which to show the others’ relative character. It is necessary to take sides.

It may be possible to say, for instance, that each of the two women is “cultural-izing” the other in this conversation—that is, attributing the other’s idiocy to her “culture,” while “interpreting” her own position as “natural.” In such a case, one might also say that the Piro woman’s argument about the “body” amounts to a kind concession to the teacher’s assumptions. still, if this were to be so, then note that the concession was not reciprocated. The Piro woman may have agreed to disagree, but the teacher in no way did the same. The former did not contest the fact that people in the city of Lima should (“maybe”) drink boiled water, while the latter peremptorily refuted the idea that people from the Santa Clara village should not.

The Piro woman’s relativism—a “natural” rather than a “cultural” relativism, it should be noted—could be interpreted with reference to certain hypotheses on the cognitive economy of nonmodern societies, or those without writing, or traditional, et cetera. Take Robin Horton’s (1993: 379ff.) theory, for example. Horton posits what he called “worldview parochialism” as a prime characteristic of these societies: contrary to Western modernity’s rationalized cosmologies’ implicit demand for universality, traditional peoples’ cosmologies seem to be marked by a spirit of great tolerance, although it would be fairer to say that they are altogether indifferent to competing worldviews. The Piro’s apparent relativism would thus not be manifesting the breadth of their views, but much to the contrary their myopia: they remain unconcerned with how things are elsewhere.36

There are a number of good grounds to resist readings such as Horton’s. Among others, one reason is that so-called primitive relativism is not only intercultural, but also intracultural and even “auto-cultural,” and, to boot, expresses neither tolerance nor indifference, but rather an absolute departure from the crypto-theological idea of “culture” as a set of beliefs (Tooker 1992; Viveiros de Castro 1993). The main reason to resist such readings, however, is perfectly prefigured in Gow’s own comments, namely, that the idea of “parochialism” translates the Santa Clara debate into the terms of the teacher’s position, with her natural universalism and (more or less tolerant) cultural particularism. There are many worldviews, but there is only one world—a world in which all children should drink boiled water (if, of course, they find themselves in a place where infant diarrhea is a threat).

Let me propose a different reading. The anecdote on different bodies raises questions as to the possible world that the Piro woman’s judgment might express. A possible world in which human bodies can be different in Lima and in santa Clara—a world in which it is necessary for white and Amerindian bodies to be different. Now, to define this world we need not invent an imaginary world, a world endowed with a different physics or biology, let us say, where the universe is not isotropic and bodies can behave according to different laws in different places. That would be (bad) science fiction. It is rather a matter of finding the real problem that renders possible the world implied in the Piro woman’s reply. The argument that “our bodies are different” does not express an alternative, and naturally erroneous, biological theory, or an imaginarily nonstandard37 objective biology. What the Piro argument manifests is a nonbiological idea of the body, an idea in which the question of infant diarrhea cannot be treated as the object of a biological theory. The argument affirms that our respective “bodies” are different, by which we should understand that Piro and Western concepts (rather than “biologies”) of the body are divergent. The Piro water anecdote does not refer to an other vision of the same body, but another concept of the body—the problem being, precisely, its discrepancy from our own concept, notwithstanding their apparent “homonimy.” Thus, for example, the Piro concept of the body cannot be, as ours is, in the soul or “in the mind,” as a representation of a body that lies beyond it. On the contrary, such a concept could be inscribed in the body itself as a perspective (Viveiros de Castro 1996). So, this would not be the concept taken as a representation of an extra-conceptual body, but the body taken as a perspective internal to the concept: the body as an implication of the very concept of perspective. And if, as Spinoza said, we do not know what a body can do, how much less do we know of what such a body could do. Not to speak of its soul.


Alexiades, Miguel. 1999. “Ethnobotany of the Ese Eja: Plants, health, and change in an Amazonian society.” PhD diss., City University of New York.

Clastres, Pierre. (1962) 1974. “Échange et pouvoir: Philosophie de la Chefferie Indienne.” In La société contre l’Etat: Recherches d’anthropologe politique, 25–42. Paris: Minuit.

Deleuze, Gilles. 1969a. Logique du sens. Paris: Minuit.

———. 1969b. Différence et répétition. Paris: PUF.

———. 1981. “Class on Spinoza.” February 17, 1981. http://www.webdeleuze.com/php/texte.php?cle=38&groupe=Spinoza&langue=2.

———. 1990. Pourparlers. Paris: Minuit.

———. (1979) 1995. Dialogues. Paris: Flammarion.

Deleuze, Gilles and Félix Guattari. 1991. Qu’est-ce que la philosophie? Paris: Minuit.

Detienne, Marcel. 1981. L’invention de la mythologie. Paris: Gallimard.

Evans-Pritchard, Edward E. 1976. Witchcraft, oracles, and magic among the Azande. Edited by Eva Gillies, abridged from 1937. Oxford: Clarendon Press.

Gell, Alfred. 1998. Art and agency: An anthropological theory. Oxford: Clarendon.

———. 1999. The art of anthropology: Essays and diagrams. London: Athlone.

Horton, Robin. 1993. Patterns of thought in Africa and the West: Essays on magic, religion and science. Cambridge: Cambridge University Press.

Ingold, Tim. 1992. “Editorial.” Man 27 (1): 694–97.

———. 2000. The perception of the environment: Essays on livelihood, dwelling and skill. London: Routledge.

Jullien, François. 1989. Procès ou création: Une introduction à la pensée Chinoise. Paris: Seuil.

Jullien, François, and Thierry Marchaisse. 2000. Penser d’un dehors (la Chine): Entretiensd’Extrême Occident. Paris: Seuil.

Lambek, Michael. 1998. “Body and mind in mind, body and mind in body: Some anthropological interventions in a long conversation.” In Bodies and Persons: Comparative Perspectives from Africa and Melanesia, edited by Andrew Strathern and Michael Lambek, 103–22. Cambridge: Cambridge University Press.

Latour, Bruno. 1991. Nous n’avons jamais été modernes. Paris: Découverte.

———. 1996a. “Not the question.” Anthropology Newsletter 37 (3): 1–5.

———. 1996b. Petite réflexion sur le culte moderne des dieux faîtiches. Le Plessis/Robinson: Institut Synthelabo.

———. 1999. Politiques de la nature: Comment faire entrer les sciences en démocratie. Paris: Découverte.

Lévi-Strauss, Claude. 1962. La pensée sauvage. Paris: Plon.

———. 1971. L’homme nu. Paris: Plon.

Lipuma, Edward. 1998. “Modernity and forms of personhood in Melanesia.” In Bodies and persons: Comparative perspectives from Africa and Melanesia, edited by Andrew Strathern and Michael Lambek, 53–79. Cambridge: Cambridge University Press.

Mimica, Jadran. 1991. “The incest passions: An outline of the logic of Iqwaye social organization (part 1).” Oceania 62 (1): 34–58.

Moore, Henrietta. 1999. “Anthropological theory at the turn of the century.” In Anthropological theory today, edited by Henrietta Moore, 1–23. London: Polity Press.

Obeyesekere, Gananath. 1992. The apotheosis of Captain Cook: European mythmaking in the Pacific. Princeton, NJ: Princeton University Press.

Prado Jr., Bento. 1998. “Sur le ‘Plan d’immanence.’” In Gilles Deleuze: Une vie philosophique, edited by Eric Alliez, 305–24. Le Plessis/Robinson: Institut Synthélabo.

Scholte, Bob. 1984. “Reason and culture: The universal and the particular revisited.” American Anthropologist 86 (4): 960–65.

Schrempp, Gregory. 1992. Magical arrows: The Maori, the Greeks, and the folklore of the universe. Madison: University of Wisconsin Press.

Simondon, Gilbert. (1964) 1995. L’individu et sa genèse physico-biologique. Paris: Millon.

Sperber, Dan. 1974. Le symbolisme en général. Paris: Hermann.

———. 1982. Le savoir des anthropologues. Paris: Hermann.

Strathern, Marilyn. 1987. “The limits of auto-anthropology.” In Anthropology at home, edited by Anthony Jackson, 16–37. London: Tavistock.

———. 1988. The gender of the gift: Problems with women and problems with society in Melanesia. Berkeley: University of California Press.

———. 1999a. Property, substance, and effect: Anthropological essays on persons and things. London: Athlone.

———. 1999b. “No limite de uma certa linguagem: Entrevista com Marilyn Strathern.” Mana 5 (2): 157–75.

Tooker, Deborah. 1992. “Identity systems of highland Burma: ‘Belief,’ Akha zan, and a critique of interiorized notions of ethno-religious identity.” Man 27 (4): 799–819.

Vernant, Jean-Pierre. (1966) 1996. “Raisons d’hier et d’aujourd’hui.” In Entre mythe et politique, 229–36. Paris: Seuil.

Veyne, Paul. 1983. Les grecs ont-ils cru à leurs mythes? Paris: Seuil.

viveiros de Castro, Eduardo. 1993. “Le marbre et le myrte: De l’inconstance de l‘âme sauvage.” In Mémoire de la tradition, edited by Aurore Becquelin and Antoinette Molinié, 365–431. Nanterre: Société d’Ethnologie.

———. 1996. “Os pronomes cosmológicos e o perspectivismo Amerindio.” Mana 2 (2): 115–44.

———. 1998. “Cosmological deixis and Amerindian perspectivism.” JRAI 4 (3): 469–88.

———. 1999. “Etnologia Brasileira.” In O que Ler na ciência social Brasileira (1970–1995). Volume I: antropologia, edited by S. Miceli, 109–223. Sâo Paulo: Ed. Sumaré/ANPOCS.

Wagner, Roy. 1981. The invention of culture (2nd ed.). Chicago: University of Chicago Press.

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

Whitehead, Alfred N. (1920) 1964. Concept of nature. Cambridge: Cambridge University Press.

Whitehouse, Harvey. 2000. Arguments and icons: Divergent modes of religiosity. Oxford: Oxford University Press.

Wittgenstein, Ludwig. (1930–48) 1982. Remarques sur le Rameau d’Orde Frazer. Translated by J. Bouveresse. Paris: L’Age d’Homme.


Eduardo VIVEIROS DE CASTRO is Professor of Social Anthropology at the National Museum, Universidade Federal do Rio de Janeiro. He has been Simon Bolivar Chair of Latin American Studies at Cambridge University (1997–98) and Directeur de recherches at the C.N.R.S. (2000–2001). His publications include From the enemy’s point of view (Chicago, 1992), Métaphysiques cannibales (PUF, 2009), The inconstancy of the Indian soul (Prickly Paradigm, 2011), and “Cosmo-logical perspectivism in Amazonia and elsewhere” (HAU Masterclass Series, 2012).


Editor’s note: This is a translation of Eduardo Viveiros de Castro’s (2002) “O nativo relativo,” published in the Brazilian journal Mana 8 (1): 113–48. We are grateful to Mana and Eduardo Viveiros de Castro for granting us the permission to publish this translation. We also thank the translators, Julia Sauma and Martin Holbraad. —Ed.

1. The original article was prefaced by the author with the following preamble: “The pages that follow have been adapted from the introductory remarks of a book, currently in preparation, in which I develop ethnographic analyses that have been sketched out in earlier work. The main one is an article published in Mana, ‘Cosmological Deixis and Amerindian Perspectivism’ (Viveiros de Castro 1996 [this appeared in English in JRAI in 1998]), whose metatheoretical premises, as it were, are rendered explicit in the present work. While the text presented here requires no previous familiarity with that earlier work, the reader may bear in mind that such notions as ‘perspective’ and ‘point of view,’ as well as the idea of ‘indigenous thought,’ are elaborated there also.” —Trans.

2. The use of the masculine is arbitrary.

3. The fact that, canonically as well as literally speaking, anthropological discourse takes the form of texts has a host of implications, which cannot be explored here, though the topic has received exhaustive attention in recent currents of auto-anthropological reflection. The same can be said of the fact that native discourse is, generally, not a text, as well as of the fact that it is often treated as if it were.

4. “Knowledge is not a connection between a subject-substance and an object-substance, but rather a relation between two relations, one located in the domain of the object and the other in the domain of the subject; …the relation between two relations is a relation itself” (Simondon [1964] 1995: 81; translation, emphases removed). I translated the word rapport, which Gilbert Simondon distinguishes from relation, as “connection”: “we can call a relation the disposition of the elements in a system, which is beyond the spirit’s simple and arbitrary target, and reserve the term connection for an arbitrary and fortuitous relation …the relation would be a connection that is as real and important as the terms themselves; consequently, we could say that the true relation between two terms is actually equivalent to the connection between three terms” (Simondon 1995: 66; translation).

5. For an analysis of the relational assumptions of this knowledge effect, see Strathern (1987). The author argues that the native’s relation with his discourse is not, in principle, the same as the anthropologist’s relation with his own discourse, and that this difference at once conditions the relation between the two discourses and imposes limits to the whole auto-anthropological enterprise.

6. We are all natives, but no one is native all the time. As Lambek (1998: 113) remarks in a comment about the notion of habitus and its analogs, “[e]mbodied practices are carried out by agents who can still think contemplatively; nothing ‘goes without saying’ forever.” Thinking contemplatively, one should say, does not mean thinking as anthropologists think: reflexive techniques crucially vary. The native’s reverse anthropology (the Melanesian cargo cult, for example; Wagner 1981: 31–34) is not the anthropologist’s auto-anthropology (Strathern 1987: 30–31): symmetrical anthropology carried out from the tradition that generated anthropology is not symmetrical to symmetrical anthropology conducted from beyond that tradition’s boundaries. Symmetry does not cancel difference, because the virtual reciprocity of perspectives that is at issue here is not a “fusion of horizons.” In short, we are all anthropologists, but no one is an anthropologist in the same way: “it’s fine when Giddens affirms that “all social actors …are social theoreticians,” but the phrase is empty when the theoretical techniques have little in common” (Strathern 1987: 30–31).

7. As a rule, it is assumed that the native does both things—natural ratiocination and cultural rationalization—without knowing what he does, at different phases, registers, or situations during his life. The native’s illusions are, one might say, taken as necessary, in the double sense of being inevitable as well as useful (or as others would say, they are evolutionarily adaptive). Such a necessity defines the “native,” and distinguishes him from the “anthropologist”: the latter may err, but the former deludes himself.

8. “Implausibility” is an accusation that is frequently raised by practitioners of the classic game, against those who might prefer other rules. But this notion belongs in police interrogation rooms, where one must indeed be careful to ensure the “plausibility” of one’s stories.

9. This is how I interpret Wagner’s (1981: 35) declaration: “We study culture through culture, and so whatever operations characterize our investigation must also be general properties of culture.”

10. See Jullien (1989: 312) on this. Other cultures’ real problems are only possible problems for our own culture; the role of the anthropologist is to give this (logical) possibility the status of an (ontological) virtuality, determining—or rather, constructing— its latent operation in our own culture.

11. Published as an appendix to The logic of sense (Deleuze 1969a: 350–72; see also Deleuze 1969b: 333–35, 360). It is reconsidered in practically identical terms in What is philosophy? (Deleuze and Guattari 1991: 21–24, 49), (almost) his final work.

12. “…Others, from my point of view, introduce the sign of the unseen in what I do see, making me grasp what I do not perceive as perceptible to an Other” (Deleuze 1969a: 355, English translation 2003: 306).

13. This “he,” as Other, is neither a person—a third person to I and you, awaiting his turn in a dialogue—nor a thing—a “this” to speak about. The Other would be the “fourthperson singular”—situated along the river’s third bank, one might say,—and is therefore logically anterior to the perspectival game of personal pronouns (Deleuze [1979] 1995: 79).

14. The anthropologist does exactly what he thinks because the bifurcation of his own nature, while admitted perhaps in principle, is ruled out of court when it comes to his own role as anthropologist. After all, for the anthropologist it is just such a bifurcation that distinguishes the “native” from the “anthropologist” in the first place. The expression “bifurcation of nature” is coined by Whitehead ([1920] 1964: chap. II) as part of his argument against the division of reality into primary qualities, that are inherent to the object, and secondary qualities, that are attributed to it by the subject. Primary qualities are the proper object of science, although, in an ultimate sense, they remain inaccessible to it; secondary qualities are subjective and, ultimately, illusory. “Thus there would be two natures, one is the conjecture and the other is the dream” (Whitehead [1920] 1964: 30; see the quote and its commentary in Latour 1999: 62–76, 315 n49 and n58). Such a bifurcation is identical to the anthropological opposition between nature and culture. And when the object is also a subject, as in the native’s case, the bifurcation of his nature transforms itself through the distinction between the anthropologist’s conjecture and the native’s dream: cognition vs. ideology (Bloch), primary vs. secondary theory (Horton), unconscious vs. conscious model (Lévi-Strauss), propositional vs. semi-propositional representations (Sperber), and so on.

15. See Strathern (1999b: 172), on the terms of the possible knowledge relation between, for example, Western anthropologists and Melanesians: “This has nothing to do with understanding, or with cognitive structures; it is not a matter of knowing if I can understand a Melanesian, if I can interact with him, behave appropriately, etc. These things are not problematic. The problem begins when we begin to produce descriptions of the world.”

16. This is Alfred Gell’s (1998: 4) suggestion. Of course, it could be applied just as well to “human nature.”

17. This argument is only apparently similar to the one Sperber (1982: chap. 2) levels at relativism. For the author does not believe that cultural diversity is an irreducible politico-epistemological problem. For him, cultures are contingent examples of the same substantive human nature. The maximum for Sperber is a common denominator, never a multiple—see Ingold’s criticism (2000: 164) of Sperber, advanced from a different perspective, but compatible with the one adopted here.

18. On these two ideas of limit, one Platonic and Euclidian, the other Archimedean and Stoic (reappearing in the infinitesimal calculus of the seventeenth century), see Deleuze (1981).

19. In the same sense, see Jadran Mimica’s (1991: 34–38) dense phenomenological argumentation.

20. Veyne inadvertently paraphrases Evans-Pritchard, when, in characterizing this (universal) condition of being a prisoner in a (particular) historical fishbowl, he writes that “when one does not see what one does not see, one does not even see that one is blind” (Veyne 1983: 127, my emphasis, for greater clarity).

21. I am obviously interpreting Veyne’s essay here with some malice. His work is much richer (because it is so much more ambiguous) than this, taking the fishbowl beyond the “fishbowl’s” sorry image.

22. This reading of the notion of Gedankenexperiment is applied by Thierry Merchaisse to the work of François Jullien on Chinese thought (Jullien and Marchaisse 2000: 71). See also Jullien (1989: 311–12), about comparative “fictions.”

23. Responding to critics of her analysis of Melanesian sociality, who accuse her of negating the existence of a “human nature” that includes the peoples of that region, Marilyn Strathern (1999b: 172) clarifies: “[The] difference lies in the fact that the modes through which Melanesians describe, cope with human nature, are radically different to our own—and the point is that we only have access to descriptions and explanations, we can only work with them. There is no means to elude this difference. So we cannot say: very well then, now I understand, it is just a matter of different descriptions, so we can turn to the commonalities between us and them …from the moment we enter into communication, we do so through these auto-descriptions. It is essential that we can account for this.” In effect, the point is essential. See also what Jullien says about the difference between the affirmation of the existence of different “modes of orientation in thought” and the affirmation of the operation of “other logics” (Jullien and Marchaisse 2000: 205–7).

24. On the “signature” of philosophical and scientific ideas, and the “baptism” of concepts, see Deleuze and Guattari (1991: 13, 28–29).

25. The quote, and the paragraph that precedes it, have been cannibalized from Viveiros de Castro (1999: 153).

26. On “non-philosophy”—the plane of immanence or life—see Deleuze and Guattari (1991: 43–44, 89, 105, 205–206), as well as Prado Jr.’s brilliant commentary (1998).

27. The expression “apparently irrational” is a secular cliché in anthropology, from Andrew Lang in 1883 (cf. Detienne 1981: 28) to Dan Sperber in 1982.

28. As the “common-sense school of anthropology” professes, as penned by authors such as Obeyesekere (1992) or LiPuma (1998), for instance.

29. Wittgenstein’s observations on the Golden bough remain pertinent in this regard. Among others: “A religious symbol does not rest on any opinion. An error belongs only with opinion”; “I believe that what characterizes primitive man is that he does not act according to his opinions (contrary to Frazer)”; “The absurd here consists in the fact that Frazer presents these ideas [about rain rituals, etc.] as if these peoples had a completely false (and even foolish) representation of nature’s course, when all they actually have is a strange interpretation about the phenomena. That is, if they could put their knowledge of nature into writing it wouldn’t be so fundamentally different from our own. It is only that their magic is different from ours” (Wittgenstein [1930–48] 1982: 15, 24, 27). Their magic or, we could say, their concepts.

30. Rendering exterior this special and artificial condition—that is, generalizing and naturalizing it—gives rise to the classic anthropological mistake: the formal eternity of the possible is transmuted onto a historical scale, rendering anthropologist and native noncontemporaneous with one another. We then get the Other as primitive, freezeframed as an object (of the) absolute past.

31. Alexiades cites his interlocutor in Spanish—“Todos los animales son Ese Eja.” We should note that there is a further twist here: “all” the animals (the ethnographer shows numerous exceptions) are not “humans,” but they are “Ese Eja,” an ethnonym that can be translated as “human people,” and understood in opposition to “spirits” and “strangers.”

32. I am not referring to the problem of ontogenetic acquisition of “concepts” or “categories,” in the sense given to these terms by cognitive psychology. The simultaneity of the ideas of peccary, human and their identity (conditional and contextual) is, from an empirical point of view, characteristic of the thought of adults in that culture. Even if we admit that children begin by acquiring or manifesting the “concepts” of peccary and human before being taught that “the peccary are human,” it remains for the adults, when they act or argue this idea, not to re-enact this supposed chronological sequence in their heads, first thinking about humans and then peccaries, and then their association. Aside from that and above all, this simultaneity is not empirical, but transcendental: it means that the peccaries’ humanity is an a priori component of the idea of peccary (and the idea of human).

33. “The notions of relevance, necessity, the point of something, are a thousand times more significant than the notion of truth. Not as substitutes for truth, but as the measure of the truth of what I am saying” (Deleuze 1990: 177, my emphasis, English translation 1997: 130).

34. The quotations from Bento Prado Jr. are translations by this article’s translators. —Ed.

35. This is a translation of the author’s translation of an email conversation with Peter Gow. —Trans.

36. In effect, the Piro woman’s response is identical to a Zande observation, which can be found in the bible for those anthropologists of a Hortonian persuasion: “I once heard a Zande say of us: ‘Maybe in their country people are not murdered by witches, but here they are’” (Evans-Pritchard 1976: 274). I must thank Ingrid Weber for reminding me of this.

37. As Gell (1998: 101) forewarned in a similar context, magic is not a mistaken physics, but a “meta-physics”: “Frazer’s mistake was, so to speak, to imagine that practitioners of magic afforded a nonstandard theory of physics, when, in fact, ‘magic’ is what one has when one goes without a theory of physics due to its superfluousness, and when one seeks support in the perfectly practicable idea that the explanation for any given event …is that it is caused intentionally.”