The title I have given this text is 'Perspectival anthropology and the method of controlled equivocation.’ This is an allusion to a famous article by Fred Eggan (1954) entitled 'Social anthropology and the method of controlled comparison,’ which made up part of the toolbox of the legendary Harvard–Central Brazil Project, of which I am one of the academic descendents. The double difference between the titles registers the general direction of my argument, which, truth be known, has little to do with Eggan’s. The substitution of 'perspectival’ for 'social’ indicates first of all that the 'anthropology’ to which I am referring is a hybrid formation, the result of a certain recursive imbrication between Western anthropological discourse (our very own ethno-anthropology), rooted in our modern multiculturalist and uninaturalist ontology, and the anthropological image conveyed by Amerindian cosmopraxis in the form of a perspectivist theory of transpecific personhood, which is by contrast unicultural and multinatural.
Secondly and more generally, this substitution expresses my conviction that our anthropology is 'social' (or, for that matter, 'cultural') only in so far as the first question faced by the anthropologist is working out what constitutes, both in extension and comprehension, the concept of 'the social' (or 'the cultural') for the people she or he studies; or, put otherwise, what is the anthropology of this people – one which configures this people as theoretical agent rather than patient. As I argued in a recent paper (Viveiros de Castro 2002: 122), anthropology's defining problem consists less in determining which social relations constitute its object, and much more in asking what its object constitutes as a social relation – what a social relation is in the terms of its object, or better still, in the terms which emerge from the relation (a social relation, naturally) between the 'anthropologist' and the 'native.'
Put concisely, what I am saying is that doing anthropology means comparing anthropologies. Nothing more – and nothing less. And if this is true, then comparison is not just our primary analytic tool – it is also our raw-material and our ultimate grounding. Since what we compare are always and necessarily, in one form or other, comparisons. If culture, as Marilyn Strathern wrote, "…consists in the way people draw analogies between different domains of their worlds" (1992: 47), then every culture is a gigantic and multidimensional process of comparison. And if anthropology – and now I cite Roy Wagner – "studies culture through culture," then "whatever operations characterize our investigations must also be general properties of culture" (1981: 35). In brief, the anthropologist and native are engaged in "directly comparable intellectual operations" (Herzfeld 2002: 7), and such operations are above all comparative operations. Intracultural relations, or internal comparisons (the Strathernian "analogies between domains"), and intercultural relations, or external comparisons (the Wagnerian "invention of culture"), are in strict ontological continuity.
But direct comparability does not necessarily signify immediate translatability, just as ontological continuity does not imply epistemological transparency. How can we restore the analogies traced by Amazonian peoples within the terms of our own analogies? What happens to our comparisons when we compare them with indigenous comparisons?
I propose the notion of 'equivocation' as a means of reconceptualizing, with the help of Amerindian perspectivist anthropology, this emblematic procedure of our academic anthropology – comparison. Not the kind Eggan had in mind, which was comparison between different spatial or temporal instantiations of a given sociocultural form; seen from the viewpoint of the 'rules of anthropological method,' this type of comparison is just a regulative rule – other forms of anthropological investigation exist. Rather, the comparison of which I am thinking is a constitutive rule of the discipline. It concerns the process involved in the translation of the 'native's practical and discursive concepts into the terms of anthropology's conceptual apparatus; that is, I am talking about the kind of comparison, more often than not implicit or automatic (and hence uncontrolled), which necessarily includes the anthropologist's discourse as one of its terms, and which starts to be processed from the very first moment of fieldwork, if not well before. Controlling this translative comparison between anthropologies is precisely what comprises the 'art of anthropology.'
Today it is undoubtedly a commonplace to say that cultural translation is our discipline's distinctive task. But the problem is knowing what precisely is, can or should be a 'translation,' and how to carry such an operation out. It is here things start to become a lot more tricky, as Talal Asad demonstrated in a noteworthy article (1986). I adopt the radical position which is, I believe, the same as Asad's, and which can be summarized as follows: in anthropology, comparison is in the service of translation and not the opposite. Anthropology compares so as to translate, and not to explain, justify, generalize, interpret, contextualize, reveal the unconscious, say what goes without saying, and so forth. I would add that to translate is always to betray, as the Italian saying goes; but a good translation – and here I am paraphrasing Walter Benjamin (or rather Rudolf Pannwitz via Benjamin) – is one which betrays the destination language, not the source language. A good translation is one which allows the alien concepts to deform and subvert the translator's conceptual toolbox so that the intention of the original language can be expressed within the new one.
I shall present a brief account (a translation) of the theory of translation present in Amerindian perspectivism in order to see whether we can succeed in modifying our own ideas about translation – and thus about anthropology – in such a way as to reconstitute the intention of Amerindian anthropology in the language of our own. In doing so I shall make the claim that perspectivism projects an image of translation as a process of controlled equivocation – 'controlled' in the sense in which walking may be said to be a controlled way of falling. Indigenous perspectivism is a theory of the equivocation, that is, of the referential alterity between homonymic concepts; equivocation appears here as the mode of communication par excellence between different perspectival positions – and therefore as both condition of possibility and limit of the anthropological enterprise.
I use 'perspectivism' as a name for a set of ideas and practices found throughout indigenous America and to which I shall refer, for simplicity's sake, as though it were a 'cosmology.' This cosmology imagines a universe peopled by different types of subjective agencies, human as well as non-human, each endowed with the same generic type of soul, i.e. the same set of cognitive and volitional capacities. The possession of a similar soul implies the possession of similar concepts, which determine that all subjects see things in the same way; in particular, individuals of the same species see each other (and each other only) as humans see themselves; that is, as beings endowed with human shape and habits, seeing their bodily and behavioral aspects in the form of human culture. What changes when passing from one species of subject to another is the 'objective correlative,' the referent of these concepts: what jaguars see as 'manioc beer' (the proper drink of people, jaguar-type or otherwise), humans see as 'blood;' where we see a muddy salt-lick on a river bank, tapirs see their big ceremonial house, and so on. Such difference of perspective – not a plurality of views of a single world, mind you, but a single view of different worlds – cannot derive from the soul, since the latter is the common original ground of being; such difference is located in the bodily differences between species, for the body and its affections (in Spinoza's sense: its capacities to affect and be affected by other bodies) is the site and instrument of ontological differentiation and referential disjunction. 
Hence, where our modern, anthropological 'multiculturalist' ontology is founded on the mutual implication of the unity of nature and the plurality of cultures, the Amerindian conception would suppose a spiritual unity and a corporeal diversity – or, in other words, one 'culture,' multiple 'natures.' In this sense, perspectivism is not relativism as we know it – a subjective or cultural relativism – but an objective or natural 'relativism' – a multinaturalism. Cultural relativism imagines a diversity of subjective and partial representations (cultures) referring to an objective and universal nature, exterior to representation; Amerindians, on the other hand, propose a representative or phenomenological unity which is purely pronominal in kind (any species of subject perceives itself and its world in the same way we perceive ourselves and our world; 'culture' is what one sees of oneself when one says 'I'), applied to a real radical diversity.
The problem for indigenous perspectivism is not therefore one of discovering the common referent (say, the planet Venus) to two different representations (say, 'Morning Star' and 'Evening Star'), but, on the contrary, one of making explicit the equivocation which would comprise imagining that when the jaguar says 'manioc beer' he is referring to the same thing as us (i.e. a tasty, nutritious and heady brew). In other words, perspectivism supposes a constant epistemology and variable ontologies; the same representations, other objects; single meaning, multiple referents. Therefore, the aim of perspectivist translation – translation being one of shamanism's principal tasks, as we know (Carneiro da Cunha 1998) – is not that of finding a 'synonym' (a co-referential representation) in our human conceptual language for the representations which other species of subject use to speak about one and the same Thing; rather, the aim is not to 'lose sight' of the difference concealed within equivocal 'homonyms' between our language and that of other species – since we and they are never talking about the same things.
This idea may at first sound slightly counter-intuitive, for when we start thinking about it, it seems to collapse into its opposite. Here is how Gerald Weiss, for instance, described the Campa world:
It is a world of relative semblances, where different kinds of beings see the same things differently; thus humans eyes can normally see good spirits only in the form of lightning flashes or birds whereas they see themselves in their true human form, and similarly in the eyes of jaguars human beings look like peccaries to be hunted (1972: 170).
Now, the manner in which Weiss 'sees things' is not an error but, precisely, an equivocation. The fact that different kinds of beings see the same things differently is but a consequence of the fact that different kinds of beings see different things in the same way. For what counts as 'the same things'? Same in relation to whom, to which species? The phantasm of the Thing-in-Itself haunts Weiss's formulation, which actually expresses an inversion of the problem posed by perspectivism – a typically anthropological inversion.
Perspectivism includes a theory of its own description by anthropology – since it is an anthropology. Amerindian ontologies are inherently comparative: they presuppose a comparison between the ways different kinds of body 'naturally' experience the world as an affectual multiplicity. They are, thus, a kind of inverted anthropology, for the latter proceeds by way of an explicit comparison between the ways different types of mentality 'culturally' represent the world, seen as the unitary origin or virtual focus of its different conceptual versions. Hence a culturalist (anthropological) account of perspectivism necessarily implies the negation or delegitimization of its object, its 'retroprojection' as a primitive and 'fetishized' kind of anthropological reasoning.
What I propose as an experimental program is the inversion of this inversion, which starts out from the following question: what would a perspectivist account of anthropological comparison look like? As I lack the time here to reply in full with detailed examples of 'controlled equivocation,' I shall discuss just its general principles.
One of the starting points for my first analysis of perspectivism, published in 1996, was an anecdote told by Lévi-Strauss in Race et histoire (see above, Introdução ao contexto do perspectivismo. It illustrates the pessimistic thesis that one of the intrinsic aspects of human nature is the denial of its own universality. A congenital and narcissistic avarice, preventing the attribution of the predicates of human nature to the species as a whole, appears to be part of these self-same predicates. In sum, ethnocentrism, just like good sense (of which it is perhaps the sociological translation) is the best shared thing in the world. Lévi-Strauss illustrates the universality of this anti-universalist attitude with an anecdote based on Oviedo's History, and which took place in Puerto Rico:
In the Greater Antilles, some years after the discovery of America, whilst the Spanish were dispatching inquisitional commissions to investigate whether the natives had a soul or not, these very natives were busy drowning the white people they had captured in order to find out, after lengthy observation, whether or not the corpses were subject to putrefaction. (1973 : 384)
The parable's lesson obeys a familiar ironic format, but is none the less striking. The favoring of one's own humanity at the cost of the humanity of another manifests a similarity with this scorned other. And since the Other of the Same (of the European) is revealed to be the same as the Other of the Other (of the Indian), the Same ends up revealing itself – unknowingly – to be exactly the same as the Other.
The anecdote was recounted by the author in Tristes tropiques. There it illustrates the cosmological shock produced in sixteenth century Europe by the discovery of America. The moral of the tale continues to be that of the previous book, namely the mutual incomprehension between Indians and Spaniards, equally deaf to the humanity of their unheard-of others. But Lévi-Strauss introduces an asymmetry, observing tongue-in-cheek, that, in their investigations into the humanity of the other, the Whites invoked the social sciences, while the Indians placed more trust in the natural sciences; and that while the first came to the conclusion that the Indians were animals, the latter were content to suspect that the Whites were gods. "In equal ignorance," concludes the author, the latter was an attitude more befitting of human beings (1955: 81-83).
Therefore, despite sharing an equal ignorance about the Other, the Other of the Other was not exactly the same as the Other of the Same. It was in pondering this difference that I began to formulate the hypothesis that indigenous perspectivism situated the crucial differences between the diversity of subjects on the plane of the body and not the spirit. For the Europeans, the ontological diacritic is the soul (are Indians humans or animals?); for the Indians it is the body (are Europeans humans or spirits?). The Europeans never doubted that the Indians had bodies (animals have them too); the Indians never doubted that the Europeans had souls (animals and spirits have them too). In sum: European ethnocentrism consisted in doubting whether other bodies have the same souls as they themselves (today we would call the soul 'the mind', and the XVI century theological problem would now be the philosophical 'problem of other minds'). Amerindian ethnocentrism, on the contrary, consisted in doubting whether other souls had the same bodies.
This anecdote from the Antilles casts some light, I think, on one of the core elements of the perspectivist 'message' – the idea of difference being inscribed in bodies, and the idea of the body as a dispositional system of affectability (do Europeans putrefy?) rather than as a material morphology. It was only very recently, though, that it dawned on me that the anecdote wasn't simply 'about' perspectivism, it was itself perspectivist, instantiating the same framework or structure (if you pardon my French – my old French) manifest in the innumerable Amerindian myths thematizing interspecific perspectivism. Here I have in mind the type of myth where, for example, the human protagonist becomes lost deep in the forest and arrives at a strange village; there the inhabitants invite him to drink a refreshing gourd of 'manioc beer,' which he accepts enthusiastically – and, to his horrified surprise, his hosts place in front of him a gourd brimming with human blood. Both the anecdote and the myth turn on a type of communicative disjuncture where the interlocutors are not talking about the same thing, and do not know this (in the case of the anecdote, the 'dialogue' takes place on the plane of Lévi-Strauss's comparative reasoning on reciprocal ethnocentrism). Just as jaguars and humans apply the same name to two very different things, both Europeans and Indians 'were talking' about humanity, that is, they were questioning the applicability of this self-descriptive concept to the Other; but what Europeans and Indians understood to be the concept's defining criterion (its intension and consequently its extension) was radically different. In sum, both Lévi-Strauss's anecdote and the myth turn on an equivocation.
If we think about it carefully, the Antilles anecdote is similar to countless others we can come across in the ethnographic literature, or in our own recollections from fieldwork. In actual fact, I think it encapsulates the anthropological situation or event par excellence, expressing the quintessence of what our discipline 'is all about.' It is possible to discern, for example, in the archi-famous episode of the death of Captain Cook, as analyzed by Marshall Sahlins, a structural transformation of the cross experiments of Puerto Rico: we are presented with two versions of the archetypical anthropological motive, intercultural equivocality. Life as always imitates art – events mime myth, history rehearses structure. [In fact we could describe the Sahlins/Obeyesekere debate as a further recursion of the Cook-Hawaiian episode: a misunderstanding about a misunderstanding about a misunderstanding.]
I shall propose one or two more examples of equivocation below. But what I wish to make clear is that equivocation is not just one among other possible pathologies which threaten communication between the anthropologist and the 'native' – such as linguistic incompetence, ignorance of context, lack of personal empathy, indiscretion, literalist ingenuity, commercialization of information, lies, manipulation, bad faith, forgetfulness, and sundry other deformations or shortcomings which may afflict anthropological discursivity at an empirical level. In contrast to these contingent pathologies, the equivocation is a properly transcendental category of anthropology, a constitutive dimension of the discipline's project of cultural translation. It expresses a de jure structure, a figure immanent to anthropology . It is not merely a negative facticity, but a condition of possibility of anthropological discourse – that which justifies the existence of anthropology (quid juris? as in the Kantian question). To translate is to situate oneself in the space of the equivocation and dwell there. Not to unmake it, since this would be to suppose it never existed, but precisely the opposite: to emphasize or potentialize it, that is, to open and widen the space imagined not to exist between the conceptual languages in contact – a space which the equivocation precisely concealed. The equivocation is not that which impedes the relation, but that which founds and impels it: a difference in perspective. To translate is to presume that an equivocation always exists; it is to communicate by differences, instead of silencing the Other by presuming a univocality – the essential similarity – between what the Other and we are saying.
M. Herzfeld recently observed that "anthropology is about misunderstandings, including anthropologists' own misunderstandings, because these are usually the outcome of the mutual incommensurability of different notions of common sense – our object of study" (2003: 2). I could not agree more. I would simply insist on the point that, if anthropology exists (de jure), it is precisely (and only) because that which Herzfeld calls 'common sense' is not common. I would also add that the incommensurability of the clashing 'notions,' far from being an impediment to their comparability, is precisely what enables and justifies it (as Michael Lambek argued ). Since it is only worth comparing the incommensurable – comparing the commensurable is a task for accountants, not anthropologists. And finally I should add that I conceive the idea of 'misunderstanding' in the specific sense of equivocality found in Amerindian perspectivist cosmology: an equivocation is not just a 'failure to understand' (O.E.D.), but a failure to understand that understandings are necessarily not the same, and that they are not related to imaginary ways of 'seeing the world' but to the real worlds which are being seen. In Amerindian cosmology, the real world of the different species depends on their points of view since the 'world in general' consists of the different species themselves; it is the abstract space of divergence between them as points of view: there are no points of view onto things – it is things and beings which are points of view (as Deleuze would say, 1981: 203). The question for Indians, therefore, is not one of knowing "how monkeys see the world" (Cheney & Seyfarth 1990), but what world is expressed through monkeys, of what world they are the point of view. I believe this is a lesson our own anthropology can learn from.
Anthropology, then, is 'about misunderstandings'. But as Roy Wagner appropriately said (1981: 20) about his early relations with the Daribi: "their misunderstanding of me was not the same as my misunderstanding of them." The crucial point here is not the empirical fact that misunderstandings exist, but the 'transcendental fact' that it was not the same misunderstanding.
The question is not discovering who is wrong, and still less who is deceiving whom. An equivocation is not an error, a mistake or a deception, but the very foundation of the relation which it implicates, and which is always a relation with an exteriority. An error or deception can only be determined as such from within a given language game, while an equivocation is what unfolds in the interval between different language games. Deceptions and errors suppose premises that are already constituted, and constituted as homogenous, while an equivocation not only 'supposes' the heterogeneity of the premises at stake – it poses them as heterogenic and presupposes them as premises. An equivocation determines the premises rather than being determined by them. Consequently, equivocations do not belong to the world of dialectical contradiction, since their synthesis is disjunctive and infinite. An equivocation is indissoluble, or rather, recursive: taking it as an object determines another equivocation 'higher up,' and so on ad infinitum.
The equivocation, in sum, is not a subjective failure, but a tool of objectification. It is not an error or an illusion – it doesn't involve imagining objectification in the post-Enlightenment and moralizing language of reification or fetishization (today better known as 'essentialization') –, but the limit-condition of every social relation, a condition which itself becomes super-objectified in the limit-case of so-called 'interethnic' or intercultural relations, where the language games diverge maximally. This divergence includes, it goes without saying, the relation between anthropological discourse and native discourse. Thus, the anthropological concept of culture, for example, as Wagner argued, is the equivocation which emerges as an attempt to solve intercultural equivocality; and it is equivocal in so far as it fixes on the "paradox created by imagining a culture for people who do not imagine it for themselves" (1981: 27). Accordingly, even when misunderstandings are transformed into 'understandings' – like when the anthropologist transforms his initial bewilderment at the natives' ways into 'their culture,' or when the natives understand that what the Whites called 'gifts' were in reality 'commodities' – even here such understandings persist in being not the same. The Other of the Others is always other. And if the equivocation is not an error, an illusion or a lie, but the very form of the relational positivity of difference, its opposite is not the truth, but the univocal, as the claim to the existence of a unique and transcendent meaning. The error or illusion par excellence consists, precisely, in imagining that the univocal exists beneath the equivocal, and that the anthropologist is its ventriloquist.
An equivocation is not an error: the Spanish theologians and the Indians of Porto Rico, the Hawaiian warriors and the British sailors were not wrong. I now wish to present another example of an equivocation, this time taken from an anthropological analysis. The example has been extracted from a recent Americanist monography of the highest quality – I wish to emphasize this –, written by a colleague who I admire greatly. Consider, then, this metacommentary by Greg Urban in his book Metaphysical Community, on Shokleng community-making discourse. Explaining Discourse's sociogenetic powers, Urban observes:
Unlike the Serra Geral mountain range or jaguars or araucaria pines, the organization of society is not a thing that is out there, waiting to be understood. The organization must be created, and it is something elusive, intangible that does the creating. It is culture – here understood as circulating discourse (1996: 65)
The author is defending a moderate constructionist position. Society, qua Shokleng social organization with its groups and emblems, is not something given, as traditional anthropologists used to think; it is something created through discourse. But discourse's powers have their limits: geographical features and biological essences are out there; they are, so to speak, bought ready-made, not made at home through circulating discourse. What to make of this? Firstly, it must be admitted there is nothing in the least bit shocking about Urban's commentary; indeed, it seems eminently reasonable, and canonically anthropological. And moreover, it also accords neatly with what some equally reasonable philosophers look to teach us about the structure of reality. Take the doctrine of John Searle (1995), for example, which argues that two and only two types of facts exist: 'brute facts,' such as hills, rain and animals, and 'institutional facts,' such as money, iceboxes or marriage. The latter are made or constructed (performed) facts, since their sufficient reason coincides entirely with their meaning; the former, however, are given facts, since their existence is independent of the values attributed to them. In a couple of words: nature and culture.
However, what do the Shokleng have to say about the matter? At the end of reading Metaphysical Community, the reader cannot but feel a certain unease in noting that Urban's splitting of the world into a given realm of jaguars and pine trees, and a constructed world of groups and emblems, is not the split made by the Shokleng. Actually, it is almost exactly the inverse. The indigenous myths magnificently analyzed by Urban tell, among other things, that the original Shokleng, after sculpting the future jaguars and tapirs in araucaria wood, gave these animals their characteristic pelts by covering them with the diacritical marks pertaining to the clanic-ceremonial groups: spots for the jaguar, stripes for the tapir (1996: 156-58). In other words, it is social organization which was 'out there,' and the jaguars and tapirs who were created or performed by it. The institutional fact created the brute fact. Unless, of course, the brute fact is the clanic division of society, and the institutional fact, the jaguars of the forest. For the Shokleng, in fact, culture is the given, and nature, the constructed. For them, if the cat is on the mat, (or rather, if "o gato está no mato," the jaguar is in the jungle) – it is because someone put it there.
In sum, we are faced with an equivocation. The discordant distribution of the given and the constructed which inexorably separates Shokleng discourse on the real and anthropological discourse on Shokleng discourse is never explicitly recognized as such by Urban. The solution which he implicitly offers for this chiasma is anthropology's classical solution. It consists of a highly characteristic operation of translation, which involves the metaphysical demotion of the indigenous distribution of the world to the condition of metaphor: "Creation of the animal world is a metaphor for the creation of community" (ibid: 158). Where would we be without this statutory distinction between the literal and the metaphoric, which strategically blocks any direct confrontation between the discourses of anthropologist and native, thereby avoiding any major unpleasantness? Urban deems that the creation of community is literal, and that of jaguars, metaphoric. Or rather, that the first is literally metaphoric and the second metaphorically literal. The creation of community is literal, but the community thereby created is metaphoric (not 'something out there'); jaguars, they will be pleased to know, are literal, but their creation by the community is metaphoric.
We don't know whether the Shokleng concur with the anthropologist in considering the creation of jaguars and tapirs as a metaphor for the creation of the community; we could hazard a guess that probably not. On the other hand, Urban deems that the Shokleng do concur with him about the metaphorical nature of the community created by themselves, or better (and literally), by their discourse. Unlike other anthropologists and/or peoples encumbered by a more essentialist mentality, the Shokleng are aware, thinks Urban, that their division into nominally but not really exogamic groups is not a brute fact, but a metadiscursive representation of the community, which merely deploys the idiom of affinity and interfamily alliance in a 'playful' way (ibid: 168). Thus, the anthropologist agrees with the Shokleng construction of the community as constructed, but disagrees with their positing of jaguars as constructed.
Later in his work, Urban interprets indigenous ceremonies as a way of representing the community in terms of relations within the family. The latter is described in its turn (though we do not know whether by the anthropologist or by the natives) as an elementary unit founded on the 'psychologically primitive' relations between the sexes and generations (ibid: 188-93). Society, metaphorized into its emblematic divisions and its collective rituals, is therefore imagined either as the result of an alliance between families, or, at a deeper level, as a nuclear family. But the family does not seem to be, in Urban's eyes at least, a metaphor of anything else – it is literal. It is a given which usefully serves as a metaphor for less literal things; it is a naturally appropriate image, due to its cognitive salience and affective pregnancy (ibid: 171, 192-93): it is more real than the community. Society is naturally metaphoric, the family is socially literal. The nuclear family, the concrete bonds of conjugality and filiation, are a fact, not a fabrication. Kinship – not the metaphoric and intergroup kind of the community, but the literal and interindividual kind of the family – is something just as out there as the animals and plants; something without whose help, furthermore, discourse would be unable to construct the community. Indeed, it may even be out there for the same reasons as the animals and plants: by being, let's say, a 'biological' phenomenon.
Urban claims that anthropologists, in general, "have been the dupes" of peoples who may have taken their own metadiscourse on social organization "too seriously," and who thus proved to be over-literalists, that is, essentialists (ibid: 137, 168-69). It may be that anthropology really has adopted a literalist attitude vis-à-vis the essence of 'society.' But in counterpart, in terms of indigenous discourse on 'nature' at least, it has never been duped, neither by the native, nor above all about the native: the so-called symbolist interpretation (Skorupski 1976) of primitive metaphysics has been in discursive circulation ever since Durkheim. And it is this same interpretation which Urban applies to Shokleng discourse on jaguars, whose literality he rejects – but rejects in favor of a completely literalist interpretation of the Western discourse on 'things out there.' In other words, if the Shokleng concur (for the sake of hypothesis) with Urban's anti-Durkheimian ontology of society, Urban concurs with Durkheim about the ontology of nature. What he is advocating is simply the extension of the symbolist attitude to the case of discourses about society, which thereby ceases to be the referential substrate of crypto-metaphoric propositions about nature (as it was in Durkheim). Now society too is metaphoric. But in this case, what serves as a real substrate, or the literal truth metaphorically distorted by discourse? – discourse itself, which turns into a substance 'out there.' The impression left behind is that discursive constructionism has to reify discourse in order to be able to de-reify society. Discourse and, as we can spot, the family.
Was Urban wrong – that is, was he making a false claim – in declaring that mountains and natural species are out there, while society is a cultural product? I don't believe so. But nor do I think he was right either. Since this is not the point. As far as any anthropological point is at stake here, the interest of his declaration lies in the fact that it counterinvents the equivocation it enables, and that gives it its objectifying power. Urban's professed faith in the ontological self-subsistence of mountains and animals and on the institutional demiurgy of discourse is, in the final analysis, indispensable for us to be able properly to evaluate the enormity of the gap separating the indigenous and anthropological ontologies.
I believe that I can indeed speak of an error or mistake on Urban's part, since I am situated within the same language game as himself – anthropology. I can therefore legitimately say (though I certainly may be wrong) that Urban was perpetrating an important anthropological error by failing to take into account the equivocation within which he was implicated. The discordant distribution of the given and constructed parts between Urban and the Shokleng is not an anodyne choice, a mere swapping of signals leaving the terms of the problem untouched. There is "all the difference in the world" (Wagner 1981: 51) between a world where the primordial is experienced as naked transcendence, pure anti-anthropic alterity – the non-constructed, the non-instituted, what opposes custom and discourse –, and a world of immanent humanity, where the Primordial takes on human form – which doesn't make it, I must stress, necessarily tranquilizing: there where everything is human, the human is something else entirely. Describing this world as though it were an illusory version of our own, unifying the two via a reduction of one to the conventions of the other, is to imagine an overly simple form of relation between them. This explanatory ease ends up producing all sorts of uneasy complications, since this desire for ontological monism usually pays with an inflationary emission of epistemological dualisms – 'emic' and 'etic,' metaphoric and literal, conscious and unconscious, representation and reality, illusion and truth, and what not. "Perspective is the wrong metaphor," fulminates Stephen Tyler in his normative manifesto for postmodern ethnography (1986: 137). The equivocation which articulates Shokleng discourse with the discourse of their anthropologist leads me to conclude, very much to the contrary, that metaphor is the wrong perspective. Above all when anthropology finds itself face-to-face with a cosmology which is literally perspectivist.
I conclude by narrating a small translational equivoque in which I became involved a few years back. Milton Nascimento, the celebrated Brazilian musician, had made a journey to Amazonia, guided by some friends of mine who work for an NGO (Non Governmental Organization). One of the high points of the trip had been a two-week stay among the Cashinahua of the Jordão river. Milton was overwhelmed by the warm welcome received from the Indians. Back on the Brazilian coast, he decided to use an indigenous word as a title for the album he was recording. The word chosen was txai, which the Cashinahua had used abundantly in addressing Milton and the other members of the expedition.
When the album Txai was due to be released, one of my ISA friends asked me to write a sleeve note. The idea was to explain to Milton's fans what the title meant: could I not say something about the sense of brotherhood expressed by the term txai, about its meaning 'brother,' and so on?
I replied that it was impossible to write the note in these terms, since txai may mean just about everything except, precisely, 'brother.' I explained that txai is a term used by a man to address certain kinsfolk: his cross-cousins, his MF, his DCh, and, in general, following the Cashinahua system of 'prescriptive alliance,' to any man whose sister Ego treats as an equivalent to his wife, and vice-versa (Kensinger 1995: 157-74). In sum, txai means something akin to 'brother-in-law;' it refers to a man's real or possible brothers-in-law, and, when used as a friendly vocative to speak to non-Cashinahua outsiders, the implication is that the latter are kinds of affines. Moreover, I explained that one doesn't need to be a friend to be txai; it suffices to be an outsider, or even – and even better – an enemy: thus the Inca in Cashinahua mythology are at once monstrous cannibals and archetypical txai with whom, we should note in passing, one should not or indeed cannot marry (McCallum 1991).
But none of this would work, complained my friend; Milton thinks that txai means 'brother,' and besides it would be fairly ridiculous to give the record a title whose translation is 'Brother-in-law', wouldn't it? Perhaps, I conceded. But don't expect me to skip over the fact txai signifies 'other' or affine. The end result of the conversation was that the album continued to be called Txai, and the sleeve note ended up being written by someone else.
Note that the problem with this misunderstanding about txai doesn't lie in the fact Milton Nascimento and my friend were wrong concerning the sense of the Cashinahua word. On the contrary, the problem is they were right – in a certain sense. In other words, they were 'equivocated,' so to speak. The Cashinahua, like so many other indigenous peoples of Amazonia, use terms whose most direct translations are 'brother-in-law' or 'cross-cousin' in various contexts in which Brazilians, and other peoples from the Euro-Christian tradition, would really expect something like 'brother.' In this sense, Milton was right. Had I remembered, I would have reminded my interlocutor that the equivocation had already been anticipated by an ethnologist of the Cashinahua. Talking about the difference between the social philosophy of this people and that held by the surrounding Whites, B. Keifenheim concludes: "The message 'all men are brothers' encountered a world where the most noble expression of human relations is the relation between brothers-in-law…" (1992: 91). Exactly. But it is for this very reason that 'brother' is not an adequate translation for txai; if there exists anyone who a Cashinahua man would be reluctant to call 'txai', it is his own brother. Txai means 'affine,' not 'consanguine,' even when used for purposes similar to our own, when we address a stranger as 'brother.' While the purposes may be similar, the premises are decidedly not so.
My translational mishap will undoubtedly sound completely banal to the ears of Americanists, who have been interested for a long time in the innumerable symbolic resonances of the idiom of affinity in Amazonia. The interest of this anecdote in the present context, however, is that it seems to me to express, in the actual difference between the idioms of 'brother' and 'brother-in-law,' two inverse modes of conceiving the principle of translative comparison: the multiculturalist mode of anthropology and the multinaturalist mode of perspectivism.
The powerful Western metaphors of brotherhood privilege certain (not all) logical properties of this relation. What are siblings, in our culture? They are individuals identically related to a third term, their genitors or their functional analogs. The relation between two siblings derives from their equivalent relation to an origin which encompasses them and whose identity identifies them. This common identity means that siblings occupy the same point of view onto an exterior world; deriving their similitude from a similar relation to a same origin, they will have 'parallel' relations (to use an anthropological image) to everything else. Thus, people who are unrelated, when conceived to be related in a generic sense, are so in terms of a common humanity which makes all of us kin; that is, siblings, or at least, to continue to use the previous image, parallel cousins, classificatory brothers: children of Adam, of the Church, of the Nation, of the Genome or of any other figure of transcendence. All men are brothers to some extent, since brotherhood is in itself the general form of the Relation. Two partners in any relation are defined as connected in so far as they can be conceived to have something in common, that is, as being in the same relation to a third term. To relate is to assimilate, to unify, to identify.
The Amazonian model of the Relation could not be more different to this. 'Different' is the apposite word, since Amazonian ontologies postulate difference rather than identity as the principle of relationality. And it is precisely the difference between the two models which founds the relation which I am attempting to establish here between them (and here we are already using the Amerindian mode of comparing and translating).
The common word for the relation, in Amazonian worlds, is the term translated by 'brother-in-law' and/or 'cross-cousin.' This is the term we call people we don't know what to call, those with whom we wish to establish a generic relation. In sum, 'cousin/brother-in-law' is the term which creates a relation where none existed; it is the form through which the 'unknown' is made known.
What are the logical properties of the connection of affinity highlighted in these indigenous usages? As a general model of relationship, the brother-in-law connection appears as a cross connection with a mediating term which is seen in diametrically opposite ways by the two poles of the relation: my sister is your wife and/or vice-versa. Here, the parties involved find themselves united by that which divides them, linked by that which separates them (Strathern 1992b: 99-100). My relation with my brother-in-law is based on my being in another kind of relation to his relation with my sister or my wife. The Amerindian relation is a difference of perspective. While we tend to conceive the action of relating as a discarding of differences in favor of similarities, indigenous thought sees the process from another angle: the opposite of difference is not identity but indifference. Hence, establishing a relation – like that of the Cashinahua with Milton Nascimento – is to differentiate indifference, to insert a difference where indifference was presumed. (No wonder, then, that animals are so often conceived as affinally related to humans in Amazonia. Blood is to humans as manioc beer is to jaguars, in exactly the same way as a sister to me is a wife to my brother-in-law. The many Amerindian myths featuring interspecific marriages and discussing the difficult relationships between the in-marrying affine and his/her allospecific parents-in-law, simply compound the two analogies into a single one.)
The implications of these two models of social relationship for an anthropological theory of translation are evident, and it is with these that I finish. Such implications are not 'metaphorical;' if anything, the opposite happens to be the case – since relations of meaning are social relations. Very well. If the anthropologist starts out from the meta-principle that 'all men are brothers,' he (or she) is presupposing that his (or her) discourse and that of the native manifest a relation of an ultimately 'brotherly' nature. What founds the relation of meaning between the two discourses – and therefore justifies the operation of translation – is their common referent, of which both present 'parallel' visions. Here, the idea of an external Nature which is logically and chronologically prior to the cultures that partially represent it acts out the role of the Parent which founds the relation between two 'siblings.' (And here we could imagine a hierarchical interpretation of this brotherly parallelism, with the anthropologist assuming the role of literal and rational 'eB,' and the native his metaphoric and symbolic 'yB' – or, on the contrary, we could adopt a radically egalitarian interpretation, with the two protagonists seen as twins, etc.). Whatever the case, in this model translation is only possible because the discourses are composed of 'synonyms:' they express the same 'parental' reference to some (indeed any) kind of transcendence with the status of Nature (Physis, Socius, Gene, Cognition, Discourse, etc.). Here, to translate is to isolate what the discourses share in common, something which is only 'in them' because it is (and was already before them) 'out there;' the differences between the discourses amount to no more than the residue which precludes a 'perfect translation,' that is, an absolute identificatory overlap between them. To translate is to presume redundancy.
However, if all men are brothers-in-law rather than brothers – that is, if the image of the social connection is not that of sharing something in common (a something-in-common acting as foundation), but, on the contrary, that of the difference between the terms of the relation, or better, of the difference between the differences which constitute the terms of the relation – then a relation can only exist between what differs and in so far as it differs. In this case, translation becomes an operation of differentiation – a production of difference which connects the two discourses to the precise extent to which they are not saying the same thing, in so far as they point to discordant exteriorities beyond the equivocal homonyms between them. (Contrary to Derrida, I believe the hors-texte perfectly well exists, de facto and de jure – but contrary to the positivists, I think each texte has its own hors-texte). In this case, cultural translation is not a process of induction – finding the common points in detriment to the differences – , much less a process of deduction – applying apriori a principle of natural unification to cultural diversity in order to determine (or decree) its meaning –, but a process of the type which the philosopher Gilbert Simondon called transduction:
Transduction functions as the inversion of the negative into a positive: it is precisely that which determines the non-identity between the terms, that which makes them disparate (in the sense held by this term in the theory of vision) which is integrated with the system of resolution and becomes the condition of signfication; transduction is characterized by the fact that the outcome of this operation is a concrete fabric including all the initial terms... (1995: 32).
In this model of translation, which I believe coincides with that present in Amerindian perspectivism, difference is therefore a condition of signification and not a hindrance. The identity between the 'beer' of the jaguar and the 'beer' of humans is posed only the better to see the difference between jaguars and humans. As in stereoscopic vision, it is necessary that the two eyes don't see the same given thing in order for another thing (the real thing in the field of vision) to be able to be seen, that is, constructed – counterinvented. In this case, to translate is to presume a difference. The difference, for example, between the two modes of translation I have presented to you here. But perhaps this is an equivocation.