Transmutating Beings: A Proposal for an Anthology of Thought




By Dr. Carlos Severi
Director of Research, Professor of Anthropology
Ecole des Hautes Etudes en Sciences Sociales


Forms of thought, from what Lévi-Strauss called the “systematization [of] what is immediately presented to the senses,” to the causal theories studied by Evans-Pritchard in witchcraft, have generally been interpreted as an expression of a specific language or “culture.” In this paper, I discuss this way of defining thought. Three classic objections are examined: (1) societies sharing the same “system of thought” may speak different languages, and vice versa; (2) if a relation between language and thought exists, it is an indirect and controversial one, and we should never take it for granted (or infer qualities of thought from language structures) without further investigation; (3) the languages that we use to qualify different kinds of thought are constantly translated. Through a discussion of the context of translation, I argue that instead of seeing the possibility of translation as a theoretical difficulty for defining thought, we could, on the contrary, consider the ethnography of translation as a chance to observe the dynamics and structure of thought processes, and to study how they operate in different cultural contexts. Using three Amazonian examples, I will try to describe the kind of cognition involved by the form of translation that Jakobson calls transmutation. I will argue that from this ethnographic analysis, we can not only derive a better (both wider and more precise) idea of some, rarely studied, cultural translation processes, but also draw from it a new way to define the concept of “cultural ontology,” both for Amazonian cultures and in more general terms.


In his Remarks on Frazer’s Golden Bough, Wittgenstein writes that a good theory of magic should “preserve its depth,” not simply condemn it as a mistake from the point of view of rationality. This “cancelling of magic”—he adds immediately after—would “have the character of magic itself” (Wittgenstein [1967] 1979: 1). Wittgenstein captures here one of the more deeply rooted ambitions of social anthropology: to reach a rational understanding of the forms of thought that we find enacted in ethnography. Classically, these forms of thought, from what Lévi-Strauss called the “systematization [of] what is immediately presented to the senses” ([1962] 1966: 11), to the causal theories studied by Evans-Pritchard, for instance in Zande witchcraft and oracles (1937), have been interpreted, at least since Boas (1989), as an expression of a specific language or “culture.” To use a more recent terminology, thought has been linked to “ontologies” associated with certain languages and societies. This is why we still commonly speak of “Chinese,” “Greek,” “African,” or “Amazonian” thought.


How to Do Things with Words / J.L. Austin

From a theoretical point of view, this way of defining thought calls for an epistemological preliminary remark and for three objections. The preliminary remark concerns the definition of thought itself From Lévy-Bruhl’s considerations on “prelogical mentality” (1949), up to Sperber’s arguments on apparently irrational beliefs (1982), a great part of the anthropological literature devoted to this topic does not really concern the study of thought as a general human activity. It concerns the opposition between rationality and irrationality. In this perspective, anthropologists usually compare an abstract definition of “rationality” with an empirical counterpart, mostly founded on the analysis of some forms of categorization and theories of causality. It is obvious, however, that there is much more to human thought than categorization, or propositional rationality. Ideas about perception and space, language and communication, right or wrong moral values, for instance, are constantly present in ethnography. It would be hard to qualify them as “rational” or “nonrational” (or even “symbolic”). As we know, at least since Austin (1975), concepts of this kind would be better qualified as “appropriate” or “inappropriate,” “felicitous” or “infelicitous” in a certain context, than as rational or nonrational. In sum, when approaching the idea of an anthropology of thought, there is a preliminary choice to make. Either one chooses what we may call a Piagetian model of thought-as-rationality, seen in its various manifestations, but defined only through the opposition between rational or nonrational (e.g., Piaget [1923] 2001, [1926] 2007); or one refers to a more extensive, and more realistic, definition of thought. One of the classic authors who have worked in this direction (and whom we could, in this respect, oppose to Piaget) is Vygotsky, the great Russian psychologist (Vygotsky 1978). Not unaware of the problems posed by cultural differences, Vygotsky elaborated a multifaceted conception of the exercise of thought, which includes not only rational inference, but also metalinguistic, metacommunicational, aesthetic (“thought through our eyes,” as Joyce defines it), and narrative thought. In this exploratory and speculative paper, I will take, as a starting point, this Vygotskian option, and try to develop it in a new direction. But let us first examine the three classic objections to the definition of thought, so common in our discipline, as directly linked to culture, language, and society.

The first objection is empirical and has been known at least since the works of Sapir (1985) on North American cultures. Societies sharing the same “system of thought” may speak different languages, and vice versa. Thus, we know of cases (consider, for example, the Quechua-speaking peoples of the Amazon—Gutierrez Choquelvica 2010, 2011) in which language is not necessarily a good key to understand culture.

The second objection has a more theoretical character. The idea that we can establish a direct relationship between thought and language has, in many ways, proven to be logically weak. In his classic essay about translation, Jakobson has an amusing episode illustrating this point. “In the first years of the Russian revolution there were fanatic visionaries who argued in Soviet periodicals for a radical revision of traditional language and particularly for the weeding out of such misleading expressions as ‘sunrise’ or ‘sunset.’ Yet we still use this Ptolemaic imagery without implying a rejection of Copernican doctrine” (Jakobson 1959: 234). Jakobson’s conclusion is clear: if a relation between language and thought exists, it is an indirect and controversial one, and we should never take it for granted (or infer qualities of thought from language structures) without further investigation.


The Structure of Scientific Revolutions / Thomas S. Kuhn

The third objection is that the languages that we use to qualify different kinds of thought are constantly translated. Despite all its difficulties, translation in all its various forms, from translation of different languages to “translation of different systems of thought” (as analyzed, for instance, by Kuhn [(1962) 2012] or Lloyd [1996, 2006, 2007]), is a cognitive task that the people we study are often and in many ways confronted with. As Jakobson again remarks: “Both the practice and the theory of translation abound in intricacies and from time to time attempts are made to sever the Gordian knot by proclaiming the dogma of untranslatability. . . . [However,] all cognitive experience and its classification is conceivable in any existing language” (1959: 232). One might think that in the classic debate that opposes relativists to universalists, Jakobson is here the taking side of universalism. However, his argument on translation is more nuanced that it may appear at first sight. His position relies on a distinction that both universalists and relativists rarely pay attention to. Jakobson remarks that since, as Boas (1938: 127) has observed, “the grammatical pattern of a language (as opposed to its lexical stock) determines those aspects of each experience that must be expressed in the given language,” languages “differ essentially in what they must convey and not in what they may convey” (Jakobson 1959: 235–36). For instance, many North American Indian languages encode a distinction between visible and invisible, as in this Kwakiutl example:

T’e’semgya     “this stone (visible, near me)”

T’e’semgya’     “this stone (invisible, near me)”

Since these languages distinguish visibility of the referent to the speaker, the speakers of such languages are forced to attend to the visibility or invisibility of the objects they refer to. To designate a stone in Kwakiutl, one must mention whether it is visible and close to the speaker, or not. This does not mean that Kwakiutl, as a language, could not express the same “cognitive experiences” that are commonly expressed in languages, like European languages, which do not encode visibility in demonstratives (cf. Hanks, this issue). Despite a number of constraints concerning what they “must convey” (aspects of reality they “have to” express in words), all languages are translatable and constantly translated. From these considerations, one can draw the conclusion that universalists are right when they affirm that “all cognitive experience . . . is conveyable in any existing language” (Jakobson 1959: 234) and relativists are not entirely wrong when they underline that different languages obey different grammatical constraints “that determin[e] those aspects of each experience that must be expressed in the given language” (ibid.: 235–36).

Once these points are granted, however, one might still wonder whether the potential translatability of all languages is a good reason to stop minding, or (as some universalists would argue) even thinking about, the kind of cultural difference which is thus expressed by language use. It is true that one difficulty in accounting for these grammatical differences in theoretical terms lies in the fact that it is often hard to understand their raison d’être. They seem to obey no general rule. Differences in grammatical patterns might be episodic. Hence, they seem to have no general (theoretical) import. One can simply disregard them, or consider them curious but hardly relevant for an analysis of thought necessarily based on general principles.

To respond to this objection, one might use the distinction, currently used in logic, between the power (the possibility to account for a limited number of features valid for a great number of cases) and the expressivity (the possibility to account for a great number of features belonging to a limited number of cases) of symbolic systems (Mangione 1964: 52–53). Any case-centered inquiry (e.g., a fieldwork-based ethnography) needs to be in some measure expressive, while any comparative or statistical analysis needs to be reasonably powerful. With this distinction in mind, one could say that all human natural languages have potentially the same logical power, while they constantly differ in degrees of expressivity. This not only means that the grammatical differences between languages are specific forms of a general logical property of all symbolic systems (“degrees of expressivity”), not simply “episodic” or contingent phenomena; it also means that (as the indecisive results of the debate about linguistic relativism also indicate) the controversy between universalists and relativists, if still formulated in traditional terms, might well prove to be quite undecidable. Seen from this perspective, the problems posed by cultural translatability would not be solved by taking a universalistic attitude, trying to eliminate different degrees of expressivity that we find in different languages. Equally unproductive would be a relativistic theory that refused to admit any general property of human languages. In this paper I would like to show that a good account of the question, and a solution of the controversy, would rather lie not in the elimination of one of the two aspects of the question, but in the possibility of understanding the many ways in which logical power and expressivity, in different languages and in different semiotic codes, may relate with each other. In short, more theoretical and empirical work is needed not only to solve the problem of the translatability of cultures, but also to formulate it correctly.

I will argue that, instead of seeing the possibility of translation as a theoretical difficulty for defining thought, we could, on the contrary, consider the ethnography of translation as a chance to observe the dynamics of thought processes, and to study how they operate, both in adapting to constraints and in exploiting possibilities, in different cultural contexts. From this ethnographic perspective, the question of understanding the kind of cognition that might be involved by the use of “a” language (with its own specific degree of expressivity), or by the formulation of “a” specific ontology (or “system of thought”), ceases to be the only question we are confronted with. Another question, equally important, arises: How are we to describe the kind of cognition that is constantly mobilized in the process of translating languages (and in passing from one “ontology” to another)? To use Jakobson’s terms, how is it possible to pass from what a language” (or any other symbolic system) must convey” to what it “may convey”?

Furthermore, the distinction between what a symbolic system “must” or “may” convey is not necessarily confined to semantic and grammatical questions, or to cultural differences. Ethnography constantly shows (and Jakobson also admits) that there is more to translation than language. Processes of “translation” (involving specific cognitive tasks) operate not only between different cultures (or languages), but also between different pragmatic contexts in the same language, and between linguistic and nonlinguistic ways of expression, even within single societies. Thus, a second series of questions related to the question of translatability arises: How can we describe these forms of cognition? Are they identical, comparable, or totally different from the cognition involved in linguistic translation processes? Does the logical distinction between what “has to” be conveyed and what “might be” conveyed also apply to this context-to-context or verbal-to-nonverbal form of translation?


Why Suyá Sing:  A Musical Anthropology of an Amazonian People / Anthony Seeger

Obviously, to try to give a full answer to all these general questions in a single paper would be unreasonable. I will, then, limit my argument to a single kind of translation, as it operates in a specific ethnographic area. Using three Amazonian examples, I will try to describe the kind of cognition involved in the form of translation that Jakobson calls transmutation. I will argue that from this ethnographic analysis, we can not only derive a better (both wider and more precise) idea of some, rarely studied, cultural translation processes, but also draw from it a new way to define the concept of “cultural ontology.” The anthropology of the Amazon offers an ideal field for this kind of analysis. In the last twenty years, at least since the publication of Descola and Taylor’s “La remontée de l’Amazone” (1993), the question of the relationship between iconographies, narrative structures, ritual chants, and, in general, the pragmatics of the transmission of knowledge has been intensely and productively debated in this area of study, The groundbreaking work of Rafael José de Meneses Bastos (1978, 1999, 2007) has shown how music performed in ritual action can function as a sort of lingua franca in the Upper Xingu, providing for a common ground of shared knowledge in a multilingual group of societies where a pidgin was never invented. A crucial corollary of this general conception is that the original source of music is not human, but essentially animal (e.g., Beaudet 1983, 1997; Brabec de Mori and Seeger 2013). Humans generally “learn” or “acquire” their music from nonhumans. This is why music is also used for communicating with spirits.

The group of researchers first gathered by Vidal, Pessis, and Guidon (2000), probably inspired by the fundamental work of Guss (1986, 1989), and subsequent work by Gow (1988, 1999), Barcelos Neto (2002, 2008, 2009, 2011, 2013), Taylor (2003), Velthem (2003, 2013), Lagrou (2007, 2009a, 2009b, 2011, 2013; Severi and Lagrou 2013), Belaunde (2009, 2013), Cesarino (2011), Fausto (2011a, 2011b, Langdon (2013), and Fausto and Penoni (2014), among many others, have shown that myths cannot be used as “captions” of iconographies, nor can images or artifacts be understood as illustrations of myths. A complementary relation exists in Amazonian iconographic practices, just like in other kind of Amerindian iconographies (Severi 2012), between myths, ritual chants, and the drawings, picture-writings, or body-decorations related to them. As a consequence, iconographies are no longer seen as redundant decorations. They are understood as “variations” of the same “conceptual imagination” that generates mythical narrations (Barcelos Neto 2013: 181; Severi and Fausto 2014). Eventually, as, for instance, Meneses Bastos (1978), Basso (1981), Beaudet (1997), Piedade (2004), Seeger (2004), Fausto, Franchetto, and Montagnani (2011), and Brabec de Mori and Seeger (2013)have also shown, synesthesia is everywhere in the Amazon. Not only do complex verbal compositions, like shamanistic chants, always presuppose the experience of vision (e.g., Luna 1992; Townsley 1993; Hill 1993, 1994, 2009; Severi [2007] forthcoming), but “what can be seen as an image” can always be perceived, by another subject and from another perspective (Viveiros de Castro 2004), as a sequence of sounds. This is why, as Barcelos Neto has recently shown, the image of a mythical anaconda can be interpreted, among the Wauja, simultaneously as a sequence of graphic themes and as a sequence of chants (Barcelos Neto 2013: 183).

How can we understand this situation of constant “synesthetic fusion” (ibid.: 187) where “what is seen” can be constantly translated into “what is heard,” and vice versa? What happens when the same concept (often expressed by a proper noun) is “translated” from verbal expressions to images and from images to sounds?

Forms of translation: Definitions


Mind and Nature:  A Necessary Unity / Gregory Bateson

Let us get back to Jakobson. He has defined three forms of translation: intralinguistic, interlingual, and transmutation. According to him, “intralinguistic translation or “rewording” is an interpretation of verbal signs by means of other signs of the same language,” “interlingual translation or translation proper is an interpretation of verbal signs by means of some other language,” and “intersemiotic translation or transmutation is an interpretation of verbal signs by means of signs of nonverbal sign systems” (Jakobson 1959: 233). A very rich literature in linguistics and in anthropology has been devoted to the intricacies, both practical and theoretical, of the two first kinds of translation (and to the general question of linguistic relativism that they imply).1 The third form has been, by far, less studied. In the paper I have just quoted, Jakobson himself tends to consider it only a theoretical possibility. He is far from being the only scholar who adopts this attitude. When approaching the subject, the great majority of authors—with few notable exceptions (among them Goodman [1976] and Baxandall [1993])—avoid any attempt at detailed analysis. Some authors simply confuse transmutation with its reverse: verbal comment on visual or acoustic images (see, e.g., McGaffey in Rubel and Rosman 2003: 257–58). Others affirm that this form of translation, because it relies upon heterogeneous codes of signs (verbal and nonverbal), can hardly generate a consistent way to represent knowledge. Others (among them Wittgenstein [(1914–16) 1974] and Bateson [1979]) think that since an iconic code is not a means of communication comparable to writing, no cultural tradition, or transmission of knowledge, can be built on it. As a result, transmutation is seen as either too arbitrary or too subjective to be really compared to linguistic translation.

We anthropologists cannot afford this attitude. The ethnography of “oral” traditions often confronts us not only with consistent, effective, and long-lasting systems of interpretation of verbal signs by means of images (like, for instance, Plains Indians picture-writings, the Andean khipus, or Nahuatl pictography—Severi 2013), but also with a fourth variety of translation, which we could call transmutation proper. In many Amerindian cultures, for instance, we find that the interpretation of signs belonging to a nonverbal system can also be realized by means of signs belonging to another nonverbal system. For instance, a statement or a notion usually expressed through words can be first “translated” into images, and then further “translated” (one should say “transmutated”) into music or ritual gestures. My purpose is to show:

  1. that in these cultures, transmutation, far from being “arbitrary” or “subjective,” has general technical (semiotic) properties that generate a specific logical form that we shall define as a multilayered four-term analogy; and
  2. that the analysis of cultural forms of transmutation can reveal a special kind of “cognition about ontology” that leads to the construction of inter-specific beings.

I will take here the example of three Amerindian iconographic traditions from the Upper Orinoco region (Yekwana, Wayana, and Wayampi) where the process of transmutation of narrations into visual images, and then of visual images into sequences of sounds (“transmutation proper”), is developed in particularly interesting ways, both in iconographic and in musical traditions.2 The Yekwana and the Wayana are Carib-speaking hunters and horticulturalists from the Upper Orinoco region of Venezuela and Brazil. The Wayampi, who speak a Tupi-Guarani language, are neighbors of the Wayana and belong to the same cultural group. I will use the first ethnographic case, Yekwana weavings, to identify some basic formal (or semiotic) features of transmutation as a nonarbitrary and nonsubjective form of translation “from verbal signs to nonverbal signs.” I will use the second and third cases, Wayana iconographies (which are an interesting and consistent development of the Yekwana visual tradition) and Wayampi music, not only to confirm the formal features of transmutation (and transmutation proper), but also to raise some new questions concerning the concept of ontology, and the kind of thought which is expressed through these iconographies.

Yekwana weavings and mythology: Two formal properties of transmutation

Yekwana mythology (de Civrieux [1970] 1997; Guss 1989) is composed of a long cycle of tales describing the various bloody episodes of a conflict that is seen as governing the entire universe. The conflict is between Wanadi, a positive being associated with the sun, who presides over human material culture (agriculture, fishing, hunting and tool-making), and his twin brother, Odosha, who is a personification of evil, misfortune, illness, and death. This cosmic battle is not simply responsible for the creation of the universe, but has continued unabated since the beginning of time, and still affects everyday human existence, often with tragic consequences. According to the Yekwana, evil always triumphs over good, which is why their ally, Wanadi, lives in a distant part of the heavens and has limited contact with the human world below. In contrast, his evil twin, Odosha, who lives surrounded by demons (often represented as invisible animal and plant “masters”), is a constant, threatening presence. This explains why representations of Odosha include a wide range of different, maleficent creatures (howler monkeys, snakes, jaguars, and foreign cannibals), whereas Wanadi, the sole defender of humans, is holed up in his corner of the heavens. Indeed, the Yekwana consider that all hunting or fishing or agricultural activities must be carried out against the will of a host of “invisible masters” of plants and animals. This world of potential enemies belongs to Odosha and his demons. This basic asymmetry between good and evil is coupled with an idea that the one constantly transforms into the other. For the Yekwana, all cultural achievements (weapons, weaving techniques, body-painting, etc.) are the result of a transformation of evil or of the creatures that rely on it. This means that all creatures and creations are necessarily ambiguous as everything that is useful or good contains a transformation of some evil being.

As David Guss (1989) has shown, the visual memory associated with this mythology is linked to a specific iconography, woven into twill-plaited basketry, which constitutes a sort of “catalogue” of the names of these creatures. Among the Yekwana, the skill of weaving baskets “measures the maturity and character of any developing male member of the society” (ibid.: 79). To weave baskets has, for Yekwana boys, a veritable initiatory role that can be understood as a cycle of ritual actions. A young man cannot become a husband without learning this technique and the knowledge of the graphic patterns that is associated with it. At his marriage, every young man must weave for his bride a series of baskets in a strict prescribed order. With each basket are associated symbols of rebirth, health, and purity, but also threats of death. Actually, depending on the design inscribed in it, a basket can feed a person, but it also can poison him or her. This is the reason why the choice of a graphic pattern for a basket has to be made with extreme care. To choose the decoration of a basket, the husband has to consult his father, who usually is the human “owner” of the design. The father will hand to him the right to weave a number of designs into the surface of the basket (ibid.: 81–82), and the young man will have to weave them for all his life. In this sense, writes Guss, a basket design might assume the importance of a family crest passed on from generation to generation, though its real function is, for the Yekwana, to define a couple’s identity, representing in a durable way what has been until then an amorphous and transitional relation. As long as husband and wife remain together, the special images woven into the basket will be a clear statement of the strength and uniqueness of their bond (ibid.: 82).

Actually, the twill-plaited baskets, decorated with designs that every man has to weave to prepare for and confirm his marriage (and to accomplish his male initiation), are strictly connected with the ritual relations that humans entertain with nonhuman and mythical beings. The baskets incorporate a complex system of symbols that acts as an index and key to the rest of the culture. This point explains, as Guss also remarks, why “the most accomplished ritual singers and the most skillful basket makers are inevitably one” (ibid.: 85). Actually, baskets are generally said to be the property of nonhuman supernatural “masters.” But this notion of property often becomes much stronger: baskets as artifacts are themselves said to be “embodiments” (ibid.: 102) of the mythical beings. Like the ancestral predators they incarnate, they are “living beings” that can attack humans. Their designs woven into their surface are the “body paints” that decorate the skin of the mythical predators (ibid.). “The identification of the baskets with the demonic forces of Odosha is reaffirmed in every story in which they appear,” writes Guss. “When a Yekwana narrated the origin of the baskets to the French explorer Gheerbrant, the power he ascribed to them was that of Odosha himself. Baskets did not simply signify death, they actually caused it” (ibid.: 103). A myth quoted by Guss confirms this point very clearly. When the artifacts appear in the narration of the origin of the world, they are immediately shown to be living beings with decorated skins: “The baskets began to walk, and they entered the water [of a river]. They were caiman-alligators—you had only to look at their skins to see that” (Gheerbrant 1954, cited in ibid.: 103).

We will get back to the kind of agency, connected to ritual action, which is attributed here to the Yekwana artifacts. We will see that many other artifacts of this kind, in all our Guyana cases, are ritually endowed with life. Let us focus, for the moment, on the interpretation of the graphic patterns appearing on the baskets and related to the chants, “mostly composed of lists of names of mythical beings” (Guss 1989: 36), that accompany their weaving. It is remarkable that, rather than trying to represent a particular mythological event in a “realistic” way, Yekwana weaving organizes mythological knowledge at a more profound level: in the iconography, each being is graphically linked to its invisible side. Let us see how. As we have already noted the two central tenets of this mythological system are a constitutive opposition between two principal types of creatures (good and bad) and the idea that a process of continual transformation affects them. These metamorphoses take two forms. On the one hand, a multiple being such as Odosha may “take the form” of a whole series of other creatures—in which case we see a movement from an individual to a series. On the other hand, this process of constant metamorphosis (wherein good is necessarily a transformation of evil) gives rise to individual creatures possessed of an inherent ambiguity that makes them simultaneously positive and negative—in which case the movement is from a series of creatures to one complex being that synthesizes them.

Yekwana iconography proposes precise visual translations of these two organizational principles. Indeed, all visual themes representing spirit names are derived from a single grapheme: a sort of inverted “T” that represents Odosha (Figure 1). A few simple geometrical transformations allow all other mythical characters to be derived from this grapheme. This conveys the idea of the creatures’ singularity (as monkeys, serpents, toads, etc.) as derivations of an elementary pattern (Figure 2). In this way, the different characters are developed out of a single basic form in a system that is capable of representing not only specific characters, but also their possible relationships. These relationships (of analogy, inclusion, and transformation) bespeak an internal organization clearly predicated on a single criterion: the representation of the potential plurality of all creatures. But this is not all. The visual technique outlined above implies the interplay between forms (or between form and background)


Figure 1. Odosha (from Guss 1989: 172).


Figure 2. The Toad and (below) the Frog (from Guss 1989: 201).

that allows for the simultaneous representation of a specific creature and one of its potential metamorphoses. Several mythical characters (e.g., monkeys, bats, toads) can thus be represented as potentially dual beings. An example of this “chimerical character” of the being represented by this iconography is the graphic theme called woroto sakedi (“jaguar mask,” Figure 3), which simultaneously represents Odosha and Awidi, one of his serpent avatars, depending on whether one focuses on the form of the T, which functions here as a frame, or on one of its segments, which represents, with its spiriform pattern, the enrolling of Awidi, the coral snake. Let us briefly analyze now the formal properties which, in this case, preside over the process of translation of narrations (“verbal signs,” here represented by proper nouns) into images. It is clear that in the Yekwana case the passage from verbal to nonverbal code does not involve a simple equivalence between code-units. Rather, this form of translation mobilizes two equivalent messages in two different codes, language and conventional iconography. Each code is organized following its own rules. In other words, there is indeed semiotic heterogeneity. Nonetheless, the relation between the two codes is not arbitrary, nor episodic. The technique that enables the Yekwana weaver to realize the passage from verbal to nonverbal signs


Figure 3. Awidi, the snake, and Odosha, combined in the pattern known as “jaguar mask” (from Guss 1989: 182).

actually follows at least two basic features. The first is selectivity. Not every sign belonging to a narration is “translated into” images in the weavings—only the nouns of the mythological creatures are translated. Verbs or adjectives, for instance, are never represented in visual terms. The second feature is visual redundancy. The visual pattern woven into the surface of a basket not only represents the name of the creatures of the myth; it also reveals, in the case of the first series (Figure 2), their relationship to the “elementary pattern” (Odosha/Wanadi) they all derive from. In the case of the “jaguar mask,”” the “chimerical” creature that associates Odosha and Awidi in a single image, the image is generated by the superposition of Odosha and the “coral snake” pattern (Figure 3). In both cases (the series and the “chimera”), the geometrical pattern on which the image is based is redundant, since it “gives hints” about the nature and mutual relationships of mythical creatures that are not present in their names.

I have shown elsewhere (Severi 2013) that the two basic features of selection and redundancy play a constitutive role in American Indian picture-writing, and that they can generate more complex configurations. But even from this first example, we can conclude that “transmutation” in Amerindian iconographic traditions, even if it does not “follow rules” in the same way that the grammar of a language does, can be shown to be—quite unexpectedly—logically consistent and, in its own way, systematic. We can draw the conclusion that in the Yekwana iconography, the passage from verbal to nonverbal signs is neither arbitrary, nor subjective.

Wayana iconographies: Logical form and ontology

Let us now turn to the Wayana. They share with the Yekwana the technique of weaving and a very similar notion of iconographic representation (and even particular graphic themes, such as that of the jaguar [Velthem 2003: 352–56]). For both groups, iconographic representation is an elaboration of simple geometrical forms such as triangles, squares, spirals, and intersecting or parallel lines, and for both groups, this type of representation concerns the commentary and memorization of myths, and has close connection with ritual action. For the Wayana too, weavings are potentially living beings, and can become active in specific situations. What distinguishes Wayana from Yekwana iconography is the complexity of the discourse surrounding visual representation. Four concepts play a central role in this context. The first is wayaman. For the Wayana, a geometric theme woven into the surface of a basket is not merely the sign or emblem of a mythological being; it is also the reflection of a specific form of knowledge known as wayaman, which is metaphorically situated in the pupil of the person who masters weaving techniques. Wayaman is an “inverted figure” of an anthropomorphic spirit present in the pupil of the basket-weaver, and it is the wayaman, not the person who made it, who is the object’s true “author.” The wayaman is conceived of as a type of “thought,” but also as a reflection of this “other” who lives in the weaver’s eyes and “guides his hand,” and it is only truly revealed when the object is created in accordance with traditional rules. Once the form is completed, then the object will reveal its true nature and show itself to be “like a living creature,” and the “property” and incarnation of an ancestral nonhuman being.

The second concept concerns what the Wayana call the “skin” of the image. Actually, in Wayana tradition, artifacts, humans, and nonhumans can, and sometimes must, be adorned in the same way. In these cases, they “adopt the same skin.” This is a key notion because for the Wayana the skin, or rather the skin painted with a recognizable pattern, represents “that element that allows for the identification of a being’s actual nature” (Velthem 2003: 129). Thus, if some ritual artifacts are thought of as “copies” or “imitations” of ancestral predator beings (such as anaconda, vulture, and jaguar), it is because they bear the same skin. Because of this “identity of design” (and of the wayaman they incarnate), artifacts (as in the Yekwana case) can “dance”, “talk,” and even “attack” like predators. To illustrate the complexity and flexibility of this notion of “skin,” Velthem cites the example of the dances held in the men’s ceremonial hut. The men’s ceremonial hut is supposed to be “inhabited by certain fishes,” who feature (alongside numerous other animals) on the central ceiling wheel of the great ceremonial hut (Figure 4). But the fishes are also represented as “bearing the skin of long-beaked hummingbirds,” and so when masked men, during their dances, “act like fishes,” they also become “long-beaked hummingbirds.” To be more precise, they then adopt the skin of a series of beings: fishes, long-beaked hummingbirds, and young male human beings.

The third concept related to Wayana iconography refers to a particular way to categorize “supernatural” beings. The idea of a potential and unceasing transformation of all beings is widespread throughout the Amazon. We have seen that among the Yekwana, this is expressed via the opposition between two enemy brothers, Wanadi and Odosha, who represent good and evil, respectively. The Wayana share this idea. However, for them, predators and nonpredators are not individual characters with distinct personalities. Where the Yekwana rely on paradigmatic personalities, the Wayana think in terms of classes. Instead of contrasting a Wanadi to an Odosha, they distinguish between different modes of existence that can be applied to all creatures, be they animal, vegetable, human, or artifact. Consider the anaconda, one of the classic predators. “Its acts of predation,” Velthem notes, “are so paradigmatic that not only do they invariably evoke the wider supernatural dimension, but they can also refer to the acts of any other species.” This notion of predator-as-paradigm is not limited to the anaconda. It is usually extended to other


Figure 4. A central ceiling wheel of a Wayana ceremonial hut. (Geneva, Musée d’Ethnographie. Photo: J. Watts.)

predators. “This conception”, Velthem continues, “allows other creatures, such as caterpillars, centipedes, fishes, and birds, to display predatory instincts in a supernatural setting via their association with jaguars, vultures, or anacondas” (ibid.: 105). In such cases, the anaconda (or the jaguar or the vulture) will “bear the name, the wayaman, and the skin” of the animals in question.

This type of categorization is also present in language. Velthem remarks that “this coupling of creatures is linguistically signaled for instance, by the suffix okoin, which means ‘qua anaconda’ and is applied to a specific species” (ibid.: 105). So kiap (the toucan) becomes koimë or “toucan-qua-anaconda” and is represented by a long-beaked serpent whose skin is covered with feathers of different colors. A similar process also exists for the jaguar, whose presence is signaled by a different suffix (kaikuxin), which marks the transformation of animals like the rodent quatipuru into “rodent-qua-jaguar.”

Sometimes, these complex definitions are interpreted as referring to “qualities,” or “gradient of qualities,” belonging to different species (Viveiros de Castro 1998; Lima 2000). However, qualities are partial properties attributed to a (logically preexisting) object. For instance, in a statement like “this butterfly is red,” I suppose the potential existence of other properties, such as “light,” “noisy,” “flying,” and so on. When I designate, as the Wayana do, “a toucan-qua-anaconda,” I am using a being, not a property, to designate the mode of existence of another being. Instead of enumerating the properties of a single being, I am connecting in a single statement two different beings. The result of this connection is a plural creature, not a series of qualities belonging to a single being. Consider the example we have just mentioned of the dances held in the men’s ceremonial hut. When masked men “act like fishes” and thus behave also “like long-beaked hummingbirds,” they give birth to new, ritually generated complex creatures, not to an enumeration of the possible qualities of a preexisting being. The same is true for the twofold (or serial) characters of the Yekwana baskets and, as we will see later, for the “acoustic complex beings” of Wayampi music.

Actually, we should understand terms like “toucan-qua-anaconda” as “verbal chimeras” that describe composite and changing beings that belong to a common class by virtue of their suffix. The concept of a “series,” which is also present in Yekwana iconography, here assumes a different aspect, for the Yekwana series are, so to speak, linear series of beings. In the Wayana series, beings are embedded in each other. Thus, in another kind of dance, linked to the initiation of young men, the initiate wears a series of masks that transform him into a composite being made up not only of different spirits (macaws, falcons, fish, sun, rainbow, etc.), but also of different forms of these spirits “qua” incarnations of different predators: jaguars, vultures, and anacondas (Velthem 2003: 212). In this double series of markings on the initiate’s body, the concept of chimerical representation reaches unprecedented levels of complexity. The ritual becomes a site of transformation wherein masked young men progressively “assume the painted skin” (and the wayaman) of a whole series of animal, vegetable, and human spirits that are themselves subject to innumerable metamorphoses.3

Let us see some examples of this kind of representation in iconography. Wayana graphic themes are divided into three distinct categories: those that “belong” to anaconda body-decorations; those that are linked to the skin of the jaguar; and those that evoke the skin of “anthropomorphic monsters” (a category that includes enemies such as white people). Particular visual motifs that retain their specific referent are then used to identify groups or entire categories of creatures. For instance, writes Velthem, “one of the paradigmatic forms of predation is the act of ‘wounding, stabbing or piercing’. The act that synthesizes such predation (’to pierce the skin with a projectile’) is characteristic of an artifact (the arrow) as well as of several animals, including cobras, wasps, scorpions, and birds such as the maguari stork (Florida caerulae)” (ibid.: 322–23). This bird is recognized as the prototype of piercing creatures and is represented by a motif called the “maguari beak” (Figure 5), whose outline depicts the animal’s “wary and attentive posture.” “In fact, this graphic theme represents both the arrow as artifact and any predatory animal that can strike its prey like an arrow. The double arrow symbol, then, describes fairly indeterminately ‘everything that pierces’” (ibid.: 183).


Figure 5. The motif called “maguari beak” (Wayana) (from Velthem 2003: 322).

In other cases, we can find the same principle differently deployed. We have seen that the “maguari beak” is a single symbol that designates several beings. But “Wayana graphic patterns can be themselves composite” (ibid.: 313). In such cases, the image can be broken down into several parts, each with its own distinct referent. So, for instance, the theme “crab” (Figure 6) also contains the theme “tapir’s eye.” Interpretation then relies on what Velthem calls an “internal dialogue” of forms that takes place within the graphic theme itself.4

The fourth fundamental Wayana concept related to iconicity is the distinction between the ukuktop (or “perceptual image” of an animal that can be observed in the forested environs of a village, with its morphology, normal behavior, food preferences, etc.) and the mirikut (the graphic theme that represents the animal


Figure 6. The theme “crab” combined with the theme “tapir’s eye” (Wayana) (from Velthem 2003: 312).

in traditional weaving). “Though all mirikuts are, of course, images . . . not all images are mirikuts” (ibid.: 317). The mirikut allows one to interpret (or “decipher from its painted skin”) the “true nature” of an animal. The geometric theme does not (just) represent the (familiar and essentially harmless) animal, but also its “normally invisible and monstrous double” (ibid.). This is further evidence of the essentially serial nature of Wayana iconography: a creature or being can never be understood only in terms of its singularity. Its “painted skin” always defines it as a member of a class or of a sequence of possible “modes of existence.”

From a technical (semiotic) point of view, we can conclude that the Wayana technique of “transmutation” is, in its own way, selective and redundant. Wayana iconography is, like the Ye’kwana’s, linked to the representation of lists of proper nouns—whether it represents specific entities or logical series. Furthermore, the Wayana have conceived a recursive principle that enables them to “embed” classes of beings in other “classes of beings.” From the ontological point of view, both the Yekwana and the Wayana cases follow the same logic. The “invisible aspect” of nonhuman beings (“what really is there” in ontological terms—beyond their appearance) is shown through the construction of composite pictures, constituted either by individual “complex” figures such as Wanadi/Odosha among the Yekwana, or by serial beings, or even “classes of serial beings embedded in each other,” as among the Wayana. It is also remarkable that images are, in both cases, always perceived as displaying their nature and power during ritual action (as in the case of Wayana dances of initiation), or in direct connection to it, as for the Yekwana weavings, which are also seen as living beings.

The music of the Wayampi: An example of “transmutation proper”

The underlying logic of this process of transmutation of concepts concerning “special beings” from names found in mythical narrations to nonverbal signs linked to ritual action can be taken yet further, to the passage from one nonverbal code to another. In the music of the Wayampi (who also have weavings, just the way the Yekwana and Wayana have music), we find a very similar way to represent the “real nature” of invisible predators as collective beings. The first point to mention is that Wayampi musicians perform names of spirits (currently used in mythological narratives) just in the same way that the Wayana and Yekwana represent them in visual terms. Actually, performing any music on an instrument, such as a flute or a clarinet, is, for the Wayampi, a precisely defined act of communication, primarily addressed to nonhuman beings. What we may call the pragmatics of ritual musical performances is, as a consequence, both complex and explicit. A Wayampi musician inherits the right to perform every single piece of music from a master or an elder member of his family. Every piece has a proper occasion in which it has to be performed, either by a soloist or by a group of performers, a link to a named place, and a relation to a specific nonhuman being (Beaudet 1997: 128). Furthermore, the music of the Wayampi is not only “addressed to” animals. It has, in itself, a nonhuman nature. To perform a piece of music, even the simplest melody, is to imitate the nonhuman “owner” (and inventor) of the music. In this sense, every musical performance, for the Wayampi, is a call. In its simplest form this “calling” involves a specific form of musical onomatopoeia, which selects a single acoustic trait (a melodic fragment, in musicological terms) to designate (and “call”) a specific animal. The call is the musical incarnation of its name. In order to “call” a toucan, for instance, one “sings” or performs (e.g., on a flute) a theme called “toucan”. Beaudet (ibid.), who gives beautiful examples of these calls, underlines that this “toucan theme” does not necessarily imitate only the “cry” of a toucan. Other aspects of the bird can be represented acoustically, like its elegance, its agility, the vividness of its colors, or the like. Such a theme can become, in other forms of daily-life compositions, the “signature” of a piece, also called “toucan,” where this precisely identified group of sounds, performed by a soloist, can be repeated and subjected to different types of variations.

When music is ritually addressed to an invisible spirit, this relationship between the performed motif and the being it addresses (and imitates) becomes stronger. In that case, the invisible spirit is no longer “only imitated” by the music. It “is” the music. The music becomes the only index of its presence during the performance. “When the anaconda hear its music,” Beaudet remarks, “he comes to listen to it” (ibid.: 137). This is the reason why performing ritual music can become dangerous. While playing, the performer knows that the spirit is there to check that “its” music is correctly performed. If the performance is wrong, the musician or even the whole village may be punished, or become ill (ibid.: 144–46). In Wayampi mythology (see, e.g., ibid.: 143), the knowledge of a certain piece of music (a song, an instrumental piece, or a long suite) is always presented as the result of a fragile agreement with the spirits, a sort of truce that rules out both sex relationships and aggression between them and human beings. Many myths narrate that animal spirits have given certain pieces of music to humans as a token of this agreement (ibid.: 156). Every performance reenacts the conditions of this agreement, and can consequently become dangerous. Not to remember correctly a piece of music, or the simple fact of performing it badly, is understood as a transgression, which might provoke a revenge, and the reactivation of a state of conflict between human and nonhuman beings.

Actually, the Wayampi distinguish between several kinds of musical “calls.” In certain cases (mostly nonritual performances), a simple group of sounds, imitating its cry, can be sufficient to “call” a bird, or a monkey, and even to establish a dialogue with them. In other cases, when the being called for is an important spirit, the structure of the music (which acquires, in this case, strong “shamanistic connotations,” ibid.: 172) becomes far more complex. Let us consider, for instance, the acoustic representation of a predator like the anaconda. To “make the anaconda present,” one has to perform a particular musical suite of themes, made of a sequence of pieces, performed by a group of clarinets (and/or a group of dancers). The structure of this composition is based on the alternation of individual pieces, each of them characterized by a theme and performed by a soloist, with the repetitions/variations of another single theme. This theme, collectively performed by a group of musicians, and repeated after the performance of each solo piece, characterizes the entire suite, and gives to it its name.

Let us have a look at the suite, called Moyotule, which acoustically represents the anaconda. From a formal point of view, it follows the Wayampi traditional pattern. It is an alternation of several pieces, each characterized by its theme and played by a soloist, with another theme which characterizes the suite and is performed collectively. Performed by a group of clarinets, the “anaconda” theme is slowed down and iterated with minor variations after the performance of each solo piece. Following the rule of alternation that we have seen, a number of pieces belonging to other beings are then inserted into the “large and collective” version of the anaconda theme. In the list that Beaudet (1983) has recorded of the names of the pieces included (Beaudet 1997: 139), a number of animals appear. Among them the falcon, the monkey, several kinds of birds, insects, mammals, and fish are “called for,” in order to construct an acoustic image (and to generate the indexical presence) of the invisible predator that “owns” this music. The “anaconda” theme collectively performed becomes thus (from the point of view of perception) a sort of musical background on which a series of themes—shorter and performed by a soloist—designating other animal species are embedded. The result of this process is a suite which “bears the name” (in Wayana terms, one could say that it “takes on the skin”) of the anaconda, where a sequence of other beings “existing in the form of the predator” is made present. An acoustic image of a “complex invisible being” is thus generated.

The formal analogy of the Wayampi music with Ye’kwana-Wayana iconographies is, of course, striking. The Yekwana represent invisible beings (like Odosha and Wanadi) either as complex compositions, in which different beings are embedded, or as series of other beings, resulting from the variation of a single form. The Wayana have developed this model, inventing more complex forms of variations involving classes of beings existing as visible manifestations of invisible predators. In Wayampi instrumental music, we find sequences of visible (perceivable) beings as indexes of other invisible beings. The sequence of their “calls” (as they acquire saliency from the background) allows one to infer the actual presence of the anaconda (or of other mythical beings) during the ritual performance of “its” music. The Wayampi seem to play with music the same game that Wayana and Yekwana play with images.

Transmutation and analogy

We can now try to draw some conclusions from the analysis of these ethnographic cases. Let us consider first the definition of transmutation as a cultural form of translation, and then the kind of “cognition about ontology” that is mobilized by these musical and visual ways to define complex nonhuman beings.

We have already seen that, in the Yekwana and Wayana cultures, “simple” transmutation (involving the passage from verbal to nonverbal signs) is both selective and redundant. In the music of the Wayampi, we can recognize the same features. The music of a Wayampi suite is selective because its use of onomatopoeia for designating a being results from a selection from all the possible aspects belonging to it (the musical form of its cry, but also its elegance, rapidity, etc.), that music “transcribes” in sounds. As we have seen, a theme (“signature”) of a piece always is a stylized portrait of a nonhuman being. But Wayampi music can also be redundant. The “anaconda” suite that we have briefly studied “tells” more about the nature of a predator than a simple name. It indicates that such an exceptional being as the spirit of the anaconda is described not by its acoustic appearance, but by a series of acoustic signals related to the different beings that indirectly designate its invisible presence. In both visual and acoustic images, the passage from verbal to iconic signs (or from one nonverbal code to another) mobilized by transmutation never limits itself to the description of the appearance of the beings it represents. On the contrary, the process of transmutation of words in images (be they visual or acoustic) makes the presence of supernatural beings indirectly perceivable through the appearances of other beings. To use again a Wayana notion, music and visual iconographies aim to construct mirikut, images of concepts and relationships, not ukuktop, imitations of appearances. Only through sequences of this kind does the nonhuman being represented (or made present) by music or graphic themes become perceptible, and thus imaginable, and even thinkable. The aim of transmutation proper is both to make relations between signs (be they technically interpreted as icons or as indexes) perceptible—and “supernatural” special beings imaginable as generated by relationships between them.

How is this realized? Can we describe a sort of method, a logical form presiding over these forms of transmutation, beyond the two basic operations of selection and redundancy that we have seen until now? Let us compare our three ethnographic cases and the cultural forms of transmutation they mobilize. In the Yekwana and the Wayana cases, iconographies woven into baskets tend to represent complex beings (designated by group of names, such as Odosha/Wanadi, Toad, Monkey, etc.) or whole categories of special beings through complex images. In the second form of transmutation (illustrated by the Wayana basketry comparison with Wayampi instrumental music), groups of sounds “translate” groups of visual themes into sequences of sounds. Music makes audible implicit relationships that collectively designate special (nonhuman and invisible) beings, which are thus “called for” and made ritually present.

It is remarkable that in all the cases we have seen, the passage from one code to another (language, iconography, music) is never direct. Never, as in the ancient and medieval theories of translation, is an “exact equivalent,” a sort of cast of every word (or sign, image, or sound), searched for. The object of the translation always is an intuitive relation between concepts. Yekwana weavings “describe” the antagonism between Odosha and Wanadi using a specific visual means to express opposition, be that the combinations of different shapes in one, or even the contrast between shape and ground. In other cases, Yekwana designs may describe connections of other sorts (relationships deriving from the transformations of Odosha) using visual ways to express them: analogies of patterns, inclusion, parallelism, and so forth. When we pass from the comparison of Yekwana and Wayana basketry to the comparison between Wayana weavings and Wayampi music, we discover that ordered sequences of images are “transmuted” into ordered sequences of sounds in the same way. Again, the object of translation (transmutation) is never the individual image, word, or sound. It is always the intuitive relation, previously established, between groups of sounds, images, and words. Using a concept formulated by Saussure, we could describe this process as the progressive construction of a four-term analogy (Saussure [1913] 2006) between relationships previously established in each semiotic code involved. This complex form of analogy would operate at two levels (Figure 7). At the lower level we could represent the relationships identified (through selection and redundancy) within each semiotic (verbal, visual, musical) code. At the second level we could represent the relationship established between these groups of relationships. To this last (and more complex) relation, which establishes a logical link between groups of analogies, and only to it, we could then give the name of transmutation. At both levels of our four-term analogy, only relationships represent relationships. Relations between sounds in music represent relations between images in iconographies; relationships between images represent oppositions (and other forms of connection) expressed in words, and so forth. When a higher-level relationship is established between groups of relationships, a transmutation is generated. What is represented then are not individuals, or qualities, or single actions, but similarities, oppositions, inclusions, derivations, and so forth. In other words, transmutation thus overcomes precisely the difficulty that we have seen so many authors affirm: the heterogeneity of semiotic codes. The four-term analogy operates, in fact, as a way to establish an order in the assemblage of these heterogeneous codes. From an abstract point of view, it thus assembles entities possessing the same logical nature and the same intuitive apprehension. The stuff transmutation “is made of” is relationships. We might add to this theoretical model the hypothesis that the distinction Jakobson has drawn between what must be expressed and what may be expressed in language applies to each level of our diagram. We could conclude that selection and redundancy are the first steps in a process of ordering relationships that can be represented, in all the ethnographies that we have studied, by a complex form of four-term analogy, constituted by two logical levels.


Figure 7. Transmutation and analogy.

Let us now turn to ontology, and to the consequences that our analysis of the transmutation processes might have for the definition of this concept.

Transmutation, cognition, and ontology

In a recent comment on Lloyd’s book Being, humanity, and understanding (2012), I have argued (Severi 2013) that the concept of ontology, as it has been defined in Western philosophy, is not fully understood by anthropologists. Many of our colleagues tend to call “ontology” any discourse about the origins and nature of the world. However, since Parmenides, the term “ontology” does not refer to the various material constituents of the universe (fire, water, air, etc.) and their different ways of combining. The ontological argument is about “being itself.” It aims to the construction of an “ontology” as a science of abstract principles (founded on the analysis of predicates of being such as necessity versus contingence, possibility versus impossibility, subsistence versus potentially, and the like) not as a discourse about the origins of what physically exists. Nor does Parmenides look for a classification of the different beings inhabiting the universe. He wants, on the contrary, to identify an abstract relationship between nous and physis, and looks for the conditions under which the world is thinkable. This is why a classification of the categories of different beings, following, for instance, the distinctions between animate/ inanimate, human/animal, male/female (which is often understood as “ontological” by anthropologists), technically does not make for an “ontology.” In the works of Aristotle, for instance, the study of these forms of knowledge belongs to the Parva Naturalia, not to the doctrine of Being, which is the object of metaphysics. Needless to say, Parmenides’ approach to being-as-being has been fundamental for the history of Western philosophy, from Aristotle to Kant, and still is discussed in very similar terms in modern philosophy (see, e.g., the famous debate on Russell’s “theory of descriptions” and its consequences for the relation between language and “what it is”: Quine 1943, 1948; Russell 2005; Carnap 2009).


Cognitive Variations:  Reflections on the Unity & Diversity of the Human Mind / Geoffrey E.R. Lloyd

I have come to the conclusion that, if we do an epistemological analysis of this concept in Western philosophy, the kind of world-visions anthropologists usually study are precisely not ontologies but natural philosophies without ontologies. In short, the risk involved in ignoring the theoretical import of this concept is to mistake Melanesian or Amazonian “conceptions of the world,” founded on “background commonly shared assumptions” about physical and social observable phenomena (Lloyd 2012: 67), for pseudo-Parmenidean ontologies. And, consequently, to understand them as coherent systems of thought: “unique, immobile, and unchanging” like the Parmenidean concept of Being.

I think that there are no empirical reasons to understand them as such. Only a wrong decision to view cosmologies as such may transform them into systematic ontology, or even (adopting an extreme Heideggerian idealism) into indigenous metaphysics. In this latter case, cosmologies would become no more than anthropological artifacts. In my view, what anthropologists tend to call “cosmologies” are de facto regularities in the establishment of a number of shared assumptions, very rarely expressed in the form of an explicit argument, and always related to specific practices, systems of relationships, and genres of discourse. They are linked to ritual, mythology, or daily-life contexts, and thus to the kind of semiotic means in which thoughts are formulated. These discourses might sometimes intersect, generating the appearance of a unitary “discourse on ‘the nature of what it is.’” But what is particularly interesting about them is precisely their semiotic complexity and their unsystematic character, the fact that they always leave a space open for different strategies of thought.

In this paper, I have tried to give an example of this strategic plurality of thought that characterizes the cultures we study, and to show what kind of “blind spot” the use of an uncritical concept of “ontology” can generate in our understanding of ethnography. By analyzing three examples of transmutation in three Amazonian traditions, we have been able to show that the passage from words to images and to music that “transmutation” enables is meant to designate the existence of plural beings. Among the Yekwana, this class of inter-specific beings is represented by key individual plural figures that dominate the myth cycle. Among the Wayana, these chimerical individual figures develop into classes (and sometimes classes of classes) of hybrid entities embedded in each other. In Wayampi music, alternation of themes, linked to the identification of different animals, replaces the Wayana embedding, but still generates an analogous form of designating complex beings. In all three cases, as elsewhere in Amazonia, this kind of ritual representation of the invisible is linked to the representation of living beings whose defining traits never entirely overlap with those of recognized human, animal, and vegetable species.

This kind of ontological ordering of beings is in sharp contrast with the way Amazonian ontology has been described until now. In a number of influential papers (e.g., 1998, 2004), Viveiros de Castro has claimed that the conception of the relationship between souls and bodies that we find throughout Amazonia should be understood as an interconnection of the different “points of view” generated by the “perspectives” belonging to human and nonhuman beings. He refers to this set of ideas (“for simplicity’s sake,” as he says) “as though it was a cosmology.” “This cosmology,” he writes,

imagines a universe peopled by different types of subjective agencies, human as well as nonhuman, each endowed with the same generic type of soul, that is, the same set of cognitive and volitional capacities. The possession of a similar soul implies 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 being endowed with the human figures and habits, seeing their bodily and behavioral aspects in the form of human culture. (Viveiros de Castro 2004: 6)

According to Viveiros de Castro, this strategic position of human culture in the relationship with nonhumans (the animals being here “the paradigmatic Other”) should generate a new epistemological paradigm. We should pass from the traditional idea of multiculturalism (where one single nature faces different cultures) to the idea of a “multinaturalism,” where many natural appearances are understood as sharing the same culture.5 As is well known, this theoretical approach, which invites the redefinition of “subjects” as “points of view” (ibid.), has been both influential and controversial. It is obviously impossible to discuss it here in full detail, but it is undeniable that Viveiros de Castro’s approach (as well as the work of Descola on this point) has transformed the somewhat immobile traditional distinction between Nature and Culture into a distributed system of differences, where both natural and cultural features or properties are interpenetrated. I would like to focus here on only one aspect of this theory: the kind of categorization which is implied by this system of differences. The central point, in this respect, is the role played by “corporeal differences” (Viveiros de Castro 1998: 470).6 In a system where there is only one culture (the human’s culture), Viveiros de Castro underlines several times that the difference between subjects, in this cosmology dominated by a “spiritual unity,” is given by the “specificities of the bodies” (ibid.: 470, 478).

In Viveiros de Castro’s argument, the concept of “corporeal diversity” has two different meanings. The first refers to the morphology of organisms. “The body of every species is unique,” writes Viveiros de Castro (ibid.: 478), for instance, a statement that, while suppressing individual differences within each species, supposes an identity between the concept of “body” and the notion of “species.” However, he attributes to the concept of the “body as differentiator” also a second meaning, which refers to “an intermediate plane” situated between “the formal subjectivities of souls” and “the substantial materiality of organisms.” “What I call body is not a synonym for distinctive substance or fixed shape; it is an assemblage of affects or ways of being that constitute a habitus” (ibid.). Viveiros de Castro here curiously uses this notion (originally introduced by Bourdieu [(1972) 1977]) to designate not only what a body looks like, but also “what a body eats, how it communicates, where it lives, whether it is gregarious or solitary, and so on” (Viveiros de Castro 1998: 478).7 However, when the concept of difference becomes crucial in his argument, species (both as “sets of habits or processes” [ibid.: 480] and as bodies that are morphologically different) are constantly referred to.8 In substance, as far as categorization of differences is concerned, Amazonian ontology is described as composed by classes of “beings,” typically humans versus nonhumans, mostly predators or the like, which define each other through their “bodily” differences. These binary oppositions critically (though, sometimes, tacitly) coincide with the appearance and typical behavior of species. This is why, in this cosmology, we constantly see human beings opposed to various forms of nonhumans, be they animals, plants, stones, or artifacts.

As is well known, the focus of this approach, and the paradigm of its epistemology, is myth. As Viveiros de Castro writes, myth is “the vanishing point where the differences between points of view are at the same time annulated and exacerbated.” This is why, for perspectivism, “myths takes on the character of an absolute discourse” (ibid.: 483). The study of our three Amazonian cases shows that, if we focus on iconographies linked to ritual action and the specific processes of transmutation that mobilize them, a radically different kind of ontological “way of ordering beings” emerges. The classes represented by Yekwana and Wayana iconographies, as well as the Wayampi “acoustic creatures,” do not coincide with species, to be opposed to one another. Quite the contrary, these images designate classes of special beings, where members of different species are associated in sequences. In all these cases of transmutation, a logic focused on the construction of composite beings possessing an invisible presence emerges. Instead of binary oppositions of “bodily” classes defined through relevant “natural” differences, we find the generation of inter-specific beings. The analysis of the forms of thought implied by transmutation leads to the conclusion that another form of “ontology,” based on very different principles, exists in the same area where perspectivism allegedly rules every cultural expression of meaning. We might call it a plural ontology for transmutating beings, linked to ritual action and visual thinking.

Translation and the anthropology of thought

This conclusion can help us in giving a more precise definition of what might become the horizon of a new anthropological theory of thought. Our analyses have shown that the crucial distinction that Jakobson has formulated between what must and what may be expressed in a language does not concern only linguistic (grammatical) patterns in interlinguistic translation, but it also applies to many other forms of translation, as the intersemiotic transmutation, and the visual and analogical thinking that it mobilizes. With this conclusion in mind, we could go a step further, and formulate the hypothesis that Jakobson’s logical distinction characterizes not only “language” and nonlinguistic codes, but also the exercise of thought itself. In this way, we could pass from an abstract opposition between “thought” (defined as rationality and categorization) and “language” (essentially defined as grammatical patterns) to the study of a set of multiple relations between forms of cognition (related, for instance, to ritual action and visual thinking) and intralinguistic, interlinguistic, and intersemiotic forms of translation. As we have seen, these forms of translation do not exist only between different languages, but also between different codes, and different pragmatic contexts, within a single culture. In this new perspective, variations in the use of codes and variations in the establishment of pragmatic contexts (not only variations between cultures and languages) would become sources for the variation of thought processes, and for the subsequent definition of “states of the world,” or of ontologies.

The essential plurality of “ontological” thought, of which we have seen an example in this paper, would thus find its general theoretical principle. “Worlds” vary when not only the abstract structure of a grammar but also many practices of language use (of which intersemiotic transmutation is an example) generate forms of thought where what must be conveyed joins, at a different level, what may be expressed. The consequence would be that if a general concept of ontology could be formulated for understanding cultures, it would refer not to “conceptions of the world” linked to different languages, but to a plural and unsystematic way of constantly activating different forms of thought. In short, it would designate not a single system, but a form of life where different systems constantly combine. To look at the relationship between language, thought, and culture in this perspective could be a way, as Wittgenstein wrote of magic in his Notes on Frazer’s Golden Bough, to fully “preserve its depth.”


Austin, J. L. 1975. How to do things with words. Oxford: Oxford University Press.

Barcelos Neto, Aristóteles. 2002. A arte dos sonhos: Uma iconografía ameríndia. Lisbon: Assírio & Alvim.

———. 2008. Apapaatai: Rituais de máscaras no Alto Xingu. São Paulo: Editora da Universidade de São Paulo.

———. 2009. The (de)animalization of objects: Food offerings and the subjectivization of masks and flutes among the Wauja of Southern Amazonia. In The occult life of things: Native Amazonian theories of materiality and personhood, edited by Fernando Santos-Granero, 128–53. Tucson: University of Arizona Press.

———. 2011. “Le réveil des grands masques du Haut-Xingu: Iconographie et transformation.” In Masques des hommes, visages des dieux, edited by Jean-Pierre Goulard and Dimitri Karadimas, 27–52. Paris: éditions CNRS.

———. 2013 “O trançado, a musica e as serpentes da transformaçâo no Alto Xingu”, in Quimeras en diálogo: Grafismo e figuração na arte indígena, edited by Carlo Severi and Els Lagrou, 181–97. Rio de Janeiro: Viveiros de Castro Editora.

Basso, Ellen B. 1981. “A musical view of the universe: Kalapalo myth and ritual as religious performance.” Journal of American Folklore 94 (373): 273–91.

Baxandall, Michael. 1993. “Pictorially enforced signification: St. Antonius, Fra Angelico and the Annunciation.” In Hülle und Fülle: Festschrift für T. Buddensieg, edited by Andreas Beyer, Vittorio Lampugnani, and Gunther Schweikhart, 31–39. Alfter: VDG Verlag.

Bateson, Gregory. 1979. Mind and nature: A necessary unity. New York: Dutton.

Beaudet, Jean-Michel. 1983. “Les orchestres de clarinettes tule des Wayãpi du Haut-Oyapock.” Doctoral dissertation, Université Paris Ouest Nanterre La Défense.

———. 1997. Souffles d’Amazonie: Les orchestres tule des Wayãpi, Vol. 3. Nanterre: Sociétéd’Ethnologie.

Belaunde, Luisa Elvira. 2009. Kené: Arte, ciencia y tradición en diseño. Lima: Instituto Nacional de Cultura.

———. 2013. “Movimento e profundidade no kene shipibo-conhibo da Amazônia peruana.” In Quimeras en diálogo: Grafismo e figuracão na arte indígena, edited by Carlo Severi and Els Lagrou, 199–223. Rio de Janeiro: Viveiros de Castro Editora.

Boas, Franz. 1938. “Language.” In General anthropology, 127–55. Boston: D. C. Heath.

———. 1989. A Franz Boas reader: The shaping of American anthropology, 1883–1911. Chicago: University of Chicago Press.

Bourdieu, Pierre. (1972) 1977. Outline of a theory of practice. Translated by Richard Nice. Cambridge: Cambridge University Press.

Brabec de Mori, Berndt, and Anthony Seeger. 2013. Introduction: Considering music, humans, and non-humans. Ethnomusicology Forum 22 (3): 269–86.

Carnap, Rudolph. 1991. Empiricism, semantics, and ontology. In The philosophy of science: An historical anthology, edited by Timothy McGrew, Marc Alspector-Kelly, and Fritz Alhoff, 356–65. Oxford: Wiley-Blackwell.

Cesarino, Pedro de Niemeyer. 2011. Oniska, poética do xamanismo na Amazônia. São Paulo: Editora Perspectiva.

de Civrieux, Marc. (1970) 1997. Watunna: An Orinoco creation cycle. Translated by David M. Guss. Austin: University of Texas Press.

Descola, Philippe, and Anne-Christine Taylor, eds. 1993. “La remontée de l’Amazone.” Special issue, L’Homme 33 (2–4).

Evans-Pritchard, Edward. 1937. Magic, witchcraft and oracles among the Azande. Oxford: Oxford University Press.

Fausto, Carlos. 2011a. “Masques et trophées: De la visibilité des êtres invisibles en Amazonie.” In Masques des homes, visages des dieux, edited by Jean-Pierre Goulard and Dimitri Karadimas, 229–54. Paris: CNRS Éditions.

———. 2011b, “Les masque de l’animiste: Chimères et poupées russes en Amérique indigene.” Special issue, “Pièges à voir, pièges à penser,” edited by Carlo Severi, Gradhiva 13: 48–68.

Fausto, Carlos, Bruna Franchetto, and Tommaso Montagnani. 2011a. “Les formes de la mémoire: Art verbal et musique chez les Kuikuro du Haut-Xingu (Brésil).” L’Homme 197: 41–69.

Fausto, Carlos, and Isabel Penoni. 2014. “Les effigies rituelles: La figuration de l’humain dans le Haut-Xingu.” In L’image rituelle, edited by Carlo Severi and Carlo Fausto, 14–38. Paris: L’Herne.

Gheerbrant, Adrian. 1954. Journey to the far Amazon. New York: Simon & Schuster.

Gilbert, Aubrey L., Terry Regier, Paul Kay, and Richard B. Ivry. 2006. “Whorf hypothesis is supported in the right visual field but not the left.” Proceedings of the National Academy of Sciences of the United States of America 103 (2): 489–94.

Goodman, Nelson. 1976. Languages of art: An approach to a theory of symbols. Cambridge, MA: Hackett.

Gow, Peter. 1989. “Visual compulsion: Design and image in Western Amazonian art.” Revindi: Revista Indigenista Americana 2: 19–32.

———. 1999 “Piro designs: Painting as meaningful action in the Amazonian lived world.” Journal of the Royal Anthropological Institute (N.S.) 5: 229–46.

Gumperz, John. J. and Stephen C. Levinson, eds. 1996. Rethinking linguistic relativity. Cambridge: Cambridge University Press.

Guss, David. M. 1986. Keeping it oral: A Yekwana ethnology. American Ethnologist 13 (3): 413–29.

———. 1989, To weave and sing: Art, symbol, and narrative in the South American rain forest. Berkeley: University of California Press.

Gutierrez Choquevilca, Andrea-Luz. 2010. “Imaginaire acoustique et apprentissage d’une ontologie animiste: Le cas des Quechua d’Amazonie péruvienne.” Ateliers du LESC 34. ; DOI : 10.4000/ateliers.8553.

———. 2011. “Sisyawaytii tarawaytii: Sifflements serpentins et autres voix d’esprits dans le chamanisme quechua du haut Pastaza (Amazonie péruvienne).” Journal de la Société des Américanistes 97 (1): 179–221.

Hill, Jonathan D. 1993. Keepers of the sacred chants: The poetics of ritual power in an Amazonian society. Tucson: University of Arizona Press.

———. 1994. “Musicalizing the other: Shamanistic approaches to ethnic-class competition in the Upper Rio Negro region.” In Religiosidad y resistencia indígenas hacia el fin del milenio, edited by Alicia Barabas, 105–28. Quito: Abya-Yala.

———. 2009. Made-from-bone: Trickster myths, music, and history from the Amazon. Urbana: University of Illinois Press.

Jakobson, Roman. 1959. “On linguistic aspects of translation.” On Translation 3: 232–39.

Joyce, James. (1922) 1972. Ulysses. London: Penguin Books/Bodley Head.

Kay, Paul, and Chad K. McDaniel (978. “The linguistic significance of the meanings of basic color terms.” Language 54: 610–46.

Kay, Paul, and Willett Kempton. 1984. “What is the Sapir-Whorf hypothesis?” American Anthropologist 86 (1): 65–79.

Kuhn, Thomas. S. (1962) 2012. The structure of scientific revolutions. Fiftieth anniversary edition. Chicago: University of Chicago Press.

Lagrou, Els. 2007. A fluidez da forma: Arte, alteridade e agência em uma sociedade amazónica (Kaxinawa, Acre). Rio de Janeiro: Topbooks.

———. 2009a. Arte indígena no Brasil: Agência, alteridade e relação. Belo Horizonte: EditoraC/Arte.

———. 2009b. “The crystallized memory of artifacts: A reflection on agency and alterity in Cashinahua image-making.” In The occult life of things: Native Amazonian theories of materiality and personhood, edited by Fernando Santos-Granero, 192–213. Tucson: University of Arizona Press.

———. 2011. “Le graphisme sur les corps amérindiens: Des chimères abstraites?” Special issue, “Pièges à voir, pièges à penser,” edited by Carlo Severi, Gradhiva 13: 69–93.

———. 2013, “O grafismo nos corpos ameríndios, quimeras abstratas?” In Quimeras en diálogo: Grafismo e figuração na arte indígena, edited by Carlo Severi and Els Lagrou, 38–66. Rio de Janeiro: Viveiros de Castro Editora.

Langdon, Esther Jean. 2013. “Perspectiva xamanica: Relaçoes entre rito, narrativa e arte grafica.” In Quimeras en diálogo: Grafismo e figuração na arte indigena, edited by Carlo Severi and Els Lagrou, 111–39. Rio de Janeiro: Viveiros de Castro Editora.

Lévi-Strauss, Claude. (1962) 1966. The savage mind. Chicago: University of Chicago Press.

Lévy-Bruhl, Lucien. 1949. Les carnets. Paris: Presses Universitaires de France.

Lima, Tânia Stolze. 2000. “Towards an ethnographic theory of the nature/culture distinction in Juruna cosmology.”. Revista Brasileira de Ciências Sociais 15 (99): 43–52.

Lloyd, Geoffrey E. R. 1996. Adversaries and authorities: Investigations into ancient Greek and Chinese science, Vol. 42. Cambridge: Cambridge University Press.

———. 2006. Principles and practices in ancient Greek and Chinese science, Vol. 849. Farnham, UK: Ashgate.

———. 2007. Cognitive variations: Reflections on the unity and diversity of the human mind. Oxford: Clarendon Press.

———. 2012. Being, humanity, and understanding: Studies in ancient and modern societies. Oxford: Oxford University Press.

Lucy, John A. 1992. Language diversity and thought: A reformulation of the linguistic relativity hypothesis. Cambridge: Cambridge University Press.

Luna, Luis Eduardo, 1992. “Icaros: Magic melodies among the Mestizo shamans of the Peruvian Amazon.” In Portals of power: Shamanism in South America, edited by E. Jean Matteson Langdon and Gerhard Bauer, 231–53. Albuquerque: University of New Mexico Press.

Mangione, Corrado. 1964. Elementi di logica matematica. Turin: Boringhieri.

Meneses Bastos, Rafael José de. 1978. Etnomusicologica Kamayura. Universidade Federal, Santa Catarina, Florianopolis.

———. 1999. “Apùap world hearing: On the Kamayurá phono-auditory system and the anthropological concept of culture.” The World of Music 41 (1): 85–96.

———. 2007. “Música nas sociedades indígenas das terras baixas da América do Sul: Estado da arte.” Mana 13 (2) : 293–316.

Piaget, Jean. (1923) 2001. The language and thought of the child. Translated by Marjorie and Ruth Gabain. London: Routledge.

———. (1926) 2007. The child’s conception of the world. Translated by Joan and Andrew Tomlinson. Lanham, MD : Rowman & Littlefield.

Piedade, Acácio Tadeu de Camargo. 2004. “O canto do Kawoká: Música, cosmologia e filosofia entre os Wauja do Alto Xingu.” Ph.D. dissertation, Universidad Federal de Santa Catarina.

Quine, Willard Van Orman. 1943. “Notes on existence and necessity.” Journal of Philosophy 40 (5): 113–27.

———. 1948. On what there is. Washington, DC: Catholic University of America, Philosophy Education Society.

Rubel, Paula G., and Abraham Rosman. 2003. Translating cultures: Perspectives on translation and anthropology. New York: Berg.Russell, Bertrand. 2005. Collected papers of Bertrand Russell, Vol. 29. Hove: Psychology Press.

Sapir, Edward. 1985. Selected writings of Edward Sapir in language, culture and personality, Vol. 342. Berkeley: University of California Press.

Saussure, Ferdinand de. (1913) 2006. Course in general linguistics. Translated by Roy Harris. Chicago: Open Court.

Seeger, Anthony. 2004. Why Suyá sing: A musical anthropology of an Amazonian people. Evanston: University of Illinois Press.

Severi Carlo. (2007) Forthcoming. The chimera principle: An anthropology of memory and imagination. Translated by Janet Lloyd. Chicago: HAU Books.

———. 2011. “L’espace chimérique: Perception et projection dans les actes de regard.” Special issue, “Pièges à voir, pièges à penser,” edited by Carlo Severi, Gradhiva 13: 5–163.

———. 2012. “The arts of memory: Comparative perspectives on a mental artifact.” HAU: Journal of Ethnographic Theory 2 (2): 451–85.

———. 2013 “Philosophies without ontologies.” HAU: Journal of Ethnographic Theory 3 (1): 192–96.

Severi, Carlo, and Carlos Fausto, eds. 2014. L’image rituelle. Paris: L’Herne.s

Severi, Carlo, and Els Lagrou, eds. 2013. Quimeras en diálogo: Grafismo e figuração na arte indígena. Rio de Janeiro: Viveiros de Castro Editora.

Sperber, Dan. 1982. “Apparently irrational beliefs.” In Rationality and relativism, edited by Martin Hollis and Steven Lukes, 149–80. Cambridge, MA: MIT Press.

Taylor, Anne-Christine. 2003 “Les masques de la mémoire: Essai sur la fonction des meintures corporelles Jivaro.” In Special issue, “Anthropologie et image,” edited by Carlo Severi, L’Homme 165: 223–48.

Townsley, Graham. 1993. “Song paths: The ways and means of Yaminahua shamanic knowledge.” Special issue, “La remontée de l’Amazone,” edited by Philippe Descola and Anne-Christine Taylor, L’Homme 33 (2–4): 449–68.

Velthem, Lucia von. 2003. O belo e a fera. Universidade de São Paulo.

———. 2013. “Homens, guaribas, mandiocas e artefactos.” In Quimeras en diálogo: Grafismo e figuração na arte indígena, edited by Carlo Severi and Els Lagrou, 139–63. Rio de Janeiro: Viveiros de Castro Editora.

Vidal, Lux Boelitz, Anne-Marie Pessis, and Niéde Guidon. 2000. Grafismo indígena: Estudos de antropologia estética. São Paulo: Studio Nobel.

Vygotsky, Lev S. 1978. Mind in society. Cambridge MA: Harvard University Press.

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

———. 2004. “Perspectival anthropology and the method of controlled equivocation.” Tipití: Journal of the Society for the Anthropology of Lowland South America 2 (1): 3–22.

Whorf, Benjamin Lee. 2012. Language, thought, and reality: Selected writings of Benjamin Lee Whorf. Cambridge, MA: MIT Press.

Wittgenstein, Ludwig N. (1914–16) 1974. Notebooks 1914–1916. Translated by G. E. M. Anscombe. Chicago: University of Chicago Press.

———. (1953) 2010. Philosophical investigations. Translated by G. E. M. Anscombe. London: John Wiley & Sons.

———. (1967) 1979. Remarks on Frazer’s Golden Bough. Translated by A. C. Miles, edited by Rush Rhees. Atlantic Highlands, NJ: Humanities Press.


Carlo SEVERI is Professor (Directeur d’études) at the École des Hautes études en Sciences Sociales and Director of Research (Directeur de recherche) at CNRS. A member of the Laboratoire d’Anthropologie Sociale of the Collège de France since 1985, he has been a Getty Scholar at the Getty Institute for the History of Art and the Humanities in Los Angeles (1994–95), a Fellow of the Wissenschaftskolleg in Berlin (2002–2003), and a Visiting Fellow at Cambridge (1990, 2012). He is the author of La memoria rituale (La Nuova Italia, 1993), Naven or the other self (with Michael Houseman, Brill, 1998; French edition: CNRS Éditions, 1994), and The chimera principle (HAU Books, forthcoming; French edition: Rue d’Ulm-Musée du Quai Branly, 2007).

Carlo Severi
Laboratoire d’Anthropologie Sociale
Collège de France
52 Rue du Cardinal Lemoine
75005 Paris


1. To give an account of this tradition of studies, one should refer at least to the classic question of linguistic relativity (from classic works by Boas, Sapir (1985) and Whorf (2012) to Kay 1978; Kay and Kempton 1984; and Gilbert et al. 2006; see also Lucy 1992 or Gumperz and Levinson 1996) and the recent debates on the nature of linguistic translation (see, e.g., Rubel and Rosman 2003). Both tasks are well beyond the scope of this paper.

2. I have written a first analysis of Yekwana and Wayana iconographies in a paper devoted to the nature of “chimerical” representations (Severi 2011). On the Yekwana pictographs as an example of the Amerindian “arts of memory,” see Severi (2013).

3. For Velthem, beings considered “qua-anacondas” are those capable of clasping and devouring humans; meanwhile those associated with caterpillars, themselves thought of “qua-jaguars,” include beings capable of “biting [humans] from within,” often almost imperceptibly, as in illnesses (ibid.: 320).

4. We find startling echoes of these representations of a single “complex being” in the documents collected by Barcelos Neto (2002) among the Waura (Xingu), which include representations of supernatural anacondas comprising series of images each of which refers to a different animal.

5. “What changes when passing from one species of subject to another,” writes Viveiros de Castro, “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 [humans] see a muddy salt-lick on a riverbank, tapirs see their big ceremonial house, and so on. Such difference of perspective—not a plurality of views of a single world, but a single view of different worlds—cannot derive from the soul, since the latter is the common original ground of being. Rather, such difference is located in the bodily differences between species, for the body and its affectations . . . is the site and instrument of ontological differentiation and referential disjunction” (ibid.: 6).

6. In Amazonian cosmologies, “the body appears to be the great differentiator, that is as that, which unites beings of the same type, to the extent that it differentiates them from other beings” (ibid.: 479).

7. The use of this notion to designate the “true origin of perspective” (ibid.) is surprisingly ambiguous. Needless to say, Bourdieu used it to define any process which leads to the constitution of a cultural tradition.

8. Among the many passages where this notion is invoked, see, for instance: “the notion of animals as people is always associated with the idea that the manifest form of each species is just an envelope” (ibid.: 470); “we would have a distinction between an anthropomorphic essence of a spiritual type . . . and a variable body appearance, characteristic of each individual species. . . . Such difference of perspective . . . is located in the bodily differences between species” (ibid.: 471).