From one ontology to (an)other

This work is licensed under the Creative Commons | © Gérard Lenclud. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau4.3.023


From one ontology to (an)other

Gérard LENCLUD, Centre national de la recherché scientifique (CNRS)

Translated from the French by Matthew Carey

Comment on Descola, Philippe. 2013. Beyond nature and culture. Translated by Janet Lloyd with a foreword by Marshall Sahlins. Chicago: University of Chicago Press.

Is the only limit to the diversity of different human communities’ experiences of the world the fantastic variety of expressions that cultural genius takes? Were this the case, could human beings still communicate across cultures and, as the anthropological project supposes, learn from one another? This, for want of space, is the only question I shall seek to answer in the following discussion of Philippe Descola’s Beyond nature and culture.

Let me get right to the point. Descola’s purpose here is to order the various ontological schemata devised by the human mind to represent “what there is” (i.e., the different entities that people our world) and the ways in which “what there is” is held to exist (i.e., the entities’ mode of being, which is linked to the qualities they are thought to possess). He proceeds to identify, more for heuristic than directly typological reasons, four grand ontological schemata. Each schema corresponds to one of the four combinatorial possibilities that emerge from the attribution or denial of particular ontological properties to the interiority and physicality of different classes of “existant.” Interiority refers, obviously enough, to what goes on “inside” these existants and thus is hidden from our gaze—this includes aspects of what we call mind [esprit] or consciousness, mental states, and manifestations of subjectivity. Physicality, meanwhile, refers to distinctive traits and visible characteristics of their material body—that which can be seen. The four grand schemata are animism, naturalism, totemism, and analogism.

When faced with a given human or nonhuman entity—what Descola describes in resolutely antinaturalist terms as an “other,” be it creature or thing—I may proceed as if we possessed a common interiority but distinct physicalities (animism), distinct interiorities but a common physicality (naturalism), a common interiority and physicality (totemism), or distinct interiorities and physicalities (analogism).

Descola offers us, in short, a “general cosmological grammar” (2013: 131), each instance of which is just that: one particular instance and nothing more. This includes the naturalist cosmology that we have illegitimately set up as a sort of standard candle by the light of which to measure the others. On the one hand, he outlines the underlying unifying logic that underpins human understanding in all four cases and, on the other hand, he establishes points of comparison between the different worlds generated by these four schemata. It is, then, crystal clear that the book is the expression of a profound universalist conviction, balk as Descola may at a term he suspects of acting as a fig-leaf for our efforts to claim a monopoly on truth.

That said, the opposition that unfolds between these different ontological schemata, each of which holds a mirror up to the others, cannot quite chase away the specter of ontological relativism and, by extension, the adoption of a relativist view of truth (though not, of course, of that which is taken to be true). Indeed, Descola applies himself to the task of broadening the breach that separates the different ontological regimes he identifies. Each is presented as a radically different [“autre”] form of human experience. What is more, Descola reaffirms the categorically unique nature of “our” vision of the world, with its sharp distinction between natural and cultural facts. Modern Western naturalism, with its binary division between types of existant, seems to Descola to be a cultural exception of sorts. This rejection of any hint of ontological ecumenism, as well as of any thought that the naturalist model might be uniquely contagious or compelling, ultimately leads Descola to rehabilitate the Great Divide between Us and Them, at least as regards the ways in which we conceive of the world.

Faced with this model, we cannot but wonder: if the inhabitants of our earth conceive of “what there is” and of the way in which “what there is” exists according to independently developed cultural guiding principles, how can a user of one ontological schema access the world as decoded through a different ontological schema? What allows, for instance, an anthropologist, whose mind has been formatted by a naturalist ontology, to learn in situ and later explain the perspective of an interlocutor raised within an animist environment?

The importance of asking this question is amply illustrated by the chapter of Descola’s book entitled “The great divide,” in which he identifies the lessons to be learned from the insular character of the four different ontological regimes. He rejects the idea that the architecture of the human mind might be responsible for the “fundamental discontinuity of the real” discussed by Claude Lévi-Strauss, and thus impose constraints on any human species classification. Descola stops short, of course, of calling into question the psychic unity of humankind (and so also, at least in principle, of human nature), but he does seem to reject the idea that this shared nature might consist of a common cognitive and epistemological capital that underpins our different cultural versions of the world. Above all, he claims in this chapter that each ontological schema tends to be associated with a particular world—a world that is constructed, rather than discovered, and that imposes itself sui generis on the humans who people it. Indeed, he mentions that people who do not share our particular cosmology, invent “realities distinct from our own” (Descola 2013: 122) and he borrows Roy Wagner’s critique of the vast majority of anthropology: that it acts as if non-Western societies constructed alternative versions of our reality, when in fact they are versions of their own realities, which differ from ours.

How, though, are we to understand this idea of a reality that is not shared by all humans? Taken in a “deflationist” sense, it is little more than a banal remark to the effect that members of different cultures have different perspectives on the world around (and encompassing) them. They see the same things differently and so when they stop to observe them, they also look different. Simply put, interpretation begins where “mere” perception ends, however informed this latter may be. The hypothesis of perceptual relativism, in contrast, runs directly counter to this position. For its advocates (declared or otherwise), members of different cultures see different things when placed in front of what they all take to be the “same” object. Each person is convinced that the others see the same things he does, when in fact they all see different things. Seeing is already conceiving. It follows that for advocates of perceptual relativism, the idea of a “same object of perception” (même chose perçue) only exists because of the parochial nature of human thought,” to borrow an idea from Thomas Kuhn. Descola seems to lean toward the relativist side of the equation when he singles out one particular function fulfilled by these collective schemata in pronouncing ontological judgments: “Structure the flow of perception in a selective fashion, granting a preeminence in signification to particular traits and processes that can be observed in the environment” (Descola 2013: 103).

If one rules out the idea of a “same object of perception,” then it raises the problem of the “knowability” of perceptual worlds other than that of the knowing subject. Can the anthropologist really put himself in the position of somebody whose percepts are differently structured from his own? Learning to conceive as another conceives is undoubtedly possible; but seeing as she sees is another kettle of fish entirely. For one thing, one cannot observe another’s seeing and so “copy” it. And for another, one cannot see at will or on request. A researcher can only know what another person sees by inferring it from her publicly observable behavior, most obviously speech. But what allows us to assume that people are even capable of revealing the content of their perceptual experiences? To do so, they must “translate” into words and phrases the sensory effect the world has upon them—an effect that operates both instantaneously and wordlessly.

Furthermore, the idea of a reality that is not identical for all humans needs to be seriously modulated if it is to avoid supposing a self-refuting relativist theory of truth. The fatal concatenation would look like this: if other cultures, other realities; if other realities, other objectively known worlds. Why objectively? Because if we grant that all humans have the capacity to conceptualize the opposition between true and false (which cannot seriously be contested), it follows that all humans have a concept of objectivity and so recognize that certain states of reality exist independently of their own beliefs, desires, or fears. It is the existence of these states of reality that makes their own thoughts or statements true or false, that acts as their truth conditions. It follows that if we take seriously the idea that reality is not the same for all people then the truth conditions of their statements (at least the propositional ones) are also necessarily culturally variable. Other cultures constitute other tribunals of experience: the same proposition can be objectively true in one place and objectively false in another.

We know, however, that any statement made in one language can be translated (with varying degrees of brio, elegance, and concision) into another. And all translation implies recourse to a standard notion of synonymy and the equally standard notion of meaning. What is it that makes statements uttered in different languages synonymous, thereby ensuring the relative success of translation? Let us take three descriptive sentences: in English, “cats meow”; in French, “les chats miaulent”; and in Japanese, “neko wa niaô to naku” (literally, “as for cats, they make a neowing sound”). These statements are only synonymous insofar as they refer to the same fact—viz. that cats meow. Despite what constructivists may claim, it is trivially obvious that this fact exists independently both of the linguistic context (that of a propositional statement) and of the thought-world within which these statements are uttered. Simply put, these statements mean the same thing because they are all true with regard to the entity mentioned (a member of the family felidae) and the quality ascribed to it. Were it the case that cats did not meow in the three countries in question, then the sentences would not be synonymous and the translation erroneous. In short, the standard notion of synonymy, which is inseparable from the act of translation, requires the application of identical truth conditions. Meaning and truth cannot be sundered as one cannot understand the meaning of a statement without knowing the conditions in which it is true and one obviously cannot know these conditions without knowing what the statement means (Newton-Smith 1988).

Let there be a relativist anthropologist who asserts, or even implies, that p is as (objectively) true among the So-And-So, where she conducts fieldwork, as not-p is objectively true in her homeland. If she has access to the truth conditions of p, this means that she has translated it, producing a more or less synonymous version of p in her own language. And if she has translated p, it follows that the truth conditions of p are the same among the So-And-So as back home. She must necessarily have translated p, because if p had remained unfathomable for her, she could not have made the slightest claim about its truth conditions, which she had to do to assert that the So-And-So inhabit a reality different from our own. The success of the translation is proof of the error of her position, which could only be demonstrated by an irremediable impossibility of translation.

It seems to me that this necessarily leads us to draw two conclusions. First, the idea of a reality that is not the same for all humans must accommodate a categorical rejection of a possible contradiction between truth conditions. Over the very long term, and provided the case is correctly formulated, the tribunal of experience cannot lay down fundamentally divergent judgments. And second, the idea that people inhabit different known worlds must be compatible with the conviction (typically dear to any anthropologist) that it is possible to know known worlds that are not known as one knows one’s own.

In other words, though one can establish a theoretical opposition between the different ontological schemata identified by Descola, they are also necessarily convergent. We might also say that human beings must perforce share a common lot of ontological commitments: communication would quite simply be impossible if their ontologies were uniquely exclusive. And needless to say, the ethnographic project would be doomed to failure in ontologically distant lands. Indeed, as Descola quite rightly insists, the ontology of a given group of people is nothing more than “the way in which a group schematizes its experience” (2013: 107), and it is precisely this experience that the ethnographer studies. How can the ethnographer pretend to gloss another’s experience of the world and capture its main lines if the different elements of that experience in no way coincide with those of her own world (supposing that she shuns any pretension of occupying an overarching position)? To which we can add that limited and meaningful disagreement can only be apprehended on a background of general agreement.

The conceptual schema of identity is best placed to allow for the convergence of distinct ontological regimes and thus allow inhabitants of different known worlds to communicate their experience. One can argue over whether this schema is sufficient but one can scarcely doubt that it is necessary. How could one even begin to access the representations of reality held by a creature for whom beings and things are not necessarily what they are? What kind of “thing” could these “things” be that were not identical to themselves? Let me briefly elaborate on this.1

Particular creatures, people, and things necessarily play a central role in our experience of the world. It is they who are the principal object of acts of predication and predication is the foundation of human thought. To predicate is to specify by affirming, in response to the question “What is this?,” that it is such and such a thing; it is also to characterize by saying that it is like this or like that. Some particular objects play key strategic roles in that they serve to singularize other categories of particular entities or things—for instance events or processes. These key reference points are those endowed with a material form: we can call them “spatio-temporal objects.” They are endowed with greater unity and unicity. They can be individually identified and so counted. And it is first and foremost to these particular objects, whose “real” existence is so blindingly obvious, that the schema of identity is applied. To exist is to differ, as scarcely any metaphysician (either descriptive or revolutionary) would deny. To differ is to possess a particular identity. To exist, then, is to exist as one and the same thing. This is what is implied by Leibniz’ famous statement, in a letter to Arnauld, April 30th, 1687: “That which is really not a being, is not really a being either.”

Briefly put, the conceptual schema of identity structures the ways in which we identify and recognize particular objects. It makes use of two concepts of identity that function as guiding principles. The first is a concept of numeric identity. It refers to the logical principle that any object is necessarily identical to itself and to no other thing than itself, however similar they may be. And this holds for as long as the object exists, just so long (as we shall see) that it maintains this relationship of identity with itself. These objects do not “come to be” in supposedly Heraclitean fashion but rather persist—i.e., exist at more than one time. As David Lewis said about persons, they are “endurers,” which means that, when present at two times, they are wholly present at these times. This idea of durable identity, of a thing with itself, is an all-or-nothing proposition. It cuts like a guillotine. A particular object is not roughly or more or less that which it is (a person, a cat, a tree, a blowpipe); it either is or is not. This principle appears to subtend all possible ontological perspectives: an experience of the world in which people and things are not identical to that which they are defies understanding.

There is also the concept of specific identity. A particular object is not only, and necessarily, that which it and it alone is; it is also, and just as necessarily, what it is: a representative of a class (or species) of beings or things, a quality that it shares with other such objects. Indeed, a particular object (be it person, cat, tree, or blowpipe) does not constitute a class unto itself. Though metaphysically possible, the idea of a “bare” object, stripped of its similars, is epistemologically indefensible. Just try to find one! In our own culturally specific way, we all organize particular objects into classes of natural kinds—“natural” in that we find it natural to class them together by virtue of certain shared properties. “Birds of a feather flock together,” seems to be the watchword for the construction of classes, at least those classes that seem intuitively obvious to us, wherever we may be.

Numeric and specific identity are indissolubly linked in our apprehension of the world. Many philosophers claim that this connection, mentioned as far back as Aristotle (“Things that are one in number are also one in species” Metaphysics V.6 1016b37) takes the form of a principle, dubbed the “theory of sortal dependency” of identity (David Wiggins). This principle has an existential and an epistemic dimension. The existential dimension implies that the preservation of a specific identity is a necessary, if not obviously sufficient, condition of the maintenance of numeric identity. In order for something to continue to be a particular human, cat, tree, or blowpipe, it must remain a human, a cat, a tree, or a blowpipe. The epistemic dimension implies that knowledge of a concept of class (or “sortal”) expressed by a general term, an enumerable name, provides a method for individuating the particular objects that people that class. In theory, it identifies the conditions under which a particular member of this class may remain what it is (and so that which it is), by laying out the changes that it may, or even must, undergo without losing its identity, as well as those which deprive it of this identity. It is trivially obvious that these identity conditions and levels of change vary according to class. To grasp the necessity of going through classes to reach particular objects, it is enough to note that we neither identify nor count particular objects as particulars but as members of classes. Identifying a particular entity does not involve distinguishing it from all other existing entities! This is why the intuitive classes discussed in scholarly texts, as well as in everyday speech, are neither conjunctive (the class of either humans or cats), negative (the class of non-blowpipes), nor disjunctive (the class of neither cats nor trees). Their members would be drowned in a sea of overly dissimilar “similars.”

In short, the notion of a particular object implies one of class, just as the notion of class implies particular members. There can be no species without specimens and no specimens without species! This is how the conceptual schema of identity, which serves as the skeleton of our grammar of self and other, functions. Is it conceivable that this grammar be excluded from a particular ontological regime?

To answer the question, let us focus our attention on the animist regime. I have chosen an animist ontology as Descola locates it at the precise opposite of our own, and yet he has managed to apprehend the particular entities that people the world as it is known through this schema. Let me just remind the reader that in the kingdom of animism, it is appearances alone that serve to establish discontinuities. With the exception of creatures without souls (most insects and fish, poultry, and numerous plant species), which Descola only mentions in another text (1986), all entities are endowed with the same interiority as humans.

And what of particular entities within animism? Unsurprisingly, they are precisely that which they are: highly individual and in no way indeterminate. The guillotine of numeric identity falls pitilessly upon them. If human clothing is not enough to distinguish between two entities, then the Achuar resort to body markings. Multiple births are a source of dread and homonymy is scrupulously avoided. Simply put, particular animist entities are refused the luxury of crises of (numeric) identity.

Let us now turn to specific identity and so to animist species. Here we find ourselves in another known world. Descola writes the following: “forms are fixed for each class of entities but are variable for the entities themselves” (2013: 135). He repeats this in an endnote (n18), where he adds that form is permanent at the level of the species/class but not at the individual level (416). One can only conclude that numeric and specific identity have here been sundered. Particular objects are not wed to “their” classes. Particular entities are and remain that which they are but are not eternally condemned to be what they are. Particular animist entities are, then, only contingently endowed with the status of members of their classes. This logical state of affairs (quite illogical for us) is the outcome of an ontological decree: any entity is hic et nunc that which it is (human, jaguar, or hummingbird) by virtue of the perspective from which it is apprehended by another entity—bearing in mind that with few exceptions, any entity is capable of having a perspective of any other, for they (we) share the same interiority. As what they are depends not on themselves but on others (so to speak), particular animist entities have no fixed residence. They are, in short, what others have them be.

There are, however, two caveats to be added to this relativization of specific identity. First, animist ontology is characterized less by its acceptance of metamorphosis than by the idea of the interchangeability of form. Metamorphosis implies that the subject is, to a certain extent, anchored to its original and terminal class or species; for instance, human being (disobedient woman) and pillar (of salt) in the case of Lot’s wife. In the kingdom of animism, a being A does not transform into being B, rather any subject has the potential to be an A, a B, and so on—like Proteus the Egyptian, the subject is not the only agent of his shape-shifts. It is not then that the bind of specific identity (like to like) is violated; it is simply ignored, at least in principle.

Second, it is worth asking whether animist ontologies admit the active presence of (what we would call) class or species in their intellectual arsenal. There are at least two reasons for this. On the one hand, the logic of specification requires that individual members of a class or species possess the properties common to this species as it is only by virtue of possessing them that they are considered specimens. The determination of a species depends, in turn, on the properties of its members. The idea of a species whose members do not possess its defining attributes, whatever they may be, goes against the very idea of a species. If hummingbirds, for instance, can change their plumage, and so their species, while the species “hummingbird” remains prisoner of its plumage, it is hard to fathom what membership of a species actually means and so in what an animist species might consist. And, on the other hand, if particular individuals are able to don (almost) as many disguises as there are perspectives upon them, then they are simply members of a quasi-universal “species,” endowed with a nature common to almost all existants. The relationship between hummingbirds, for instance, is not intrinsically different to the relationship between humans and jaguars. The indiscriminate animist notion of species does not derive from some division of reality: there are no real articulations. Animists even manage to avoid dividing existants into those ontological categories that developmental psychologists have identified as structuring the perceptual frame by means of which young children apprehend the world: animate/inanimate, human/ animal/vegetable.

It remains, however, the case that animists do divide the world into species. How could they not? Species function as “seven-league boots” (William James), which we can use to blaze a trail through the backwoods of multiplicity. It is quite simply impossible to identify, generalize, predict, or narrate the unknown to the known without recourse to species or classes. On what ground could one stand in a notional world without classes? And yet Descola managed to master the nomenclature relative to beings and things in the world of the Achuar. His success is proof, if proof were needed, of the fact that the world of animism partially coincides with our own. The conceptual schema of identity demonstrably constitutes just such a point of coincidence. Indeed, Descola admirably documented the Achuar science of the concrete, more attuned to the thirst for knowledge than to the urge to produce, in his book, In the society of nature. They thus distinguish between six hundred animal species, including thirty-three butterfly species and forty-two types of ant. This appears to give rise to a contradiction: by sparing particular individuals the burden of lifetime membership of a single species, animist ontology undermines the very principle of specification, but animists themselves do classify their world, and do so more than is strictly necessary.

To resolve this contradiction, one might imagine that the Achuar have two ontologies, one for everyday existence and the other reserved for philosophical speculation. The first ontology, which satisfies the need to order reality, would endow existants with the status of specimens. This ontology would also coincide with our own, which metaphysicians and experimental psychologists consider universal. The second ontology, meanwhile, would serve for reflection about the status and fate of existants. It would ultimately annul by fiat the boundaries between species. This ontology would also be culturally marked. Descola rejects out of hand the hypothesis of ontological duality.

Perhaps we need to phrase things differently. Fair enough: the Achuar do not have two competing, contrasting ontologies. Does this mean that they have a single ontology in the form of a unified and coherent corpus of propositions? In both In the society of nature and Beyond nature and culture, Descola stresses that Achuar experience of the world is categorically not supported by this sort of coherent conceptual scaffolding. The same is true of our experience of the world … and everybody else’s. The only ontologies that have this unified character are those deployed in scientific activity and presupposed by particular theoretical versions of the world (whose purpose is not even to be extensible from one field of research to another).

If one grants, on top of the idea that human beings do not have one ontology, a) that they do not have an ontology in the same way that they have a nationality or landed property, b) that their imputations of existence and of mode of being are not entirely transparent to them, and c) that the concept of ontology is a mediocre class term, possessed of as little individuating power as its close cousin, cosmology … if one grants all this, then it follows that one would do better to speak of a patchwork of ontological commitments that is forever unfinished and that can never be described in all its detail.

One might further hazard that such commitments are of two sorts. On the one hand, those associated with naïve or common sense ontologies, and built into the ordinary usage of conceptual and natural languages. They are expressed in the form of intuitive beliefs and are generally tacit. Perceptual “beliefs” are the prototype of such commitments. And on the other hand, there are those commitments that form part of a well thought out cultural ontology, that one might also describe as theoretical insofar as it goes beyond the limits of the concrete, of “everyday perceptions” (Thomas Kuhn) or “natural interpretations” (Paul Feyerabend). They take the form of reflexive beliefs.

Intuitive beliefs tend to go unnoticed by researchers, because they most likely share them. These are the sorts of belief that descriptive metaphysicians try to identify using semantic analysis and logic. Reflexive beliefs, in contrast, catch the researcher’s attention because they are culturally alien. They call for interpretation.

Naïve and theoretical ontologies only rarely engage in open battle. Descola is doubtless correct when he asserts that animals do not change their mode of being depending on whether they are the target of an Achuar hunter’s weapon (and so are members of edible species) or his speculative thought (when their specific identity is uncertain). He does, however, mention in In the society of nature those not so infrequent cases where actual hunting practice does violence to “official norms.”

As for myself, I do not think it belittles the variety of cultural genius to suggest that Descola’s four ontological regimes share common foundations based in common sense conceptions of beings and things that fix their identity conditions. These four different “sciences of the world” are theoretical versions of the ways in which human beings schematize their experience.

And besides, people cleave far more closely to their reflexive beliefs than to their instinctive certainties …


Descola, Philippe. 1986. La nature domestique. Paris: Editions de la Maison des sciences de l'homme.

———. 2013. Beyond nature and culture. Translated by Janet Lloyd. Chicago: University of Chicago Press.

Lenclud, Gérard. 2013. L'universalisme ou le pari de la raison. Paris: Hautes Etudes, Gallimard/Seuil.

Newton-Smith, William. 1988. “Relativism and the possibility of interpretation.” In Rationality and relativism, edited by Martin Hollis and Steven Lukes, 106–23. Oxford: Basil Blackwell.



Gérard Lenclud
CNRS—National Centre for Scientific Research
Laboratoire d’Anthropologie Sociale
52 rue du Cardinal Lemoine, 75005 Paris, France


1. I have developed elsewhere (Lenclud 2013) the argument presented here in more schematic form.