On selves, forms, and forces

This work is licensed under the Creative Commons | © Bruno Latour. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau4.2.014


On selves, forms, and forces

Bruno LATOUR, Sciences Po, Paris

Comment on Kohn, Eduardo. 2013. How forests think: Toward an anthropology beyond the human. Berkeley: University of California Press.

Tout réel se possède lui-même; autrement qui le posséderait?

—Raymond Ruyer

I read How forests think as part of a vast movement to equip anthropologists, and more importantly, ethnographers, with the intellectual tools necessary to handle a new historical situation: the others are no longer outside; nonhumans have to be brought back in the description in a more active capacity. Both of those features, naturally, mark the disappearance of older notions of nature and of its counterpart, namely culture; disappearance, that is itself due to the fact that everybody— ethnographers as well as former informants—are pulled deeper and deeper into the same ecological maelstrom. Whatever the term—is it an ontological or a semiotic turn?—the importance of the book relies on the most crucial turn of all: that is, a turn to experience and how to describe it empirically.

Mind you, the book is not empirical in the sense of bringing in a vast array of data, but rather in the sense of concentrating, onto a very narrow set of microevents, the full weight of a shift in perspective regarding what it is to be attentive to the world (a change in perspective underlined by the choice of black and white photographs working as powerful attention-shifters—see especially Kohn 2013: 70). I am not completely sure of my count, but the 228 pages of analysis amplifies and develops probably less than 35 seconds of interactions between entities, some of whom are recognizable as human interlocutors, and most of the others involving sonic vibrations of some sort. That is to say that the book, in its very style and construction, does what it says we should do, that is, shift “beyond human” and “beyond language” but not—that’s the key point—beyond meaning. Vibrations include river catchments, forests, the dead, dogs, colonial history, biological lineages, and of course pumas, and dreams, in addition to the writings of a fairly cryptic philosopher: C. S. Peirce.

“We humans are not the only ones who do things for the sake of a future by re-presenting it in the present. All living selves do this in some way or another. Representation, purpose, and future are in the world—and not just in that part of the world we delimit as human mind. This is why it is appropriate to say that there is agency in the living world that extends beyond the human” (Kohn 2013: 41–42).

The whole book manages to establish relations between those entities without ordering them into language and world and without relying on cause and effect connections when materiality comes in. The universal connector, because it seems that you need one, is called “semiosis” in the concrete sense of what folds the world into one direction or directs attention to some of those folds (very much like the French word “sens”). Hence, many sentences use “semiosis” repetitively to erase artificial distinctions: “Indeed, the startled monkey’s jump, and the entire ecosystem that sustains her, constitutes a web of semiosis of which the distinctive semiosis of her human hunters is just one particular kind of thread” (ibid.: 42).

So “semiosis” is supposed to tear anthropologists away from their obsession with language and representation1 enabling them to get fully immersed into other materialities without, however, having to take these materialities as “matter.” Whenever you encounter an entity, the reference of a former language game, you can, by taking the entity itself as part of the semiosis, reverse the direction of description and start instead with the entity, letting the vibration from this entity traverse other entities, sometimes provoking what used to be called linguistic utterances, but most of the time passing through language and humans without stopping at them. In that sense plants have just as much representational abilities as the demoted humans: “Plants in the tropics, as opposed to those in temperate regions, come to form relatively more nuanced representations of the characteristics of their environments. They make more differentiations among soil types because of the ways they are caught up in a relatively denser web of living thoughts” (ibid.: 82, my emphasis).

This is to me the most interesting aspect of the book: whereas “beyond human” could imply that you begin with the subject and immerse it into the world of things (this is after all what naturalist descriptions most often do), the book (and I take it also as a work of art) manages to describe situations where neither the human, nor its subjectivity, nor its linguistic capacities are central. The book could have been called: “How to bypass talking humans.” The central thesis is that you will be able to detect small differences once you have made all entities into selves. (Well not exactly “all,” since rocks and nonliving entities are treated quite traditionally as what is talked about but not as what has semiosis. until chapter five brings them into the vibrating medium as forms).

Since my interest in the book does not come from any ethnographic knowledge of Latin America, I will underline its importance for what I define as a diplomatic dimension of contemporary ethnography: how do we equip those who take upon themselves the description of the worlds of others (which now most of the time means describing the crisis facing their worlds) to enter into relations with those who are thus addressed—and sometimes even with their enemies? On that ground, I think that Eduardo Kohn’s book succeeds fairly well.

I am of course directly interested in such an enterprise since the book criticizes my own attempt at establishing a similar diplomatic scene. If ANT fails as a connector of entities, it’s crucial that semiosis, in Kohn’s hands, succeeds. For me, the two projects are actually much closer than the author thinks. The following definition of an attempt to bring the reality of forms into the picture, would fit marvel-ously ANT’s project:

Form, then, is crucial to lives, human and otherwise. Nevertheless, the workings of this vague entity remain largely undertheorized in anthropological analysis. This, in large part, is due to the fact that form lacks the tangibility of a standard ethnographic object. Nevertheless, form, like the basic intentionality of the pig and the palpable materiality of its meat, is something real. Indeed, its particular mode of efficacy will require us to think again what we mean by the “real.” If, as anthropologists, we can find ways to attend ethnographically to those processes of form amplification and harnessing as they play out in the Amazon, we might be able to become better attuned to the strange ways in which form moves through us. This, in turn, can help us harness form’s logics and properties as a conceptual tool that might even help us rethink our very idea of what it means to think. (Kohn 2013: 160)

Rematerializing thinking as well as rethinking what is real, that’s what is common to science studies and to anthropology. So why is it presently better to handle connections among entities through semiosis rather than through associations?2 For one thing, the generalized use of the connector “semiosis” to handle all vibrations among entities, is enormously facilitated in Kohn’s fieldwork by the abilities of his informants to extend the notion of “selves” to all beings. “Selves are signs. Lives are thoughts. Semiosis is alive. And the world is thereby animate” (99). And that’s an ability we never had in science studies. When you study situations where “per-spectivism” is involved, it’s much easier to manoeuver so as to shift perspectives in a surprising way. When you study animists (in the new sense given to the word by Descola [2013]), it’s less difficult to “animate” entities well “beyond” human souls. Science studies could not use the same repertoire to study naturalists who make use of many “phusi-” or “techno-morphisms” instead of those “psycho-morphisms” or “auto-morphisms”3 so generously distributed to the Runa by the beings entering in relation with them. When you study scientists and engineers, to be sure you never come across the famous “materialism” and “cause and effect relations” they are supposedly stuck with, but you are faced with a bewildering heterogeneity of registers. And that’s what makes their study so difficult (Latour 2014).

So there is an ambiguity here in the treatment of the Runa register, in their ontological claims, so to speak, made by Kohn in the name of their dead, forests, dreams, dogs, puma, rivers, and plants. In spite of what Kohn often says, automorphisms do not define the background of the world as it is, but only one register of how to handle connections by treating all of them as being relations among selves. For the Runa, it seems to work fine, but it will not allow them to make themselves understood by, for instance, the forest engineers, the agronomists, or other “whites” visiting Avila. To those outsiders, such a register would sound like exoticism, not as a chance to understand anew how their own reality is actually constituted. A diplomatic encounter would have backfired. No wonder, since the extreme violence of colonial history forms the background of every chapter of the book. Their ontological claims have been pushed away as so many beliefs.

And that’s the problem I have with the powerful counterpoison Kohn had to rely on to avoid exoticism, namely the use of Peirce’s semiotics. Since ANT has made large use of another semiotics to escape the narrow “realism” that passes as a description of “societies,” I understand the move. But it’s not the same semiotic at all. Whereas Greimas’ semiotics allows multiple registers since every actant can be played out by many actors, Peirce’s semiotics (at least in Kohn’s treatment of it) claims to be an alternative description of what the world is. Each semiotics risks losing what the other gains. If it’s true that Greimas could have difficulties making ontological claims, he can entertain a vast diversity of registers well beyond relations among selves; while Peirce allows strong ontological claims but has to stabilize much too fast all connections into auto-morphisms. And yet, no matter how good Kohn’s book is, the Runa qua Peircian ontology have not become for everybody else the definition of their common world. Hence the danger of stabilizing too quickly what the furniture of the world is, and the necessity of having a semiotic toolkit able to restart the negotiation whenever it has stalled. Such is for me the advantage of Greimas over (Kohn’s) Peirce.4

A detailed analysis of the advantages and disadvantages of those two approaches in How forests think would take too long, but it would be of great interest to see when the appeal to semiosis works and when it fails to record important variations. For instance, chapter five—a very beautiful description of homologies between river catchments, the rubber trade, Runa colonial history, plant dispersal, and power structures—could have been written exactly in the same way by many historians, environmentalists, or geographers. The addition of semiosis here adds nothing except a funny turn of phrase.5 By contrast, chapters two and four recount how some sound vibrations connecting barking dogs, predatory jaguars, linguistic structures, and habits of thought fully benefit from the continuity allowed by treating all entities as trafficking in the future (the beautiful definition of selves that Kohn has taken from Peirce).

The problem as I see it, is that Kohn stumbles upon the same obstacle that has made ANT inoperative after a few years. Any attempt at choosing a homogeneous concept to establish connections among all entities (association for ANT, semiosis for Kohn) has a powerful but short-lived effect. Powerful because it allows not to make artificial distinctions (human and nonhuman for ANT, language and world for Kohn), but short lived because inevitably the differences that had been recorded slowly fade, turning out to be the same way for everything to be different. Yes, it is admirable to begin to think of the forest as what is able to align growing seeds, falling trees, startled monkeys, shooting hunters, dreaming fathers, dead mothers, moving glottises, onomatopoeias, words as symbols, ethnographic documents, all the way down to the beautifully written text, yes, but when you sum it up by saying “the forest thinks” you are bound to say “wait a minute, now that you have given us such amazing continuities, I want my discontinuities back, not the old ones to be sure, I don’t care about abandoning nature/culture, object/subject, world/words, but still, I don’t wish to confuse a seed with a dream or a predator with a prey.” And that’s why, in my experience, the half-life of descriptive theory is about five years! That’s at least what has happened to ANT’s quickly diminishing returns. And I am afraid that’s what will happen to this influx of semiosis into ethnography. Ontologi-cal pluralism cannot be achieved through only one mode of existence, no matter how encompassing it appears to be.6

But that does not matter much, because the aim of the book, as I see it, does not depend, in spite of the enthusiastic and quite convincing borrowing of quotes from Peirce, upon a unified ontology but rather on establishing a writing strategy that removes the speaking humans away from center stage. Human sometimes do speak in symbolic forms, but not always. They sometimes are more than a prop but always less than the only dramatis personae of the plots. What a relief! An anthropology that learns how to escape anthropocentrism without, for that, being in any way reductionist, antihumanist, or deterministic, that’s quite a wonder. Such is the immensely refreshing move for which we should be grateful to the Runa’s ways of life transformed, through Kohn’s aesthetics, into an alternative philosophical anthropology.

But, naturally, to really evaluate Kohn’s attempt, the test is still to come: how could an ethnographer, or, for that matter, a Runa scholar, equipped with such a philosophical anthropology find ways to make his or her ontological claims understood in negotiating what a forest is made of, when faced with forestry engineers, loggers, tourists, NGOs, or state administrators? That’s where the so-called ontological turn finds its moment of truth. Not on the epistemological scene but on the bittersweet attempts at negotiating alternative ways to occupy a territory, being thrown in the world, designating who is friend and who is enemy. That’s where Kohn is leading us, and that’s where I yearn to join him.


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

Kohn, Eduardo 2013. How forests think: Toward an anthropology beyond the human. Berkeley: University of California Press.

Latour, Bruno 1988. “Irreductions.” In The pasteurization of France, part 2. Translated by John Law. Cambridge, MA: Harvard University Press.

———. 2013. An inquiry into modes of existence: An anthropology of the moderns. Translated by Catherine Porter. Cambridge, MA: Harvard University Press.

———. 2014. “Agency at the time of the anthropocene.” New Literary History 45: 1–18.


Bruno Latour
Sciences Po
27 rue Saint-Guillaume
75007 Paris Cedex 07, France


1. Or rather, “representation” shifts from being solely linguistic or psychological to something that is a general property of living things (the nonliving are excluded from this capacious definition): “Selves relate the way that thoughts relate: we are all living, grow-ing thoughts” (Kohn 2013: 89).

2. ANT is not often sourced where it is most usefully defined, at least philosophically, that is in Latour (1988).

3. It is important to pluralize the prefixes that make up the words ending with “morphism” to alert the readers to the oddity of stabilizing one register for instance by believing that “anthropomorphism” possesses a clear and coherent meaning. On this point see Latour (2014).

4. Kohn is right to criticize ANT for sticking to the linguistic register (especially 2013: 92) but he fails to notice that auto- or psycho-morphism have the same limits as logo-morphisms and that ANT, at least in its radical intent, is a study of all morphisms, not of one in particular. So when he writes, “selves, not things, qualify as agents” (92), a semiotician of my persuasion will see this as a shift from semiotics to the expression of “world view” and thus an abandonment of the commitment to description. Of course, it does not matter for the Runa, but if you were studying brain scientists, psychologists, psychiatrists, or New Age healers, you will want to make sure not to define selves in advance.

5. And an adept of ANT will have resisted losing the description into “overarching form” and “structure”: “The result was that all these more basic regularities came to be part of an overarching form—and exploitative political-economic structure whose grasp was very difficult to escape” (Kohn 2013: 165). Ouch! Indeed “a grasp difficult to escape.”

6. In the vocabulary of the inquiry into modes of existence, Kohn’s book is an invaluable exploration of the specificities of two modes [REP] and [HAB] (Latour 2013).