The Wrong Question?

: The Wrong Question? is the response by an anthropologist to a question posed by a philosopher concerning the intelligibility of alien forms of thought. I argue that it is wrong to describe the problem of intelligibility as one of logic or rationality. Indeed, foreign practices (no less than our own) may become intelligible only once they are not evaluated according to abstract criteria of rationality. To ask of a given practice or form of life whether it is rational is an error of grammar (nonsense) in Wittgenstein’s sense. I describe how intelligibility emerges over the course of ethnographic ﬁeldwork but also argue that we must work on our own concepts in order to make foreign ones intelligible. The response draws from both Gadamer and Wittgenstein as well as anthropologists Geertz, Evans-Pritchard, and L é vi-Strauss. Following Cora Diamond, I suggest further that the ethical and rational dimensions of understanding another are indissociable.

Problems that are properly anthropological should never be put either in the psychologistic terms of belief or in the logistic terms of truth-value. Eduardo Viveiros de Castro [1] (p. 490) If one mission of anthropology is to draw or develop theory from our ethnographic subjects and objects, it is also the case that one can contribute to theory by reflecting on ethnography itself as a practice. Here I do so in responding to an invitation to speak at a workshop on the topic of "the problem of the intelligibility of alien forms of thought" as formulated by philosopher Jean-Philippe Narboux for a workshop exploring "new avenues of mutual illumination between anthropology and philosophy", in particular by means of attention to Wittgenstein, Evans-Pritchard, and Lévi-Strauss. 1 My response is framed primarily through a hermeneutically shaped perspective, though I take up Narboux's "three crucial figures of the previous century" (Wittgenstein, Evans-Pritchard, and Lévi-Strauss) in the second half of the essay. The gist of my argument concerns the fact that the intelligibility of alien forms of human thought has little to do with rationality per se, or at least is not best addressed directly by means of logic or recourse to universals. I think that with sufficient hermeneutic and contextual work, alien forms of (human) thought are not unintelligible. Indeed a major achievement of a leading figure of the socalled "ontological turn" in anthropology, Eduardo Viveiros de Castro, is precisely to show how to understand what he calls Amerindian cosmological deixis or perspectivism [3]. 2 Just as an English speaker, especially if she starts young enough and with the right attitude, may acquire reasonable facility in speaking Cree or Hungarian, and conversely, and more likely, as Cree and Hungarian speakers regularly do acquire English, so we have access to other peoples' worlds. 3 As bilingual speakers move easily between languages, one can move between forms of thought. However, this takes time, and for most anthropologists, this access is modest and must be approached with due humility. In Gadamer's metaphor, we open and gradually enlarge a common arc of understanding rather than share a complete horizon [5]. Furthermore, I take it that the hermeneutic process is a reflexive one, such narcissism, and the like. I only mention these as axes of intelligibility but will not pursue them here.
What I have just said about my experience in the field might sound like a kind of naïve confession (or confession of naiveté), and my sense of mutual understanding and sympathy could be but a (narcissistic) illusion. In other words, such claims to understanding could be taken as indices of their opposite. Moreover, complacency is disastrous for an anthropologist; once you think you understand, you stop asking questions. I think that claims by anthropologists to have gone native or to understand "from the inside" are often naïve and premature. Like Socrates, ethnographic wisdom should be premised on acknowledging that we do not know.
I take what I consider an Aristotelian path on this matter, trying to be scrupulous about not exaggerating the proximity or depth of my understanding but equally not exaggerating the difficulties or the strangeness. Ideally, I continue to work at understanding, and I do not understand understanding or intelligibility to be an endpoint or a container such that one could say that it is either "reached" or "full".
I see the practice of the ethnographer as one that operates simultaneously on an intellectual and ethical plane. My character is such that in the field, I practice or profess a kind of asceticism or aloofness with regard to what is sometimes called "participation". I rarely engage so completely that I forget my anthropological position. I take this to be somewhat akin to the ethical stance of a psychoanalyst, which Janet Malcolm memorably labeled as the "impossible profession" [15]. Intimacy and identification have their limits. To cite Geertz once again, here, after Wittgenstein, and in a way that is precisely the inverse of psychoanalysis, we access other peoples' worlds only insofar as they are public [16]. However, I acknowledge that every ethnographer must handle this differently, according to who they are and contingently, according to how relations unfold and what your interlocutors give you access to know. The ethics of intelligible participation is complex and a matter of continuous practical judgment rather than, as North American research boards seem to think, of establishing and following procedures and rules.
Intelligibility is not a direct product of experience. My closest friends in Mayotte engage in spirit possession, a practice or set of practices that were initially alien to me and that I have spent much of my career exploring. I think I understand some aspects of possession quite well, and my published work is available to assess this claim. However, I have never become possessed by a spirit myself and never tried to become possessed. Although I can understand what is going on, I cannot perform it. Much as I can appreciate a musician or athlete without being able to perform as they do, so I cannot be a competent spirit medium. This could be partly a matter of simple fear or absence of the physical capacity, just as I could not and would not climb a coconut palm to knock off the nuts (not that anyone expected me to do this either). I also appear to lack the unconscious motivation or have too much resistance to become possessed. One could say that possession has become part of my world of understanding but not part of my practice; intelligible in some ways and inaccessible in others, just as playing violin or sky diving are somewhat intelligible but inaccessible to me. I should add that many citizens of Mayotte understand or acknowledge possession without becoming possessed themselves. Some actively resist it, while others simply do not feel the call and might even regret their incapacity. However, everyone engages with spirits when they actively possess their hosts, and everyone acknowledges their presence behind the scenes at other times. Even with respect to hosts or mediums (words I use interchangeably), I have come to think that a better description than one of possession is that of cohabitation. I can make sense of cohabiting with spirits, and, to a degree, I have cohabited with them as well. One could call living with spirits a form of life, and one can participate in it without becoming possessed oneself.
One thing spirit mediums say is that they have no awareness when the spirit takes active possession and speaks through them and no recollection of what transpired. 8 Elsewhere I have analyzed how this premise underlies much of what follows with respect to communication, forming at base a minimal triad composed of the medium, the spirit, and the trusted third party who engages with the spirit and reports back to the medium and conversely engages with the medium or client and reports back to the spirit. This is a role I have played on occasion. For example, the male spirit possessing a female friend once asked me to tell her to be more circumspect with men who accosted her when she was working alone in the fields. When I passed on the message to my friend, she proceeded to tell me about a time she had narrowly escaped being sexually assaulted. The spirit spoke to me rather than to my friend's husband, as the husband would have wanted to know the identity of the interloper.
The fact of loss of consciousness also means that I probably would not gain a deeper understanding of possession even were I to become actively possessed, at least not one I could report back on and not with respect to its social nature. 9 The closest I can come to this is to report on my experience waking up in a Toronto hospital after having suffered a concussion, with no memory either of having been knocked fully unconscious by a cyclist or of the conversation I had thereafter with the health workers, one in which I was unable to give my name or address but, according to what I was subsequently told, reiterated with some urgency that I was a professor. Perhaps I was, and am, possessed by or cohabiting with a professor spirit.
Anthropological fieldwork is sometimes described as participant observation. At one extreme are those ethnographers who try to participate fully, which one can think of either as admirable openness to experience or as sentimental self-indulgence, and who may confuse knowing how for understanding what is at stake. At the other extreme are those few anthropologists who think one should merely observe, as though the behavior of another species, and who sometimes go so far as to argue that any conversation with the objects of observation, by treating them as if they were subjects, distorts the data. 10 I consider this ethological position both ethically odious and intellectually stupid. Situated somewhere between these poles, I find myself carrying out what I might describe as simply listening and engaging in sustained conversations, albeit neither listening nor engaging in sustained conversations are either simple or straightforward. With respect to participation, anthropologists might focus less on their individual experience than on considering the capacity for using concepts within a given form of life. 11 There are also more resources in the word 'observation' than those conveyed by the ethologist or than ethnographers have explored. After all, to observe a holiday is not to watch it from a distance but to respect it. I like to think that observation, in this sense, conveys something of my position or action qua ethnographer. To observe here conveys attention and respect-for the dignity, intelligibility, and sensibility of others and for the value of their practices.
To this position, Hans-Georg Gadamer's precept concerning conversation is apposite, namely to "Always recognize in advance the possible correctness, even the superiority of the conversation partner's position". Furthermore, he adds, "If we want to understand, we shall try to make [the other's] arguments even more cogent". 12 These are things that good anthropologists try to do, although, in the end for ethnographers, it is generally not a matter of a mutually exclusive choice between two positions, if only because we engage with numerous interlocutors who may hold divergent positions.
Gadamer's ideals, albeit considerably undermined in his encounter with Derrida [19], are, however, not the whole story. Ten pages after the last remark, Gadamer poses or exposes what I consider a central dilemma of ethnographic conversation. In this passage of Truth and Method, Gadamer opposes two different kinds of conversation. There is, . . . a conversation that we have with someone simply in order to get to know him, i.e., to discover his standpoint and his horizon. This is not a true conversation, in the sense that we are not seeking agreement concerning an object, but the specific contents of the conversation are only a means to get to know the horizon of the other person. Examples are oral examinations, or some kinds of conversation between doctor and patient . . . Just as in a conversation, when we have discovered the standpoint and horizon of the other person, his ideas become intelligible, without our necessarily having to agree with him, the person who thinks historically comes to understand the meaning of what has been handed down, without necessarily agreeing with it, or seeing himself in it.
But this means, continues Gadamer, that . . . we have as it were, withdrawn from the situation of trying to reach an agreement . . . . By including from the beginning the other person's standpoint in what he is saying to us, we are making our own standpoint safely unattainable. . . . The text that is understood historically is forced to abandon its claim that it is uttering something true. We think we understand when we see the past from a historical standpoint, i.e., place ourselves in the historical situation and seek to reconstruct the historical horizon. In fact, however, we have given up the claim to find, in the past, any truth valid and intelligible for ourselves. Thus this acknowledgment of the otherness of the other, which makes him the object of objective knowledge, involves the fundamental suspension of his claim to truth 13 .
The suggestion seems to be that insofar as we render the other's practice intelligible in its own terms, we do so at the expense of rendering it significant for us. "Intelligibility", therefore, is by no means a straightforward concept. Questions of the ethical and intelligible are intertwined.
Older anthropology, say that of Evans-Pritchard or Lévi-Strauss, would probably advocate conversation in this limited sense. During the encounter, the point is only to gain the "native's point of view". It is back home and writing up that anthropologists, now anticipating a different audience, attempt to make it "valid and intelligible for ourselves" in terms of the theoretical concepts and models at our disposal.
I suspect many contemporary anthropologists would reject this as unethical and opt for the full or true conversation, one in which we do not suspend the other's claim to truth and do not render our own standpoint "safely unattainable", and they would say this should take place directly in the field, among coevals, as anthropologist Johannes Fabian memorably put it [21]. There is much moralizing about this. Whereas, in the notso-distant past, most anthropologists, for their own good reasons of moral concern, did not want to adopt the missionary position; now, with respect to their political knowledge, anthropologists increasingly take it up in the name of advocacy. When I was first in Mayotte in 1975, I encountered political campaigning to keep Mayotte French rather than join the emerging independent nation of Comoros. Full of righteous anti-colonial sentiment, I declared my support for independence, only to be told that if I was to continue to advocate that position, I would be asked to leave the village where I had settled as I might implicate my hosts. I shut my mouth. The anti-independence movement prevailed, and 36 years later, in 2011, Mayotte became a full overseas département of France, the long, drawn-out process having brought both enormous benefits and enormous problems. I do not think it is my business to draw up an accounting in the sense of determining whether, in the wisdom of hindsight, the political decision was correct, but to be honest, I am not sure where criticism should enter. Some anthropologists do intervene with positive intentions and sometimes even positive results, but we are often inclined to exaggerate our importance in such matters.
To speak one's mind or one's truth is not always the position of intellectual openness or ethical sensibility. As we know from seminars, a certain pausing, holding back, or bracketing is necessary to enable others to speak; to take the risk of speaking, the risk both of being misunderstood, insofar as the interlocutor recognizes that the anthropologist does not yet know enough in order to understand what she says, and the risk of being humiliated by a scornful response. Certainly, my understanding of possession, and the potential for increasing my understanding, would have collapsed had I taken the path of speaking my truth. The art of the anthropologist, like that of the psychoanalyst, is primarily one of listening and waiting until they have learned what are the right questions, in the sense of both what questions to ask and how and when to ask them. It is an art of postponement.
I have led us to a kind of impasse, but I think there are ways out. If I understand him, Gadamer's two kinds of conversation are only ideal types, and he proposes that hermeneutics itself is able to transcend or finesse the distinction between them. First, there is the empirical fact that closed horizons are a romantic fiction; 14 there is a space for understanding in any conversation, and conversation itself is a tool to pry open more understanding. As a saying in Mayotte goes, "conversation begets conversation" (kuraña miteraka kuraña). Language is intrinsically open to the world, not closing things off. Second, even where I do not try honestly to reach agreement with my interlocutor, I expose my own ideas and prejudices to challenge. The point is not that I should challenge my interlocutor but that I should challenge myself; receptiveness to Others entails recognizing how their concepts and opinions are situated with respect to my prejudices, and in that reception, my prejudices are realized for what they are, as prejudices rather than truth. 15 In a parallel fashion, the withholding psychoanalyst can still confront their counter-transference.
I reached another way out of the dilemma, this one afforded by the specific standpoint of the other. I came to understand that practitioners' own understandings of possession could be described in a certain fashion as ironic (rather than literal) and indeed that possession itself, insofar as it is intrinsically heteroglossic, is ironically constituted. 16 A remark I often heard in Madagascar from people who did not know me well, attempting at once to gauge and shape my response to possession, and always uttered in French, was "incroyable . . . mais vrai!" So possession is both unbelievable and true rather than either unbelievable or true. The irony is central to the intelligibility of possession, and possession provides, in this respect, a reading of other practices and stances, including that of anthropology.
The problem may lie with truth itself as the grounding concept for intelligibility. Or one might say that I have found truth in irony or even irony in truth, and yet, insofar as I am ironic, I must also note the insufficiency of irony itself.
Putting these questions of truth and irony aside, what I am trying clumsily to get at is that with respect to intelligibility, the ethical and rational dimensions or criteria are indissociable. I take this to be a point developed by Gadamer and also an insight developed independently by certain Wittgensteinian scholars like Stanley Cavell and Cora Diamond. As Diamond puts it, "In our cognitive activities, we may be aiming at getting things right, but how we carry on these activities, how we understand what 'getting it right' might involve or might cost in the particular case, reveals our own nature as moral beings. Our moral nature expresses itself in all our modes of consciousness" [25] (p. 103). 17 It follows, Diamond says, that "If thought is seen as inherently or ubiquitously moral, then we need to reject the idea that moral thought is a department of thought, and moral discourse a department of discourse" [25] (p. 104). As an aside, this is why I find the phrase "anthropology of ethics" a misleading conceptualization of the intellectual project in which I have been engaged [14].
I would add that with respect to possession, and likely certain other topics, modes, and genres of expression, the literal and metaphorical or the literal and ironic are also indissociable from one another. It is a matter of phronesis, of practical wisdom, to know in what proportion to receive them. It is not simply a question of "what we say when" but a question of what and how we receive when. 18 To go one step further, I suggest it is again a category mistake, in the sense that Ryle diagnoses body and mind to fully distinguish between the rational and the ethical, to ask of a given practice, act, or utterance whether it is either rational or ethical. The same could, at times, be true for the question of whether it is either literal or ironic. You may perceive that my own statement here is of this kind. To attempt such discrimination might appear to increase intelligibility while actually working at its cost.
I want to add one brief caveat to all I have said, namely that what I find least intelligible, on both rational and ethical grounds, are certain political ideas and practices "at home". Furthermore, I do not think these are matters on which one should take an open Gadamerian position, if only because we are already long past that. ** ** In his abstract for the workshop, Jean-Philippe Narboux raised three important thinkers, namely Wittgenstein, Evans-Pritchard, and Lévi-Strauss. In the remainder of the essay, I will briefly address each.
I am on shaky ground when I speak about Wittgenstein and defer to other contributors to this issue. One thing I take from Wittgenstein is that to ask of any given practice or tradition whether it is rational is to ask the wrong question. It is simply not the right way to address the kinds of things people do and say in the course of ordinary life or the forms through which they do and say them. In other words, the question itself is neither logically correct or incorrect, rational or irrational, but a kind of misapplication or misapprehension, an error of grammar. One could say that the act of posing it is infelicitous in Austin's sense or that the question itself is nonsense in Wittgenstein's. 19 The question is nonsense because most of what anthropologists observe, and despite our claims to the contrary, is in fact not rational in the logicians' sense of that word or concept. However, by the same token, it is not irrational either. These are the wrong criteria to apply to forms of life. Moreover, most voiced utterances are not in the form of propositions; as both philosophers and linguistic anthropologists have described, people are doing many other things with their words and concepts.
It follows that what the anthropologist observes might itself be considered nonsense. This is not a criticism and not to say it is therefore unintelligible. On the contrary, it could be intelligible as poetry is rather than as mathematical proof is. However, taken in their own terms, people might themselves sometimes mistake their nonsense for sense, their poetry for proofs, and certainly, their illocutionary acts for locutionary statements. 20 I think that in our good intentions to show that people who have otherwise been disparaged for their differences from dominant Euro-American society or from a class segment of that society, or from self-proclaimed intellectuals, are our intellectual equals, anthropologists may have unintentionally concealed this point. In demonstrating that they are not our inferiors, a point (must I say it?) with which I obviously agree, perhaps we have not always clarified that they are not necessarily our superiors either. We are all fallible beings, all subject to human finitude, all liable to make category mistakes or other mistakes of grammar, and all liable to mistake those mistakes for truth.
Put in other words, if philosophers have seen as one of their missions to point out faulty thinking among themselves, anthropologists have seen as one of their missions to do the obverse: to point out nonfaulty thinking among others. That has been a good thing, and it has led to much insight, but perhaps it has also meant that we have ignored the fact that people in other societies are no more privileged than we to think and act rationally, to live, for example, without making category mistakes or without practicing the kind of nonsense that Wittgenstein diagnoses and disdains among philosophers and among the precursors of modern anthropology such as Frazer.
I take a primary insight of Wittgenstein to be, as I stated earlier, that it is wrong to describe the problem of intelligibility as one of rationality or to reduce it to that. 21 Indeed, practices, statements, and actions may become intelligible only once it is realized that they are not fully rational according to the abstract standards or criteria of logic. Additionally, intelligibility requires context; hence, we do not take such practices, statements, and actions in isolation from one another but as part of language games and forms of life. However, we quickly discover that the ideas, practices, language games, and forms of life found within any given society are never fully consistent or commensurable with one another. Furthermore, they will always be plural since no one of them can be complete, can satisfactorily answer all our questions, or can address all human concerns.
This brings me to Evans-Pritchard, who is famous for pointing out that answering why something happened is not the same as answering how it happened [31]. Material explanation may be effective for addressing 'how' questions-the granary collapsed because it could not bear the weight of the harvest once termites had eaten away at the wooden support posts-but it cannot address the 'why' question, why the collapse happened just when Grandmother was napping under it. Material or scientific explanations are unable to address existential concerns. Another language game is necessary, and possibly more than one, since some questions can never be adequately answered by any given system. We are speaking here not only of the intelligibility of other systems of thought for us but how they each address the intelligibility of the universe for those who hold them. Evans-Pritchard thought the material and non-material explanations did fit neatly together in practice for the Azande, such that, as summaries of his argument put it, "witchcraft explains unfortunate events". However, in fact, as he himself showed, witchcraft for Azande is also material, a given part of the world. Moreover, if witchcraft explains some things, it raises new questions: How can I know whether I am a witch? Why am I a witch? And so forth.
There is no body of thought that is fully consistent or fully comprehensive in addressing human concerns, no doctrine that does not generate further inquiry or confront its adherents with aporia. In Judaeo-Christian-Muslim traditions, one manifestation of this is evidently that if God is omnipotent and loving, why do bad things happen? Many compelling answers have been given, but I do not think any of them has persuaded everyone or held consistently over time. One of the stronger answers has been to dispense with intelligibility altogether-to argue that God's will is not intelligible to humans. People who hold this view must supplement it with other, more practical answers.
Evans-Pritchard effectively described what we might ourselves describe, after the fact, as life with concepts. Within a three-year period, he produced two classic works on two distinct societies, Witchcraft, Oracles and Magic among the Azande (1937) [31] and The Nuer: A Description of the Modes of Livelihood and Political Institutions of a Nilotic People (1940) [32].These books are not usually talked about in the same space, one ostensibly dealing with matters of 'religion' and the other with 'ecology' and 'social structure'. It is largely the first book that has drawn the attention of philosophers, due in part to the assumption that questions of rationality or intelligibility hold more with respect to what gets classed as 'religion' or 'magic' than to what gets classed as 'politics' or 'kinship.' In fact, however, Evans-Prichard makes the same general point in each book, and his achievement, his powers of abstraction in grasping the intelligibility of a foreign order, are if anything even more impressive in the book on the Nuer. The general point is complementary to Wittgenstein's, namely that the anthropologist who stands back and tries to understand a given system in the abstract is approaching things in the wrong way. Nuer and Azande no more live in abstraction that we do. Or, like us, let us say, most of the time, in most modes of attention. To understand what they are saying or doing, we need to understand the context in which they are doing it at that moment, under those circumstances, within particular forms of life. Thus Evans-Pritchard asserts that Zande ideas of witchcraft only become intelligible when they are not laid out like artifacts in a museum display case but understood as they follow one another in the practice of everyday life-someone falls ill, treatment is ineffective, a new diagnosis is made, a culprit is sought, an accusation is made, a response is given, and so forth. In effect, he is pointing to a Wittgensteinian form of life.
Likewise, Evans-Pritchard declares that the ways Nuer answered questions about their social identities were unintelligible when he interviewed them in his tent. This was not, as he first thought, because they were lying or being unnecessarily difficult (though there may have been some of that as well) but because they did not understand the question-or rather, that the question he was posing itself made no sense, was unintelligible, in that context. Evans-Pritchard eventually realized that he was asking them a question whose answer could only be context dependent and that he was doing so in a situation that was itself decontextualized or outside of known context, outside any form of life in which it might be recognized by Nuer. In the tent, there literally was no single, right, or consistent answer to the question.
Evans-Pritchard came to understand that Nuer did not hold absolute, categorical, or objectified social identifications but lived in a world where the salient identification was always indexed to the context. In Peircean terms, Nuer social identities and attributions were indexically constituted. Nuer answers were unintelligible until he understood this point. Conversely, Evans-Pritchard's question was unintelligible to the Nuer until he asked it in a specific social context. To expect a consistent and stable answer to the question of who one is or to which group one belongs is as wrongheaded as expecting a Euro-American to give a consistent referent to 'home'. It could be a particular house, a city, or a country, where I grew up or where I live now, and it could be more than one of each, depending on the context in which the question is asked. To say I am going home could mean to my flat in Willmersdorf, back to Canada, back to the Germany that my father fled, and so on, each correct, but each apparently inconsistent to a questioner who expects a single answer and who fails to realize that the ostensibly nominative form is actually deictic. It is a mistake of grammar in the literal sense. The great achievement of The Nuer is, first of all, to perceive this and, secondly, to reconstruct the logic of context, a kind of structural relativity, such that the answers to the question do become intelligible in the sense of being both consistent and comprehensive.
A general point here is that an intelligible answer depends on an intelligible question. It is not just a matter of what we think or understand about alien ideas and practices but about what they think or understand by our questions.
Like Zande witchcraft, Malagasy spirit possession is not irrational. As Evans-Pritchard showed, once one accepts an initial premise (concerning, respectively, the existence of witches for Zande or autonomous spirits for people in Mayotte), everything falls into place. The first premise is particular and arbitrary, but the rest is largely not. Furthermore, again, as Evans-Pritchard also presciently pointed out, the logic here is not fully abstract nor, in the first instance, a system of rationalization but practical within a form of life. Moreover, like the Zande with respect to witchcraft, Malagasy speakers are not naïve about possession; any given performance can be evaluated for its felicity, with some people or performances unmasked as duplicitous. More interestingly, as I have noted, possession is deeply ironic, both in its form or structure and in its awareness of itself.
Evans-Pritchard wrote against Lévy-Bruhl, who was, in turn, responsive to the criticism. I am not sufficiently versed in Lévy-Bruhl to take this up except to say that I think the debate raises the following points. First, if we explore deeply enough another world that seems at first encounter to make, from our perspective, logical mistakes, like conflating subject and object, we will eventually find that these are largely mistakes on our part, made from our own initially naïve perspective, a product of our prejudices rather than theirs. What this illustrates about anthropology is that we walk a path, as I like to picture it, along a steep ridge where we risk falling down one slope of unduly exoticizing our subjects of study and down the other slope of unduly banalizing them, of reducing others to our common sense, dullest, or most ethnocentric assumptions; for example, the idea that the natives are bent on maximizing self-interest or that their practices can be explained in maximizing genetic reproduction, or whatever the reductive and ethnocentric formulation. Staying on the ridge, occasional and inevitable miss-steps notwithstanding, requires the kind of balancing act of practical reason or judgment that Aristotle describes as phronesis and that applies to intellectual no less than ethical or aesthetic contexts and aspects of thought. As I suggested earlier, it is a path without an endpoint or final resting place.
The ethnographic path also requires us (anthropologists, philosophers), as Gadamer realized, to return to our own concepts, both as a challenge to them and as a challenge to us to find the best language with which to describe the worlds of others. Thus, one criticism that has been applied to Evans-Pritchard is that, with respect to his corpus of three books on the Nuer, he apparently distinguishes their political sphere from what he calls Nuer kinship and Nuer religion. However, kinship, politics, and religion are "our" domains, not those of the Nuer, and not natural or logical categories [33]. Thus, a challenge to us is how to overcome these ingrained distinctions. This is very difficult both to do and to appreciate. In my book, Knowledge and Practice in Mayotte [34], I specifically did not distinguish between the religious and the medical or therapeutic, and I treated Islam and spirit possession equally within the same volume. This meant the book fell between stools; it was not evident where or by whom it should be reviewed or in what kind of anthropology course it could be taught. These courses have been, until recently, labeled by the same kinds of domaining terms-as the anthropology of kinship, religion, etc.-even though the first and primary thing we do in such courses is unpack or deconstruct the category term itself. The book was, in a sense, unintelligible to the anthropological public. In Island in the Stream [35], I argue that Islam, kinship, and community membership have been indissociable from one another in Mayotte. Further, in northwest Madagascar, spirit possession is equally "religion", "politics", "kinship", and "history" in our domaining terms and equally both serious business and entertainment [36]. By continuing to think in terms of such concepts and domains, we inevitably enact category mistakes by means of such questions as, "What is the relationship between kinship and religion?" or "between religion and politics?" We also render foreign concepts ostensibly intelligible to us at the expense of distorting their original locations.
I could continue to discuss hermeneutics, but my perception of what I do has changed somewhat from interpreting to describing. Put it this way: I now consider the latter the better description of what I do. By describing, I mean that I put things that I observe and overhear or converse about with my interlocutors under description. Furthermore, what those things are that I observe or talk about with them, are largely their acts of putting things under description as well. Thus a Zande puts a given condition under the description of "mangu" or a given person under the description of "a witch", and an anthropologist puts these acts under the description of "witchcraft" and places witchcraft under the description of "religion", and so forth.
Putting under a description is a relatively old concept for philosophers, but hardly taken up by anthropologists. Yet it covers much of what we observe people doing and much of what anthropologists themselves do, both in writing their field notes and in organizing or analyzing their material. As Elizabeth Anscombe showed [37], while some descriptions may simply be false, others are mutually compatible, and the choice among them is a matter of what is relevant at hand. Hence, an understanding of description renders many of our theoretical debates null. Arguments that appear to be between divergent explanations or interpretations can be seen as unnecessary once they are understood as different levels or angles of description. Moreover, the series of misunderstandings to which this gives rise is of the same order as many of the misunderstandings and disagreements we observe in the field. Some of the ostensible irrationality in the thought of others is simply that we, among ourselves, and they, among themselves, and we and they, between us, frequently operate at different levels or angles of description. Once again, it is not a matter of rationality in the abstract.
In other words, I am suggesting that we sometimes confuse different levels of description for competing arguments or interpretations. What we need to be clearer about are the criteria and reasons for operating at one level of description rather than another. It is a matter of sensibility to ascertain the level of description appropriate to what we are observing and a matter of practical reasoning for selecting the description under which, in turn, to place it. The anthropological description of so-called magical practices became infinitely more sophisticated once we realized that we were observing people putting things under descriptions by means of illocutionary acts and persuading others to accept them by means of analogical reasoning [38]. This is one way in which I think anthropological understanding and, indeed, that of social science more broadly can be informed by philosophy.
Analogical reasoning brings me finally to Lévi-Strauss. One of the things that Wittgenstein, Evans-Pritchard, and Lévi-Strauss share is an interest not in the hyper-rationalized discourse of experts but in what was so badly translated from Lévi-Strauss as "the savage mind", although La Pensée Sauvage is actually neither savage nor mind. 22 Rather than the hypostatized and static concept of mind, it is precisely something active and dynamic, thought or reasoning, that Lévi-Strauss is after. Furthermore, rather than savage, it is undomesticated thought, or, we might say, ordinary thought-thought not subject to the rationalization or intellectualization of professionals. However, perhaps none of these three authors discriminate as explicitly as they might between what Alfred Schutz distinguished as distinct attitudes or perspectives toward knowledge, which can be held at different moments by one and the same thinker. Schutz distinguishes among perspectives he calls the expert, the well-informed citizen, and the man on the street [41]. These are descriptions not of individual kinds of thinkers or levels of rationality but of thinking relevant to different situations. The same person who, when she is functioning as an expert, might use sleight of hand to extract sorcery from a client will also act as a compliant client (man on the street) when she is ill and diagnosed with sorcery, and it is from her body that the sorcery is extracted. 23 Lévi-Strauss, and those writing in his tradition, albeit very originally, like Mary Douglas [42,43] and more recently Eduardo Viveiros de Castro [1,3] and Philippe Descola [12,44], share certain assumptions about structure. They illustrate binary oppositions in the thought of others, and they work themselves by means of binary oppositions, in the case of Douglas and Descola, with two distinct binary oppositions depicted as perpendicular intersecting axes. These models vastly increase the scope of intelligibility of foreign worlds, but they are too tidy and, although outside the Weberian tradition, work, as Anne-Christine Taylor says positively of Descola, by means of ideal types [45]. In many ways, this runs counter to more existential and phenomenological perspectives and to the circumstantiality of history, albeit the latter is something that Lévi-Strauss acknowledges and that another of his distinguished interlocutors, namely Marshall Sahlins, tries to incorporate.
In their attention to and reliance upon binary oppositions as the basis for thought and hence for intelligibility, structuralists risk overlooking the prevalence and significance of the incommensurability of concepts, that is, the absence of a common measure between them. Incommensurability is itself a concept, one developed by Thomas Kuhn [46] and other philosophers of science, but equally, as well shown by Richard Bernstein in his book with the perspicacious title Beyond Objectivism and Relativism [47], taken up by Gadamer [5] and Richard Rorty [48]. 24 I also find it complementary with, or congenial to, ideas in Wittgenstein.
I have said that Lévi-Strauss overlooks incommensurability, but in fact, what he does is show people busily commensurating. They commensurate by replacing incommensurable concepts such as nature and culture with analogical and ostensibly commensurable terms or objects, commensurable in the sense they can be marked by binary distinctions, black and white, animals of land and animals of sea, upper and lower, and so forth. I suggest that not all concepts can be placed in binary opposition to other concepts nor submit to this kind of science of the concrete, as Lévi-Strauss describes it, without remainder or without instability. This is a point he acknowledges, but whereas he argues that the remainders are, in turn, put back into the analogical machine-I suggest that commensuration has its limits.
When the binary quality of oppositions is false and reified, one of the effects, as Ryle puts it, is the enactment of category mistakes. However, some incommensurable concepts and paradigms resist the process; in that case, practical reason entails moving between them, both intelligently and ethically. As Diamond argues, classification is insufficient to describe how we live with concepts. She writes, after Wittgenstein, "grasping a concept (even one like that of a human being, which is a descriptive concept if any are) is not a matter just of knowing how to group things under that concept; it is being able to participate in life-with-the-concept". 25 She suggests that we "get rid of the idea that using a concept is a matter of using it to pick out what falls under the concept and what does not . . . [and] see instead that life with a concept involves doings and thinkings and understandings of many sorts, into which one's grasp of the concept enters in different ways" [18] (p. 276).
To conclude, intelligibility means recognizing in the first instance that what other people do is not hold beliefs or speak in propositions (most of the time) but deploy concepts (Viveiros de Castro [1]) and place things under description. Second, it necessitates recognizing that many concepts, ideas, and practices are incommensurable with one another and, hence, are deployed according to practical judgment. Third, it means considering the rational and the ethical-or the intellectual and the practical-together rather than fully distinguishing them. Fourth, it means understanding that concepts are deployed within specific forms of life. Fifth, it entails acknowledging the inevitability of mistakes, not only those of our own making but also those among our subjects. These include confusions of levels and angles of description, inappropriate domaining, category mistakes, over-extending commensuration, and perhaps most generally, as Diamond puts it, the "underlying inadequacy in a philosophical view of language that ties description to classification" [18] (p. 266).
The broad question is, what are the implications for (Gadamerian) hermeneutics once the object of ethnography is conceived as a (Wittgensteinian) form of life rather than as cultural text, conversation, or practice? And conversely, what might hermeneutics have to offer a Wittgensteinian approach to intelligibility?

Conflicts of Interest:
There are no conflicts of interest.

1
The workshop took place at the wonderful facilities of the Wissenschaftskolleg zu Berlin, June 7-8, 2018. My thanks to Jean-Philippe Narboux for the invitation and provocation; to the Wissenschaftskolleg for their hospitality; and to fellow participants for much stimulation. I have kept the informal tone of the oral presentation as well as Narboux's request to focus on the three thinkers named above. Some of the ideas have been further developed in a Tanner Lecture delivered at the University of Michigan and published as Concepts and Persons [2]. Thanks also to Juliet Floyd, Sandra Laugier and the additional discerning reviewers. 2 This is true also of Marilyn Strathern's work on Melanesian gender and personhood [4]. 3 Since writing this I've learned from American anthropologist Thomas Csordas that he has some knowledge of both Hungarian and Navajo. My argument is not to suggest either that learning particular languages is the equivalent of coming into language in the first place or that proficiency in writing or reading a given language is equivalent to entering the forms of life of its primary speakers. 4 I continue to speak of "culture" and "cultural difference", Viveiros de Castro notwithstanding. 5 I am speaking here of human thought. When the concept 'alien' points beyond the human that is a matter on which I cannot comment, but see the discussion in Miguens, ed [11]. 19 And perhaps it is nonsense in the way that Austin shows that truth and falsity are not the criteria to evaluate illocutionary utterances [27]. 20 See Rappaport [28]) on the significance of Austinian performativeness for establishing truth and certainty but also on the mystification of performativeness. 21 For further anthropological discussion of Wittgenstein see the essays collected in da Col and Palmié, eds. [29]. Veena Das, especially, [30], offers the deepest appreciation of Wittgenstein by an anthropologist. 22 Lévi-Strauss [39] has been retranslated and retitled [40]. 23 I describe such an instance in Lambek [34], where I develop the three perspectives more fully. 24 For Rorty, incommensurable ideas or discourses are those not "able to be brought under a set of rules which will tell us how rational agreement can be reached . . . where statements seem to conflict" [48] (p. 316); see also Bernstein [47] (pp. 85-86). 25 Cora Diamond [18] (p. 266). The passage continues: "What kinds of descriptive concept there are is a matter of the different shapes life-with-a-concept can have. Life with the concept human being is very different from life with the concept member of the species Homo sapiens. To be able to use the concept 'human being' is to be able to think about human life and what happens in it; it is not to be able to pick human beings out from other things or recommend that certain things be done to them or by them. The criticism I am making could be put this way: linguistic philosophers have brought to their study of language an impoverished view of what can be involved in conceptual life" [18] (p. 266).