In anthropology, it’s emic all the way down

This work is licensed under the Creative Commons | © Marshall Sahlins. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau7.2.020


In anthropology, it’s emic all the way down

Marshall SAHLINS, University of Chicago

Response to comments on Sahlins, Marshall. 2017. “The original political society.” HAU: Journal of Ethnographic Theory 7 (2): 91–128.

My profound thanks to all the commentators for taking this text so seriously and thoughtfully. Given the large theoretical pretensions of “The original political society”—not to mention HAU’s own recent attentions to the “ontological (re)turn”—it is perhaps not surprising that most of the commentators took the occasion to recount their own ethnographic experience with these issues. The result is a set of documents each worthy of extensive contemplation on its own. All I have been able to manage here is to take them up at the point they engage directly with the “original” text. Also, I tried to put them in a sequence that might afford a connected reply, hopefully with some positive cumulative effects for elucidation of the issues at stake.

First, Carlo Ginzburg, particularly on “etics” and “emics.” True, Carlo and I have had this exchange more than once. But since I have never been able to make my point, any more than I have been convinced by his, I try once more.

Human worlds are meaningfully constituted. Verum-factum: the methodology of anthropology is essentially exegesis. What humans make is humanly intelligible. The many works of Carlo Ginzburg—himself a compatriot of Vico—are shining examples. But when he tells us that they entail an “etic” analysis of “emic” phenomena, he risks entangling the intersubjective conversation characteristic of the human sciences (as Lévi-Strauss described it) with the positivist epistemology of the natural sciences or of what God makes (as Vico described it). It is as if Ginzburg were searching for some culture-free language of description, whereas what is evidently needed to understand other peoples and other times is a culture-rich one.[158]

To be sure, Ginzburg explicitly disavows Kenneth Pike’s apparent conflation of the observer’s “etic level” with the “scientific approach.” Historians and anthropologists he says, engage in a “more tortuous, complex trajectory,” in the course of which their initially ethnocentric or anachronistic categories are “duly reworked” so as to “rescue from the actors some emic answers” (p. 139). Ginzburg does not say what this reworking of inapplicable “etic level” concepts consists of. What is clear, however, is that having started with avowedly emic concepts—ethnocentric or anachronistic as they may be—which are then modified in the encounter with the others, the result is going to be categories that are still emic (if something less anemic). So-called “etic” is “emic.” Of course, the human sciences have their appropriate languages by means of which they engage in empirical projects that augment our understanding of the phenomena at issue. Except that in the human sciences, most particularly in anthropology, it’s emic all the way down.

Revisiting the phonetic/phonemic distinction that originally gave rise to Kenneth Pike’s influential generalization of “etic” and “emic” will help us understand how a general matrix of culturally constituted emic categories provides the means for explicating the particular ethnographic case—whether a particular language or particular cultural scheme. While providing the phonetic grid for the determination of the sound system of any natural human language, the International Phonetic Alphabet (IPA) is not itself a thing of natural science, not a machine-recorded compendium of all possible humanly produced sound waves. The IPA is a systematic assemblage of all those sounds known to differentiate meanings in recorded human languages, all those phones known to function phonemically. The phonetic grid is phonemically constituted. True, these sounds are described and classed according to their production by the human vocal apparatus: operations of the tongue, lips, teeth, and so forth. But of all possible permutations of the vocal apparatus, only those involved in the production of semantic differences are inscribed in the IPA. Hence for purposes of scientific description, the natural attributes of sound production as well as the sounds as such are under “emic” control.1 It thus becomes possible to accurately describe the phonemic pattern of any given language by reference to the positional values of the IPA. For all that it is meaningfully ordered—that is, by relative cultural values—the science is objective as well as empirical.

Mutatis mutandis, something quite similar can be said in respect of cultural practices. And happily Marilyn Strathern’s generous contribution to this discussion is a case directly in point. Marilyn is one of anthropology’s Sky People. The intellectual life-force in the ethnographic seedlings she plants and tends develop into whole new clans of anthropological knowledge. Just so, her commentary here is an enrichment of the original. But besides the results, I first call attention to the method.

In relation to an earlier study (with Andrew Strathern) of Hagen spirit cults, Strathern wites, “Sahlins pulls me up sharp . . . what strikes me now is the classificatory scheme of the writers” (p. 150). Moreover, the fruitful transformations she now makes of Hagen cultural schemes also depend importantly on categorical innovations developed in Philippe Descola’s analyses of Achuar. As summarized [159]in “The original political society,” these include the observation that the common Euro-American notion of “production” does not adequately describe human praxis in a metahuman cosmos where plants (inua) are children of the women who cultivate (nurture) them, even as the godly mistress and original mother of cultivated plants is the source of their abundance and the success of the women’s gardens.

Thus endowed with a new classificatory grid informed by previously unremarked cultural schemes, Strathern offers new empirically supported insights into Hagen ethnography. Among these: that plants, pigs, and people here are no more “domesticated” (removed from the control of nature) than they are “produced” by humans (rather than made to flourish or fail by the diverse metaperson sources of vitality); accordingly, that sacrifice is a relation of “production” (so-called); and that materially (as in subsistence) and socially (as in clan exogamy), life comes from outside.2 Moreover, adopting still further cultural notices and notions from Françoise Panoff’s study of life-force transformations in the taro ritual of Maenge people of New Britain, Strathern offers an alternative to the political emphasis on a cosmic polity in “The original political society,” suggesting instead a paradigmatic emphasis on the ways that people, while seeking prosperity and avoiding dangerous contact, become involved in one another’s bodies, thereby influencing one another for good or particularly for ill. At the least, Strathern opens a new inquiry into the “egalitarian” character of such societies. For myself, however, I would retain the salience of the cosmic polity, precisely because it uniquely entails the encompassment and subjection of the human order in a larger society inhabited by powerful metapersonal beings wielding universal life-forces—like the Maenge god in the python whose feces convey the soul of the taro to the people’s gardens. Not that one need chose; this is all one pattern of complementary parts. Of particular interest here is how Strathern works to develop our understanding of it.

Analogous to the IPA, the grid of ethnographic categories Strathern deploys to analyze Hagen and Maenge comes not only from her own field experience, but explicitly and implicitly from a wide range of alternative cultural possibilities. Using “metaperson” and “inua,” she engages the cosmologies of Inuit, Chewong, Arawete, and others; she develops notions of enchanted persons from reports of the reciprocal identities of men and gods in Papua New Guinea; from accounts of trans-species kinship and the divine source of garden growth among Achuar, she totally revises in a corollary vein the modes of human and subsistence generation in Hagen, and so forth. But then, neither does the grid of categorical permutations and relations stop there. For in calling on the works of Descola, Sahlins, Panoff, Schwimmer, and the untold others she has studied, each working from his or her own cross-cultural experience, Strathern brings a veritable encyclopedia of cultural schemes to bear on the particular case of the Hageners. One is reminded of Lévi-Strauss’ aspirations for a Mendeleev-like table of the structural elements—and of some reasons they were never achieved. Cultural variation is always in a state of becoming, even as the knowledge thereof is dispersed among members of the discipline, each of whom can master and advance only some relevant part of it. The entire table of cultural [160]forms is only virtual, but for all that indispensable for the science of them. Cross-cultural reference is a condition of anthropological possibility.

Now, for another example: Yup’ik. If you’re lucky, your students, by their own work, will become your teachers. I have been very lucky in Annie (Ann Fienup-Riordan). She turns her gentle reproach of my lament about studying disappearing and disappeared cultures into a revealing ethnography of the perduring animism of Yup’ik modernism. Speaking to distinctive qualities of the Yup’ik cosmos—the dominance of Weather as universal being as opposed to Sedna of Central Inuit, and the relative insignificance of metaperson “ownership,” among other things—her work makes evident a critical lacuna in “The original political society”: the inattention to the diversity of cosmic polities. If, indeed, we admit to the existence of such larger societies of which the human component is but a part, a great deal of ethnographic work remains to be done—or redone. If I may be allowed to carry the argument to its logical extreme (or perhaps beyond that), it’s almost as if we would be advised to start anthropology all over again.

There is also one particular aspect of Fienup-Riordan’s comment that not only reveals a critical problem in “The original political society” but also goes a long way toward resolving it. This is the vexed issue of transcendence, of whether it is appropriate to speak of an “other world” or “other worlds” in regard to the cosmic schemes at issue. Fienup-Riordan’s description of Yup’ik generally comes down on the side I had argued: that humans and metapersons inhabit the same reality, that the latter are commonly present in the experience of the former, and that metapersons interact with people socially, frequently, and in humanly customary ways such as marriage, sex, exchange, and various other interpersonal relations. Here there is indeed no “supernatural,” as in the Judeo-Christian sense of the absolute transcendence of God in relation to the “this-world” of humanity: a world that was made by Him out of nothing and is otherwise without subjectivity (cf. Frankfort et al. [1946] 1977: 167 et passim).

Still, among Yup’ik and many other peoples presently under consideration, there are divisions between humans and the metapersons of their ken, and I have given no account of what these divisions consist of or how they are effected. “Division,” not incidentally, is how Marilyn Strathern described the way the concentration of powers among Sky People separated them from Hagen people. On the other, human side, as reported by Signe Howell for Chewong, their taboo on ridiculing forest animals would divide them from the animals on the matter of familiarity. What Fienup-Riordan importantly adds to the discussion is the observation that many of the very many Yup’ik rules of behavior “were designed to both create boundaries and passages between worlds. While the Yup’ik, like the Chewong, view the cosmos as all-inclusive and ‘make no meaningful separation between nature and culture’ (Sahlins, p. 93), boundary-making—and boundary-breaking—activities were ubiquitous in the past” (p. 136). Speaking of relations between “worlds” would not seem inappropriate here, since what is at issue are limits between domains and beings of the same inclusive world. This, I offer, is the alternative to transcendence in the great range of societies that human share with other subjects: permeable boundaries between the personkinds occupying the one ontological space of universal dimensions. And if I may be allowed even higher flying: this notice of negotiable boundaries, on the one hand, complements Strathern’s insights on how [161]people, dealing with life-and-death forces, become involved in each other’s bodies; while, on the other, it responds to the ground of the cosmic polity as argued in “The original political society,” that the vitality, mortality, and prosperity of human societies depend on external forces beyond their control.

But then, precisely, Marc Abélès, having deftly and accurately decoded the relationship between “The original political society” and “The original affluent society” of ancient memory, and having, moreover, developed the implications of both for a critique of the contemporary political economy of capitalism, questions whether the cosmic polity of metapersons could actually be credited with governing the human fate in the societies discussed in these essays, could actually be the agents thereof. He asks: “Can we say that the spirits, the divinities, the ancestors, and so on, to which the indigenous traditions attribute these powers really exercise them? In the ethnographies quoted by Sahlins, I don’t find any clue of the agency of these metapersons . . .” (p. 131). I have to admit I am puzzled, since it seems obvious that for the purpose of anthropological understanding of cultural-cum-ontological others, it’s not up to us to say—least of all to say what “really” happened, for what’s at stake is what it is that happened. Was it a wink, or a blink? Nor is it, then, the question of “can the subaltern speak?,” but rather, can we hear her?

The evident “clue” to the agency of Tanko, the maker of violent storms, is that Chewong say he felled the tree that killed the people. The clue to the agency of the goddess Sedna is that Inuit say she is hiding the seals from the hunters. The clue to the agency of Afek is that Mountain Ok people say she made the taro flourish this year. (Note that these are Kantian objective judgments on the order of “the sun warms the stone”; and inasmuch as the proposition does not depend on an affirming or doubting subject, that is, as a question of faith, one could argue it is a matter of knowing rather than believing.) Nor is this the error for which Lévi-Strauss (1966: 38) famously reproached Marcel Mauss, that is, for (supposedly) allowing himself to be “mystified by the native.” One simply has to credit what the people say in such matters in order to understand their practices in regard to them, for such are the premises and principles on which they act. Asserting that the ancestors did not really make the pigs grow fat will not help explain why Hageners are compelled to sacrifice to them. But more importantly in the present context, agency in such endeavors as hunting, gathering, and gardening is culturally and historically relative, and accordingly variable: for it is not the empirical question of who is actually performing the work—it is not in this sense “etic”—but who is socially assigned the responsibility for creating the outcome—it is in this sense “emic.”

As already said too often, in one significant sense no human is responsible for the natural processes and conditions that cause plants grow or make animals available: these are things God makes, in the Vichian acceptation—which in a broad way is also what the Inuit, Mountain Ok, and others say. In this regard the modern industrial farmer is no more the agent of the growth of crops than some people’s ancestors may be, or even less so, if the ancestors are customarily buried in garden land. (Iban say that rice is the transubstantiation of the ancestors.) No doubt there are even American farmers who thank God for a good harvest—or blame Obama’s policies for a bad one. On the other hand, a modern industrial farmer may be deemed the responsible agent of production where immigrant Mexican laborers actually do the work: for in a capitalist system, it is notoriously the owners [162]of the means of production who are the “producers,” the “entrepreneurs,” the “job-creators.” (Talk of fetishism.) In other cultural orders it may be the garden magician, the shaman, the clan elder, the big-man, or the chief who is credited with the success or failure of people’s subsistence endeavors, or it may be the hunter or the gardener him- or herself: in any and all cases the attribution of agency being dependent on privileged access to the metaperson authors of natural fertility. The game will come when the native Australian elder ritually renews the painting of the wonjina on the rock shelter wall. Agency is always meaningfully constituted and socially attributed, even as it is often independent of the physical action of which it is the acknowledged cause.

Marc Abélès and the other commentators, in expressing the possibility that “The original political society” could lead to some rethinking of their own work, do the essay great honor. Yet in thus demonstrating their openness to evidence and alternative interpretations, which is to say their scholarly qualities of the highest order, they do even greater honor to themselves—and none more so than Signe Howell. For of all the ethnographic notices in the piece, I uniquely challenged her characterization of Chewong as an egalitarian order, since included under that description were the people’s relations to metaperson authorities who exercised life-and-death powers over them. I hasten to acknowledge, however, that Howell’s extraordinary ethnographic work was an inspiration and encouragement for many of the arguments in “The original political society.” Critically significant have been her observations on the inclusion of the human community in a larger cosmic society; on the ubiquity of metapersonal beings in the Chewong world and their reality in people’s experience; on the “cosmo-rules” (not “cosmic rules,” my bad) covering even the minutiae of people’s daily existence laid down by these metaperson authorities and sanctioned by them with misfortunes ranging unto death. Reciprocally and generously, Howell recognizes these hierarchical aspects of the Chewong world—but interestingly she maintains it is universally and ideologically egalitarian, even in respect of people’s relationship to the punitive godly powers who wreak natural disasters upon them.3 Since the apparent contradiction—or “conundrum,” as Howell has called it—is difficult to square, it allows me to conclude these remarks with another tribute to her scholarship: it opens a lot of questions.

Some are recurrent questions about “egalitarian societies” so-called. Is there an explicit value to the effect that all persons are or should be equal—what’s the term or phrase? is it distinct from “same”?—or is it a negative prescription to the effect that no one should get ahead of others (as in the oft-repeated, modern Hawaiian proverbial, “we are all crabs in a bucket; when someone tries to crawl out, we grab him and pull him down”). More particularly, since “equality” is said to characterize [163]relations between Chewong and all other “people like us,” how exactly do they relate to the godly metapersons who set and enforce the rules of their behavior? There seem to be two answers. One suggests an explicit value of equality, as when Howell writes: “the relationship between human and superhuman is predicated upon a profound encompassing value of equality” (p. 146). Rather more intriguing, however, is a suggestion of denial: “Chewong displayed a perverse tendency to ignore all implications of hierarchical differences. What emerged as important was a necessity to keep that which is different apart; to separate like from nonlike in a flat nonhierarchical space” (p. 144). If this sounds like a version of “society against the state,” maybe the big difference the Chewong are perversely ignoring, precisely by creating a “nonhierarchical space,” is between themselves and the Malay sultanates they have probably known for centuries. Still, possible critics and skeptics of “The original political society” will be happy to learn that it all ends here with, “I don’t know the answer.”


Frankfort, H., H. A. Frankfort, John A. Wilson, Thorkild Jacobsen, and William A. Irwin. (1946) 1977. The intellectual adventure of ancient man: An essay on speculative thought in the Ancient Near East. Chicago: University of Chicago Press.

Graeber, David, and Marshall Sahlins. 2017. On kings. Chicago: HAU Books.

Lévi-Strauss, Claude. 1966. “Introduction à l’oeuvre de Marcel Mauss.” In Sociologie et anthropologie, Marcel Mauss, xiv–xvii. Paris: Presses Universitaires de France.

Sahlins, Marshall. 1976. “Colors and cultures.” Semiotica 16: 1–22.


1. For an analogous discussion of the semiosis of color percepts, see Sahlins (1976).

2. The alterity of power and of the means of life and death are major themes in On kings (Graeber and Sahlins 2017), of which work “The original political society” is the opening chapter.

3. A relatively minor disagreement is raised by Howell’s objection to my remark on the existence of kingly beings in heaven even where there are no chiefs on earth: for the Chewong “superhumans” in question, Tanko and the Original Snake, she contends, are not really rulers. But I did mean “kingly,” not “kings”, and these metapersons do have kingly attributes, including judicial functions, powers of capital punishment, some control of natural phenomena, and Tanko himself is the effective (pro)creator of humans—not to mention that also at issue are Afek, Magalim, Sila, Sedna, and their like in other societies noted in the essay.