On the very possibility of mutual intelligibility

This work is licensed under the Creative Commons | © G. E. R. Lloyd. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau4.2.010


On the very possibility of mutual intelligibility

G. E. R. LLOYD, University of Cambridge

This paper acknowledges that the issues of translation and of translatability are general and concern the possibility of mutual intelligibility in many registers and including within a single natural language. Both anthropologists and ancient historians are faced with such problems, where the historians are at a disadvantage in not being able to check their understandings with those whom they are seeking to understand. But faced with seemingly paradoxical statements, beliefs, or practices, we must and can avoid the apparent dilemma (either those statements must be rendered in or reduced to our terms or we must admit they are strictly incomprehensible) by insisting on the revisability of our existing conceptual framework, especially in relation to such key terms as personhood, agency, causation, and nature. Again, instead of insisting on the dichotomy of literal and metaphorical, we should allow that any term may exhibit what is here called semantic stretch. Moreover, if we accept (as is argued here) that the phenomena or realities described are multidimensional, then the goal of a single definitive translation is a mirage. The open-endedness of translation is no threat to mutual intelligibility but its precondition.

Keywords: intelligibility, incommensurability, paradigms, ontology, relativism, translation

Like the ethnographer, the student of ancient societies is faced with a recurrent problem of translation, and in one important respect suffers from an obvious considerable disadvantage. Modern ethnographers can question members of the groups they study to get some reaction to the question of whether or how well they have understood them, though the quality of the response will reflect the relationship the ethnographer has been able to build up. He or she may be told, simply out of politeness, how brilliantly he or she has grasped the meaning of their words and of their actions. The ancient historian is just confronted by documents and texts, those that have survived the vagaries of transmission and are mostly now buried beneath a pile of previous interpretations.

In both cases we have to be realistic about the level of understanding that can be achieved. But we have also to be realistic about the level of understanding attainable even when the conditions are optimal, when we are dealing with someone who shares with us the same natural language, maybe also the same upbringing and environment. The particular problems of trying to grasp the meaning of some ancient text in a foreign language, or the work of an author as a whole, are mirrored, even if less severely, in our efforts to understand some writing in English. What is it to say that we have grasped the meaning of King Lear, let alone of Shakespeare as a whole, or, to come down to today, of Salman Rushdie or A. S. Byatt? To state the obvious, it is never the question of “the” meaning, just the one, but always of multiple meanings. The point is familiar from religious hermeneutics, and although in that context we may sometimes suspect deliberate mystification, the lesson that readings are open-ended is surely obvious across the board. Not even scientific communications are immune to multiple interpretations.

But if being realistic means we have to acknowledge difficulties, it also means that we should not be unduly pessimistic, which is certainly part of the message Hanks and Severi have insisted upon. One important point about translation and about understanding in general is that, although always difficult and always imperfect, it is seldom (they would say never) the case that we have to admit to complete and utter defeat. That is true even with terms that are admitted to have no exact single equivalent in any but the natural language in which they occur, like German Gemütlichkeit, Welsh hwyl, Russian toska, or Ifaluk fago, where we can get at least some inkling of what they cover.1 No ethnographer returns from the field to say that he or she understood nothing of the society that was the subject of investigation. No student of ancient Greek philosophy admits to understanding Plato not at all. It is only if we have no grasp whatsoever of a particular language that we must admit to total incomprehension, of the words at least, and even then the body language of our interlocutors may leave us in little doubt about some of their feelings.

What the first line of the Daodejing means, dao ke dao fei chang dao, has been the subject of countless commentaries down the ages in many different languages.2 But if you have some classical Chinese, you will know, for instance, that dao can mean not just “way,” but also “guide,” that is, show as the way. So while the first dao is “the Way,” with all its multiple associations, the second dao can be taken as “shown as a dao” so that the whole says, very roughly, “the Way that can be spoken of as a way is not the constant way.” The constant (chang) Way is thereby contrasted with others that fail the requirement of constancy, because they can be spoken of. That illustrates what an approximate translation can be like, with the added bonus, perhaps, of the substantive message to do with (in)expressibility that this particular famous line conveys.

We can study the range of usage of dao and that of chang. At that point some might attempt to cordon off “literal” from “metaphorical” or “figurative” uses. But that is one of the dichotomies that cannot be taken for granted. I have rehearsed my reasons for challenging it on other occasions.3 My preferred alternative is to make use of the notion of “semantic stretch,” which has the advantage that it allows that all terms have some stretch. Dao may be exceptional: there is a dao of butchery and even one of robbery, though it is the dao of the Sages to which one aspires. Yet it is as well to make allowance for stretch in every term in any communicative exchange, and that is even before we factor in further complex points from the pragmatics of the situation. When a live conversation is in question, there is the body language of the speakers to consider and the relationships between them, of friendliness or hostility, cooperativeness or competitiveness, superiority or deference.

That may seem to open up an infinite number of options for interpretation, far beyond anything that Wittgenstein contemplated when discussing “family resemblances” in his famous discussion of the concept of “game” (not forgetting his own notion of “language games,” e.g., Wittgenstein 1963: paras. 66–71, 83, 86). Yet semantic stretch, so far from precluding progress in understanding, may even be a necessary condition for it. The possibilities in interpretation are not limitless, even when we are faced with the Daodejing, but they are liable to be prematurely circumscribed if we start from the assumption that there is just the one, correct, understanding to be secured by homing in on a single “literal” meaning. We must acknowledge that much always escapes us; but that does not mean that we are always in a state of complete bafflement. Reminding ourselves that we may not have got it right is always salutary, but should encourage us not simply to give up in despair, but to renew our efforts. This is of course what we naturally do, at least when we are not inhibited by some sense that we must be able to resolve the philosophical issues before we can even start.

Texts look as if they ought to say something that we shall have some opportunity to construe in other terms, even if that may involve heavy paraphrasing. But what about pieces of music, works of art, ritual performances? We should not say that they do not signify anything, even when, as often, they do not refer. But to begin to put into words what we believe them to signify is always difficult, often seemingly impossible. What Beethoven’s Fifth Symphony meant for the different members of the audience whom E. M. Forster described as listening to it in Howards End is expressed, in that novel, in what now seem rather jejune terms. But in any case what a piece of music means for one person on one occasion is never going to be precisely what it means for another on that occasion, or even for the same person on a different occasion. Again the hermeneutic temptation is to elide all that diversity to get at some essential, core, understanding, but to do so is always going to be reductionist, always to miss the opportunity to explore other possible resonances and associations.

Warnings as to the difficulty and imperfection of understanding are always needed, but it is amazing how much we can understand, including across different languages and dealing with unfamiliar subject-matter. We should never underestimate the capacity of humans to learn and to adapt, even in the face of pressures that urge us to stay with the views and practices of our elders and betters. Sure, we sometimes delude ourselves that we are on the right track. Sure, we are sometimes the victim of deliberate deception on the part of our interlocutors. But that would not be possible if there was always deception. Our default assumption is that we are not being deliberately misled by our partners in conversation. Obscurantism, in turn, is only recognizable by contrast with the relatively plain and clear. The very fact that, with the help of an interlocutor, or just on our own, we can improve our grasp of what is being communicated, and correct some of our misunderstandings, should encourage us to continue our efforts, including even in the face of an insistence, on the part of those interlocutors or the apparent message of a text, that what we are dealing with is the inexpressible. Dao ke dao fei chang dao takes away with one hand, but gives us something with the other.

It is true that some aspects of modern philosophical discussion of the problems have not helped as much as they might. Quine’s inscrutability of reference (Quine 1960) and Kuhn’s incommensurable paradigms (Kuhn [1962] 1970) certainly underline the difficulties. But we can concede that ultimately reference is inscrutable without conceding that it is arbitrary. Even if there is always an element of indeterminacy, we can narrow down the possibilities by a process we may compare to bracketing—in particular by excluding what lies outside the brackets. There was no ostrich on the scene when our friend announced “gavagai,” while there was indeed a rabbit, so it is less likely that “gavagai” has to do with an ostrich or even temporal ostrich slices than with something to do with rabbits, even though it may not be the creature in mind, but rather the event, or again it may be neither. Even when a rabbit event occurs, there are always plenty of other items and occurrences in the scene that may have occasioned the comment. Faced with the first exclamation “gavagai,” the outsider will be baffled: sufficient repetition of the word should get the process of bracketing under way. Similarly I learnt to recognize hwyl (roughly, inspired speech) by being exposed to it on many occasions and registering that people acknowledged it in some speakers, but not in others, in some performances, but not in others.

As for Kuhn’s incommensurabilities, they do not preclude, but may even presuppose, the possibility of comparison, where at least we can make a start. There is no common measure for the side and the diagonal of a square. But we can certainly say that the diagonal is longer than the side. It is only if both are recognized as lengths that we can say they are incommensurable. We do not bother to remark that there is no common measure between a length and a color, for that involves a straightforward category mistake. Paradigm shifts always pose tough problems of interpretation, especially when the same term, say “force”, or “weight,” or “mass,” comes to be used with quite new senses and referents. But in the stock historical instances used to illustrate such shifts, we should not say there was total lack of comprehension between the parties. Copernicus certainly had a fair grasp of Ptolemy’s astronomical system, Galileo of Aristotle’s idea of natural motion, Einstein of Newton’s classical dynamics, even when the definitions of key terms were being transformed and new ones had to be coined to convey the new understanding.

But while the ideas of Quine and Kuhn have often been construed as threatening to undermine mutual intelligibility, conversely other attempts to come to its defense likewise may suffer from shortcomings. Faced with such famous but muchabused examples as the Nuer belief that twins are birds (Evans-Pritchard 1956) or the Dorze’s that the leopard is a Christian animal (Sperber [1974] 1975), some adopt Davidson’s hermeneutic principle of charity in interpretation (Davidson 2001), which recommends that whatever statements are reported should be construed, so far as possible, as making sense in our terms.4 Maybe we can find points of similarity between twins and birds and again between leopard behavior and Christian behavior to see how the reported belief can be made to make sense, without our having to follow those interpreters who have recourse to the idea that the statements were not meant “literally” but only “metaphorically.” Trying to decide between those alternatives regularly led to an impasse, the problem being compounded by the fact that the actors themselves, the Nuer and the Dorze, had no such explicit categories.

On the one hand, the supposition that those holding what seem to us counterintuitive beliefs are just foolish or irrational obviously will not do as a general methodological principle. There are fools in every society, not excluding our own, but attempting to diagnose wholesale folly in whole communities is not just racist, but hardly compatible with their evident ability to survive, often in difficult circumstances, including many where the average urbanized citizens of “advanced” industrial societies would simply perish. Yet that of course is not to say that every custom and belief that is maintained in any human group is well adapted to the aims of survival or of flourishing. That would be straightforwardly to commit the functionalist fallacy.

But on the other hand the translation of Nuer or Dorze beliefs into terms that make sense according to our given categories presupposes that those categories are already up to the job, and there is no reason a priori to go along with that. Evans-Pritchard was a great ethnographer, but, truth to tell, some of his interpretations now reek of some of his own preoccupations, theological ones, perhaps, especially. Faced with those counterintuitive statements, whether in ethnography or in ancient texts, we may need to revise our own categories and understandings, quite substantially perhaps, on such matters as the notion of a person, for instance, or of agency and causation.5 It cannot be assumed that our existing concepts will be adequate, and to do so is to miss the opportunities for learning that ethnography and the study of ancient societies both present. I believe my exposure to ancient Greek and Chinese ideas has taught me a thing or two. I shall give an example—nature—shortly.

Such general points are particularly germane to the recent ontological turn in anthropology, to explorations of radically different ontologies in Descola or of perspectivism in Viveiros de Castro.6 In the perspectivism of Viveiros de Castro it is not nature that is universal while cultures differ. Rather, all beings share culture while their natures differ, so this is monoculturalism and multinaturalism as opposed to multiculturalism and mononaturalism. Moreover, the key categories that we might suppose to be given in nature, “animal” and “human” among them, turn out to be inherently relational. For while humans see themselves as human, and animals as animals, and even spirits (if they see them) as spirits, animals see themselves as human and humans as animals whether as predators or as prey (Viveiros de Castro 1998: 470–71; contested, however, by Turner 2009).

Descola ([2005] 2013) invokes some of the same ethnographic data in defining what he calls “animism,” but that for him is only one of four different ontological regimes, varying according to the continuity or discontinuity they assume with regard to interiority and physicality. He uses those two differentiae to give him his fourfold schema, animism, totemism, analogism, and naturalism, which he uses to investigate practices of giving, taking, exchange, producing, protecting, and transmitting. While Descola calls these “ontologies,” Severi has pointed out that they differ from the philosophical usage of that term in being much looser and less explicit. “What is particularly interesting about them is precisely their unsystematic character, the fact that they always leave a space open for different strategies of thought” (Severi 2013: 195). This was a point he made to drive a wedge between some of my ancient ontologies (Lloyd 2012) and Descola’s regimes (ideal types, as Anne-Christine Taylor [2013] has insisted).

The implications of both Descola’s and Viveiros de Castro’s views for translation and understanding are momentous. From the jaguar’s perspective the blood of his prey (as we see it) is manioc beer. This controlled equivocation, as Viveiros de Castro calls it, might appear to a naturalist completely to undermine mutual understanding, to the point where the only thing we can understand is that the jaguar’s perception is indeed radically different from “ours,” though that is not to say anything about his understanding, except that it appears that jaguars, like us, enjoy drinking “beer.” Yet that is to miss the whole point of perspectivism, which is that both the senses and the referents of terms (including “beer”) shift across the languages of different kinds of creatures, being relative to the creatures in question, in particular being determined by the bodies they have. Translation is, then, a matter not of finding equivalent words to convey information about a single world, but of identifying different worlds to which the same words apply.

But how is the jaguar’s perspective to be accessed? It is only shamans who are in a position confidently to pronounce on that, for they alone can cross species boundaries, which in turn means that what is taken for a jaguar may be a shaman in disguise, or vice versa. It is certainly not easy to know where you are. But that is precisely the fundamental message. Rather than conclude that this difficulty, for us, of accessing others’ perspectives undermines the whole enterprise, we should reflect on what we can learn from considering what it would be for persons and substances to be relational, where we can start from, and use as a model, our familiar—banal—acceptance that the same individual can be both father (of one person), brother (of another), and son (of yet another), though in the perspectivist process “the same individual” gets to be radically reinterpreted. The Achuar and the Araweté and many other groups, not limited to Amazonia, of course, hold that other beings besides themselves are defined by the culture, rituals, rules of exchange, and so on, that constitute their way of being in the world. But what each kind of being apprehends depends on the bodies they have.

The same applies, these people would say, to the Whites who come to study them, for they (we) have the customs they (we) have because their (our) bodies are as they are. It is clear that the Achuar and the Araweté themselves puzzle over the Whites studying them. Indeed they can be said to do anthropology on the Whites, as much as the Whites do anthropology on them, as Viveiros de Castro suggested in his Cannibal metaphysics ([2009] forthcoming).7 But their anthropology does not presuppose the same commonalities and divergences as ours does, for, as I noted, their commonalities relate to culture, their divergences to nature.

Now I must recognize my own limitations in the face of the ethnographic aspects of the issues of translation raised by the “ontological turn.” But I do not think it is impossible to bring to bear some historical points from the study of ancient ontologies. At least my investigations of those ontologies prompt me to propose a different way of bypassing the treacherous dichotomy between nature and culture, which was of course what Descola’s book set out to transcend and was radically revised by Viveiros de Castro. My historical analysis of how the concepts of nature, natura, Greek phusis, originate in the West yields what may be a crucial point, that they were very much the product of a particular polemical situation, for which we have direct evidence in Greece.8

Those who went into battle, there, under the banner of phusis were dubbed the phusikoi or natural philosophers (they begin to be prominent in the fifth and fourth centuries BCE), and they claimed “nature” as the domain over which they were to be the acknowledged experts. Where traditionally in Greece such phenomena as earthquakes, thunder and lightning, eclipses, and diseases had generally been assumed to be the work of the gods (they were not natural phenomena then), the phusikoi argued that that was a category mistake. It was to ignore that those phenomena had regular causes, natures, in fact, that could be investigated; and they, the natural philosophers, could supply the correct theories and explanations (though most of those they proposed were quite fantastical). Where the traditionalists faced with lightning saw Zeus at work, the naturalists spoke of clashing clouds or whatever. Where the traditionalists saw eclipses as omens, the naturalists said they were regular and predictable. And note that in this instance it would be hard to apply the point that these different perceptions were due to differences in the bodies of the actors concerned. We move into a different mode of discourse, one that depends heavily, most would say, on literacy and a certain level of complexity of social organization.

Yet reference to other ancient societies shows that literacy by itself cannot be the whole answer. Ancient China had no single concept that covered what phusis covered in Greek or natura in Latin, and the same applies to every other ancient society with which I am familiar—Mesopotamia, Egypt, India, though I know far less about them. The Chinese recognized the spontaneous as the spontaneous, heaven and earth as heaven and earth, the different characteristics that different creatures (including humans) are born with as those characteristics, and so on. But they were not tempted to suppose that they were dealing with the same problem in all those instances. Nor were they tempted to read off value judgments from “nature” as such, even though they certainly debated moral questions and took different views on whether humans are inherently good, bad, or indifferent.9

Nature, I conclude, is not natural at all, but a cultural artifact,10 as much as a political regime or a set of religious beliefs and practices is. And that should release us from any assumption that whatever people thought they were dealing with in the physical world has to correspond to “nature,” has to be shoehorned into our category in other words. Of course some idea of the regularities in the phenomena is universal: we rely on that when we plant crops or light a fire. But which phenomena they are, and how exceptions are to be accounted for, are questions where we cannot just assume that an explicit concept of nature will provide the answers.

So when I claim that nature is an invention, I am sometimes understood to mean that just the understanding of nature is. But in fact I make a much stronger claim, that our notion of nature is not fit for purpose, but a trap. If we hold that nature is out there waiting to be discovered, we have simply not been critical enough of our own pet assumptions. That is a conclusion that the ontological turn in anthropology partly agrees with, but maybe for different reasons.

So the way I recognize the pluralism in ontologies proceeds rather differently from Descola’s or Viveiros de Castro’s approach. I agree with them in rejecting the assumed privileged status of a naturalist ontology. But my brand of ontological pluralism is a matter of what I call the multidimensionality of the phenomena or, alternatively, the multidimensionality of reality. It may seem shocking to consider those two formulations interchangeable. But the appearance/reality dichotomy is another one that needs to be pensioned off. In many contexts what is real is what appears, and conversely, though, to be sure, appearances may deceive. But then reality too may hide. The more important point, in both cases, is the possible, indeed the likely, multidimensionality.11

Let me cite one of my favorite examples, color, to illustrate that multidimensionality.12 In that case, the three dimensions of hue, luminosity, and saturation provide three different sets of differentiations, although that point is not always taken into account, particularly by those on the hunt for cross-cultural universals. There is no one correct way to talk about color. We should not privilege one of the three ways and exclude the other two. All three are valid, and it is a mistake to think we have to choose between them. That does not mean that this introduces ambiguities that preclude generalization. It is true that a multidimensional phenomenon cannot be given a per genus et differentiam definition. But it can and should be given a disjunctive definition, where each of the disjuncts (in the color case, hue, luminosity, and saturation) can be identified unambiguously, though each has its considerable semantic stretch.

But if thus far my position is “relativist,” it is not at all relativist insofar as I recognize that in each case there are more or less correct ways of doing the differentiations. Color talk is certainly not merely arbitrary: it is not the case there (or anywhere else) that just anything goes, whatever Feyerabend may have claimed. Nor is it impossible to allow communication between different modes of color talk, even though the particular vocabulary for the particular differentiations in view may be distinctive for that particular mode. Multidimensionality does not rule out mutual intelligibility. In this case it may even be a necessary condition for it.

The consequences for translation and understanding go like this. A vocabulary that simply differentiates hues will not by itself be up to the task of capturing the other two modes of differentiation. Thus far translation is not possible if we stick to that single vocabulary. But why should we have to? What we need is complementary modes of discourse to do justice to all three dimensions of differentiation, yielding a more comprehensive understanding of the subject-matter. But then there is nothing to prevent our conveying the complementary character of the phenomena in different natural languages, with more or less adequate translations between them. In some cases neologisms will be needed. In many cases the idiosyncratic range of particular color terms in one language will need a gloss or a paraphrase in another: the French “brun” is an odd-ball from an Anglo-Saxon point of view, and so too is “blond” used of tobacco and of beer. But of course the phenomenon is not confined to French: ancient Greek xanthon and chloron (conventionally inadequately rendered as “yellow” and “green”) are other examples. Chloron, for instance, denotes what is fresh, unripe, full of sap.

In many cases what may be represented as color terms may come to be seen as not color terms at all, but rather terms to pick out the living from the dead, or the succulent from the dry.13 Within a single natural language, there will be shifts between one register and another, and other possible sources of ambiguity. But at least we shall not be driven, as the former British prime minister William Gladstone thought he was driven, from the observation that ancient Greek focused especially on luminosity to conclude that they were all color blind and could not distinguish hues (Gladstone 1877). He never read, or if he read, he ignored, Aristotle on the rainbow.

Where, I may now ask more generally, do my suggestions leave the epistemology of anthropology or of ancient history? Does the ontological turn in anthropology spell the demise of its epistemology (as some have thought), leaving us perhaps with the conclusion that epistemologies are relative to the distinctive ontologies in play? Once again that gestures toward a conclusion of mutual unintelligibility. But once again that may be resisted.

It is true that the traditional dichotomies or dualisms within epistemology, between reason and perception, or between the a priori and the empirical, once again exhibit their limitations if considered as alternatives. But the correct response is not to abandon both sides of each pair, nor to plump for one to the exclusion of the other, but to combine them. That corresponds to my own practice here, for on the one hand I have been discussing what understanding is possible on the basis of an abstract analysis, but on the other I do that with as much attention as I can manage to empirical case histories which can act as a check on where the analysis needs correcting and complexifying. Epistemology, on this view, is not the bogey man it has been represented as being by those who suspect that it is merely a covert way of denying others’ claims to know. The multidimensionality of what is there to be known cannot help but generate a multidimensionality of ways of knowing.

But several possibly fundamental objections to my use of those notions of multidimensionality and of semantic stretch must now be met in conclusion. First it might be argued that I am still trapped in my own particular conceptual framework, even that everyone always is. Is it not the case that others’ categories either will be reduced to mine, or will forever remain beyond reach? Of course my particular conceptual resources are whatever they are at any moment in time (though I can see they have changed over time, as I said). But my answer to that first question would be to reject the alternative. Provided I am indeed allowed the point that conceptual frameworks (like languages themselves) are revisable, reductionism can be avoided.

Then a second objection might be that I am somehow presupposing some transcendent metalanguage into which all others can be parsed, which surely savors of Western hegemonic pretensions: give me a place to stand and I can move the world. Again I would resist, for I would claim both that I allow a voice to each pluralist rendering of multidimensionality, and that I recognize the revisability, indeed the imperfections, of my own understanding. A single metalanguage, let alone one that imposes a rule of strict univocity, is a chimera; rather, we can and should exploit the full resources of every understanding to which we can have access, and that will include those expressed in actions rather than words.14

I am reminded of a Chinese story in Zhuangzi (13: see Graham 1989: 187) about carpenter Bian. He had the cheek to reprimand his employer, Duke Huan, for reading old books. The Duke protests: “What business is it of a carpenter to criticize what I read? If you can explain yourself, well and good, but if not, you die.” Whereupon Bian refers to his own work as a carpenter. When he makes a wheel, if he chips at it too slowly, the chisel slips and does not grip; but if he is too fast, the chisel jams and catches in the wood. The right way to do it is something he feels in his hands but he cannot put it into words. That is how one becomes an expert carpenter, and the books of the past are just “dregs.” In such a case the feeling is the understanding.

However, even if we reject the notion of a metalanguage, the problem does not go away, indeed it might be thought aggravated by that rejection. How, I asked, is any translation across ontologies to be achieved? It is all very well, some will say, to insist that reality is multidimensional and that every term exhibits some degree of semantic stretch, but how does that bridge the gulf between divergent perspectives? If we reject the realist option, according to which only one ontology is correct, and the others must be dismissed as mistaken, are we not forced to the relativist view, according to which not only should we say all are correct, but also that there is no neutral way of judging them?

Once again that dichotomy obstructs a resolution. The multidimensional move would have it first that each ontology deals with a particular manner of ordering reality, including the key considerations of interiority and physicality. Secondly each may do so in a more or less satisfactory, and certainly not arbitrary, way. But more or less satisfactory to whom? In the first instance, of course, to the actors themselves, to the Achuar and the Araweté and to all those others whose cosmologies have been explored by the ethnographers, allowing, as before, that some are more explicit and comprehensive than others. But it is up to observers, to us, to see what we can learn from the exercise of investigating their ways of being in the world. That means suspending disbelief and being prepared to revise just about everything we normally take for granted about those key concepts of person, agency, causation, space, time and, yes, certainly, nature. Clearly we must abandon the assumption that reality is a given, to which unmediated access is possible. Yet neither actors nor observers will settle for “anything goes.”

But then does that not amount to some wishy-washy politically correct liberal relativism? Not if we can still provide for the possibility of diagnosing error in ourselves and others, allowing that both of those are tricky—for different reasons, for we may be insufficiently self-critical in the first case and overcritical in the second. But evidently we are not infallible, no more was any given ancient Greek or Chinese thinker, no more are present-day Achuar or Araweté. But before we conclude that we or they have made a mistake, we have first to consider the complexities of interpretation that I have been talking about, and there may always be an element of doubt about how thorough that exploration has been. In many cases we may say it can never be complete. But in some simple ones it is certainly possible to draw a line and to reassure ourselves that the job has been adequately done at least for the occasion in question. We have to get on with our everyday transactions, and indeed we do so.

Of course it is up to other researchers, not just ethnographers and ancient historians, but also linguists, cognitive scientists, evolutionary psychologists, to react to these proposals. I attempt no general theory of translation, of course, let alone a general formula for how understanding is possible. But I offer these thoughts as suggestions about what can reasonably be expected and about how one can go about the job. The very open-endedness of translation may look to be a menace to mutual intelligibility; but that is only so if we hanker after definitive results. Abandoning that will of the wisp, we can rather welcome that open-endedness as a positive resource for increasing understanding.


Berlin, Brent, and Paul Kay. 1969. Basic color terms: Their universality and evolution. Berkeley: University of California Press.

Carrithers, Michael, Steven Collins, and Steven Lukes, eds. 1985. The category of the person. Cambridge: Cambridge university Press.

Davidson, Donald. 2001. Essays on actions and events. Second edition. Oxford: Clarendon Press.

Delpla, Isabelle. 2001. Quine, Davidson: Le principe de charité. Paris: Presses Universitaires de France.

Descola, Philippe. (2005) 2013. Beyond nature and culture. Translated by Janet Lloyd. Chicago: University of Chicago Press.

Evans-Pritchard, Edward E. 1956. Nuer religion. Oxford: Clarendon Press.

Gladstone, William E. 1877. “The colour sense.” The Nineteenth Century 2: 360–88.

Graham, Angus C. 1989. Disputers of the Tao. La Salle, IL: Open Court.

Henare, Amiria, Martin Holbraad, and Sari Wastell, eds. 2007. Thinking through things: Theorizing artefacts ethnographically. London: Routledge.

Heywood, Paolo. 2012. “Anthropology and what there is: Reflections on ‘ontology.’” Cambridge Anthropology 30: 143–51.

Hocart, Arthur M. 1915. “Psychology and ethnology.” Folk-Lore 26: 115–37.

Holbraad, Martin. 2012. Truth in motion. Chicago: University of Chicago Press.

Holbraad, Martin, Morten Axel Pedersen, and Eduardo Viveiros de Castro. 2014. “The politics of ontology: Anthropological positions.” Fieldsights—Theorizing the Contemporary, Cultural Anthropology Online._January 13. http://culanth.org./ fieldsights/ 462-the-politics-of-ontology-anthropological-positions.

Ingold, Tim. 2000. The perception of the environment. London: Routledge.

Kuhn, Thomas S. (1962) 1970. The structure of scientific revolutions. Revised edition. Chicago: University of Chicago Press.

Laidlaw, James. 2012. “Ontologically challenged.” Anthropology of This Century 4. http:// aotcpress.com/articles/ontologically-challenged.

Laidlaw, James, and Paolo Heywood. 2012. “One more turn and you’re there.” Anthropology of This Century 5. http://aotcpress.com/articles/turn.

Latour, Bruno. (2012) 2013. An inquiry into modes of existence. Translated by Catherine Porter. Cambridge, MA: Harvard University Press.

Lloyd, Geoffrey E. R. 1970. Early Greek science: Thales to Aristotle. London: Chatto & Windus.

———. 1990. Demystifying mentalities. Cambridge: Cambridge University Press.

———. 1991. Methods and problems in Greek science. Cambridge: Cambridge University Press.

———. 2002. The ambitions of curiosity. Cambridge: Cambridge University Press.

———. 2004. Ancient worlds, modern reflections. Oxford: Clarendon Press.

———. 2007. Cognitive variations. Oxford: Oxford University Press.

———. 2012. Being, humanity, and understanding. Oxford: Oxford University Press.

Marett, Robert R. 1912. “The study of magico-religious facts.” In Notes and queries on anthropology, edited by Barbara W. Freire-Marreco and John L. Myres, 251–60. Fourth edition. London: Royal Anthropological Institute.

Mauss, Marcel. 1938. “Une catégorie de l’esprit humain: La notion de personne, celle de ‘moi.’ Un plan de travail.” Journal of the Royal Anthropological Institute 68: 263–81.

Mosko, Mark. 2010. “Partible penitents: Dividual personhood and Christian practice in Melanesia and the West.” Journal of the Royal Anthropological Institute (N.S.) 16: 215–40.

Pedersen, Morten A. 2011. Not quite shamans: Spirit worlds and political lives in northern Mongolia. Ithaca, NY: Cornell University Press.

———. 2012. “Common nonsense: A review of certain recent reviews of the ‘ontological turn.’” Anthropology of This Century 5. http://aotcpress.com/articles/common_nonsense.

Quine, Willard van Orman. 1960. Word and object. Cambridge, MA: MIT Press.

Rivers, William H. R. 1912. “The primitive conception of death.” Hibbert Journal 10: 393–407.

Schaffer, Simon. 2010. “Opposition is true friendship.” Interdisciplinary Science Reviews 35: 277–90.

Severi, Carlo. 2013. “Philosophies without ontology.” HAU: Journal of Ethnographic Theory 3: 192–96.

Sperber, Dan. (1974) 1975. Rethinking symbolism. Translated by Alice L. Morton. Cambridge: Cambridge University Press.

Stocking, George W. 1996. After Tylor: British social anthropology 1888–1951. London: Athlone.

Strathern, Marilyn. 1988. The gender of the gift. Berkeley: University of California Press.

———. 1999. Property, substance and effect: Anthropological essays on persons and things. London: Athlone.

———. 2005. Kinship, law and the unexpected. Cambridge: Cambridge University Press.

Taylor, Anne-Christine. 2013. “Distinguishing ontologies.” HAU: Journal of Ethnographic Theory 3: 201–4.

Turner, Terence S. 2009. “The crisis of late structuralism. Perspectivism and animism: Rethinking culture, nature, spirit and bodiliness.” Tipití: Journal of the Society for the Anthropology of Lowland South America 7: 1–42.

Vilaça, Aparecida. 2011. “Dividuality in Amazonia: God, the Devil and the constitution of personhood in Wari’ Christianity.” Journal of the Royal Anthropological Institute (N.S.) 17: 243–62.

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

———. (2009) Forthcoming. Cannibal metaphysics. Translated by Peter Skafish. Minneapolis: Univocal.

———. 2010. “In some sense.” Interdisciplinary Science Reviews 3–4: 318–33.

Wagner, Roy. 1991. “The fractal person.” In Big men and great men, edited by Maurice Godelier and Marilyn Strathern, 159–73. Cambridge: Cambridge University Press.

Wittgenstein, Ludwig. 1963. Philosophical investigations. Translated by G. E. M. Anscombe. Reprinted second edition. Oxford: Blackwell.

De la possibilité d’une intelligibilité mutuelle

Résumé : Cet article s’appuie sur le constat que les problèmes concernant la traduction et la possibilité de traduire se posent fréquemment, notamment en ce qui concerne la possibilité d’une intelligibilité mutuelle dans des registres différents, y compris au sein d’un même langage naturel. Les anthropologues et les historiens de l’antiquité ont fréquemment été confrontés à de tels problèmes. Les historiens le sont tout particulièrement lorsque ils se trouvent dans l’incapacité de vérifier leurs interprétations avec ceux qu’ils tentent de comprendre. Mais face à des déclarations, des croyances ou des pratiques apparemment paradoxales, nous devons et pouvons éviter le dilemme apparent (ou bien ces faits doivent être restitués et/ou réduits dans nos termes, ou bien nous devons admettre qu’ils sont strictement incompréhensibles) en insistant sur le caractère modifiable de notre cadre conceptuel, en particulier en relation à des notions cruciales tels que la subjectivité individuelle (personhood), Y agency, la causation, et la nature. Au lieu d’insister sur la dichotomie entre le sens littéral et le sens métaphorique, nous devons admettre que les termes dont nous nous servons possèdent ce que nous appelons ici une élasticité sémantique (semantic stretch). De plus, si nous acceptons (comme nous le faisons dans cet article) que le phénomène ou les réalités décrits sont multidimensionnels, il apparaît clairement que la perspective d’une seule et définitive traduction est un mirage. Louverture de la traduction n’est pas une menace pour l’intelligibilité mais sa condition préalable.

Geoffrey E. R. LLOYD is Professor Emeritus of Ancient Philosophy and Science at the University of Cambridge and is now based at the Needham Research Institute. He has published twenty-six books, initially concentrating on ancient Greek philosophy and science and then embarking on detailed comparative studies with ancient Chinese thought. His three most recent books, which tackle the underlying philosophical problems of such comparisons and related issues in cognitive science, are: Cognitive variations: Reflections on the unity and diversity of the human mind (2007), Disciplines in the making (2009), Being, humanity and understanding (2012), and The ideals of inquiry (2014), all from Oxford University Press.

Geoffrey E. R. Lloyd
University of Cambridge
Senior Scholar in Residence
Needham Research Institute
8 Sylvester Road
Cambridge, CB3 9AF


1. The opacity of those last three terms was discussed in Lloyd (2007: ch. 4).

2. I discussed the interpretation of this first line of the Daodejing in Lloyd (2002: ch. 5).

3. I analyzed the historical background to the introduction of the literal/metaphorical dichotomy in Lloyd (1990: ch. 1), and proposed “semantic stretch” as an alternative in Lloyd (2002: ch. 5). I introduced “multidimensionality” in my sense in Lloyd (2004: ch. 7).

4. For a recent review of the different ways in which the principle of charity has been taken, see Delpla (2001).

5. Thus the concept of a person, which was already problematized by Mauss (1938), has in recent years become even more of a field for contending interpretations (see, e.g., Carrithers, Collins, and Lukes 1985). We have been introduced to the notion of “fractal persons” and to the view that persons are not individuals but, in the phrase made famous by Marilyn Strathern, “dividuals,” divisible into multiple components formed from relations with others and subject to constant disequilibrium (e.g., Strathern 1988, 1999, 2005; Wagner 1991; Mosko 2010; Vilaça 2011).

6. Latour’s ([2012] 2013) monograph, stimulated in part by his reflections on the impasse of modernity, introduces further considerable possibilities for the exploration of multiple ontologies. The recent literature on the ontological turn comprises notable contributions, including some revisions of previously held views, by Pedersen (2011, 2012), Heywood (2012), Holbraad (2012), Laidlaw (2012), Laidlaw and Heywood (2012), and Holbraad, Pedersen, and Viveiros de Castro (2014).

7. The trope of indigenous peoples doing anthropology on anthropologists goes back to Rivers (1912), and was taken up by Hocart (1915) in his criticisms of Marett (1912). Cf. Stocking (1996: 236) and Schaffer (2010: 286–87).

8. My original proposal that “nature” was invented (not discovered) by the ancient Greeks was first published in my Methods and problems in Greek science (Lloyd 1991: ch. 18), though in earlier work (Lloyd 1970) I had indeed talked of a discovery.

9. There is a hard-hitting debate between Mencius, Gaozi, and Xunzi on precisely that issue in the fourth and third centuries BCE (see Graham 1989: 117–23, 244–51).

10. Others tackling the problems from other perspectives might agree with this conclusion, though without depending, as I do, on a historical argument.

11. See Viveiros de Castro (2010), where he explored the “multidimensionality of incommensurability.”

12. Lloyd (2007: ch. 1) discusses the problem with some background information concerning earlier and ongoing controversies.

13. This was a point that I argued against Berlin and Kay (1969) in Lloyd (2007: ch. 1).

14. Again this point was emphatically made by Wittgenstein (1963) and has been taken up by many others, notably by Ingold (2000) and the contributors to Henare, Holbraad and Wastell (2007).