On Value

This work is licensed under the Creative Commons | © Louis Dumont. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 DOI: http://dx.doi.org/10.14318/hau3.1.028


On Value

The Radcliffe-Brown Lecture in Social Anthropology, 1980



I saw Radcliffe-Brown only once, in this very room. In my memory I can still see him today, though somewhat hazily, delivering the Huxley Memorial Lecture for 1951.2 I must have made it to London for the occasion, from Oxford where I was a new, if not that young, lecturer. As I listened to him, he seemed to have made one step in the direction of Lévi-Strauss, and I felt comforted in my recent structural allegiance. In fact it was only a limited, passing convergence.3

In those days I was busy learning a good deal from him, and from British anthropology at large, which had reached unprecedented heights partly under his influence. Yet I must confess that, for one whose imagination had been initially fired by Mauss’s genial humanism, Radcliffe-Brown’s constricted version of Durkheimian sociology was not very attractive.

[208] Today, one feels the need to insist, beyond all divergences, on continuity on one basic point. Reading his Natural Science of Society, one is struck by Radcliffe-Brown’s decided holism.4 Whatever the shortcomings of his concept of ‘system’, the point—should I say the importation?—was probably decisive in the development of anthropology in this country, and it made possible the dialogue with the predominant sociological tradition of the French.

There is relatively little about values in Radcliffe-Brown’s writings.5 Yet the expression was very much in the air in British departments of anthropology in the last years of his life. My impression was that it figured largely as a substitute for ‘ideas’, which stressed the relation to action and was therefore less unpalatable to the empiricist temper. No doubt the situation is quite different today. But to state plainly the reason for my use of the term, preferably in the singular, and for my choice of topic: I have been trying in recent years to sell the profession the idea of hierarchy, with little success, I may add. I thought of making another bid, this time by using the professionally received vocable, which I had instinctively shunned heretofore, I suppose because of the forbidding difficulties the term seems to present. May the attempt be taken as an effort to come closer to the Radcliffe-Brownian heritage.

In fact, my intention is solely to offer an observation bearing on the relation between ideas and values, or rather to comment on [209] that observation and draw some consequences from it. The modern type of culture in which anthropology is rooted and the non-modern type differ markedly with regard to value, and I hold that the anthropological problems relating to value require that the two be confronted. We shall start from the modern configuration, which represents an innovation, then introduce in contrast some fundamental features of the more common non-modern configuration, and finally return to the modern predicament with a view to setting it ‘in perspective’ and to thus, it is hoped, throwing some light on the position and task of anthropology as a mediating agency.

The modern scene is familiar. In the first place, modern consciousness attaches value predominantly to the individual, and philosophy deals, at any rate predominantly, with individual values while anthropology takes values as essentially social. Then, in common parlance, the word, which meant in Latin healthy vigour and strength and in medieval times the warrior’s bravery, symbolizes most of the time the power of money to measure everything. This important aspect will be present here only by implication. (Cf. Dumont (1977)).

As to the absolute sense of the term, the modern configuration is sui generis and value has become a major preoccupation. In a note in Lalande’s Philosophical Dictionary, Maurice Blondel said that the predominance of a philosophy of value characterizes the contemporary period, following a modern philosophy of knowledge and an antique and medieval philosophy of being.6 For Plato the supreme Being was the Good. There was no discord between the Good, the True and the Beautiful, yet the Good was supreme, perhaps because it is impossible to conceive the highest perfection as inactive and heartless, because the Good adds the dimension of action to that of contemplation. In contrast we moderns separate science aesthetics and morals. And the nature of our science is such that its existence by itself explains or rather implies the separation between the true on the one hand, the beautiful and the good on the other and in particular between being and moral value, what is and what ought to be. For the scientific discovery of the world was premised on the banning as secondary of all qualities to which physical measurement was not applicable. Thus for a hierarchical cosmos was substituted our physical, homogeneous universe (Koyré (1958)). The value dimension which had been spontaneously projected on to it was relegated to what is to us its proper locus, that is, man’s mind, emotions, and volition.

[210] In the course of centuries, the (social) Good was also relativized. There were as many Goods as there were peoples or cultures, not to speak of religions, sects or social classes. ‘Truth this side of the Pyrenees, error beyond’ noted Pascal; we cannot speak of the Good when what is held as good on this side of the Channel is evil on the other, but we can speak of the value or values that people acknowledge respectively on one and the other side.

Thus, value designates something different from being, and something which, while the scientifically true is universal, is eminently variable with the social environment, and even within a given society, according not only to social classes but to the diverse departments of activity or experience.

I have listed only a few salient features, but they are enough to evoke the complex nexus of meanings and preoccupations to which our word is attached, a tangle to which all kinds of thoughtful efforts have contributed, from the romantic complaint about a world that has fallen asunder to the various attempts at reuniting it, and to a philosophy of despair, Nietzsche’s, contributing to spreading the term. I do not think that anthropology can disregard this situation. Yet it is no wonder that there is something unpleasant about the term. Being comparative in essence, it seems doomed to emptiness: a matter of values is not a matter of fact. It advertises relativism, as it were, or rather both the centrality of the concept and its elusive quality, to which a considerable literature testifies. It smacks of euphemism or uneasiness, like ‘underdevelopment’, ‘methodological individualism’ and so many other items in the present-day vocabulary.

Yet there is a positive counterpart, modest but not insignificant, for the anthropologist: we have at our disposal a word that allows us to consider all sorts of cultures and the most diverse estimations of the good without imposing on them our own: we can speak of our values and their values while we could not speak of our good and their good. Thus the little word, used far beyond the confines of anthropology, implies an anthropological perspective and invests us, I think, with a responsibility. But of this more later.

We begin with a few introductory remarks about the study of values in anthropology. The prevailing use of the word in the plural—values—is indicative not only of the diversity of societies and of the modern compartmentalization of activities but also of a tendency to atomize each configuration that is in keeping with our culture in general. This is certainly the first point that requires attention. In a paper published in 1961,7 Francis Hsu criticized [211] some studies of the American character for their presenting a bare catalogue of traits or values without bothering about the relations prevailing between those items. He saw conflicts and inconsistencies between the different values listed, wondered at the lack of serious attempts to explain them, and proposed to remedy the situation by identifying one fundamental value and by showing that it implied precisely the contradictions to be explained. The ‘American core value’, Hsu suggested, is self-reliance, itself a modification or intensification of European, or English, individualism. Now self-reliance implies contradiction in its application, for men are social beings and depend heavily on each other in actual fact. Thus is produced a series of contradictions between the level of conception and the level of operation of the main value and of the secondary values derived from it or allied with it.

I for one cannot but applaud both the search for a cardinal value and its identification in this case as some form of individualism. One also notes that Hsu implies, if he does not state it explicitly, a hierarchy between conception and operation. Yet in the end Hsu’s distinction between the two levels is still insufficient. He uses a classification of Charles Morris, who had listed three uses of value or sorts of value, among them conceived and operative value, and he goes some way toward ranking these two levels. Finally however, he speaks of ‘values’ for both, and thus lumps them together again in the8 same way as the atomizing authors he had begun by criticizing. In fact the two levels should be firmly distinguished. For we have here a universal phenomenon. Surely all of us have encountered this characteristic complementarity or reversal between levels of experience where what is true on the more conceptual level is reversed on the more empirical level, a reversal which bedevils our attempts at unifying, for the sake of simplicity, the representation and its counterpart in action. Whatever the peculiarities of the American case, the end cannot be its own means: either what is called ‘operative values’ are not values at all, or they are second-order values that should be clearly distinguished from first-order values or values proper.

In general, there is perhaps a surfeit of contradictions in contemporary literature in general. An author belonging to a different era or milieu will frequently be taxed of contradicting himself simply because a distinction of levels obvious to him and therefore implicit in his writings, but unfamiliar to the critic, is missed.9 It will be seen later on that where non-moderns [212] distinguish levels within a global view, the moderns know only of substituting one special plane of consideration for another and find on all planes the same forms of neat disjunction, contradiction, etc. Perhaps there is a confusion here between individual experience which, while crossing different levels, may be felt as contradictory, and sociological analysis, where the distinction of levels is imperative in order to avoid the short-circuit that results in tautology or incomprehension. Apart from Clyde Kluckhohn, the late Gregory Bateson is one of the rare anthropologists who clearly saw the necessity of recognizing a hierarchy of levels.10

There has been in the history of anthropology at least one sustained attempt at advancing the study of values. In the late forties, Clyde Kluckhohn chose to focus attention on values and to concentrate efforts and resources on a vast co-operative long-term project devoted to their study, the Harvard ‘Comparative Study of Values in Five Cultures’. There seems to have been in the United States, at the end of World War II, a wide renewal of interest in social philosophy and in the understanding of foreign cultures and values.11 Kluckhohn may have found in the circumstances of the time the occasion to develop what was undoubtedly a deep personal concern. In the late forties, he launched his project, which assembled a number of scholars and issued in an imposing array of publications over the next decade. Today this considerable effort seems largely forgotten. Unless I am deeply

[213] mistaken, it has not left a deep mark in American cultural anthropology. Is this one more example of those fashions that disconcertingly displace one another in our discipline, particularly in the US; or are there internal reasons to the discredit, and, in the worst case, are values a mistaken focus or a ‘non-subject’, something I could hardly believe? I am not able to answer this complex question. I shall only try to draw from Kluckhohn’s endeavour a lesson for our benefit. There must be such a lesson if we believe with him that values are a central problem. For Kluckhohn was not naïve: he was obviously a man of wide culture (with a German component, I suppose, as is the case with several of the early American anthropologists), and, moreover, he anticipated much of what I shall have to say here. Yet, whatever contributions the project may have brought to the knowledge of each of the particular groups or societies studied, the results seem disappointing as regards Kluckhohn’s main aim, namely the advancement of comparative theory. How can we account for the fact?

Clyde Kluckhohn was closely associated with Parsons and Shils in the symposium that was published as Towards a General Theory of Action and to which he contributed an important theoretical essay which can be taken as the chart of the Harvard Project.12 It is clear that Kluckhohn developed his own position while agreeing on the broad ‘conceptual scheme’ of the symposium. He dissented only from the rigid separation between social and cultural systems.13 To be brief I shall mention only three main points in Kluckhohn and two of his main associates. First, that (social) values are essential for the integration and permanence of the social body and also of the personality (p. 319)—we might say with Hans Mol for their identity14—is perhaps obvious, but it is in practice too easily forgotten, either by anthropologists insisting unilaterally on change, or by philosophers abstracting individual values from their social background. Saint Augustine said somewhere that a people is made up of men united in the love of something.

Second, the close link between ideas and values—here ‘cognitive’ and ‘normative’, or ‘existential’ and ‘normative’, aspects—is clearly acknowledged, as it was by Parsons and Shils (1951, pp. 159–89), under the central concept of ‘value-orientation’ [214] as defined by Kluckhohn (1951, pp. 410–11). (The concept is open to criticism on another score, as was shrewdly noticed by an anthropologist.)15 Thus the scheme for the classification of values used by Florence Kluckhohn includes by the side of values proper a minimum of ideas and beliefs. One may prefer the more ample treatment of the Navaho case by Ethel Albert, which includes not only the normally unverbalized ‘value-premises’, but also a complete picture of the worldview as the ‘philosophical context’ of the value system strictly defined.16

The third point is the clear recognition of the fact that values are ‘hierarchically organized’. Clyde Kluckhohn’s programmatic article had a very lucid and sensitive page on the question (p. 420), but it is perhaps Florence Kluckhohn that most developed this aspect. Early in the research, she proposed a grid for the comparison of ‘value-orientations’. It is a scheme of priorities distinguishing, in each instance under three terms, different stresses relative to relations between man and nature, to the conception of man, to relations between men, to time, and to action.17 The author underlines the importance of hierarchy and of nuances in hierarchy. Each value system is seen as a hierarchical combination sui generis of elements which are universal in the sense of being found everywhere. This was a solution to a problem that much concerned Clyde Kluckhohn himself. He was reacting against an excessive stress on relativity in anthropological literature. He wanted to avoid falling into (absolute) relativism, and he tried to salvage a modicum of universal values.18 Florence Kluckhohn found this universal basis in the very material which was elaborated into different value systems, in each case, by an original combination of particular value emphases.

Let me briefly articulate a double criticism. The scheme does [215] not yet apply broadly enough the recognition of hierarchy, and therefore gets stuck in a measure of atomism: no relation is posited between the five subdivisions. What for instance about the relative stress between relations to nature and relations between men (items 1 and 3)? A universal basis seems here to be unduly assumed. The scheme remains thus inevitably sociocentric. Indeed, it is actually centered on a White American and even a Puritan model. Other cultures may make different choices, but only in terms derived from the American choices.

A later text by Clyde Kluckhohn adds a new scheme of classification of his own to a presentation of those of Ethel Albert and Florence Kluckhohn. The paper,19 apparently Kluckhohn’s last word on the question, would deserve longer consideration than can be given to it here, less for the scheme itself than for the preoccupations that lead up to it. The general, universal bearing of the project is stressed, while the provisional character of the particular scheme is granted. The effort is to make the scheme purely relational: it consists of a series of qualitative, binary oppositions. What is more, it is supplemented by an effort to bring out, by tabulation, the associations between features and thus to reconstitute to some extent the systems analysed.

How is it, then, that a considerable effort containing so many correct perceptions leaves one finally unsatisfied? We are left, on the abstract side, with grids into the pigeon-holes of which we should be able to distribute the elements of any value system. It is clear that, notwithstanding Clyde Kluckhohn’s last and pathetic effort to affirm a structural, or structuralist, approach and to recapture the living unity given at the start, the whole has vanished into its parts. Atomization has won the day. Why? Because, I submit, the attempt has been unwarily to unite fire and water, structure, hierarchical structure, and classification, that is, classification through individual features. The need for classification was certainly reinforced by the attempt to compare five cultures in one compass, and the most valuable products of the project are probably the monographic pictures in the manner of Albert that it produced. A somewhat unpalatable conclusion follows, namely that a solid and thorough comparison of values is possible only between two systems taken as wholes. If classification is to be introduced further on, it will have to start from wholes and not from itemized features. For the time being we are closer to

[216] Evans-Pritchard’s ‘historiography’ than to Radcliffe-Brown’s ‘natural science of society’.

Kluckhohn noticed that the term ‘value’, chiefly used in the plural, had come recent into the social sciences from philosophy. He saw in it a kind of interdisciplinary concept20 and, probably largely for this reason, mingled occasionally individual and group values. The term ‘value-orientation’ itself is indicative of a commanding concern with the individual actor (see p. 214 n.1 above). Of course, all this tallies with a behavioural approach, but it is above all an index of the philosophical background of our anthropological problems. The philosophical debate is of intimidating dimension and complexity. Yet we cannot possibly leave it out entirely in an attempt to clarify the anthropological question. Fortunately, I believe that, conversely, an anthropological perspective can throw some light on the philosophical debate and that it is thus possible to take a summary and yet not ineffectual view of it.

There are two kinds of philosophers, or rather two kinds of philosophizing in the matter. One locates itself within modern culture and is careful to work in accordance with its constraints, its basic inspiration, its inner logic and its incompatibilities. From that point of view the conclusion follows that it is impossible to deduce what ought to be from what is. No transition is possible from facts to values. Judgements of fact and judgements of value are different in kind. It is enough to recall two or three major aspects of modern culture to show that the conclusion is inescapable. First, science is paramount in our world, and, as we recalled at the start, to make scientific knowledge possible the definition of being has been altered by excluding from it precisely the value dimension. Second, the stress on the individual has led to internalizing morality, to finding it exclusively within the individual’s conscience while it is severed from the other ends of action and distinguished from religion. Individualism and the concomitant separation between man and nature have thus split the good, the true and the beautiful, and have produced a theoretically unbridgeable chasm between is and ought to be. This situation is our lot in the sense that it lies at the core of modern culture or civilization.

Now, whether this situation is comfortable or reasonable is quite another question. The history of thought seems to show that it is not, for no sooner had Kant proclaimed this fundamental split that his gifted successors, and the German intelligentsia as a [217] whole, hastened in various attempts to re-establish unity. It is true that the social milieu was historically backward and that German intellectuals, while inspired by individualism, were still imbued with holism in the depth of their being. But the protest has continued down to the present day. It must be admitted that, for one who turns away from the environment and attempts to reason from first principles, the idea that what man ought to do is, let us say, unrelated to the nature of things, to the universe and to his place in it, will appear queer, aberrant, incomprehensible. The same holds true of someone who would take into account what we know of other civilizations or cultures. I have said elsewhere that most societies have believed themselves to be based in the order of things, natural as well as social; they have thought they were copying or designing their very conventions after the principles of life and the world. Modern society wants to be ‘rational’, to break away from nature and set up an autonomous human order.21 We may thus be inclined at first flush to sympathize with those philosophers who have tried to restore unity between facts and values. Their attempts testify to the fact that we have not entirely broken away from the more common mould of mankind, that it is still in some manner present with us, underlying and perhaps modifying the yet compelling modern framework. But we should be on our guard…

The attempt can take different forms. One consists in annihilating values entirely. Either value judgements are declared meaningless, or the expression of mere whims or emotional states. Or, with some pragmatists, ends are reduced to means: having construed a category of ‘instrumental values’, they proceed to deny the distinct existence of ‘intrinsic values’ that is, of values proper.22 Such attempts seem to be an index of the inability of some philosophical tendencies to take into account real human life, to mark a dead end of individualism. Another type can be taken as a desperate attempt at transcending individualism by resorting to a modern ersatz of religion. In its Marxist form, and through it and somewhat similarly in totalitarian ideologies in general, this doctrine has proved fateful; it is sometimes regarded as sinister, at least in continental Europe, and rightly so. Here we must firmly side with Kolakowski in his impassioned [218] condemnation of the trend, as against certain rambling intellectuals.23

We follow Kolakowski especially on one point; the danger does not arise only from the violent attempt to implement such doctrines, but is contained in the doctrine itself under the form of value incompatibilities that call for violence on the level of action. To confirm this point: in an article of 1922, which in retrospect appears prophetic of later developments in Germany, Karl Pribram has noted the parallel incongruity and structural similarity of Prussian nationalism and Marxist socialism. Both, Pribram pointed out, jumped from an individualistic basis to an illegitimate, holistic (‘universalistic’) construct, the State in the one case, the proletarian class in the other, which they endowed with qualities incompatible with their presuppositions.24 Totalitarianism is present in germ in such encounters. Philosophers themselves are not always sensitive to such incompatibilities, but their constructions are seldom applied to society.25 Here a question arises: it is convenient to link totalitarianism with such incompatibilities—yet there exist incompatibilities in societies without their developing into that scourge. Tönnies insisted that both Gemeinschaft and Gesellschaft are present as principles in modern society. My provisional answer is that they are found on different levels of social life, while it is characteristic of modern artificialism to disregard such levels altogether and thus to make for collision between what it consciously introduces and the substratum which it does not really know. There may well be, indeed there actually is, a need for reintroducing some measure of holism into our individualistic societies, but it can be done only on clearly articulated subordinate levels, so that major clashes with the predominant or primary value are prevented. It can be done, that is, at the price of introducing a highly complex hierarchical articulation, which can be imagined, mutatis mutandis, as a parallel to the highly elaborate Chinese etiquette.26 This point will become clearer in what follows. At any rate, we should in the first place, as citizens of the world and of a particular state within it, abide with Kolakowski by the Kantian distinction as an integral part of the modern make-up.

[219] Now what are the consequences of the distinction for social science? Let us take as vanished the times when a behavioural science banned the study of social values together with that of conscious representations at large. We do study social representations as social facts of a particular kind. Here two remarks are called for. First, it is clear that we maintain this ‘value-free’ attitude on the basis of the Kantian distinction, for otherwise our own native view of ‘facts’ would command value judgements and we should remain locked up in our own system, sociocentric as all societies are except, in principle, our own. The point simply confirms the link between science in general and the is/ought separation. But then our approach is philosophically questionable. It may be argued that we should distinguish tyranny from legitimate rule. Leo Strauss maintained against Max Weber that social science could not escape evaluation,27 and it is true that Weber was led by this ‘value-free’ stand to undesirable admissions, such as his ‘ethic of conviction’. More radically, one may contend that values cannot really be understood without our adhering to them (note the proximity to the Marxist plea), and that to relativize values is to kill them. In a discussion, A. K. Saran maintained the thesis in its full consequence.28 According to this view, cultures cannot communicate, which means cultural solipsism, a return to sociocentrism. And yet, there is point in it in the sense that comparison implies a universal basis: it must appear in the end that cultures are not as independent from each other as they would claim to be and as their internal consistency seems to warrant.

Stated otherwise, our problem is: how can we build a bridge between our modern ideology that separates values and ‘facts’ and other ideologies that embed values in their worldview?29 Lest our quest should appear futile, let us not forget that the problem is in a way present in the world as it is. Cultures are in fact interacting, thus communicating in some mediocre manner. It behoves anthropology to give a conscious form to that groping and thus to answer a contemporary need. We are committed to reducing the distance between our two cases, to reintegrating the modern case within the general one. For the moment, we shall try to formulate more precisely and thoroughly the relation between them.

[220] Values are in general intimately combined with other, non-normative representations. A ‘system of values’ is thus an abstraction from a wider system of ideas-and-values.30 This is true not only of non-modern societies, but also of modern societies, with one cardinal exception, namely that of (individual) moral values in their relation to scientific, ‘objective’ knowledge. For all that we said previously about ought bears exclusively on individual, ‘subjective’ morality. That this morality is, together with science, paramount in our modern consciousness does not hinder its cohabiting with other norms, or values of the common sort, namely traditional social ethics, even if some transition, some substitution of the former for the latter is taking place under our eyes. Thus the modern value of equality has spread in the last decades in European countries to domains where traditional ethics were still in force; from the French Revolution, in whose values it was implied, up to our days, the equality of women had not imposed itself against subordination as entailed by a whole nexus of institutions and representations. Now the struggle between the two ‘systems of values’ has intensified, and the outcome has still to be seen: our individualist values are at loggerheads with the considerable inertia of a battered social system that is gradually losing its own justification in consciousness.

A convenient example of the inseparability of ideas and values is found in the distinction between right and left. It is widespread, if not universal, and is still found with us in some manner, although our attitude to it is highly consonant with modern ideology. We are in the habit of analysing it into two components. We see it essentially as a symmetrical opposition, where the two poles have equal status. The fact that the two poles are unequally valued, that the right hand is felt to be superior to the left hand, appears to us an arbitrary, superadded feature, which we are at pains to explain. Such was the frame of mind of Robert Hertz when he wrote his classical essay, and it has prevailed ever since. It is wholly mistaken. As I argued elsewhere, the reference to the body as to a whole to which right and left hands belong is constitutive of the right, the left and their distinction.31 The contention should be obvious: take a polar opposition at random, add to it a difference [221] in value, and you will not get right and left. Right and left, having a different relation to the body (a right relation and a left relation, so to speak) are different in themselves. (They are not two identical entities situated in different places, as we know pretty well from sensuous experience). Being different parts of a whole, right and left differ in value as well as in nature, for the relation between part and whole is hierarchical, and a different relation means here a different place in the hierarchy. Thus the hands and their tasks or functions are at one and the same time different and ranked.32

There is something exemplary about this right-left relation. It is perhaps the best example of a concrete relation indissolubly linked to human life through the senses, which physical sciences have neglected and which anthropology may presumably retrieve or rehabilitate. I believe it teaches us in the first place that to say ‘concrete’ is to say ‘imbued with value’. That is not all, for such a difference in value is at the same time situational, and the point will require attention. The fact is that, if certain functions are allotted to the left hand, then, in relation to their performance, the right hand will come second notwithstanding its being on the whole superior.

The right-and-left pair is indissolubly both an idea and a value, it is a value-idea or an idea-value. Thus at least some of the values of any given people are enmeshed in that people’s conceptions. To discover them, it is not necessary to go about eliciting people’s choices. These values have nothing to do with the preferable or the desirable—except in that they suppose that the naïve perception of the relation between whole and parts, that is, of order as given in experience, has not been obliterated. The moderns tend to define value in relation to arbitrary will, Tönnies’s Kürwille, while we are here in the realm of Naturwille or natural, spontaneous will. The whole is not, strictly speaking, preferable to its parts, it is simply superior to them. Is the right ‘preferable’ to the left? It is only apposite in some circumstances. What is ‘desirable’, if one insists, is to act in accordance with the nature of things. As to the modern tendency to confuse hierarchy with power, who will pretend that the right hand has power over the left? Even its pre-eminence is, on the level of action, limited to the accomplishment of its proper functions.

[222] The case also gives us a clue as to how we moderns manage to avoid the ranked nature of things, for we have not ceased to possess a right and a left hand and to deal with our body and with wholes in general. Not only have we developed permissivity in the matter in accordance with our de-valuation of the hands and with our individualism. We also tend to decompose the original relation by separating value from idea, and in general from fact, which means separating ideas and facts from the whole(s) in which they are actually to be found. Rather than relating the level under consideration—right and left—to the upper level, that of the body, we restrict our attention to one level at a time, we suppress subordination by pulling apart its elements. This shunning of subordination, or, to call it by its true name, of transcendence, substitutes a flat view for a view in depth, and at the same time it is the root of the ‘atomization’ so often complained of by romantic or nostalgic critics of modernity. The point holds in general: in modern ideology, the previous hierarchical universe has fanned out into a collection of flat views of this kind. But I am anticipating.33

[223] In the non-modern view that I here tried to retrieve, the value of the right or the left hand is rooted in their relation to the body, i.e. to a higher level of being: the value of an entity is dependent upon or intimately related to a hierarchy of levels of experience in which that entity is situated. Here is perhaps that main perception that the moderns miss, or ignore, or suppress without being fully conscious of so doing.34

The point has a bearing on the problem of evil. Two different conceptions are currently contrasted: for some, evil is only the absence or insufficiency of good, vice the limit or zero degree of virtue; for others evil is an independent principle pitched against its opposite as the will of Satan defying that of God.35 Yet, if Leibniz’s Theodicy is compared with Voltaire’s discussion of the Lisbon earthquake, one senses a contrast of a perhaps different nature. Let me interpret freely. For Leibniz, the fact that there is evil locally, here and there, in the world does not prevent the world from being, globally considered, the best of all possible worlds. Voltaire concentrates on a massive example of evil and refuses to look elsewhere or beyond; or rather he simply cannot. Voltaire will not ask himself what are the conditions for a real world to exist.

[224] He might well say that it is a question beyond the reach of human reason. For Leibniz36 good and evil are interdependent to begin with, the one inconceivable without the other. But that is not enough, for surely they are no more equal than are right and left. If I may make use of the definition I proposed of the hierarchical opposition, good must contain evil while still being its contrary.

In other words, real perfection is not the absence of evil but its perfect subordination. A world without evil could not possibly be good. Of course we are free to call this a universe of faith as against a universe of common sense, of modern common sense. But it is also a universe of rich concreteness as against one of desiccated principle. More precisely, a universe thick with the different dimensions of concrete life, where they have not yet come apart. The different dimensions of life do of course exist for Voltaire, but his thought sorts them out, it cannot embrace them all at once. And no doubt we live in Voltaire’s world, and not in Leibniz’s. It is just a matter of advancing in the perception of the relation between them.

Now let us suppose that, enlightened by the right-and-left example, we agree not to separate an idea and its value but to consider instead as our object the configuration formed by idea-values or value-ideas. It may be objected that such complex entities will be difficult to handle. Can we really come to grips with such multidimensional objects in their interrelations? Certainly the task is not easy, as it goes against our most ingrained habits. Yet, we are not entirely deprived of clues to make a beginning. We start with three remarks. First the configuration is sui generis, idea-values are ranked in a particular fashion. Second, this ranking includes reversal as one of its properties. Third, the configuration is thus normally segmented. I shall comment in turn on these three characteristics.

First about ranking. ‘High’ ideas will both contradict and include ‘low’ ideas. I called this peculiar relation ‘encompass-[225]ment’. An idea that grows in importance and status acquires the property of encompassing its contrary. Thus I found that in India purity encompasses power. Or, to take an example closer to us, from those that came up in the course of studying economic ideas: economists speak of ‘goods and services’ as one overarching category comprising, on the one hand, commodities and on the other, something quite different from commodities but assimilated to them, namely services.37 This is incidentally an example of relations between men (services) being subordinated to relations to things (goods), and if we were to study, say, a Melanesian system of exchanges, it would come nearer to the mark to reverse the priority and speak of prestations and goods, I mean prestations (relations between men) including things or encompassing their contrary, things.

We have already alluded to the second characteristic, reversal. The logical relationship between priest and king, as found in India or, nearer to us, in Christianity itself, five centuries after Christ, under the pen of Pope Gelasius, is exemplary in this regard. In matters of religion, and hence absolutely, the priest is superior to the king or emperor to whom public order is entrusted. But ipso facto the priest will obey the king in matters of public order, that is, in subordinate matters.38 This chiasmus is characteristic of hierarchy of the articulate type. It is obscured only when the superior pole of the hierarchical opposition is coterminous with the whole and the inferior pole is determined solely in relation to the former, as in the instance of Adam and Eve, Eve being created from a part of Adam’s body. What happens here is that it is only on the empirical level—and thus not within the ideology proper—that a reversal can be detected, as when the mother comes to dominate in fact the family in which she is in principle subordinate to her husband. The reversal is built-in: the moment the second function is defined, it entails the reversal for the situations belonging to it. That is to say, hierarchy is bidimensional, it bears not only on the entities considered but also on the corresponding situations, and this bidimensionality entails the reversal. As a consequence, it is not enough here to speak of different ‘contexts’ as distinguished by us, for they are foreseen, inscribed or implied in the ideology itself. We must speak of different ‘levels’ hierarchized together with the corresponding entities.

In the third place, values are often segmented or rather, I [226] should say, value is normally segmented in its application, except in specifically modern representations. I shall give a few examples of a striking contrast between non-modern and modern cultures, which bears on the way distinctions are organized or configurated. Impressionistically, on one side, as I said of India, distinctions are numerous, fluid, flexible, running independently of each other, overlapping or intersecting; they are also variably stressed according to the situation at hand, now coming to the fore and now receding. On the other side, we think mostly in black and white, extending over a wide range clear either/or disjunctions and using a small number of rigid, thick boundaries defining solid entities.39 It is noteworthy that the same contrast was recently found between early Christianity and the late Middle Ages in political theology. According to Gerard Caspary, the ‘slow growth of scholastic and legal modes of thinking’, emphasizing ‘clarity and distinctions rather than interrelationships’ has disembedded the political dimension while the ‘multifaceted and transparent symbols … have become one-dimensional and opaque emblems’.40

A similar contrast has been pointed out in modern psychology by Erik Erikson. Discussing the adolescent’s identity formation he contrasts two possible outcomes of the process, which he calls ‘wholeness’ and ‘totality’, as two different forms or patterns of ‘entireness’:

As a Gestalt, then, wholeness emphasizes a sound, organic, progressive mutuality between diversified functions and parts within an entirety, the boundaries of which are open and fluent. [Note the plural!] Totality, on the contrary, evokes a Gestalt in which an absolute boundary is emphasized; given a certain arbitrary delineation, nothing that belongs inside must be left outside, nothing that must be outside can be tolerated inside. A totality is as absolutely inclusive as it is utterly exclusive: whether or not the category-to-be-made-absolute is a logical one, or whether the parts really have, so to speak, a yearning for one another.41

We cannot at this point follow any further Erikson’s fine discussion. We retain essentially the perception of two conceptions or definitions of a whole, one through a rigid boundary, the other through internal interdependence and consistency. From our [227] point of view, the former is modern and arbitrary or somewhat mechanical, the second traditional and structural.42

It should be clear that such contrasts between segmented and unsegmented representations have not taken us away from values. In the first approximation the opposition is between holistic values in the former and individualistic values in the latter.

I owe to Robert Bellah a superb reference to hierarchy in Shakespeare. In the third scene of Troilus and Cressida Ulysses pronounces a long eulogy of order as degree:

The heavens themselves, the planets, and this centre

Observe degree, priority, and place

Insisture, course, proportion, season, form,

Office and custom, in all line of order…

There is one egregious example of the segmentation of value. It is the representation of the universe as a linear hierarchy of beings that is called the Great Chain of Being. It was influential all through our history from neoplatonism to the nineteenth century, as was shown in the well-known book which Arthur Lovejoy devoted to it.43 It pictures the world as a continuous series of beings, from the greatest to the least. It combines, Lovejoy tells us, plenitude, continuity, and gradation. It is a kind of ladder with a secret. The rungs of the ladder are so multiplied that the distance between two successive rungs shades into insignificance and leaves no void; the discontinuity between different sorts of beings is thus seen as a continuity of Being as a whole. The hierarchical aspect is evident, yet it appears on reflection that Lovejoy did not do it full justice. As most moderns, he was unable to see the function of hierarchy in the scheme. He gave scant attention to the only treatise we have on hierarchy, that of the Pseudo-Dionysius Areopagitus, in fact a double treatise on celestial and on terrestrial hierarchy. Let us have a look at Dionysius’s definition:44

I mean by hierarchy a holy ordering, a knowledge, and an activity, which assimilates itself as closely as possible to the divine form, and [228] which raises itself to the imitation of God in proportion to the lights which God has granted to it; the beauty which is worthy of God, being simple and good and the principle of initiation, is on the one hand absolutely pure of any dissimilarity, and on the other grants to each one, according to his desert, a share of its own light, while it initiates each one in the most divine initiation, forming them to an harmonious and indistinguishable likeness of itself. The aim of hierarchy is therefore the attainment, as far as possible, of likeness and of union with God.

It is worth stressing that in Dionysius the emphasis throughout is on communication if not on mobility (at least not in our sense of the term). The angels and other creatures situated between men and God are there to transmit or relay the word of God which men could not otherwise perceive, as well as to pave the way, as it were, for the ascent of the soul.45

It is not enough, then, to speak of a transformation of discontinuity into continuity. More widely and deeply, the Great Chain of Being appears as a form for acknowledging differences while at the same time subordinating them to and encompassing them in unity.

Nothing can be more remote from this grand scheme than the American ‘color-bar’. Of course there is no homology, for the latter representation is limited to men (in accordance with the modern split between man and nature). Yet it is as characteristic of the modern as the Great Chain is of the traditional mode of thought. All men, instead of being divided into a number of estates, conditions or statuses as previously, in harmony with a hierarchical cosmos, are now equal, but for one discrimination. It is as if a number of distinctions had coalesced into one absolute, impassable boundary. Characteristic is the absence of the shades still found elsewhere or previously: no half-breeds, mulattoes or mestizos are recognized here: what is not pure white is black.

Clearly we reach here the perfect opposite of segmentation. The contrast is so decisive that one might as well speak of antisegmentation, and the similarity with the other examples adduced tends to show that this form is characteristic of modern ideology.

With ranking, reversal, and segmentation, we have gained [229] some insight in the common, non-modern, I am tempted to say ‘normal’, configuration of value. Such a configuration is part and parcel of the system of representations (ideas-and-values) which I call, for the sake of brevity, ideology. This type of configuration appears very different from the modern type: more precisely, granted that it is not completely absent from modern society, but survives in it in parts and in some degree, it is a fact that modern ideology itself is of quite different type, is indeed as exceptional as Polanyi said of an aspect of it. Now, as we have seen, science has a predominant place and role in modern ideology. It follows that modern scientific, and to a large extent philosophic, ideas, linked as they are with the modern system of values, are often ill-fitted for anthropological study and sociological comparison. Actually it follows from the connection between ideas and values that, just as we must be ‘value-free’ in our ‘laboratory’, we should in principle at the same time be wary of applying our own ideas, especially our most habitual and fundamental ideas to our subject-matter. To do this is of course difficult, and at the limit impossible, for we cannot work ‘idea-free’. Actually we are caught between the Scylla of sociocentrism and the Charybdis of obscurity and incommunicability. All our basic intellectual tools cannot be replaced or modified at one stroke. We have to work piecemeal, and that is what anthropology has done, as its history shows. The reluctance one feels against putting oneself in question— for in the end this is precisely what the effort amounts to—inclines us to do too little, while self-aggrandizement to the neglect of the scientific community counsels to do too much.

Regarding our use of a given concept, it might be of some help to get a clearer view of its place among modern values. I shall take an example. Clearly the absolute distinction between subject and object is fundamental for us and we tend to apply it everywhere, even unknowingly. Its link with some of the ideas already mentioned is obvious, and it clearly bears a value stress.46 At the same time, it has a bearing on a contemporary problem. We badly need a theory of exchanges, for they enshrine a good deal of the essence of certain societies, as in Melanesia. Now, judging from recent literature, we seem condemned either to subordinate exchanges to the social morphology, or the reverse. The two [230] domains or aspects collide and we have no means of subsuming them under a unified framework. Have we not here a case where our absolute subject/object distinction obtrudes? When Lévy-Bruhl spoke of ‘participation’ between men and objects, was he not trying to circumvent the distinction? Mauss’s Essay on Gift [sic], so celebrated nowadays, is largely busy acknowledging two facts, first that exchanges cannot be sliced up into economic, juridical, religious, and other aspects, but are all that at one and the same time (a point not irrelevant here, but one that is now widely admitted); and second that men do not exchange things as we would think but, inextricably and fluctuatingly mixed up with those ‘things’, something of themselves.

I am not pleading for cancelling all distinction between subject and object; but only for releasing the value stress that bears on the matter, thus suspending its absolute character and allowing the boundary to fluctuate as the case may be, and/or other distinctions to come into play in keeping with native values.47

But is such an approach practicable? It has been attempted. A young scholar, André Itéanu, has taken such a course in his reanalysis of the Orokaiva, a Papuan society, from Williams’ and Schwimmer’s writings. In my reading of his thesis,48 he has found an alternative principle for ordering the data in an assumption that again contradicts our received conceptions, although it should not seem so surprising after all, namely that the society has to be thought of as including the dead, the relations with them being constitutive of it and offering the global framework within which not only all the detail of ritual and festive exchanges, but also what there is of social organization proper make sense.

The Orokaiva do not have moneys in the classical Melanesian sense. Yet, as Melanesian money has generally to do with life and the ancestors, the paramount place that the Orokaiva give to the dead reminds us of the cases where ceremonial exchanges do make use of institutional money. Here I am inclined to bring together two problems that can hardly be entirely left out of a discussion bearing on value. Those ‘primitive’ moneys have to do with absolute value. Therefore their relation to money in the modern, restricted sense of the term is somehow homologous to the relation, [231] among us, between value in the general, moral or metaphysical sense and value in the restricted economic sense. In the background of both lies the contrast between cultural forms that are essentially global and those in which the field is separated out or decomposed into particular domains or planes, that is, roughly speaking, between non-modern and modern forms.

Perhaps two features of the contrast may prove significant. Is it a fact in tribal societies that, where we have elaborate systems of exchange making use of one or more traditional moneys—mostly shell-moneys—to express and seal a wide range of ceremonial transitions and important rituals, we do not have permanent elaborate chiefship or rulership, and conversely that where the latter is found the former are absent? Melanesia and Polynesia seem clearly contrasted in that regard. If this were so, we might suppose that one thing can replace the other, that there is a certain equivalence of function between them. Now, in modern Europe the predominance of economic representations has resulted from the emancipation of economics from politics and has demanded, at some stage, the curtailment of political prerogatives.49 Is there here, despite the vast difference in the backgrounds, more than a chance parallelism, an index to a more general relation between two aspects of society?

Another feature drew the attention of Karl Polanyi. He contrasted the fixed ‘equivalencies’ between objects of exchange in primitive or archaic societies with the fluctuating price of goods in market economies. Mainly in the former case, the sphere of equivalence and possibly of exchange may be restricted to a few types of objects, while in the second, money tends to be a universal equivalent. But the question I want to raise is about the contrast between fixed and fluctuating rates of exchange. Polanyi attributed the fixity encountered in Dahomey to royal regulation.50 But the phenomenon was probably widespread. In the Solomon Islands, where regulation by political authority was out of the question, the rate of exchange between native money and the Australian dollar remained unaltered over a long period even though the devaluation of the dollar entailed very unpleasant consequences.51 At the other end of the spectrum, in the case of a high civilization and a complex society, Byzantium offers a spectacular case of fixity. The purchasing power of gold money [232] remained practically unchanged from the fifth to the eleventh century.52 The fact seems unbelievable if one thinks of the vicissitudes of the Empire during that period, where it was repeatedly, in every century, threatened in its very existence. Given the circumstances, the admitted excellence of imperial revenue administration is perhaps not a sufficient explanation of this remarkable phenomenon. I propose a different hypothesis which may or may not be tested but which I see other reasons for putting forth. When the rate of exchange is seen as linked to the basic value(s) of the society it is stable, and it is allowed to fluctuate only when and where the link with the basic value and identity of the society is broken or is no longer perceived, when money ceases to be a ‘total social fact’ and becomes a merely economic fact.53

It remains to recapitulate the foregoing and set in perspective the modern ideological framework and the anthropological predicament. The picture will be perforce incomplete and provisional, the language very approximate. The aim is to assemble a number of features, most of which have found stray recognition here or there, in order to perceive, or merely to sense, some of the relations between them. I insisted elsewhere on man as an individual as being probably the cardinal modern value, and on the concomitant emphasis on relations between men and things as against relations between men.54 These two features have notable concomitants regarding value.

First, the conception of man as an individual entails the recognition of a wide freedom of choice. Some of the values, instead of emanating from the society, will be determined by the individual for his own use. In other words, the individual as [233] (social) value demands that society should delegate to him a part of its value-setting capacity. Freedom of conscience is the standard example.55 The absence of prescription which makes choice possible is actually commanded by a superior prescription. Let me say in passing that it is therefore idle to suppose that men must have in all societies a similar range of choices open to them. Contrariwise, and very generally, value is embedded in the configuration of ideas itself. As we saw with right and left, this condition prevails as long as the relation between part and whole is effectively present, as long as experience is spontaneously referred to degrees of totality; and there is no place here for freedom of choice. We are faced, once again, with two alternative configurations; either value attaches to the whole in relation to its parts,56 and value is embedded, prescribed, as it were, by the very system of representation, or value attaches to the individual, which results, as we have seen, in the separation between idea and value. The antithesis is economically seen in terms of Tönnies’s Naturwille and Kürwille, the crux of the matter being that freedom of choice or Kürwille is exercised in a world without wholes, or rather in a world where the assemblages, sets or empirical wholes that are still encountered are deprived of their orientating function or value function.

Let us turn to the complex link between the modern value configuration and the relation between man and nature. Relations between men have to be subordinated for the individual subject to be autonomous and ‘equal’; the relation of man to nature acquires primacy, but this relation is sui generis, for, whether or not the independence of the individual demands it, man is indeed separated from nature: the free agent is opposed to nature as determined,57 subject and object are absolutely distinguished. Here we encounter science and its predominance in the culture as a whole. To cut the story short, let us say that the dualism in question is artificialist in essence: man has distanced himself from nature and the universe of which he was a part, and has asserted his capacity to remodel things according to his will. Again it makes full sense to say that Naturwille had been superseded by Kürwille, the latter being taken here less as arbitrary will than as detached, [234] disembedded, independent will. Given the close link between the will and values, it is worthwhile asking whence came this unprecedented type of will.

I surmise that it was forged in the otherworldliness, or rather outworldliness of early Christianity, from which issues finally the figure of Calvin, a prototype of modern man, with his iron will rooted in predestination. Only this Christian gestation seems to me to make understandable the unique and strange ‘prometheism’ of modern man.58

At any rate, with Kürwille as human will detached from nature and applied to its subjugation, we are in a position to appreciate the deep anchorage of the dichotomy between is and ought in modern ideology and life.

Finally, our two configurations embody two different relations between knowledge and action. In the one case, the agreement between the two is guaranteed on the level of the society:59 ideas are in conformity with the nature and order of the world, and the subject can do no better than consciously insert himself in this order. In the other case, there is no humanly significant world order, and it is left to the individual subject to establish the relation between representations and action, that is to say broadly speaking, between social representations and his own action. In the latter case, this world devoid of values, to which values are superadded by human choice, is a subhuman world, a world of objects, of things. One can know it exactly and act on it on condition of abstaining from any value imputation. It is a world without man, a world from which man has deliberately removed himself and on which he is thus able to impose his will.

This transformation has been made possible only by the devaluation of relations between men, relations which generally commanded the relation to things. They have lost, in the predominant ideology, their concrete character; they are especially seen from the viewpoint of relations to things (remember the Parsonian variables) except for one residue, namely moral action. Hence the abstract universality of the Kantian imperative.

So much for the subject side of the matter. Despite our absolute distinction between subject and object, there is some homology between the ways we look at both sides. I wish to add a few notes on the object side to complete the picture and to draw attention to a few features of the modern configuration of knowledge. It is a [235] commonplace to say that modern knowledge is distributed into a number of separate compartments, to speak of a high degree of division of labour, of scientific specialization. I shall try and characterize the modern model more precisely in contrast to the traditional one, of which we recalled some major aspects in the foregoing.

The modern configuration can be taken as resulting from the break-up of the value relation between element and whole. The whole has become a heap. It is as if a bag containing balls had volatilized: the balls have rolled away in all directions. This again is commonplace. The fact is that the objective world is made up of separate entities or substances in the image of the individual subject, whose empirically ascertained relations are taken as external to them.60 Yet the image is poor. In the first place, it suggests that the final distribution of the elements is random. Actually a complex, multidimensional world of ordered and fluctuating relations has been analysed, decomposed by the effort of (philosophical and) scientific reason into simpler components whose inner constitution and relations are quite peculiar. A somewhat better image is that of a multidimensional solid bursting out into a number of discrete, straight surfaces or planes that can accommodate only level linear figures and relations. Those planes have, I think, three characteristics: they are absolutely separate and independent, they are homologous to each other, and each of them is homogeneous throughout its extension.

The bursting out in general is relatively familiar: the history of modern painting from Impressionism onwards provides an example. The means that were until then subordinated to the descriptive reference were emancipated and each of them could in turn occupy the foreground. Nor is there any doubt as to the perfect separation of the ‘planes’ of knowledge: do we speak of physics or chemistry, psychology or physiology, psychology or sociology? But what is it that determined the identity of each of the disciplines among which the constituents of the world have been distributed? The answer seems to be that the instrumental point of view is decisive.61 Correlatively, we had occasion to point out the [236] extreme and striking weakness of the notion of a ‘whole’ in philosophical thought.

Secondly, the ‘planes’ on which knowledge and progress are concentrated remain ‘homogeneous’ throughout their extent. All the phenomena considered are of the same nature, have equal status, and are essentially simple. The paradigm here would be Galilei’s model of rectilinear uniform motion: a single material point moving through empty space. As a consequence, the planes have a tendency to split when the development of science reveals an (instrumental) heterogeneity.

Yet all planes are homologous, at least in principle, in the sense that the methods applied to diverse kinds of phenomena are identical. There is only one model of the natural sciences. It is true that with time and experience the model may be altered, but only with difficulty (witness biology and psychology). The model is mechanistic, quantitative, it rests on cause and effect (one individual agent, one individual result).62 It is essential to note that scientific rationality is present and at work only on each of these distinct planes, and that its exercise supposes that the whole has been put to pieces. It cannot reach beyond the relation of means to ends.

Successful as they have been in ensuring the mastery of man over the natural world, the sciences have had other results, one of which is to confront us with what Alexandre Koyré called ‘the enigma of man’. If anthropology is dealing, in its own manner, with this ‘enigma’, then it is both an integral part of the modern world, and in charge of transcending it, or rather of reintegrating it within the more common human world. I hope that our observations on value have pointed in that direction. There remains to face squarely the question of our relation to value: anthropology is poised between a ‘value-free’ science and the necessity to restore value to its proper and universal place. The philosophical critic of social science demands that it should be evaluative. He may grant us the ability to go beyond mere neutrality in the matter and yet maintain that we are unable to get rid of it completely and to evaluate or prescribe.

That is true in practice. It is not quite so, I suggest, in principle, and the point is worth making. What happens in the anthropol-[237]ogical view is that every ideology is relativized in relation to others. It is not a matter of absolute relativism. The unity of mankind, postulated and also verified (slowly and painfully) by anthropology, sets limits to the variation. Each particular configuration of ideas and values is contained with all others in a universal figure of which it is a partial expression.63 Yet this universal figure is so complex that it cannot be described, but only vaguely imagined as a kind of sum integral of all concrete configurations.

It is thus impossible for us to grasp directly the universal matrix in which the coherence of each particular value system is rooted, but it is perceptible in another way: each society or culture carries the trace of the inscription of its ideology within the human predicament. It is a negative mark, carved below the surface in intaglio. Just as an action has unforeseen consequences or ‘perverse effects’, or as each individual choice in our societies is immersed in a milieu of greater complexity and thus brings forth involuntary effects, so each ideo-normative configuration has its specific obscure yet compelling concomitants, which accompany it as its shadow and which manifest the human condition in relation to it. These concomitants are what I called in a somewhat different context the ‘non-ideological features’ that we find by comparison and which we see as non-conscious aspects, unsuspected by the people themselves.64

There is thus in each concrete society the imprint of this universal model, which becomes perceptible to some degree as soon as comparison begins. It is a negative imprint, which authenticates, so to speak, the society as human, and whose precision increases when comparison proceeds. It is true that we cannot derive a prescription from this imprint, but it represents the reverse side of the prescription, or its limit. In principle, anthropology is thus fraught with progress in the knowledge of value, and hence of prescription itself, and this should lead in the end to a reformulation of the philosopher’s problem.

But what about here and now? Granted that the meaning of ‘prescription’ is made more complex in our perspective, to the extent that we should prefer to speak of counsel rather than injunction, can we not offer something of the sort on the basis of our factual conclusions? We found that the modern configuration, however opposed to the traditional, is still located within it: the modern model is an exceptional variant of the general model and remains encased, or encompassed, within it. Hierarchy is [238] universal, at the same time it is here partially but effectively contradicted. What is it, then, that is necessary in it? A first and approximate answer is that there are things that equality can and things that it cannot do. A contemporary trend in public opinion, in France and elsewhere, suggests an example.

There is much talk round about of ‘difference’, the rehabilitation of those that are in one way or an other ‘different’, the recognition of alter. This may mean two things. In so far as it is a matter of enfranchisement in general, equal rights and opportunities, equal treatment of women, or of homosexuals, etc.,—and such seems to be the main import of the claims put forward on behalf of such categories —there is no theoretical problem. It should only be pointed out that in such equalitarian treatment difference is disregarded, neglected or subordinated, and not ‘recognized’. Given the easy transition from equality to identity, the long-range outcome is likely to be in the direction of the erasing of distinctive characteristics in the sense of a loss of the meaning or value previously attributed to the corresponding distinctions.

But there may be more in these claims. The impression is that another meaning is also subtly present in them, namely the recognition of alter qua alter. I submit that such recognition can only be hierarchical—as was keenly perceived by Burke in his Reflections on the French Revolution. Here, to recognize is the same as to value or to integrate (remember the Great Chain of Being). This statement flies in the face of our stereotypes or prejudices, for nothing is more remote from our common sense than Thomas Aquinas’s dictum that ‘order is seen to consist mainly in inequality (or difference: disparitate)’.65 Yet it is only by a perversion or impoverishment of the notion of order that we may believe contrariwise that equality can by itself constitute an order. To be explicit: alter will then be thought of as superior or inferior to Ego, with the important qualification of reversal (which is not present in the Great Chain as such). That is to say that, if alter was taken as globally inferior, he would turn out as superior on secondary levels of consideration.66

[239] What I maintain is that, if the advocates of difference claim for it both equality and recognition, they claim the impossible. Here we are reminded of the American slogan ‘separate but equal’ which marked the transition from slavery to racism. To be more accurate, however, I should say that the above is true on the level of pure representation—equality or hierarchy—and thus make room for an alternative of a different kind. As to the practical forms of integration, most of those we can think of either assemble equal, principially identical agents, as in co-operation, or refer to a whole and are implicitly hierarchical, as the division of labour. Only conflict qualifies, as Max Gluckman has shown, as integrator. We should then say, speaking roughly, that there are two ways of somehow recognizing alter: hierarchy and conflict. Now, that conflict is inevitable and perhaps necessary is one thing, and to posit it as an ideal, or as an ‘operative value’, is quite another67 —although it is in keeping with the modern trend. Did not Max Weber himself grant more credibility to war than to peace? Conflict has the merit of simplicity while hierarchy entails a complication similar to that of Chinese etiquette. The more so, as it would here have to be encompassed within the paramount value of individualism and equality. Yet I must confess my irenic preference for the latter.


Albert, Ethel M. (1956). ‘The Classification of Values: A Method and Illustration’, American Anthropologist, 58.

Bateson, Gregory (1972). Steps to an Ecology of Mind. Ballantine Books.

Buber, Martin (1938). Je et tu. Paris.

Caspary, Gerard E. (1979). Politics and Exegesis: Origen and the Two Swords. Berkeley, University of California Press.

Dumont, Louis (1953a). ‘The Dravidian Kinship Terminology as an Expression of Marriage’, Man, March 1953, no. 54.

——— (1953b). ‘Dravidian Kinship Terminology’, Man, Sept. 1953, no. 224.

——— (1966). ‘A fundamental problem in the Sociology of Caste’, Contributions to Indian Sociology. ix. Dec. 1966; pp. 17–32.

——— (1975). La civilisation indienne et nous. Paris, Colin.

——— (1977). From Mandeville to Marx, University of Chicago Press.

——— (1979). ‘The Anthropological Community and Ideology’, Social Science Information, 18, 6. 1979.

——— (1980a). Homo Hierarchicus. Complete and revised edition. University of Chicago Press. (In French: Gallimard, coll. Tel, 1979).

——— (1980b). ‘A Modified View of Our Origins. The Christian beginnings of modern individualism’. Deneke Lecture 1980. (To appear.)

Erikson, Erik H. (1964). Insight and Responsibility. New York, Norton,

Gauchet, Marcel (1980). ‘Tocqueville, l'amérique et nous’, Libre, 7. (Paris, Payot); pp. 43–120.

Gierke, Otto (1958). Political Theories of the Middle Age. Trans. by Maitland (1900), Beacon Press.

Hsu, Francis L. (1961). ‘American Core Value and National Character’. In Hsu, ed. Psychological Anthropology, Homewood, Ill. 1961; pp. 209–30.

Itéanu, André (1980). Échanges, circulation et valeurs chez les Orokaiva. Thèse de 3e cycle, Univ. de Paris VII (unpublished).

Jacob, Francois (1970). La logique du vivant. Une histoire de l'hérédité. Paris, Gallimard.

Kluckhohn, Clyde, et al. (1951). ‘Values and Value-orientation in the Theory of Action. An exploration in definition and classification’, in Talcott Parsons and Edward A. Shils, eds., Toward a General Theory of Action. Cambridge, Mass. 1951.

———. (1952). ‘Categories of Universal Culture’. Draft, Wenner-Gren symposium, June 1952. Tozzer Library, Harvard University.

———. (1959). ‘The Scientific Study of Values’. Offprint in the Tozzer Library. (Kluckhohn file).

Kluckhohn, Florence (1961). ‘Dominant and Variant Value Orientations’, in F. Kluckhohn and F. L. Strodtbeck, eds. Variations in Value Orientations, Evanston, Ill. 1961.

Koestler, Arthur (1967). The Ghost in the Machine. London, Hutchinson.

Kolakowski, Leszek (1977). ‘The Persistence of the Sein-Sollen Dilemma’, Men and World, vol. 10, no. 2.

Koyré, Alexandre (1958). From the Closed World to the Infinite Universe. Harper Torchbooks.

Lalande, André (1968). Vocabulaire technique et critique de la philosophie. 2e ed., Paris, PUF.

Leach, Edmund (1961). Rethinking Anthropology. London, Athlone Press.

———. (1976). Social Anthropology: A Natural Science of Society? Radcliffe-Brown Lecture 1976. (From the Proceedings of the British Academy, vol. lxii, 1976). Oxford University Press.

Lepley, Ray. ed. (1949). Value: a Co-operative Inquiry. New York, Columbia University Press.

Lovejoy, Arthur (1973). The Great Chain of Being (1933). Oxford University Press, 1973.

Mol, Hans (1976). Identity and the Sacred. A Sketch for a New Social Scientific Theory of Religion. Agincourt, The Book Society of Canada.

Morris, Charles (1956). Varieties of Human Value. University of Chicago Press.

Northrop, F. S. C. (1946). The Meeting of East and West. New York, Macmillan.

Ostrogorsky, Georges (1969). Histoire de l’état byzantin. Paris, Pavor.

Parsons, Talcott and Shils, Edward A., eds. (1951). Toward a General Theory of Action. Harvard University Press, quoted in Harper Torchbooks edn. 1962.

Phillips, D. C. (1976). Holistic Thought in Social Science. Stanford University Press.

Plenge, Johann (1930). Zur Ontologie der Beziehung (Allgemeine Relationstheorie). Münster i.W. (Forschungsinstitut für Organisationslehre).

Polanyi, Karl and Rotstein, Abraham (1966). Dahomey and the Slave Trade: an Analysis of an Archaic Economy, Seattle and London, University of Washington Press.

Pribram, Karl (1922). ‘Deutscher Nationalismus und deutscher Sozialismus’. Archiv für Socialwissenschaft und Sozialpolitik, 49: pp. 248–376.

Radcliffe-Brown, A. R. (1953). ‘Dravidian Kinship Terminology’. Man, July 1953, no. 169.

———. (1957). A Natural Science of Society. Glencoe, Ill., The Falcon’s Wing Press.

———. (1958). Method in Social Anthropology. Edited by M. N. Srinivas, Chicago.

Ritter, Joachim (1977). ‘Hegel und die französische Revolution’ [1957] in Metaphysik und Politik. Frankfurt, Suhrkamp.

Sahlins, Marshall (1972). Stone Age Economics. Chicago, Aldine-Atherton.

Serres, Michel (1968). Le système de Leibniz et ses modèles mathématiques. 2 vols. Paris, PUF.

Sheldon, Richard C. (1951). ‘Some Observations on Theory in Social Science’, in Talcott Parsons and Edward A. Shils, eds. Toward a General Theory of Action. Cambridge, Mass., 1951.

Strauss, Leo (1954). Droit naturel et Histoire. Paris, Plon. (Original: Natural Right and History. Chicago, 1953).

Williams, Raymond (1976). Keywords. New York, Oxford University Press.



Louis DUMONT (1911–1998) was a prominent French anthropologist and Director of Research at the École des hautes études en sciences sociales (EHESS) in Paris. A specialist in the study of cultures and societies in India, western philosophy, and the anthropology of kinship, his works include From Mandeville to Marx: The genesis and triumph of economic ideology (University of Chicago Press, 1977), Homo Hierarchicus: The caste system and its implications (University of Chicago Press, 1980), Affinity as a value: Marriage alliance in South India, with comparative essays on Australia (University of Chicago Press, 1983), and Essays on individualism: Modern ideology in anthropological perspective (University of Chicago Press, 1987).


Publisher’s note: This is a reprint of Louis Dumont, 1980. “On Value.” Proceedings of the British Academy 66: 207–41. We are very grateful to the British Academy for granting us the permission to reprint the lecture. We remind the reader that we retain the style of the original, and indicate in the text the original page numbers in square brackets.

1. The author is grateful to Joseph Erhardy who has helped, here as so often in the past, to give his English a more acceptable shape than it would otherwise have had. Thanks are also due to Alan Montefiore, who has kindly suggested some improvements in that regard.

2. Radcliffe-Brown (1958).

3. Actually, Sir Raymond Firth tells me that such developments were habitual in Radcliffe-Brown’s teaching, from early days onwards (in Australia in the thirties). Radcliffe-Brown said in the lecture ‘the kind of structure with which we are concerned is one of the union of opposites’ (Radcliffe-Brown (1958), p. 123, my emphasis). It was thus a particular case, not the application of a general principle, which required speaking of ‘oppositions’. Cf. Leach (1976), p. 9. Accordingly, my first and limited attempt at structuralist analysis (Dumont (1953a)) drew shortly afterwards a magisterial rebuke from the aging Radcliffe-Brown (Radcliffe-Brown (1953), my reply Dumont (1953b)). My paper was a piece of that ‘Parisian heresy’ which, as Sir Edmund Leach said (ibid.), was mostly ignored in this country for ten years or more. Yet, let it be said for the record that Radcliffe-Brown’s strictures did not alter the friendly protection and non-committal encouragement of Evans-Pritchard who, of all colleagues, showed most understanding for the effort at a systematic retrieval of affinity.

4. Sir Edmund Leach has discussed at length (Leach (1976)) this posthumous presentation of Radcliffe-Brown’s widest views (Radcliffe-Brown (1957)). In it the positive aspects of Radcliffe-Brown’s teaching appear clearly, together with what appears to us now (or to me) as its shortcomings. In retrospect, he is seen to have gone in the right direction, but not quite far enough. Yet his articulate holism (pp. 22, 110, etc.) coupled with the consequent stress on ‘relational analysis’ and on synchrony (pp. 14, 63), and, remarkably enough, with the downgrading of causality (p. 41, cf. below, n. 62), appears as very meritorious if one looks at it against the background of the nominalism which permeates his own thought and the predominant orientation in British ideology. In this perspective, it is not surprising that Radcliffe-Brown’s holism remains narrowly functional, that the distinction between ‘culture’ (somewhat reluctantly ushered in, p. 92) and ‘social structure’, sound in principle, in fact reduces the former to a mere means of the latter (p. 121). Also Radcliffe-Brown did not– probably could not–perceive that relational analysis demands that the boundaries of the ‘system’ be rigorously defined and not left to arbitrary choice or expediency (p. 60), and that such analysis is incompatible with the primary emphasis he put on classification or taxonomy (pp. 16, 71) (See Leach’s early dismissal of ‘butterfly collecting’, Leach (1961)). I shall refer to a few other points in the following (‘natural kinds of systems’, n. 61 below; fixed equivalencies in exchange, n. 53 below).

5. See Radcliffe-Brown (1957), pp. 10–1, 119, 136–40 (economic value).

6. Lalande (1968), p. 1183.

7. Hsu (1961), pp. 209–30.

8. Morris (1956).

9. Arthur Lovejoy sees in some passages of Plato a contradiction between the Good (or God) being self-sufficient in its perfection and its being the ground and source of this world: the same entity cannot be both complete in itself and in any degree dependent on something else. (Lovejoy (1973), pp. 43–50). But Lovejoy comes to this contradiction by erasing the philosopher’s progress and flattening its result. In a first step one must turn away from the world to come to grasp the Idea of the Good (and True and Beautiful). In a second step, once the Good is correctly understood–as limitless generosity or irrepressible fecundity–one finds that it explains and justifies the world as it is. These two conclusions are not at the same level: on an inferior level God is absolutely distinct from the world, on a superior level the world itself is contained in God; the Good transcends the world and yet the world has no being but through it. The world depends on God, God does not depend on the world. The crux of the matter is that Lovejoy stops at the inferior level. He does not and probably cannot accept hierarchy, or transcendence. He looks at Plato with egalitarian eyes.

10. Gregory Bateson (1972), pp.271–8 (double bind), 336, and passim; cf. Kluckhohn (1951) ‘what appear superficially as incompatibilities are seen on closer examination to be functions of different frames of reference’, (p. 399 n. 19); the difference is between seeing things-in-themselves and seeing things in relation, i.e. within a ‘frame of reference’.

11. Cf. F. S. C. Northrop (1946) and see p. 257; Ray Lepley, ed. (1949); Clyde Kluckhohn himself alludes to the circumstances (ibid., pp. 388–9).

12. Kluckhohn (1951), p. 388–433. Kluckhohn reiterated his basic position in a number of papers.

13. See the note in Parsons and Shils (1951), p. 26–7.

14. Mol (1976).

15. ‘In the working out of the theory by far the major attention is paid to value-orientations (as against ideas and beliefs) because much of the theory is concerned with the selection by actors of objects and gratifications’ writes Richard Sheldon in what is actually a minute of dissent (Sheldon (1951), p. 40). Sheldon went on to say that this stress on personality and on the ‘social system’ resulted in cutting culture in two.

16. Albert (1956), pp. 221–48.

17. The reference is to a later version of Florence Kluckhohn (1961).

18. Cf. especially Clyde Kluckhohn (1952). It must be added that Florence Kluckhohn was particularly keen on nuances in the hierarchical make-up, which enabled her to grasp variations not only between cultures but also within a given value system, thus securing an opening toward the question of changes in values.

19. Kluckhohn (1959). The text is apparently a part of a volume of Installation Lectures, which I have not been able to identify (pp. 25–54). It would not have been earlier than 1959.

20. See Kluckhohn (1959) section II, and (1951), p 389.

21. See Homo Hierarchicus (Dumont (1980a)), App. A.

22. It is the fulcrum of the discussion in the symposium edited by Lepley (Lepley, 1949). The pragmatists’ attempt goes against the means/ends distinction, which is akin to the others we referred to and is as fundamental as they are to modern culture.

23. Kolakowski (1977). I alluded to the problem in Dumont (1977).

24. Pribram (1922).

25. See Dumont (1979), p. 795. A caricatural example: according to Ritter, Hegel succeeded in building up an Aristotelian philosophy of the French Revolution (Ritter (1977).

26. Dumont (1979), p. 796. It goes without saying that, to be successful, such a distinction of levels should be present in the consciousness of the citizens.

27. Leo Strauss (1954) ch. 2 and p. 85.

28. Discussion and references in Dumont (1966), pp. 25–7.

29. As the reference to ‘embeddedness’ may remind the reader, we have been following in the footsteps of Karl Polanyi and simply widening his thesis on the exceptional character of modern civilization.

30. We found the point stressed by Parsons and Shils (1951) as well as by Kluckhohn (above). The latter analysed the interplay between normative and ‘existential’ statements (1951, pp. 392–4); he quoted (p. 422). Herskovits on ‘cultural focus’ as linking the distribution of values and the configuration of ideas (also Dumont (1977), pp. 19–20 and (1979), p. 814).

31. For details and references, see Dumont (1979), pp. 806–15.

32. Therelation between whole and part was previously defined as the hierarchical opposition, or the encompassing of the contrary (ibid. and 1980). For Thomas Aquinas, difference by itself suggested hierarchy. So that ‘order is seen to consist mainly in inequality (or difference: disparitate)’, cf. Otto Gierke (1958), (= DGR III) n. 88.

33. To assert that the modern mode of thought is destructive of the wholes with which man had until then seen himself surrounded may seem excessive or incomprehensible. Yet I think it is true in the sense that each whole has ceased to be value-providing in the above sense. If one turns to our philosophies with the simple question: What is the difference between a whole and a collection, most of them are silent, and when they give an answer, it is likely to be superficial or mystical as in Lukàcs, cf. Kolakowski (1977). I take it as exemplary that the constitution of Hegel’s system results from a shift in the location of the Absolute, or of infinite value, from the Whole of Being (in the writings of his youth) to the Becoming of the individual entity–a point I intend to argue elsewhere. There is a small current of holistic thought, but it also bears the mark of the difficulty that modern minds experience in the matter, see D. C. Phillips (1976)–the discussion is sometimes tendentious. A book of Alfred Koestler’s (1967) represents an exception. To quote from a summary (p. 58): ‘Organisms and societies are multi-levelled hierarchies of semi-autonomous sub-wholes branching into sub-wholes of a lower order, and so on. The term “holon” has been introduced [by the author] to refer to these intermediary entities which, relative to their subordinates in the hierarchy, function as self contained wholes; relative to their superordinates as dependent parts.’ Koestler is seen to stress hierarchy as a chain of levels, while I have insisted on the elementary relation between two successive levels. The definition of ‘holon’ is valuable. I would only rank the two faces of this ‘Janus’ in relation to each other: the integration of each sub-whole as a unit in the next higher one is primary, its self-integration or ‘self-assertion’ is secondary (Dumont (1980a), p. 403)·

We have already noted Gregory Bateson’s recognition of the hierarchy of levels (n. 10 above). A biologist, Francois Jacob, introduced the ‘integron’ in a sense somewhat similar to Koestler’s ‘holon’ (Jacob (1971) p. 323).

34. Is it possible that what is true of particular entities or wholes (sub-holes or ‘holons’ in Koestler’s terms) is true also of the great Whole, the universe or whole of wholes? Is it possible that the Whole in its turn needs a superior entity from which to derive its own value? That it can be self-integrative only by its subordination to something beyond itself? Clearly religions have a place here, and one could even try to deduce what the Beyond should be like in order to be final. Then we could say not only that men feel a need for a complement to the ‘empirically’ given, as Durkheim supposed, but that the need bears on an apex of valuation. This speculation arises from an exactly opposite view put forward by Lovejoy. He begins his classical book The Great Chain of Being (1973, see below) by positing ‘otherworldliness’ as a general attitude found in different forms in some of the world religions and consisting in taking refuge outside the world from its incoherence and wretchedness. Lovejoy states an absolute separation between this attitude and the world: it is only a place to get away from and about which otherworldliness has nothing to say (ibid., pp. 28–30). Here we may wonder. Let us take, as Lovejoy tends to do, an extreme form of ‘otherworldliness’ such as Buddhism. No doubt Buddha was not busy justifying the world. Yet he offers a kind of explanation of it, if a negative one. In general, the beyond is more than a refuge, It is a distant place from which, so to speak, one looks back with detachment upon human experience in the world; it is finally a transcendence that is posited and in relation to which the world is situated. Has not this transcendent glance been historically necessary to the understanding of the world as a whole? At any rate history shows abundantly, in India and perhaps in the West as well, that otherworldliness has powerfully acted on life in the world, and this process would be incomprehensible if an absolute heterogeneity was presupposed.

35. Lovejoy (1973), ch. 7.

36. Cf. Michel Serres. (1968). Leibniz’s world should not be simply identified with the traditional world. Perhaps theodicies are an index of individualist questioning and an effort, more or less successful, to reassert the holistic view. On the other hand, the Voltairian mood has had to stomach certain lessons, to learn, for instance that one pole of a magnet could not be separated from the other as some would have wished. ‘Jadis, en brisant les aimants, on cherchait à isoler le magnétisme nord et le magnétisme sud. On espérait avoir deux principes différents d'attraction. Mais à chaque brisure, si subit, si hypocrite que fût le choc, on retrouvait, dans chacun des morceaux brisés, les deux pôles inséparables’ (Bachelard in his preface to Buber (1938), p. 9).

37. Cf. Dumont (1977), index, s.v. Hierarchy, instances.

38. Cf. Dumont (1980b).

39. Cf. Dumont (1975), p. 30.

40. Gerard Caspary (1979), pp. 113–14, 189–91. The whole conclusion should be read.

41. Erikson (1964), p. 92.

42. Eriksontakes both forms as normal, although one is obviously inferior (‘more primitive’), to the other. At the same time he points out acutely the possible transition from the mechanical form to the totalitarian disease. In that regard the weakness or the very absence of the structural form in philosophical discourse is remarkable.

43. Lovejoy (1973).

44. Pseudo-Dionysius Areopagitus, Celestial Hierarchy, ch. 3, §§ 1–2, 164d–165a. The translation is by Jasper Griffin, Fellow of Balliol College, Oxford, to whom I am grateful, and who also provided the following translation (n. 45).

45. Very similar is the function of Love (Eros) as defined in Plato’s Symposium by Diotima: he is a daemon, that is, a being intermediary between gods and men: ‘He interprets and makes a communication between divine and human things, conveying the prayers and sacrifices of men to the Gods, and communicating the commands and directions concerning the mode of worship most pleasing to them, from Gods to men. He fills up that intermediate space between these two classes of beings, so as to bind together, by his own power, the whole universe of things.’ (202e, Shelley’s translation.)

46. The distinction accompanies in particular the priority of the relation between man and nature, and on that account is already eccentric for a system stressing relations between men. The value stress is seen even in the contradictory valuations of subject and object in positivism and in idealism, of which Raymond Williams reminds us (1976, pp. 259–60).

47. There is a precedent in German philosophy, in Schelling’s philosophy of nature, where he wanted to transcend the Kantian duality, and downgraded this distinction to one of mere degree or complementarity within a class. I am not advocating Schelling’s perhaps primitive and inefficient device. For us each particular context should be decisive.

48. Itéanu (1980).

49. Dumont (1977), p. 6.

50. Polanyi (1966).

51. Oral communication from Daniel de Coppet (about the 'are’ Are on Malaita).

52. Ostrogorsky (1969), pp. 68, 219n., 317, 371.

53. Radcliffe-Brown had already attracted attention to fixed equivalences as against the action of supply and demand ((1957) pp. 112, 114, 138).–The hypothesis may seem unwarranted, coming after the careful and thoughtful study of Marshall Sahlins (1972, ch. 6). As formulated here, however, it is not straightforwardly contradicted by Sahlins’ conclusion. We may read him as stating only that contact with a market economy and/or radical economic changes have directly or indirectly an action on fixed equivalences in the long run. Also there may be, between the two conditions that the hypothesis contrasts, intermediary transitional stages with a complex interaction between norm and fact.

54. Startingfrom these two kinds of relations and tracing their application and combinations, the German sociologist Johann Plenge developed a complete– ahierarchical–and impeccable classification of relations in a brochure published in 1930: Plenge (1930).

55. The individual’s capacity is obviously limited. Analytically, either he exerts his choice between existing virtual values, or existing ideas, or he constructs a new idea-value (which must be rare).

56. Koestler allows for more precision: ‘the whole’ is mostly a sub-whole or holon, itself part of a higher sub-whole.

57. Descartes’s penseé et étendue, etc.

58. See Dumont (1980b).

59. The relation is intrinsically problematic. To ensure it is the essential and distinctive function of religion (cf. the note in Dumont (1977), p. 214).

60. Predominantly at any rate. About ‘internal relations’, see Phillips (1976), cf. n. 33 above.

61. Radcliffe-Brownwrote of ‘natural kinds of system’ (1957, p. 23); thus implicitly admitting that the separation between scientific disciplines is grounded in nature. The relation with the predominance of nominalism in science is obvious. The Cartesian difficulty of conceiving the relations between soul and body is perhaps the archetype of such fission. Hence the surfeit of contradictions and of simple oppositions badly subsumed.

62. It is noteworthy that Radcliffe-Brown saw the incompatibility between a holistic or systemic approach and causal explanation and rejected causality from his ‘theoretical social science’ (1957, p. 41).

63. Dumont (1979), p. 793.

64. Dumont (1980a), § 118.

65. Cf. n. 32, above.

66. For the application to societies, see Dumont (1979), p. 795. If we suppose the levels to be numerous, and the reversal multiplied, then we have a fluctuating dyadic relationship which may statistically give the impression of equality. In a quite different context, Sahlins’ analysis of exchange in the Huon Gulf is pregnant with meaning (Sahlins (1976), pp. 322 f.). [Editor’s note: the reference to the English edition should be Sahlins (1972), pp. 285 ff.] Briefly: (1) between two commercial partners, each of the exchanges in a series is unbalanced, alternatively in one and the other direction, in approximation to a balance reached in the end, i.e. for the series as a whole; equality is thus reached through a succession of somewhat unequal exchanges; (2) each particular exchange is thus not closed but remains open and calls for the next one: the stress is on the continuing relationship more than on instantaneous equivalence between goods. All aspects of our problem are here contained in a nutshell: the difference between hierarchy and equality is not at all what we are wont to suppose.

67. Thisis what, in my view, Marcel Gauchet does in a thoughtful reappraisal of Tocqueville; Gauchet (1980, see esp. pp. 90–116).