cover of The Savage Mind

Claude Lévi-Strauss 1962

History and Dialectic

Source: final chapter of “The Savage Mind,” October 1961, translated by George Wiedenfeld and Nicholson Ltd 1966. Footnotes and references omitted

In the course of this work I have allowed myself, not without ulterior motive, to borrow a certain amount of Sartre’s vocabulary. I wanted to lead the reader to face a problem, the discussion of which will serve to introduce my conclusion. The problem is to what extent thought that can and will be both anecdotal and geometrical may yet be called dialectical. The savage mind totalizes. It claims indeed to go very much further in this direction than Sartre allows dialectical reason, for, on the one hand, the latter lets pure seriality escape (and we have just seen how classificatory systems succeed in incorporating it) and, on the other, it excludes schematization, in which these same systems reach their consummation. In my view, it is in this intransigent refusal on the part of the savage mind to allow anything human (or even living) to remain alien to it, that the real principle of dialectical reason is to be found. But my idea of the latter is very different from Sartre’s.

In reading the Critique it is difficult to avoid feeling that Sartre vacillates between two conceptions of dialectical reason. Sometimes he opposes dialectical and analytical reason as truth and error, if not as God and the devil, while at other times these two kinds of reason are apparently complementary, different routes to the same truths. The first conception not only discredits scientific knowledge and finally even leads to suggesting the impossibility of a science of biology, it also involves a curious paradox; for the work entitled Critique de la raison dialectique is the result of the author’s exercise of his own analytical reason: he defines, distinguishes, classifies and opposes. This philosophical treatise is no different in kind from the works it examines and with which it engages in discussion, if only to condemn them. It is difficult to see how analytical reason could be applied to dialectical reason and claim to establish it, if the two are defined by mutually exclusive characteristics. The second conception is open to a different objection: if dialectical and analytical reason ultimately arrive at the same results, and if their respective truths merge into a single truth, then, one may ask in what way they are opposed and, in particular, on what grounds the former should be pronounced superior to the latter. Sartre’s endeavour seems contradictory in the one case and superfluous in the other.

How is the paradox to be explained, and avoided? Sartre attributes a reality sui generis to dialectical reason in both the hypotheses between which he hesitates. It exists independently of analytical reason, as its antagonist or alternatively its complement. Although in both our cases Marx is the point of departure of our thought, it seems to me that the Marxist orientation leads to a different view, namely, that the opposition between the two sorts of reason is relative, not absolute. It corresponds to a tension within human thought which may persist indefinitely de facto, but which has no basis de jure. In my view dialectical reason is always constitutive: it is the bridge, forever extended and improved, which analytical reason throws out over an abyss; it is unable to see the further shore but it knows that it is there, even should it be constantly receding. The term dialectical reason thus covers the perpetual efforts analytical reason must make to reform itself if it aspires to account for language, society and thought; and the distinction between the two forms of reason in my view rests only on the temporary gap separating analytical reason from the understanding of life. Sartre calls analytical reason reason in repose; I call the same reason dialectical when it is roused to action, tensed by its efforts to transcend itself.

In Sartre’s terminology I am therefore to be defined as a transcendental materialist and aesthete. I am a transcendental materialist because I do not regard dialectical reason as something other than analytical reason, upon which the absolute originality of a human order would be based, but as something additional in analytical reason: the necessary condition for it to venture to undertake the resolution of the human into the non-human. And I count as an aesthete since Sartre applies this term to anyone purporting to study men as if they were ants. But apart from the fact that this seems to me just the attitude of any scientist who is an agnostic, there is nothing very compromising about it, for ants with their artificial tunnels, their social life and their chemical messages, already present a sufficiently tough resistance to the enterprises of analytical reason ... So I accept the characterization of aesthete in so far as I believe the ultimate goal of the human sciences to be not to constitute, but to dissolve man. The pre-eminent value of anthropology is that it represents the first step in a procedure which involves others. Ethnographic analysis tries to arrive at invariants beyond the empirical diversity of human societies; and, as the present work shows, these are sometimes to be found at the most unforeseen points. Rousseau foresaw this with his usual acumen: ‘One needs to look near at hand if one wants to study men; but to study man one must learn to look from afar; one must first observe differences in order to discover attributes’. However, it would not be enough to reabsorb particular humanities into a general one. This first enterprise opens the way for others which Rousseau would not have been so ready to accept and which are incumbent on the exact natural sciences: the reintegration of culture in nature and finally of life within the whole of its physico-chemical conditions.[1]

However, in spite of the intentionally brutal turn given to my thesis, I am not blind to the fact that the verb ‘dissolve’ does not in any way imply (but even excludes) the destruction of the constituents of the body subjected to the action of another body. The solution of a solid into a liquid alters the disposition of its molecules. It also often provides an efficacious method of putting them by so that they can be recovered in case of need and their properties be better studied. The reductions I am envisaging are thus legitimate, or indeed possible, only if two conditions are satisfied. First, the phenomena subjected to reduction must not be impoverished; one must be certain that everything contributing to their distinctive richness and originality has been collected around them. For it is pointless to pick up a hammer unless to hit the nail on the head.

Secondly, one must be ready to accept, as a consequence of each reduction, the total overturning of any preconceived idea concerning the level, whichever it may be, one is striving to attain. The idea of some general humanity to which ethnographic reduction leads, will bear no relation to any one may have formed in advance. And when we do finally succeed in understanding life as a function of inert matter, it will be to discover that the latter has properties very different from those previously attributed to it. Levels of reduction cannot therefore be classed as superior and inferior, for the level taken as superior must, through the reduction, be expected to communicate retroactively some of its richness to the inferior level to which it will have been assimilated. Scientific explanation consists not in moving from the complex to the simple but in the replacement of a less intelligible complexity by one which is more so.

Seen in this light, therefore, my self is no more opposed to others than man is opposed to the world: the truths learnt through man are ‘of the world’, and they are important for this reason.[2] This explains why I regard anthropology as the principle of all research, while for Sartre it raises a problem in the shape of a constraint to overcome or a resistance to reduce. And indeed what can one make of peoples ‘without history’ when one has defined man in terms of dialectic and dialectic in terms of history? Sometimes Sartre seems tempted to distinguish two dialectics: the ‘true’ one which is supposed to be that of historical societies, and a repetitive, short-term dialectic, which he grants so-called primitive societies whilst at the same time placing it very near biology. This imperils his whole system, for the bridge between man and nature which he has taken such pains to destroy would turn out to be surreptitiously re-established through ethnography, which is indisputably a human science and devotes itself to the study of these societies. Alternatively Sartre resigns himself to putting a ‘stunted and deformed’ humanity on man’s side, but not without implying that its place in humanity does not belong to it in its own right and is a function only of its adoption by historical humanity: either because it has begun to internalize the latter’s history in the colonial context, or because, thanks to anthropology itself, historical humanity has given the blessing of meaning to an original humanity which was without it. Either way the prodigious wealth and diversity of habits, beliefs and customs is allowed to escape; and it is forgotten that each of the tens or hundreds of thousands of societies which have existed side by side in the world or succeeded one another since man’s first appearance, has claimed that it contains the essence of all the meaning and dignity of which human society is capable and, reduced though it may have been to a small nomad band or a hamlet lost in the depths of the forest, its claim has in its own eyes rested on a moral certainty comparable to that which we can invoke in our own case. But whether in their case or our own, a good deal of egocentricity and naivety is necessary to believe that man has taken refuge in a single one of the historical or geographical modes of his existence, when the truth about man resides in the system of their differences and common properties.

He who begins by steeping himself in the allegedly self-evident truths of introspection never emerges from them. Knowledge of men sometimes seems easier to those who allow themselves to be caught up in the snare of personal identity. But they thus shut the door on knowledge of man: written or unavowed ‘confessions’ form the basis of all ethnographic research. Sartre in fact becomes the prisoner of his Cogito: Descartes made it possible to attain universality, but conditionally on remaining psychological and individual; by sociologizing the Cogito, Sartre merely exchanges one prison for another. Each subject’s group and period now take the place of timeless consciousness. Moreover, Sartre’s view of the world and man has the narrowness which has been traditionally credited to closed societies. His insistence on tracing a distinction between the primitive and the civilized with the aid of gratuitous contrasts reflects, in a scarcely more subtle form, the fundamental opposition he postulates between myself and others. Yet there is little difference between the way in which this opposition is formulated in Sartre’s work and the way it would have been formulated by a Melanesian savage, while the analysis of the practico-inert quite simply revives the language of animism.[3]

Descartes, who wanted to found a physics, separated Man from Society. Sartre, who claims to found an anthropology, separates his own society from others. A Cogito – which strives to be ingenuous and raw – retreats into individualism and empiricism and is lost in the blind alleys of social psychology. For it is striking that the situations which Sartre uses as a starting point for extracting the formal conditions of social reality – strikes, boxing matches, football matches, bus-stop queues – are all secondary incidentals of life in society; and they cannot therefore serve to disclose its foundations.

This axiomatic, so far removed from the anthropologist’s, is all the more disappointing when he feels himself very close to Sartre whenever the latter applies himself, with incomparable artistry, to grasping, in its dialectical movement, a present or past social experience within our own culture. Sartre then does what every anthropologist tries to do in the case of different cultures: to put himself in the place of the men living there, to understand the principle and pattern of their intentions, and to perceive a period or a culture as a significant set. In this respect we can often learn from him, but these are lessons of a practical, not a theoretical, nature. It is possible that the requirement of ‘totalization’ is a great novelty to some historians, sociologists and psychologists. It has been taken for granted by anthropologists ever since they learned it from Malinowski. But Malinowski’s deficiencies have also taught us that this is not where explanation ends. It only begins when we have succeeded in constituting our object. The role of dialectical reason is to put the human sciences in possession of a reality with which it alone can furnish them, but the properly scientific work consists in decomposing and then recomposing on a different plane. With all due respect to Sartrian phenomenology, we can hope to find in it only a point of departure, not one of arrival.

Furthermore, dialectical reason must not let itself be carried away by its own elan, nor must the procedure leading to the comprehension of an other reality attribute to it, in addition to its own dialectical features, those appertaining to the procedure rather than to the object: it does not follow from the fact that all knowledge of others is dialectical, that others are wholly dialectical in every respect. By making analytical reason an anti-comprehension, Sartre often comes to refuse it any reality as an integral part of the object of comprehension. This paralogism is already apparent in his manner of invoking history, for one is hard put to it to see whether it is meant to be the history men make unconsciously, history of men consciously made by historians, the philosopher’s interpretation of the history of men or his interpretation of the history of historians. The difficulty becomes even greater, however, when Sartre endeavours to explain the life and thought of the present or past members not of his own society but of exotic societies.

He thinks, rightly, that this attempted comprehension stands no chance of succeeding unless it is dialectical; and he concludes, wrongly, that the relationship between native thought and his knowledge of it, is that of a constitutive to a constituted dialectic, and thus, by an unforeseen detour, he repeats all the illusions of theorists of primitive mentality on his own account. It seems even less tolerable to him than to Levy-Bruhl that the savage should possess ‘complex understanding’ and should be capable of analysis and demonstration. Of the Ambrym native, made famous by Deacon’s work, who was able to show the field-worker the functioning of his marriage rules and kinship system by a diagram in the sand (an aptitude in no way exceptional as plenty of similar cases are recorded in ethnographic literature) Sartre says: ‘It goes without saying that this construction is not a thought: it is a piece of manual work governed by unexpressed synthetical knowledge’. Granted: but then the same must be said of a professor at the Ecole Polytechnique demonstrating a proof on the blackboard, for every ethnographer capable of dialectical comprehension is intimately persuaded that the situation is exactly the same in both cases. So it would follow that all reason is dialectical, which for my part I am prepared to concede, since dialectical reason seems to me like analytical reason in action; but then the distinction between the two forms of reason which is the basis of Sartre’s enterprise would become pointless.

I must now confess to having myself unintentionally and unwittingly lent support to these erroneous ideas, by having seemed all too often in Les structures élémentaires de la parenté as if I were seeking out an unconscious genesis of matrimonial exchange. I should have made more distinction between exchange as it is expressed spontaneously and forcefully in the praxis of groups and the conscious and deliberate rules by which these same groups – or their philosophers – spend their time in codifying and controlling it. If there is anything to be learnt from the ethnographic enquiries of the last twenty years, it is that this latter aspect is much more important than has generally been realized by observers, who labour under the same delusion as Sartre. Thus we must, as Sartre advocates, apply dialectical reason to the knowledge of our own and other societies. But we must not lose sight of the fact that analytical reason occupies a considerable place in all of them and that, as it is present, the approach we adopt must also allow us to rediscover it there.

But even were it not present, Sartre’s position would not be improved. For in this case exotic societies would merely confront us, in a more general manner than others, with an unconscious teleology, which, although historical, completely eludes human history: that of which certain aspects are revealed by linguistics and psycho-analysis and which rests on the interplay of biological mechanisms (structure of the brain, lesions, internal secretions) and psychological ones. There, it seems to me, is ‘the bone’ (to borrow a phrase from Sartre) which his critique does not manage to break, and moreover cares nothing about, which is the most serious charge one could level at it. For language does not consist in the analytical reason of the old-style grammarians nor in the dialectic constituted by structural linguistics nor in the constitutive dialectic of individual praxis facing the practico-inert, since all three presuppose it. Linguistics thus presents us with a dialectical and totalizing entity but one outside (or beneath) consciousness and will. Language, an unreflecting totalization, is human reason which has its reasons and of which man knows nothing. And if it is objected that it is so only for a subject who internalizes it on the basis of linguistic theory, my reply is that this way out must be refused, for this subject is one who speaks: for the same light which reveals the nature of language to him also reveals to him that it was so when he did not know it, for he already made himself understood, and that it will remain so tomorrow without his being aware of it, since his discourse never was and never will be the result of a conscious totalization of linguistic laws. But if, as speaking subject, man can find his apodictic experience in an other totalization, there seems no longer any reason why, as living subject, he should not have access to the same experience in other, not necessarily human, but living beings.

This method could also lay claim to the name ‘progressive-regressive’; in fact, what Sartre describes as such is the very method anthropologists have been practising for many years. But Sartre restricts it to its preliminary step. For our method is progressive-regressive not once but twice over. In the first stage, we observe the datum of experience, analyse it in the present, try to grasp its historical antecedents as far as we can delve into the past, and bring all these facts back to the light of day to incorporate them into a meaningful totality. The second stage, which repeats the first on a different plane and at a different level, then begins. This internalized human thing which we have sought to provide with all its wealth and originality, only fixes the distance analytical reason must cover, the leap it must make, to close the gap between the ever unforeseen complexity of this new object and the intellectual means at its disposal. It must therefore transform itself as dialectical reason, in the hope that once flexible, widened and strengthened, by its agency this unforeseen object will be assimilated to others, this novel totality will be merged into other totalities and that thus little by little clambering on to the mass of its conquests, dialectical reason will descry other horizons and other objects. No doubt the procedure would go astray if it were not, at every stage and, above all, when it seemed to have run its course, ready to retrace its steps and to double back on itself to preserve the contact with that experienced totality which serves both as its end and means. This return on itself is in my view a verification, rather than, as Sartre regards it, a demonstration, for, as I see it, a conscious being aware of itself as such poses a problem to which it provides no solution. The discovery of the dialectic subjects analytical reason to an imperative requirement: to account also for dialectical reason. This standing requirement relentlessly forces analytical reason to extend its programme and transform its axiomatic. But dialectical reason can account neither for itself nor for analytical reason.

It will be objected that this expansion is illusory since it is always accompanied by a contraction in meaning, and we should abandon the substance for the shadow, clarity for obscurity, the manifest for the conjectural, truth for science fiction. Again, Sartre would have to show that he himself avoids this dilemma, inherent in every attempt at explanation. The real question is not whether our endeavour to understand involves a gain or a loss of meaning, but whether the meaning we preserve is of more value than that we have been judicious enough to relinquish. In this respect Sartre seems to have remembered only half of Marx’s and Freud’s combined lesson. They have taught us that man has meaning only on the condition that he view himself as meaningful. So far I agree with Sartre. But it must be added that this meaning is never the right one: superstructures are faulty acts which have ‘made it’ socially. Hence it is vain to go to historical consciousness for the truest meaning. What Sartre calls dialectical reason is only a reconstruction, by what he calls analytical reason, of hypothetical moves about which it is impossible to know – unless one should perform them without thinking them – whether they bear any relation at all to what he tells us about them and which, if so, would be definable in terms of analytical reason alone. And so we end up in the paradox of a system which invokes the criterion of historical consciousness to distinguish the ‘primitive’ from the ‘civilized’ but – contrary to its claim – is itself ahistorical. It offers not a concrete image of history but an abstract schema of men making history of such a kind that it can manifest itself in the trend of their lives as a synchronic totality. Its position in relation to history is therefore the same as that of primitives to the eternal past: in Sartre’s system, history plays exactly the part of a myth.

Indeed, the problem raised by the Critique de la raison dialectique is reducible to the question: under what conditions is the myth of the French Revolution possible? And I am prepared to grant that the contemporary Frenchman must believe in this myth in order fully to play the part of an historical agent and also that Sartre’s analysis admirably extracts the set of formal conditions necessary if this result is to be secured. But it does not follow that his meaning, just because it is the richest (and so most suited to inspire practical action), should be the truest. Here the dialectic turns against itself. This truth is a matter of context, and if we place ourselves outside it – as the man of science is bound to do – what appeared as an experienced truth first becomes confused and finally disappears altogether. The so-called men of the Left still cling to a period of contemporary history which bestowed the blessing of a congruence between practical imperatives and schemes of interpretation. Perhaps this golden age of historical consciousness has already passed; and that this eventuality can at any rate be envisaged proves that what we have here is only a contingent context like the fortuitous ‘focusing’ of an optical instrument when its object-glass and eye-piece move in relation to each other. We are still ‘in focus’ so far as the French Revolution is concerned, but so we should have been in relation to the Fronde had we lived earlier. The former will rapidly cease to afford a coherent image on which our action can be modelled, just as the latter has already done. What we learn from reading Retz is that thought is powerless to extract a scheme of interpretation from events long past.

At first sight, there seems no doubt: on one side the privileged, on the other the humble and exploited; how could we hesitate? We are Frondeurs. However, the people of Paris were being manoeuvred by noble houses, whose sole aim was to arrange their own affairs with the existing powers, and by one half of the royal family which wanted to oust the other. And now we are already only half Frondeurs. As for the Court, which took refuge at Saint-Germain, it appears at first to have been a faction of good for nothings vegetating on their privileges and growing fat on exactions and usury at the expense of the collectivity. But no, it had a function all the same since it retained military power; it conducted the struggle against foreigners, the Spaniards, whom the Frondeurs invited without hesitation to invade the country and impose their wills on this same Court which was defending the fatherland. The scales, however, tilt the other way again: the Frondeurs and Spaniards together formed the party of peace. The Prince de Conde and the Court only sought warlike adventures. We are pacifists and once again become Frondeurs. But nevertheless did not the military exploits of Mazarin and the Court extend France to its present frontiers, thus founding the state and the nation? Without them we should not be what we are today. So here we are on the other side again.

It suffices therefore for history to move away from us in time or for us to move away from it in thought, for it to cease to be internalizable and to lose its intelligibility, a spurious intelligibility attaching to a temporary internality. I am not however suggesting that man can or should sever himself from this internality. It is not in his power to do so and wisdom consists for him in seeing himself live it, while at the same time knowing (but in a different register) that what he lives so completely and intensely is a myth – which will appear as such to men of a future century, and perhaps to himself a few years hence, and will no longer appear at all to men of a future millennium. All meaning is answerable to a lesser meaning, which gives it its highest meaning, and if this regression finally ends in recognizing ‘a contingent law of which one can say only: it is thus, and not otherwise’ (Sartre), this prospect is not alarming to those whose thought is not tormented by transcendence even in a latent form. For man will have gained all he can reasonably hope for if, on the sole condition of bowing to this contingent law, he succeeds in determining his form of conduct and in placing all else in the realm of the intelligible.

Sartre is certainly not the only contemporary philosopher to have valued history above the other human sciences and formed an almost mystical conception of it. The anthropologist respects history, but he does not accord it a special value. He conceives it as a study complementary to his own: one of them unfurls the range of human societies in time, the other in space. And the difference is even less great than it might seem, since the historian strives to reconstruct the picture of vanished societies as they were at the points which for them corresponded to the present, while the ethnographer does his best to reconstruct the historical stages which temporally preceded their existing form.

This symmetry between history and anthropology seems to be rejected by philosophers who implicitly or explicitly deny that distribution in space and succession in time afford equivalent perspectives. In their eyes some special prestige seems to attach to the temporal dimension, as if diachrony were to establish a kind of intelligibility not merely superior to that provided by synchrony, but above all more specifically human.

It is easy to explain, if not to justify, this preference. The diversity of social forms, which the anthropologist grasps as deployed in space, present the appearance of a discontinuous system. Now, thanks to the temporal dimension, history seems to restore to us, not separate states, but the passage from one state to another in a continuous form. And as we believe that we apprehend the trend of our personal history as a continuous change, historical knowledge appears to confirm the evidence of inner sense. History seems to do more than describe beings to us from the outside, or at best give us intermittent flashes of insight into internalities, each of which are so on their own account while remaining external to each other: it appears to re-establish our connection, outside ourselves, with the very essence of change.

There would be plenty to say about this supposed totalizing continuity of the self which seems to me to be an illusion sustained by the demands of social life – and consequently a reflection of the external on the internal – rather than the object of an apodictic experience. But there is no need to resolve this philosophical problem in order to perceive that the proposed conception of history corresponds to no kind of reality. As historical knowledge is claimed to be privileged, I feel entitled (as I would not otherwise feel) to make the point that there is a twofold antinomy in the very notion of an historical fact. For, ex hypothesi, a historical fact is what really took place, but where did anything take place? Each episode in a revolution or a war resolves itself into a multitude of individual psychic movements. Each of these movements is the translation of unconscious development, and these resolve themselves into cerebral, hormonal or nervous phenomena, which themselves have reference to the physical or chemical order. Consequently, historical facts are no more given than any other. It is the historian, or the agent of history, who constitutes them by abstraction and as though under the threat of an infinite regress.

What is true of the constitution of historical facts is no less so of their selection. From this point of view, the historian and the agent of history choose, sever and carve them up, for a truly total history would confront them with chaos. Every corner of space conceals a multitude of individuals each of whom totalizes the trend of history in a manner which cannot be compared to the others; for any one of these individuals, each moment of time is inexhaustibly rich in physical and psychical incidents which all play their part in his totalization. Even history which claims to be universal is still only a juxtaposition of a few local histories within which (and between which) very much more is left out than is put in. And it would be vain to hope that by increasing the number of collaborators and making research more intensive one would obtain a better result. In so far as history aspires to meaning, it is doomed to select regions, periods, groups of men and individuals in these groups and to make them stand out, as discontinuous figures, against a continuity barely good enough to be used as a backdrop. A truly total history would cancel itself out – its product would be nought. What makes history possible is that a sub-set of events is found, for a given period, to have approximately the same significance for a contingent of individuals who have not necessarily experienced the events and may even consider them at an interval of several centuries. History is therefore never history, but history-for.[4] It is partial in the sense of being biased even when it claims not to be, for it inevitably remains partial – that is, incomplete – and this is itself a form of partiality. When one proposes to write a history of the French Revolution one knows (or ought to know) that it cannot, simultaneously and under the same heading, be that of the Jacobin and that of the aristocrat. Ex hypothesi, their respective totalizations (each of which is anti-symmetric to the other) are equally true. One must therefore choose between two alternatives. One must select as the principal either one or a third (for there are an infinite number of them) and give up the attempt to find in history a totalization of the set of partial totalizations; or alternatively one must recognize them all as equally real: but only to discover that the French Revolution as commonly conceived never took place.

History does not therefore escape the common obligation of all knowledge, to employ a code to analyse its object, even (and especially) if a continuous reality is attributed to that object.[5] The distinctive features of historical knowledge are due not to the absence of a code, which is illusory, but to its particular nature: the code consists in a chronology. There is no history without dates. To be convinced of this it is sufficient to consider how a pupil succeeds in learning history: he reduces it to an emaciated body, the skeleton of which is formed by dates. Not without reason, there has been a reaction against this dry method, but one which often runs to the opposite extreme. Dates may not be the whole of history, nor what is most interesting about it, but they are its sine qua non, for history’s entire originality and distinctive nature lie in apprehending the relation between before and after, which would perforce dissolve if its terms could not, at least in principle, be dated.

Now, this chronological coding conceals a very much more complex nature than one supposes when one thinks of historical dates as a simple linear series. In the first place, a date denotes a moment in a succession: d2 is after d1 and before d3. From this point of view dates only perform the function of ordinal numbers. But each date is also a cardinal number and, as such, expresses a distance in relation to the dates nearest to it. We use a large number of dates to code some periods of history; and fewer for others. This variable quantity of dates applied to periods of equal duration are a gauge of what might be called the pressure of history: there are ‘hot’ chronologies which are those of periods where in the eyes of the historian numerous events appear as differential elements; others, on the contrary, where for him (although not of course for the men who lived through them) very little or nothing took place. Thirdly and most important, a date is a member of a class. These classes of dates are definable by the meaningful character each date has within the class in relation to other dates which also belong to it, and by the absence of this meaningful character with respect to dates appertaining to a different class. Thus the date 1685 belongs to a class of which 1610, 1648 and 1715 are likewise members; but it means nothing in relation to the class composed of the dates: 1st, 2nd, 3rd, 4th millennium, nor does it mean anything in relation to the class of dates: 23 January, 17 August, 30 September, etc.

On this basis, in what would the historian’s code consist? Certainly not in dates, since these are not recurrent. Changes of temperature can be coded with the help of figures, because the reading of a figure on the thermometer evokes the return of an earlier situation: whenever I read 0°C, I know that it is freezing and put on my warmest coat. But a historical date, taken in itself, would have no meaning, for it has no reference outside itself: if I know nothing about modern history, the date 1643 makes me none the wiser. The code can therefore consist only of classes of dates, where each date has meaning in as much as it stands in complex relations of correlation and opposition with other dates. Each class is defined by a frequency, and derives from what might be called a corpus or a domain of history. Historical knowledge thus proceeds in the same way as a wireless with frequency modulation: like a nerve, it codes a continuous quantity – and as such an asymbolic one – by frequencies of impulses proportional to its variations. As for history itself, it cannot be represented as an aperiodic series with only a fragment of which we are acquainted. History is a discontinuous set composed of domains of history, each of which is defined by a characteristic frequency and by a differential coding of before and after. It is no more possible to pass between the dates which compose the different domains than it is to do so between natural and irrational numbers. Or more precisely: the dates appropriate to each class are irrational in relation to all those of other classes.

It is thus not only fallacious but contradictory to conceive of the historical process as a continuous development, beginning with prehistory coded in tens or hundreds of millennia, then adopting the scale of millennia when it gets to the 4th or 3rd millennium, and continuing as history in centuries interlarded, at the pleasure of each author, with slices of annual history within the century, day to day history within the year or even hourly history within a day. All these dates do not form a series: they are of different species. To give just one example, the coding we use in prehistory is not preliminary to that we employ for modern and contemporary history. Each code refers to a system of meaning which is, at least in theory, applicable to the virtual totality of human history. The events which are significant for one code are no longer so for another. Coded in the system of prehistory, the most famous episodes in modern and contemporary history cease to be pertinent; except perhaps (and again we know nothing about it) certain massive aspects of demographic evolution viewed on a world-wide scale, the invention of the steam-engine, the discovery of electricity and of nuclear energy.

Given that the general code consists not in dates which can be ordered as a linear series but in classes of dates each furnishing an autonomous system of reference, the discontinuous and classificatory nature of historical knowledge emerges clearly. It operates by means of a rectangular matrix,


where each line represents classes of dates, which may be called hourly, daily, annual, secular, millennial for the purposes of schematization and which together make up a discontinuous set. In a system of this type, alleged historical continuity is secured only by dint of fraudulent outlines.

Furthermore, although the internal gaps in each class cannot be filled in by recourse to other classes, each class taken as a whole nevertheless always refers back to another class, which contains the principle of an intelligibility to which it could not itself aspire. The history of the 17th century is ‘annual’ but the 17th century, as a domain of history belongs to another class, which codes it in relation to earlier and later centuries; and this domain of modern times in its turn becomes an element of a class where it appears correlated with and opposed to other ‘times’: the middle ages, antiquity, the present day, etc. Now, these various domains correspond to histories of different power.

Biographical and anecdotal history, right at the bottom of the scale, is low-powered history, which is not intelligible in itself and only becomes so when it is transferred en bloc to a form of history of a higher power than itself; and the latter stands in the same relation to a class above it. It would, however, be a mistake to think that we progressively reconstitute a total history by dint of these dovetailings. For any gain on one side is offset by a loss on the other. Biographical and anecdotal history is the least explanatory; but it is the richest in point of information, for it considers individuals in their particularity and details for each of them the shades of character, the twists and turns of their motives, the phases of their deliberations. This information is schematized, put in the background and finally done away with as one passes to histories of progressively greater ‘power’.[6] Consequently, depending on the level on which he places himself, the historian loses in information what he gains in comprehension or vice versa, as if the logic of the concrete wished to remind us of its logical nature by modelling a confused outline of Gödel’s theorem in the clay of ‘becoming’. The historian’s relative choice, with respect to each domain of history he gives up, is always confined to the choice between history which teaches us more and explains less, and history which explains more and teaches less. The only way he can avoid the dilemma is by getting outside history: either by the bottom, if the pursuit of information leads him from the consideration of groups to that of individuals and then to their motivations which depend on their personal history and temperament, that is to say to an infra-historical domain in the realms of psychology and physiology; or by the top, if the need to understand incites him to put history back into prehistory and the latter into the general evolution of organized beings, which is itself explicable only in terms of biology, geology and finally cosmology.

There is, however, another way of avoiding the dilemma without thereby doing away with history. We need only recognize that history is a method with no distinct object corresponding to it to reject the equivalence between the notion of history and the notion of humanity which some have tried to foist on us with the unavowed aim of making historicity the last refuge of a transcendental humanism: as if men could regain the illusion of liberty on the plane of the ‘we’ merely by giving up the ‘I’s that are too obviously wanting in consistency.

In fact history is tied neither to man nor to any particular object. It consists wholly in its method, which experience proves to be indispensable for cataloguing the elements of any structure whatever, human or non-human, in their entirety. It is therefore far from being the case that the search for intelligibility comes to an end in history as though this were its terminus. Rather, it is history that serves as the point of departure in any quest for intelligibility. As we say of certain careers, history may lead to anything, provided you get out of it.

This further thing to which history leads for want of a sphere of reference of its own shows that whatever its value (which is indisputable) historical knowledge has no claim to be opposed to other forms of knowledge as a supremely privileged one. We noted above that it is already found rooted in the savage mind, and we can now see why it does not come to fruition there. The characteristic feature of the savage mind is its timelessness; its object is to grasp the world as both a synchronic and a diachronic totality and the knowledge which it draws therefrom is like that afforded of a room by mirrors fixed on opposite walls, which reflect each other (as well as objects in the intervening space) although without being strictly parallel. A multitude of images forms simultaneously, none exactly like any other, so that no single one furnishes more than a partial knowledge of the decoration and furniture but the group is characterized by invariant properties expressing a truth. The savage mind deepens its knowledge with the help of imagines mundi. It builds mental structures which facilitate an understanding of the world in as much as they resemble it. In this sense savage thought can be defined as analogical thought.

But in this sense too it differs from domesticated thought, of which historical knowledge constitutes one aspect. The concern for continuity which inspires the latter is indeed a manifestation, in the temporal order, of knowledge which is interstitial and unifying rather than discontinuous and analogical: instead of multiplying objects by schemes promoted to the role of additional objects, it seems to transcend an original discontinuity by relating objects to one another. But it is this reason, wholly concerned with closing gaps and dissolving differences, which can properly be called ‘analytical’. By a paradox on which much stress has recently been laid, for modern thought ‘continuity, variability, relativity, determinism go together’.

This analytic, abstract continuity will doubtless be opposed to that of the praxis as concrete individuals live it. But this latter continuity seems no less derivative than the former, for it is only the conscious mode of apprehending psychological and physiological processes which are themselves discontinuous. I am not disputing that reason develops and transforms itself in the practical field: man’s mode of thought reflects his relations to the world and to men. But in order for praxis to be living thought, it is necessary first (in a logical and not a historical sense) for thought to exist: that is to say, its initial conditions must be given in the form of an objective structure of the psyche and brain without which there would be neither praxis nor thought.

When therefore I describe savage thought as a system of concepts embedded in images, I do not come anywhere near the robinsonnades of a constitutive constituent dialectic: all constitutive reason presupposes a constituted reason. But even if one allowed Sartre the circularity which he invokes to dispel the ‘suspect character’ attaching to the first stages of his synthesis, what he proposes really are ‘robinsonnades’, this time in the guise of descriptions of phenomena, when he claims to restore the sense of marriage exchange, the potlatch or the demonstration of his tribe’s marriage rules by a Melanesian savage. Sartre then refers to a comprehension which has its being in the praxis of their organizers, a bizarre expression to which no reality corresponds, except perhaps the capacity which any foreign society presents to anyone looking at it from the outside, and which leads him to project the lacunae in his own observation on to it in the form of positive attributes. Two examples will show what I mean.

No anthropologist can fail to be struck by the common manner of conceptualizing initiation rites employed by the most diverse societies throughout the world. Whether in Africa, America, Australia or Melanesia, the rites follow the same pattern: first, the novices, taken from their parents, are symbolically ‘killed’ and kept hidden in the forest or bush where they are put to the test by the Beyond; after this they are ‘reborn’ as members of the society. When they are returned to their natural parents, the latter therefore simulate all the phases of a new delivery, and begin a re-education even in the elementary actions of feeding or dressing. It would be tempting to interpret this set of phenomena as a proof that at this stage thought is wholly embedded in praxis. But this would be seeing things back to front, for it is on the contrary scientific praxis which, among ourselves, has emptied the notions of death and birth of everything not corresponding to mere physiological processes and rendered them unsuitable to convey other meanings. In societies with initiation rites, birth and death provide the material for a rich and varied conceptualization, provided that these notions (like so many others) have not been stripped by any form of scientific knowledge oriented towards practical returns – which they lack – of the major part of a meaning which transcends the distinction between the real and the imaginary: a complete meaning of which we can now hardly do more than evoke the ghost in the reduced setting of figurative language. What looks to us like being embedded in praxis is the mark of thought which quite genuinely takes the words it uses seriously, whereas in comparable circumstances we only ‘play’ on words.

The taboos on parents-in-law furnish the matter for a cautionary tale leading to the same conclusion by a different route. Anthropologists have found the frequent prohibition of any physical or verbal contact between close affines so strange that they have exercised their ingenuity in multiplying explanatory hypotheses, without always making sure that the hypotheses are not rendered redundant by one another. Elkin for instance explains the rarity of marriage with the patrilateral cousin in Australia by the rule that as a man has to avoid any contact with his mother-in-law, he will be wise to choose the latter among women entirely outside his own local group (to which his father’s sisters belong). The aim of the rule itself is supposed to be to prevent a mother and daughter from being rivals for the affections of the same man; finally, the taboo is supposed to be extended by contamination to the wife’s maternal grandmother and her husband. There are thus four concurrent interpretations of a single phenomenon: as a function of a type of marriage, as the result of a psychological calculation, as protection against instinctive tendencies and as the product of association by contiguity. This, however, still does not satisfy Elkin, for in his view the taboo on the father-in-law rests on a fifth explanation: the father-in-law is the creditor of the man to whom he has given his daughter, and the son-in-law feels himself to be in a position of inferiority in relation to him.

I shall content myself with the last explanation which perfectly covers all the cases considered and renders the others worthless by bringing out their naivety. But why is it so difficult to put these usages into their proper place? The reason is, I think, that the usages of our own society which could be compared with them and might furnish a landmark to identify them by, are in a dissociated form among ourselves, while in these exotic societies they appear in an associated one which makes them unrecognizable to us.

We are acquainted with the taboo on parents-in-law or at least with its approximate equivalent. By the same token we are forbidden to address the great of this world and obliged to keep out of their way. All protocol asserts it: one does not speak first to the Queen of England or the President of the French Republic; and we adopt the same reserve when unforeseen circumstances create conditions of closer proximity between a superior and ourselves than the social distance between us warrants. Now, in most societies the position of wife-giver is accompanied by social (and sometimes also economic) superiority, that of wife-taker by inferiority and dependence. This inequality between affines may be expressed objectively in institutions as a fluid or stable hierarchy, or it may be expressed subjectively in the system of interpersonal relations by means of privileges and prohibitions.

There is therefore nothing mysterious about these usages which our own experience enables us to see from the inside. We are disconcerted only by their constitutive conditions, different in each case. Among ourselves, they are sharply separated from other usages and tied to an unambiguous context. In exotic societies, the same usages and the same context are as it were embedded in other usages and a different context: that of family ties, with which they seem to us incompatible. We find it hard to imagine that, in private, the son-in-law of the President of the French Republic should regard him as the head of the state rather than as his father-in-law. And although the Queen of England’s husband may behave as the first of her subjects in public, there are good reasons for supposing that he is just a husband when they are alone together. It is either one or the other. The superficial strangeness of the taboo on parents-in-law arises from its being both at the same time.

Consequently, and as we have already found in the case of operations of understanding, the system of ideas and attitudes is here presented only as embodied. Considered in itself, this system has nothing about it to baffle the anthropologist. My relation to the President of the Republic is made up entirely of negative observances, since, in the absence of other ties, any relations we may have are wholly defined by the rule that I should not speak unless he invites me to do so and that I should remain a respectful distance from him. But this abstract relation need only be clothed in a concrete relation and the attitudes appropriate to each to accumulate, for me to find myself as embarrassed by my family as an Australian aborigine. What appears to us as greater social ease and greater intellectual mobility is thus due to the fact that we prefer to operate with detached pieces, if not indeed with ‘small change’, while the native is a logical hoarder: he is forever tying the threads, unceasingly turning over all the aspects of reality, whether physical, social or mental. We traffic in our ideas; he hoards them up. The savage mind puts a philosophy of the finite into practice.

This is also the source of the renewed interest in it. This language with its limited vocabulary able to express any message by combinations of oppositions between its constitutive units, this logic of comprehension for which contents are indissociable from form, this systematic of finite classes, this universe made up of meanings, no longer appears to us as retrospective witnesses of a time when: ‘... le ciel sur la terre Marchait et respirait dans un peuple de dieux[when heaven walked and breathed on earth among a population of Gods], and which the poet only evokes for the purpose of asking whether or not it is to be regretted. This time is now restored to us, thanks to the discovery of a universe of information where the laws of savage thought reign once more: ‘heaven’ too, ‘walking on earth’ among a population of transmitters and receivers whose messages, while in transmission, constitute objects of the physical world and can be grasped both from without and from within.

The idea that the universe of primitives (or supposedly such) consists principally in messages is not new. But until recently a negative value was attributed to what was wrongly taken to be a distinctive characteristic, as though this difference between the universe of the primitives and our own contained the explanation of their mental and technological inferiority, when what it does is rather to put them on a par with modern theorists of documentation.[7] Physical science had to discover that a semantic universe possesses all the characteristics of an object in its own right for it to be recognized that the manner in which primitive peoples conceptualize their world is not merely coherent but the very one demanded in the case of an object whose elementary structure presents the picture of a discontinuous complexity.

The false antinomy between logical and prelogical mentality was surmounted at the same time. The savage mind is logical in the same sense and the same fashion as ours, though as our own is only when it is applied to knowledge of a universe in which it recognizes physical and semantic properties simultaneously. This misunderstanding once dispelled, it remains no less true that, contrary to Levy-Bruhl’s opinion, its thought proceeds through understanding, not affectivity, with the aid of distinctions and oppositions, not by confusion and participation. Although the term had not yet come into use, numerous texts of Durkheim and Mauss show that they understood that so-called primitive thought is a quantified form of thought.

It will be objected that there remains a major difference between the thought of primitives and our own: Information Theory is concerned with genuine messages whereas primitives mistake mere manifestations of physical determinism for messages. Two considerations, however, deprive this argument of any weight. In the first place, Information Theory has been generalized, and it extends to phenomena not intrinsically possessing the character of messages, notably to those of biology; the illusions of totemism have had at least the merit of illuminating the fundamental place belonging to phenomena of this order, in the internal economy of systems of classification. In treating the sensible properties of the animal and plant kingdoms as if they were the elements of a message, and in discovering ‘signatures’ – and so signs – in them, men have made mistakes of identification: the meaningful element was not always the one they supposed. But, without perfected instruments which would have permitted them to place it where it most often is namely, at the microscopic level – they already discerned ‘as through a glass darkly’ principles of interpretation whose heuristic value and accordance with reality have been revealed to us only through very recent inventions: telecommunications, computers and electron microscopes.

Above all, during the period of their transmission, when they have an objective existence outside the consciousness of transmitters and receivers, messages display properties which they have in common with the physical world. Hence, despite their mistakes with regard to physical phenomena (which were not absolute but relative to the level where they grasped them) and even though they interpreted them as if they were messages, men were nevertheless able to arrive at some of their properties. For a theory of information to be able to be evolved it was undoubtedly essential to have discovered that the universe of information is part of an aspect of the natural world. But the validity of the passage from the laws of nature to those of information once demonstrated, implies the validity of the reverse passage – that which for millennia has allowed men to approach the laws of nature by way of information.

Certainly the properties to which the savage mind has access are not the same as those which have commanded the attention of scientists. The physical world is approached from opposite ends in the two cases: one is supremely concrete, the other supremely abstract; one proceeds from the angle of sensible qualities and the other from that of formal properties. But if, theoretically at least and on condition no abrupt changes in perspective occurred, these two courses might have been destined to meet, this explains that they should have both, independently of each other in time and space, led to two distinct though equally positive sciences: one which flowered in the neolithic period, whose theory of the sensible order provided the basis of the arts of civilization (agriculture, animal husbandry, pottery, weaving, conservation and preparation of food, etc.) and which continues to provide for our basic needs by these means; and the other, which places itself from the start at the level of intelligibility, and of which contemporary science is the fruit.

We have had to wait until the middle of this century for the crossing of long separated paths: that which arrives at the physical world by the detour of communication, and that which as we have recently come to know, arrives at the world of communication by the detour of the physical. The entire process of human knowledge thus assumes the character of a closed system. And we therefore remain faithful to the inspiration of the savage mind when we recognize that the scientific spirit in its most modern form will, by an encounter it alone could have foreseen, have contributed to legitimize the principles of savage thought and to re-establish it in its rightful place.

16 October 1961


1. The opposition between nature and culture to which I attached much importance at one time (1, ch. 1 and 2) now seems to be of primarily methodological importance.

2. This even holds for mathematical truths of which a contemporary logician, however, says that ‘The characteristic of mathematical thought is that it does not convey truth about the external world’ (Heyting, pp. 8-9). But mathematical thought at any rate reflects the free functioning of the mind, that is, the activity of the cells of the cerebral cortex, relatively emancipated from any external constraint and obeying only its own laws. As the mind too is a thing, the functioning of this thing teaches us something about the nature of things: even pure reflection is in the last analysis an internalization of the cosmos. It illustrates the structure of what lies outside in a symbolic form: ‘Logic and logistics are empirical sciences belonging to ethnography rather than psychology’ (Beth, p. 151).

3. It is precisely because all these aspects of the savage mind can be discovered in Sartre’s philosophy, that the latter is in my view unqualified to pass judgment on it: he is prevented from doing so by the very fact of furnishing its equivalent. To the anthropologist, on the contrary, this philosophy (like all the others) affords a first-class ethnographic document, the study of which is essential to an understanding of the mythology of our own time.

4. Quite so, will be the comment of the supporters of Sartre. But the latter’s whole endeavour shows that, though the subjectivity of history-for-me can make way for the objectivity of history-for-us, the ‘I’ can still only be converted into ‘we’ by condemning this ‘we’ to being no more than an ‘I’ raised to the power of two, itself hermetically sealed off from the other ‘we’s. The price so paid for the illusion of having overcome the insoluble antinomy (in such a system) between my self and others, consists of the assignation, by historical consciousness, of the metaphysical function of Other to the Papuans. By reducing the latter to the state of means, barely sufficient for its philosophical appetite, historical reason abandons itself to a sort of intellectual cannibalism much more revolting to the anthropologist than real cannibalism.

5. In this sense too, one can speak of an antinomy of historical knowledge: if it claims to reach the continuous it is impossible, being condemned to an infinite regress; but to render it possible, events must be quantified and thereafter temporality ceases to be the privileged dimension of historical knowledge because as soon as it is quantified each event can, for all useful purposes, be treated as if it were the result of a choice between possible pre-existents.

6. Each domain of history is circumscribed in relation to that immediately below it, inscribed in relation to that above it. So each low-powered history of an inscribed domain is complementary to the powerful history of the circumscribed domain and contradictory to the low-powered history of this same domain (in so far as it is itself an inscribed domain). Each history is thus accompanied by an indeterminate number of anti-histories, each complementary to the others: to a history of grade 1, there corresponds an anti-history of grade 2, etc. The progress of knowledge and the creation of new sciences take place through the generation of anti-histories which show that a certain order which is possible only on one plane ceases to be so on another. The anti-history of the French Revolution envisaged by Gobineau is contradictory on the plane on which the Revolution had been thought of before him. It becomes logically conceivable (which does not mean that it is true) if one puts oneself on a new plane, which incidentally Gobineau chose clumsily; that is to say: if one passes from a history of ‘annual’ or ‘secular’ grade (which is also political, social and ideological) to a history of ‘millennial’ or ‘multi-millennial’ grade (which is also cultural and anthropological), a procedure not invented by Gobineau which might be called: Boulainvilliers’ ‘transformation’.

7. The documentalist neither disallows nor disputes the substance of the works he analyses in order to derive the constitutive units of his code or to adapt them, either by combining them among themselves or, if necessary, decomposing them into finer units. He thus treats the authors as gods whose revelations are written down, instead of being inscribed into beings and things, but which have the same sacred value, which attaches to the supremely meaningful character that, for methodological or ontological reasons, it is ex hypothesi necessary to recognize in them in both cases.