Lévi-Strauss; drawing by David Levine

The problem is, Just where do we fit in? Are we better or worse or indeed in any way different from our prehistoric ancestors or our primitive contemporaries? We are animals and, therefore, a part of Nature, but we are also self-conscious human beings who can somehow or other conceive of ourselves as outside observers, looking on. And then again the process we call “thinking” is quite clearly something that goes on inside our heads, a function of electro-chemical processes in the brain, yet it is also a response to signals which we receive from outside through our senses. How can I believe that I can think about Nature, when quite clearly thinking is a phenomenon in which I and Nature interact?

Claude Lévi-Strauss is a prolific writer, but his commentators are even more so and, as each new volume appears, the whole business becomes more and more a private dialogue for the initiated. It is esoteric stuff, and although this review is addressed to a more general reader it can scarcely avoid some touches of gobbledegook.

The Savage Mind is a translation, in plain cover, of La Pensée sauvage (1962) which started out in a baffling jacket illustrated with a picture of wild pansies—“there is pansies, that’s for thoughts.” The translation is anonymous, but it is common knowledge that it has proved a publisher’s nightmare. At least three distinguished hands had their share in the final product. Professor Geertz has declared the result to be “execrable”; Dr. Rodney Needham, on the other hand, reports that “a word-by-word collation of the first chapter with the French text yields only a few inconsequential slips.” Perhaps both are right. A passage such as:

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 [p. 264]

can hardly be said to make much sense though it is a literally exact transformation of the corresponding passage at p. 349 of the French original. Perhaps you have to be an initiate even to understand the French. All the same, I do feel that the translators have fuzzed things up quite unnecessarily in places. Take the title, for example. The obvious translation of La Pensée sauvage is “savage thought”—as in the quotation above—whereas The Savage Mind recalls Lévi-Strauss’s much used, but decidedly ambiguous, l’esprit humain, which might suggest that the book is about metaphysics, whereas, in fact, it is about logic.

THE FUNDAMENTAL THEME of this book is that we are at fault if we follow Lévy-Bruhl (and by derivation Sartre) in thinking that there is an historical contrast between the “pre-logical” mentality of Primitives and the “logical” mentality of Modern Man. Primitive people are no more mystical in their approach to reality than we are. The distinction rather is between a logic which is constructed out of observed contrasts in the sensory qualities of concrete objects—e.g., the difference between raw and cooked, wet and dry, male and female—and a logic which depends upon the formal contrasts of entirely abstract entities—e.g. + and – of logex and ex The latter kind of logic, which even in our own society, is used only by highly specialized experts is a different way of talking about the same kind of thing. Primitive thought differs from scientific thought much as the use of an abacus differs from mental arithmetic, but the fact that, in our present age, we are coming to depend on things outside ourselves—such as computers—to help us with our problems of communication and calculation makes this an appropriate moment to examine the way in which primitive peoples likewise are able to make sense of the events of daily life by reference to codes composed of things outside themselves—such as the attributes of animal species. Lévi-Strauss’s investigation is complicated, ingenious, and, in my view anyway, persuasive, though the uninitiated will certainly find it easier going if they take a preliminary bite at the cherry by first tackling Le Totémisme aujourd’hui (1962; English version by Rodney Needham, Totemism, 1964).

What excites the anthropologist is the way in which Lévi-Strauss sees that quite disparate kinds of ethnographic fact, culled from entirely different geographic regions, are ordered according to the same kind of logical principles. Thus Indian caste systems consist of endogamous groups which are distinguished by cultural criteria (occupations), whereas Australian aboriginal societies consist of exogamous groups which are distinguished by natural criteria (totems), yet considered as total systems, that is as structured arrangements of categories, these entirely different patterns of culture are strictly comparable. Lévi-Strauss’s demonstration that this is the case has the appearance of a highly sophisticated conjuring trick, and his problem throughout is how to convince the reader that the logic of primitive thought, which he exhibits, is a genuine characteristic of the way the human brain operates—an aspect of l’esprit humain—rather than just an exemplification of the ingenuity of Claude Lévi-Strauss.


NON-INITIATE READERS who are intrigued but not wholly persuaded will either have to take the argument on trust or tackle the detailed evidence which Lévi-Strauss is providing in his huge study of the mythology of the tribal peoples of South America. This is planned as a four volume work, of which two volumes have so far appeared: Le Crux et le Cuit (1964) and Du Miel Aux Cendres (1966). It is an extraordinary performance. In Le Cru et le Cuit Lévi-Strauss claims to demonstrate, by reference to 187 different myths, that in this region of the world there is felt to be a logical similarity between polarities of cooking, the polarities of sound, and the polarities of human self-consciousness of the form:

the transformation Raw—Cooked:

the transformation Fresh—Putrid::

Silence: Noise:: Culture: Nature

In the new book, which includes references to an additional 166 myths, the themes of honey and tobacco are seen as the “penumbra of cooking” (les entours de la cuisine), and their contrasts are said to correspond, in the logic of mythology, to contrasts “internal to the category of noise,” such as the opposition: continuous sound versus discontinuous sound or modulated sound versus unmodulated sound. The argument is that objects and sensory characteristics of things “out there” are manipulated by the brain, through the thought system incorporated in myth, just as if they were symbols in a mathematical equation. As an illustration of just how complicated this mathematics is said to be I provide a quotation from the end of the book:

When used as a ritual rattle (hochet) the calabash is an instrument of sacred music, utilized in conjunction with tobacco which the myths conceive under the form of (an item of) culture included within nature; but when used to hold water and food, the calabash is an instrument of profane cooking, a container destined to receive natural products, and thus appropriate as an illustration of the inclusion of nature within culture. And it is the same for the hollow tree which, as a drum, is an instrument of music whose summoning role is primarily social, and which when holding honey, has to do with nature if it is a question of fresh honey being enclosed within its interior, and with culture if it is a question of honey being put to ferment within the trunk of a tree which is not hollow by nature but hollowed artificially to make it into a trough [pp. 406/7].

If Lévi-Strauss is justified in believing that primitive people “think like that” then quite clearly the Frazer-Lévy-Bruhl-Sartre notion that primitive thought is characterized by naïveté, childishness, superstition, and so on is wholly misplaced. Lévi-Strauss’s primitives are just as sophisticated as we are; it is simply that they use a different system of numeration.

But is he justified? The ethnographic worry is that Lévi-Strauss may have selected his evidence so as to fit his theory, that if he had used other evidence the thesis might fall to pieces; for it must be understood that, despite the formidable list of 353 myths, there is a lot of other rather similar stuff which might have been used. Of course, there are two volumes still to come, but it worries me that at least one important source has so far been largely neglected, namely von den Steinen’s materials on Bakairi myth. These data are unusually detailed, and since the Bakairi live next door to the Bororo whom Lévi-Strauss utilizes as his basic “type” of Amazonian Indian, one would have thought that this evidence must be highly relevant.*

But that is by the way. I have put in that cautionary paragraph for the benefit of my anthropological colleagues some of whom seem to think that I am such a devoted Lévi-Straussian that I can no longer exercise any scholarly detachment on the subject at all. In fact I think the case does stand up; the fault, if there is one, is that Lévi-Strauss tries to make his mathematics of manipulated sensory objects too systematic. He fails to allow for the fact that, whereas the symbols used by mathematicians are emotionally neutral—ix is not more exciting than x just because i is an imaginary number—the concrete symbols used in primitive thought are heavily loaded with taboo valuations. Consequently psychological factors such as “evasion” and “repression” tend to confuse the logical symmetries. This does not mean that Lévi-Strauss’s calculus must be “invalid,” but it may be much less precise than he seems to suggest. Or to put the same point another way: Because he takes his cue from Jakobson-style linguistic theory and the mechanics of digital computers, Lévi-Strauss tends to imply—as is clearly shown in the long quotation cited above—that the whole structure of primitive thought is binary. Now there is not the slightest doubt that the human brain does have a tendency to operate with binary counters in all sorts of situations—but it can operate in other ways as well. A fully satisfactory mechanical model of the human mind would certainly contain many analog features which do not occur in digital computers. For example most human beings make a distinction between b sounds and p sounds, also between green color and yellow color, and also between the value good and the value bad. These distinctions are of the either/or (binary) kind, but they cannot be mechanical in any simple sense because different human beings can cut up the cake of experience in quite different ways. But quite apart from that we can also make distinctions of intensity—louder-softer, brighter-darker, better-worse, and these are not binary discriminations at all. So far, the Lévi-Straussian scheme of analysis cannot take such factors into account.


Oh dear! Once again I seem to be talking polemic for the initiates! Novices who tackle The Savage Mind as their introduction to the mind of Lévi-Strauss will, if they are patient, get an enormous brain-twisting enjoyment out of the first eight chapters, which provide a basic introduction to the logic of “the science of the concrete.” They will not be in a position to judge whether Lévi-Strauss is correct in claiming that this logic is a universal human characteristic, but they will certainly begin to see some of their own familiar behavior in a new light. May I commend in particular the extensive references to be found in the Index under: “Dogs, names given to”? The basic point here is that, with us, dogs, as pets, are a part of human society but not quite human, and this is expressed when we give them names which are like human names but nearly always slightly different from real human names. Well, perhaps? But then Lévi-Strauss goes on to argue that the names we give to racehorses are of quite a different kind (as certainly they are) because racehorses “do not form part of human society either as subjects or objects. Rather they constitute the desocialised condition of existence of a private society; that which lives off racecourses or frequents them.” The train of thought is to me quite fascinating, but what sort of “truth” is involved? Even if we grant that the names given to racehorses form a class which can be readily distinguished, is this juxtaposition of the type of name and the type of social context anything more than a debating trick? The question needs to be asked. Whether it can be fairly answered I am not sure. Each reader needs to consider the evidence and think it out for himself.

WHAT WILL doubtless puzzle the novice reader—more particularly when he comes to Du Miel aux Cendres—is how on earth Lévi-Strauss comes upon his basic oppositions in the first place. How could it ever occur to anyone that an opposition between roast pork and boiled cabbage might reflect a fundamental characteristic of human thinking, or that honey and tobacco (of all things) might come to have a significance as fundamental as that which opposes rain and drought? The answer, I think, is that Lévi-Strauss starts at the other end. He first asks himself: How is it and why is it that men, who are a part of Nature, manage to see themselves as “other than” Nature even though, in order to subsist, they must constantly maintain “relations with” Nature? He then observes, simply as a fact of archaeology rather than of ethnography, that ever since the most remote antiquity men have employed fire to transform their food from a natural raw state to an artificial cooked state. Why is this? Men do not have to cook their food, they do so for symbolic reasons to show they are men and not beasts. So fire and cooking are basic symbols by which Culture is distinguished from Nature. But what about the honey and tobacco? In the case of cooked food the fire serves to convert the inedible natural product into an edible cultural product; in the case of honey the fire is used only to drive away the bees, that is, to separate the food, which can be eaten raw, from its natural surroundings; in the case of tobacco it is the conversion of the food by fire into a non-substance—smoke—which makes it a food. So here already we have a set of counters of different shapes and sizes each with a front and a back which can be fitted together into patterns, and which could be used to represent the exchanges and transformations that take place in human relations, as when a boy becomes an adult, or the sister of A becomes the wife of B. With some such framework of possibilities in his mind, plus the basic proposition that mythology is concerned to make statements about the relations between Man and Nature and between man and man, Lévi-Strauss looks at his evidence and the pieces of the puzzle begin to fit together.

Because the game is unfamiliar the whole business at first seems very astonishing; there must be a catch in it somewhere. On the other hand, if Lévi-Strauss’s basic assumptions are valid—as I myself think they are—it could hardly be otherwise! Even if his argument eventually has to be repudiated in certain details, we simply must accept certain fundamental parts of it. Any knowledge that the individual has about the external world is derived from structured messages which are received through the senses…patterned sound through the ears, patterned light through the eyes, patterned smell through the nose, and so on. But since we are aware of a single total experience—not a sound world plus a sight world plus a smell world—it must be because the coding of the various sensory signal systems can be made consistent—so that hearing and sight and smell and taste and touch, etc., seem all to be giving the same message. The problem then is simply to devise a means of breaking the code. Lévi-Strauss thinks he has solved this problem; even those who have doubts should be fascinated by the ingenuity of the exercise.

The ninth chapter of The Savage Mind is of a different kind from the rest. It is a direct critical attack on the presuppositions of Sartre’s Critique de la raison dialectique and it is, for me anyway, extremely difficult to pin down just what all the pother is about. What Lévi-Strauss seems to be saying is that Sartre attaches much too much importance to the distinction between history—as a record of actual events which occurred in a recorded historical sequence—and myth, which simply reports that certain events occurred, as in a dream, without special emphasis on chronological sequence. History records structural transformations diachronically over the centuries; ethnography records structural transformations synchronically across the continents. In either case the scientist, as observer, is able to recorded the possible permutations and combinations of an interrelated system of ideas and behaviors. The intelligibility of the diachronic transformations is no greater and no less than the intelligibility of the synchronic transformations.

The argument is close-woven and full of splendidly distracting rhetorical flourishes—at one point Sartre’s views are rated “a sort of intellectual cannibalism much more revolting to the anthropologist than real cannibalism”! The synchrony of ethnographic variation corresponds to a characteristic of savage thought itself which Lévi-Strauss compares to the effect of a hall of mirrors in which “a multitude of images forms simultaneously, none exactly like any other, so that no single one furnishes more than a partial knowledge…but the group is characterized by invariant properties expressing a truth.” The truth is a totality, the sum of many overlapping partial images. History, on the other hand, sacrifices totality in the interest of continuity. “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…a history of the French Revolution cannot simultaneously and under the same heading be that of the Jacobin and that of the aristocrat.” In a formal sense Lévi-Strauss does not denigrate history; he simply argues that “historical knowledge has no claim to be opposed to other forms of knowledge as the supremely privileged one,” but there is a strong suggestion that the only really satisfactory way to make sense of history would be to apply to it the method of myth analysis which Lévi-Strauss is exhibiting in his study of South American mythology! Whether such an argument could possibly have any appeal to professional historians or philosophers of history it is not for me to say. Certainly it lies far off the beaten track of conventional anthropology.

The anthropologist’s main lament is rather different. The technique of analysis which Lévi-Strauss employs with such skill and seeming success in the analysis of myth grew out of his early exercises in the analysis of formal marriage systems, notably in Les Structures élémentaires de la parenté (1949). In this work the conventions of matrimonial exchange were shown to constitute a system of communication between social groups; in the same way conventions of cooking have now been shown to constitute a system of communication between Man and Nature. This analogy is not preposterous. Human brains devised styles of cooking and human brains devised rules of marriage—the two codes of thought may well link up. But Lévi-Strauss has not tried to show that they in fact link up; he has simply moved off to study other things.

This is a pity. Compared with cooking and music and the peculiarities of naming systems the study of kinship and marriage is dull and pedestrian stuff, but, for an anthropologist, kinship is the hard core, and for some of us Lévi-Strauss’s retreat to the land of the Lotus Eaters is, to some extent, a matter for regret.

This Issue

October 12, 1967