Notes on: Levi-Strauss. C. (1976). The Savage Mind. London: Weidenfeld and Nicholson

Dave Harris

[A massively detailed study of other people’s anthropological findings, impossible to summarise in full detail. His point is to summarise and organise them as a much simpler qualitative logical structure. This is very impressive but impossible to check or disagree with of course. Much of the detail must be omitted here. Luckily, the end of each chapter provides the punchiest examples of the approach. My orientation has been to contrast the work to the banalities of the accounts of indigenous cultures in the stuff I have been reading lately – Smith or Chilisa on indigenous research methods and their supporters. Generally, I find the summaries in those works of indigenous cultures as centred on spirituality, community, respect for all things living and non-living etc about as useful as using the Ten Commandments as a description of American or British culture]

Chapter 1 The Science of the Concrete

Languages that lack terms to express concepts, or which seemed to lack abstract words are sometimes seen as primitive, even if they offer detailed inventories of things like species or varieties. Boas has argued that Chinook contains some the most complex inventories he has encountered. Discourse and syntax always supplant any inadequate vocabularies. There is a similar myth that argues that simple natives only name things that are relevant to their immediate needs, this is sometimes contrasted to theoretical interests. However this does not reflect greater or lesser intellectual capacity, but rather different cultural characteristics. Abstract words are usually defined in terms of their own language anyway — tree for example, whereas more specific words like oak or beech could be equally abstract and refer to concepts and technical issues.

Primitive people are as thirsty for objective knowledge as we are, although not directed towards facts of the same level as those of modern science, but certainly implying 'comparable intellectual application and methods of observation… The universe is an object of thought at least as much as it is a means of satisfying needs' (3). We seem equally disinterested in objective orientations to other factors, for example not classifying all aspects of winds, varieties of water, or particular kinds of crop [compared to the Hawaiians in this case, and with lots more examples of a Philippine people]. This interest in, say, plants which are not of immediate use arises because they do have 'significant links with the animal and insect world'. This not only help such people survive on natural resources, often much better than Europeans could, producing, for example, substantially detailed lists of ethnobotanical categories [lots of examples]

These are based on experience with the actual plants, however, connected with knowledge of healing and people and their environment. As a result, knowledge does not just relate to practical purposes, and extends, for example to animals that have no economic benefit, such as snakes and other reptiles. It is almost the reverse of the usual view — plants and animals 'are deemed to be useful or interesting because they are first of all known' (9). Classifications like this meet intellectual requirements. They fulfil needs to develop points of view where, for example 'a woodpecker's beak and a man's tooth can be seen as "going together"', with a therapeutic purpose being only one example of a possible connection. Classifying has a value of its own. Scientists are interested in classification because they want to pursue the notion of order, and so is 'the thought  we call primitive' (10). We have here 'properties common to all thought'. It may be that sacred objects contribute to order precisely because they occupy the places allocated to them, and this might explain the function of ritual which reinforces order. Sometimes conventionally scientifically valid results, such as the correct prediction of the onset of spring, even though the methods used to predict it [examining the foetuses of bison] were 'illusory'.

The same goes for determinism — magical thought can express 'more imperious and uncompromising demand for it which can at the most be regarded as unreasonable and precipitate from the scientific point of view' (11), so that if a buffalo gores a man, the Azande say that witchcraft must also be involved in producing the particular situation in which buffalo and man are brought together [citing Evans Prichard]. There is a complete and all embracing determinism for magic, but the notion of levels for science where only some of these are deterministic. In that sense, magic can be seen as 'so many expressions of an act of faith in a science yet to be born'. Sometimes these anticipations succeed sometimes they anticipate science itself, for example in systematising from an early stage what was presented to the senses [as a kind of empiricism? The example given is a particular development in modern chemistry that tried to separate out secondary characteristics of particular chemicals by relying on data provided by the aesthetic senses — weird, page 12]. The section ends by insisting that taxonomy '"has eminent aesthetic value"', and so it follows that aesthetics can be useful in taxonomies (13).

This is not to argue that magic is a timid early form of science, a mere stage. It is complete in itself, already finished and coherent, not a beginning not a sketch, already well articulated and independent of science except through an analogy [one can be seen as a metaphor of the other]. The two approaches emerged in different objective conditions. We see this in the 'Neolithic paradox', when human beings first mastered operations like pottery, weaving, agriculture, the domestication of animals [and, later metal smelting]. This could not have arisen as a result of chance discoveries [my favourite example is the clever breeding of maize from an original wild grass that had to involve human intervention]. All these involved methodological observation, hypotheses and experiment, which we can see when new plants are introduced in places like the Philippines. Useful properties here were originally absent. Techniques had to be worked out and these were often long and complex. 'All these achievements required a genuinely scientific attitude, sustained and watchful interest and desire for knowledge for its own sake' (14). These products are the result of a long scientific tradition, and yet 'several thousand years of stagnation' followed.
This tells us that there must be two modes of scientific thought, not different developments of the human mind, but 'two strategic levels at which nature is accessible to scientific enquiry' — one involving perception and imagination, the other quite removed from it, remote from intuition. Both depend on classification and rational ordering, even on the basis of aesthetics, and it is convenient if aesthetic qualities like taste and use also correspond to objective reality. This led to a useful cultural memory bank, often preserved in the form of myths and rites. We can call this a 'science of the concrete' necessarily restricted in the results it could achieve but no less scientific and providing no less genuine results than the exact sciences: 'they were secured 10,000 years earlier and still remain at the basis of our own civilisation'.
We can also understand speculation or bricolage, originally a matter of extraneous movement, more recently working with your hands using devious means, drawing upon a 'heterogeneous repertoire' which may be extensive but still limited, but to be drawn on because nothing else is available (17). Mythical thought is intellectual bricolage, and can reach 'brilliant unforeseen results'[we can apparently see this in some artistic examples, which include Dickens in Great Expectations and the example of the suburban castle — pass] [there follows the famous distinction between bricoleur and engineer, apparently criticised by Derrida]. Bricoleurs are good at the large number of diverse tasks but do not refer back to the availability of raw materials and tools related to the purpose of the project, instead making do with whatever is at hand, not limited to a particular project, but rather using bits and pieces collected because they might always come in handy. These materials never have only one definite and determinate use and 'represent a set of actual and possible relations' (18).

The same goes for elements of mythical thoughts, 'halfway between percepts and concepts' [turning on whether thoughts can be separated from concrete situations in which they appeared, or, conversely, put into brackets]. Between images and concepts there are signs, as in Saussure on linguistic signs, where images and concepts are the signifying and the signifying [I thought for one minute that this was going to lead to a connection with the logic of myth as obeying structuralism, just as kinship practices do, as in his other work, but I don't think it is developed in this particular work. There is a connection later in the arbitrary nature of things used as signs -- content does not matter]. Signs are like images in that they are concrete, but they also have general powers of reference like concepts and like concepts they can be substituted for something else although not in an unlimited way. We see this with the way that the bricoleur works: first he turns back to an already existing set of tools and materials and engages in a kind of dialogue with it before choosing elements to 'index the possible answers which the whole set can offer to his problem', interrogating all the different objects to discover what each of them could '"signify"'[explores the machinic phylum?]. This in turn implies that there is a set, 'yet to materialise', but not necessarily the same as an instrumental set — a particular piece of wood could be used as a wedge or a pedestal, for example, although there are always limits according to the 'particular history of each piece' [its social or mythical history?] and its features and properties, like those arising from any earlier modifications. These are forms of pre-constraint, 'like the constitutive units of myth', restricted in this case because they are drawn from a language where they already possess a sense. Each element has a possibility of being replaced, so there is a choice involved, and that may well involve 'a complete reorganisation of the structure', perhaps in ways which had never been envisaged or previously preferred.

Engineers also 'cross-examine' their resources, because they also have limits to their means and knowledge and will meet resistance, although in effect they are questioning 'the universe' (19), or having a dialogue with nature rather than a subset of culture [or rather with a relationship between nature sanctified by his culture]. However, engineers also have to establish a set of theoretical and practical knowledge and technical means. So there are some similarities, although there are real differences in that engineers tried to go beyond the constraints imposed 'by a particular state of civilisation', while bricoleurs try to remain within them 'by inclination or necessity'. This follows from the differences of working with concepts as opposed to signs — the sets composed of concepts are on a different location from those of signs 'on the axis of opposition between nature and culture' (20) [gives far too much to the abstract nature of concepts and their ability to escape cultural constraints]. Certainly signs allow and even require human culture to intervene in reality.

However both  bricoleurs and scientists search for messages, codes, either those that have been codified already into experience, or those that are available to open up sets, provide new significations and reorganisations. Images can also play a part coexisting with ideas in science and can reserve a place for ideas. They may not yet be comprehensible and easily linked with other entities but they can be related to other entities as long as they form a system. This is the same as the logical processes of extension and intension, and is a way in which mythical thought generalises: it can therefore become 'scientific, even though it is still entangled in imagery' (20). There are also analogies and comparisons even though [for] the bricoleur these are new arrangements of elements which remain in the same set or arrangement, even though the effects might be to destroy mythological worlds altogether and require that they be built up again from the fragments, from the same materials, following the same ends [he puts this rather confusingly — 'the signified  changes into the signifying and vice versa' (21)]. This implies there must be some notion of the total, some understanding of the relation between an instrumental set and a project. It is also inevitable that the bricoleur 'always put something of himself' into the project.

Science is based on the distinction between the contingent and the necessary, leading to a distinction between event and structure. It focuses on qualities which 'formed no part of living experience and remained outside and, as it were, unrelated to events' — primary qualities. Mythical thought, on the other hand, and bricolage in its practical phase, builds up structured sets not only with other structured sets like language but by using 'the remains and debris of events' (22), odds and ends, bits and pieces of the history of an individual or a society, reversing the usual [scientific] notion of 'the relation between the diachronic and the synchronic'. This builds up structures. For science it is the reverse, creating means and results as events thanks to the structures which it elaborates in the form of hypotheses and theories. Both approaches are equally valid, not sequences in some evolution [then an odd bit about how sciences are already 'striving to become qualitative again… To account also for secondary qualities… Means of explanation' one example is biology trying to explain life itself. Unlike myth, science is trying to extend meaning from what it originally agreed as a compromise could be considered meaningless — maybe].

[Then a diversion on art]. Art lies halfway between science and mythical or magical thoughts, with artists as both scientist and bricoleur. This is illustrated with close analysis of a particular portrait and the way it generates 'very profound aesthetic emotion' despite its highly realistic reproduction [in this case of a lace collar]. This goes on to discuss why small-scale models,  miniatures and so on have an aesthetic quality, why reducing the dimensions adds to understanding by making objects appear easier to understand, less resistant to understanding, and obviously man-made experiments, helping us see how methods of construction have produced effects, just as with bricolage. This immediately raises issues of choice and modification so that observers become active participants 'without even being aware of it' (24) [reminds me of Barthes and the punctum of the photograph], so that intelligible dimensions are increased, by metaphor. Painting also includes [narrative], some element of becoming, a hint of social events which adds to the aesthetic emotion.

Going back to myths we can see them as abstract relations and aesthetic contemplation. Myths approach creation from the other end compared to art, using a structure to produce an object consisting of a set of events rather than using an object in order to discover a structure. There might be an objection in an example which he has of [primitive] art which seems to offer a straightforward integration of structure, myth and function — it is a carving of a mythical animal which is used as a club to kill fish. He proposes that we can modify the definition of art he offered above to now include not just events but the contingent itself, which may be depicted as an event or as contingency itself [very weird.] It ends with the classification of art according to whether it has or thinks it has mastered contingency of technique, or other contingencies of form, and this can help us distinguish between easel art and ‘primitive’ art, the same dialogue with materials and means that we found with bricolage, only developed further as a balance between structure and event, necessity contingency, the external and the internal.

Finally he finds relations of the same kind with games and rites. Games have rules which allow any number of matches to be played. Ritual can be seen as 'a favoured instance of a game' which has to produce a particular result, an equilibrium between the two sides, for example games in New Guinea which have to be played until both sides reach the same score, or until a preferred side win [a side representing death in this particular case,p 31]. Competitive games often accompany rites. In terms of symbolism, winning a game is the same as killing an opponent, so the right people should always win. Games therefore 'appear to have a disjunctive effect' (32). establishing a difference between players or teams where there originally was no inequality, and producing winners and losers. Rituals do the opposite, conjoining, relating to initially separated groups . Games produce events from structures, so it is not surprising that they 'should flourish in our industrial societies', while rites and myths, and bricolage take to pieces and reconstruct events and use them to create structural patterns which can be either ends or means [and some have died out]

[A lot of this stuff about opposite functions ensues later ]

Chapter 2 the logic of totemic classifications

The items used in totems look like odds and ends, almost arbitrary, but that's because we are focusing on the content. There is a logic that connects them nevertheless, to do with the form, analogies, developed by bricolage. We start to see this if we look at first of all the use which these elements have as images, how they get detached from their immediate context and used for various purposes. They acquire the rigour of any term used in language or in a technological system, that is they become 'condensed expressions of necessary relations which impose constraints' when they are used, expressing a semantic or aesthetic order (36).

It is like the way in which a kaleidoscope works, bits and pieces have to be homologous in different respects, but they can also become parts of the new type of entity, patterns produced by the play of mirrors or reflections, possibilities, which may be considerable, although not unlimited. The kaleidoscope produces combinations of contingent events according to a law, by which the kaleidoscope is constructed and the models it produces are intelligible but still provisional. The patterns do not correspond to the observer's experience, even though they might resemble objective structures like snow crystals.

We can observe this sort of logic at work in ethnography and it has an affective and an intellectual logic. For example there is a belief in supernatural beings among the Ojibwa, which are part of the natural order of the universe just like men whom they resemble in being male and female, have the same sort of intelligence and even families, and can be identified with, more like men than different. The same might go for animals, long known and understood, perhaps even able to teach, maybe even once married to the natives. This is 'concrete knowledge', shared, perhaps by current colleagues who work with animals like directors of zoos, who have theoretical knowledge as well as affective.

Let's turn to the systematic nature of native classification, which seems extensive to put it mildly [examples of the Dogon and the way they divide plants, page 39, also found in America, often based on binary classifications, sometimes trinary — see example later]. They do not correspond to contemporary zoology but may be based, for example on size, function in magical ritual, but native taxonomy is often 'precise and unambiguous' (40) types of animal plants correspond to various rites leading to considerable complexity [example page 41 –  a 6×8 table]. These were often avoided by earlier ethnologists who believed they were studying simple societies because those had low economic and technical levels — we now know that complex 'forms of science and thought… Are extremely widespread in so-called primitive societies' (42). Nor is there any evidence for 'consciousness governed by emotion lost in a maze of confusion and participation'. Instead, we have considerable speculation at least as advanced as those of 'the naturalists and alchemists of antiquity and the Middle Ages', and close to the systems of Greeks and Romans or the mediaeval church.

Native classifications are methodological and 'based on carefully built up theoretical knowledge. There are also times comparable from a formal point of view to those still in use in zoology and botany… [Some]… are able experimenters in the preservation of foodstuffs' (43)  and the American army borrowed some techniques from the Bolivians to reduce to powder their potatoes. Elaborate taxonomies of maize were developed based on form texture or 'sex', and these were superior to more recently developed ones that simply over-generalised [marvellous examples 44, 45. The Navajo were particularly adept classifiers]. Artemisia is an important plant in a number of rituals across North America, but that is because there are in fact several varieties which should be carefully distinguished, undetected at first and the different varieties given different connotations [in a triumph of structural analysis, Lévi-Strauss organises these different associations as a diagram based on dichotomous poles of female and male, with the female further divided into female birth and male birth, page 48]. Lots of other examples of using various rituals to hunt or incorporate different animals, such as eagles or wolverines and the rituals again display different but important contradictory characteristics of the animals, again revealing important dualisms, say between the eagle as celestial prey and the Indians who hunt them who conceal themselves in pits subterranean hunters: the contradictions are resolved in complex rites to manage the symbolic differences, involving transformations of people into arrows, or ambiguous animals, with menstruating women as important mediators [roughly because their blood symbolises the blood of the necessary bait to attract the prey, one of the few occasions on which menstruation has a positive significance]. Other associations ramify outwards to provide a general system of reference 'allowing the detection of homologies between themes whose elaborate forms do not at first seem related in any way' (52) which makes eagle hunting extremely important.

For the natives, the accurate identification of large numbers if not 'every animal, plant, stone, heavenly body or natural phenomena mentioned in myths and rituals' is necessary (54) but well beyond most ethnographers. To acquire this knowledge involves 'long and constant tension, painstaking use of senses, ingenuity which does not despise the methodological analysis of droppings of animals to discover their eating habits, et cetera'. Among this huge amount of knowledge, selections are then used to provide knowledge, in what looks like a rather arbitrary way, again we need to consider the whole system here, however — those chosen usually 'lend themselves to anthropomorphic symbolism and… Are easy to distinguish from each other… [And]… Can be combined to fabricate more complex messages' (54). All sorts of other possibilities could have been used, however and none of the terms have 'any intrinsic significance'. Everything turns on the position of the terms, the structural system in which they are set, and the history and cultural context. In one example, Navajo, insects are grouped under a generic term, referring to the larval state, while larks are described in reference to the extended hind claw. An early classification of colours among a Philippine group was baffling at first but became clear after it was realised that it was based on specimens rather than abstract axes like our own [and I think of Bourdieu on the calendar among the Kabylia]. The [bird] tree creepers are classified among those who hide from birds of prey by Australians, but as redheaded birds by North American Indians, and by birds associated with tempests and storms by another group of American Indians, while Borneo persons see their triumphal song as the most important for its symbolic role. [There are lots of other examples with crows, say, or bees, carrying a different 'semantic load' (56) and one interesting example of emergent categorisation where languages lose particular classes and so categories come to be grouped together, in this case animals and manufactured goods, bees and canoes, since both are '"manufactured"', or at least honey is].

Sometimes observers can offer interpretations which 'cut across the natives' interpretation', coming up with a master plan, although they still need to be tested against a suitably rich ethnographic context [and examples of recalcitrant data are given on 57 and 58]. In general, 'the principle underlying classification can never be postulated in advance. It can only be discovered a posteriori by ethnographic investigation' (58). Some classifications are more systematic than others, but even here it helps to know that particular associations are in play, even though there may be 'nothing to suggest this in advance' [an example is the symbolic role of the pelican which because it lives to a great age is associated with metal on account of its hardness, or the body of the elk, whose components represent a wide variety of components of the landscape], In one case, data has been collected by an insider himself aware of 'all the intricacies of native thoughts', but in other cases, tribes are almost extinct and difficulties are insurmountable and answers to fieldwork questions are 'hopelessly vague' (60).

We may simply lack knowledge of what native peoples have observed, and what they take to be facts or principles, real or imaginary. We must work with 'small but precious clues' including native texts [one example explains that a native text explains that an interest the Ojibwa have in  squirrels is really 'an interest in a kind of tree', while New Guinea tribesman are interested in squirrels because they are fruit eaters and headhunters feel a fraternal relationship to them. There are lots of intriguing examples of homologies concerning the burrowing habits of animals applied to states of pregnancy]. Another difficulty is that logic is often 'polyvalent' applying to different types of connection at the same time. In one example, clans have the names of animals, plants or manufactured articles, but these are linked in pairs rather than being strictly totemic: they illustrate a joking relationship, where leopard and goat clans are related because leopards eat goats, iron clans joke with all the others with animal names because animals are killed by metal spears. Plants may take on particular virtues because they resemble parts of the body, they grow near important medicinal plants, they are associated with animals, they make the right sort of colour or odour, they are found near trees struck by lightning and so on — there are several axes for systems of logic, contiguity, resemblance, sometimes appearance, sometimes functions, sometimes close or distant connections, synchronic or diachronic.

Much depends very often on precise and detailed observations, and the same item can take on different roles. We need not only ethnographic data but also 'zoological, botanical geographical et cetera' data (64) there seems to be no limit to the variety of interpretations. Sometimes a structure of opposites can be reversed — red and white colours, for example, associated with death in one case, life in the other, or sometimes the whole opposition is replaced with another one of black and white, or colour and its absence, or sound can replace colour. This makes the point that it is form not content that counts, and helps dispel the emphasis on archetypes or collective unconsciousness — any common contents depend on 'the objective properties of a particular nature or artificial entities or in diffusion and borrowing, in either case, that is, outside the mind' (65).

Concrete logic is complex because anything which comes to hand can be used. It is a qualitative logic, just like structural linguistics ['arbitrary']. There is also demographic change, diachronic alterations despite synchronic systems, as populations become progressively smaller, for example. Languages persist if they are protected by practical purposes, communication, but conceptual systems are not dependent on communication but rather 'means of thinking' (67) and ways of remembering. [There is a section which suggests, I think, that speculation is also easier than actually remembering past forms of communication].
In the example, we begin with three clans each of which has the name of an animal, but demographic changes lead to the extinction of one clan and the increase in the population of another which splits into two sub clans which eventually become clans in their own rights. The old structure disappears and is replaced by the new structure. It's impossible to detect the original structure and it might have disappeared from native thoughts leaving only the original names as traditional titles. It might be possible to reconstruct a system in theory, and perhaps the original tripartite system can be identified in the new one. More importantly, the original system 'rested on myths of creation and origin and permeated ...ritual'(68) and myths and rites might survive demographic collapse, at least for a time, especially if they had a certain vigour, and were to some extent compatible with the new structure. There would be a kind of feedback system where any discordance would be directed back towards an equilibrium, 'a compromise between the old state of affairs and the confusion brought in from outside  '(69) [an argument for saying that there would be some residues of the old system even after colonisation, that it would be possible to gain some knowledge of pre-colonial cultures?]. Some 'traditional legends of the Osage apparently show this sort of interpretation and adjustment going on, turning on the adjustments made by the ancestors and how they encountered others and produced original clans, how the system encountered disequilibria, how the number of clans were adjusted, and how camps still represent the original number of clans, to integrate past and present forms of structure, 'at once historical and structural… Symmetrical and asymmetrical, stable and top-heavy' (70). Instead of choosing between alternatives, as modern practice in academic debates requires, the Osage take the opposition 'as a point of departure' accept both and try to work out 'a single scheme which allowed them to combine the standpoint of structure with that of event'.

This can also explain other mixtures of 'divergences and parallels' found in other American Indian societies, and other systems where clan names are 'almost always midway between order and disorder' where demographic changes push towards disorganisation, while 'speculative inspiration' push towards reorganisation as close as possible to the earlier state of affairs [more examples follow]. So-called primitive peoples therefore, are 'constantly negotiating diachrony and synchrony, event and structure, the aesthetic and the logical', and it is fruitless to try to explain their social lives in terms of one or the other only. 'Between the basic absurdity Frazer attributed to primitive practices and beliefs and the specious validation of them in terms of the supposed common sense invoked by Malinowski, there is scope for a whole science and the whole philosophy' (74).

[Another example turns on some botanical implications, where socially mixed people in Mexico, who were Spanish-speaking and did not consider themselves Indians grew far more mixed variants of maize. More traditional groups who had retained their old languages and their own cultures also grew far more true to type kinds of maize as well, which is difficult to do because maize easily crossbreeds and you have to be really 'finicky' in selecting seed and pulling out crossbred plants. The traditional groups even grew different varieties among themselves, showing 'a fanatical adherence to an ideal type' (73)]

Chapter 3 systems of transformations

On theoretical and practical planes 'the existence of differentiating features is of much greater importance than their content' (75): we can use them to form a system or grid to de-cipher text which was originally unintelligible and we can note divisions and contrasts which will produce a significant message. We can also organise a sociological field to grasp historical and demographic processes of evolutionary transformation. This also suggests a 'theoretically unlimited series of different contexts'. Opposition is the basis of the logical principle, suggested by the empirical totality. The issue then becomes how to oppose, identifying a formal character, turning say totemic items into signs or codes which can be transposed into other codes and express messages received by different codes. It was an early mistake to reify this form and tie it to a specific content. What it actually is is 'a method for assimilating any kind of content' (76) [just like the dispositions in the habitus]. Totemism is not autonomous, but derives from a deeper '"socio-logic"' in Durkheim's terms.

[In an example, Frazer diagnosed totemic beliefs in Melanesia as revealing a more primitive form from which all the specific variants were derived. A soul, in effect, was discovered during pregnancy when an object was mysteriously connected with the unborn child, and this led to, say, prohibitions connected with that object, abstaining from it if it was a food object, sheltering it if it was an animal. Frazer saw a connection with a similar prohibition where a dying man would indicate an animal in whose form he would be reincarnated which led to a prohibition on harming the animal, a food taboo. He 'elevated [this] to the status of a natural and universal phenomenon… The ultimate origin of all totemic beliefs and practices' (78). This was based on a false generalisation, the cravings of Melanesian women and those of European society at the time, which led Fraser to see this as natural — he couldn't see it is cultural because this would allow 'alarming resemblances' between European societies and cannibals [he was OK with natural resemblances?] . We now know that these cravings are temporary and have probably disappeared now anyway. It is not clear why these cravings of pregnant women were seen as prior to those of dying men either. The two systems are not exact counterparts, nor is one chronologically prior to the other. The relation is best understood as a structural one, a triple opposition between birth and death on the one hand, individual and collective nature of a diagnosis or a prohibition on the other (diagram page 80). There is only a 'homology between natural distinctions and cultural distinction. They are not universal and do not apply to all members of the society, but to a sample]

Other ethnologists have also noticed relations between widely dispersed social arrangements [in this case in Australia]. There seems to be, for example systematic reversals with intermediate cases as you go from the north to the south of Australia, with accompanying rules of marriage [very detailed, diagram on 83]. A methodological problem is raised here, since anthropologists have disagreed, but one saw the culture in question 'when it was still intact' while the other saw it 'only in an already advanced state of decay' (84)

We also are introduced to the notion of 'functional yield' which turns on transitivity. That is, some marriage systems produce different social groups for the offspring, which permits totemic groups to survive, while totemic systems that reproduce only the mother's group are more at risk. The same goes for matters such as the significance of locality: one group treated it as having 'a real absolute value' because it belongs to a totemic species, while another sees the spirits as owning the locality and they are much more flexible and mobile, so 'the totemic places are ports of call rather than ancestral homes' (85).

Further variations are introduced among groups that are split into moieties, which can have reciprocal rights with different implications for transformations. For example some cult groups perform rites for the benefit of other groups, for example eating food that is taboo for them. Benefiting from increases may be regulated in different ways, periodically or non-periodically, regulated by living communities or spiritual ones, and again there are relations of symmetry and opposition balanced with similar symmetries and oppositions in other areas like marriage rights.

[There are some interesting differences in how totemic ancestors or spirits are conceived. Some see them as 'single individuals who are half human half animal' others have a 'multiplicity of ancestors (for each totemic group) who are, however incomplete human beings', while others are a mixture of incomplete human beings and proper men. These are balanced by mythologies which contrast multiple ancestors with individualised rituals and the reverse, or ascribing properties to the earth which are religious or social].

Lévi-Strauss claims that 'all these transformations could be systematically set out'. For example one group has a man dreaming the totemic affiliation of its future child, but in another group it is the reverse [note the structuralist notion of an opposite or reverse] , where the woman experiences it. There may be exact prohibitions, complemented by detailed restraints on marriage. Totemic taboos can extend across generations and so can marriage prohibitions extend across corresponding clans. Sometimes food is prohibited on the grounds that it might incarnate an ancestor, but at other times names of the deceased are avoided by the descendants even if there is only a remote resemblance. Overall 'the same ideas appear and disappear in different societies either identical or transposed from one level of consumption to another, sometimes applying to the treatment of women, sometimes that of foods, sometimes the words used in speech' (88).

Anthropologists have often focused only on small areas but even then have gathered enormous amounts of data and thus have abandoned the idea of synthesis. Intuitive methods in particular are unable to manage large amounts of data across a number of dimensions and Lévi-Strauss hopes that the use of punched cards and computers will eventually do the trick, especially in Australia.

[This next section summarises his entire approach, embracing complexity but also promising structure]

Australian societies have developed complexity not only because Australia is isolated but because this was 'desired and conceptualised, for a few civilisations seem to equal the Australians in their taste for erudition and speculation and what sometimes looks like intellectual dandyism... These shaggy and corpulent savages… Were, in various respects, real snobs. They have indeed been referred to as such by a specialist, born and brought up among them speaking their language… As soon as they were taught accomplishments of leisure, they prided themselves on painting that dull studied watercolours one might expect of an old maid… Theorising discussion was all the rage in this closed world and the influence of fashion often paramount… Each community had its own dress… And [this] was never called in question. It was in wealth or ingenuity of detail alone that people try to distinguish themselves from, and to outdo, the neighbouring village… Culture [was] treated like themes and variations in music… Australian cultures… [Stand]… In relation of transformation with each other, possibly more completely and systematically than those of other regions of the world' (89 – 90). But there is still the same internal relation at the general level in different levels of a single culture, codes which can lead to conceptual systems, referring to human relations to each other, to technical and economic matters and the relations to nature. Durkheim and Malinowski tried to reduce totemism either to social or natural domains, but actually it is 'preeminently the means (or hope) of transcending the opposition between them' (91).

[An example of an origin myth follows, where two women set off naming places animals and plants. One is pregnant after committing incest. They finally encountered a great snake who emerges from a water hole who swallows the women and caused a flood, so became associated with the rainy season which occurs on a regular basis. When the snake raises his head above the water hole the Earth is flooded, and this pattern can be plotted on a graph, which looks very much like a snake {only to us of course}. The floods are necessary. The snake is a male token, so the wet season becomes a male, while the original sisters, women and the uninitiated are associated with the dry season. They also blamed because they committed incest to get pregnant in the first place, but without doing that there would have been no life cycle of seasons. This is an obvious homology between natural and social conditions, but an obvious contradiction: men are superior to women in every respect, except that the rainy season also brings famine and isolation and danger, while it is the low status dry season that brings all the good things and women are obviously the fertile ones.  The contradiction has to be managed. Males are superior, but at a price-- they cannot be happy — except when they get old. This gives privileges to old men, including sexual privileges and control over culture and initiation rites. The whole example shows that relations with humans and nature is not obvious has to be thought out and compounded. The 'naturalist school' thought that 'natural phenomena are what myths seeks to explain, when they are rather the medium through which myths try to explain facts which are themselves not of a natural but a logical order' (95). Natural infrastructure is provided as if it were the deck of cards which human beings play with and which set limits on the games that can be played. Form determines this, not content. Contradictions exist of course — 'the poverty of religious thought can never be overestimated… [There are always]… "Structures of contradiction"'

Rituals can also act out contradictions — the rainy season engulfs the dry season so men possess women, and the initiated '"swallow up" the uninitiated' (96). Other totemic terms code natural situations, so that thunder is symbolised as a bird among the Ojibwa and bird species wintering in the south appear in April and disappear later in October which corresponds to meteorological data. This is common in other cultures where sequences of weather events are personified [in animals].

Totemism also offers 'an ethic which prescribes or prohibits modes of behaviour' at least prohibitions or rules of exogamy. There is here a petitio principi [a begged conclusion]. [I think, that totemism involves selection of animals and plants which involves prohibitions of different ones, and it follows that marriage between people of the same name can also be forbidden]. Eating prohibitions are very complex, usually, and extend beyond totemic prohibitions. [Lots of examples follow — some sorcerers may not eat particular animals because they have irregularly spotted hides which would mean irregular divinations, and eating zebra have the same prohibitions, fish with sharp bones might destroy the liver, the main organ of divination, particular trees symbolising the erect penis must not be desecrated, animals with innards the colour of blood are eschewed, or those with sharp teeth 'symbolising the painful after-effects of circumcision' (98) {surely these are straightforward indexable signs}. They can be intermediate types and a variety of reasons — mice are forbidden to girls because they steal things and girls might be stolen, or, elsewhere, because they are regarded as members of the family. Horned animals are associated with the moon. Some taboo foods are forbidden to initiates of particular cults but permitted to novices and so on.  Sometimes eating taboo foods produce psychosomatic allergic disorders. There may be mock prohibitions].
 Again it is about distinction. If there are no food prohibitions there are other kinds of differentiation — bodypaint, clothes, how you wash your clothes, whether you build dams, how you bury people, whether you wear feathers, swim rivers and so on. Eating prohibitions may not be distributed evenly even among neighbouring cultures — some may be patrilineally transmitted, others matrilineally [the latter association seems more general]. Prohibitions can sometimes be obligations — 'the totem may be killed and eaten but not insulted'(102).

Examples show that the forbidding of some species 'is not attributable to the belief that the former have some intrinsic physical mystic property… But to the concern to introduce a distinction between "stressed" and "unstressed" species (in the sense linguists give to these terms)' (102). Prohibiting a species is to make it significant, and to subject it to a qualitative logic which can refer to images and modes of behaviour (marriage as well as eating). Sometimes transformations occur by 'inverting all its terms', making terms relevant to marriage suddenly become relevant to eating [the tribe in question 'has of course long been extinct and the data on it are contradictory'. However, the data gain some validity by being symmetric with existing institutions in another group]. Sometimes a kind of inversion takes place if the number of stressed foods increases — then what is prohibited, the negative term as it were,  becomes the significant term.

Overall what we have here is important 'means of "denoting significance" and a logical system some or all of whose elements are edible species' (103) although the systems are of different types. For example South African bushman have lots of prohibitions but no totems. Instead, their prohibitions turn on things like forbidding all game killed by bows and arrows until the chief has eaten a piece, except the liver. Prohibitions remain for some functional social categories, for example the wife of the man who killed the animal can only eat parts of it, as can boys. In this case, the system can be transformed into a totemic one simply by replacing parts of the animal for separate animals, and by extension, the functional classes are seen as parts making up the society — 'natural and social groupings are homologous in both cases' (104).
There are other empirical connections between marriage rules and eating prohibitions. Among the Tikopia and the Nuer, men abstain from eating food prohibited to their wives, because such food may may be introduced into their wives bodies through sperm. It is the reverse for the Fang, where intercourse makes men feeble if women continue to eat their own food.

These examples show 'the very profound analogy which people throughout the world seem to find between copulation and eating' (105)'. They may even be called by the same name, and even French uses the term consuming for food and women. Eating the totem becomes a kind of cannibalism in some societies and real or symbolic cannibalism is the punishment: the same goes for absconding with a woman forbidden by the law of exogamy. 'These associations could be multiplied indefinitely'. There is no relation of priority. It is metaphorical connection. It is found today in slang French.

We can often grasp the universal metaphor by 'semantic impoverishment': the 'Laws of Manu' provides an example: '"what is destitute of motion is the food of those endowed with locomotion, (animals) without fangs (are the food of) those with fangs, those without hands of those who possess hands, and the timid of the bold"' (106). We often see men as devourers, although the theme of the vagina dentata occasionally occurs to invert it. We see resemblance to contrast in other examples, involving the parts of the bodies of animals, for example — totemic animals may be divided into edible parts and emblematic parts, sometimes justified on the grounds that the emblematic bits are those in which animals and humans differ. Some anthropologists have simply seen the identity between the human and the edible parts as the crucial thing, but 'matters are in fact infinitely more complex: there is an exchange of similarities and differences between culture and nature' (107). Human beings sometimes have to deny a common nature with animals, and it is particularly important to deny that any animal species, or any part of particular animal species can be foodstuff. Instead human beings 'have to assume the symbolic characteristics by which they distinguish different animals (and which furnish them with a natural model of differentiation) to create differences among themselves' (108).

Chapter 4 Totem and caste

[Even more esoteric here, clearly referring to rather specialist debates among anthropologists about the relative merits of the two forms of classification]

The exchange of women and the exchange of food help interlock social groups with one another and are often found together as procedures of the same type or two aspects of the same procedure. They can reinforce each other, perform the actual function or represent it symbolically, or act as alternatives. Exogamy however is never entirely absent because it is necessary, it has 'real substance' (109) compared to exchanges of food. Totemic species on the other hand are only increased in imagination, by saying something. Nevertheless, there are parallels between eating prohibitions and rules of exogamy, sometimes complementarity, sometimes supplementarity.

In an example discussed earlier, we may exchange women but not seeds, resulting in extreme purity of seeds: the seed contains the spirit of the plant and there is a fear that that will disappear from its locality. It is apparently 'common in Melanesia' (110). In other societies, particular plants, yams, are seen as persons and like women they give birth to children, so successful cultivators are magicians and unsuccessful ones will not succeed in marriage: apparently similar beliefs are found in France until recently. Exogamy can be particularly feared and this can be reinforced by 'endo-agriculture'. In Australia, it is the reverse. There is a connection between matrilineality and patrilineality. [Lots of other examples of parallels and differences  ensue -- complex articulations, 112 – 113]

[There are implications for whether you classify societies as totemic clans or castes. They seem different, not least in terms of how developed societies are, and there is a difference in terms of support for exogamy in the first case and endogamy in the second. However there have been mixed examples and 'at least superficial analogies' (113) in Australian cases, especially those with moieties, where these are given special functions over food production. The argument ends with a triumph of homology again between two systems of difference, one in nature and one in culture. This will give two pure systems, one totemic, where there is a straightforward correspondence between species and social groups, and a more developed form where the implicit content takes over, as it were, so that clans actually become like the species animal, instead of just being homologous to them. Such a transformation 'can sometimes be directly observed' (115) {in Torres Strait Islanders}, to the extent that members of particular clans were supposed to display characteristic behaviours of their species animals, to be fighters or peaceloving, to be unpredictable or to be good runners, for example, like their clan animals — sharks, rays, dogs, or cassowaries respectively. Other beliefs are found elsewhere.

There are implications for social solidarity and marriage rules. If nature and culture are formally analogous, each domain provides a solidarity and 'exogamy furnishes the means of resolving this opposition balanced between diversity and unity' (116). If however social groups have developed on their own account, as it were, then diversity prevails over that unity and each group is seem to have particular differentiating properties with weakened ties of solidarity. It then becomes difficult to exchange women with other groups which are seen as belonging to different species].

Lévi-Strauss is aware that he might have overstressed ideology in superstructures here, stressing ideological transformations as giving rise to social ones. However, 'the reverse is in fact true. Men's conception of the relations between nature and culture is a function of modifications of their own social relations' (117). However he wants to outline a theory of superstructures, singling them out for attention and that involves bracketing other major phenomena. However, he is still 'merely studying the shadows on the wall of the Cave, without forgetting that it is only the attention we give them which lends them a semblance of reality' (117). So it is only conceptual transformations that he has outlined in the passage from exogamy to endogamy or vice versa. Similarly, there may be hybrid forms between totemic groups and castes [the real social structures] as well as hybrids between endogenous and exogenous forms, sometimes differences between moieties, including hierarchies [examples 118 – 119. Note that 'this material was collected at the time the traditional institutions no longer existed except in old informants' memories and it is plain that it is partly made up of old wives' tales' because such societies would be dysfunctional and would soon split up into independent hostile bands.

[More examples follow from considering [Asian] Indian castes (120 F). They also have totemic names and eating prohibitions, and their clan names may feature plants and animals but also manufactured goods. {There is a marvellous example of pragmatism in a clan where turmeric is prohibited but as it is 'inconvenient to be deprived of so essential a condiment , the Korra grain replaces it as the forbidden food' (121). Objects can also become clan names in parts of Africa. There is a similarity again here in that clans control totemic foods, and castes monopolise particular manufactures, and modern farmers assist totemic animals to produce themselves]

Some wonderful examples of complexity can also be introduced. Some exogamous groups develop endogenous subgroups (123). Some groups have different arrangements for natural and manufactured objects and thus for women and goods and services. Contradictions arise from 'the trap reality sets for the imagination' (123) and this is evaded 'by seeking real diversity in the natural order… The only objective model on which they can draw' (124). One model of concrete diversity is the diversity of species and the other, on the cultural plane, is the diversity of functions. Marriage exchanges inhabit both and so are particularly ambiguous — in some systems, the natural similarities outweigh the differences, so they can or cannot be exchanged between different castes, on the grounds of either natural or cultural heterogeneity (the latter being only an illusion of heterogeneity — presumably powerful nevertheless)

There are contradictions and oppositions. In a caste system, the basis is a cultural model and there is a symmetry between nature and culture. Women become equally diverse, seen as naturally diverse and thus cannot be exchanged any more than species can cross with each other. Totemic systems think the reverse, that they are defined on a natural model and exchange natural objects including women. However, heterogeneity is introduced among natural species, including women, from the point of view of culture, ultimately because 'they have the common feature that man has the power to control and increase them' (125). [Hard to grasp the point here, especially why this is a sacrifice — perhaps it prevents marriage with women from different cultures?]. Generally, the point seems to be that systems that are heterogeneous in function can be homogeneous in structure [castes]: they already have diverse functions and so marriage exchanges between diverse units would have no practical value. The converse applies for totemic groups, whose cultural function 'makes no real yield and amounts to no more than a repetition of the same illusion for all groups. They therefore have to be heterogeneous in structure each being destined for the production of women of different social species' (125) [it all seems very Darwinian]

As a French witticism: 'castes picture themselves as natural species, while totemic groups picture natural species as castes' (127)

Naturally there are complications in actual cultures, for example there may be imitation food prohibitions at the level of preparation, and although 'culture finds the field open for the great game of differentiation' (126), both positive and negative, there are limits because there is not infinite substitution of foods, for example, or occupational functions. Despite intentions, 'while species are different, no one can make them identical… Subject in the same way to human will' (127) [nor women I would have thought]

Manufactured objects look like a different set involving relations between nature and culture, but they have been managed in various ways, for example through a myth spread among hunting tribes in North America where parts of a hunted animal [once butchered] were taken as representatives of particular natural function, not the animal itself, creating 'a second nature over which man has a hold'. [So nothing stops the ingenuity of concrete thinking]

So totemism could be seen as the same as the notion of a caste via various transformations. None of the features of totemism are distinct to it, like food prohibitions. None of these are separate, and they should be seen instead as ways to 'give concrete expression to practice' (129. 'There is an analogy between sexual relations and eating in all societies' (130). There is no suggestion that 'social life is just a projection of a conceptual game taking place in the mind': conceptual schemes do govern and define practices, but it is convenient for ethnologists to study them and they should not be confused with praxis.

Praxis, 'constitutes the fundamental totality for the sciences of man' (130).'[and he agrees with Sartre and Marx but accuses Marx of blurring the distinction between praxis and practices]. But practices do not follow directly from praxis and he is not 'questioning the undoubted primacy of infrastructures'. There is always a mediator — 'a conceptual scheme by the operation of which matter and form, neither with any independent existence, are realised as structures, that is as entities which are both empirical and intelligible. It is to this theory of superstructures, scarcely touched on by Marx, that I hope to make a contribution'. We should study these with history, demography, technology, historical geography and ethnography, but not ethnology, 'for ethnology is first of all psychology' (131).

Superstructures have a dialectic like language. They have can constitutive units which must be defined by contrasting them in pairs and then used to elaborate the system which 'plays the part of a synthesising operator between ideas and facts, thereby turning the latter into signs' (131). We can then pass from 'empirical diversity to conceptual simplicity and then from conceptual simplicity to meaningful synthesis'.

He wants to illustrate this with a Yoruba myth. It starts with a series of rules involving naming a child turning on the creature or thing which it worships and which contains a prohibition about marrying someone with the same name. The creature named bears a spirit [ewaws] passed onto the descendants of the child. The son of the child takes a second spirit, 'his father's wife's animal' spirit. The son of that son takes his father's wife's third or vegetable spirit and so on. These are complicated rules but can be seen as based on a division of people into six groups — fishermen, Hunter, Farmer with their various omens, fish snake bird, quadrupeds, plants, each group has men and women so there are 12 in each case. An initial incestuous union between the brothers and sisters is represented in a table, with developments where the incestuous couple took female products of the next incestuous couple, followed by war, appropriation and so on to produce still further intermarriage and interbreeding. These original connections are preserved in the idea of inheriting various kinds of totemic spirits [131 – 132]. The whole thing shows 'institutions and rules which in their society, as in many others, are of an intellectual and deliberate character. Sensible images undoubtedly come in but they do so as symbols: they are counters in a game of combinations which consists in permuting them according to rules without ever losing sight of the empirical significance for which, provisionally, they stand'(133).

Chapter 5 Categories, elements, species, numbers.

It is a mistake to think that human culture is based on animals or celestial bodies or personifications of nature [as Boas did]. Totemism instead turns on relationships and classificatory schemes which allowed things to be grasped as organised wholes. There are different preferences for schemes of classification, but they all have a common characteristic. They must allow for other levels, analogies, whole systems of references which reflect underlying contrast between general and particular on the one hand, and nature and culture and the other. Early anthropologists mistakenly tried to isolate one level of classification, for example emphasis on natural species, and turn it into an institution, whereas it is only one form of classification among others, no more important than any other. What we should be searching for instead is a classification which is adjustable and enables a focus on all the planes from abstract to concrete, from cultural to natural 'without changing its intellectual instrument' (136).

Even though Boas saw problems in that classification is based on natural models often could not be adequately explained by sufficient distinctions between species of animals, he failed to go on to discover the actual scheme of oppositions behind mythical discourses, nor to see how classifying biological species gave access to other distinctive systems. It is more likely that zoological and botanical typologies are useful as intermediates between extreme forms -- categorical and singular, species and individuals. Species are a collection of individuals, but in relation to other species, a species is 'a system of definitions'(136).

The species has another quality — it is 'the operator which allows (and even makes obligatory) the passage from the unity of a multiplicity to the diversity of the unity' [more below, including the marvellous diagram] . Bergson apparently discovered this, or at least its logical structure, but he tended to restrict things to subjective and practical ways in which men related to the natural world [I don't understand the example, but it seems to turn on utilitarian reasons for locating individuals as unproblematic examples of species, unproblematic metonyms, taking particular items as answers to questions about what is for lunch. There is 'presumptive objectivity']. The notion of species instead offers 'the sensible expression of an objective coding', and there are parallels with communication theory in that the diversity of species can be analysed in terms of variations of the chromosomes which offer combinations of four distinct groups. Lévi-Strauss proposes that a similar approach would explain the fascination of totemism.

Back to natural sciences and the notion of kingdoms, independent and sovereign domains, with the idea that species were inert and separate classes.Societies were never seen like this, but rather as 'stages or moments in a continuous transition' among primitive peoples [examples given on 138 extend to an ingenious form of binary division which starts with dividing things from persons and animals, then into classes of plants and animals, then further subdivisions into herbaceous plants, then woody plants, then pepper plants, and eventually houseyard plants until you get specific individual variants of pepper [diagram on 141]. The system eventually gives names to each of 1625 plant types and are in some ways better than the classic Linnean categories, and closer to the 'popular botany practice by the gardener or housewife' (139). Other examples involve the classification of diseases on the same principles].


Overall we don't get separate domains but an 'all embracing dynamic taxonomy' with a perfectly homogenous structure 'the unity of which is assured by the perfect homogeneity [of its components]'. We can pass from levels, such as species to category, or system and lexicon. The universe is a 'continuum made up of successive oppositions' (139). Another example concerns seasonal rites among the Pawnee, or the classification of clans in Melanesia, or birds in the Solomons.

There may be other oppositions like high/low which can be implicit although not always explicitly formulated, and can be found in 'the ternary aspect' rather than the binary, or compounded into quinary systems (142). Dualist structures can also be decomposed into systems of two pairs. Classifiers can be convertible, such as bows and arrows, or the colours they are painted which symbolise day and night. Sometimes a ritual accompanies the production or ownership of an object because they contain a concealed contradiction — in the example, possession of moccasins is held to oppose 'evil' grass, but the land represents a particular moiety, so a ritual is needed to overcome this 'logical instability'.

A particular opposition may assume the greatest logical power and this can develop a complex grammar by means of correspondences with more variables. Sometimes there may be a 'mystic numerology' where particular numbers belong to particular moieties. It is not uncommon to see a 'meticulous rigour in the practical application of a logical system' (144), for example in a mourning ritual, the very clothes worn by the participants would express the opposition between high and low, receptacles containing food would not be covered by particular objects that may have been walked on, it was forbidden to sit on a pillow, and so on. Odd and even numbers may be treated differently and organised into opposing systems. In this way an abstract numerology may be developed as well as a concrete classification system. Even classification based on species seems to encounter no limits and can progress 'through new detotalisations and retotalisations which can take place on several planes' (147) [lots of examples — animals can be decomposed into their parts and the parts are grouped according to morphology, colour and so on]. The movements can take place on a diachronic plain as well -- eg the seasons are anticipated with signs of emaciation and renewal, or longer term temporal processes like origins are depicted. As all these examples indicate, the totem is far from just a specific biological entity, but more like 'a conceptual tool with multiple possibilities for detotalising or retotalising any domain, synchronic and diachronic, concrete or abstract, natural or cultural' (149). The tool can be applied to empirical situations, but it also always preserves general properties. It doesn't always have the same form, although 'those which come from the centre remain in the centre, those which come from the periphery on the periphery'.

The species is an important 'medial classifier' with the greatest yield. It can operate upward, heading towards elements, categories and numbers, or downward with proper names, and the levels it produces can appear in very different manners and with lots of ramifications — 'nomenclature, differences of clothing, bodily paintings or tattoos, ways of being or behaviour, privileges and prohibitions'. There are always two axes one horizontal and one vertical, similar to Saussure and syntagmatic and associative [not paradigmatic?]. However, just as with mythical and poetical thought generally, 'the principle of equivalence acts on both planes', meaning that codes can operate without any alteration by means of different lexical elements — a categoric opposition such as high/low, an elemental one like skies/earth, or a specific one like eagles/bear (150). We could in principle classify systems according to the number of categories employed and the number and choices of elements and dimensions. They might be macro and micro, according to the number of species that become totems, or whether the same species provided totems, whether systems referred purely to animals or plants, or manufactured articles, whether there was just one totem per plane or many, whether there were homogenous systems, with elements of the same type or heterogeneous systems with different elements. A full classification would require 'the aid of machines'.

We can offer only a simple [!] figure (152).

diagram of

Three species, each with different types, then parts. The parts can then be grouped according to type — 'all heads, all necks' and then we can regroup so as to restore the individual [maybe. I think I get it]. We can then move between unity, multiplicity, diversity, and identity each way. This is only a fraction of the ideal model, given that there are 2 million natural species and an unlimited number of individuals, and lots of organs or parts of the body which could be classified as well. For that matter, there are a large number of terms which could be used to describe the zoological and botanical environment. Generally though, the species and varieties actually recorded 'seem to be in the order of several hundred, around 300 to 600' (153) although there may well be more — 'the best works confirm this' [and he thinks 2000 would not be unreasonable].

One issue is whether these classifications are 'motivated at all levels', governed by 'strict and invariable relations' or after some extent showing 'a certain measure of contingency and arbitrariness at the most concrete level' (154). Certainly, as soon as we think we have closed the system, we come up against difficulty, and 'mechanisms never function perfectly; and they are also endangered by wars, epidemics, and famines… A constantly repeated battle between synchrony and diachrony from which it seems that diachrony must emerge victorious every time' (155). The nearer we get to concrete groups the more we will find arbitrary distinctions 'which are explicable primarily in terms of occurrences and events and defy any logical arrangement'. New events can be turned into totems including white fellows and sailors after first contact [some can be deified as in those accounts of Hinduism]. Some totems have been borrowed rather than established to fit features of the environment into existing systems [we are getting close to cargo cults here]. Taxonomy is '"also an art"' (156).

Saussure recognises that signs were arbitrary, but also saw that some can be 'relatively motivated' [examples include words which seem to be more strongly connected, say as the opposites of accepted words], and a general argument that words are not always equally irrational, but are governed by certain principles of order and regularity. Formally, 'some languages are more lexicological [to do with vocabulary], and others more grammatical [to do with rules] ' (156). Saussure himself saw these as two poles, and detected a general move from arbitrariness to motivation. Totemism often moves the other way around, with systematic schemes constantly broken open under the influence of elements taken from elsewhere: some are incorporated, but some seriously disturb the system. One example is what happened to Australian tribes who were resettled — they adopted common terminology and rules to try and re-harmonise tribal structures, but this was not uniform and different forms of logic were applied to try to preserve their original classifications [157 – 8]. There was 'social chaos' but a certain attempt to preserve the theoretical structure, a syncretism, which could have worked.

Generally, we can never be sure exactly what is arbitrary, because it may have a logical origin, and later elements could still have reciprocal relations with earlier ones. [He is pursuing an analogy with a tree here, and the connection between trunk and branches]. However, the part played by motivation diminishes and the role of arbitrariness increases. There are also 'statistical fluctuations'. The structure might be 'intelligible at the start' but it can reach 'a sort of inertia or logical indifference'. Even after 'the effect of multiple and varied incidents', however, these might be 'too late to prevent an attentive observer from identifying it and classifying it in a genus'.

Chapter 6 Universalisation and particularisation

[This begins with a sneaky rebuke to empiricists — Evans Prichard in particular]. It is quite wrong to see a contrast between empirical history and the attempt to demonstrate a logical system, because there is a dynamic relation between them. We start with a binary opposition and then at each pole there can be further developments, new terms, sometimes developed by 'opposition, correlation or analogy'(161). They do not have to be homogeneous, but can exist locally [what we're talking about here is the difference between tactics and strategy]. The logic between immediately associated terms is not necessarily the same for 'every link in the semantic chain' [Lévi-Strauss's own analogy is playing dominoes where you just react to the immediate piece being played, as opposed to the logic of the whole game]. So we may not see the logic of the system at every point. The general logic is 'of a different order… definable by the number and nature of the axes employed' and by the rules of transformation. These systems are relatively inert, moral less receptive to 'unmotivated factors'.

Totemic classifications are just one aspect of general systematic activity. Those based on natural objects may be important, but so might those based on cosmology or the occupational system. This is a departure from Durkheim who thought that natural totemism was the basis of everything else [attributed to Van Gennep]. However, even totemism is not an institutional reality, nothing particularly distinctive, not comparable to a species, for example, and certainly not comparable to a chemical compound [another of VG's examples]. Chemical compounds change their properties from those of their constituent elements but totemism does not — the presence or absence of specific elements have no specific effects, although there may be 'relative inflation' of particular elements like natural classifications. Even so we never know if this is down to the objective properties of the scheme or just the particular conditions under which the observation was conducted [that is attributed by the anthropologists, a point of view]

Rousseau and Comte had earlier speculations, and Comte even discussed taboo, although he did not specifically address totemism, even though it would have fitted nicely in the passage from fetishism to polytheism where he did discuss the notion of species, especially deified species. The thing about classifications, however, is they can go 'beyond their limits: either extending to domains outside the initial set — by universalisation, or alternatively by particularisation' (164) and even to individuation.

There is no restriction to categories of social life, but, for example, an extension to domains of diseases and remedies, seeing as originating in a conflict between men, animals and plants by certain American Indians [really nice, and involving some homologies]. Some classifications are extended to territory or geography — for example to particular sites which embody myths and orders of ceremonies for Australian aboriginals, or more generally the northern part of territories represent physiography, and the southern one civilisation. Both time, via myths, and space, via topography can be classified, and this can extend to the status of persons within groups, and to expand the group itself beyond its traditional confines.

Totemic classifications do enclose members of the group and tend to treat outsiders as less than human, but also prevent this closing and can 'promote an idea something like that of a humanity without frontiers' (166) and this is actually widespread, reflected in names and rituals shared among tribes, notions of shared linguistic families and beliefs, relations that extend beyond different villages or tribes, arising from shared totems. This can even extend beyond humanity to totemic animals — some Australian tribes treat dogs as brothers or sons, some American Indians do the same for dogs or horses. The classifications can also 'shrink to filter and imprison reality' and this is to be shown.

There is also a role to be played by proper names, sometimes producing a whole 'personified geography' where trails or houses or places are swapped with personal names. Thus for the Aranda, divine beings were shapeless until individuated by a particular God who taught them civilisation and the basis of classifications. He also territorialized them, and gave children particular individual features. [Lots of other examples 169]. There may well be a general 'organicism' at work, a general correspondence between the members of the society and at least some of the attributes of natural species, perhaps 'parts of the body, characteristic details, ways of being or behaviour' (169). Many languages equate parts of the body, for example — 'morphological classifiers' which are common, revealing 'anatomical detotalisation and organic retotalisation' (169). These morphologies may be further developed via a 'correlative tendency', shown, for example, in the way in which certain American Indians cut their children's hair to evoke a feature or aspect of an animal or natural phenomenon 'which serves as an eponym' [lovely diagrams on 171 — haircuts on the back of the head that look like the heads of bears, shells of turtles and so on]. Sometimes social and moral attitudes can also be incorporated, especially if there is an organising metaphor between high and low, sky and land, day and night — in one example, particular members of a moiety were deemed to be black, while others were deemed to be white and these categories influenced people's temperament vocation, and whole aspects of personal destiny [an interesting early example of colour coding (171)]

The final level classification is individuation proper. Common membership of a classification implies that everyone has a distinct position in it so there is some homology between individuals and the class and classes within superior categories: it is the same type of logical operation. That extends even to peripheral domains which might be thought at first to have escaped it and overflow the mould — and that includes diverse individual and collective beings. It is a mistake to think that they have been 'named only because they could not be signified' [attributed to Gardiner] (172).

Proper names do raise real problems, fitting them in to systems. Lots of people, including JS Mill saw them as meaningless, without signification, while the argument here is that all forms of thought are totalising and transformable into each other — so here is a problem provided by application to the concrete. Is concreteness itself irreducible?

'Almost all' the example so far have proper names from clan terms, related to the clan animal, for example either the name of the animal or mentioning its habits or attributes or characteristics, real or mythical. Sometimes they refer to associated animals or objects, and these are plentiful — for example a list of Osage proper names 'is long enough to occupy 42 quarto pages' (173) [lots of nice examples 173 – five]. Those are easy to trace back to the more general categories [and reference is made back to the great diagram linking species and individuals reproduced above].

However, sometimes there is what seems to be an arbitrary element where it is not easy to see what exactly the totem is, and the name could describe '"an action or condition that might apply equally well to other totems"' [citing Kroeber]. This is also common, a 'relative indeterminacy' affecting the retotalisation phase. However, there is still a unity in the process of detotalisation. There can also be similar prohibitions so that eponymous plants or animals may not be used as food, for example by that individual, or perhaps the name produced by an 'imaginary dissection of the body of the animal' or linguistic dissection can have behavioural consequences — [the part of the body may not be eaten, and any word that sounds like the name of the animal may not be used — I think]. This again shows 'an indisputable analogy with food prohibition' and this gives names 'an ad hoc reality comparable to those of animal or plant species' (177).

Nevertheless, there are still some people who seem to have proper names 'entirely distinct from the system of clan appellations' [citing the Iroquois]. Their names have a verb with an incorporated noun or a noun followed by an adjective such as 'in – the – centre – of -  the – sky', only ever a reference to activities, or natural phenomena celestial bodies. These names might well have been 'arbitrarily created' [but Lévi-Strauss suggests that that is one reason why tribes like the Mohawk faced a rapid decay in their clan organisation]. Also, despite their arbitrary nature, Iroquois names are still clan possessions, and then names are subject to the same distributions and classifications as clan appellations.

African tribes like the Baganda are even more confusing with thousands of names belonging to a clan, sometimes reflecting the state of mind of the parent, or their behaviour or character. These are often uncomplimentary among the Lugbara, usually because they are chosen by the grandmother, an expression of the resentment of forced marriages, perhaps, or perhaps a reluctance, due to custom. Different names may be given in different circumstances like birth order [even more complications page 180 – 81].

So it is possible to find extreme types of proper name where the name is an identifying mark establishing that an individual is a member of her preordained group, while at the other extreme the name is a free creation on the part of an individual who gives the name expressing 'a transitory and subjective state of his own' (181). Neither are really celebrating the individual — both class someone else, obeying conventions [L-S stresses it is the same as obeying conventions in naming a dog, where personal wishes and tastes form only one domain to which one is responsible: there is social convention, Kennel Club rules etc]. So proper names are still both personal and socially connected to a complex system, which may include kinship terms, sacred names and others. This combination makes an individual, while totemic appellations and kinship terms are group terms, sacred [the combination makes them profane] (183).

There can even be different categories of names — for one group there are kinship terms, names indicating status, nicknames, and true proper names, and different terms can be used in different social situations such as terms of address, during mourning, and so on. Some names are owned by clans. Some names relate only to parts of their body, or are given according to parts of the body, even the placenta. The general argument is that there are formal procedures behind all these different varieties of names — they are 'always signifying membership of an actual virtual class which must either be that the person named or of the person giving the name' (185) and all differences between names can be reduced to this distinction. [An aside points out that we only know historical figures like Vercingetorix by one name because we are ignorant of Gallic society].

Returning to what looked like a complex example, we can now decode it as a combination of two class indicators which already delimits a subset, further individuated by an even more specific name based on placental characteristics. Lévi-Strauss says it is just like the naming procedures in botany where three names are given to individual specimens, often referring to class, subclass and then the discoverer's name, which has a logical as well as a moral function, he insists. Natives might have a strange technique but it's a familiar structure, and the three name process is found in lots of other examples — it might consist of a clan name, an ordinal name referring to birth order and a military title, for example, and even the clan name might be chosen according to different procedures.

While he is here, he says that our own system of giving people proper names also has logical functions placing people in families in a way that dispels ambiguities [so that patronyms in effect become nicknames] [I read a piece recently that said most surnames in fact started as nicknames — John the Smith, Bill the Weaver, Fred long shanks and so on.]

Some names remain perplexing like those reserved for particular occurrences like twins, or children who are born adjacent to twins, birth order,or names given to novices according to their grade. Again there are parallels if eldest sons are given paternal grandfathers Christian names as was once common in Europe. 'There is an imperceptible transition from names to titles' again showing a connection with 'their structural role in a classificatory system from which it would be vain to bring to separate them' (194).

[Of course an obvious conclusion also strikes me now. Individual names are indeed located in general systems, may be several of them. So are apparently individual actions of the kind studied by qualitative researchers and ethnomethodologists. These look individual because they are studied in dyadic interactions, but of course they operationalise whole wider systems which obey social rules, not least linguistic rules, and rules of etiquette, rules of distinction, and all the rest of it. Why on earth should the local be emphasised? Just because the whole context can't be studied?]

Chapter 7 The individual as a species

[I think we're going to flesh out the implications of the great diagram above here] in one example, the Penan of Borneo, individual names contain personal names, 'teknonyms' — '"father of so and so, mother of so-and-so"', and finally what one feels like calling a necronym which expresses the kinship relation of the deceased relative to the subject: "father dead", "Niece dead", et cetera' (191).  One group has 26 distinct necronyms according to degree of kinship, age of the deceased, sex and the order of birth of children. There are rules governing the use of these names, such as when a child might be known by its proper name, the sequence in which particular relatives die, what happens if particular children die, how siblings are named and so on [192]. There is a notion here of a self and of other selves, and necronyms refer not to selves but to relationships, in this case, extinct relationships. The only slightly familiar feature is the use in English and French of terms such as the 'Widow Smith' to refer to a particular woman [further complications 194 – 5]. The overall implications are that proper names are not a category apart but used with other terms united by structural relations; these terms are class indicators [that is categories]; you can understand them by reference to a complex system; some persons may not even enter the system temporarily if there is no logical place for them, and some terms are the logical inverse of the others.

There is a common prohibition on pronouncing the names of the dead, but again this can be seen as 'a structural property of certain systems of naming' rather than a custom in its own right (197). Normal citizens fit into the system, newcomers raise a problem because it is difficult to classify them first, and so do dead people. In some systems, proper names given to those dead people literally have to be eliminated: in the system described above, dead people live on in the form of a necronym, but these enter the system only when necronyms are to be awarded. These are examples of ways in which structures are imposed 'on the continuous flux of generation', a matter of 'individuating the continuous' (199), a 'preliminary condition of classification [very complex and detailed discussion here of different systems — Lévi-Strauss is taking tremendous complexity and explaining it in terms of variations of logical system again, arguing that names are far from just terms for kinship, for example].

Another argument is that names can be generalised away from people together, for example applied to birds: in French, human names are given to sparrows (Pierrot) or to swans (Godard) [my favourites — CF English magpie, Robin redbreast and so on]. These names may have an origin as terms of address [!].

We also find this in scientific terminology — Brassica rapa, the scientific name for the turnip. This might be seen as a proper name, the name of a type, it implies a similarity to other vegetables of its kind. There is a similarity to the ways in which human proper names are constructed, because their role is often that of class indicators. Latin names for species again express something, and are not arbitrary, at least once you know the meaning of the Latin words and the rules of the taxonomy.

There are general conclusions. Proper names, like scientific names are paradigmatic, 'only with difficulty… included in the syntagmatic chain' (203). It's common to indicate this break for example in French by not including an article before them and writing them with a capital letter. The Navajo indicate the same problems in a myth [which I don't understand, 203 — apparently a conflict between a mouse and a bear arises because a mouse offends the bear by addressing them incorrectly. For botanical terms, the Navajo have a trinome: the first part is the name proper, the second describes the use, and the third its appearance, but most people only know the descriptive term, while the 'real name' is used only by priests. But it is essential to be aware of it. I still don't get it — presumably the mouse addresses the bear incorrectly and this is a metaphorical moral tale?].

We often give animals and plants some names used normally for people, although birds particularly seem to attract this kind of behaviour. It is not the same for [French] dogs who cannot be given human Christian names 'without causing uneasiness or even mild offence' (205). Birds apparently resemble humans because [although] they're so different, they are physically separated from human society so they seem to form an independent community, some sort of utopian society, loving freedom, building their own homes, having nice social relations and so on — 'a metaphorical human society', as demonstrated in countless myths and folklore. Giving them human forenames expresses this similarity via a metonym.

For French dogs, they are not independent but domestic, but so low that we should not designate them as human beings, although some Australians and Amerindians do. They should have a special series of names ['Azor', 'Medor', 'Sultan', 'Fido'], stage names, metaphorical names. Cattle are more openly treated as objects — dogs have a subjective side, for example we don't eat them. The names of cattle often referred to their colour, their bearing or temperament, again a metaphor but coming from the syntagmatic chain [the examples include 'Rousset', 'Douce'] [quite different in England].

Horses, especially racehorses occupy a different sociological position again, they are an independent society, like birds but are produced by human industry, living as isolated individuals, in a private society, like the humans who frequent racecourses. Their names are chosen with particular rules and have a wider range to the others, more eclectic, often drawing on literature. They have rigorously individualised names and draw again on the syntagm, but have no descriptive connotation. As long as they are individuated all is well, and they do not need to be described, perhaps given nicknames.

Overall, birds are metaphorical human beings, 'dogs are metonymical human beings, cattle metonymical inhuman beings, racehorses metaphorical inhuman beings' (207). We see here converse images and inverted symmetry. Bird and dog names are derived from language, although the former are real and the latter still conventional Christian names but rarely used, rather than ordinary Christian names, the names of cattle and horses derived from speech, the syntagm, with the closest those of names of cattle, descriptive, something talked about. The names of racehorses belong to speech but stripped out units of speech — naming them is 'the most frankly inhuman of all and the technique of linguistic demolition employed to construct it likewise the most barbarous' (208).

This diversion shows us the nature of proper names, even in our society, and explains why we experience other cultures with a mixture of strangeness and proximity. Going back to  one of the systems discussed earlier, which awards several names, but requires them to be distinct, which retires names when a person dies but which also requires new names, there are obvious problems in finding new names. Sometimes a proper name can be put back into circulation since names of the dead are prohibited only for a certain period. Other procedures involve prohibited names that are similar to the names of the dead instead, withdrawing them from circulation, or making them part of sacred languages. These might be the names of objects. A third option is to take sacred words and turn them back into proper names by adding a suffix. There is thus a cycling — proper names contaminate common nouns, common nouns are banished from ordinary language and become sacred language and that sacred language then turns back into proper names: semantic charges are swapped. The connection between common nouns, proper names and sacred language often turns on phonetic similarities, or there may be metonymic links.

We can even draw implications for our own language from this. Latin terms in botany offer a kind of sacred language -- parts can appear as having the function of a a proper name, as part of an overall system. The same goes with the way in which we transform names, for example the names of flowers into proper names for girls, or the ways in which horticulturalists give their discoveries the proper names from human beings, and the transfers back from the language of social distinctions [again must be French? — flowers are called Queen Elizabeth, Madame Bridget Bardot -- not always consistent in English]. He sees parallels with naming in Melanesian groups.

Again we can see in Western practices metaphors — a woman is fair as a rose, and also metonyms — a new rose is given the name of a particular Empress, a person in a special social role, or the names we give to birds, which are often diminutives, implying that birds are 'a humble and well-behaved subgroup' (213).

There are still different levels of generality — the names given to birds apply to any members of the species, but the names given to flowers only apply to one variety or even sub- variety so the field of application varies. Animal names can refer only to an individual, especially so with racehorses. But this still shows that proper names are the same thing as species names with no fundamental differences. Both express the ways in which reality is  divided up by culture, and what classification is trying to do, the sorts of appellations which are required. However, Durkheim is not correct to think that there is a social origin of all logical thoughts — rather a 'dialectical relation between social structure and systems of categories' (214) where, 'each at the cost of laborious mutual adjustments, translate certain historical and local modalities of the relations between man and the world' (214). Those relations are the 'common substratum'.

So all human beings are members of the same species and this is the same as any other animal or plant, but social life 'effects a strange transformation in this system for encourages each biological individual to develop a personality'. That leads to not only different specimens but to 'types of varieties or of species probably not found in nature'. When a particular 'personality' dies, so does 'the synthesis of ideas and modes of behaviour as exclusive and irreplaceable as the one floral species develops out of the simple chemical substances common to all species'. There are some modes of classification universally implied. They might include some forms of totemism — even we use them in a suitably humanised way, assuming, for example that 'every individual's own personality were his totem: it is the signifier of his signified being'. [so individualism is the unique combination of collective elements etc]

Proper names always derive from a paradigmatic set [that is never simply an individual matter] and thus 'form the fringe of a general system of classification' serving as the 'last act of the logical performance'  -- that performance, 'the length of the play and the number of the acts are a matter of the civilisation, not of the language' [that is the naming] (215). Proper names cannot be determined nor discovered by comparing them with other words. They are a matter of classification, especially the point at which classification is complete, that there is no further level required.

Proper names are not indices as in Peirce [nor are they the logical unit found in demonstrative pronouns, apparently for Russell]. Proper names actually reflect different levels of generalisation and this will vary with different cultures because 'each culture fixes its thresholds differently'. There is thus no 'imperceptible passage from the act of signifying to that of pointing' as Peirce believed (215). The natural sciences stop at the level of species, varieties or sub- varieties and give those proper names. Others including native sages and scientists give proper names to individual members of the social group, single positions which individuals can occupy. There is no fundamental difference 'from a formal point of view'. A botanist who gives a scientific two-stage name to a recently discovered plant and an Omaha priest who confers an available four-stage name on a new member of the group are doing the same thing: 'they know what they are doing in both cases'.

Chapter 8 Time regained [!]

[Like Bourdieu's habitus] the system both has 'internal coherence and… Practically unlimited capacity for extension' (217) there is a vertical axis connecting the general with a particular, the abstract with the concrete, and limits in either direction. Classification 'proceeds by pairs of contrasts' and ceases if it is no longer possible to establish oppositions. This means that 'the system knows no checks' but it lacks dynamism as it proceeds and eventually comes to a halt once it has 'wholly fulfilled its function'.

It is never defeated by diversity since 'reality undergoes a series of progressive purifications' leading to 'a final term in the form of a simple binary opposition' beyond which it is 'useless as well as impossible to go' and this is endlessly repeatable whether it is applied to the internal organisation of social groups which can extend to international societies, or to a mythical geography, 'an inexhaustible variety of landscapes' which finally end in a binary opposition between directions and elements, or between land and water. Qualitative diversity is seen as symbolic material of an order, and the concrete and particular and individual, 'even proper names' are still 'terms for a classification'.

Ethnologists have made the mistake of trying to pull the system to pieces, to reduce it to total institutions, especially totemism. Some have tried to overcome the problems by talking of classificatory totemism [Elkin], for example, to reintroduce classification as a form of totemism, where it is really the other way about. Comte [who might have coined the term savage mind --nah] referred to a unity of method in early history, but Lévi-Strauss sees this as a product of 'mind in its untamed state as distinct from mind cultivated or domesticated for the purpose of yielding a return', nothing primitive or archaic (219), and this appears at other locations and at certain points in history. Indeed, there is a need  to 'coexist and interpenetrate' just as natural and domesticated species can, — and the latter threatens the former. Art is an example which offers relative protection 'through indifference or inability'.

[Comte apparently called it spontaneous — LS is calling it savage]. The system is exceptional in being extensive. It analyses and synthesises to the furthest limits and at the same time mediates between two poles. Comte himself noted the analytic power and the persistent observation and exploration of spontaneous thought but was mistaken about its synthetic aspect, ignoring symbolism, as a result of 'positivist preconceptions'. On the contrary, there is a 'consuming symbolic ambition such as humanity has never again seen rivalled and… Scrupulous attention directed entirely toward the concrete' (220) and 'the implicit conviction that these two attitudes are but one', theoretical and practical point of view. The whole thing is not driven by arbitrary superstitions or the vagaries of chance, any more than the development of science was.

Comte also deals with a paralogism in arguing that all evolution really proceeds from an early theology. But Lévi-Strauss argues [a bit like Feuerbach] that this assumes a prior internalisation of the forces of nature in human beings in order to people nature with will in the first place, establishing a similarity between practical action and magical ritual action — 'inserting supplementary links through… rites' (221), misrecognised as coming from outside. This also explains why it doesn't matter if fraud or trickery are involved — 'the sorcerer never "cheats"' because it doesn't really matter what he actually does.

Religion might be a humanisation of natural laws, and magic naturalisation of human actions, but these are not alternatives, certainly not stages in an evolution, but rather two components which 'vary only in proportion', and 'each implies the other. There is no religion without magic', and vice versa, so no need to invoke vanished faculties or the development of some extra sensibility. Indians tracking a trail 'is no different from our procedure when we drive a car' (222) — we are also deciphering signs using our intellect, even though our actions are 'immeasurably increased by the mechanical power of the engine'.  A visiting exotic ethnographer would certainly declare traffic in the centre of a large town to be beyond human faculties 'and so in effect it is'. It certainly is no longer the operation of an agent on an object, but something far more complex, the confrontation of people as subjects and objects at the same time.

Attentive observation is immediately connected with symbolism in savage thought, just as it is in ordinary understanding of signs in our culture. We can see this if we discuss some examples. The first one is totemism as a possible origin of sacrifice — this looks surprisingly contrasting and incompatible as many anthropologists thought, since animal totem is at least involve an affinity between people and animals. Totemism involves identifying with an animal, but for sacrifice 'the fundamental principle is that of substitution' (224) as long as the intention persists [and in the example, the Nuer have been known to sacrifice a cucumber instead of an ox!]. The difference is whether there is 'a continuous passage between its terms', including a gradation — it would be absurd to sacrifice an ox instead of a cucumber. In the clan system, the relation between men and animals is one of homology, but in sacrifice, the relation between men and deity is not homology but 'contiguity… Successive identifications' (225)., And sacrifice is absolute, extreme, involving an intermediary object, and invoking irreversibility and reciprocity. [Lots more examples 226F]. Clans are metaphoric, sacrifices are metonymic. Sacrifices involve a third term, the deity.

The second example turns on myths of origin and there is lots of similarity, which he has collected. Elsewhere he has argued that there seems to be a global correspondence between two series, a metaphorical one, and this becomes evident only after we have suppressed elements to reveal their internal structure, a discontinuity. There is a methodological problem because we often know about myths 'only in abridged or mutilated versions' (228), often only available in pidgin, which has produced 'sketchy and ridiculous versions' (229). Nevertheless, we find a great deal of similarity: the half human half animal ancestor attributed to a totemic group, who appeared at a particular period, followed a course, performed particular actions originating geographical features which can still be seen — a particular course, water points, thickets or rocks which become of sacred value and which provide an affinity with some species or rather like Caterpillar or Kangaroo.

There are now better texts and adaptations provided by 'competent specialists' and myths where linguistic difficulties have been overcome — like America or Brazil. He takes one from the Menomini [USA] [229--30 — very nice one]. Then a South American one, the Bororo, involving tobacco and the sun and the moon.

The common characteristics are listed. First there is brevity with no apparent digressions, essential outlines no surprises. They are 'falsely aetiological' — 'the kind of explanation they give is reducible to a scarcely modified statement of the initial position; from this point of view they appear redundant. Their role appears to be demarcative… They do not really explain an origin or indicate a cause; what they do is to invoke an origin or a cause (insignificant it itself), to make the most of some detail or to "stress" a species" (231). A particular detail becomes important because it's endowed with an origin, almost modestly — one detail is privileged with the past and it only acquires significance because it forms a system.

The problem is these myths are synchronic and are therefore 'engaged in a never-ending struggle with diachrony' (231) — the myth tends to have its destination run away with it and constantly encounters new doubts when it meets actual structures. There is a problem of interpretation. Do philosophers deal with actual structures or just those that represent 'the congealed image, by means of which native philosophers give themselves the illusion of fixing a reality which escapes them?' (Sociological theorists and CRT activists have the same problem!]. This provided analytic problems for some early anthropologists who were unsure about the relationship between totemism and systems of exogamy [turning on whether totemism was a real structure, I think — Lévi-Strauss says it isn't, it's a grammar]

So there is 'a permanent conflict between the structural nature of the classification and the statistical nature of its demographic basis' (232). The totemic system is active it is 'an hereditary system of classification' in the sense that the classification is always confronted with emergent combinations. The classification is often dismantled and recombined. What is important is its function rather than its structure. It is even the case that 'the form of the structure can sometimes survive when the structure itself succumbs to events'

This can even explain the strange absence of anything that seems to relate to totemism, even in the form of historic remains, in the great civilisations of Europe and Asia. This is because they 'have elected to explain themselves by history and that this undertaking is incompatible with that of classifying things and beings (natural and social) by means of finite groups'. Totems operate with origins and derivatives, but the original can still be traced alongside the human and is always there is a system of reference so 'in theory, if not in practice, history is subordinated to system' (233).

But if society stress the history, classification into finite groups is impossible because there is a single series of interrelated elements with one term derivative from another, no homology between some external series and human history. This can be seen in some Polynesian mythologies which are already at the critical point where diachrony is about to prevail. And the myths become a prolongation of some original events, a kind of evolutionism.

The crucial difference is between what he calls '"cold" and "hot" societies' — cold ones trying to annul historical factors, and hot ones internalising the historical process and trying to make it central to their development. It might be possible to see historical sequences in different ways — as an annual cycle between the seasons, or life and death, or exchanges of goods and services. These sequences are periodically repeated and do not produce change, so they can coexist in cold societies, even if they do not succeed completely. So of course all societies are in history, and all change, but they do not all react in the same way — some accept it with good reluctance, others eagerly embrace it, others deny and try to reduce it. It is not enough just to have institutions designed to regulate the incidence of demographic factors and smooth down antagonisms, but what's needed is that possibly threatening effects should be broken up as rapidly as possible, or even prevented from forming in the first place. One procedure is to admit the historical process as a 'form without content', without significance [I think] (235) [examples seem to indicate people who go on performing activities that the totemic ancestors once performed, without ever thinking of changing or improving them. Lévi-Strauss says that the example, an Australian one, is provided by 'ethnologist born and brought up among the natives, speaking their language fluently and remaining deeply attached to them', and therefore not likely to disapprove, unlike us, of such 'obstinate fidelity to a past conceived as a timeless model' (236). Nor is this to be seen as some intellectual deficiency.

Mythical history is paradoxical in that it is both remote from and conjoined with the present, just as the mythical ancestors were different from us in being creators rather than just imitators, and there are risks of conflicts still appearing [external disruptions?] One anthropologist has talked about different sorts of rites — those of control aim to increase or restrict species or to tame phenomena by fixing spirits or substances allowed to emanate from totemic centres, managing the benefits accruing to the group; those historical rites which 'recreate the sacred and beneficial atmosphere of mythical times, the "dream age"' (237); 'mourning rites' convert dead people into ancestors, linking the past and the present, personification and myth. The churinga plays an important part giving tangible confirmation to the diachronic processes transforming past and present. It is important especially in particular groups which overemphasise the synchronic dimension [I think]: the Churinga represents the past and offers a 'means of reconciling empirical individuation and mythical confusion' (238).

Each Churinga represents the physical body of a definite ancestor, and it is given to the person believed to be the ancestor's reincarnation. Normally they are are hidden but are sometimes taken out to be handled, polished and have prayers addressed to them. They act just like documentary archives in our society, which are also semi-sacred, and occasions to recite the deeds and achievements of ancestors. We can see them as sacred without going as far as Durkheim — we can simply explain it in terms of a 'distorted reflection of a familiar image which we confusedly recognise as such without yet managing to identify it' (239). They do not need to have totemic marks necessarily drawn on them [or engraved — lovely examples on page 240]: they can be just smooth pebbles that necessarily take on an emblematic nature. It is the same argument that there is no real totem, but rather it is that the sacredness attaches to the signified not the icon. Nor is the Churinga actually the ancestor's body — the Aranda apparently told this to anthropologist but they meant it as a metaphor. Instead it is tangible proof of the relation between the ancestor and the descendant, just like archival documents in our society. They represent especially 'diachronic significance which they alone attest in a system' which is otherwise synchronic: our past would similarly disappear without our archives.

Archives are also 'the embodied essence of the event' for us, representing all the contradictions of past and present, something which surmounts the present. By this analogy we can understand 'pure history' at the centre of the Savage mind. Whether actual events are recounted accurately or just symbolically is not the issue, since much will depend anyway on historians. The point is whether the related history displays 'the characteristic traits of an historical event' (242). This will depend both on contingent status [that something happened in a particular spot] and the 'power of arousing intense and varied feelings' (243) [sounds like a counterstory!]. The whole system shows that 'so-called primitive peoples have managed to evolve not unreasonable methods for inserting irrationality in its dual aspect of logical contingents and emotional turbulence into rationality'. Classificatory systems allow the incorporation of history even if it defies the system, as so many totemic myths do. Some societies have the most elaborate social organisation in marriage requiring 'the efforts of mathematicians for their interpretation' a cosmology that 'astonishes philosophers' and pursue 'lofty theorising' in those domains, but they have a history which resembles 'conducted tours to Goethe's or Victor Hugo's house, the furniture of which inspires emotions as strong as they are arbitrary. As in the case of the Churinga, the main thing is not that the bed is the selfsame one on which it is proved Van Gogh slept: all the visitor asks is to be shown it' (244).

Chapter 9 History and dialectic

[This begins with a discussion with Sartre which I can't follow very well because I have not really read much Sartre, especially the particular book at issue here, his Critique of Dialectical Reason. There are some marvellous general arguments about how social science should proceed however]

The first problem is the extent to which thought can be both 'anecdotal and geometrical' (245) and yet at the same time 'dialectical'. The 'savage mind totalises' much more than Western thinking, while dialectical reason is not concerned with 'pure seriality' or schematisation'. Nothing human, or even living remains alien to the savage mind. Paradoxically this should be 'the real principle of dialectical reason', but this is not the case with Sartre.

Sartre actually seems to have two conceptions of dialectical reason. One is opposed to analytic reason, while the other saw the two kinds as complementary, different routes to the same truth. The first conception discredits scientific knowledge, although, ironically, it results from Sartre's own exercise of analytic reason, defining, classifying and opposing. If the two are complementary, the difficulty lies in showing how they are opposed and why one is superior to the other.

Sartre seems to think that there is a reality in itself revealed by dialectical reason, existing independently of analytic reason, either as an antagonist or a complement. This might be rooted in Marx, although, for Lévi-Strauss, the opposition between these two sorts of reason is relative and corresponds to at tension within human thought which may exist in practice but has no basis 'de jure' (246). Indeed, for Lévi-Strauss, dialectical reason is a kind of bridge, reforming analytical reason trying to endlessly account for 'language society and thought', for understanding life itself, rousing analytic reason to action, trying to transcend it, something additional in analytic reason. In particular, it is 'the necessary condition for its to venture to undertake the resolution of the human into the nonhuman' [otherwise known as the material, but Lévi-Strauss means this in quite a distinctive sense, potentially to encompass the very cells and atoms of which we are made].

For him, 'the ultimate goal of the human sciences [is] not to constitute, but to dissolve man' (247). Ethnography gets at invariance beyond empirical diversity, for example, often found by observing differences. Nor can we just posit a general humanity instead of particular ones. Culture must be ultimately reintegrated into nature and finally 'life within the whole of its physico – chemical conditions' [and a note says that the opposition between nature and culture once seemed so important to him but 'now seems to be of primarily methodological importance''. However 'dissolve' here does not mean destroy the constituents, rather to think of rearranging and recovering them. If we reduce phenomena we must not impoverish them but preserve their 'distinctive richness and originality'.

We must rethink preconceived ideas about levels, especially about general humanity. If we achieve proper ethnographic reduction, the general humanity to which we proceed 'will bear no relation to any one may have formed in advance'. The same goes for life as a function of inert matter — 'the latter has properties very different from those previously attributed to it'. Reduction is not a matter of finding superior or inferior levels. An allegedly superior level has to communicate some of its richness to the inferior one, and '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' (248) [greater explanatory power].

It is not a matter of opposing self to others, or man to the world, because the truths learnt at one level are important at the other [a note claims that this applies even for mathematical truths, since they tell us about the functioning of the mind, the 'activity of the cells of the cerebral cortex' which in turn tells us something about the nature of the mind,  the thing therefore about the nature of things and ultimately 'an internalisation of the cosmos and thus 'the structure of what lies outside in a symbolic form']. It follows that 'anthropology [of his kind is] the principle of all research'.

For Sartre, it is to be replaced by history and dialectic. He deals with so-called primitive societies only in terms of having a short-term dialectic, which placed them near to biology and produces a stunted and deformed version of humanity. This in turn seems to have derived either because more advanced forms have discovered and begun to colonise primitive forms, or granted them meaning in terms of more advanced humanity. In either case, 'the prodigious wealth and diversity of habits, beliefs and customs is allowed to escape' (249), as has the claim that each society that has existed has already claimed to represent 'the essence of all the meaning and dignity of which human society is capable… a moral certainty' comparable to our own. It requires 'egocentricity and naïveté' to take one particular mode of existence as the most human — 'the truth about man resides in the system of their [the modes] differences and common properties'.

[Now some perceptive remarks about identity and introspection, again in the context of Sartre's method I think, but with excellent applicability to my interests in the politics of identity]. 'He who begins by steeping himself in the allegedly self evident truths of introspection never emerges from them'. If you get caught 'in the snare of personal identity', you 'shut the door on knowledge of man: written or undervalued "confessions" form the basis' [of all social science knowledge — he actually says of all ethnographic research]. You become the 'prisoner of [your] cogito' even though this might be the cogito  of the group and period, rather than a Cartesian psychological and individual one [rebuke to the notion of a black episteme]: instead of a timeless consciousness, we have one for 'each subject's group and period'. We have exchanged 'one prison for another'. [Sartre's — but many another] 'view of the world and man has the narrowness which has been traditionally credited to closed societies', partly because he traces 'the distinction between the primitive and the civilised with the aid of gratuitous contrasts' [compare the distinction between the indigenous and the Eurocentric]. In his case this is a reflection of the opposition between myself and others, found in Sartre and also in the formulations of 'a [any] Melanesian savage'. In this case, it is because Sartre… 'Separates his own society from others' (250), basically because he chooses as a starting point to describe social reality, mere 'secondary incidentals of life', trivial examples such as 'strikes, boxing matches, football matches, bus stop queues' rather than going for things that will disclose the foundations. [Certainly the case with modern examples of trivial covert symbolic racism].

Sartre does do something similar to what anthropologists do, in putting himself in the place of people living in different societies and trying to understand the pattern of their intentions, but this is not really theoretical enough and does not proceed towards totalisation, and effort taken for granted by anthropologists although still a novelty to other social scientists. However, we have to do it properly after constituting our object — 'the properly scientific work consists in decomposing and then re-composing on a different plane' (250), and phenomenology [in the most general sense, looking at subjective reactions and so on] is not enough. The procedure is exciting and seductive, but it would be a mistake to assume that 'others are wholly dialectical in every respect' and that there can be no other element of reality in them.

This appears, for example when he tries to explain the life and thoughts of exotic societies. [I think the argument here is that you can only comprehend exotic societies through your own constituted dialectic, a relationship between native thought and his knowledge of it, which 'repeats all the illusions of theorists of primitive mentality', particularly that 'the savage should possess "complex understanding" and should be capable of analysis and demonstration' (251)]. To take one example, Sartre is sceptical about the diagram produced by a particular native explaining the functioning of rules and kinship systems, which he insists is just a piece of manual work — Lévi-Strauss says that the same could be said of a diagram made by a professor at the École Polytechnic demonstrating a proof on the blackboard.

Lévi-Strauss says he has also fallen into some of these areas in his own work on the elementary structures of kinship by seeking out some unconscious genesis for matrimonial exchange. Instead, he should have made a distinction between the praxis of groups, exchange expressed 'spontaneously and forcefully' and the 'conscious and deliberate rules' with which the practice was codified and controlled. He now realises that the latter aspect is far more important than observers had realised, and thus that analytic reason is also important.

Even if these biases were not apparent, there would still be a problem in that the 'unconscious teleology' connecting exotic societies and our own would be unavailable through human history, especially those bits which depend on 'linguistics and psychoanalysis and which rest on the interplay of biological mechanisms... And psychological ones' (252). Language in Sartre is far too simple, ignoring those elements 'outside (or beneath) consciousness and will… Human reason which has its reasons and of which man knows nothing', 'an unreflective totalisation'. The human subject may be able to speak this language and make himself understood, but may not be aware of the totalisation of linguistic laws or 'have access to the same experience in other, not necessarily human, but living beings'.

This might qualify for the '"progressive – regressive" method in Sartre, although for anthropologists it happens twice over — observing the data of experience, analysing it in the present and then trying to grasp its historical antecedents, then bringing all these 'facts' together into a 'meaningful totality', only to begin a second stage at a different level. Here, possible analytic reason becomes possible dialectical reason, where 'this unforeseen object is assimilated to others'(253), a 'novel totality', which will help us 'descry other horizons and other objects'. It will require a lot of retracing of steps, doubling back on itself in order to 'preserve the contact which that experienced totality which serves both as its ends and means'. This will be a form of verification not just a demonstration in Sartre of how conscious beings pose problems — they also show dialectical reasoning, an extension of analytic reason transformation of its axiomatic. However, 'dialectical reason can account neither for itself nor for analytical reason'.

To some extent, there is also always a contraction in meaning [empirical detail?], a change to the 'conjectural', 'truth for science fiction' for Sartre. This is inherent in 'every attempt an explanation' for LS, and the issue is whether the meaning preserved is of more value than that which has been relinquished. Like Marx and Freud, the argument here is that [humanly available] meaning is never 'the right one' (254) — superstructures are 'faulty' even though they have become accepted socially, so we cannot use mere historical consciousness to get to the truth. We cannot know analytically what has produced them [I think], but must resort to 'hypothetical moves about which it is impossible to know'. There is a particular paradox here if we use our own developed consciousness to distinguish primitive from civilised societies, because our own historical consciousness is 'ahistorical' in the sense of being 'an abstract schema', a 'synchronic totality' developing history of a particular kind, no different from the way in which so-called 'primitives' relate to the 'eternal past' — 'in Sartre's system, history plays exactly the part of a myth'.

For example, the real question is 'under what conditions is the myth of the French Revolution possible?' It must be believed in if contemporary French people are to play their part as historical agents, and Sartre analyses the implications in a way which is rich and 'most suited to inspire practical action'. But that does not make it necessarily 'the truest', since 'truth is a matter of context', it is as experienced. In our particular period of history there is still the idea of the 'congruence between practical imperatives and schemes of interpretation', but this 'golden age of historical consciousness has [perhaps] already passed'. We can still focus on the French Revolution like this, but so we could have done on earlier French movements [the Fronde], but this is ceased to offer a coherent image as we learned more about it and the complexity of the alliances involved on either side of the struggle [255].

Historical events offer only 'a spurious intelligibility attaching to a temporary internality' [between the meaning of past events and contemporary meanings]. These are important connections and cannot be avoided, but we all know in a different 'register' that we are living 'a myth' and this will become clear to future historians and perhaps even to ourselves in a few years. All grand meanings regress until one can only say '"it is thus and not otherwise"' to quote Sartre, that we can gain all we can reasonably hope for from history, that there is no other form of transcendence.

However, contemporary philosophers often value history above all the other human sciences, and have 'an almost mystical conception of it' (256) [and LS wants to include Sartre in this after all]. This is not the case for anthropology, where history is a complementary study, to one that arranges vanished societies in space. These are equivalent perspectives, but this is denied by those who want to attach special prestige to the temporal dimension is offering some superior intelligibility.

Temporal dimension seems to offer a continuous transformation discontinuity which confirms 'the evidence of inner sense'. We seem to gain some 'flashes of insight into internalities', some connection 'outside ourselves, with the very essence of change'. This is 'an illusion sustained by the demands of social life — and consequently a reflection of the external on the internal'. Experience is not 'apodictic'. It can be demonstrated that there is 'a twofold antinomy in the very notion of an historical fact' (257). Where exactly did anything actually take place? Episodes can be reduced into 'a multitude of individual psychic movements' and further into 'cerebral, hormonal or nervous phenomena, which themselves have reference to the physical or chemical order'. The historian 'constitutes [these events as historical facts] by abstraction and as though under the threat of an infinite regress'.

There is an obvious selection, 'for a truly total history would confront [historians] with chaos'. There is a multitude of individuals who totalise differently, summarising inexhaustible number of incidents. Even universal history still only juxtaposes a few local histories. There is always selection of 'regions, periods, groups of men and individuals'[treated these days as a licence to develop endless professional commentary]. History is always 'history – for', those who treat accounts as significant [a note says that Sartre would agree with this, but traces the development of groups only through a dyadic development of egos, which LS calls 'intellectual cannibalism']. History is therefore always partial and biased, but each account must be treated as 'equally true' [leaving only decisionism, or some strange claim that you can totalise 'the set of partial totalisations'] (258).

History like all knowledge has to employ some code to analyse its object, and has particular problems with grasping the nature of continuous reality [a note explains that it is impossible to actually reach the continuous without infinite regress, and if you attempt it you must quantify events and therefore restrict temporality since quantification means you have to treat each event 'as if it were the result of a choice between possible pre-existents' (258)]. Historians claim there is no code, but of course they rely on chronology, dates, which remain 'its sine qua non', despite recent denials. In fact chronological coding is quite complex, since dates are not only ordinal but cardinal numbers, expressing a distance. This is usually glossed when 'we use a large number of dates to code some periods of history; and fewer for others' (259) which really expresses 'the pressure of history'. In '"hot"' periods, 'numerous events appear as differential elements', but in others, 'very little or nothing took place'. Sometimes, nothing at all seems to have taken place, as when we put dates in sets like 'first… millennium'.

Thus strictly speaking, historical dates have no meaning in themselves, except when they are placed in classes of dates, and here we are hinting at 'complex relations of correlation and opposition with other dates… frequency… a corpus or domain of history'. The best analogy is 'a wireless with frequency modulation… Frequencies of impulses proportional to its variations'. Thus 'history is a discontinuous set composed of domains of history' each of which has a characteristic frequency and different notion of before and after. It's difficult to pass between the barriers thus 'the dates appropriate to each class are irrational in relation to all those of other classes'. The dates do not just form a series 'they are of different species' (260) like the dates in prehistory compared to those of contemporary history. There is always a reference back to other classes to add intelligibility — so that a particular domain located in a century relates to earlier and later centuries [the example is the notion of the history of the 17th century as '"annual"' (261)].

There are different levels of [explanatory] power as well. 'Biographical and anecdotal history' is at the bottom of the scale, 'low powered… Not intelligible in itself' having to be transferred to a higher power of history. There is no smooth dovetailing between these levels however since we lose richness in terms of information [a note refers to anti-histories in  some strange French example which I do not understand]. There is always a choice between 'history which teaches us more and explains less, and history which explains more and teaches less' (262). The only way forward is to go down to the level of individuals and their motivations, 'an infra-historical domain in the realms of psychology and physiology' or up, into prehistory and general evolution which will lead to 'biology, geology and finally cosmology'.

An alternative is to avoid the whole argument that history is the only way to preserve 'transcendental humanism' and thus 'preserve the illusion of liberty on the plane of the "we"' merely by giving up the "I"s that are to obviously wanting in consistency' [Sartre's particular project?]. We need to see that history is not tied to 'man' nor to any particular object, but consists only as a method to catalogue any structure whatever. It is therefore a point of departure in seeking intelligibility, but it's not sufficient, and has no claim to be privileged.

It is already found 'in the savage mind' but is not developed fully there. There is timelessness, the world is seen as both a synchronic and diachronic totality, through a multitude of images of the world. This leads to the development of mental structures which both resemble the world and help to understand it — 'in this sense savage thought can be defined as analogical thought' (263). History seems quite domesticated by comparison, seeking to unify rather than pursue analogies, to transcend, close gaps, and dissolve differences. This opens the door to determinism as [someone called Auger has argued].

It is no good to suggest that praxis by concrete individuals will serve to oppose analytic abstract continuity because this is also derivative as only the conscious mode of grasping processes and domesticating them [maybe]. Of course reason 'develops and transforms itself in the practical field', but thought has to develop first as an objective structure [I think he means biological objects here, in the brain]. Sartre thinks that praxis is a kind of primaeval functionalism [I think, with references to Robinsonade form of argument to relate to the origin of social rituals (264)].

By contrast, proper anthropology has noticed the commonalities between initiation rites in quite diverse societies. There is a pattern where novices are symbolically killed taken away, put to the test and then reborn. When they are returned, the parents 'simulate all the phases of a new delivery' and even begin to re-educate the child as if it were an infant. It would be wrong to see this as embedded in [functionalist?] praxis, however because this would be to miss the significance of notions of death and birth which came first and provided material for all sorts of conceptualisation not oriented to practical returns. Initiation rituals mark thoughts which takes the words involved quite seriously.

The same might be said of certain taboos and parents-in-law, the frequent prohibition of any contact with them. This has been explained in various ways, [as an extension of the incest taboo, to prevent intergenerational rivalry, to express particular reservations about the wife's maternal grandmother — quite different interpretations] Lévi-Strauss thinks there is one underlying interpretation which we can arrive at if we consider the parallel with taboos and approaching our social superiors directly — the wife-giver is the same as the social superior. Again, the same ideas and attitudes are embodied differently, and natives are more systematic, seeing more implications while we operate with more detached pieces.

We are now far more interested in the language which has a limited vocabulary can express 'any message by combinations of oppositions between its constituent of units… [Where] contents are indissoluble from form… [With a] universe made up of meanings'. Indeed, 'the laws of savage thought reign once more' (267) we now realise that this is not a negative way of looking at the world but one which is on a par with 'modern theorists of documentation' [a note explains that this means people who decode various works, seemingly sacred works]. It is the same way that physical science had to discover a semantic universe to describe the characteristics of objects. We now realise that primitive conceptualisations are coherent and 'the very one demanded in the case of an object to is elementary structure presents the picture of a discontinuous complexity' (268).

'The savage mind is logical in the same sense the same fashion as ours', trying to establish a knowledge of the universe 'in which it recognises physical and semantic properties simultaneously'. 'Its thought proceeds through understanding and not affectivity', using distinctions and oppositions. It is 'a quantified form of thought'.

Is savage thought mistaken in taking 'mere manifestations of physical determinism for messages'? [In the first place there is now confusion about this in modern information theory, I think he's saying]. Secondly, totemism might be illusory, but it does take sensible properties of plants or animals as elements of a message or signs, and this eventually enabled the correct identification of elements at the microscopic level, so totemism was a kind of preliminary form of science, 'principles of interpretation whose heuristic value in accordance with reality have been revealed to us only through very recent inventions'. Despite mistakes, they did get some of the properties right, and were right to insist on dimensions outside the consciousness of transmitters and receivers, a 'universe of information [as] part of an aspect of the natural world' (269). The approach has been relatively successful for millennia, allowing 'men to approach the laws of nature by way of information'.

The properties of the savage mind are not the same as those of science, and approach the physical world from opposite ends, one 'supremely concrete' the other 'supremely abstract,' sensible qualities on the one hand and formal properties on the other. However, they were combined in the Neolithic period which provided a sensible order for the whole of subsequent civilisation, and 'contemporary science is the fruit'. There may now be a new crossing of the paths, via the detour of communication theory, which will 'have contributed to legitimise the principles of savage thought and to re-establish it in its rightful place'.