The Elementary Forms of the Religious Life (1912)

[Excerpt from Robert Alun Jones. Emile Durkheim: An Introduction to Four Major Works. Beverly Hills, CA: Sage Publications, Inc., 1986. Pp. 115-155.]

Outline of Topics

  1. Durkheim's Two Problems
  2. Defining Religion
  3. The Most Primitive Religion
    1. Animism
    2. Naturism
    3. Totemism
  4. Totemic Beliefs: Their Nature, Causes, and Consequences
  5. Totemic Rites: Their Nature and Causes
  6. The Social Origins of Religion and Science
  7. Critical Remarks

Durkheim's Two Problems

Durkheim's primary purpose in The Elementary Forms was to describe and explain the most primitive1 religion known to man. But if his interests thus bore some external similarity to those of the ethnographer or historian, his ultimate purpose went well beyond the reconstruction of an archaic culture for its own sake; on the contrary, as in The Division of Labor and Suicide, Durkheim's concern was ultimately both present and practical: "If we have taken primitive religion as the subject of our research," he insisted, "it is because it has seemed to us better adapted than any other to lead to an understanding of the religious nature of man, that is to say, to show us an essential and permanent aspect of humanity."2

But if Durkheim's goal was thus to understand modern man, why did he go to the very beginning of history? How can the crude cults of the Australian aborigines tell us anything about religions far more advanced in value, dignity, and truth? And if he insisted that they can, wasn't he suggesting that Christianity, for example, proceeds from the same primitive mentality as the Australian cults? These questions were important, for Durkheim recognized that scholars frequently focused on primitive religions in order to discredit their modern counterparts, and he rejected this "Voltairean" hostility to religion for two reasons. First, alluding to the second chapter of The Rules, Durkheim insisted that such hostility was unscientific; it prejudges the results of the investigation, and renders its outcome suspect. Second, and more important, he considered it unsociological; for it is an essential postulate of sociology that no human institution can rest on an error or a lie. If an institution is not based on "the nature of things," Durkheim insisted, it encounters a resistance in nature which destroys it; the very existence of primitive religions, therefore, assures us that they "hold to reality and express it." The symbols through which this reality is expressed, of course, may seem absurd; but we must know how to go beneath the symbol, to uncover the reality which it represents, and which gives it its meaning: "The most barbarous and the most fantastic rites and the strangest myths translate some human need, some aspect of life, either individual or social. The reasons with which the faithful justify them may be, and generally are, erroneous; but the true reasons," Durkheim concluded, "do not cease to exist" and it is the duty of science to discover them."3

In this sense, all religions are "true"; but if all religions are thus equal with respect to the reality they express, why did Durkheim focus on primitive religions in particular? Briefly, he did so for three "methodological" reasons. First, Durkheim argued that we cannot understand more advanced religions except by analyzing the way they have been progressively constituted throughout history; for only by placing each of the constituent elements of modern religions in the context within which it emerged can we hope to discover the cause which gave rise to it.4 In this analysis, as in Cartesian logic, the first link of the chain was the most important; but for Durkheim, this link at the foundation of the science of religions was not a "conceptual possibility" but a concrete reality based on historical and ethnographic observations. Just as biological evolution has been differently conceived since the empirical discovery of monocellular beings, therefore, religious evolution is differently conceived depending upon what concrete system of belief and action is placed at its origin.

Second, Durkheim suggested that the scientific study of religion itself presupposed that the various religions we compare are all species of the same class, and thus possess certain elements in common: "At the foundation of all systems of belief and all cults," Durkheim thus argued,

there ought necessarily to be a certain number of fundamental representations or conceptions and of ritual attitudes which, in spite of the diversity of forms which they have taken, have the same objective significance and fulfill the same functions everywhere. These are the permanent elements which constitute that which is permanent and human in religion; they form all the objective contents of the idea which is expressed when one speaks of religion in general.5

Again, therefore, Durkheim was trying to answer a time-honored philosophical question (the "essential nature" of religion) by new, sociological means (the ethnography of primitive societies); and the special value of such ethnographies was that they captured religious ideas and practices before priests, prophets, theologians, or the popular imagination had had the opportunity to refine and transform them:

That which is accessory or secondary... has not yet come to hide the principal elements. All is reduced to that which is indispensable to that without which there could be no religion. But that which is indispensable is also that which is essential, that is to say, that which we must know before all else.6

Primitive religions are privileged cases, Durkheim thus argued, because they are simple cases.

But if this simplicity of primitive religions helps us to understand its nature, it also helps us to understand its causes. In fact, as religious thought evolved through history, its initial causes became overlaid with a vast scheme of methodological and theological interpretation which made those origins virtually imperceptible. The study of primitive religion, Durkheim thus suggested, is a new way of taking up the old problem of the "origin of religion" itself -- not in the sense of some specific point in time and space when religion began to exist (no such point exists), but in the sense of discovering "the ever-present causes upon which the most essential forms of religious thought and practice depend."7

This description and explanation of the most primitive religion, however, was only the primary purpose of The Elementary Forms; and its secondary purpose was by far the most ambitious of Durkheim's attempts to provide sociological answers to philosophical questions. At the base of all our judgments, Durkheim began, there are a certain number of ideas which philosophers since Aristotle have called "the categories of the understanding" -- time, space, class, number, cause, substance, personality, and so on.8 Such ideas "correspond to the most universal properties of things. They are like the solid frame which encloses all thought; this does not seem to be able to liberate itself from them without destroying itself, for it seems that we cannot think of objects that are not in time and space, which have no number, etc."9

How are these ideas related to religion? When primitive religious beliefs are analyzed, Durkheim observed, these "categories" are found, suggesting that they are the product of religious thought; but religious thought itself is composed of collective representations, the products of real social groups. These observations suggested to Durkheim that the "problem of knowledge" might be posed in new, sociological terms. Previous efforts to solve this problem, he began, represent one of two philosophical doctrines: the empiricist doctrine that the categories are constructed out of human experience, and that the individual is the artisan of this construction, and the a priorist doctrine that the categories are logically prior to experience, and are inherent in the nature of the human intellect itself.

The difficulty for the empirical thesis, Durkheim then observed, is that it deprives the categories of their most distinctive properties -- universality (they are the most general concepts we have, are applicable to all that is real, and are independent of every particular object) and necessity (we literally cannot think without them); for it is in the very nature of empirical data that they be both particular and contingent. The a priorist thesis, by contrast, has more respect for these properties of universality and necessity; but by asserting that the categories simply "inhere" in the nature of the intellect, it begs what is surely the most interesting and important question of all: "It is necessary," Durkheim insisted, "to show whence we hold this surprising prerogative and how it comes that we can see certain relations in things which the examination of these things cannot reveal to us."10 In sum, if reason is simply a variety of individual experience, it no longer exists; but if its distinctive properties are recognized but not explained, it is set beyond the bounds of nature and thus of scientific investigation. Having planted these (allegedly) formidable obstacles in the paths of his philosophical adversaries, Durkheim then offered his frustrated reader an attractive via media: "... if the social origin of the categories is admitted," he suggested, "a new attitude becomes possible which we believe will enable us to escape bath of the opposed difficulties."11

How, then, does the hypothesis of the social origin of the categories overcome these obstacles? First, the basic proposition of the a priorist thesis is that knowledge is composed of two elements -- perceptions mediated by our senses, and the categories of the understanding -- neither of which can be reduced to the other. By viewing the first as individual representations and the second as their collective counterparts, Durkheim insisted, this proposition is left intact: for "between these two sorts of representations there is all the difference which exists between the individual and the social, and one can no more derive the second from the first than he can deduce society from the individual, the whole from the part, the complex from the simple."12 Second, this hypothesis is equally consistent with the duality of human nature -- just as our moral ideals are irreducible to our utilitarian motives, so our reason is irreducible to our experience. In so far as we belong to society, therefore, we transcend our individual nature both when we act and when we think. Finally, this distinction explains both the universality and the necessity of the categories -- they are universal because man has always and everywhere lived in society, which is their origin; and they are necessary because, without them, all contact between individual minds would be impossible, and social life would be destroyed altogether: "... society could not abandon the categories to the free choice of the individual without abandoning itself. If it is to live," Durkheim concluded, "there is not merely need of a satisfactory moral conformity, but also there is a minimum of logical conformity beyond which it cannot safely go."13

But one might still object that, since the categories are mere representations of social realities, there is no guarantee of their correspondence to any of the realities of nature; thus we would return, by a different route, to a more skeptical nominalism and empiricism.14 Durkheim's rationalist and rather metaphysical answer is that society itself is a part of nature, and "it is impossible that nature should differ radically from itself... in regard to that which is most essential. The fundamental relations between things -- just that which it is the function of the categories to express cannot be essentially dissimilar in the different realms."15

Defining Religion

In order to describe and explain the most primitive religion known to man, Durkheim observed, we must first define the term "religion" itself: otherwise we risk drawing inferences from beliefs and practices which have nothing "religious" about them, or (and this was the greater danger to Durkheim) of leaving many religious facts to one side without understanding their true nature.16 In fact, Durkheim had already made such an attempt in "Concerning the Definition of Religious Phenomena" (1899), where he argued that religion consists of obligatory beliefs united with definite practices which relate to the objects given in the beliefs."17 While this definition achieved a number of aims,18 however, Durkheim soon became displeased with its overriding emphasis on "obligation"; and, as he later acknowledged,19 the definition offered in 1912 is significantly different.20

Following The Rules21 and Suicide,22 Durkheim's 1912 definition is reached by a two-step process. First, he insisted, we must free the mind of all preconceived ideas of religion, a liberation achieved in The Elementary Forms through a characteristic "argument by elimination": "it is fitting," Durkheim suggested, "to examine some of the most current of the definitions in which these prejudices are commonly expressed, before taking up the question on our own account."23 Second, Durkheim proposed to examine the various religious systems we know in their concrete reality, in order to determine those elements which they have in common; for "religion cannot be defined except by the characteristics which are found wherever religion itself is found."24

The first of the prejudicial definitions of religion to be eliminated by this procedure was that governed by our ideas of those things which surpass the limits of our knowledge -- the "mysterious," the "unknowable," the "supernatural" -- whereby religion would be "a sort of speculation upon all that which evades science or distinct thought in general."25 Durkheim saw at least four discernible difficulties in such a definition. First, while he admitted that the sense of mystery has played a considerable role in the history of some religions, and especially Christianity, he added that, even in Christianity, there have been periods -- e.g., the scholastic period (tenth to fifteenth centuries), the seventeenth century, etc. -- in which this sense was virtually non-existent. Second, while Durkheim agreed that the forces put in operation by some primitive rite designed to assure the fertility of the soil or the fecundity of an animal species appear "different" from those of modern science, he denied that this distinction between religious and physical forces is perceived by those performing the rite; the abyss which separates the rational from the irrational, Durkheim emphasized, belongs to a much later period in history. Third, and more specifically, the very idea of the "supernatural" logically presupposes its contrary -- the idea of a "natural order of things" or "natural law" -- to which the supernatural event or entity is presumably a dramatic exception; but the idea of natural law, Durkheim again suggested, is a still more recent conception than that of the distinction between religious and physical forces.26 Finally, Durkheim simply denied that the object of religious conceptions is that which is "exceptional" or "abnormal"; on the contrary, the gods frequently serve to account for that which is constant and ordinary -- "for the regular march of the universe, for the movement of the stars, the rhythm of the seasons, the annual growth of vegetation, the perpetuation of species, etc. It is far from being true," Durkheim concluded, "that the notion of the religious coincides with that of the extraordinary or the unforeseen."27

The second prejudicial definition rejected by Durkheim was that based upon the idea of "gods"28 or, more broadly, "spiritual beings."29 The relations we can have with such beings, Durkheim observed, are determined by the nature attributed to them -- they are conscious beings, and thus we can act upon them only through conscious processes (invocations, prayers, offerings, sacrifices); and since the object of religion is to govern our relations with such beings, there can be no religion except where such conscious processes are at work. The difficulty for this definition, Durkheim insisted, is that it fails to acknowledge two categories of undeniably religious facts. First, there are great religions (e.g., Buddhism, Jainism, Brahminism, etc.) from which the idea of gods or spirits is almost absent, and in which the "conscious processes" indicated above play a minor role at best. Second, even within those religions which do acknowledge such beings, there are many rites which are completely independent of that idea, and in some cases the idea is itself derived from the rite rather than the reverse.30 "All religious powers," Durkheim concludes, "do not emanate from divine personalities, and there are relations of cult which have other objects when uniting man to a deity. Religion is more than the idea of gods or spirits, and consequently cannot be defined exclusively in relation to these latter."31

Definition by the ideas of "spiritual beings" and "the supernatural" thus eliminated, Durkheim turned to the construction of his own definition. Emphasizing that religion is less an indivisible whole than a complex system of parts, he began by dividing these parts into rites (determined modes of action) and beliefs (collective representations); and since rites can be distinguished from other actions only by their object, and the nature of that object is determined by the beliefs, Durkheim insisted on defining the latter first. "All known religious beliefs," he observed, "present one common characteristic: they presuppose a classification of all the things, real and ideal, of which men think, into two classes or opposed groups, generally designated by two distinct terms which are translated well enough by the words profane and sacred.32 The characteristic by which the latter is distinguished from the former, moreover, is simply that it is distinguished absolutely:33 "In all the history of human thought," Durkheim emphasized, "there exists no other example of two categories of things so profoundly differentiated or so radically opposed to one another."34 Durkheim thus arrived at his preliminary definition of the essential parts of any religious system: sacred things are those isolated and protected by powerful interdictions; profane things are those which, according to those interdictions, must remain at a distance from their sacred counterparts; religious beliefs are representations which express the nature of sacred things and their relations, either with one another or with profane things; religious rites are rules of conduct which prescribe how one should behave in the presence of sacred things; and finally, where "a certain number of sacred things sustain relations of co-ordination or subordination with each other in such a way as to form a system having a certain unity," the beliefs and rites thus united constitute a religion.35

The seemingly insuperable obstacle to the immediate acceptance of this definition was its subsumption of a body of facts ordinarily distinguished from religion -- i.e., magic. Indeed, magic is also composed of beliefs and rites, myths, dogmas, sacrifices, lustrations, prayers, chants, and dances as well; and the beings and forces invoked by the magician are not only similar to those addressed by religion, but are frequently the same. Yet historically, magic and religion have frequently exhibited a marked repugnance for one another,36 suggesting that any definition of the latter should find some means of excluding the former. For Durkheim, this means was Robertson Smith's insistence, in his Lectures on the Religion of the Semites,37 that religion was a public, social, beneficent institution, while magic was private, selfish, and at least potentially maleficent. "The really religious beliefs," Durkheim could thus argue, "are always common to a determined group or `Church,' which makes a profession of adhering to them and of practicing the rites connected with them.... The individuals which compose it feel themselves united to each other by the simple fact that they have a common faith."38 The belief in magic, by contrast,

does not result in binding together those who adhere to it, nor in uniting them into a group leading a common life.... Between the magician and the individuals who consult him, as between these individuals themselves, there are no lasting bonds which make them members of the same moral community, comparable to that formed by the believers in the same god or the observers of the same cult.39

Hence Durkheim's definition: " A religion is a unified system of beliefs and practices relative to sacred things, that is to say, things set apart and forbidden -- beliefs and practices which unite into one single moral community called a Church, all those who adhere to them."40

The Most Primitive Religion

Armed with his "preliminary definition" of religion, Durkheim set out in search of its most primitive, elementary form. Almost immediately, however, another difficulty arose -- even the crudest religions of which we have any historical or ethnographic knowledge appear to be the products of a long, rather complicated evolution, and thus exhibit a profusion of beliefs and rites based upon a variety of "essential" principles. To discover the "truly original" form of the religious life, Durkheim observed, it is thus necessary "to descend by analysis beyond these observable religions, to resolve them into their common and fundamental elements, and then to seek among these latter some one from which the others were derived."41 The problem, in short, was less one of describing and explaining a system of observable beliefs and practices than of constructing the hypothetical, essential origin from which these later religions presumably derived.

This was a problem for which two contrary solutions had been proposed, based upon the two common elements found universally among the observable religions. One set of beliefs and practices, for example, is addressed to the phenomena of nature, and is thus characterized as naturism; while a second body of religious thought and action appeals to conscious spiritual beings, and is called animism. The problem of accounting for the confusing properties of the observable religions thus resolved itself into two mutually contradictory evolutionary hypotheses: either animism was the most primitive religion, and naturism its secondary, derivative form; or the cult of nature stood at the origin of religion, and the cult of spirits was but a peculiar, subsequent development.42 It was through the critical examination of these traditional theories -- another argument by elimination -- that Durkheim hoped to reveal the need for a new theory altogether.


According to the animistic theory, the idea of the human soul was first suggested by the contrast between the mental representations experienced while asleep (dreams) and those of normal experience. The primitive man grants equal status to both, and is thus led to postulate a "second self" within himself, one resembling the first, but made of an ethereal matter and capable of traveling great distances in short periods of time. The transformation of this soul into a spirit is achieved with death, which, to the primitive mind, is not unlike a prolonged sleep; and with the destruction of the body comes the idea of spirits detached from any organism and wandering about freely in space. Henceforth, spirits are assumed to involve themselves, for good or ill, in the affairs of men, and all human events varying slightly from the ordinary are attributed to their influence. As their power grows, men increasingly consider it wise to conciliate their favor or appease them when they are irritated, whence come prayers, offerings, sacrifices -- in short, the entire apparatus of religious worship. Reasoning wholly by analogy, the primitive mind also attributes "second selves" to all non-human objects -- plants, animals, rivers, trees, stars, etc. -- which thus account for the phenomena of the physical world; and in this way, the ancestor cult gives rise to the cult of nature. In the end, Durkheim concluded, "men find themselves the prisoners of this imaginary world of which they are, however, the authors and models."43

If this animistic hypothesis is to be accepted as an account of the most primitive religion, Durkheim observed, three parts of the argument are of critical significance: its demonstration that the idea of the soul was formed without borrowing elements from any prior religion; its account of how souls become spirits, and thus the objects of a cult; and its derivation of the cult of nature from ancestor worship. Doubts concerning the first were already raised by the observation, to be discussed later,44 that the soul, though independent of the body under certain conditions, is in fact considerably more intimately bound to the organism than the animistic hypothesis would suggest. Even if these doubts were overcome, moreover, the animistic theory presumes that dreams are liable to but one primitive interpretation -- that of a "second-self" -- when the interpretive possibilities are in fact innumerable; and even were this objection removed, defenders of the hypothesis must still explain why primitive men, otherwise so unreflective, were presumably driven to "explain" their dreams in the first place.

The "very heart of the animist doctrine," however, was its second part -- the explanation of how souls become spirits and objects of a cult; but here again Durkheim had serious doubts. Even if the analogy between sleep and death were sufficient to suggest that the soul survives the body, for example, this still fails to explain why the soul would thus become a "sacred" spirit, particularly in light of the tremendous gap which separates the sacred from the profane, and the fact that the approach of death is ordinarily assumed to weaken rather than strengthen the vital energies of the soul. Most important, however, if the first sacred spirits were souls of the dead, then the lower the society under investigation, the greater should be the place given to the ancestor cult; but, on the contrary, the ancestor cult is clearly developed only in relatively advanced societies (e.g., China, Egypt, Greece and Rome) while it is completely lacking among the most primitive Australian tribes.

But even if ancestor worship were primitive, Durkheim continued, the third part of the animist theory -- the transformation of the ancestor cult into the cult of nature -- is indefensible in itself. Not only is there little evidence among primitives of the complicated analogical reasoning upon which the animist hypothesis depends; neither is there evidence among those practicing any form of nature worship of those characteristics -- anthropomorphic spirits, or spirits exhibiting at least some of the attributes of a human soul -- which their derivation from the ancestor cult would logically suggest.

For Durkheim, however, the clearest refutation of the animistic hypothesis lay in one of its unstated, but implied, consequences; for, if it were true, not only would it mean (as Durkheim himself believed) that religious symbols provide only an inexact expression of the realities on which they are based; far more than this, it would imply that religious symbols are products of the vague, ill-conceived hallucinations of our dream-experience, and thus (as Durkheim most certainly did not believe) have no foundation in reality at all. Law, morals, even scientific thought itself, Durkheim observed, were born of religion, long remained confounded with it, and are still somewhat imbued with its spirit; it is simply inconceivable, therefore, that "religions, which have held so considerable a place in history, and to which, in all times, men have to receive the energy which they must have to live, should be made up of a tissue of illusions."45 Indeed, the animistic hypothesis is inconsistent with the scientific study of religion itself; for a science is always a discipline applied to the study of some real phenomenon of nature, while animism reduces religion to a mere hallucination. What sort of science is it, Durkheim asked, whose principle discovery is that the subject of which it treats does not exist?


In sharp contrast to animism, the naturistic theory46 insisted that religion ultimately rests upon a real experience -- that of the principal phenomena of nature (the infinity of time, space, force, etc.) -- which is sufficient to directly arouse religious ideas in the mind. But religion itself begins only when these natural forces cease being represented in the mind in an abstract form, and are transformed into personal, conscious spirits or gods, to whom the cult of nature may be addressed; and this transformation is (allegedly) achieved by language. Before the ancient Indo-European peoples began to reflect upon and classify the phenomena of nature, Durkheim explained, the roots of their language consisted of very general types of human action (pushing, walking, climbing, running, etc.). When men turned from the naming and classifying of actions to that of natural objects, the very generality and elasticity of these concepts permitted their application to forces for which they were not originally designed. The earliest classes of natural phenomena were thus metaphors for human action -- a river was "something that moves steadily," the wind was "something that sighs or whistles," etc. -- and as these metaphors came to be taken literally, natural forces were quite naturally conceived as the product of powerful, personal agents. Once these agents had received names, the names themselves raised questions of interpretation for succeeding generations, producing the efflorescence of fables, genealogies, and myths characteristic of ancient religions. Finally, the ancestor cult, according to this theory, is purely a secondary development -- unable to face the fact of death, men postulated their possession of an immortal soul which, upon separation from the body, was gradually drawn into the circle of divine beings, and eventually deified.47

Despite the contrast mentioned above, Durkheim's objections to this naturistic hypothesis followed much the same line as those objections to its animistic counterpart. Leaving aside the numerous criticisms of the philological premises of the naturistic theory, Durkheim insisted that nature is characterized not by phenomena so extraordinary as to produce a religious awe, but by a regularity which borders on monotony. Moreover, even if natural phenomena were sufficient to produce a certain degree of admiration, this still would not be equivalent to those features which characterize the "sacred", and least of all to that "absolute duality" which typifies its relations with the "profane." The primitive, in any case, does not regard such forces as superior to his own; on the contrary, he thinks he can manipulate them to his own advantage by the exercise of certain religious rites. And in fact, the earliest objects of such rites were not the principal forms of nature at all, but rather humble animals and vegetables with whom even the primitive man could feel himself at least an equal.48

Durkheim's major objection, however, was that the naturistic theory, like animism, would reduce religion to little more than a system of hallucinations. It is true, he admitted, that primitive peoples reflect upon the forces of nature from an early period, for they depend on these forces for their very survival. For precisely this reason, however, these forces and the reflections upon them could hardly be the source of religious ideas; for such ideas provide a palpably misleading conception of the nature of such forces, so that any course of practical activity based upon them would surely be unsuccessful, and this in turn would undermine one's faith in the ideas themselves. Again, the important place granted to religious ideas throughout history and in all societies is evidence that they respond to some reality, and one other than that of physical nature.49


Whether from dreams or from physical nature, therefore, animism and naturism both attempt to construct the idea of the sacred out of the facts of our common, individual experience; and for Durkheim, whose argument again parallels Kant's attack on empiricist ethics, such an enterprise is simply impossible: "A fact of common experience," he insisted, "cannot give us the idea of something whose characteristic is to be outside the world of common experience."50 Durkheim's largely negative assessment of rival theories of religious origins thus led to his first positive conclusion: "Since neither man nor nature have of themselves a sacred character," he argued,

they must get it from another source. Aside from the human individual and the physical world, there should be some other reality, in relation to which this variety of delirium which all religion is in a sense, has a significance and an objective value. In other words, beyond those which we have called animistic and naturistic, there should be another sort of cult, more fundamental and more primitive, of which the first are only derived forms or particular aspect.51

This more fundamental and primitive cult is totemism.

The peculiar set of beliefs and practices known as totemism had been discovered among American Indians as early as 1791; and though repeated observations for the next eighty years increasingly suggested that the institution enjoyed a certain generality, it continued to be seen as a largely American, and rather archaic, phenomenon. J. F. McLennan's articles on "The Worship of Animals and Plants" (1870-71) showed that totemism was not only a religion, but one from which considerably more advanced religions had derived; and L.H. Morgan's Ancient Society (1877) revealed that this religion was intimately connected to that specific form of social organization that Durkheim had discussed in The Division of Labor -- the division of the social group into clans. As the same religion and social organization were increasingly observed and reported among the Australian aborigines, the documents accumulated until James Frazer brought them together in Totemism (1887). But Frazer's work was purely descriptive, making no effort to understand or explain the most fundamental aspects of totemism. The pivotal work in the explanation and interpretation of this institution, therefore, was Robertson Smith's Lectures on the Religion of the Semites (1889), which made totemism the origin of sacrifice, and thus of the ritual apparatus of higher religions generally; and in The Golden Bough (1890), Smith's protégé Frazer connected the same ideas to the gods of classical antiquity and the folklore of European peasants. All these works, however, were constructed out of fragmentary observations, for a true totemic religion had not yet been observed in its complete state. This hiatus was filled, however, in Baldwin Spencer and F.J. Gillen's Native Tribes of Central Australia (1899), a study of totemic clans almost definitively primitive; and, together with the studies they stimulated, these observations were incorporated within Frazer's four-volume compendium, Totemism and Exogamy (1910).

The initial contribution of The Elementary Forms to this rapidly growing literature was simply its methodological approach. As a member of the "anthropological" school, for example, Frazer had made no effort to place the various religious systems he studied within their social and historical context; rather, as the name of the school implies, he assumed that man has some sort of innate, religious "nature" regardless of social conditions, and thus "compared" the most disparate beliefs and rites with an eye to their most superficial similarities.52 But for the sociologist, Durkheim emphasized, social facts vary with the social system of which they are a part, and cannot be understood when detached from that system. For this reason, two facts from different societies cannot be usefully compared simply because they seem to resemble one another; in addition, the societies themselves should resemble each other -- be varieties of the same species.53 Moreover, since the number of societies with which a sociologist can be genuinely familiar is quite limited, and since, in any case, he regarded the alleged "universality" of totemism as a question of only residual interest, Durkheim ultimately concentrated on the aboriginal societies of central Australia almost exclusively.54 These societies, indeed, suited Durkheim's purposes admirably -- the ethnographic reports of their totemic institutions were easily the most complete, their structural features were all of a single type (the "single-segment" societies of The Division of Labor and The Rules), and, since this type of societal organization was the most rudimentary known, it seemed to Durkheim the best place to search for that "most primitive" religion whose description and explanation was the central purpose of The Elementary Forms.

Totemic Beliefs: Their Nature, Causes, and Consequences

But where, in such totemic societies, was one to look first? At their rites, as had Robertson Smith and the early Frazer? Or at their beliefs, following Tylor and Frazer's later work? The fact that myths are frequently constructed after the rite in order to account for it suggested the first; while recognition that rites are often the sole expression of antecedent beliefs argued for the second. On this contemporary controversy in the scientific study of religion, Durkheim ultimately leaned heavily toward the second alternative; and on the ground that it is impossible to understand a religion without a firm grasp of its ideas, his discussion of Australian totemism in The Elementary Forms thus began with its beliefs.55

The most fundamental of these beliefs is that the members of each clan56 consider themselves bound together by a special kind of kinship, based not on blood, but on the mere fact that they share the same name. This name, moreover, is taken from a determined species of material objects (an animal, less frequently a plant, and in rare cases an inanimate object) with which the clan members are assumed to enjoy the same relations of kinship. But this "totem" is not simply a name; it is also an emblem, which, like the heraldic coats-of-arms, is carved, engraved, or designed upon the other objects belonging to the clan, and even upon the bodies of the clan members themselves. Indeed, it is these designs which seem to render otherwise common objects "sacred," and their inscription upon the bodies of clan members indicates the approach of the most important religious ceremonies.

The same religious sentiments aroused by these designs, of course, are aroused by the members of the totemic species themselves. Clan members are thus forbidden to kill or eat the totemic animal or plant except at certain mystical feasts (see below), and the violation of this interdiction is assumed to produce death instantaneously. Moreover the clan members themselves are "sacred" in so far as they belong to the totemic species, a belief which gives rise to genealogical myths explaining how men could have had animal and even vegetable ancestors. Durkheim thus rejected McLennan's interpretation of totemism as a form of animal worship; for man belongs to the sacred world himself, and thus his relations with his totem are much more like those uniting members of the same family.57

Totemism is thus a religion in which three classes of things -- the totemic emblem, the animal or plant, and the members of the clan -- are recognized as sacred; but in addition, totemism constitutes a cosmology, in which all known things are distributed among the various clans and phratries, so that everything is classified according to the social organization of the tribe.58 In short, because men themselves are organized socially, they are able to organize things according to their societal model; thus one of the essential "categories of the understanding" -- the idea of class -- appears to be the product of certain forms of social organization.59 But this is not merely a logical or cognitive classification, it is also moral -- all things arranged in the same clan are regarded as extensions of the totemic animal, as "of the same flesh," and thus as themselves "sacred" in some degree.60 Finally, since all of these beliefs clearly imply a division of things between sacred and profane, we may call them "religious"; and since they appear not only related, but inseparably connected, to the simplest form of social organization known, Durkheim insisted that they are surely the most elementary forms of the religious life.

How, then, were these beliefs to be explained?61 The first step toward an answer to this question, Durkheim suggested, is to recognize that, while all the things discussed above (emblems, animals, clan members, and all other objects) arc sacred in different degrees (their sacredness declines in roughly that order), they are all sacred in the same way; thus, their religious character could hardly be due to the special properties of one or the other, but rather is derived from some common principle shared (albeit, again, in different degrees) by all. Totemism, in short, is not a religion of emblems or animals or men at all, but rather of an anonymous, impersonal "force,"62 immanent in the world and diffused among its various material objects.

But, surely, such a conception surpasses the limits of the primitive mind? On the contrary, Durkheim argued, whether it is described as mana, wakan, or orenda, this belief in a diffused, impersonal force is found among the Samoans, the Melanesians, various North American Indian tribes, and (albeit less abstracted and generalized) among the totemic clans of central Australia.63 Indeed, this explains why it has been impossible to define religion in terms of mythical personalities, gods, or spirits; for these particular religious things are simply individualized forms of this more impersonal religious principle, and their sacredness is thus non-intrinsic. And quite aside from its purely religious significance, Durkheim argued that this was the original form under which the modern, scientific idea of force was conceived.64

To explain totemism is thus to explain this belief in a diffused, impersonal force. How might such a belief arise? Obviously, not from sensations aroused by the totemic objects themselves, Durkheim argued, for these objects -- the caterpillar, the ant, the frog, etc. -- are hardly of a kind to inspire powerful religious emotions: on the contrary, these objects appear to be the symbols or material expressions of something else. Of what, then, are they the symbols? Durkheim's initial answer was that they symbolize both the "totemic principle" and the totem clan; but if this is the case, then surely that principle and the clan are one and the same thing: "The god of the clan, the totemic principle," he insisted, "can therefore be nothing else than the clan itself, personified and represented to the imagination under the visible form of the animal or vegetable which serves as totem."65

This hypothesis -- that god is nothing more than society apotheosized--was supported by a number of characteristically Durkheimian arguments. It was insisted, for example, that a society has all that is necessary to arouse the idea of the divine, for it is to its members what a god is to his worshippers. It is both physically and morally superior to individuals, and thus they both fear its power and respect its authority; but society cannot exist except in and through the individual conscience, and thus it both demands our sacrifices and periodically strengthens and elevates the divine "principle" within each of us -- especially during periods of collective enthusiasm, when its power is particularly perceptible.66 Indeed, it is during such extremely rare gatherings of the entire Australian clan, Durkheim suggested, that the religious idea itself seems to have been born, a fact which explained why its most important religious ceremonies continue to be observed only periodically, when the elan as a whole is assembled. It is this succession of intense periods of "collective effervescence" with much longer periods of dispersed, individualistic economic activity, Durkheim suggested, which gives rise to the belief that there are two worlds -- the sacred and the profane -- both within us and within nature itself.67

But how does this belief give rise to totemism? Briefly, the individual who is transported from his profane to a sacred existence in a gathering of the clan seeks some explanation for his altered, elevated state. The gathering of the clan itself is the real cause, though one too complex for the primitive mind to comprehend; but all around him, the clan member sees symbols of precisely that cause -- the carved engraved images of the totem -- and fixes his confused social sentiments on these clear, concrete objects, from which the physical power and moral authority of society thus seem to emanate. Just as the soldier who dies for his flag in fact dies for his country, so the clan member who worships his totem in fact worships his clan.

To the classical formula Primus in orbe deos fecit timor -- the fear-theory defended in various ways by Hume, Tylor, and Frazer -- Durkheim thus added a decisive, if not entirely original, dissent. Following Robertson Smith -- indeed, it was probably this idea, seized upon in his anti-Frazerian mood of 1900, that so dramatically altered Durkheim's conception of religion itself -- Durkheim insisted that the primitive man does not regard his gods as hostile, malevolent, or fearful in any way whatsoever; on the contrary, his gods are friends and relatives, who inspire a sense of confidence and well-being. The sense thus inspired, moreover, is not an hallucination, but is based on reality; for however misunderstood, there actually is a real moral power -- society -- to which these beliefs correspond, and from which the worshipper derives his strength.68

This argument -- the very heart of The Elementary Forms69 -- was also intimately bound to Durkheim's important conception of the role of symbols in society. Their utilitarian value as expressions of social sentiments notwithstanding, Durkheim's more ambitious claim was that such symbols serve to create the sentiments themselves. For collective representations, as we have seen, presuppose the mutual reaction of individual minds upon one another, reactions inexplicable in the absence of collective symbols; and, once formed, such representations would quickly dissipate in the absence of symbols which serve to maintain them in the individual mind. Thus, society, "in all its aspects and in every period of its history, is made possible only by a vast symbolism."70

This explanation of totemism in turn helps us to understand a phenomenon that had recently been discussed by Lucien Lévy-Bruhl in Les Fonctions mentales dans les sociétés inférieures (1910)71 -- that "law of participation" whereby primitive peoples ignore the distinctions between animals, vegetables, and inanimate objects, granting rocks a sex, for example, or a star a soul, thus giving rise to elaborate mythologies in which each being partakes of the properties of others. Inexplicable on the basis of ordinary experience -- nowhere do we see beings "mixing their natures" or "metamorphizing themselves into each other" -- such participation was explained by Durkheim as a consequence of the symbolic representations just described: once the clan became "represented" by a species of animal or plant, the latter were thought of as relative of men, and both were assumed to "participate in the same nature."72

Finally, like the concept of mana, this notion of "participation" had significance for the evolution of scientific as well as purely religious thought. To say that one thing is the cause of another, Durkheim explained, is to establish relations between them, to suggest that they are bound together by some natural, internal law. Like Hume, Durkheim insisted that sensations alone can never disclose such law-like connections; and like Kant, therefore, he argued that the human reason must supply them, thus enabling us to understand cause and effect as necessary relations. The great achievement of primitive religion, Durkheim then suggested, is that it constructed the first representation (the "law of participation") of what these "relations of kinship" might be, thus rescuing man from his enslavement to mere appearance, and rendering science and philosophy possible; and religion could do this, he added, only because it is a mode of collective thought, which imposes a new way of representing reality for the old manipulation of purely individual sensations. Between religion and science, Durkheim thus concluded, there can be no abyss; for, while the former applies its logical mechanisms to nature more awkwardly than the latter, both are made up of the same collective representations.73

Throughout his discussion of the nature and causes of totemic beliefs, Durkheim insisted that no idea of the soul, spirits, or gods plays any role. In order to complete this discussion, therefore, it was necessary to show how such ideas -- universal among the known religions -- could have evolved out of "the more essential conceptions" just described. Every known society, for example, acknowledges the existence of the human soul -- a second, ethereal self, which dwells within and animates the body; and since the Australian aborigines provide the most primitive instance of this belief, Durkheim's search for its origin began by asking how the aborigines themselves explained it.

According to the Australians, Durkheim observed, the souls which enter and animate the bodies of new-born children are not "special and original creations"; on the contrary, they are the old souls of dead ancestors of the clan, whose reincarnation explains the phenomena of conception and birth. To such ancestors superhuman powers and virtues are attributed, rendering them sacred; and most important, they are conceived under the form not of men, but of animals and plants. Durkheim thus concluded that the human soul is simply a form of "individualized mana," the totemic principle incarnate, and the most primitive form of that conception of the "duality of human nature" which has perplexed the philosophers and theologians of more advanced societies for centuries.

The last point is important, for Durkheim claimed that this explanation of the belief in the soul helps us to understand two more advanced ideas: the theological conception of immortality, and the philosophical idea of personality. The first belief, Durkheim argued, cannot be accounted for by the moral demand for a future, just, retribution,74 for primitive peoples make no such demand; neither can it be explained by the desire to escape death,75 an event to which the primitive is relatively indifferent, and from which, in any case, his particular notion of immortality would offer little relief; and finally, it cannot be explained by the appearance of dead relatives and friends in our dreams,76 an occurrence too infrequent to account for so powerful and prevalent a belief. The failure of these explanations, Durkheim added, is particularly embarrassing in that the idea of the soul itself does not seem to imply its own survival, but rather seems to exclude it -- since the soul is intimately connected with the body, the death of the latter would seem to bode ill for the former. This embarrassment is relieved, however, if one accepts Durkheim's explanation, in which the belief in the immortality of the soul and its subsequent reincarnations is literally required if the phenomena of conception and birth are to be explained. And in holding this belief, Durkheim again asserted, the primitive is not misled; for the soul is simply the individualized representation of the clan, and the clan does outlive its individual members. The belief in the immortality of the soul is thus the earliest, symbolic means whereby men represented to themselves the truth that society continued to live while they must die.77

The philosophical idea of personality would seem to have posed greater difficulties for Durkheim; for the modern notion of what is "personal" seems to imply what is "individual" --an association which would surely confound any sociological explanation. But Durkheim insisted that the terms were in no way synonymous, a distinction clearly evident in their most sophisticated philosophical formulations. In his suggestion that all reality is composed of "monads," for example, Leibniz had emphasized that these psychic entities are personal, conscious, autonomous beings; but he had also insisted that these consciousnesses all express the same world; and since this world is itself but a system of representations, each particular consciousness is but the reflection of the universal consciousness, the particularity of its perspective being explained by its special location within the whole.78 Similarly, for Kant, the cornerstone of the personality was the will, that faculty responsible for acting in conformity with the utterly impersonal reason; and to act in accordance with reason was to transcend all that is "individual" within us (our senses, appetites, inclinations, etc.).79 For both Leibniz and Kant, therefore, a "person" was less an individual subject distinguished from all others than a being enjoying a relative autonomy in relation to its immediate environment; but this, Durkheim concluded, is precisely the description of the primitive idea of the soul -- that individualized form of society within, yet independent of, the body.80

The subsequent evolution of totemic beliefs is one from souls to spirits, spirits to "civilizing heroes," and heroes to "high gods," in which the focus of religious worship becomes increasingly powerful, personal, and international. Since the idea of souls is inexplicable without postulating original, "archetypal" souls from which the others are derived, for example, the primitive imagines mythical ancestors or "spirits" at the beginning of time, who are the source of all subsequent religious efficacy.81 When the clans come together for the tribal initiation ceremonies, the primitive similarly seeks an explanation for the homogeneity and generality of the rites thus performed; and the natural conclusion is that each group of identical ceremonies was founded by one great ancestor, the "civilizing hero" of the clan, who is now venerated by the larger tribe as well.82 And where the tribe as a whole, gathered at such initiation ceremonies, acquires a particularly powerful sentiment of itself, some symbol of this sentiment is sought; as a result, one of the heroes is elevated into a "high god," whose authority is recognized not only by the tribe thus inspired, but by many of its neighbors. The result is a truly "international" deity, whose attributes bear a marked similarity to those of the higher religions of more advanced civilizations.83 But this "great tribal god," Durkheim emphasized, retracing his evolutionary steps, "is only an ancestral spirit who finally won a pre-eminent place. The ancestral spirits are only entities forged in the image of the individual souls whose origin they are destined to explain. The souls, in their turn, are only the form taken by the impersonal forces which we found at the basis of totemism, as they individualize themselves in the human body. The unity of the system," Durkheim concluded, "is as great as its complexity."84

Totemic Rites: Their Nature and Causes

Despite an occasional dalliance with the ritual theory of myth, Durkheim's most consistent position was that the cult depends upon the beliefs; but he also insisted that beliefs and rites are inseparable not only because the rites are often the sole manifestation of otherwise imperceptible ideas, but also because they react upon and thus alter the nature of the ideas themselves. Having completed his extensive analysis of the nature, causes, and consequences of totemic beliefs, therefore, Durkheim turned to a somewhat shorter discussion of the "principal ritual attitudes" of totemism.

Sacred things, as we have seen, are those rather dramatically separated from their profane counterparts; and a substantial group of totemic rites has as its object the realization of this essential state of separation. In so far as these rites merely prohibit certain actions or impose certain abstentions, they consist entirely of interdictions or "taboos";85 and thus Durkheim described the system formed by these rites as the "negative cult."86 The interdictions characterizing these rites were in turn divided into two classes: those separating the sacred from the profane, and those separating sacred things from one another according to their "degree of sacredness";87 and even the first class alone assumes a variety of forms -- certain foods are forbidden to profane persons because they are sacred, while others are forbidden to sacred persons because they are profane; certain objects cannot be touched or even looked at; certain words or sounds cannot be uttered; and certain activities, particularly those of an economic or utilitarian character, are forbidden during periods when religious ceremonies are being performed. For all their diversity, however, Durkheim argued that all these forms are reducible to two fundamental interdictions: the religious life and the profane life cannot coexist in the same place, and they cannot coexist in the same unit of time.

Although literally defined in terms of these interdictions, however, the negative cult also exercises a "positive" function -- it is the condition of access to the positive cult. Precisely because of the abyss which separates sacred things from their profane counterparts, the individual cannot enter into relations with the first without ridding himself of the second. In the initiation ceremony, for example, the neophyte is submitted to a large variety of negative rites whose net effect is to produce a radical alteration of his moral and religious character, to "sanctify" him through his suffering, and ultimately to admit him to the sacred life of the clan. But here, again, religion is only the symbolic form of society which, while augmenting our powers and enabling us to transcend ourselves, demands our sacrifice and self-abnegation, suppresses our instincts, and does violence to our natural inclinations. There is a ruthless asceticism in all social life, Durkheim argued, which is the source of all religious asceticism.88

But if this is the function of the negative cult, what is its cause? In one sense. of course, this system of interdicts is logically implied by the very notion of "sacredness" itself, for respect habitually begets inhibition; but many things acquire and retain our respect without becoming the objects of religious interdictions. The peculiar attribute of sacred things which renders them, in particular, the objects of the negative cult is what Durkheim called "the contagiousness of the sacred" -- religious forces easily escape their original locations and flow, almost irresistibly, to any objects within their range. In so doing, of course, they contradict their own essential nature, which is to remain separated from the profane; and thus a whole system of restrictions is necessary in order to keep the two worlds apart.89

But how was this contagiousness itself to be explained?90 Observing that such contagiousness is characteristic of those forces or properties (heat, electricity, etc.) which enter a body from without rather than being intrinsic to it, Durkheim reminded his reader that this is precisely what his own theory implies -- religious forces do not inhere in physical nature, but rather are brought to it by the collective representations of society. Indeed, it is by an initial "act of contagion" (see above) that ordinary objects receive their sacred character in the first place; and it is natural that, in the absence of rigorous interdictions, they should lose this character just as easily.91

This brings us to the most crucial phase of Durkheim's treatment of totemic rites, that based upon those materials which, in the 1899-1901 period, so dramatically altered his understanding of religion.92 Until 1899, as Durkheim himself observed, the negative cult just described was virtually all that was known of the religious aspect of totemism; but the negative cult, as we have seen, contains no reason for existing in itself, and merely introduces the individual to more positive relations with sacred things. The significance of Spencer and Gillen's Native Tribes of Central Australia (1899), therefore, was that it described one ceremony in particular that exhibits the essential features of the "positive cult" found in more advanced religions -- the Arunta Intichiuma ceremony.

In central Australia, Durkheim explained, there are two sharply divided seasons: one is long and dry, the other short and very rainy. When the second arrives, the vegetation springs up from the ground, the animals multiply, and what had been a sterile desert abounds with luxurious flora and fauna; and it is at the moment when this "good" season seems near at hand that the Intichiuma is celebrated. Every totemic clan has its own Intichiuma, and the celebration itself has two phases. The object of the first is to assure the abundance of that animal or plant which serves as the clan's totem, an object obtained by striking together certain sacred stones (sometimes drenched with the blood of clan members), thus detaching and scattering grains of dust which assure the fertility of the animal or plant species. The second phase begins with an intensification of the interdictions of the negative cult -- clan members who could ordinarily eat their totemic animal or plant if they did so in moderation now find that it cannot be eaten or even touched -- and concludes with a solemn ceremony in which representatives of the newly increased totemic species are ritually slaughtered and eaten by the clan members, after which the exceptional interdictions are lifted and the clan returns to its normal existence.93

To Durkheim, the significance of this system of rites was that it seemed to contain the essential elements of the most fundamental rite of the higher religions -- sacrifice; and equally important, it seemed in large part to confirm the revolutionary theory of the meaning of that rite put forward by Robertson Smith twenty-four years earlier.94 The more traditional theory of sacrifice, epitomized in Tylor's Primitive Culture (1871), was that the earliest offerings were "gifts" presented to the god by his worshippers; but such a theory, Durkheim observed, failed to account for two important features of the rite: that its substance was food, and that this food was shared by both gods and worshippers at a common feast. Noting that in many societies such commensality is believed to create (and re-create) a bond of kinship, Smith had suggested that the earliest sacrifices were less acts of renunciation and expiation than joyous feasts, in which the bond of kinship uniting gods and worshippers was periodically reaffirmed by participation in the common flesh.

Insisting (pace Smith) that it was participation in sacred flesh that rendered the rite efficacious, Durkheim argued that the ceremony concluding the second phase of the Intichiuma was precisely such a rite; and in the Australian rite, he added, the object of such communion was clear -- the periodic revivification of that "totemic principle" (society) which exists within each member of the clan and is symbolized by the sacrificial animal or plant. This in turn explained the temporal aspect of the rite -- the totemic principle would seem most thoroughly exhausted after a long, dry period, and most completely renewed just after the arrival of the "good season," and analogous practices were found among many, more advanced peoples: "... the Intichiuma of the Australian societies," Durkheim concluded, "is closer to us than one might imagine from its apparent crudeness."95

But if Durkheim shared Smith's view that the earliest sacrifices were acts of communion, he did not share his view that this was all they were, nor did he share Smith's reasons for holding these views in the first place. For Smith was a devout Scottish Calvinist who found the very idea that the gods receive physical pleasure from the offerings of mere mortals a "revolting absurdity," and insisted that this conception had no part in the original meaning of the rite, emerging only much later with the institution of private property.96 In the first phase of the Intichiuma, however, Durkheim found precisely this idea -- the totemic species requires the performance of certain rites in order to reproduce itself -- and thus he suggested that the complete sacrifice is both an offering and a communion.97

To Durkheim, moreover, this was not a minor, antiquarian quibble, for his subsequent explanation of this "mutual interdependence" of gods and worshippers was a major part of his sociological theory of religion. We have already seen, for example, how the "logic" of the Intichiuma corresponds to the intermittent character of the physical environment of central Australia -- long dry spells punctuated by heavy rainfall and the reappearance of animals and vegetation. This intermittency, Durkheim now added, is duplicated by the social life of the Australian clans -- long periods of dispersed, individual economic activity, punctuated by the intensive communal activity of the Intichiuma itself.98 Since sacred beings exist only at the sufferance of collective representations, we should expect that their presence would be most deeply felt precisely when men gather to worship them, when they "partake of the same idea and the same sentiment"; in a sense, therefore, the Australian was not mistaken -- the gods do depend upon their worshippers, even as the worshippers depend upon their gods, for society can exist only in and through individuals, even as the individual gets from society the best part of himself.99

In addition to this primitive form of sacrifice, Durkheim discussed three other types of positive rites -- imitative, representative, and piacular -- which either accompany the Intichiuma or, in some tribes, replace it altogether. The first consists of movements and cries whose function, guided by the principle that "like produces like," is to imitate the animal or plant whose reproduction is desired. These rites had been interpreted by the "anthropological school" (Tylor and Frazer) as a kind of "sympathetic magic," to be explained, as in Jevons's account of the contagiousness of the sacred, by the association of ideas; and as it had been to Jevons, Durkheim's response to Tylor and Frazer was that a specific, concrete social phenomenon cannot be explained by a general, abstract law of psychology. Indeed, restored to their actual context, Durkheim argued, imitative rites are fully explained by the fact that the clan members feel that they really are the animal or plant of their totemic species, that this is their most essential trait, and that this should be demonstrated whenever the clan gathers. In fact, it is only on the basis of such demonstrations that the Intichiuma could have the revivifying efficacy which Durkheim attributed to it.100 Far from being a magical, utilitarian rite to be explained by psychology, Durkheim concluded, the imitative rite was a moral. religious phenomenon to be explained sociologically.101

The principle "like produces like" thus has its origin in collective representations rather than the association of ideas in the individual mind, a fact of as much interest to the history of science as to the history of religion and magic; for Durkheim insisted that this principle was itself a primitive version of the more recent, scientific law of causality. Consider the two essential elements of that law: first, it presumes the idea of efficacy, of an active force capable of producing some effect; and second, it presupposes an a priori judgment that this cause produces its effect necessarily. The prototype of the first idea, as we have seen, is that collective force conceived by primitive peoples under the name of mana, wakan, orenda, etc., and then objectified and projected into things. The second is explained by the fact that the periodic reproduction of the totemic species is a matter of great concern to the clan, and the rites assumed to effect it are thus obligatory. What is obligatory in action, however, cannot remain optional in thought; thus society imposes a logical precept -- like produces like -- as an extension of the ritual precept essential to its well-being. Durkheim thus repeated the claim of his introduction -- that the sociological theory of the categories could reconcile the a priorist with the empiricist epistemology -- by showing how their necessity and universality could be both retained and explained: "The imperatives of thought," Durkheim concluded, "are probably only another side to the imperatives of action."102

Nonetheless, throughout Durkheim's discussion of sacrificial and imitative rites, there is a palpable air of insecurity; for the Intichiuma gave almost every indication of being precisely what Frazer, Spencer, and Gillen had said that it was -- a cooperative system of magic designed to increase the supply of the totemic animal.103 It was with an almost audible sigh of relief, therefore, that Durkheim turned to those "representative" or "commemorative" rites found in the Intichiuma of the Warramunga tribe (but which Durkheim insisted were "implicit" in the Arunta rite as well), for these rites contained no gesture whose object was the reproduction of the totemic species, consisting entirely of dramatic representations designed to recall the mythical history of the clan. But a mythology, Durkheim observed, is a moral system and a cosmology as well as a history; thus, "the rite serves and can serve only to sustain the vitality of these beliefs, to keep them from being effaced from memory and, in sum, to revivify the most essential elements of the collective consciousness."104 So the efficacy of the rites, again, was moral, religious, and social rather than economic, magical, and material.105 Indeed, Durkheim insisted that here we see the origin of the aesthetic element in the higher religions -- all that is artistic, expressive, and re-creative, rather than economic and utilitarian -- which, far from being a trivial accoutrement, plays an essential role in the religious life.106

Durkheim concluded his treatment of the positive cult with a discussion of "piacular" rites -- those which, in sharp contrast to the confident, joyous celebrations just described, are characterized by sadness, fear, and anger. Such rites ordinarily follow some disaster that has befallen the clan (the death of one of its members), and may involve the knocking out of teeth, severing of fingers, burning of skin, or any number of other self-inflicted tortures; but Durkheim insisted (pace Jevons again) that none of these acts were the spontaneous expression of individual emotion. On the contrary, such "mourning" appeared to be a duty imposed by the group and sanctioned by severe penalties, which the primitive says are required by the souls of ancestors; but in fact, Durkheim argued, the obligatory character of these rites is to be explained in the same way as their more joyous counterparts -- when someone dies, the clan assembles, thus giving rise to collective representations which reflect its sense of loss while simultaneously reaffirming the sense of its own permanence and solidarity.107

Finally, this account of piacular rituals also accounted for what Durkheim called the "ambiguity of the sacred." In his Lectures (1889), Robertson Smith had already suggested that primitive peoples distinguish not simply between sacred and profane, but also between sacred things which are good, pure, benevolent, and propitious, and those which are evil, impure and malicious. and unpropitious; and also that, while these two categories of things are separated like the sacred and the profane, there is also a certain "kinship" between them -- they both have the same relations of absolute heterogeneity with respect to profane things, and frequently, through a mere change of external circumstances, an evil, impure sacred thing may be transformed into its good, pure counterpart. But while Smith had an "active sentiment" of this ambiguity, Durkheim observed, he never explained it.108 How are evil forces also sacred? And how are they transformed into their counterparts? Durkheim's answers were that evil powers are the symbolic expression of those collective representations excited by periods of grief or mourning and consequent assemblies of the clan, and that they are transformed into their more benign opposites by that reaffirmation of the permanence and solidarity of the group effected by the ceremonies thus celebrated. These two extremes of the religious life thus reflect the two extremes through which all social life must pass. "So, at bottom," Durkheim concluded, "it is the unity and diversity of social life which make the simultaneous unity and diversity of sacred beings and things."109

The Social Origins of Religion and Science

Social scientists who have attempted to "explain" religion have typically regarded it as a system of ideas or beliefs, of which the rites are an external, material expression; and this has naturally led to a concern for whether these ideas and beliefs may or may not be reconciled with those of modern science. The difficulty for this approach, Durkheim argued, is that it does not correspond to the religious believer's own account of the nature of his experience, which is less one of thought than of action: "The believer who has communicated with his god," Durkheim observed, "is not merely a man who sees new truths of which the unbeliever is ignorant; he is a man who is stronger. He feels within him more force, either to endure the trials of existence, or to conquer them."110 The mere ideas of the individual are clearly insufficient to this purpose; and, in this sociological version of the principle extra ecclesia nulla salus, Durkheim thus insisted that it is the repeated acts of the cult which give rise to "impressions of joy, of interior peace, of serenity, of enthusiasm which are, for the believer, an experimental proof of his beliefs."111

Durkheim thus agreed with William James,112 who, in The Varieties of Religious Experience (1902), had argued that religious beliefs rest upon real experiences whose demonstrative value, though different, is in no way inferior to that of scientific experiments. As with such experiments, Durkheim added, it does not follow that the reality which gives rise to these experiences precisely corresponds to the ideas that believers (or scientists) form of it; but it is a reality just the same, and for Durkheim, the reality was society. This, indeed, explained why the cult rather than the idea is so important in religion -- "society cannot make its influence felt unless it is in action, and it is not in action unless individuals who compose it are assembled together and act in common."113 But Durkheim also felt that all societies need such periodic reaffirmations of their collective sentiments, and that there is thus something "eternal" in religion, destined to outlive the particular symbols -- totemic, Christian, or otherwise -- in which it had been previously embodied. The difficulty for a society living through the period of "transition" and "moral mediocrity" described in The Division of Labor and Suicide was in imagining what form its future symbols might assume.

But if religion is thus a mode of action, it is also a mode of thought -- one not different in kind from that exercised by science. Like science, for example, religion reflects on nature, man, and society, attempts to classify things, relates them to one another, and explains them; and as we have seen, even the most essential categories of scientific thought are religious in origin. Scientific thought, in short, is but a more perfect form of religious thought; and Durkheim thus felt that the latter would gradually give way before the inexorable advances of the former, including those advances in the social sciences extending to the scientific study of religion itself. In so far as it remains a mode of action, however, religion will endure, albeit under yet unforeseen forms.114

Science is thus religious in its origins; but if religion is itself only the apotheosis of society, then all logical, scientific thought originates in society. How is this possible? All logical thought, Durkheim explained, is made up of concepts -- generalized ideas which are distinguished from sensations by two important characteristics. First, they are relatively stable -- unlike our sensations, which succeed one another in a never-ending flux and cannot repeat themselves, our concepts remain the same for long periods of time. Second, they are impersonal -- again unlike our sensations, which are held privately and cannot be communicated, our concepts are not only communicable but provide the necessary means by which all communication becomes possible. These two characteristics in turn reveal the origin of conceptual thought. Since concepts are held in common and bear the mark of no individual mind, they are clearly the products of the collective mind; and if they have greater permanence and stability than our individual sensations, it is because they are collective representations, which respond much more slowly to environmental conditions. It is only through society, therefore, that men become capable of logical thought -- indeed, of stable, impersonal "truth" altogether; and this explains why the "correct" manipulation of such concepts carries a moral authority unknown to mere personal opinion and private experience.115

In one sense, the "categories of the understanding" are simply concepts so stable and impersonal that they have come to be seen as immutable and universal; but in another sense, the social explanation of the categories is more complex, for they not only have social causes but also express social things:

the category of class was at first indistinct from the concept of the human group; it is the rhythm of social life which is at the basis of the category of time; the territory occupied by the society furnished the material for the category of space; it is the collective force which was the prototype of the concept of efficient force, an essential element in the category of causality. 116

How is it that these categories, the pre-eminent concepts by which all of our knowledge is constructed, have been modeled upon and express social things? Durkheim's answer was that, precisely because the categories must perform this permanent, pre-eminent function, they must be based upon a reality of equally permanent, pre-eminent status -- a function for which our shifting, private sensations are clearly inadequate.

It might be argued, of course, that society is also inadequate for this function, that there can be no guarantee, for example, that categories modeled on social things will provide an accurate representation of nature; but this would be to deprive society of those attributes which Durkheim had laboriously attached to it throughout his productive and distinguished career. Society, for example, is itself a part of nature, and since "nature cannot contradict itself" we can expect that categories modeled on its realities will correspond to those of the physical world;117 and, in so far as the concepts founded in any particular group reflect the peculiarities of its special situation, we can expect that the increasing "internationalization" of social life will purge such concepts of their subjective, personal elements, so that we come closer and closer to truth, not in spite of the influence of society, but because of it.118

Like Kant, therefore, Durkheim denied any conflict between science, on the one hand, and morality and religion, on the other; for, also like Kant, he felt that both were directed toward universal principles, and that both thus implied that, in thought as in action, man can lift himself above the limitations of his private, individual nature to live a rational, impersonal life. What Kant could not explain (indeed, he refused to do so) is the cause of this dual existence that we are forced to lead, torn between the sensible and intelligible worlds which, even as they seem to contradict each other, seem to presume and even require each other as well. But to Durkheim the explanation was clear -- we lead an existence which is simultaneously both individual and social, and as individuals we can live without society no more than society can live without us.

Critical Remarks

Had Durkheim written nothing but The Elementary Forms, his place in the history of sociological thought would have been secure. It is a work of stunning theoretical imagination, whose two major themes and more than a dozen provocative hypotheses have stimulated the interest and excitement of several generations of sociologists irrespective of theoretical "school" or field of specialization. Nonetheless, it is not without flaws; indeed, it contains most of the indiscretions discussed earlier, and a few others besides. In sharp contrast to Max Weber, for example, Durkheim largely ignores the role of individual religious leaders, as well as the way religion functions in social conflict and asymmetrical relations of power. The "collective effervescence" stimulated by religious assemblies presumes a social psychology never made explicit, and Durkheim's account of how such gatherings generate totemic symbols is dubious to say the least. His definition of religion, preceded by an extended argument by elimination and containing a massive petitio principii, bears little relation to anything that the central Australians themselves understand by their beliefs and behavior; and students of aboriginal religion like W.E.H. Stanner have spent months looking for instances of the sacred-profane dichotomy, even to the point of questioning their own competence, before admitting that the Australian facts simply do not fit.119

Indeed, if there is a single feature of the work more disturbing than any other, it is Durkheim's treatment of the ethnographic evidence. The choice of the "single case" of central Australia has an intrinsic appeal to anyone familiar with the "scissors and paste" method of comparative religion epitomized in The Golden Bough; but in practice, this focus led Durkheim either to ignore counter-instances among the neighboring Australian tribes, or to interpret them arbitrarily according to some ad hoc, evolutionary speculations, or to "correct" them in light of the more advanced, and hence allegedly more edifying, American tribes. In fact, there is no evidence that Australian totemism is the earliest totemism, let alone the earliest religion; and, though technically less advanced than the North American Indians, the Australians have a kinship system which is far more complex. But then, pace Durkheim, there is no necessary relationship between the "simplicity" of a society (however that is defined) and that of their religious beliefs and practices; nor, for that matter, is there any necessary relationship between religion and totemism generally (wakan and mana have no discernible relationship to the "totemic principle").

Even if we limit ourselves to Australian tribes, we find that the central tribes are atypical (the Intichiuma, for example, scarcely exists elsewhere, and where it does, its meaning is altogether different); that the major cohesive force among aborigines is the tribe rather than the clan; that there are clans without totems (and totems without clans); that most totems are not represented by the carvings and inscriptions on which Durkheim placed so much weight; and that the "high gods" of Australia are not born of a synthesis of totems.120 Finally, we may choose to circumvent the details of Durkheim's interpretation of the ethnographic literature altogether, observing that its raison d'être -- the notion that the "essence" of religion itself may be found among the Arunta -- is, in Clifford Geertz's words (1973: 22), "palpable nonsense." What one finds among the Arunta are the beliefs and practices of the Arunta, and even to call these "religious" is to impose the conventions of one's own culture and historical period.

Criticisms such as these have led some scholars121 to suggest that the Australian data were introduced simply to illustrate Durkheim's theories, rather than the theories being constructed or adopted to account for the data. But this suggestion requires at least one major qualification in the light of what we know about the historical development of Durkheim's ideas on religion. The "largely formal and rather simpliste " conception of religion characteristic of Durkheim's early work,122 for example, persisted at least until the appearance of Baldwin Spencer and F.J. Gillen's Native Tribes of Central Australia (1899); and it was the largely psychologistic and utilitarian interpretation of these ethnographic data proposed by Frazer (and endorsed by Spencer himself) which led Durkheim back to the more determinedly sociological (even mystical) views of Robertson Smith, and to the contrived and even grotesque evolutionary interpretation of these data found in "Sur le totémisme" (1902b). If the theories of The Elementary Forms do not "explain the facts" of Australian ethnography, therefore, it is because so much of their original purpose was rather to explain them away.

The most ambitious claim of The Elementary Forms, of course, is that the most basic categories of human thought have their origin in social experience; but this claim, Steven Lukes has argued, is not one but six quite different claims which Durkheim did not consistently and clearly distinguish -- the heuristic claim that concepts (including the categories) are collective representations; the causal claim that society produces these concepts; the structuralist claim that these concepts are modeled upon, and are thus similar to, the structures of society; the functionalist claim that logical conformity is necessary to social stability; the cosmological claim that religious myth provided the earliest systems of classification; and the evolutionary claim that the most fundamental notions of modern science have primitive religious origins. The structuralist, cosmological, and evolutionary claims, Lukes observed, have been both challenging and influential. But the heuristic claim, by conflating the categories with concepts in general, confuses a capacity of mind with what is better described as its content. In so far as society is literally defined in terms of collective representations (as the later Durkheim increasingly did), both the causal and the functionalist claims seem simply to restate the heuristic claim, and are vulnerable to the same objection; but, in so far as society is construed in structural terms (as in The Division of Labor), the causal claim in particular is open to serious objections. The very relations proposed between the structures of primitive societies and their conceptual apparatus, for example, would seem to presuppose the primitives' possession of precisely those concepts; and the causal hypothesis itself cannot be framed in falsifiable form -- we cannot postulate a situation in which men do not think with such concepts, Lukes observes, because this is what thinking is.123 Finally, Durkheim's sociology of knowledge seems susceptible to at least as many empirical objections as his sociology of religion.124


  1. A religious system may be said to be "most primitive" when: (a) it is found in a society whose organization is surpassed by no others in simplicity; and (b) it is possible to explain this system without using any element borrowed from any previous religion (cf. 1912: 13).
  2. 1912: 13.
  3. 1912: 14-15.
  4. Cf. Durkheim's "Rules for the Demonstration of Sociological Proof" (1895: 145-158).
  5. 1912: 17.
  6. 1912: 18.
  7. 1912: 20.
  8. Although Durkheim's arguments here reveal his enormous debt to Kant and the French neo-Kantian Charles Renouvier (1815-1903), his more specific source was Renouvier's follower Octave Hamelin (1856-1907). Where Kant had described time and space as "forms of intuition" rather than among the twelve categories of the understanding, for example. Durkheim's decision to treat them as categories reflects the influence of Hamelin's Essai sur les éléments principaux de la représentation (1907) (cf. Durkheim, 1912: 21 n. 4).
  9. 1912: 22.
  10. 1912: 27.
  11. 1912: 28.
  12. 1912: 28.
  13. 1912: 30. Durkheim suggested that this explains the exceptional authority inherent in our reason -- it is simply the authority of society, transferred to that part of our thought which is the indispensable foundation of all common action.
  14. More strongly, Lukes (1972: 437) suggests that this implies a conventionalism and extreme relativism.
  15. 1912: 31.
  16. This second remark is particularly directed at James Frazer, whose essay, "The Origin of Totemism (1899) had argued that the beliefs and practices of the Australian aborigines, also the focus of The Elementary Forms, were largely magical rather than religious in nature.
  17. 1899: 93.
  18. Cf. Lukes, 1972: 241-243.
  19. Cf. 1912: 37 n. 1, 63 n. 68.
  20. In a recent essay (Jones, 1985), I have tried to explain the reasons tor this displeasure, placing particular emphasis on the ethnographic data presented in Spencer and Gillen's Native Tribes of Central Australia (1899), and especially its Frazerian interpretation (1899).
  21. 1895: 72-75.
  22. 1897: 42-46.
  23. 1912: 39.
  24. 1912: 38.
  25. 1912: 39. Among the works in which such a definition is entertained, Durkheim suggests, are Herbert Spencer's First Principles (1862) and F. Max Muller's Introduction to the Science of Religion (1873) and Lectures on the Origin and Growth of Religion (1878).
  26. This observation provides Durkheim with the opportunity for some caustic remarks on that intellectual field from which such"primitive" thinking has not yet disappeared -- the social sciences: "There are only a small number of minds," he suggests, "which are strongly penetrated with this idea that societies are subject to natural laws and form a kingdom of nature. It follows that veritable miracles are believed to be possible there" (1912: 41).
  27. 1912: 43.
  28. Cf. Albert Réville's Prolégomènes de l'histoire des religions (1881).
  29. Cf. E.B. Tylor's Primitive Culture (1871).
  30. The "ritual theory of myth" advanced by Robertson Smith (1889: 18-22) suggested that the earliest religions consisted primarily of actions rather than ideas, and that the latter were adapted as ex post facto rationalizations of the former.
  31. 1912: 50.
  32. 1912: 52.
  33. Durkheim considered the suggestion that sacred things are superior to profane things, either in dignity or in power, but concluded that: (a) there is nothing inherently "religious" in most distinctions between superior and inferior: (b) many sacred things have little dignity or power: and (c) even where men are dependent upon, and thus in some sense inferior to, their gods, the gods are also dependent upon their worshippers. The relationship is reciprocal (cf. 1912: 52-53).
  34. 1912: 53. Such emphasis is not without qualifications. Within the two classes, for example, there are "secondary species" equally opposed to one another (cf. 1912: 434 461 and below); and even between the two classes, the separation must be such as to permit some degree of communication of the sacred with the profane. Such communication requires elaborate ritual precautions, however, and even so, the two classes "cannot... approach each other and keep their own nature at the same time" (1912: 55).
  35. 1912: 56.
  36. Here Durkheim profited from Hubert and Mauss's "Esquisse d'un théorie generale de la magie" (1904).
  37. 1889: 90, 264-265.
  38. 1912: 59. This criterion itself faced difficulties in the so-called "private religions" which individuals establish and celebrate by themselves; but, while acknowledging the recent and increasing popularity of "these individual cults," Durkheim insisted that they "are not distinct and autonomous religious systems" but merely aspects of the common religion of the whole Church of which the individuals are members...." (1912: 61).
  39. 1912: 60. Here again, Durkheim faced difficulties in the so-called "societies of magicians"; but such societies, he responded, are not essential to the efficacy of the magical rite itself, nor do they include the "layman" for whom such rites are celebrated (1912: 60-61).
  40. 1912: 62.
  41. 1912: 64.
  42. Neither the problem nor Durkheim's account of it was quite so simple. First, Durkheim at least temporarily ignored all theories which appealed to "supernatural" data -- e.g., Andrew Lang's appeal, in The Making of Religion (1898), to a direct intuition of the divine, or the similar view of Wilhelm Schmidt developed in "L'Origin de l'idée de Dieu" (1908-10) -- though he took these up later (cf. 1912: 322-332). Second, Durkheim admitted the possibility of combining the two theories -- as Fustel de Coulanges had done in La Cité antique (1864) -- in which case we must decide what place to grant each. And third, Durkheim also acknowledged the efforts of some scholars to use certain propositions in these theories without adopting either in its systematic form (cf. 1912: 65).
  43. 1912: 68. Durkheim's summary of the animistic hypothesis constructed from Tylor's Primitive Culture (1871) and Spencer's Principles of Sociology, Parts I (1876) and Vl (1885). The two treatments are largely similar, although Spencer rejects Tylor's appeal to primitive analogical reasoning as an explanation of the cult of nature; the belief that non-human objects have souls, Spencer argued, was rather the result of metaphorical names (e.g., of a plant, animal, star, etc.) being given to children at birth, but later taken as literal in a "confusion of language," so that the non-human object was assumed to be the child's ancestor, and becomes the object of worship. Durkheim rejects Spencer's explanation, however, arguing that errors of language are insufficiently general to account for an institution of the universality of the cult of nature (cf. 1912: 70-71).
  44. Cf. 1912: 273-308.
  45. 1912: 87.
  46. Durkheim provides a brief survey (pp. 89-91) of the literature here but takes the work of F. Max Muller (1823-1900) as paradigmatic. Cf. Muller's Oxford Essays (1856), The Origin and Development of Religion (1878), Natural Religion (1889), Physical Religion (1890), Anthropological Religion (1892), Theosophy, or Psychological Religion (1893), and Contributions to the Science of Mythology (1897).
  47. 1912: 91-97.
  48. 1912: 103-5.
  49. 1912: 97-102.
  50. 1912: 106.
  51. 1912:107.
  52. Durkheim admitted that Totemism and Exogamy (1910) represented some progress away from such a view; for, by setting the domestic institutions of each tribe within their geographic and social environments, Frazer there departed from "the old methods of the anthropological school" (1912: 113n.32).
  53. Cf. 1895: 108-118.
  54. Durkheim did not deny himself occasional comparisons with the totemic institutions of the North American Indians, whose clarity, stability, and more advanced stage of development allowed him some rather speculative reconstructions of the historical evolution of totemism (cf. 1912: 115-117). Durkheim's method was thus quite similar to that of Robertson Smith, whose focus was almost exclusively on Semitic religions.
  55. 1912: 121.
  56. Australian clans are also joined together in "phratries" which have their own totems, and which Durkheim regarded as former clans (cf. 1912: 128-130).
  57. 1912: 156-163.
  58. Durkheim and his nephew, Marcel Mauss, had developed this argument in much greater detail in Primitive Classification (1903).
  59. Durkheim did not deny that the individual intellect can perceive similarities and differences among various objects, but he did deny that such vague, fluctuating images are themselves sufficient to the idea of class, which is the more concrete framework within which such images are distributed (cf. 1912: 170-172).
  60. To these beliefs, Durkheim added descriptions of "individual totemism" and "sexual totemism," which he regarded as secondary, derivative forms (cf. 1912: 183-193).
  61. Durkheim paved the way for his own explanation with a characteristic argument by elimination, in which competing accounts of the origin of totemism (including those of Tylor, G. A. Wilken, Wilhelm Wundt, F.B. Jevons, Frazer, Hill Tout, Alice Fletcher, Franz Boas, and Andrew Lang) were successively taken up, criticized, and discarded. The argument is fascinating in itself and provided a superb example of Durkheim's dialectical virtuosity; but to do justice to it here would require an extended and, in relation to the mainline of Durkheim's argument, unjustified digression (cf. 1912: 195-215).
  62. This is not intended metaphorically. Whether construed physically (it can mechanically produce physical effects) or morally (the individuals sharing it feel mutually bound to one another), this principle behaves like a real "force" (cf. 1912: 218-219).
  63. Durkheim insisted that an additional "presumption of objectivity" is provided by the multiple independent and almost simultaneous discovery of the same conception by a variety of scholars including R.R. Marett, Hubert and Mauss, Karl Preuss, and even Frazer himself (cf. 1912: 230- 232).
  64. 1912: 232-234.
  65. 1912: 236.
  66. 1912: 236-245.
  67. 1912: 245-251.
  68. 1912: 255-262.
  69. Cf., for example, Durkheim's defence of the "two principal ideas" of this work before the Société française de philosophie, transcribed and published as "Le Problème religieux et la dualité de la nature humaine" (1913).
  70. 1912: 264. The choice of animals and vegetables as the particular symbols of totemic clans are explained by their economic importance, their constant proximity and familiarity, and the ease with which they can be represented (cf. 1912: 266).
  71. Translated by L. A. Close as How Natives Think (1923).
  72. 1912: 269.
  73. In Les Fonctions mentales, Lévy-Bruhl had argued that the representations of primitives are emotional and mystical rather than intellectual, and that primitive thought is thus "prelogical," indifferent to modern conceptions of self-contradiction and causality. Durkheim's response, of course, insisted that the differences of primitive and modern thought are those of degree (e.g., more care in observation, more method in explanation) rather than kind, and that the former is the evolutionary origin of the latter (cf. 1912: 270-271).
  74. Cf. Kant's Critique of Practical Reason (1788) and Critique of Judgment (1790).
  75. Cf. Hume's "On the Immortality of the Soul" (1777).
  76. Cf. Tylor's Primitive Culture (1871).
  77. 1912: 303-304.
  78. Leibniz's theory of "monads," developed after 1695, can be found in Philip P. Wiener (ed.), Leibniz: Selections. New York: Charles Scribner's Sons, 1951, pp. 533-552.
  79. Again, cf. Kant's Critique of Practical Reason (1788).
  80. 1912: 305-308.
  81. Durkheim suggested that this is the explanation for what he, and Frazer before him, had called "individual totemism" (cf. 1912: 315-316).
  82. This is a clear example of Durkheim's application of the "ritual theory of myth," first proposed by Robertson Smith.
  83. For example, such gods are considered immortal, eternal, and providential; as the "creators" of all living beings; as having lived on earth themselves, before ascending into "heaven"; and as moral judges, even prophesying a "judgment day" on which good and evil men will receive their appropriate deserts (cf. 1912: 322-326).
  84. 1912: 332-333.
  85. Here, with significant reservations (cf. 1912: 338), Durkheim adopted the common idiom inaugurated by Frazer's Encyclopedia Britannica article "Taboo" (1887).
  86. Consistent with his earlier, rather ad hoc distinction, Durkheim's analysis of the negative cult explicitly avoided any discussion of those interdictions associated with magic. Two reasons were offered: first. while both magical and religious interdictions presuppose sanctions which follow violations of the interdict automatically, religious interdictions presuppose additional sanctions imposed by human agents; and second, while the objects to which magical interdictions are applied are kept separate as the consequence of fear, those to which religious interdictions are applied are kept separate out of respect for their "sacredness." To use the Kantian terminology, magical interdictions are hypothetical imperatives, while religious interdictions are categorical.
  87. Many of the second type, Durkheim observed, can be traced back to the first (cf. 1912: 340 n. 7).
  88. 1912: 355-356.
  89. 1912: 356-361.
  90. Durkheim briefly entertained the hypothesis proposed by F.B. Jevons in his Introduction to the History of Religions (1896), which emphasized the association of ideas. The difficulty for Jevons's argument, Durkheim observed, is not simply that civilized men, who continue to honor sacred objects, would be unlikely to be deceived by such associations; more important, primitive men themselves seem equally undeceived in all but their most religious preoccupations. The special status of the latter thus remains untouched by any purely psychological hypothesis (cf. 1912: 361-362).
  91. This, of course, is Durkheim's explanation for that "law of participation" observed by Lévy-Bruhl (cf. 1912: 365, and above).
  92. Durkheim himself said that his " révélation " came in 1895, upon his reading of the works of "Robertson Smith and his school"; but Smith's influence is evident in Durkheim's work only after 1900, when he attempted to countermand Frazer's interpretation of Spencer and Gillen's observations of the Arunta Intichiuma. This surely led him back to Smith, whose views he then grasped for the first time (cf. Durkheim, 1902b).
  93. 1912: 368-377.
  94. Cf. Smith's Encyclopedia Britannica article "Sacrifice" (1886), as well as the more detailed statement of the theory in his Lectures on the Religion of the Semites (1889).
  95. 1912: 380.
  96. Smith (1889: 392-396). The idea of private property, Smith observed (1889: 396) "materializes everything that it touches.... "
  97. 1912: 381-385.
  98. While the social environment thus "duplicates" the periodicity of its physical counterpart, Durkheim hardly granted them equal status in his explanation. The fluctuations of physical nature merely provide the "occasions" upon which men assemble or disperse, the "outer framework" within which the truly determinative action of society operates (cf. 1912: 391).
  99. 1912: 385-392.
  100. For this reason, Durkheim eventually concluded that imitative rites were the earliest form of the cult, from which all other forms were subsequently derived (cf. 1912: 432-433).
  101. 1912: 393-405. Durkheim's argument reversed the first two stages of the evolutionary sequence (i.e., magic-religion-science) proposed by Frazer in The Golden Bough (1890; 1900), relying heavily on Mauss's General Theory of Magic (1904).
  102. 1912: 412.
  103. Cf. Frazer's "The Origin of Totemism" (1899).
  104. 1912: 420.
  105. Durkheim acknowledged the fact that many "representative" rites are construed as having the same physical efficacy as their "imitative" counterparts; but he insisted that such interpretations are "accessory" and "contingent," not essential to the rite itself (cf. 1912: 423).
  106. Cf. 1912: 424-428.
  107. 1912: 442-449. Durkheim thus rejected Robertson Smith's view that the sense of sin and expiation was a late product of higher religions, arguing instead that it was implicit in the sentiments of fear and misery which gave rise to the earlier piacular rituals (cf. 1912: 453-455).
  108. This was not quite fair to Smith, who argued that the difference between gods and demons was that the first enjoyed stable, institutionalized relations with men, while the latter were approached only by individuals for private, asocial purposes; the first was religion, the second, magic (cf. Smith, 1889: 90, 264). Durkheim, for whom magic and sacredness were virtually antithetical categories, could hardly have been attracted by this explanation.
  109. 1912: 460.
  110. 1912: 464.
  111. 1912: 464.
  112. Though not with his epistemological views: cf. Durkheim's Pragmatism and Sociology (1913-1914).
  113. 1912: 465.
  114. 1912: 474-479.
  115. 1912: 479-487.
  116. 1912: 488.
  117. Lukes (1972: 440) notes that here Durkheim anticipates both Lévi-Strauss and the early Wittgenstein.
  118. 1912: 493.
  119. Stanner, 1967: 229.
  120. For more detailed accounts of these and other empirical shortcomings, see Lukes (1972: 477-480) and Pickering (1984).
  121. Cf., for example, Seger (1957: 69). This interpretation gathers force from the famous story about Durkheim told by Henri Bergson: "When we told him that the facts were in contradiction with his theories," Bergson observed, "he would reply: `The facts are wrong.'"
  122. Cf. Lukes, 1972: 240.
  123. Cf. Lukes (1972: 435-449). For a contrasting effort to resurrect Durkheim's argument, however, see Bloor (1984).
  124. On these in particular see Needham (1963: vii-xlviii).

To return to the Durkheim Home Page, click here.
For additional information, send a message to