1 Explaining magico-religious beliefs was one of the major preoccupations of late twentieth-century anthropology because, through the study of primitives’ ways of thinking, it constituted an effort, within a dominant evolutionist framework, to identify the gaps marking the border between “us” and “them,” thus identifying the superior forms of civilization. Émile Durkheim, well abreast of the anthropological literature of his day, did not dodge this debate—to such an extent that his book The Elementary Forms of the Religious Life might seem like the vivid reflection of a bygone era. The mark of a great book, however, is that it surpasses the context that brought it into being. This assessment applies to The Elementary Forms, which, hailed as an event upon its publication while immediately drawing caustic criticism, has ended up having a growing influence on anthropological thought.
2 This article briefly skims through the stages of this career and attempts to define anthropology’s debt to The Elementary Forms, anthropology being here defined as an academic discipline whose research deals mainly with archaic societies. This discipline, long suspected by Durkheim of amateurism in its data collection practices and epistemologically reduced to a mere auxiliary of sociology (Karady 1988; Affergan 2008), gradually gained the French sociologist’s respect, to the point that it became the center of his studies. This was notably due to the growing list of reports produced by ethnographic research on Australian populations. From the works of Lorimer Fison and William Howitt (1880) to those of Baldwin Spencer and Francis Gillen (1899; 1904), not to mention James Frazer’s early summaries (1887; 1910; 1911-1915 ), this vast anthropological worksite allowed Durkheim to discover the consistency of the first systematically organized religions. Based on this discovery, the project of a sociological theory of belief became possible: the same project, undertaken upon the founding of L'Année sociologique,  was completed with The Elementary Forms, following its painstaking composition.
The Book’s Reception
3 In 1913, at the start of his career, Bronislaw Malinowski published an article in which two interwoven sentiments prevailed: first, admiration, as the publication of Émile Durkheim’s work was being greeted as a “scientific event” (1963 , 283); and second, an assured skepticism toward its thesis. He expressed his doubts on two points that he deemed essential. First, he highlighted the gap between the ambitions of Durkheim’s project--determining the origin of religious feeling--and the narrowness of his empirical base, which rested primarily on a few Australian tribes. Nor did he believe it methodologically possible to define religion based on the separation of the sacred and the profane, for there is no indication that such a separation, though perhaps at work in Australian societies, is universal. His criticisms were concentrated, however, on the heart of the Durkheimian system, namely, the thesis according to which the object of religious beliefs was the hypostatized society, called the “metaphysical concept.” The identification of religion with society on the one hand, and the reduction of religion to a “crowd phenomenon” on the other, are, for Malinowski, reminiscent of the Hegelian method in their obscurity (285-287). He was never to stray from that mistrust of Durkheimian constructions, which he suspected of underestimating the role of psychology in the production of beliefs. Admittedly, after giving a favorable assessment of The Elementary Forms in his 1913 work on the Australian family (Malinowski 1913), in a 1925 article, “Magic, Science, and Religion,” he appeared for a time to endorse the distinction between the sacred and the profane (1992 , 20), only to put it into perspective by granting a leading role to psychological factors. Indeed, Malinowski’s definitions of magic (a technique with utilitarian aims, a formula for reducing uncertainty) and of religion (a belief with no immediate purpose, preoccupied with the question of meaning) owe nothing to Durkheim’s work, to such a degree that, to the anthropologist, totemism seems more like a magical belief than a religious one (24).
4 In 1913, Arnold Van Gennep also took a stance on Durkheim’s book, with criticisms similar to Malinowski’s. His 1909 book, The Rites of Passage, showed a certain distance from the works of the French sociological school, as Durkheim was not even mentioned. A line was crossed in 1913, as Van Gennep’s reading seems to lay waste to The Elementary Forms. In his view, the ethnographical data mentioned are not just fragile (since the sources are doubtful), but are poorly understood by Durkheim to boot, as the latter failed to realize the degree to which Australian societies are highly complex groups (Van Gennep, 1994 , 206-207). The proof of Durkheim’s poor methodological habits is that his definition of totemism as the religion dedicated to an anonymous force embodied in plants or animals is not only the definition of fetishism, but also contributes to eliminating the central question that should deal with the diversity of forms of totemic belief. According to this point of view, Durkheim, the victim of his own naturalistic illusion, sought in Australia an equivalent of unicellular structure in human societies, thus tacking concepts and language onto a reality and granting them a realistic signification (208). In the end, Van Gennep found that the French sociologist’s perspective was mistaken. Admittedly, religion encompasses numerous aspects of social life in primitive societies; however, the double process of secularization and individualization that characterizes our modern societies ran counter to religion and, consequently, cannot result from this belief. The publication of The Current State of the Totemic Problem would in no way modify Van Gennep’s point of view on The Elementary Forms, a book that, in his words, offered “no real solution;” was based on reviving the theses of Andrew Lang,  an important interlocutor of Durkheim; and had no validity beyond the perimeter of Central Australia (Van Gennep, 1920, 41, 43-44, 46).
5 Alexander Goldenweiser, author of an early study of totemism in 1910 (1933a ), published an analysis of The Elementary Forms in 1915 and extended it later (1916, 1933b ). He deemed that the scope of Durkheim’s contribution “commands attention” (1994 , 209). However, though there is clear evidence he understood that Durkheim’s intention was not so much to explain a belief, but rather to make a major contribution to five interrelated questions (theories of religion, totemism, social control, ritual, and thought (Ibid., ), Goldenweiser himself appears particularly circumspect regarding these five aspects. In particular, he did not believe in the relevance of a definition of religion based on the sacred/profane dichotomy. In his view, Durkheim did not successfully grasp the primitive mentality that draws a distinction, rejected by the French sociologist, between the natural and the “strange” or “extraordinary,” a distinction which implies recourse to an explanation in terms of forces or mysteries (211). Goldenweiser also keeps his distance from the Durkheimian method that overevaluates social factors in the explanation of beliefs, factors which bring us back to the thesis according to which religion is born from states of “collective effervescence,” the empirical confirmation of which lies solely in the Intichiuma ceremony of the Arunta tribe (218). Finally, Goldenweiser finds the theory of thought supported by Durkheim too narrow and exclusive, inasmuch as it underestimates the contribution of “profane experience” in the formation of the categories of human thought (226).
6 In Early Civilization in 1922, Goldenweiser systematized the critical view that a generation of anthropologists took of the Durkheimian conception of totemism. Indeed, Durkheim believed it possible to establish a link between the opposition of the sacred and profane and the clan organization of Australian society, marked by alternating phases in which social life slows or intensifies depending on the density of social interactions. However, according to Goldenweiser (1922, 361-388), the root of the demonstration--clan organization and totemic belief--is particularly fragile. On the one hand, Durkheim loses his way by maintaining that all Australians are organized into clans, a false assertion from an ethnographic point of view, and one to which he adheres by following, for once, James Frazer’s developments as put forth in Totemism (1887). On the other hand, he thinks that where there are clans, there must be observable totemic beliefs, and that where there are totemic beliefs, one will observe clans. In particular, this relationship between social organization and religious beliefs does not hold up to the study of facts as conclusive as the existence of hunting peoples whose organization has the same simplicity as that of the Australians and who have no religious beliefs in totemic form.
7 Alfred Radcliffe-Brown, who had already participated in the first debate on totemism (Radcliffe-Brown 1914) and who had argued with English-language anthropologists in favor of Durkheim (Kuper 1983), authored an important 1929 article that, once again, revealed an ambivalent attitude toward The Elementary Forms. From an empirical point of view, he remarked, Durkheim’s thesis suffered from ignorance of the variety of religious forms inventoried in Australia and from an overestimation of the importance of totemic belief in social life (Radcliffe-Brown 1972 , 194-197). But “totemism,” a nominal label that does not designate any reality as such and whose origins are unknown, poses a theoretical problem to which anthropology must offer a response: one concerning the ritual, not sacred, relationship that links individuals to natural objects. However, according to Radcliffe-Brown, this response is found in part in the 1912 work in that Durkheim, for the first time, laid the groundwork for a sociological explanation of this belief (195):but only in part, to the extent that the relation to the totem is but one form among others of the ritual relationship that unites mankind with natural species. From the particular case of totemism, Radcliffe-Brown arrives at the following formulation which, according to him, reveals the general “law” of ritual relationship: “Any object or event which has important effects upon the well-being (material or spiritual) of a society [. . .] tends to become an object of the ritual attitude” (202) Radcliffe-Brown’s article revisits, and applies to totemism, a conception of anthropological work that he had expressed in his study of the Andaman Islanders (1922), in which he already indicated his indifference toward overly narrow definitions of religion and of magic and his concern for relating the analysis of belief to the analysis of rituals. 
8 By virtue of their repetition, early criticisms of The Elementary Forms ended up constituting a doxa within the anthropological community. Although Durkheim’s contribution drew respect and attention, it did not get much support. Having fallen, like an entire generation, into the trap of totemism and the fashion of mana (Rosa 2003), Durkheim appeared to have produced a book in which metaphysical hypotheses flourished in the absence of empirical proof, and he was stuck in semantic quarrels on account of his taste for the preliminary definitions of social phenomena.
9 Half a century after the publication of The Elementary Forms, Claude Lévi-Strauss combined all of these arguments in one striking synthesis. Following Radcliffe-Brown, he indicated that totemism is a specific borderline case in which two orders coincide: first, the identification of humans with natural species, and second, the denomination of kinship relations using plant or animal names (1991 , 19). Yet it is on the basis of this specific case that Durkheim tried to establish a general theory of religious belief, while logically he should have reconstructed the possible world in four dimensions of relationships combining two series, one natural and the other socio-cultural, according to two modes, individual and collective (27). Consequently, Lévi-Strauss’s critique is based less on Durkheim’s misinterpretations than on this theoretical error that led him to elevate a specific case to the level of the universal. Furthermore, that error was redoubled by the delusion that led Durkheim to make his entire construction of the phenomenon contingent upon collective effervescence when, according to Lévi-Strauss, that effervescence is only rendered possible by religion itself (107). Never having truly strayed from the “genetic” method that makes “the objectivity of the symbol” an “inherent property of social facts” (Lévi-Strauss 1945, 518), Durkheim consequently could not understand the thought process of the “savages,” which activates categories such as mana, not because they believe in the reality of an impersonal, anonymous force present in sacred objects, but because they are obliged to resort to a “value of indeterminate meaning” (1989 , XLIV), the function of which is to fill in the gap between a signifier and a signified--a function that provides the “savages” with the possibility of representing to themselves the universe around them.
10 Edward Evans-Pritchard reaffirmed the points that make up the doxa of anthropologists with regard to The Elementary Forms in an initial contribution in the 1950s (Evans-Pritchard 1956) and, most notably, later, in his review of primitive religions (2001). He made a long presentation of Durkheim’s thesis, describing it as “more than just neat; it is brilliant and imaginative, almost poetical” (110). Unlike Lévi-Strauss, whose criticism was based on questions of method--genetic vs. structuralist--Evans-Pritchard intends to oppose Durkheim on the level of ethnographic facts. He d emonstrates, with a concrete example, that the separation between the sacred and the profane is ineffective because these two domains are irremediably entwined. For the primitives, a disease is at once a physical symptom, punishment for a mistake, and the mark of supernatural intervention, a unity that leads Evans-Pritchard to say that he has never seen how that distinction could be “useful” in the analysis of a society.  He likewise evoked the case of the Azande, a people whom he had studied thirty years previously (I will return to the topic of this research). The Azande do not impose any restrictions upon their practice of religious rituals, even using a tomb as a sacred or utilitarian space (112). Finally, the graphic instinct that was supposed to have led the primitives to visualize their totem in order to make a concrete and vivid image of the unity of their group was not validated in the field, as many totems were not represented in figurative form (115).
What Anthropology Owes to The Elementary Forms of the Religious Life
11 Lévi-Strauss wrote: “All that makes the greatness and the weakness of Durkheim’s work is found in Les Formes élémentaires de la vie religieuse” (1945, 515). The book left its mark on anthropology as soon as “the better Durkheim,” again in Lévi-Strauss’s terms (1991 , 142), managed to surface from beneath the critiques. His influence can be summarized as follows. From a negative point of view, The Elementary Forms offers a scathing repudiation of the two childhood illnesses of anthropology: evolutionism and the thesis according to which there is such a thing as a “primitive mentality.” From a positive point of view, Durkheim’s book created theoretical functionalism and laid the groundwork for the cognitive approach in anthropology. More essentially, it introduced in original terms the thematics of collective beliefs, which eventually subsumed the question of the relationship between magic and religion.
12 In 1912, the image of the primitive, as it appeared in the writings of Lewis Morgan, John Lubbock, Herbert Spencer, Edward Tylor, and James Frazer, in religious history textbooks, was monochromatic (Sanchez 2007). In this literature, primitive man represented the subhuman pole of our common human condition, as opposed to Western man. Spurred by his passions, ignorant of the principle of causality, and confusing the symbolic efficacy of his rituals with the instrumental efficacy of technologies, the primitive seemed to be plunged into darkness, making do with the prescriptions of archaic tradition to guide his behavior. He revealed a passivity in stark contrast to the self-mastery of the civilized man.
13 The Elementary Forms is a definitive refutation of this representation. As Durkheim very well demonstrates in his critique of the Tylorian theory of animism, this idea of the primitive is inconsistent, as in the end it discredits itself:
Finally, the animistic theory implies a consequence that is perhaps its best refutation. If it were true, it would be necessary to admit that religious beliefs are so many hallucinatory representations, without any objective foundation whatsoever. [. . .] One might even wonder if under these conditions the phrase science of religions can be employed without impropriety. A science is a discipline which, in whatever manner it is conceived, is always applied to some real data.
15 Moreover, the methodical refutation of that image of primitive man allowed Durkheim to distance himself from the theses of Lucien Lévy-Bruhl (1910) and to assert his methodological relevance (Boudon 2000). Lévy-Bruhl’s thesis, by which the primitive and the Western mentalities were considered immeasurable, was admittedly aimed at reevaluating the status of the primitive, whose right to be different was recognized, but this concession was made at the price of a representation of his cognitive functions in which an insensitivity to the principle of contradiction and the resonance of mystical bonds between living and nonliving things held sway. At odds with this concept, Durkheim supports, in his conclusion, the hypothesis of a continuity of forms of human thought, which is updated according to the state of knowledge and which must, whether it is religious or scientific in nature, expose its hypotheses to reality and confirm their validity in order to be legitimately endorsed by the members of a community. As Durkheim himself put it:
A collective representation is necessarily submitted to a control that is repeated indefinitely; the people who accept it verify it by their own experience.
17 This methodological position, whose radicality must be measured if it is related to the context of its formulation, contained the seeds of all subsequent developments in field anthropology, starting with those of Malinowski and Evans-Pritchard.
18 Thus, from the very moment when Lévy-Bruhl (1922) was developing the notion of the primitive mentality, Malinowski (1989 ) was completing his report on his expedition to the Trobriand Islands, which included his contextualization of beliefs. Far from discovering “primitives” on these islands, Malinowski met people who differed from Europeans only in their degree of technological development. Here as elsewhere, society fulfilled the same functions of social integration, transmission of values, and reproduction of order. Here as elsewhere, people knew perfectly well how to tell religious and technological rituals apart. Only the division of religious life undertaken by Lévy-Bruhl in How Natives Think (1910), isolating the beliefs of institutions and social representations, resulted in the assertion of a world governed by the “law” of participation. In this sense, Durkheim had mapped out the path: he who had, in contradiction to the majority of writings of his day, refused to collate, like an entomologist, the realities of the entire world and centered his subject on Australian societies alone, in order to be able to establish the connection between beliefs and their social foundations.
19 But it is to Evans-Pritchard (1972 ), who relied on a meticulous analysis of the ritual practices of the Azande and of their inference procedures regarding the validation of beliefs, that we owe the most complete refutation of Lévy-Bruhl’s prelogism and the most successful empirical illustration of Durkheim’s thesis. No mystical category and no predisposition to emotion are perceptible in Evans-Pritchard’s observations of the Azande. On the contrary, this tribe had developed a whole series of proofs and objections and interpreted certain events, sometimes with great subtlety, in order to grant themselves a supernatural quality (or not). The cognitive resources upon which they drew demonstrated that they practiced the principle of noncontradiction. Let us take the example of the ritual of the poison test, which might seem particularly symptomatic of a primitive mode of thought. Evans-Pritchard showed that it cannot be interpreted independently of the concepts that surround the substance. Indeed, poison is not considered a natural material that would have predictable effects on animals or on humans. It is an element that delivers its message only within the context of control procedures that incorporate performances and have to do with the rhythm and order of succession of the phases of the ritual, the distribution of roles among the actors, and the conditions under which the poison itself is ingested.
20 Here, Evans-Pritchard is in accordance with one of the methodological inspirations behind The Elementary Forms: the idea of a primitive mentality arose at a time when the anthropologist, through lack of contextualization, proved incapable of reproducing indigenous categories comparable to our own. By penetrating the Azande system of thought, Evans-Pritchard was thus able to identify twenty-two reasons that led this people to lend credence to their beliefs, applying sophisticated questioning and verification methods (538-539). However, he added, this exercise of reason must be understood as a function of the major orientation of this culture. The Azande actually manifested a limited interest in generalizations. For the Azande, the issue is thinking about the peculiar, the never-before-seen, and not the everyday reality. The belief in sorcery and magic opens for them a space of interpretation that is flexible enough to give meaning to events, with no necessity for them to resort to the development of any other explanatory system.
21 The path mapped out by Durkheim in The Elementary Forms enriched numerous subsequent works in cognitive anthropology that endeavored to show that the noted gaps between primitives and civilized men, concerning reasoning and the exercise of logical thinking, are the consequence of a contingent situation linked to the absence of writing or a low level of education. In particular, A. R. Luria (1976 ), Michael Cole and Sylvia Scribner (1974), Jack Goody (1993 ), Edwin Hutchins (1980), and Eleanor Rosch (1980) established through experiments that the propensity of individuals to solve simple inference problems depends less on their culture of origin than on their degree of literacy, and that writing is a technique that goes hand in hand with an increase in capacity for abstract thinking. For example, Goody demonstrated that the arithmetic skills of unschooled children in Ghana are often substantial, but that they run into significant difficulty when moving on to multiplication and division, abstract operations in the context of their number system.
22 The Elementary Forms was not just a formidable discursive machine fighting against evolutionist anthropology. Durkheim also caused anthropology of the time to make a major epistemological leap by asserting the primacy of theory over data, the accumulation of which is no more capable of laying the basis for a demonstration than a heap of haphazardly thrown stones would constitute a wall, to adapt Poincaré’s suggestive expression. What image of Durkheim’s work was reflected in contemporary anthropological productions? For the most part, a few postulates presented as a point of departure and point of arrival for discussion and, between the two, an erudition contest, a bibliographical obstacle course, and a hike through cultures bearing the seal of exoticism. For Durkheim, on the other hand, only a sociological theory completed by anthropological fact would provide the conditions making a reasoned comparative approach possible, even though it was developed from a limited geographical context. It is by constructing a theory of totemism that anthropology would be likely to teach us something about religious beliefs. The epistemological vector departs from theory to go in the direction of facts, and not the reverse; or, more precisely, it can go in the direction of facts only as and when the theory becomes denser and incorporates more and more data. On the other hand, the method based on the principle of collage, far from proving anything at all, reestablishes a rhapsodic image of the object of analysis, to the point that the whole eventually loses all intelligibility (Durkheim 1912, 132-133).
23 This primacy of theoretical work is, as always with Durkheim, in the service of a guiding line: beliefs are first and foremost social phenomena, that is to say, collective representations that enjoy autonomy with regard to any other authority, whether psychological or physiological. Durkheim applied a point of view that he had already systematized in his article “Représentations individuelles et représentations collectives” (Durkheim 1898b) to the field of beliefs. Presenting collective beliefs as social phenomena may seem obvious to us today as the immediate given in any anthropological investigation. In fact, though, Durkheim’s work was necessary in order for this obvious fact to take shape. Thus, without Durkheim’s preparatory work, an anthropologist like Malinowski surely could not have enjoyed such a heuristic main thread when he threw himself into his plan “to give a clear and coherent outline of the social structure and to reveal, from the jumble of facts, the laws and norms of all cultural phenomena” (Malinowski 1989 , 67).
24 This main thread leads to observance of the preeminence of the functional over the genetic in the explanatory framework provided by Durkheim in The Elementary Forms. Durkheim’s argument is as follows: Australian totemism can admittedly be embodied in a multitude of plants or animals, but this diversity is misleading because, through plants or animals, one addresses an anonymous and impersonal force that is found in each and every being, and in which all living things participate. The totem, hypostasis of the clan, is a cult object because society, unable to be directly cultic itself, imagines mediations and produces perceptible images through which beliefs can more easily pass. Thus, by a kind of trick of reason, what men are celebrating through their gods is society itself, which possesses all of the attributes necessary to awaken the powerful feeling of the sacred in consciousness. Through rituals and ceremonies, the reproduction of the social order is established. This thesis is among the first substantial formulations of functionalism in anthropology. It manifestly contains all of the elements that characterize a functionalist explanation, inasmuch as one finds in it most of the postulates highlighted by Robert Merton (1953 ), who also relied on The Elementary Forms.
25 The functionalist program was to become the dominant language of anthropology over the course of the period from 1920 to 1960, during which time the figure of Durkheim appeared as its founder.  In a more general way, this influence is easily perceptible among anthropologists for whom the big issue remained the clarification of the link between beliefs and social organization. It is visible, even if one were to consider only the researchers of the new generation, in the works of Monica Wilson (1951) on the Nyakyusa and the Pondo, M. G. Marwick (1952) on the Cewa, J. Clyde Mitchell (1956) on the Yao, Meyer Fortes (1974 ) on the Tallensi, and Godfrey Lienhardt (1961) on the Dinka, as well as in the studies gathered by John Middleton and Edward H. Winter (1963) dealing with the explanation of beliefs, and in Victor Turner’s (1972 ) reflections on the function of rituals.
26 It is undoubtedly inappropriate to exaggerate the preeminence of The Elementary Forms in the whole of functionalist literature; Malinowski’s functionalism (1992 ) is undoubtedly different from Durkheim’s, in that he claims that every social institution is dependent upon the satisfaction of a psychological or physiological need; and Durkheim was undoubtedly eclipsed in France during the structuralist movement, Lévi-Strauss (1989 ) having initially declared his preference for Durkheim’s nephew Marcel Mauss and his book The Gift.  Neither is it excessive, for all that, to say that to varying degrees, the anthropology of beliefs is still adopting the fundamental questions of social functionalism, whether in its weak version (understanding how a collective representation is linked to a given social organization) or its strong version (understanding how a social organization determines such and such a collective representation), two questions that were clearly raised in The Elementary Forms.
27 As we have noted, in this book Durkheim was still dependent upon nineteenth-century debates regarding the nature of religion and magic, but the intelligibility of his thesis is independent of the definition that he gave to these two phenomena, particularly the former. The imaginary exercise that consists in striking this very controversial (Warner 1937) definition of religion from the book would in no way modify its content. Indeed, from the introduction to the conclusion, Durkheim asks himself how to explain collective beliefs, totemism being only one of the forms that proves suitable for study, due to its simplicity. However, whether we are dealing with the religious or the magical, the aim is to gain similar understandings of these two phenomena by tying them to their social basis and indicating the arguments deployed by the individuals who subscribe to them. The unity of this category of collective beliefs is the result of the method that demands that a “utility” be accorded to every belief, even if it seems marred by a strong element of irrationality. Indeed, with the spread of the functionalist model of explanation, the theme of diversity of religious beliefs and of their definitions lost interest to the point of gradually disappearing from the literature (Goode 1949; Spiro 1972 ); Geertz 1973). Thus the great service that The Elementary Forms did for anthropology was perhaps in promoting the emergence of a field of study--collective beliefs--that allowed for the abandonment of false issues and the formulation of new questions.
See in particular Durkheim 1898 a; 1899; and 1902; and Durkheim and Mauss 1903.
Durkheim and Lang clashed over a controversy beginning in 1904. Lang devoted a chapter of his book, The Secret of the Totem (Lang 1905, 91-110), to Durkheim’s theory of totemism (1902). Durkheim rejected the “nominalistic” explanation of totemism proposed by Lang in The Elementary Forms (1912, 262-266).
For Radcliffe-Brown’s attitude toward The Elementary Forms as seen in two letters to Marcel Mauss written in 1912 and 1930, see Fournier (2007, 804).
See the same assessment in Evans-Pritchard 1981, 160.
Evans-Pritchard 1969 , 51-54.
However, see Lévi-Strauss 1973 (1960), 61-62.