Excerpt from Emile Durkheim's The Elementary Forms of Religious Life. Translated and with an Introduction by Karen E. Fields. New York: The Free Press (Simon & Schuster), 1995 [1912], pages 39-44.

The excerpt below is Section IV of Chapter One, "Definition of Religious Phenomena and of Religion." The chapter concludes with the following definition of religion:

We arrive thus at the following 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. The second element thus holds a place in my definition that is no less essential than the first: In showing that the idea of religion is inseparable from the idea of a Church, it conveys the notion that religion must be an eminently collective thing. (p. 44)
He establishes the first part of this definition in the first three sections. Section IV below establishes the second part of the definition concerning what Durkheim calls Church -- namely, the collective or communal aspect of religion.


Even so, this definition is not yet complete, for it fits equally well two orders of things that must be distinguished even though they are akin: magic and religion.

Magic, too, is made up of beliefs and rites. Like religion, it has its own myths and dogmas, but these are less well developed, probably because, given its pursuit of technical and utilitarian ends, magic does not waste time in pure speculation. Magic also has its ceremonies, sacrifices, purifications, prayers, songs, and dances. Those beings whom the magician invokes and the forces he puts to work are not only of the same nature as the forces addressed by religion but very often are the same forces. In the most primitive societies, the souls of the dead are in essence sacred things and objects of religious rites, but at the same time, they have played a major role in magic. In Australia (1) as well as in Melanesia, (2) in ancient Greece as well as among Christian peoples, (3) the souls, bones, and hair of the dead figure among the tools most often used by the magician. Demons are also a common instrument of magical influence. Now, demons are also surrounded by prohibitions; they too are separated and live in a world apart. Indeed, it is often difficult to distinguish them from gods proper. (4) Besides, even in Christianity, is not the devil a fallen god? And apart from his origins, does he not have a religious character, simply because the hell of which he is the keeper is an indispensable part in the machinery of the Christian religion? The magician can invoke regular and official deities. Sometimes these are gods of a foreign people: For example, the Greek magicians called upon Egyptian, Assyrian, or Jewish gods. Sometimes they are even national gods: Hecate and Diana were objects of a magic cult. The Virgin, the Christ, and the saints were used in the same manner by Christian magicians. (5)

Must we therefore say that magic cannot be rigorously differentiated from religion-that magic is full of religion and religion full of magic and, consequently, that it is impossible to separate them and define the one without the other? What makes that thesis hard to sustain is the marked repugnance of religion for magic and the hostility of magic to religion in return. Magic takes a kind of professional pleasure in profaning holy things, (6) inverting religious ceremonies in its rites. (7) On the other hand, while religion has not always condemned and prohibited magic rites, it has generally regarded them with disfavor. As messieurs Hubert and Mauss point out, there is something inherently antireligious about the maneuvers of the magician. (8) So it is difficult for these two institutions not to oppose one another at some point, whatever the relations between them. Since my intention is to limit my research to religion and stop where magic begins, discovering what distinguishes them is all the more important.

Here is how a line of demarcation can be drawn between these two domains.

Religious beliefs proper are always shared by a definite group that professes them and that practices the corresponding rites. Not only are they individually accepted by all members of that group, but they also belong to the group and unify it. The individuals who comprise the group feel joined to one another by the fact of common faith. A society whose members are united because they imagine the sacred world and its relations with the profane world in the same way, and because they translate this common representation into identical practices, is what is called a Church. In history we do not find religion without Church. Sometimes the Church is narrowly national; sometimes it extends beyond frontiers; sometimes it encompasses an entire people (Rome, Athens, the Hebrews); sometimes it encompasses only a fraction (Christian denominations since the coming of Protestantism); sometimes it is led by a body of priests; sometimes it is more or less without any official directing body. (9) But wherever we observe religious life, it has a definite group as its basis. Even so-called private cults, like the domestic cult or a corporate cult, satisfy this condition: They are always celebrated by a group, the family or the corporation. And, furthermore, even these private religions often are merely special forms of a broader religion that embraces the totality of life. (10) These small Churches are in reality only chapels in a larger Church and, because of this very scope, deserve all the more to be called by that name. (11)

Magic is an entirely different matter. Granted, magic beliefs are never without a certain currency. They are often widespread among broad strata of the population, and there are even peoples where they count no fewer active followers than religion proper. But they do not bind men who believe in them to one another and unite them into the same group, living the same life. There is no Church of magic. Between the magician and the individuals who consult him, there are no durable ties that make them members of a single moral body, comparable to the ties that join the faithful of the same god or the adherents of the same cult. The magician has a clientele, not a Church, and his clients may have no mutual relations, and may even be unknown to one another. Indeed, the relations they have with him are generally accidental and transient, analogous to those of a sick man with his doctor. The official and public character with which the magician is sometimes invested makes no difference. That he functions in broad daylight does not join him in a more regular and lasting manner with those who make use of his services.

It is true that, in certain cases, magicians form a society among themselves. They meet more or less periodically to celebrate certain rites in common in some instances; the place held by witches' meetings in European folklore is well known. But these associations are not at all indispensable for the functioning of magic. Indeed, they are rare and rather exceptional. To practice his art, the magician has no need whatever to congregate with his peers. He is more often a loner. In general, far from seeking company, he flees it. "He stands aloof, even from his colleagues." (12) By contrast, religion is inseparable from the idea of Church. In this first regard, there is already a fundamental difference between magic and religion. Furthermore, and above all, when magic societies of this sort are formed, they never encompass all the adherents of magic. Far from it. They encompass only the magicians. Excluded from them are the laity, as it were -- that is, those for whose benefit the rites are conducted, which is to say those who are the adherents of regular cults. Now, the magician is to magic what the priest is to religion. But a college of priests is no more a religion than a religious congregation that worships a certain saint in the shadows of the cloister is a private cult. A Church is not simply a priestly brotherhood; it is a moral community [note the first use in this book of this fundamentally important Durkheimian concept which can be thought of as "imagined community"; see pp. xxii-xxxii, xiv] made up of all the faithful, both laity and priests. Magic ordinarily has no community of this sort. (13)

But if one includes the notion of Church in the definition of religion, does one not by the same stroke exclude the individual religions that the individual institutes for himself and celebrates for himself alone? There is scarcely any society in which this is not to be found. As will be seen below, every Ojibway has his personal manitou that he chooses himself and to which he bears specific religious obligations; the Melanesian of the Banks Islands has his tamaniu; (14) the Roman has his genius; (15) the Christian has his patron saint and his guardian angel, and so forth. All these cults seem, by definition, to be independent of the group. And not only are these individual religions very common throughout history, but some people today pose the question whether such religions are not destined to become the dominant form of religious life -- whether a day will not come when the only cult will be the one that each person freely practices in his innermost self. (16)

But, let us put aside these speculations about the future for a moment. If we confine our discussion to religions as they are in the present and as they have been in the past, it becomes obvious that these individual cults are not distinct and autonomous religious systems but simply aspects of the religion common to the whole Church of which the individuals are part. The patron saint of the Christian is chosen from the official list of saints recognized by the Catholic Church, and there are canonical laws that prescribe how each believer must conduct this private cult. In the same way, the idea that every man necessarily has a protective genie is, in different forms, at the basis of a large number of American religions, as well as of Roman religion (to cite only these two examples). As will be seen below, that idea is tightly bound up with the idea of soul, and the idea of soul is not among those things that can be left entirely to individual choice. In a word, it is the Church of which he is a member that teaches the individual what these personal gods are, what their role is, how he must enter into relations with them, and how he must honor them. When one analyzes the doctrines of that Church systematically, sooner or later one comes across the doctrines that concern these special cults. Thus there are not two religions of different types, turned in opposite directions, but the same ideas and principles applied in both cases -- here, to circumstances that concern the group as a whole, and there, to the life of the individual. Indeed, this unity is so close that, among certain peoples, (17) the ceremonies during which the believer first enters into communication with his protective genie are combined with rites whose public character is incontestable, namely, rites of initiation. (18)

What remains are the present-day aspirations toward a religion that would consist entirely of interior and subjective states and be freely constructed by each one of us. But no matter how real those aspirations, they cannot affect our definition: This definition can be applied only to real, accomplished facts, not to uncertain possibilities. Religions can be defined as they are now or as they have been, not as they may be tending more or less vaguely to become. It is possible that this religious individualism is destined to become fact; but to be able to say in what measure, we must first know what religion is, of what elements it is made, from what causes it results, and what function it performs -- all questions whose answers cannot be preordained, for we have not crossed the threshold of research. Only at the end of this study will I try to look into the future.

We arrive thus at the following 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. The second element thus holds a place in my definition that is no less essential than the first: In showing that the idea of religion is inseparable from the idea of a Church, it conveys the notion that religion must be an eminently collective thing. (19)


1. See [Sir Baldwin] Spencer and [Francis James] Gillen, The Native Tribes of Central Australia [London, Macmillan, 1889], pp. 534ff., and Northern Tribes of Central Australia [London, Macmillan, 1904], p. 463; [Alfred William] Howitt, Native Tribes of South East Australia [London, Macmillan, 1904], pp. 359-361.

2. See [Robert Henry] Codrington, The Melanesians [Studies in Their Anthropology and Folklore, Oxford, Clarendon Press, 1891], chap. 12.

3. See Hubert, "Magic;" in Dictionnaire des antiquités.

4. For example, in Melanesia the tindalo is a spirit that is sometimes religious and sometimes magical (Codrington, The Melanesians, pp. 125ff., 194ff.).

5. See Hubert and Mauss, "Esquisse d'une théorie générale de la magie:" AS, vol. VII [1904]. pp. 83-84.

6. For example, the Host is profaned in the Black Mass.

7. See Hubert, "Magia" in Dictionnaire des antiquités.

8. Hubert and Mauss, "Esquisse," p. 19.

9. Certainly it is rare for each ceremony not to have its director at the moment it is conducted; even in the most crudely organized societies, there generally are men designated, due to the importance of their social role, to exercise a directive influence upon religious life (for example, the heads of local groups in certain Australian societies). But this attribution of functions is nevertheless very loose.

10. In Athens, the gods addressed by the domestic cult are only specialized forms of the gods of the City ( , ). [Zeus, protector of property, Zeus, the household god. Trans.] Similarly, in the Middle Ages, the patrons of brotherhoods are saints of the calendar.

11. For the name of Church ordinarily applies only to a group whose common beliefs refer to a sphere of less specialized things.

12. Hubert and Mauss, "Esquisse," p. 18.

13. [William] Robertson Smith had already shown that magic is opposed to religion as the individual is to the social ([Lectures on] the Religion of the Semites, 2d ed. [London, A. & C. Black, 1894], pp. 264-265). Further, in thus differentiating magic from religion, I do not mean to set up a radical discontinuity between them. The frontiers between these two domains are often blurred.

14. [Robert Henry] Codrington, "Notes on the Customs of Mota, Bank Islands in RSV; vol. XVI [1880], p. 136.

15. [Augusto] Negrioli. Dei Genii presso i Romani, [Bologna. Ditto Nicola Zanichelli, 1900].

16. This is the conclusion at which [Herbert] Spencer arrives in his Ecclesiastical Institutions [Part VI of The Principals of Sociology, New York, D. Appleton, 1886], chap. 16. It is also the conclusion of [Auguste] Sabatier, in his Esquisse d'une philosophie de la religion d'après la Psychologie et l'Histoire, [Paris, Fischbacher, 1897], and that of the entire school to which he belongs.

17. Among numerous Indian peoples of North America, in particular.

18. However, that factual point does not settle the question of whether external and public religion is anything other than the development of an interior and personal religion that would be the primitive phenomenon, or whether, on the other hand, the personal religion is the extension, inside individual consciousnesses, of the exterior one. The problem will be taken up directly below (Bk. II, chap. 5, §2. Cf. Bk. II, chap. 6 and Bk. II, chap. 7, §1). For now I merely note that the individual cult presents itself to the observer as an element and an appendage of the collective cult.

19. It is there that my definition picks up the one I proposed some time ago in the Année sociologique.In that work, I defined religious beliefs exclusively by their obligatory character; but that obligation evidently arises, as I showed, from the fact that those beliefs belong to a group that imposes them on its members. Thus the two definitions partly overlap. If I have thought it necessary to propose a new one, it is because the first was too formal and went too far in downplaying the content of religions representations. In the discussions that follow, we will see the point of having placed in evidence immediately what is characteristic of this content. In addition, if the imperative character is indeed a distinctive feature of religious beliefs, it has infinite gradations; consequently, it is not easily perceptible in some cases. There arise difficulties and troublesome questions that are avoided if this criterion is replaced by the one I have used above.