Encyclopedia of Religion
and Society

William H. Swatos, Jr. Editor

Table of Contents | Cover Page  |  Editors  |  Contributors  |  Introduction  |  Web Version

Defining religion is problematic, with some definitions threatening to drive theories and determine conclusions. Some writers recommend proceeding with study and fashioning definitions only afterward (Harrison 1912, Weber 1922).

With new religions, revitalization movements, and quasi-religious pursuits coming to the fore in modern societies, the definitional issue takes on renewed importance (Hervieu-Léger 1987; see Greil and Rudy 1990 on "quasi-religion"), and the more aspects of the social world that are deemed religious, the less one can see any secularization process taking place.

One does not know if certain conduct is religious unless one begins with a concept of religion, and one cannot conceptualize it unless one already is familiar with cases. Moreover, participants in religion, and their critics, already have working definitions of religion. As the Verstehen school social scientists point out, a depiction that does not reflect such definitions would distort rather than report (Horton 1960).

Whether and when to define religion also raises "how" questions. Should one set up sharp lines around religion, to ascertain certainly whether a phenomenon is religious? Or should it have ambiguous boundaries, allowing for the unfamiliar and surprising? Should a definition be broad, including all that might be religious, or narrow so that religion serves as a variable? Should one have a preliminary definition that is fuzzyedged and broad, and a later one that is sharply cut and narrow? Or does scientific progress begin with narrowly defined, clear cases and proceed to broad, diffuse categories? Should a definition be based primarily on the conceptions held by religious participants, or should it be based on concepts in detached minds?

The definitional literature can be organized around four approaches: substantive, functional, verstehende , and formal.

Substantive Definitions

Early in the history of social science, Tylor (1871) wanted a minimum definition that would prevent categorizing primitive religions with spiritualism, which was in disrepute in Europe. The primitives, he proposed, were explaining the difference between life and death, and they conceived of life as animation by spirits. Tylor used religion , which had favorable connotations, to refer to a "belief in spiritual beings" with this in mind; such beliefs were primitives' equivalents of today's life sciences. Tylor's definition is often cited as the first substantive one.

Substantive definitions are often used in otherwise functionalist analyses. Ross (1901:197) saw religion as something that would exert a certain social control, but he defined it as belief about the Unseen, with such attendant feelings as fear, wonder, reverence, gratitude, and love, and such institutions as prayer, worship, and sacrifice. Similarly, Parsons's early work (1937:665 ff.) developed functionalism while using a substantive concept of religion. He gave Weber's writings about charisma a functional reading: Charisma would be that which functioned to legitimate power by associating governance with teleological meanings. Religious beliefs, defined by their reference to the supernatural, would characterize a particular kind of charisma. Later commentators, not interpreting Weber as a functionalist, would still see him implicitly taking religion to be a patterning of social relationships around beliefs in supernatural powers, creating ethical consequences (Swatos and Gustafson 1992).

Sorokin (1947:225) straddles the formal and substantive approaches. Taken literally, his definition is formal: a set of ultimate values expressed in a credo, objectified by vehicles of a cult, and socialized by conduct complying with religious norms that unite members into one religious group. However, his examples suggest that he had the supernatural in mind as the religious content. Nevertheless, he had a functionalist concern; his reference to uniting members into one religious group echoes the functional portion of Durkheim's (1912) definition.

Substantive definitions often appear in critiques of functionalism. Horton (1960:211) argued that people reject as not genuine religiouslike conduct done for (functional) reasons of social symbolism. He modeled his own definition after Tylor's: an extension of the field of people's social relationships beyond the confines of purely human society.

Glock and Stark (1965:4) also used a substantive definition in preparing for their influential questionnaire studies. Their key term was ultimate meaning , and they felt obliged to explain that belief systems of ultimate meaning that had supernatural referents were religious.

Berger (1967:175-177) at first merely noted his own preference for a substantive definition. He cited the concept of the holy as described by Otto (1923), but saw it all as a matter of personal preference having little scientific import. Revealingly, Otto's phenomenology of the holy rests upon the experiences of social actors, and it was this fact (rather than culturally distant allusions by anthropologists to beliefs in supernatural beings) that appealed to Berger. Later he spelled out his methodological preference for the Verstehen approach in social science that grounded scientists' definitions in those of social actors, but he still called his definitional preference "substantive" (Berger 1974; see Weigert 1974). Meanwhile, Garrett (1974) also revived interest in the substantive approach, similarly referring to Otto and expressing concern about giving adequate accounts of religious participants' experiences. He found concepts analogous to Otto's in the works of Simmel, Weber, and Troeltsch, and citations of Otto in those of Wach, Scheler, and Schutz. Significantly, it is the effects of the religious rather than the transcendent itself that was being described in all these works—an inherently "troublesome" circumstance. The implication was that even the substantive definitions falter.

James did not develop a general definition but merely pointed to his topic. There could be a list of features, and any several could serve as criteria for considering something religious but need not all appear in any one religion (1902:39). James was interested in personal religion: feelings, acts, and experiences of individuals in their solitude, so far as they apprehend themselves to stand in relation to whatever they may consider the divine (1902:42). As he defined divine , it was quite broad: "only such a primal reality as the individual feels impelled to respond to solemnly and gravely" (1902:47).

Substantive definitions are sometimes thought to resist a certain ideological, passive image of humans. If religion defined functionally would benefit society, the social actor takes no directive role in the process. If religion seems to bring some cognizance of society itself into the consciousness of the adherent, religion itself is a transmitter of charisma or legitimacy to a passive social actor. But if religion is defined substantively, the religious person is seen as the active agent maintaining a particular viewpoint.

Functional Definitions

Featuring what religions do rather than their contents, functional definitions enjoy the advantage of dodging the issue of the truth-value of beliefs. Kidd (1894), influential in his day, wrote in the social evolution tradition, with a focus on conflict. Societies having features that gave them advantages in conflict survived; he called such features "functions." Religion was a function; it was any belief that provided an ultrarational sanction for the prosocial whenever individual interests and those of the social organism were antagonistic. Although similar to Comte's concept, this seemed acceptable to scholars because it was not associated, as was Comte's, with any religious program. Kidd's approach was readily incorporated into the far different sociologies of Small and Vincent (1894), Ward (1898), and Ross (1901). Only Ward made such a function definitive of religion; for him, religion would be a substitute in the rational world for instinct in the subrational world. Cooley defined religion in terms of a microfunction: a need of human nature, centering in a craving to make life seem rational and good (1909:372).

The functionalist element of Durkheim's definition—"beliefs and practices that unite into a single moral community called a Church all those who adhere to them" (1912)—was macrofunctional, too much so because it identified the religious group with the total society (Dobbelaere and Lauwers 1973). Updated applications of Durkheim's macrofunctionalism in the "civil religion" literature give that aspect of his definition a continuing relevance. Some functionalists would distinguish religion from magic, with religion serving latent functions in public activity and magic serving manifest instrumental purposes in private activity (Malinowski 1925 [1954]). Parsons, who thought religion central to "the integration of cognitive systems in their implications for action," defined religious ideas as answers to problems of meaning (1951:367 ff.).

Definitional functionalists would generally phrase religion's benefits in individualist terms in the manner of Cooley, given that evidence contradicted the macrofunctional theories (Friedrichs 1985). For Luckmann (1967), religion would be the transcending of human biological nature and the formation of a self—an inevitable occurrence that all societies effect in individuals. For Yinger (1970), religion is social but relativizes evils and desires for individuals; he defines religion as a system of beliefs and practices with which a group struggles with ultimate problems of human life. For Geertz (1966:4), religion is a system of symbols that establishes powerful, pervasive, and long-lasting moods and motivations by formulating conceptions of a general order of existence and by clothing those conceptions with an aura of factuality.

No doubt, various human phenomena could be "religious" in the functionalist manner. If functional definitions did not specify identifiable functions, one would want to label the definitions formal rather than functional . Schneider (1970) saw a potential for theoretical development in the sociology of religion in functionally examining all kinds of conduct as if it were religion, and vice versa.

Stark and Bainbridge (1979), like Durkheim, used a two-part definition when they introduced the idea of compensators , postulations of reward according to explanations that are hard to evaluate. Religion would be a system of general compensators based on supernatural assumptions. The "compensator" part is microfunctional, and the "supernatural" part, substantive.

The turn to the micro level lends an entirely different flavor to functionalism. Suddenly religion embodies the utopian spaces that the major institutions of society neglect. Religion comes to be defined as an imaginative enterprise that addresses the unfulfilled promises of life (Hervieu-Léger 1989:73).

Verstehende Definitions

Verstehen approaches to definition have not often emerged in theoretical statements in the study of religion. Searching for understandings within particular social worlds is implicit in ethnographic and participant observation methodology. Runciman (1969) criticized both substantive and functional definitions by pointing out that social actors decide such matters as whether there is a sacred-profane divide or an empirical-nonempirical distinction in their world, and that, moreover, any two members of a society might disagree (see Weigert 1974).

Again, how can we recognize the social actors' definitions as defining religion ? Swatos (1990) advocates beginning with a very minimal substantive definition, a sensitizing concept we might call it, and then using the verstehen or "definition of the situation" approach to find out what the social actors do with what had been tentatively identified as religious. The preliminary definition could use the supernatural as its criterion, with either the transcendent or the immanent being supernatural.

Formal Definitions

Formal definitions have a long history in the study of religion, but they have received little attention. Writers often cite Durkheim's Elementary Forms of the Religious Life , written late in his career (1912), as the locus of a definition that has both substantive and functional elements (Dobbelaere and Lauwers 1973), but his earlier definition (1899) exemplifies the formal approach. One seeks to find how religious facts can be grouped, even on the basis of secondary traits. Whatever is not found in all cases could not be definitive; consequently, he ruled out the unknowable as a criterion, because primitives have no such category, as well as belief in gods, because Buddhists do not have them. Durkheim ended up with a definition that he later judged valid but too formal and hard to use: obligatory beliefs connected with clearly defined practices that are related to the beliefs' objects.

Simmel, the founder of formal sociology, related religion to a moral imperative rather than to knowledge. By distinguishing such forms as moral imperative and knowledge from their contents, he was able to argue that "the religious state of the soul produces no logically necessary, pre-determined content" and that "no content possesses in itself the logical necessity to become religion" (Simmel 1984 [1903]: 69). He observed that religion, like morality, resided in the person's response to an object, and this was key to the religious form as he saw it.

Wach (1951) specified the religious response—elicited by an experience of ultimate reality, response by the whole person and not merely a cognitive or affective response, an experience having the potential of becoming the most intense of all, and leading to an urge to act. He maintained that such religion was a human universal. So long as religion would be a response, any particular content, such as the holy, would be a secondary, nondefining feature.

Formal definitions can be used with other kinds of theoretical problematics (Problemstellungen) . We have already observed that Durkheim deemed his earlier, formal definition compatible with his later functional theory, and O'Dea (1966:1) used a formal concept of religion in a functionalist presentation.

The formal strategy often looks for a structure that resembles known cases. Zeldin (1969) points to the narrative structure of a fall from and return to an ideal state—a structure she sees in Soviet Marxism and in world religions. Lemert (1975) and Blasi (1980) take the structure of related discontinuity between an empirical, mundane order and a superempirical, cosmic-level order as definitive of religion. Richard (1978) observed that this can lead to an analytical program of seeing the cosmic Otherness being domesticated (e.g., Berger 1974) or a domestic signaling of the relevance of the Other. Turner (1976) accepted the basic structure of a discontinuous relatedness but rejected the empirical-superempirical phraseology. Maduro (1982:6) shows religion's role at the nexus of two kinds of consciousness in social critique and transformation.

Anthony J. Blasi


P. L. Berger, The Sacred Canopy (Garden City, N.Y.: Doubleday, 1967)

P. L. Berger, "Some Second Thoughts on Substantive Versus Functional Definitions of Religion," Journal for the Scientific Study of Religion 13(1974):125-133

A. J. Blasi, "Definition of Religion and Phenomenological Approach Towards a Problematic," Cahiers du centre des recherches en sociologie religieuse 3(1980):55-70

C. H. Cooley, Social Organization (New York: Scribner, 1909)

K. Dobbelaere and J. Lauwers, "Definition of Religion," Social Compass 20(1973):535-551

É. Durkheim, "De la définition des phénomènes religieux," Année sociologique 2(1899):1-28

É. Durkheim, Les formes élémentaires de la vie religieux (Paris: Alcan, 1912)

R. W. Friedrichs, "The Uniquely Religious," Sociological Analysis 46(1985):361-366

W. Garrett, "Troublesome Transcendence," Sociological Analysis 35(1974): 167-180

C. Geertz, "Religion as a Cultural System," in Anthropological Approaches to the Study of Religion , ed. M. Banton (New York: Praeger, 1966): 1-46

C. Y. Glock and R. Stark, Religion and Society in Tension (Chicago: Rand McNally, 1965)

A. L. Greil and D. R. Rudy, "On the Margins of the Sacred," in In Gods We Trust , 2nd ed., ed. T. Robbins and D. Anthony (New Brunswick, N.J.: Transaction, 1990): 219-232

J. Harrison, Themis (Cambridge: Cambridge University Press, 1912)

D. Hervieu-Léger, "Faut-il définir la religion?" Archives de sciences sociales des religions 63(1987): 11-20

D. Hervieu-Léger, "Tradition, Innovation and Modernity," Social Compass 36(1989):71-81

R. Horton, "A Definition of Religion, and Its Uses," Journal of the Royal Anthropological Institute 90(1960):201-226

W. James, The Varieties of Religious Experience (New York: Modern Library, 1902)

B. Kidd, Social Evolution (London: Macmillan, 1894)

C. C. Lemert, "Defining Non-church Religion," Review of Religious Research 16(1975):186-198

T. Luckmann, The Invisible Religion (New York: Macmillan, 1967)

O. Maduro, Religion and Social Conflicts (Maryknoll, N.Y.: Orbis, 1982)

B. Malinowski, Magic, Science and Religion and Other Essays (Garden City, N.Y.: Doubleday, 1954 [1925])

T. F. O'Dea, The Sociology of Religion (Englewood Cliffs, N.J.: Prentice Hall, 1966)

R. Otto, The Idea of the Holy (London: Oxford University Press, 1923)

T. Parsons, The Structure of Social Action , vol. 2 (New York: Free Press, 1937)

T. Parsons, The Social System (New York: Free Press, 1951)

R. Richard, "Le concept de religion," Cahiers de centre des recherches en sociologie religieuse 2(1978):3-17

E. A. Ross, Social Control (New York: Macmillan, 1901)

W. G. Runciman, "The Sociological Explanation of 'Religious' Beliefs," Archives européennes de sociologie 10(1969):149-191

L. Schneider, "The Sociology of Religion," Sociological Analysis 31(1970):131-144

G. Simmel, "On Religion from the Point of View of the Theory of Knowledge," New England Sociologist 5(1984) [1903]):66-77

A. W. Small and G. E. Vincent, An Introduction to the Study of Society (New York: American Book Company, 1894)

P. A. Sorokin, Society, Culture, and Personality (New York: Harper, 1947)

R. Stark and W. S. Bainbridge, "Of Churches, Sects, and Cults," Journal for the Scientific Study of Religion 18(1979):117-133

W. H. Swatos, Jr., "Renewing 'Religion' for Sociology," Sociological Focus 23(1990):141-153

W. H. Swatos, Jr., and P. M. Gustafson, "Meaning, Continuity, and Change," in Twentieth-Century World Religious Movements in Neo- Weberian Perspective , ed. W. H. Swatos, Jr. (Lewiston, N.Y.: Mellen, 1992): 1-20

R. G. Turner, "Consciousness, Valuation, and Religion," Review of Religious Research 18(1976):25-35

E. B. Tylor, Primitive Culture (London: Murray, 1871)

J. Wach, Types of Religious Experience (Chicago: University of Chicago Press, 1951)

L. F. Ward, "The Essential Nature of Religion," International Journal of Ethics 8(1898):169-192

M. Weber, Economy and Society (Berkeley: University of California Press, 1978 [1922, original German edition])

A. J. Weigert, "Functional, Substantive, or Political?" Journal for the Scientific Study of Religion 13(1974):483-486

J. M. Yinger, The Scientific Study of Religion (New York: Macmillan, 1970)

M. Zeldin, "The Religious Nature of Russian Marxism," Journal for the Scientific Study of Religion 8(1969):100-111.

return to Encyclopedia Table of Contents

Hartford Institute for Religion Research   hirr@hartsem.edu
Hartford Seminary, 77 Sherman Street, Hartford, CT 06105  860-509-9500