A history of the secularization issue

Karel Dobbelaere
Karel Dobbelaere was born in 1933²⁹⁸. After studying at the Catholic University of Louvain (Belgium), he worked on a number of empirical research projects in the sociology of religion. Already in these first empirical studies, Dobbelaere resented the absence of a sound theoretical framework to guide his steps. In 1963-64, he did postgraduate work at the UC at Berkeley²⁹⁹. In 1968, back in Belgium, he published a trend report on the state of sociology of religion, and expressed his dissatisfaction with its narrowly empirical focus. He concluded by welcoming Berger and Luckmann's recent attempts at theoretical formulation (Dobbelaere 1968). As already mentioned, Dobbelaere early became a very active member in the CISR³⁰⁰, and one of the most fierce opponents of the rearguard of Catholic sociologists. At the 1969 CISR meeting, together with his colleague Jan Lauwers, he again hailed Berger and Luckmann's sociology of knowledge as a promising new approach allowing to move beyond sociologie religieuse (Dobbelaere and Lauwers, CISR 1969, p. 123).

Dobbelaere's original position with regard to secularization was less than enthusiastic. In a paper written with Lauwers, he expressed reservations about different types of secularization theories, uncovering a variety of ideological elements in them (Dobbelaere and Lauwers 1973, pp. 547-51). In spite of these different interventions, by the end of the 70s, Dobbelaere was not, as yet, a significant figure in the study of secularization. His contribution to the paradigm hinges mainly on one work, the trend report written at Wilson's All Souls College in Oxford in 1977, which appeared in 1981 under the title Secularization: A Multi-Dimensional Concept (Dobbelaere 1981). As the title and the genre indicates, this report did not aim at presenting a new theory, nor even an original contribution. And indeed, Dobbelaere's contribution to the secularization debate is limited to matters of systematization and conceptual clarification. This, of course, is no menial task; moreover, for our present concerns, this contribution is particularly important. As a matter of fact, Dobbelaere's contribution has perhaps been the most efficient of all in promoting the secularization issue to a paradigmatic status. By offering an overview of the field, by mapping the terrain, identifying the most important names in a quasi-official manner, and organizing them into a clear pattern, Dobbelaere has transformed a somewhat shapeless "debate" into a quasi-institutionalized arena.

Dobbelaere's most important theoretical innovation is conceptual. He proposed to consider secularization a multi-dimensional concept, which could be broken down in three separate dimensions: laicisation (societal level), religious change (institutional level), and religious involvement (individual level). In a later paper, however, he acknowledged that his choice of the term laicisation had been rather unfortunate: "From reviews and discussions with colleagues, I became aware that the term 'laicisation', which I used in the sense of the French term, laïcisation, was confusing. [...] I now return to the term 'secularization'" (Dobbelaere 1984, p. 200). Preferring, in the present context, precision to historical accuracy, I will follow the same path, and replace "laicisation" by "secularization" in the discussion of the 1981 trend report.

To repeat: Secularization must be considered as a multi-dimensional concept, referring to three separate dimensions. When these three dimensions are considered together (which is rare, and does not really make much sense in terms of a sociological analysis), Dobbelaere refers to "secularization in general" (1984, p. 200). These terminological discussion are important for our purposes. I would contend that "secularization in general", as a somewhat unwieldy multi-dimensional concept, corresponds to the pre-paradigmatic phase of secularization theory. "Secularization", as one of the three dimensions, purged, as it were, of parasitic accompanying phenomena, corresponds to the paradigmatic phase of secularization theory. In other words, by "secularization" (or "laicisation" in the 1981 report) as over against "secularization in general" (or "secularization" in the 1981 report), Dobbelaere, borrowing Shiner's (1967) somewhat confusing terminology, refers to the most central exemplars of the paradigm: "desacralization [rationalization*³⁰¹], differentiation, and transposition [generalization*]" (Dobbelaere 1981, p. 11). The two other dimensions, "church involvement" and "religious change", which were very important in the pre-paradigmatic phase, have become definitely peripheral in the CISR paradigm.

Let us start by taking a brief look at the two dimensions which were, as it were, discarded by Dobbelaere³⁰². First, "religious change" refers to all the changes, generally in the direction of a modernization of religion (worldliness*) occurring inside religious groups. The classical expression of this approach is the church-sect dichotomy, which was so central to the sociology of religion for several decades. In proposing to replace the notion of secularization with that of religious change, Dobbelaere follows in the steps of Yinger³⁰³ (Dobbelaere 1981, p. 7), and answers a long-standing complaint by many sociologists to the effect that it is wrong to judge negatively the adaptation of religion to worldly values. Second, "religious involvement" refers mainly to changes in religious practice, as measured in surveys. This dimension, which was central in the period of sociologie religieuse, is considered by Dobbelaere not to be a very reliable indicator of secularization. Thus, Dobbelaere's conceptual clarification essentially amounts to this: Secularization must be considered a phenomenon occurring at the societal level. Phenomena taking place at the institutional level (religious change) or at the individual level (religious involvement) can be no more than accompanying dimensions within the phenomenon of secularization in general.

As Dobbelaere's main contribution consists in the drawing of the map of the field in which the secularization debate is held, we must take a close look at the structure of the central part of his trend report, devoted to secularization. The chapter is divided into five sub-chapters. The first deals with the problem of the "historical base-line": If we want to assert that society has become secularized, we must contrast this state of affairs with some previous historical period, in which society was not secularized. Quoting Martin, Parsons, Gabriel Le Bras, Jean Delumeau, Mary Douglas and Wilson, Dobbelaere argues that the secularization thesis does not imply that we have fallen from some previous golden age of religion: The Middle Ages, for instance, were certainly not as unanimously Christian as sometimes thought. The second sub-chapter concerns the definition of religion. After discussing the functional definitions (Yinger, Andrew Greeley, Parsons, Bellah), Dobbelaere argues in favor of the use of a substantive definition. Next, he turns to the problem of social integration, and presents the arguments of people like Parsons and Bellah, who think that modern society needs religion for its integration, and of people like Wilson, Fenn, and Luhmann, who think that it does not. Dobbelaere sides with the latter, and concludes: "According to my definition, civil religion is not a 'religion'" (1981, p. 48). The fourth sub-chapter is devoted to the question of the unilinearity of the secularization process. Referring mainly to Martin and to Fenn, Dobbelaere shows that secularization must not be construed as a unilinear, straightforward evolutionary process. The last sub-chapter puts forward a distinction between two lines of analysis in secularization theory: the "Durkheimian" line (Bellah, Parsons, and, to a certain extent, Luckmann), and the "Weberian line" (mainly Wilson and Berger). Finally, Dobbelaere briefly considers some other approaches, most notably Daniel Bell's, Niklas Luhmann's, and Joachim Matthes'.

Dobbelaere's general argument can be summarized as follows. First, there are a number of misconceptions concerning secularization that must be cleared away. On these points, discussion is not possible: Secularization is not equivalent to a fall from grace, it is not a process of religious pathology, and it is not an unilinear evolutionary process. Second, there are a number of points on which discussion is possible. Religion can be viewed either as necessary to social integration, or as not necessary. The first approach considers secularization from a Durkheimian point of view; the second approach takes a Weberian perspective. Finally, there is one rather peculiar point, concerning the definition of religion. Dobbelaere argues that all working definitions of religion are substantial; that even people who define religion functionally resort to a substantial definition once they settle down to write an historical account, and that therefore, it is best to opt for an explicitly substantive definition.

On this basis, Dobbelaere decides who has put forward "good" secularization theories, and who has put forward "bad" theories, which must be rejected (of course, he does not quite present the choice in these terms; but that is what he in effect does). In doing this, he draws the boundaries of the CISR paradigm. Let us first take a closer look at the people who are rejected. First, of all, and most importantly, Catholic sociologists, as well as Marxists, were wrong because they equated secularization with decline in practice: "We should make a clear distinction between secularization on the one hand and religious participation on the other" (1981, p. 4). Similarly, a number of researchers in the area of religion (for instance Harold Pfautz) were wrong because they equated secularization with religious change. Finally, among those who did none of these mistakes, Sorokin was wrong because his historical base-line was unacceptable.

Among the people who are right, we find, first of all, those who work with a substantive definition, and in a Weberian line of analysis: Wilson, Fenn, and Martin. Fenn's and Martin's theories are especially adequate since they underscore the complexity of the secularization process. People working in the Durkheimian line - Luckmann, Parsons, and Bellah - are not as close to doing "good" theory as those working in the Weberian line, but, given certain qualifications, their accounts can also be accepted. Their only mistake is that they pretend to be working with a functional definition. Fortunately, however, they revert to a substantive definition once they settle down to do historical work.

Dobbelaere's contribution thus represents the most explicit recognition of a tendency characteristic of the mature phase in the life of the paradigm. Whereas in the first phase (Luckmann, Berger, and Wilson), theorists refused or ignored the civil religion hypothesis, in the second phase, as we saw first with Martin, then more clearly with Fenn and with Dobbelaere, civil religion became an accepted constituent part of secularization theory. Now of course, Parson's generalization theory, as well as the first formulations of civil religion theory by Herberg and Marty, were all published well before the first secularization theories of the CISR paradigm. The reason I chose to defer discussion of Parsons and of Bellah's hypotheses to the end of this presentation is that I wanted to avoid giving the spurious impression that there was any historical continuity from Parsons and Bellah to Berger and Wilson. As a matter fo fact, quite the reverse is true. Even though Berger and Wilson were, of necessity, acquainted with Parsons' work, their actual sources of inspiration were quite different. Parsons and Bellah were drawn into the paradigm against their will, and at a later stage, most notably by Dobbelaere's intervention. We will therefore presently turn to Parson and Bellah. Before that, however, a few more words on Dobbelaere are in order.

It is not necessary to engage into a more detailed discussion of Dobbelaere's trend report. The bulk of it consists in detailed presentations of the arguments of the secularization theorists. Since I have already presented them, I could only repeat myself. It is clear, however, that Dobbelaere's reconstruction is different from mine: instead of highlighting the exemplars, he highlights a number of theoretical questions. Although the general picture is different, I would contend that our accounts are not at odds, but rather complementary.

In later papers (Dobbelaere 1984; 1985; 1987a; 1987b), Dobbelaere essentially reiterated the arguments put forth in his 1981 report. In the first two of these papers, he adopted a somewhat different approach, looking for divergences - in terms of different "paradigms"³⁰⁴ - among secularization theorists instead of looking at convergences. The list of important secularization theorists was reduced to four: Berger, Luckmann, Wilson, and Luhmann³⁰⁵. However, Fenn and Martin were eliminated only for reasons of expediency: they were not really necessary for the purpose at hand (1985, p. 383). The only theoretical shift that is noticeable as far as Dobbelaere's conception of secularization goes is an increased stress on differentiation as a central theoretical element - which is probably the reason why Luhmann (see chap. 3) was brought into the picture (or the result of his being brought in).

Talcott Parsons
The two carriers of the secularization paradigm which have not been discussed yet are Parsons and Bellah. These two men developed their approaches in close cooperation, and acknowledged their mutual debts in several places³⁰⁶. Their links with theorists explicitly working in the framework of the paradigm, however, are almost non-existent³⁰⁷. Starting with Parsons - some of whose early writings were already discussed in chapters 9 and 10 - we will discuss only that small fraction of his later writings which bears directly on two exemplars: differentiation and generalization.
Differentiation
Differentiation is not specifically a part of Parsons' sociology of religion, but it is fundamental to his whole sociological approach. According to Parsons, differentiation is an "evolutionary universal", that is, "a complex of structures and associated processes the development of which so increases the long-run adaptative capacity of living systems in a given class that only systems that develop the complex can attain certain higher levels of general adaptative capacity" (1964, pp. 340-41). In other words, Parsons draws an analogy between the "human" and the "sub-human", or biological realms, and argues that in both, any major transformation that will dramatically increase the society's, or the organism's, adaptative capacity, is very likely to occur, and will do so even in different evolutionary branches that are maintained separated. Thus for instance, vision developed independently in the molluscs, in the insects, and in the vertebrates.

In this framework, Parsons distinguishes three types of differentiation. The two first types (vertical and horizontal differentiation) occur at the social-structural level; the third occurs between the social and the cultural levels. Vertical differentiation, which he also refers to as stratification, occurs when members of cadet lineages in the kinship system, for instance, are put into subordinate positions. Horizontal differentiation is a process which allows for the concentration of responsibilities for certain functions of vital importance - most importantly politics and religion (1964, pp. 342-43). The best example of the emergence of a distinctly religious sphere is the emergence of the notion of "church", which "constituted the differentiation of Christianity as a religious system (a cultural system) from the conception of a 'people' as a social system" (1963a, p. 40).

As we can see, the notion of differentiation is not limited to a division of labor in society, but also includes a differentiation between the "cultural" and the "social" spheres. To understand this, we must start from the idea that "the differentiation inherent in stratification creates new sources of strain and potential disorganization" (1964, p. 345), and that this strain in the social system can be eased only by the rise of cultural legitimations explaining the "necessity" of this incipient stratification. Thus, legitimation itself becomes a distinctively differentiated social function: "Legitimation is the differentiation of cultural definitions of normative patterns from a completely embedded, taken-for-granted fusion with the social structure" (1964, p. 346). One precise example of such differentiation is the emergence of a "moral order" during the Reformation. With this moral order, the individual was no longer controlled directly by the Church hierarchy, but by the "normative order of secular society" (Parsons 1962, pp. 148-150).
Generalization
The first sketch of Parsons' generalization thesis was put forward in 1957, in an article we discussed in chapter 10. Rather than presenting this early version again, I will directly present the theory in its mature form, as it appeared, most notably, in Christianity and Modern Industrial Society (Parsons 1963a). In this paper, Parsons proposes to consider that, although religion has to a large extent become privatized as a result of differentiation - a fact which can be interpreted as a loss for religion - this is compensated for by another, complementary process: generalization. To quote a formulation used by Parsons in another paper, "there is indeed now a sense in which church religion has come to be largely privatized, but concomitantly religious or quasi-religious significance generalized to an immense range of what previously were defined as more of less purely secular concerns" (Parsons 1971, p. 230). Privatization "need not [...] imply that 'religion' has declined. It has, to be sure, assumed a more restricted place in the cultural and social system, but this may in fact strengthen rather than weaken its strategic importance" (Parsons 1966, p. 125).

By generalization, Parsons refers to the idea that, in the course of its development, Christianity succeeded in institutionalizing in society at large a number of originally religious values, thus giving them a more general significance and a larger applicability. This process started with early Christianity, and was much furthered by the Reformation (Parsons 1963a, p. 51). For our own purposes, it will suffice to consider the last stage in this process, that of American denominational pluralism: "From a religious point of view, [the voluntary principle and denominational pluralism represent] the discrimination of two layers of religious commitment. One of these is the layer which defines the bases of denominational membership [...]. The other is a common matrix of value-commitments which is broadly shared between denominations, and which forms the basis of the sense in which the society as a whole forms a religiously based moral community" (1963a, p. 62). In other words, the American situation of fragmentation in different denominations can continue to exist only because of the existence of another, non-fragmented level of commitment, which makes for the possibility of tolerance and allows individuals to transfer from one denomination to another (Parsons 1966, pp. 136-38). This is the view which justifies the use of the term "generalization"³⁰⁸.

Not only is the religious commitment thus generalized, but the religious and the secular spheres become closer to each other. If the secular world were regarded as radically evil, the tolerance demanded by pluralism could never be attained. "For such a religious constitution to function, on the institutional level the society must present not a less but a more favorable field for the Christian life than did the society of earlier periods of Western history; its moral standards must in fact be higher. [...] In a whole variety of respects modern society is more in accord with Christian values than its forebears have been" (Parsons 1963a, p. 65).

Thus, Parsons uses the term "secularization" with a meaning exactly opposite to the usual. As he himself remarks, he is being "deliberately paradoxical in attributing to the concept secularization what has often been held to be its opposite, namely not the loss of commitment to religious values and the like, but the institutionalization of such values" (Parsons 1971, p. 217, n. 11). Thus, in a moral community characterized by what Parsons, following Bellah (Parsons 1968, pp. 442-43), calls a "civil religion", every individual - even a "secular humanist" (Parsons 1971, p. 228) - can be considered a believer (1971, p. 224). Another consequence of this view is that Marxism can be viewed as another form of religion, and that the cold war can be interpreted as a new "war of religion" (Parsons 1974, pp. 207-09).

Robert N. Bellah
As noted above, Bellah's approach was developed in close cooperation with Parsons'. Not surprisingly, it is also very close theoretically to the latter's. Both approaches are based on an evolutionary view which involves two complementary processes: privatization, and generalization. In Parsons' theory, the complementarity of these two processes is clearly set forth. In Bellah's work, it is only latent³⁰⁹. The first half of the picture is contained in a paper entitled Religious Evolution (Bellah 1970, pp. 20-50); the second half is contained in the paper on American Civil Religion (Bellah [1967] 1970, pp. 168-89).
Privatization*
Bellah's paper on religious evolution represents an explicit effort at reviving an evolutionary approach to the study of religion (1970, pp. 20-21). In keeping with Parsons, Bellah defines evolution as "a process of increasing differentiation and complexity of organization that endows the organism, social system, or whatever the unit in question may be with greater capacity to adapt to its environment" (1970, p. 21). Unlike Parsons, however, he applies this idea not to the evolution of the social system, but to that of religion as a symbol system that "relate[s] man to the ultimate conditions of his existence" (1970, p. 21). The resulting evolutionary view applies exclusively to the structure of this particular system: "Neither religious man nor the structure of man's ultimate religious situation evolves [...], but rather religion as a symbol system" (1970, p. 21).

Religious evolution thus conceived combines two trends: complexification and individualization. On the one hand, starting from an early undifferentiated cosmos, religious symbolization first becomes dualistic (with a distinction between this world and another world), then splits into an "infinitely multiplex" structure (1970, p. 40). On the other hand, starting with the first rejections of the world around the first millenium B.C., the human self starts to become conscious of itself. The ultimate consequence of this rise of individual consciousness is the privatization* of religion in modern society. These two trends are analyzed in terms of a succession of five stages: primitive, archaic, historic, early modern, and modern religion.

For our present purposes, it will suffice to make some comments on the last stage of this evolution. According to Bellah, since we are becoming increasingly aware of the nature of symbolization itself and subject it to systematic intellectual inquiry, it is doubtful whether we can even speak of a modern religious symbol system. Hierarchic dualism has completely collapsed, as has become evident in the works of theologians like Tillich, Bultmann, and Bonhoeffer. This breakdown, however, has not entailed a return to what would be an equivalent of primitive monism. Humans do not look at themselves in an unfragmented mirror: "It is not that a single world has replaced a double one but that an infinitely multiplex one has replaced the simple duplex structure. It is not that life has become again a 'one possibility thing' but that it has become an infinite possibility thing. The analysis of modern man as secular, materialistic, dehumanized, and in the deepest sense areligious seems to me fundamentally misguided, for such a judgment is based on standards that cannot adequately gauge the modern temper" (1970, p. 40).

In this new situation, the task of religious symbolization of the world "is no longer the monopoly of any groups explicitly labeled religious" (1970, p. 42). What we have, instead, is a recognition of the fact that every individual must interpret the received tradition in a personal manner. "The fundamental symbolization of modern man and his situation is that of a dynamic multidimensional self capable, within limits, of remaking self-transformation and capable, again within limits, of remaking the world, including the very symbolic forms with which he deals with it" (1970, p. 42). These trends must not be interpreted as a "secularization", but as a new way of producing religious symbolization, in which it is the individual, and no longer the group, which must "work out his own ultimate solutions" (1970, p. 43).

Generalization*
Bellah's views on generalization* are almost entirely contained in the paper on civil religion³¹⁰ ([1967] 1970, pp. 168-89). The civil religion hypothesis is very close to several earlier formulations, most notably Herberg's notion of operative religion³¹¹ (see chap. 10). Bellah's thesis is that "there actually exists alongside and rather clearly differentiated from the churches an elaborate and well-institutionalized civil religion in America" ([1967] 1970, p. 168). To sustain his claim, Bellah analyzes excerpts from presidential speeches and from the Declaration of Independence. He notices that they contain very frequent references to God - but not to the God of any denomination in particular, only to an anonymous God. Through this reference, political power comes to be responsible not only before the people, but also before God - thus gaining a new legitimation, which is in some ways superior to democratic legitimacy. But marks of religious symbolization in American public life are not limited to this reference. Civil religion is a very complex entity, which can be described as "a set of beliefs, symbols, and rituals" ([1967] 1970, p. 171). As for the beliefs, they are centered on the idea that America is God's chosen people - a new Israel. American is called to "carry out God's will on earth" ([1967] 1970, p. 172). The symbols of the civil religion are mostly drawn from the Bible: "Exodus, Chosen People, Promised Land, New Jerusalem, and Sacrificial Death and Rebirth" ([1967] 1970, p. 186). Finally, good examples of national religious rituals are Thanksgiving Day, which serves to "integrate the family into the civil religion" ([1967] 1970, p. 179), and the inauguration of a new president.

American civil religion has always been closely tied to the unfolding of American history. Religion, and particularly the idea of God, played an important part in the thought of early American statesmen. Until the Civil War, "the Declaration of Independence and the Constitution were the sacred scriptures and Washington the divinely appointed Moses who led his people out of the hands of tyranny" ([1967] 1970, p. 176). With the Civil War, new themes entered the civil religion: death, sacrifice, and rebirth. Lincoln, the assassinated president, was likened to Jesus, and the huge number of war casualties led to the establishment of national cemeteries, around which a new rite, Memorial Day, came to be performed. Thus, although civil religion is not a religion in the conventional sense, it has "its own prophets and its own martyrs, its own sacred events and sacred places, its own solemn rituals and symbols" [1967] 1970, p. 186).