A history of the secularization issue

Beyond the Secularization/Revival Debate: Danièle Hervieu-Léger
In the framework of the paradigm, Wilson's and Stark and Bainbridge's views positions are not incompatible. But the paradigm does not answer all questions. On the question of the nature of the impulse that lies behind new religious movements, we have noted that Wilson and Stark fundamentally diverge. This is the point at which Danièle Hervieu-Léger enters the arena.

Hervieu-Léger's 40 pages on secularization (Hervieu-Léger and Champion 1986, pp. 187-227) are the most complete account on secularization and new religious movements to have appeared in France to date. This analysis allows us to replace the Wilson-Stark divergence in a somewhat broader perspective. On the European continent, the debate in the sociology of religion mostly turns around the antinomy between dechristianization and popular religion, whereas in the anglo-american world, the antinomy, as we know, is between secularization and religious revival. Both these antinomies, Hervieu-Léger argues, pose the problem of the nature of modernity.

Having carefully analyzed both the continental and the Anglo-American debate, Hervieu-Léger proposes a solution of her own to the dilemma, thus explicitly positioning herself beyond the two competing approaches. Her proposal hinges on the notion of the paradoxical nature of modernity. Historically, modernity is a process which can be described in terms of progressive rationalization. But at the same time, modernity functions as a utopia showing humans the way for a better world³⁴⁶. These two processes always occur simultaneously: the second permanently recreates what has been destroyed by the first. "Modernity abolishes religion as a meaning system and as the motive behind human efforts, but at the same time, it recreates a room in space and in time for a utopia which, in its very structure, remains related to a religious problematic of accomplishment and salvation. [...] This utopian space [...] constitutes a room for the permanent redeployment of 'religious representations', which rationalism also permanently decomposes. [...] Secularization is not the disappearance of religion confronted to rationality: it is the process of permanent reorganization of the work of religion in a society which is structurally unable to fulfil the expectations that must be raised for it to exist. The analysis of this reorganization process could be the principal task of a sociology of belief which, for our modern societies, has yet to be formulated" (Hervieu-Léger and Champion 1986, pp. 224-27; see also Hervieu-Léger 1990). In other words, to use an expression proposed by Hervieu-Léger earlier in her book, secularization is a process of permanent recomposition of the forces of religion in society (1986, pp. 140 and 213).

In a later paper, setting out to accomplish the program of the "sociology of belief" proposed in her book, Hervieu-Léger not only refuses to side with either school in the issue of secularization and revival: her refusal extends to the question of the definition of religion. Escaping the functional versus substantive definition dilemma, she proposes a new definition of religion that combines two simple elements: belief and tradition. "The 'religious belief' of modern societies refers neither to objects of particular belief, nor to specific social practices, nor even to original representations of the world, but is characterized by a particular mode of legitimation of the belief act, that is the appeal to a tradition" (Hervieu-Léger 1989, p. 77). Hervieu-Léger thus escapes the fallacious opposition between "religion" and "rationality" which, from the XVIII^th century to our days, has caused an artificial division of the social scientific field in two parties, one side claiming that rationality is in the process of eliminating religion, the other maintaining that religion is back and will ultimately vanquish rationality. Building on the inherent ambiguity of modernity, which was apparent as early as in the works and life of Comte, founder of positivism and of a new religion, she proposes an approach which allows us to understand that man is inseparably a rational and an irrational being, and that society inevitably and simultaneously produces both rationality and new myths.

The Systems Theory Reformulation: Niklas Luhmann
As mentioned before, Luhmann is generally considered Habermas' prime opponent in Germany. This is not surprising, since Habermas is an heir of the Frankfurt school, whereas Luhmann is following in the steps of Parsons, whose systems theory he seeks to update. In this paper, I will not dwell on this relationship, but rather concentrate on the problem of secularization. Unlike Habermas, Luhmann examines the problem of secularization in detail. One of the chapters of his Funktion der Religion (Luhmann 1977), over fifty pages in length, is entirely devoted to secularization. Moreover, he has devoted a paper specifically to the problem of civil religion (Luhmann 1981). His knowledge of the field appears wider than Habermas'. As a matter of fact, he often cites the works of several carriers of the CISR paradigm: most prominently Fenn³⁴⁷, but also Wilson, Berger, Dobbelaere, Bellah, and of course Parsons. He is also acquainted with some of the peripheral literature. However, he does not really start from the formulations of the CISR paradigm, as is evident from the fact that he does not acknowledge the centrality of differentiation in this paradigm. When the totality of his very numerous sources is taken into account, it appears that he is influenced at least as much by German sociology of religion (Rendtdorff, Lübbe, Matthes, Schelsky, etc.) as by the anglo-american debate. Moreover, the essays contained in Funktion der Religion were written as a basis for a debate with theologians (Luhmann 1977, p. 7). As a result of this peculiar insertion, Luhmann presents himself as putting forward a "new" idea, to wit: Secularization must not be considered a transition from the "sacred" to the "profane", but a result of functional differentiation.

Before discussing Luhmann's secularization theory, I must give a brief account of his general approach to the problem of religion. Luhmann defines religion in functional terms. However, his definition is different from those that attribute to religion an integrating function, but also from those that attribute to religion an interpreting function (1977, pp. 9-13). To understand how his functional definition differentiates itself from the first tradition, we must understand his global systems theory. To understand how it differentiates itself from the second tradition, we must turn to his theory of religion.

Starting with the systems theory, it must be noted at the outset that Luhmann's views are very difficult to present - even more than Parsons'. In the latter case, the problems posed by the abstractness of the theoretical system were greatly alleviated by the fact that, when Parsons discussed religion, he usually came down from the analytic to the historical level. In contrast, Luhmann never bothers with detailed historical analyses: The historical examples he gives are nothing but convenient illustrations of his systems theory. My attempt to present Luhmann's theory will therefore inevitably result in some shortcuts and simplifications. Furthermore, as Luhmann's terminology never speaks for itself, a certain amount of "translation" will also be necessary. I will therefore not be able to stick as closely to Luhmann's own formulations as would be desirable.

Bearing this in mind, we can start with the notion of contingency. The world is experienced by humans as contingent because it is largely unpredictable. Thus for example, in a social relation, ego can never be certain what alter expects of him, and vice versa (Kiss 1986, p. 4). Of course, in practice, this contingency is not total. It is largely brought under control by social systems. Contingency can be reduced not only by conforming to institutionalized behavior-patterns (as Parsons would have it), but also through communication. Communication can act as a system-building mechanism (Kiss 1986, p. 5). This explains the pervasiveness of informal behavior models even in the midst of the most rigidly institutionalized orders (Kiss 1986, p. 66).

Because the world is contingent, it is experienced by humans as complex (Komplex). This means not so much that the world is complicated (Kompliziert) - which would only mean that it consists of a great number of different elements - but that it offers a variety of options for choice (Kiss 1986, p. 7). Every choice we make in life ipso facto leaves out a number of alternative choices we could have made, and others around us may be making. This consideration applies not only to the individual, but also to social systems and subsystems. As soon as a given subsystem is no longer unequivocally related to another subsystem, the system as such has become complex. In other words, any particular buildup of a system necessitates a choice among a number of contingent buildup possibilities.

It is also necessary to briefly present Luhmann's view of "society". A society is not made of individuals, nor even of interactions, but of communication (Kiss 1986, p. 17). Society as a system does not include individuals (as psychological and as organic systems). Individuals are part of the environment of the social system (Kiss 1986, p. 75). Thus, the social system is not more, but less than a sum of individuals: it "contains" only the communications between them. But even the notion that something is "contained" in a system is misleading. A system is defined not by its parts, but by its border. Society is a system to the extent that it marks the border separating the communications between a number of persons on the one hand, and the environment (that is, everything that does not belong to this system of communications: other communications, but also psychological and organic factors) on the other hand (Kiss 1986, p. 19). By tracing this border, the system ipso facto reduces the complexity of the environment. In other words, a social system (which can be as small as a communication system between two persons³⁴⁸) determines its very shape by excluding some possibilities, which are therefore relegated to the environment. (Thus for example, the different subsystems become differentiated when they develop each their own communication medium - "money" in the case of the economy, "truth" in the case of science - which has no currency in the other subsystems.) The environment is thus not a fixed entity: to each system corresponds a different environment. And each system is always less complex than its environment.

We can now understand in which sense Luhmann's functional analysis of religion differs from those that are based on the notion of integration. Luhmann considers that society does not need religion to be integrated. However, he redefines integration in a particular way: it is nothing more than the capacity to resist disintegration. As a result of this definition, to solve the problem of integration, it is sufficient to ensure that "the operations of a subsystem do not create insoluble problems in another subsystem" (Luhmann 1977, p. 242). Briefly, in contrast to the Parsonian notion of integration, where all subsystems are subordinated to a single element, Luhmann maintains that modern society is characterized by the fact that the different subsystems cannot be subordinated to a single principle, but must themselves work out ways to coordinate their actions. To use a convenient image, integration does not come from "above", but from "below"³⁴⁹. The "hold", or "grip" (Zugriffsform) of the system on the subsystems shifts from the subsystems' structures to their intrasocietal environment. In other words, integration is ensured by the fact that each subsystem constitutes a part of the environment of all other subsystems (Luhmann 1977, p. 243). In the new formulation put forward by Luhmann in the 1980s (which is not entirely compatible with his former views), this idea has been reformulated: Social systems and subsystems are autopoietic (that is: they build themselves autonomously). In other words, social subsystems are considered closed self-referential systems (Kiss 1986, p. 83) that adapt themselves to each other³⁵⁰.

Let us now turn to Luhmann's theory of religion. As noted above, a system is defined by its border, which allows it to control the difference in complexity between the outside and the inside, thus allowing action to take place. The control of complexity by a system (i.e., the choice it operates in the contingency of its environment) gives rise to meaning (Sinn). Meaning is the element that relates a system to its environment. Through the constitution of meaning, a given system constructs an image of "the world", that is, an horizon in which the system itself and its environment are comprised and related to each other (Luhmann 1977, p. 22). The social function of religion is related to the creation of meaning. To understand how this creation is possible, we must start from the notion that all that has been excluded from the system through complexity reduction nevertheless remains present to the system, and can be referred to as that which is appresented, in contrast to that which is represented³⁵¹: "That among which the selection has been operated remains as active appresented horizon" (Luhmann 1977, p. 22). The creation of meaning typically involves two distinctive processes. First, some possibilities present in the contingency are eliminated through self-reference. Any system (individual or social) that acts in an environment perceives itself as acting in that environment. Therefore, it can self-referentially take into account this appresented level. Thus for example, "when one communicates on different themes in a social situation, one always keeps in mind that the communication takes place in the given concrete situation among given partners and must therefore take into account more than its theme" (Luhmann 1977, p. 23). Second, that which is appresented can be brought back into the picture: it can be represented. In other words, it is possible to discuss reflexively about problems of meaning. However, these two procedures do not cover the whole range of the meaning-construction processes empirically available in human societies. The reason for this incompleteness is that there always remains some part of that which is appresented which cannot be represented.

Religion sidesteps this problem by using a different procedure. Religion is not based on the dichotomy between representation and appresentation, but on an opposition between representation and counter-representation (1977, p. 25). By this device, the problem of the incompleteness of that which is appresented dissolves. That which is represented possesses its exact, complete counterpart: the counter-representation (for example: the counterpart of "this world" is the "other world"). Thus, religion possesses the capability to deal with problems posed by the very act of representing. To reformulate: the very act of representing presents certain risks. Choices among the contingency can be operated through self-reference and through representation of the appresented. But ultimately, these procedures are bound to fail, because that which is appresented can never be completely represented. Religion reduces this specific risk by replacing all that is appresented (all the contingent possibilities) by a counter-representation (that is: no longer a variety of contingent possibilities, but the counterpart of that which is represented). Religion thus has a stabilizing function in the economy of meaning. To use a later formulation, religion "deparadoxalizes" the world (Luhmann 1985, p. 8).

In modern societies, the function of religion remains the same as ever. "The complexity of the world that can be represented and is accessible to operations based on meaning can be enormously increased when the social system correlatively reaches an adequate structural complexity. However, through this operation, the horizons that define the problem of reference of religion are only pushed aside, not surpassed. The world that is constantly implicit and taken into account in the processing of meaning through appresentation remains impossible to formulate, and this is the precise point to which religion's formulation problems refer" (Luhmann 1977, p. 25). As we can notice, although the function of religion in Luhmann's theory superficially resembles the "interpretation" thesis, it is much more specific.

Bearing these introductory remarks in mind, we can now turn to the problem of secularization. Luhmann's secularization theory is based on the notion of differentiation, which is another of the core elements of his broader theory. Although it is expressed in rather complicated terms, Luhmann's notion of differentiation in itself is hardly original (although some of its consequences are). In undifferentiated societies, religion was equivalent with the social system as a whole, and religion's environment was extrasocietal; in differentiated societies, religion has become a subsystem, whose environment now is intrasocietal (Luhmann 1977, pp. 227-28). As a result of this functional differentiation, intrasocietal communication - and the media through which this communication is made possible - takes on a crucial importance. Thus for instance, "money" as the economic subsystem's media, and "truth" as the scientific subsystem's media, have gained extra significance for the functioning of the system.

Secularization is nothing but the consequence of this general process as seen from the point of view of the religious subsystem. Secularization is not a process endogenous to religion. It is not even primarily a transformation of religion itself - even though, as we will see, it does induce some changes in religion. Secularization does not even result in a transformation of the functions of religion: religion is as relevant for the social system now as it was before. One of the results of this perspective is that secularization can be studied independently from the definition of religion one is working with (1977, p. 227). Strictly speaking (although Luhmann does not explicitly say so), secularization is not a sociological category. It is a term used by members of the religious subsystem to describe this evolution as they perceive it (1977, p. 229).

Historically, the first consequence of differentiation for religion was positive: as religion took the form of a religion of faith (Glaubensreligion), it became itself one of the intra-societal communication media, and took on a dominant position (1977, pp. 230-31). As noted above, in this position, it fulfilled a very important function for the system as a whole. Difficulties started when, as a result of further differentiation, this function itself was drawn from the outside to the inside of the system (1977, p. 231). As a result of complexification, a need for contingency-control appeared inside the system, because each subsystem must now reduce the complexity of its (intrasocietal) environment. The consequences of this development for religion are three-fold. I will express them in standard terminology. First, the relationship of the individual to the social order is altered (section II in Luhmann's chapter on secularization in 1977). Second, the integration of the social system must be based on new premises (section III). And third, the religious subsystem must modify itself in order to adapt to its complexified environment (section IV).

Let us turn, first, to the problem of the individual's relation to the social order. As a result of functional differentiation (as opposed to differentiation in terms of social strata, or of social segments), it is no longer possible to ascribe each individual authoritatively a role (1977, p. 236). At the same time, each individual must be guaranteed access to all the subsystems. It is not tolerable, for example, that someone be restricted his access to one subsystem (for example education) on the grounds of his participation in another subsystem (for example a particular religion). (It should be noted that this does not imply that everyone, for instance, can become a physician, but only that everyone can become a patient [Kiss 1986, p. 40]) Now, it must be noted that this problem is general, and does not exclusively concern religion. The ways in which the problem has been solved for the economy (through the free market and the private ownership of the means of production) and for the polity (through democracy) gives us a clue to the solution applicable to religion. The solution consists in the individuation of decisions³⁵² (Privatisierung des Entscheidens). As this solution applies to all the institutional spheres, and not only to religion, it must not be confused with the notion of privatization*. The individuation of religious decisions does not imply that religion has become irrelevant to the workings of the social system. Quite the contrary: it is exactly as relevant as the economy or politics, where the individuation of decisions also applies. Exactly as with the economic sphere's invisible hand, the individuation of religious decisions guarantees the stability of the social system through the statistical neutralization of individual decisions (1977, p. 238). However, the consequences of the individuation of decisions are similar to those of privatization* to the extent that individuation results in religion becoming part of the sphere of leisure (Freizeit), where it is subjected to a harsh competition (1977, p. 239).

The second point concerns social integration. Luhmann's position with regard to social integration has been described above, and is roughly equivalent to Fenn's. The only point that must be briefly touched upon is the place of civil religion in this framework. Civil religion does indeed play an integrative role, but only in the integration between the political and the religious subsystems, not for society as a whole. In the period during which societies became functionally differentiated, civil religion provided a much-needed link between the newly segregated subsystems, in that it guaranteed every participant in the political subsystem equal access to the religious subsystem, and vice versa (Luhmann 1981, pp. 294-300).

Finally, the last consequence of secularization concerns the inner transformations of religion. As noted above, the complexification of society results in the emergence of differentiated subsystems, which each have their own communication media (1977, p. 251). The question that then arises is: To what extent has this brought with it a transformation of the function of religion? Luhmann presents an answer which is not to be found in the CISR paradigm, but which is reminiscent of secular theology: as a result of differentiation, religion has become "purer" [eine Art Selbstreinigungsprozess der religiösen Funktion ist im Gang]³⁵³ (1977, p. 39). As other subsystems have taken over the secondary functions religion fulfilled, it is now possible to find answers to religious problems which are no longer contaminated by considerations external to the sphere of religion. Religion can find a new function in society if it evolves a dogmatic theology (Dogmatik) adapted to the complexification of its environment (1977, p. 253).

To conclude, how does Luhmann's approach compare with the CISR paradigm? The first difficulty with this comparison is that Luhmann does not use exemplars. His pronouncements do not take the form of analogies which help to see the world in a new way, but the form of formal definitions which, at least for those not thoroughly acquainted with his system, rather tend to obscure any understanding of the real world. Moreover, he is not preoccupied with puzzle-solving, but with system-building. As he does not indulge in empirical research, he does not build up his theoretical concepts to answer particular problems encountered in research, but to fit them into a preconceived grand theoretical scheme. Finally, Luhmann is not primarily a sociologist, but a systems theorist. He is interested in the social world only to the extent that this world can be reduced to abstract systems applicable along the entire range of empirically available phenomena. Therefore, in Luhmann's case, we must not ask which exemplars he uses, but rather with which exemplars his system is compatible.

As noted above, Luhmann reads a lot. He is obviously familiar with most exemplars used in the sociology of religion. Those that can fit into his scheme, he adopts and, if necessary, modifies to suit his needs; the others, he rejects. We will not attempt to follow all the theoretical ramifications of a particular exemplar adopted by Luhmann throughout his theoretical system: All we need to be concerned with is the problem of secularization. The most important exemplar used by Luhmann is obviously differentiation. As noted above, even if the particular theoretical system in which it is inserted is novel, the notion of differentiation itself as used by Luhmann is not original. Several exemplars related to the notion of differentiation are also used, albeit in a modified form. Thus, autonomization* takes place (but the idea is presented in a theoretical framework it would have been tedious to present here); there are some connections between the individuation of religious decisions and privatization*; and Luhmann accepts a weakened formulation of the notion of generalization* (in his discussion of civil religion). But Luhmann radically rejects the notion of religious roots*, of rationalization, and of worldliness*. Finally, he adds to his theory another idea, generally considered foreign to the sociological debate: the purification of religion.

To conclude, Luhmann's treatment of secularization is certainly one of the most original of those we have examined. It is one of the very few that proposes ideas not fully compatible with the CISR paradigm. How much of it can be integrated into mainstream sociology of religion remains to be seen.