A history of the secularization issue

Olivier Tschannen

Université de Fribourg

Switzerland


A HISTORY OF THE SECULARIZATION ISSUE

Original unpublished English version

of the book published in French under the title

Les théories de la sécularisation, Genève: Droz, 1992
Written at the University of Wisconsin at Madison

1989-1990
INTRODUCTION
The main purpose of this study is to provide a sociologically informed history of the secularization issue as is has been discussed among sociologists. The implications of this formulation are threefold. First, this is a study in the history of ideas. It is not a discussion of the problem of religion in the modern world. Nor is it a defense, a refutation, or even an evaluation of the validity of "secularization theory". Although the analysis sometimes uncovers the weaknesses or strengths of certain theories, and does occasionally reflect on their relation to empirical data, this is done only incidentally. Furthermore, even though I do dismiss claims put forward by certain theories, this is done only from the point of view of the history of ideas. Certain theories in effect present themselves as almost revolutionary, whereas in reality, they contain nearly nothing new. Claims of this sort have to be dismissed.

Second, the history presented here is sociologically informed. An interpretation can be properly understood only if the social background in which it plunges its roots is taken into account. Accordingly, as far as this was possible, I have written a history not only of ideas, but of the persons who put forward these ideas, and of the social contexts, in particular the scholarly communities, in which these persons were embedded. But perhaps more significantly, the whole account given here is organized according to a sociological conception of sociological theory. This conception is not normative, but descriptive: it does not tell us what sociological theory ought to be, but rests content with an empirical description of a particular type of theory construction. Specifically, I will be arguing that the recurrent emphasis on theory-construction in sociology is exaggerated; that the sociological enterprise - at least as exemplified by the secularization issue - rests not so much on formal theories and definitions, but on simple analogies and examples. In other words, this book is not only a contribution to the history of sociology, but also to the sociology of intellectual knowledge.

Third, the focus of this study is on sociological theories. The numerous theological and philosophical implications of secularization are only briefly touched upon. Unlike the great majority of sociologists who concern themselves with religion, I have no personal religious commitment. However, I am not an atheist, nor even a very systematic agnostic. As a matter of fact, until I started this work, I had never been interested in religion - not even from an academic viewpoint. The original incentive for this work was an interest in what Berger and Luckmann have called the social construction of reality - a topic that can, but need not, be discussed with reference to religion. In carrying out the research, I have remained faithful to this initial commitment to sociology as a frame of reference. Accordingly, if this book can claim to contribute to anything besides a better knowledge of the history of secularization theory, it is not to our knowledge of religion in the modern world, but to our understanding of the sociological enterprise.

Of course, these considerations do not suffice as yet to delineate my field of investigation. For it is not immediately clear what is to be understood under the "secularization issue". The notion of secularization can be defined in a number of ways and, no matter which definition one adopts, one finds that some authors speak of "secularization" without ever using the word, while others use the word while actually speaking of something else. Accordingly, any investigation in the history of this issue must rest on two preliminary decisions. The first is the definition of secularization. For reasons that will become evident in the course of this study, I will consider that a secularization theory is a body of propositions that gives an account of the situation of religion in the modern world, and that describes the evolution which led to this situation. The second decision concerns the fate of all uses of the term "secularization" that are not consistent with the chosen definition. Strictly speaking, they could be ignored. But this would greatly restrict the significance of the findings. Typically, debates among sociologists are defined by the addition of a problematic and of a term. Experience shows that, even though some uses of a term are irrelevant to a given problematic, they still impose themselves upon the discussion. Accordingly, this study is focused on the sociological theories that concern the situation of religion in the modern world (although some of these are not designated as "secularization theories"), but will not neglect uses of the term "secularization" which should in principle not concern this problematic.

Throughout this historical investigation, I will be arguing that the meaning of a text cannot be properly understood if one ignores the intentions of its author. As the same principle applies to the present text, it is only fair that I should outline what were my own intentions in writing it. This is especially important since these intentions might easily be misinterpreted, which might in turn give rise to a biased reading of the present book. As I do not spare my criticisms against certain facile and superficial reproaches addressed to "secularization theory", some readers might be tempted to conclude that my main intention has been to defend and rehabilitate the "secularization thesis". Nothing could be further from the truth. Not only because I am not very deeply involved personally in the secularization issue. But, more significantly, because my intention when I started this work was exactly the opposite. As the original dissertation proposal out of which this book grew testifies, one of my intentions was to demonstrate that secularization theory was nothing but a modern myth, or ideology of modernity - not that I had any personal reasons to dislike "secularization theory" (I did not really know it); but this approach appeared to me a convenient way to address the problematic which I was interested in: the social construction of reality. At the beginning of my research, I had been very taken in by certain criticisms of "secularization theory", and eagerly set out to uncover the mythical aspects contained in it. I hoped to be able to show that, in modern societies, sociology had taken over the role once played by religion in the construction of a coherent cosmos. However, very much to my dismay, the more I read the works of modern proponents of the "secularization thesis", the less I found reason to consider them ideological. During a few months, I felt that I would be unable to live up to the intention encapsulated in my dissertation proposal, and that I would have to abandon the whole enterprise. But I pressed on, and in doing so, my intention drastically altered. My aim was no longer to show that secularization theorists were wrong, but only to find a way to reorient my dissertation! Eventually, I found a solution to my dilemma: The people who considered secularization theory mythical were wrong, and I had been tricked by them, but at least, I felt that I was able to explain why. Not only why they were wrong (factually), but what, in the history of the field, helped explain why they had come to hold these opinions. And the only way to succeed in explaining this was in giving as complete as possible an account of the genesis of "modern secularization theory". Such, I feel, is the intention that lies behind my enterprise.

The main sources used for this history are the usual written ones: books, scattered articles, main journals in the sociology of religion, textbooks, acts of conferences, bulletins of professional associations, and the Social Science Citation Index. But an additional and very valuable source was provided by a series of formal and informal discussions with specialists in the field as well as with some other sociologists and theologians. The formal part of the investigation included twenty tape-recorded interviews with sociologists of religion, carried out in the framework of international conferences. I have made no systematic content analysis of these interviews, but have used them freely to bolster my understanding of the problem at hand. Some of the information obtained in this way is very personal - and sometimes polemical - and could not be printed, but used only as background. Some blunt remarks made during the interviews had to be revised afterwards to allow for publication (but this is always indicated). In most cases, however, quotations from the interviews have been directly integrated in the text or in the footnotes.

The most difficult and painful (as well as exhilarating) phase in any research is the entry into a completely new and unknown field. In this research, I have repeatedly had to immerse myself into unfamiliar areas of knowledge and foreign intellectual communities. These adventures would have been impossible without the help and the support of people familiar with these topics and at home in these social circles. The first incentive to study the subject of secularization came from my dissertation supervisor, Giovanni Busino, in long and repeated discussions over lunch. Without his help, I would never have been able to transform the vague problematic that interested me into a concrete research. And, I might add, without his advice and his tolerant attitude all along my endeavor, I might have abandoned the enterprise altogether.

As a result of the choice of secularization as a topic, the next step was the familiarization with religion as a subject of research. I was helped in this by my colleague Maya Burger, with whom I attended a very enlightening seminar on the history of religion given by Carl Keller. At a later stage, I received support and advice from a number of theologians and sociologists who had more intimate contact than I with religion, most notably Shaffique Keshavjee, Bernard Raymond, Richard Schoenherr, Jacques Waardenburg, and a number of participants in the Séminaire romand de troisième cycle de théologie. While in the United States, I took advantage of the religious pluralism prevalent in this country to attend a number of religious services in different congregations. I hope that, through these contacts, I have been able to develop a certain measure of empathy for the religious experience.

The next step was to penetrate into the international community of researchers in religion. I was greatly helped in this by Roland Campiche, then Secretary General of the Conférence Internationale de Sociologie des Religions. He proposed that I conduct a series of interviews to gain greater insight into the social processes behind the scene, an idea that had a great impact on my research. Further familiarization was made possible through participation in major international conferences on the sociology of religion: the CISR meetings in Tübingen and Helsinki, the meeting of the Comité de recherche sociologie de la religion of the AISLF in Geneva, the SSSR meeting in Salt Lake City, and the ASR meeting in Washington D.C.

When I was ready to start writing the dissertation out of which this monograph grew, I moved to the University of Wisconsin at Madison, where I benefited from support and friendly criticism. Charles Camic introduced me to the world of Quentin Skinner, and helped me tremendously all along by sharing with me his experience in writing intellectual history and his knowledge of American sociology. Warren Hagstrom guided my steps in my investigations into the sociology of science, and drew my attention to many a point I had overlooked in my utilization of the works of Thomas Kuhn and others. Joyce Sexton, from the Writing Lab, assisted me very efficiently in dealing with my difficulties with the English language and in my search for a more concise and precise style. I have also benefited much from exchanges in a seminar on the history of social science directed by Victor Hilts (department of history of science), and from discussions in an informal seminar with other dissertators: Su-Jen Huang, Steven Lybrand, and Richard Randell. Other friends and advisors have read my papers and helped me with their comments. I am grateful especially to Pierre Ammann, Godehard Baeck, Alexander Bergmann, Meerten ter Borg, Claudia Dubuis, Robert Alun Jones, Franco Panese, Jean-Pierre Sironneau, Laurent Thévoz, and Stephen Warner.

In addition to providing support and advice, several of the persons mentioned above have read and thoroughly criticized hundreds of pages of the working papers and drafts out of which this manuscript grew. That their name should be mentioned twice is by no means excessive. The final form of this study owes much to the constructive criticism of the members of the dissertation committee and of my advisors in Madison: Giovanni Busino, Karel Dobbelaere, Pierre Gisel, René Lévy, Charles Camic, Warren Hagstrom, Richard Lachmann, and Richard Schoenherr. To all of them, I wish to express my warmest thanks for their prolonged and patient effort, and my regrets for not always having been able to improve this study along the lines they had suggested.

Finally, I am very grateful to the twenty persons who took one or two hours from their time in Lausanne, Geneva, Konstanz, Helsinki, and Salt Lake City for interviews: Yoshiya Abe, Eileen Barker, Jim Beckford, Roland Campiche, Karel Dobbelaere, Richard Fenn, Antonio Grumelli, Danièle Hervieu-Léger, Christian Lalive d'Epinay, Leo Laeyendecker, Thomas Luckmann, Emile Poulat, José Prades, Jean Rémy, Roland Robertson, Wade Clark Roof, Jean Séguy, Liliane Voyé, Jean-Paul Willaime, and Bryan Wilson.

This research would not have been possible without the support of a number of institutions. The Department of Sociology of the University of Wisconsin at Madison invited me to be an Honorary Fellow for two years (fall 1989-fall 1991), during which I enjoyed the most nearly perfect conditions I could hope for to write this study. Adequate financial support was provided by the Swiss National Science Foundation (Commission de la recherche de Lausanne), the Fondation du 450e anniversaire de l'Université de Lausanne, the Fondation Van Walsem, the Société Académique Vaudoise, and the Faculté des Sciences Sociales et Politiques de l'Université de Lausanne.

CHAPTER 1

SECULARIZATION THEORY IN A KUHNIAN PERSPECTIVE
Since Max Weber at least, sociologists have been aware that the objects of scientific enquiry are not simply given, but must be constructed, and that this can be done only at the cost of what French sociologists call an epistemological break (Bourdieu et al. 1968, pp. 27-34; pp. 51-54), that is, a break with the common, socially given, perception. But when they turn to studying their own discipline, sociologists often fall back to a more comfortable position. They accept, as would-be "external" observers, the understanding of the field which they have as insiders. In other words, they take their familiarity with sociology as practitioners as sufficient to allow them to dispense with the usual epistemological standards. In historical investigations, this attitude often results in "presentism", that is, a reduction of past theoretical preoccupations to those current in the discipline at the time of the analysis (Jones 1977).

Nothing is more revealing of this situation than the usual approaches to "secularization theory". Sociologists who investigate this field never construct "secularization theory" as an object: they consider it as given¹. But, as disagreements and divergences among them show, in reality there is no such pre-defined object as "modern secularization theory". Indeed, the very definition of "secularization theory" is itself part of the secularization debate.

Secularization as a Theory
The debate about secularization is twofold. Not only do sociologists debate about whether secularization has happened or not; they also debate over the question whether "secularization theory" exists or not, and what this theory exactly is. This second aspect of the debate is just as important as the first one. As Pierre Bourdieu has shown in La distinction (1979), scientific definitions of reality must always be situated in a social field. This is true not only for concepts like "class", which seek to account for "social reality" through theories, but also for meta-discursive concepts, which seek to describe these theories themselves. The different definitions of the field of "secularization theory" given by different actors in the social field of sociology cannot be reduced to purely methodological or theoretical divergences: in Bourdieu's terminology, they must also be understood as efforts in the struggle for the monopoly of symbolic violence inside of the field (1976, pp. 113-20).

To illustrate: Criticisms against "the secularization thesis" are sometimes based on the contention that the concept of secularization is unclear, that the theory is contradictory, incomplete, indeed inexistent. Thus for instance Jeffrey Hadden (a critic of "secularization theory") contends that secularization is "a hodgepodge of loosely employed ideas rather than a systematic theory", and that "the theory has not been systematically stated" (1987, pp. 598-99). To this charge, Frank Lechner (a defender of "secularization theory") answers that "by strongly positivist criteria, it is hard to see what current sociological 'theories' in any domain would pass muster. [...] It is no longer clear, if it ever was, what qualifies as theory in sociology". However, he goes on to comment, "even by Hadden's standards there is a body of work that deserves to be called secularization theory" (1990, p. ?).

I am not very interested, at this particular point, in taking position in this debate (for a discussion, see chap. 13). It will suffice, for our purposes, to note that in the dispute over secularization, "secularization theory" is at stake just as much as "secularization" as such, and that there is no agreement as to what that theory is.

But - if we provisionally accept Hadden's point of view - does the absence of a unified theory really mean that there is no "secularization thesis"? The ambiguities displayed in the position of Hadden - who contends that there is no theory, but at the same time proceeds to show that the theory is wrong² (1987, pp. 598-99) - are a clear sign that the two things are quite distinct. For the absence of a unified theory does not prevent sociologists from doing research, writing books and articles, and even from obtaining certain results. Consideration of our discipline as a whole suggests that unified theories are by no means always necessary - nor even very helpful, as some famous examples should remind us.

It is indeed curious that so many sociologists should have spent so much time and energy classifying secularization theories in a seemingly desperate attempt to discover the underlying unifying theory. I would contend that these efforts can at least partially be explained by an overestimation of the importance of theory induced by an obsolete philosophy of science. For in research on secularization as in most other sub-sectors of the discipline, formal theory plays only a minor role in conducting research. In fact, for every new research, sociologists take bits and pieces from different theories, and assemble them eclectically.

I would therefore suggest that a more useful way of looking at the whole secularization issue would be to take into account the developments that have occurred in the philosophy of science with the transition from Popper (1934) to Kuhn (1962). This will induce us to stop looking for a unified theory which apparently isn't there, and to consider instead that secularization is a paradigm. The test of the usefulness of thinking in terms of paradigm will have been passed with success if it can be shown that this approach allows us to make sense of the material which is the matter for discussion, and thus to find a way out of the blind alley in which so many authors got stuck when they tried to clarify "secularization theory".

One of the most interesting consequences of the approach in terms of paradigm, as will be shown, is that even sociologists who appear to completely disagree (for instance Stark and Wilson; see chap. 13) in reality operate strictly within the same paradigm, and that their squabbles are nothing more than what Thomas Kuhn calls "mopping-up operations" (1962, p. 24), by which new elements - in the case of Stark and Wilson, the rise of the new religious movements - are made to fit into the paradigm.
Secularization as a Paradigm
In applying the idea of paradigm to the secularization issue, I will adapt somewhat loosely Kuhn's schema (1962), together with the re-reading proposed by Margaret Masterman (1970). In the Kuhn-Masterman model of science, a paradigm can be roughly characterized as follows. First, as Masterman explains, a paradigm is something "which can function when the theory is not there" (1970, p. 66). It is not a theory, but an entity which consists of at least three distinct elements: 1) a set of philosophical or metaphysical assumptions; 2) a set of universally recognized scientific achievements, which are translated into a set of scientific habits; 3) artefacts that can be used as puzzle-solving devices (1970, pp. 65-69). In the 1969 postscript to his Structure of Scientific Revolutions, Kuhn reacted positively to Masterman's reinterpretation of his concept, and proposed to distinguish between two main meanings of the term paradigm: 1) "the entire constellation of beliefs, values, techniques, and so on shared by the members of a given community"; 2) "one sort of element in that constellation, the concrete puzzle-solutions which, employed as models or examples, can replace explicit rules as a basis for the solution of the remaining puzzles of normal science" (1969, p. 175). Kuhn thus agrees with Masterman that a paradigm is not a theory, but a complex structure that includes what might be called a supra-theoretical, as well as an infra-theoretical level.

Second, if it is true that a paradigm contains a very abstract as well as a very concrete level, the latter is much more central to the Kuhnian conception. Although this did not appear very clearly in the 1962 Structure, Kuhn has since become quite explicit on this point: "The paradigm as shared example is the central element of what I now take to be the most novel and least understood aspect of this book" (1969, p. 187).

Third, the main function of the central device at the infra-theoretical level, the shared example, or exemplar, is to socialize the new student into the discipline (1969, pp. 187-89). This aspect of the Kuhnian approach has been best illustrated by Barry Barnes. "Kuhn considers a boy taking a walk with his father, in the course of which he will learn something of the different kinds of birds. We are to imagine that the child can already recognize birds, including some specific kinds of them; on the walk he will learn to recognize the hitherto unknown kinds, ducks, geese and swan. The child's father, who can be taken as a source of the accepted usage of his community, teaches that usage to the child by ostension. He points to particular birds and names them, say, as swans. When the child in his turn points to birds and identifies them as swans, the father confirms the identification, or rejects it: 'No, that's a goose'. Having seen a number of cited instances of 'swan', 'goose' and 'duck', and having himself practiced their identification under the guidance of his parent, the child becomes himself competent in identifying the three different kinds of bird. At this point the instruction is completed; the child knows the three kinds of bird" (1982, p. 23). Likewise, in the education of the would-be scientist, instruction relies not so much on formal definitions as on concrete examples of past scientific achievements. Learning is based on ostension.

Fourth, unlike a theory - which can be produced by any solitary thinker - a paradigm must be shared by an existing scientific community in order to exist at all (Kuhn 1969, p. 176). Paradigms are the result of a struggle among different groups of scientists or scholars over the cognitive models (or exemplars) that shape scientific activity. And a new paradigm can be considered to have been established only when it is routinely employed by a concrete community of researchers as a guide for the solving of the day-to-day puzzles occuring in normal science. These communities are not to be confused with the discipline as a whole. Kuhn speaks of communities of "perhaps one hundred members" (1969, p. 178), and Eckberg and Hill draw from his writings the conclusion that these communities are sub-specialities. Paradigms are never discipline-wide (Eckberg and Hill 1979, p. 929). Let us briefly note that, due to the increasing division of sociology into myriads of sub-specialities (Collins 1986, p. 1340), the notion that a paradigm concerns only a small group of specialists and not the whole discipline has become even more relevant today than it was when Kuhn was writing his Structure.

Finally, the dominance of a new paradigm must be linked to a revolution. Needless to say, revolutions are not as violent in science as they are in politics. Yet they involve a certain measure of social conflict. On the one hand, Kuhn's thinking on scientific revolutions is definitely internalist: The factors that provoke the overthrow of one paradigm and the acceptance of another are internal to the logic of science. The main factor in this process is the accumulation of anomalies (Kuhn 1962, pp. 52-65). But on the other hand, Kuhn also acknowledges that a scientific revolution cannot be brought about without a change of personnel. To illustrate this, he approvinlgy quotes Max Planck: "A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it" (quoted in 1962, p. 151). Why is that so? Paradigm changes are "changes of world views". Such changes are extremely difficult from a cognitive point of view. Thus for example, "contemplating a falling stone, Aristotle saw a change of state rather than a process" (1962, p. 124). It is doubtful whether any of us moderns can really contemplate a falling stone with Aristotle's eyes. Accordingly, "the competition between paradigms is not the sort of battle that can be resolved by proofs" (1962, p. 94). For this reason, it is often only the younger generation in a scientific community that accepts a new paradigm. Thus, all things considered, there are many similarities between scientific and political revolutions (1962, p. 94). In political revolutions, one must finally "resort to the techniques of mass persuasion, often including force" (1962, p. 93). Although Kuhn does not explicitly say so, it cannot be excluded that the same should apply, in certain cases, to the realm of science. And, I would like to add, this is all the more likely in the social sciences (especially in a field so emotionally charged as the study of religion), where one's world view almost necessarily influences one's scientific work.

The approach to the notion of paradigm just outlined is, I believe, faithful to the Kuhnian model. As a result, it is radically at odds with the reading of Kuhn most usual in the sociology of sociology. It is at odds, in particular, with one of the most well-known of these readings, that of Robert Friedrichs (1970). The approach proposed by Friedrichs and others - which, I should hasten to say, I do not think wrong, but only very un-Kuhnian - has been adopted by most sociologists. Of particular interest to us is the fact that it has been adopted by Karel Dobbelaere, the best expert on "secularization theory" as well as its most successful systematizer. According to this analysis, different secularization theories can be classified according to the paradigm on which they are based (Dobbelaere 1984). Following George Ritzer (1975), Dobbelaere broadly distinguishes between the "social facts" and the "social definition" paradigms. Which (Dobbelaere's denial notwithstanding³) is a mere variation on the old distinction between the Weberian and the Durkheimian lines of analysis, that is, between two "general sociological orientations", as Merton much more aptly calls them (1968, pp. 141-43).

Thus for Dobbelaere as well as for Firedrichs and Ritzer, a paradigm is a general orientation to sociological research based on a set of meta-theoretical assumptions. If one wishes to remain within a Kuhnian framework, several reasons militate against this use of the notion of paradigm. First, if the Weberian and the Durkheimian approaches, being grounded in existing scientific communities, loosely organized around the Archiv für Sozialwissenschaft und Sozialpolitik in one case and around the Année Sociologique in the other, could perhaps be considered as paradigms at the beginning of this century⁴, they can no longer be considered as such today. For, even if one is reluctant to follow Parson's "convergence thesis", or if one is not convinced by Berger and Luckmann's effort to bridge the two approaches, sociologists today very rarely fit exclusively into either of these two tendencies. Modern sociology has reached a point where the two approaches have somehow been eclectically combined. This view is confirmed by the difficulties into which Dobbelaere runs in his attempt. Having classified both Berger and Luckmann in the "social definition" paradigm, and finding that they both use the terms "private sphere" and "public sphere" as "objective structural elements of society" (1984, p. 207), he concludes that they have failed to live up to their own definition⁵. But to say this is to ignore that Berger and Luckmann's aim is precisely to bridge the two approaches. Of course, Dobbelaere is aware of this - he refers to this attempt by citing the Social Construction on p. 206 - but explains that, in his article, he refers only to the parts of the work of these authors which are relevant for secularization theory. Is this procedure legitimate? I think not: Authors do not write their different books in water-tight compartments, but there is always a great degree of continuity in their writings.

Second, by taking into account only the most abstract - and least significant - level in the Kuhnian paradigm, this approach totally bypasses the fundamental innovation brought about by the Kuhnian approach (Masterman 1970, p. 71; Eckberg and Hill 1979). The fact that different theories are grounded in different preconceptions can hardly be denied, but misses Kuhn's main point. The interest of his approach is that it teaches us that science does not function as its legislators, for example Popper (1934), would have it. Science is not based so much on the hypothesis-falsification cycle as on the routine solving of puzzles on the basis of exemplars. I intend to show that the same applies to sociology, and that this point of view results in a better understanding of the secularization issue.

It is my contention that, if there are any paradigms in modern sociology, they are to be found around specific sets of problems, through a combination of different "general sociological orientations". I consider that "secularization" must not be considered as grounded in different paradigms, but that "secularization" itself is a paradigm. Now this is an ambitious claim, that needs to be substantiated. Specifically, I must show how I can justify the application to sociology of a concept developed by Kuhn for the understanding of the natural sciences. If sociologists have always avoided applying the infra-theoretical part of Kuhn's argument to sociology, is it not precisely because it cannot be applied to it, or to any of the social sciences? This raises the question of the differences between the natural and the social sciences. I will be as brief as possible in addressing this thorny issue.

My use of Kuhn's model is based on the assumption that there are important similarities between the natural and the social sciences. Which does not mean that the social sciences are coming ever closer to an ideal hypothetico-deductive model, embodied in the natural sciences. Quite the contrary. The similarity lies in the fact that, all things considered, in their actual functioning, the natural sciences are not so different from the social sciences. The very revolution brought about by Kuhn must be understood in this sense. By showing the importance of world-views, socialization processes and power struggles, Kuhn has to a large extent demystified the natural sciences. This tendency has recently been pushed even further by the proponents of the "strong program" in the sociology of science (Bloor 1976). Latour and Woolgar's analysis (1979), for instance, brings to light many similarities in the ways the natural and the social sciences actually function. In both, persuasion and personal credibility play an important role. In both, "facts" are socially constructed. This is not to deny, of course, that the techniques used differ very widely. But the cognitive and the social processes involved are very similar.

Now, if these similarities warrant the use of the paradigm concept in sociology, the very important technical and epistemological differences necessitate some caution in the application of the model. The central idea underlying my attempt is that what I call a paradigm in sociology fulfills the same social and cognitive functions a paradigm fulfills in the natural sciences. To bring out this fundamental parallel, I will discuss three important characteristics that serve to define the Kuhnian paradigm: the notion that a paradigm is a shared example, the puzzle-solving function of the paradigm, and the existence of a research tradition.

To repeat, according to Kuhn, the most important element of a paradigm is the exemplar, or shared example. As an illustration, he gives Newton's Second Law of Motion, where the force is equal to the mass, multiplied by the acceleration (f=ma). This law is used as a model that can be applied analogically to a number of situations. As Kuhn explains, "it is not quite the case that logical and mathematical manipulations are applied directly to f=ma. That expression proves on examination to be a law-sketch or law-schema". In different situations, f=ma will be translated into formulas which, for the layman, have very little in common with the original exemplar, and which are not simply deduced from it through mathematical manipulations. What, then, is the function of the "law", if it is not directly applicable? As mentioned above, such exemplars play an important role in the socialization process of the future scientist: "The student discovers, with or without the assistance of the instructor, a way to see his problem as like a problem he has already encountered. [...] The law-sketch, say f=ma, has functioned as a tool, informing the student what similarities to look for, signaling the gestalt in which the situation is to be seen" (1969, p. 189).

Let us now examine an example drawn from sociology. The general context is no longer the relationship between force, mass and acceleration, but the relationship between religion and social change. Suppose an instructor in an introductory sociology course wants to make clear to his students that religion is not always a conservative force. He might choose among a variety of historical examples to make his case. But most likely he will take this opportunity to present Weber's Protestant Ethic (as he will have to present it anyway - it is on the curriculum). The core of the analysis presented in the Protestant Ethic is not much more difficult to grasp or to memorize than f=ma: To relieve the unbearable psychological stress put on him by the question Am I predestinated or not?, the puritan subjected himself to a rigorous work and life ethic, the practical results of which he then took as a sign of divine grace, which in turn relieved his psychological anxieties. This example of an unintended consequence of a theological system has roughly the same function as f=ma in the socialization process of the sociologist. It can be applied analogically to many situations where the problem of theodicy has a practical impact on social life through the mediation of psychological processes. Pondering over this example, the student will discover that a number of situations, which were until then unrelated in his mind, possess striking similarities. As a result of this, he will probably be able to generate some "original" answers of his own to the problem of the relationship between religion and social change when the time has come for him to write his exam.

At this point, it is probably necessary to briefly discuss the distinction that must be made between an exemplar and a concept. In the course of this study, I will for example be asserting that, in Parsons' theory of evolution, one encounters the differentiation exemplar (see chap. 10 and 12). This needs clarification. Surely, in Parsons' framework, differentiation is not a mere exemplar, but a concept. I do not wish to contest this. What, then, is the difference between an exemplar and a concept, and why do I reduce the concepts I encounter to the level of exemplars? I would contend that an exemplar becomes a concept when it is: 1) formally defined; 2) coherently integrated into a theoretical framework. Now, in some cases (Parsons' being the clearest exemple), this actually happens. But in some other cases, it does not. Differentiation is certainly not a concept in Richard Fenn's theory. It is never defined (nor even designated as "differentiation"), and its integration in the theoretical framework is rather approximative (see chap. 12).

Under these circumstances, why refuse to distinguish between exemplar and concept? Because, at the analytical level at which I am working, this distinction is immaterial. All concepts - even Parsons' - are based on exemplars, as is demonstrated most evidently by the fact that Parsons' formal definitions generally only make sense to the reader once they have been explained with reference to a concrete example⁶. (As a matter of fact, it is the very rarity of examples in Parsons' prose that renders it so unintelligible.) As I am working at the paradigmatic level - the only level that truly allows us to compare the different theories - I need not concern myself with concepts.

The second important element of the Kuhnian paradigm I will discuss is the puzzle-generating and puzzle-solving function. The main problem in the comparison between the natural and the social sciences again lies in the technical differences between them. I will directly turn to what I consider a good example of a puzzle in sociology of religion: the "American paradox". This paradox, as it has been expressed by Will Herberg, is that "America seems to be at once the most religious and the most secular of nations" (1955, p. 3). Now this is probably not a puzzle exactly in the same sense as in the natural sciences. But it is a puzzle, and moreover, an empirical puzzle: it concerns discrepancies in the assertions made by Americans in answer to survey questions. As Herberg explains: "When asked 'Would you say your religious beliefs have any effect on your ideas of politics and business?', a majority of the same Americans who had testified that they regarded religion as 'very important' answered that their religious beliefs had no real effects on their ideas or conduct in these decisive areas of everyday life" (1955, p. 73). This puzzle can be solved by different devices. Herberg's answer is that "the American Way of Life is the operative faith of the American people" (1955, p. 75; see chap. 10). Thomas Luckmann's is that religion in America has become "internally secularized" (Luckmann 1967, p. 37; see chap. 12). Now what is really important to note is that these two complementary answers both fit into the framework provided by the secularization paradigm (as I will describe it in chap. 4). People working in the pre-paradigmatic phase would have been hard put to solve this puzzle - if they had formulated it at all. As we can see, even if the solutions are given conceptually rather than in the form of mathematical models and/or of technical artefacts, there are puzzles in sociology, and these puzzles can best be solved in the framework provided by a paradigm.

The third important element in the Kuhnian paradigm is the existence of a research tradition. Eckberg and Hill (1979, p. 935) express the view that few such traditions exist in sociology. This may be true, but I would like to argue that, in the case of secularization, there is indeed an important research tradition, which is grounded in Weber's Protestant Ethic (1904-05), and has produced such fundamental reassessments as Lenski's Religious Factor (1962), and Berger's Sacred Canopy (1967). But what is the precise nature of this traditional link? Does it follow from the fact that a paradigm contains concrete examples embedded in a research tradition that the scientific achievement must remain tied to the context of its production? In other words, if "rationalization" is considered an examplar, or recognized scientific achievement (see chapter 4), does this imply that the sociologist who uses it today must refer to Weber and to the context in which Weber used it? The answer to that question is by no means clear from an examination of Kuhn's writings. Presumably, every student knows that f=ma is a Newtonian formula, but how much does he know about what this formula really meant to Newton, and about the context of its production?

My own answer to this question is that the more widely a scientific achievement is recognized, the more likely it is that the context in which it was produced gets forgotten. The exemplar may, or may not, remain attached to its inventor's name - this consideration does not seem to be very important. What seems more important is the broader context surrounding the invention. Latour and Woolgar have shown how "statements" (which can be more or less plausible, and are subjected to a constant process of reevaluation in an agonistic field) are transformed into "facts" (the only difference between a statement and a fact being that the latter is no longer contested, and that it is taken for granted as representing some reality which is "out there"). In this process, the most dramatic transformation occurs when someone is able to produce evidence that will dramatically improve the plausibility of a given statement (such as: "TRF is Pyro-Glu-His-Pro-NH₂"; Latour and Woolgar 1977, p. 147), pushing it beyond the point at which it becomes a "fact". At that moment, Latour and Woolgar tell us, the whole context in which the fact was constructed - the polemics that surrounded it, the technical conditions of its production -disappear. Nothing remains but an unproblematic and timeless "fact": "TRF is Pyro-Glu-His-Pro-NH₂". The "fact" has become a black box⁷: its internal structure, all the debates which have surrounded its production - and which still operate at the empirical level - disappear into oblivion.

I would like to contend that the same process - although obviouly less clear-cut - is at work in the production of "facts" and exemplars in sociology. The sociologist who uses the concept of rationalization may, or may not, refer to Weber. But even if he refers to Weber, he is very likely to use the idea of rationalization like a black box. The context in which the idea was born - Weber's personal fight against "the forces that unite to rationalize society and destroy individual autonomy" as they were embodied in his mother's religious ethic and in his father's "authoritarian despotism at home and bureaucratic subservience in his office" (Mitzman 1969, p. 179) - is forgotten and, if it is not forgotten, it is most of the time discarded as irrelevant. As a result of this process, the elements that make up paradigms in sociology seem to float in a timeless Platonician realm of ideas. Although the exemplars which undergird these elements may still be invoked, they are detached from the context in which they were produced.

This has important consequences for this study. As we will see, the exemplars used by modern secularization theorists can generally not be traced directly through footnotes to the sociologists who created them. Thus for example, all modern secularization theorists use the notion of differentiation (see chap. 4 and 12), but none of them (except Parsons) explains where this notion comes from. The reason for this is evident: differentiation is so fundamental an exemplar in the education of any sociologist that there is no need to refer to Spencer, to Durkheim, or to Parsons. "Differentiation" is used as a black box. Consequently, in tracing the genesis of the modern approach to secularization, we will not seek to trace the origins and development of such exemplars through footnotes. We do not need footnotes to tell us that modern views on differentiation are the result of many successive contributions, among which those of Spencer, Durkheim, and Parsons are foremost. In other words, to the extent that modern secularization theorists do not take the trouble to indicate the origin of the terms they use, they signal that they are the stock in trade of the sociologist.

Any sub-discipline is built on a few traditions. If the view I have sought to express is correct, these traditions find their strength, not so much in the empirical relevance of the theories, but in the practical usefulness of a few exemplars which are somehow interrelated and can be used analogically to understand the world. This becomes quite clear when we compare what Weber and Parsons, on the one hand, and what Sorokin and Becker (see chapter 9), on the other, have written on the relationship of religion and modernization. Weber and Parsons immediately "make sense" - even if we disagree with them - whereas the writings of the other two simply strike us as irrelevant and outmoded. This is not, I would contend, because Becker's or Sorokin's theories are stupid, or wrong: when one pauses to think of it, they are just as clever theoretically and relevant empirically. But we have learned to see the world through exemplars drawn from Weber and Parsons, and not from Sorokin and Becker.

To conclude, my approach to paradigms is centered on (but not limited to - see chap. 4) the notion of exemplar. An exemplar is a relatively simple and briefly formulated "image" of the world, which is widely recognized as useful and used analogically in a great number of situations. Furthermore, the context in which the analogy was first made is only of secondary importance: the exemplar has acquired a timeless quality. Therefore, in the framework of this study, the history of the secularization issue is largely the history of the exemplars used by modern secularization theorists. Before writing this history, we must become acquainted with the works of these scholars to determine what exemplars they have been using.