A history of the secularization issue

Empiricism and the Catholic reaction in American sociology (1918-1949)
As, in the precedeing chapter, we moved from Ward and Sumner to Small and Giddings, and then to Park, Sorokin and MacIver, we noticed that the space devoted to religion steadily increased. This increase of interest was due in part to the European influence; but I would like to argue that it was also related to the rise of empiricism. After World War I, American sociology developed an increasingly empirical character (House 1936, p. 294). Although the sociology of religion remained a marginal area¹¹⁶ (Lenski 1962, pp. 308-09; Glock 1959, p. 153), empiricists did not write off religion a priori, and devoted some attention to it. However, this was done in a manner that could not please Catholic sociologists. The new attitude to religion can be illustrated by the treatment reserved to it in the leading sociology journal in these years, the American Journal of Sociology. Starting in 1928, religion found a recognized place in the AJS in the yearly special issues devoted to "recent social changes" in areas such as population, education, rural life and a few others (see for instance Holt 1928). These special issues appeared again every year¹¹⁷ until 1935. The purpose of this innovation, as stated in the foreword (1928, AJS 34 [1]), was "to present records of fact rather than speculative or exhortative articles". In 1935, "Church attendance" became one of the social trends regularly investigated by William Ogburn - alongside suicides, telephone conversations and first class mail (!) - and presented in systematic tables (see for instance Ogburn 1935). The picture of religion which emerged from these figures was not very optimistic. Increasingly, the yearly reports tended to indicate a decline in religion. In 1942, Hornell Hart, in a new report, summarized the trend of the decade as a whole. He documented the decline of religion in several detailed tables, using indices like the decline in religious articles in magazines, the decline in Church membership, in Church attendance, in religious education, the increasing criticisms against the Church, the crisis in Christian missions, and the decline in the proportion of scientists believing in God. As in the previous reports, no comments accompanied these dry figures (Hart 1942).

It is against this general background that the emergence of Catholic sociology must be understood. For one thing, Catholics felt that an exclusively empirical approach to the study of society was totally improper. As one of the Catholic participants in the ASS meetings recalls: "We were pretty much disgusted with the meetings of the A.S.S. First of all, the papers were largely research topics, and to all appearances, it was largely research for the sake of research. Secondly, the outlook of these sociologists was poles away from ours. They were just in that period when Sociology was a science copying its procedures from the natural sciences. For these secular sociologists the approach was supposedly scientific and objective but, unconsciously, for all practical purposes, anti-moral and anti-religious" (Dayton Friedel, quoted in Morris 1989b, pp. 329-30). Furthermore, Catholics resented the "hostile condescension of academia and the scholarly community" (Morris 1989b, p. 329). Thomas Imse, who claims to have been the first person acknowledging being a Catholic "in the sociology department at any of the major state universities of the country" (Imse 1989, p. 404), explains: "Many universities did not want Catholics on their faculties in fields such as sociology, since they could not be trusted. They would probably teach religious morality and not be able to stick to the facts on a variety of issues. The intolerance was quite openly expressed in classrooms and even in public discussions" (Imse 1989, p. 403).

In 1938, reacting to the view held by lay sociologists that religion was nothing but a hindrance to scientific study (Reiss 1970, p. 120), a few Catholic sociologists founded the American Catholic Sociological Society (ACSS). This particular Catholic reaction in the field of sociology was part of a larger reaction against modernism in American Catholicism. "The establishment of the American Catholic Sociological Society was but one instance between the wars of a trend in American Catholic intellectual circles to create learned societies as counterparts to existing organizations in the larger society" (Kivisto 1989, p. 352).

To justify the creation of a distinctively Catholic type of sociology, Ralph Gallagher, the founder of the ACSS, stressed that sociology was "not in the full sense of the word an exact science" (Gallagher 1938, p. 319), and that every sociology was implicitly based on a social philosophy. "Thus Catholic sociology was looked upon as one of several kinds of sociology each one being determined by the social philosophy associated with it" (Reiss 1970, p. 121). Paul Mundie¹¹⁸, in his comments introducing the Journal of the newly founded association, explained: "Catholics will not be poorer scientists, but rather better scientists for knowing that man has a supernatural destiny and that society has as its goal the organization of social life in such a manner as to help man attain this destiny" (Mundie 1940, p. 321). This view resulted in a sociology in which Encyclicals of the Popes were used as basic texts in sociology courses (Reiss 1970, p. 121).

The rest of the history of the ACSS, which became the Association for the Sociology of Religion (ASR) in 1970, can be compressed in a few sentences for our purposes. The initial confidence in the virtues of a sociology based on Catholic doctrine did not last. As early as 1941, Catholic sociologists began feeling ill at ease in this framework. They started to ponder their role, and the relationships between social philosophy, social science and social reform. From 1949 on, these doubts led the leaders of the association to insist very heavily in their presidential addresses on the shortcomings of Catholic sociologists as empirical sociologists: No one could be satisfied any longer with the "second-rate scholarship" and "apathy toward social research" (Thomas Harte, cited in Reiss 1970, p. 124). Eventually, abandoning its original aim of providing a general sociology based on Catholic doctrine, the ACSS began to devote its energies to a sociology of Catholicism based on standard sociology (Reiss 1970, p. 126), thus rejoining the mainstream of lay sociology.
Empiricism and Catholic sociology in Europe (1929-1959)
The tensions we have encountered in America between Catholic and lay sociology have their counterpart in Europe. However, as already mentioned, the situation in Europe was rather different from what it was in America. For one thing, the rise of empiricism occcurred later. In France at least, the classical period lasted until 1929: Despite Durkheim's death in 1917, the group identified with L'Année sociologique remained dominant for a number of years (Stoetzel 1957, p. 630). The end of this dominance signaled a shift in the interests of sociologsts. According to Jean Stoetzel¹¹⁹, after the end of this period, the number of empirical studies increased steadily. In parallel to this evolution, the overall number of studies bearing on religion declined significantly. But while studies relying on the "humanist approach" declined sharply, an interest in religion reappeared "in an empiricist form" (1957, p. 626).

Another important difference between Europe and the United States concerns the European Catholic association - the Conférence internationale de sociologie religieuse (CISR). First, the CISR never tried to promote an all-encompassing Catholic sociolgy as the ACSS had, but from the start restricted itself essentially to religion as a subject-matter¹²⁰ (Dobbelaere 1989, p. 385). Furthermore, on the old continent, Catholic sociology was not a reaction against empirical sociology, but participated in its very development. In 1931, the man who was to become the "father" of French "religious sociology", Gabriel Le Bras¹²¹, published a very influential paper in which he called for a far-reaching empirical analysis of Catholicism (Steeman 1958, p. 2). His call was heeded, even if it took over a decade (Hervieu-Léger and Champion 1986, p. 30) to really set off what was to be the most important series of empirical studies on religion ever attempted. The Catholic Church's support for these studies, which was first somehow slow to come, was bolstered by a growing awareness of the alienation of the European working class from the Church. H. Godin and Y. Daniel's book, La France, pays de mission?, published in 1943, played a very important role in this new awareness¹²². Thus, as David Martin explains: "The relegation of religion to a footnote of the basic political text has resulted in the attempt to create for 'religious sociology' an equally independent watertight compartment concerned with its own specific order of phenomena. [...] Catholics have been perhaps specially prone to this empiricist version of the sociology of religion, and to a reduction of its scope to sociologie religieuse" (Martin 1966, p. 359).

As mentioned above, the most distinguished representative of European Catholic sociology was Gabriel Le Bras. His importance was not due only to the impact of his 1931 article, but also to the very important empirical work he did himself (Steeman 1958, p. 2). As Stoetzel explains: "Le Bras has formed a genuine school; and the monographs in sociology of religion now are models for monographs in urban and rural sociology, as well as in electoral sociology" (1957, p. 656). If we want to understand Le Bras' influence, we must take into account the fact that ha was not merely a professor of Canon Law, but also an academic holding many prestigious positions¹²³. Owing to this great intellectual and moral prestige, Le Bras was called to preside over the birth of the CISR¹²⁴, but also, paradoxically, over that of the rival lay Groupe de sociologie des religions (Poulat 1969, p. 7). But Le Bras was not the only Catholic to play a prominent role in the development of general sociology. To give just one example, Jacques Leclercq, the founder of the CISR, introduced the teaching of sociology in Louvain (Poulat 1990, p. 12; Dobbelaere 1989, p. 377, n. 1). As a result of these initiatives, by the beginning of the 50s in the field of sociology of religion, "Catholic religious sociology occupied the ground and had outrooted a Durkheimian school which was without posterity" (Poulat 1990, p. 25).

The CISR was founded in 1948 in Louvain, Belgium, by Jacques Leclercq, a Catholic Bishop (Monseigneur; Poulat 1990, pp. 12-13). At that time, the CISR had no more than the 16 members who participated in the first meeting. The first full-fledged conference was held in 1951 and, from this date onwards, conferences were held roughly every second year throughout Europe.

Originally, the group was non-confessional (Poulat 1990, p. 15). The founder of the Conference - although he was a Bishop - deemed it necessary to leave the door open to persons from other faiths. But very soon, he found himself isolated: Most members were overwhelmingly hostile to this open spirit. The 1951 Conference was attended by 49 churchmen and 17 laymen. Under the influence of the Dutch KASKI (Catholic Institute for Socio-Ecclesiastical Research), the executive committee decided that the association would henceforth further contacts exclusively among Catholics preoccupied with religious sociology (1990, p. 21). The association was to retain this confessional character until 1971, and at every conference, the few lay sociologists who attended found themselves stifled in a heavily Catholic atmosphere. Karel Dobbelaere, the present president of the CISR, remembers: "The older members may remember the typical Sunday morning: everyone went to mass and in rows - one after the other - to communion, and, at the same time, colleague priests were saying mass at the different sidealtars, a new mass starting every 15 to 20 minutes. And in Rome, in 1969, we were housed in a convent, each day at noon there was a mass, and we had rooms with doors we could not lock, maybe because it was the year of Humanae Vitae" (Dobbelaere 1990, p. 44, n. 1).

The theoretical orientations of the CISR were concordant with this membership. In the first conferences (3 to 5¹²⁵, or 1951 to 1956), issues all revolved around institutional religion: the Church, its ministry, its membership. The proportion of papers directly based on empirical research was very high (particularly in 1951 and in 1956). Another feature of these first conferences was the effort made to take stock of the state of religious sociology. This effort was apparent in the contents of the articles as well as in the general themes of the conferences: "Etat de la sociologie religieuse"; "Sociologie religieuse, sciences sociales"; "Vocation de la sociologie religieuse [...]". This feature must be understood in the light of the hopes entertained by the Catholic hierarchy that religious sociology could be used as a tool in the fight against dechristianization: "'Sociologie religieuse' was clearly a sociology at the service of the Catholic Church. [...] Clerics were not interested in scientific discussions, which they interrupted with questions about results; and scientists interrupted the discussion of results with methodological questions" (Dobbelaere 1989, pp. 378-79). Indeed, the proponents of religious sociology did not miss one opportunity to point out the practical results and the theological legitimacy of their approach¹²⁶. Sociology had become an ally of the Catholic church in its endeavor to bring the lost masses back to Christianity. In the poetic words of a country priest and amateur sociologist: "Today, one can no longer conceive a priest who would not first embrace as closely as possible the sociological reality to which his apostle's heart is called to devote itself" (Schmitt-Eglin 1952, p. 12).

Religious sociology was also a moralistic enterprise. Authors frequently made references to moral or pastoral dimensions. Talking of dechristianization, some members used the phrase: "l'évolution du mal" (Georges Delcuve, CISR 1951, p. 377). In this context, sociology was considered a remedy for the ailments from which the Church was suffering, not as a tool to analyze the crisis. Researchers took the temperature of the patient, but were not interested so much in the "illness" itself as in how to fight it. They wanted their research to be practically applicable in the daily life of the Church, to provide ready-made solutions¹²⁷.

The relationships between Catholic and classical sociology have best been summarized by Poulat: "As the Durkheimians were anticatholic, it was only natural that the Catholics should be antidurkheimian" (Poulat 1990, p. 16). This "natural antagonism" explains why religious sociology refused to establish itself in the framework provided by classical sociology, and preferred to start from scratch (Goddijn 1958-59, p. 18). The opposition between these two views was sharp, and revolved around the central question whether being a believer hurt or helped a sociological understanding of religion. A sociologist who was present at the CISR conference in Bologna in 1959 testifies that "this point has raised particularly sharp discussions in the lobbies of the Bologna conference" (Maître 1960, p. 79). Similarly, at the 5^th World Congress of the ISA in Washington, a quarrel erupted between the agnostic Wilson, who raised "the question of whether religion can be studied at all by a religious man", and the Catholic Fichter, who "testified to the possibilities of retaining both scientific objectivity and religious dedication" (Moberg 1962, pp. 568-69).

To illustrate the drawbacks of the kind of sociological approach vindicated by Catholic sociology, I will use a few excerpts from the only book stemming from this period whose title (Le mécanisme de la déchristianisation) indicates an effort in the direction of a global secularization theory (Schmitt-Eglin 1952). In this book, the inadequacies of Catholic sociology are, admittedly, presented under their worst light. But, even though this sociology was not always so extreme, this particular book gives a fair representation of the general direction in which European Catholic sociology was drawn.

Schmitt-Eglin's approach can only be described as naive and obsolete; the author apparently knows nothing - or wants to know nothing - about the modern sociological approach. In a tone reminiscent - at best - of Plato, he explains: "Normally a society is organized in such a way that the most evolved members take into their hands the direction of affairs and provide the common good to the profit of the least able. Beneath them lies the most numerous part of the population, deprived of culture, of training, and of initiative" (1952, p. 69). The individuals belonging to this strata he describes as "beings who have been stopped in their intellectual and moral growth, who act by intuition rather than by reason and tend to mistake their wishes for reality" (1952, p. 70).

The phenomenon of secularization itself is not really analyzed by Schmitt-Eglin. Indeed, this is not necessary, since dechristianization "is immediately apparent" (1952, p. 201). In his view, dechristianization is linked to a social decline, a sad retreat from the golden age of Catholicism, which he describes in idyllic terms: "In the state of Christianity, the role of the priest is to be the chief of the people. He is the lord of the village and the father of his congregation" (1952, p. 223). The moralistic attitude clearly shows through the vocabulary used: Instead of speaking of the seeds of secularization, the author speaks of "les bacilles de la déchristianisation" (1952, p. 14). However, he acknowledges that, even in this golden age, "approximately 20%" of the population did not fit into this framework. The vocabulary in which he describes this unchristened minority is a world view in itself: "This 20% can represent a human waste of retarded beings, of embittered and vicious personalities as the best parishes inevitably drag along, beings of reduced vitality" (1952, p. 225).

Thus, we now stand at the beginning of the 50s, the period in which the seeds of the CISR paradigm were sown. To summarize what has been said thus far: On both sides of the Atlantic, the study of religion was undertaken mainly by Catholic sociologists, whose approach was far from adequate from a sociological point of view. Lay sociologists either largely ignored religion, or imprisoned it into theoretical frameworks which precluded any study of the religious institution as such. Thus, the divorce between Catholic and lay sociology was almost complete. But at the same time, on both sides, an increasing number of people were dissatisfied with this situation. We will now turn to the evolution which allowed the progressive closing of this gap. We will first follow the rise of interest in sociology of religion among lay sociologists, then turn to the deconfessionalization of Catholic sociology, which will be studied mainly through the example of the CISR. At the same time, we will follow the evolution of the use of the term secularization.

The growth of interest in sociology of religion in America (1929-1966)
The 50s were marked by a resurgence in interest in sociology of religion in the United States (Glock 1959, p. 154). This resurgence was related to a specifically American phenomenon: the "religious revival" which occurred after World War II (see below). As interest in religion rose in the public, it was quite natural that interest in the scientific study of religion should rise among academics - especially as this apparent revival seemed to come as a dismissal of the reasons these academics could have had for ignoring religion.

In Europe, similar tendencies were also at work, but they were much less pervasive than in America. Writings showing a return of sociology of religion to mainstream sociology - and, by the same token, an increased involvement of lay sociologists in the field - appeared earlier in America than in Europe. In his 1968 trend report, Karel Dobbelaere remarked that in America, "41% was socio-ecclesiastical research and 46% was genuine sociology of religion", while in Europe, "genuine" sociology of religion made up no more than 21% of the works analyzed (1968, p. 334). Moreover, the kind of sociology of religion practiced in Europe was more parochial and less theoretically informed than in the U.S.A.: up to 82% of the articles had no reference to "previous research concerning the conceptual framework" (1968, p. 340). Similar conclusions had been reached two years earlier in a trend report by Schreuder (1966, p. 211).

The increase in interest in religion in America was also closely related to the rise of the functional model in all branches of sociology. The advantages in terms of recognition in the academia brought about by functional theory were claimed by sociologists of religion as well as by specialists in other sub-disciplines. Of course, the advantages were not only institutional, but also theoretical. So for instance, Glock, writing in 1959, criticized the "inability [of the framework bequeathed us by the classics] to account fully for the place of religion in modern, complex, changing societies". He deplored the fact that studies based on this tradition had "a tendency to ignore the functional alternatives to religion in meeting social needs" (1959, p. 155). Glock suggested that the structural-functionalist approach, in particular as represented by Bellah, might provide the necessary framework.

The functional model exerted a tremendous influence on the development of sociology of religion. This influence is so pervasive that it is simply impossible to imagine what would have become of the new discipline had it not been for functionalism. How, and why, did functionalism exert such a powerful influence? And when did this change take place? In reality, the change advocated in the late 50s by Glock had to a large extent taken place some time before as a result of the gradual emergence of specializations rendered possible by the sheer growth in the number of professional sociologists in America. This is apparent even in the textbooks¹²⁸. In the 1930s, there was a marked change in the general organization of the textbooks. The new pattern which emerged in this period was based on a functional view heavily influenced by recent developements in anthropology. "For a time after its publication in 1937 the single most popular introductory text was that by Robert L. Sutherland and Julian L. Woodward. The authors organized the bulk of their material around major 'social institutions' (an unfortunately ambiguous term), meaning the family, economy, polity, education, and so on. The outstanding feature of this text was the introduction of each of these 'core' topics with descriptive material from exotic (mainly tribal) societies" (Moore 1978, p. 335). From this time on¹²⁹, a large majority of the texts came to be organized around a core of chapters analyzing the different institutions playing major functions in societies. As a consequence of this new pattern, a majority of the texts appearing after 1930 included a separate chapter on religion¹³⁰. Generally, these chapters presented religion as an institution alongside other institutions, all of which were frequently included in some more general subdivision¹³¹.

This new pattern represented more than a mere reorganization of the material. It allocated religion a totally new place in sociology. From the moment religion was considered one of the major institutions in any society, it tended to be treated as an independent variable, and the necessity for a specialized sub-discipline studying this institution became apparent. But not only did religion become one among several important social institutions; with Parson's systems theory, it was placed at the top of the pyramid of cybernetical control, thus becoming even more important than the other institutions. Whereas the evolutionary perspective had mainly tried to explain religion away as a remnant of the past, the functional perspective sought to explain it as a present feature. The shift from an evolutionary to a functional perspective almost automatically entailed a reassessment of the importance of religion in modern society.

The most important trends in the evolution of the treatment of religion in the textbooks appearing after 1930 are all related in some way to the emergence of the structural-functional model. We will first analyze in some detail a text which was very influential in the development of American sociology (Moore 1978, p. 336; Coser 1971, pp. 567-68): Kingsley Davis' Human Society (TXT 1948). In this text, Davis - one of the most vocal and well-known proponents of functionalism - first presented the evolutionary theories of religion (under the title "Outmoded theories of religion"), only to better criticize them. The first critique addressed to evolutionary theory was that the accounts of the origins of religion were highly speculative, "based on imagination rather than direct evidence" (TXT 1948, p. 515). More important, however, was the contention that present-day social institutions could not be explained by an account of their origin, but must be explained by their actual functioning. The evolutionary school was also criticized for having a rationalistic explanation of religion, which tended to view religion as an "error" rather than as a means of fulfilling useful functions. Finally, in the "faded pages of evolutionary theory", the whole account of religion was based on an individualistic point of view: "All that is required is a single individual sleeping, dreaming, looking at nature, and speculating" (1948, p. 518).

Davis then presented the functional alternative, based on William Robertson Smith, Durkheim, Malinowski, and Parsons. The adoption of this new view severely undermined the disappearance thesis. One of the reasons why Davis disapproved of evolutionary theory was the following: "The logical conclusion from the evolutionary theory is that religion will ultimately disappear, being replaced by science" (1948, p. 517). In contrast, in the functional view, religion and science are no longer viewed as antithetical (1948, p. 535). "So long as one attributes religion merely to an error in reasoning, its future seems dim. One simply imagines that science will increasingly expose the error and religion will then disappear. One may even cite the decline of religion in modern society as evidence for this contention. But this rationalistic point of view itself falls into an error. It fails to see that the nonrational character of religion performs a function for both the society and the personality" (1948, pp. 541-42).

Davis' chapter on religion closes with two pages in which he sketches a secularization theory, which he himself presents as containing five core elements. "First, the gods tend to be gradually withdrawn from the local scene" (1948, p. 542). Which means that trans-local, more remote gods, make their appearance. Second, there is a diminishing of anthropomorphism. Third, there is an increasing separation between religion and everyday life: Religion becomes a category by itself. Fourth, with the appearance of new faiths, there is a diminution in the homogeneity of religious life, and fifth, "some of the essentially religious sentiments and ceremonies become attached to the state [...]. The stage is set for church-state conflict" (1948, p. 543). Although the analysis remains very sketchy, it draws the inevitable conclusion a consciously functionalistically informed analysis of religion in modern society cannot fail to draw: There is an increasing differentiation* between religion and the world.

The general framework provided by this analysis of Davis' textbook is found again, with some variations, in a great number of the subsequent texts. The organization of the chapter on religion, starting with a presentation of evolutionism, followed by a presentation of the functional approach, is repeated in TXT 1950. Similarly, although it does not start with evolutionary theory, TXT 1951 also presents criticisms of evolutionary theory, followed by a presentation of functionalism. Later texts¹³² generally omit any detailed discussion of evolutionary theory, and simply start from the functional point of view - a clear sign that the transition to the new framework had been accomplished. In many cases, the functional point of view results in a very broad definition of religion, including "such doctrines as humanism, nationalism, imperialism, internationalism, fascism, communism, democracy, and finally science itself" (TXT 1954, p. 562)¹³³.

Some of these chapters also close on some considerations on religion in modern society. Many exemplars make their appearance in some of these texts, for instance worldliness*¹³⁴, scientificization*¹³⁵, differentiation*¹³⁶, generalization*¹³⁷, autonomization*¹³⁸, and pluralism*¹³⁹. Progressively¹⁴⁰, empirical data concerning religious affiliation and attendance were being added¹⁴¹. Another trend was, from TXT 1956a onwards, to take into account the American "religious revival" thesis¹⁴². However, as mentioned in chapter 2, "secularization" itself was very rarely mentioned in these chapters.

Thus, the emergence of the functional viewpoint provided the general framework for the growing interest in sociology of religion. But this evolution was also marked by a number of important studies and writings, to which we turn now¹⁴³. We will pay attention in particular to two themes: the growth of interest in sociology of religion, and the secularization question.
Richard Niebuhr (1929)
The first book on American religion to reach a wide audience was Richard Niebuhr's Social Sources of Denominationalism (1929). Niebuhr was a theologian¹⁴⁴, and his book is, at heart, an exhortation to his coreligionists, asking them to revive the true spirit of Christianity, which has been so badly compromised by its subservience to the secular realm. Denominationalism - the subject of the book - is considered by Niebuhr "a compromise, made far too lightly, between Christianity and the world" (1929, p. 6). Denominationalism "signalizes the defeat of the Christian ethics of brotherhood by the ethics of caste" (1929, p. 22). But although the underlying aim of the book is theological, the author uses the sociological approach to make his point - and, it should be added, he uses it very skillfully. Sociology allows Niebuhr to bring through to his readers the fact that divisiveness in the church is not of theological provenance, but reflects "the effect of social tradition, the influence of cultural heritage, and the weight of economic interest [...] in the definition of religious truth" (1929, p. 17).

The core of the book is devoted to an analysis of the different churches: the churches of the disinherited, of the middle class, of the immigrants, of the different races, and so forth. Niebuhr brings to light the relationship between the emergence of new denominations (Anabaptists, Quakers, Methodists, Salvation Army) and the rise of new classes - thus providing the first influential analysis of the influence of society on religion. He also brings out the characteristics of the religiosity of the disinherited: emotional fervor, but also moral discipline - the very discipline which allowed these classes to rise. Concomitantly, he retraces the process by which these classes and denominations lost their original fervor, and became more conservative, once they had become dominant.

In this framework, Niebuhr does make some use of the term secularization. And he uses it, precisely, to describe this compromise of religion with the world which he criticizes so vehemently: "[...] the denominations [...] are the emblems [...] of the victory of the world over the church, of the secularization of Christianity, of the church's sanction of that divisiveness which the church's gospel condemns" (1929, p. 25; see also p. 113). Secularization, however, is not used as a central theoretical concept, but merely as a convenient equivalent to phrases like the "moral ineffectiveness of Christianity", the "defeat of Christian ethics", or the "ethical failure of the church". The problem Niebuhr wants to come to terms with is moral and theological, and most of the time he expresses it in these terms. The sociological approach is a mere tool at the service of this moral crusade - which might explain the subordinated role of the concept of secularization.

Secularization is also loosely used to refer to two related phenomena. The secularization of the church can lead, in turn, to the "secularization of the masses" (1929, p. 32): If a new religious movement proposing an ethical ideal fails to appear, the disinherited transfer their religious fervor to secular movements. This is what usually happens in modern times - the last truly religious revolution having been Methodism (1929, p. 72). Finally, secularization can also refer to the "secularization of the state" (1929, p. 123).

Pitirim Sorokin (1937-1966)
Unlike Niebuhr, Sorokin did not provide a full-fledged analysis of religion in the United States. But he developed a number of views which had a bearing on the problem of secularization. The views expressed in Sorokin's 1966 paper, on which I will base the present analysis, only made more explicit and available in a convenient form ideas which he had developed much earlier, most notably in his very lengthy Cultural and Social Dynamics (1937). In this work, Sorokin outlined a cyclical theory of social evolution in which three stages, the ideational¹⁴⁵, the idealistic¹⁴⁶, the sensate¹⁴⁷, then again the ideational, necessarily followed each other. In this framework, the 1966 article discusses the "trend of secularization", that is, the demise of religion as culture passed from the idealistic synthesis in the Middle Ages to the present sensate phase. This paper shows some influence of the secularization debate which has just erupted: The term secularization is used much more systematically than in the 1937 book, where it was used very sparsely, most of the time in quotation marks, and mainly to refer to the very circumscribed phenomenon of the decline in the proportion of works of art based on a religious theme (see for instance Sorokin 1937, p. 384).

Although it is far from fitting into the CISR paradigm, Sorokin's theory is very interesting for our purposes: It conveniently summarizes the view implicitly held by most sociologists in the 30s and 40s on the subject of secularization - and still held today by those who are not familiar with the CISR paradigm. In Sorokin's view, secularization is a very simple and straightforward process: it is an overall decline of religion, which started "at the end of the twelfth century A.D." (Sorokin 1966, p. 9). Linked with this simplistic view is an implicit definition of religion very close to the views held by conservative clerics: Religion is what separates man from animality. When he is no longer controlled by "deeply interiorized religious and ethical imperatives", man is quickly overcome by "biological urges, passions and lusts" (1966, p. 25).

The first cause of the decline of religion is the emergence of science, which has been carefully computed by Sorokin: 4 inventions in the VII^th century, 12 in the XII^th, 53 in the XIII^th, 429 in the XIV^th and 8527 in the XIX^th. The decline has been computed just as precisely in the other areas: rise of materialist philosophies (0% in the X^th century, 53% in the XVI^th), emergence of non-religious forms of art, and so forth. These trends resulted mainly in a loss of power of the Church on society: 1) religious dogmas lost their status as God-given revelations, and were turned into man-made rules of conduct; 2) the Church lost its power to define the contents of art; 3) Christian norms lost their influence on social life; 4) with the emergence of different confessions, the autonomization of the state, and the privatization of religion, the Church declined as a social institution.

But this trend toward a "negative religious and moral polarization" - which resulted in the atrocities of World War II - is, according to Sorokin, only one aspect of the contemporary situation. Counterbalancing this negative trend, there is a "positive religious and moral polarization", which - "despite a lack of statistical data" - manifests itself most notably in the fact that a growing number of people are in search of "the supreme reality and values" (1966, p. 32). Thus, we witness around us the first clues of the crumbling of the sensate culture, which will, of necessity, be followed again by an ideational phase.

The early Parsons (1937-1944)
As might be expected, Talcott Parsons strongly disagreed with Sorokin's theory. As a matter of fact, he explicitly refuted it in the volume which was published in 1963 in honor of Sorokin (Parsons 1963a). Although he did not provide any complete study on religion in the United States, Parsons developed a great interest in religion, and played a significant role in the development of sociology of religion. Parsons' influence on the new sub-discipline was many-sided. His first and most decisive contribution was indirect, and resulted from the fact that he played an important role in introducing the European classical tradition into American sociology and that he helped to institutionalize sociology itself through his personal commitment and writings. He also indirectly contributed to the development of sociology of religion by establishing functionalism as the official creed of American sociology. Finally, anticipating on what I will be saying later concerning the institutionalization of the discipline, it must be noted that, at the moment new journals in the sociology of religion were launched, Parsons lent them his authority by contributing to some of the early issues¹⁴⁸. But, as we will see, he also influenced this evolution through his own writings on religion.

In view of the tremendous impact of The Structure of Social Action (1937) on American sociology, the question of the treatment of secularization by Parsons in this book must be investigated carefully. After all, one could easily imagine that it was through this channel that Weber's ideas on secularization were introduced in America. As we will see, this is not the case.

The word "secularization" appears in two different places in the Structure: in the discussion of the positivist theory of action (1937, Vol. I), and in the discussion of the idealistic tradition (1937, Vol. II). The first use of the term refers to the secularization of social thought (sociologization*) in the seventeenth century in Europe (1937, p. 53; p. 61; p. 88; p. 94). Whereas "in Christian thought the necessity of the state and its coercive authority had been explained by the Fall and the consequent sinfulness of man", in secularized social thought, "the sinful element of human nature was brought into the framework of a concept of natural law" (1937, p. 88). Parsons uses this sense of secularization very systematically in the discussion of the positivist theory of action, to show that the problem of order emerged when social thought freed itself from religious assumptions: "The immediate practical animus of Hobbe's social thought lay in the defense of political authority on a secular basis" (1937, p. 94). Let us briefly note that in the Social System, Parsons will use this sense of secularization again, and speak of the necessity of "cognitive stabilization of ideological beliefs from other than religious sources" (Parsons 1951, p. 518).

The second sense in which Parsons uses secularization is derived from Weber: In the process which saw the emergence of the spirit of capitalism from the Protestant ethic, there was a "gradual dropping out of the religious background of the attitudes concerned and in place of the religious [the] substituing [of] a utilitarian motivation" (1937, p. 532). Parsons is careful to distinguish the process of secularization in Weber from "the mutual relations of prophecy, rationalization and traditionalization". These two aspects of social change are, Parsons tells us, "radically different". Secularization, then, is equated with the "emancipation from ethical control, the setting free of interests and impulses from normative limitations, traditional or rational ethic. [...] It is essentially the process involved in Pareto's process of transition from dominance of the residues of persistence to those of combinations, equally in Durkheim's transition from solidarity or integration to anomie" (1937, pp. 685-86). What Parsons has in mind here is entirely different from what he meant by secularization in the first sense. In his first discussion, secularization was the replacement of religious ideological justifications of the state by secular ideological justifications. It is a transition from a religiously based political order to a secularity based order. In the second use of the term, Parsons speaks of a transition from a religiously prescribed order to disorder.

But secularization is used in still a further sense by Parsons, when he speaks of the submission, in ancient Greece, of the "hereditary professional priestly class" to the political structure of the polis (1937, p. 613). Here, secularization no longer refers to a change in social thought or in the moral order, but in the balance of power between a class of priests and lay political authority.

The diversity of the meanings of the term (together with the fact that Parsons, usually so pedantically aware of the necessity of defining his terms, never formally introduces the term secularization) makes it clear that it is not linked to any "secularization theory", but is simply used as a descriptive term. As with most other writers in his time, it refers to any change - whatever its theoretical or historical implications or the level of analysis - from "religion" to "secularity".

However, even if the term itself is not linked to a theory, it is clear that "secularization theory" as it is understood today is very much present in Parsons' analysis, if only because it was very present in the writings of the authors he analyzes. "It is one of the most striking facts about the group of writers being considered in this study that with the exception of Marshall they all bring forward ideas involving a process of disintegration of the kind of social control associated with a faith" (1937, p. 288). But at this point, interestingly enough, Parsons does not speak of secularization. Such broad general use of the term in not to be found in the Structure.

The reasons for this failure to use what would appear to us to be the appropriate term are clear if we remember that the ultimate consequence which Parsons attached to secularization (as used in Vol. II) was anomie. Therefore, secularization could not be considered by Parsons a stable product of social change. This appears most clearly in his discussion of Pareto's process of the circulation of the elites. At one phase of the cycle, the instinct of combination residues come to predominate, inside of the governing elite, over the persistence of agregates residues. At this moment, society enters into an "age of skepticism" characterized by the absence of control by value elements over conduct. But this stage of the cycle, like the others, is intrinsically unstable: Logico-experimental science, which is associated with the age of skepticism, cannot produce really scientific theories (in social thought), but only pseudo-scientific ones, "for action rests on premises for which factual data are necessarily lacking. [...] The pseudo-scientific theories, those of 'skepticism', do not indicate a state of the sentiments which is stable any more than do those which surpass experience, hence another reaction back to the pole of 'faith'" (1937, pp. 286-87). To use a terminology which will be used by Parsons in his later formulations, secularization is intrinsically unstable because it is unable to fulfil at least two of the functional prerequisites essential to the existence of a social system: social control (integration), and the definition of ultimate ends (goal-attainment).

How were these premises laid down in Parsons' first major work developed later, in his other writings? In a 1944 paper, Parsons comments on the development of the sociology of religion. As other functional theorists around the same period, he criticizes the assumptions held by the evolutionary school concerning the fate of religion. He explains to the reader that, in the views developed by Tylor and Spencer, religious belief would "naturally disappear as an automatic consequence of the advance in scientific knowledge" (Parsons 1944, p. 199). He then goes on to show that the writings of Pareto, Malinowski, Durkheim, and Weber cast an entirely different light on the fate of religion in modern society. In short, Parsons contends, these writers have shown that empirical knowledge will never solve the problems of meaning, and that religion will thus never disappear (1944, p. 211).

But it is not, to my knowledge, before 1957 that Parsons explicitly tackled the problem of secularization. Before presenting this paper, we must turn to the first complete study of American religion written by a person who defined himself as a sociologist.

Will Herberg (1955)
The second very important analysis of American religion, Will Herberg's Protestant-Catholic-Jew (1955), appeared a quarter of a decade after Niebuhr's book. Although he taught as a sociologist, Herberg did not really have any training in sociology and, with his "casual statistics", he was "not a model for colleagues in that discipline"¹⁴⁹ (Martin E. Marty, in Herberg 1955, pp. x and viii). His book, however, had a tremendous influence on subsequent American sociology and, as we will see, many of the points of view which were developed in it were very influential.

The book is famous, first of all, for its discussion of the American "religious revival" which analysts had begun to become aware of at the end of the previous decade, and which Herberg was the first to give widespread publicity. As already mentioned, this "revival" - no matter how it is to be interpreted - played a very important role in reviving interest in the sociology of religion. Herberg's book was therefore inseparably a result, and a factor, in the revival of sociology of religion.

The religious revival trend, as presented by Herberg, can be summarized as follows: "In the quarter of a century between 1926 and 1950 the population of continental United States increased 28.6 per cent; membership of religious bodies increased 59.8 per cent; in other words, church membership grew more than twice as fast as population" (Herberg 1955, p. 47). Other indicators, like the growth in Sunday school enrollment, the expansion in church construction, the increasing prestige of religious leaders, the new intellectual prestige of religion as measured by the best-selling books list, pointed in the same direction (1955, pp. 50-53). Herberg's figures were contested, most notably by Seymour Lipset. A hot debate followed, Lipset and Herberg arguing against each other in papers bearing revealing titles: "Religion in America: What Religious Revival?" (Lipset 1959); "There is a Religious Revival!" (Herberg 1959). The question whether there indeed was a religious revival or not¹⁵⁰ need not concern us here; let it suffice to be noted that, with this debate, religion suddenly became a fashionable subject of inquiry, and that, as a result, in the following years, a number of empirical investigations were launched, most notably by Lenski (1961) and by Glock and Stark (1965).

Let us now turn to the interpretation of these trends, as attempted by Herberg. The starting point of Herberg's analysis is the American religious paradox¹⁵¹: Religion seems at once to be very important in America - more important than in any other modern country - and irrelevant to the functioning of society. "This is at least part of the picture presented by religion in contemporary America: Christians flocking to church, yet forgetting all about Christ when it comes to naming the most significant events in history; men and women valuing the Bible as revelation, purchasing and distributing it by the millions, yet apparently seldom reading it themselves. Every aspect of contemporary religious life reflects this paradox - pervasive secularism amid mounting religiosity, the strengthening of the religious structure in spite of increasing secularization¹⁵²" (Herberg 1955, p. 2). "Secularization" played a much more important role in Herberg's analysis than in any other previously published work. And, as we will see later on, the very definition of secularization used by Herberg gives us the clue to the resolution of the paradox.

Briefly stated, Herberg's thesis is that religion in America can only be understood if, by analyzing religion while bearing in ming a historical background - the process of assimilation of the succeeding waves of immigrants - we come to understand which role religion played, as well as which functions it fulfilled, and still fulfils. For the immigrant, religion represented continuity; it allowed him to define his identity, to answer to the question "what am I?" (1955, p. 12). For the first and the second generation the ethnic Churches provided a means of integration into American society. Things were somewhat different for the third generation, who were de facto integrated, and had no need to identify themselves with an ethnic Church. Increasingly, their allegiance shifted to one of the major "American faiths": Protestantism, Catholicism, or Judaism. Each of these religions worked as a melting pot: "The traditional 'single melting pot' idea must be abandoned, and a new conception, which we will term the 'triple melting pot' theory of American assimilation, will take its place" (Ruby Jo Kennedy, quoted in Herberg 1955, p. 33). Seen against this background, the American paradox can be resolved: Religion fulfils the important social function of "belonging" - which explains its importance - but need not be very important from a strictly religious point of view as defined by the man of faith.

The same phenomenon can also be viewed in terms of "operative" versus "conventional" religion. The operative religion is the set of rituals, images, and symbols that is necessary in any society to give it an overarching sense of unity, to provide individuals with a normative context. In most historical cases, this operative religion is simply "conventional religion", that is, religion as it is institutionally defined. In America, however, operative and conventional religion no longer coincide, and this characteristic is precisely what Herberg calls secularization: "We may indeed define a secularized culture as a culture in which the conventional religion is no longer the operative religion" (Herberg 1962, p. 148). In America, Herberg tells us, the operative religion is the American Way of Life: "Sociologically, anthropologically, if one pleases, it is the characteristic American religion, undergirding American life and overarching American society despite all indubitable differences of religion, section, culture, and class" (Herberg 1955, p. 77).

The relationship between culture and religion has thus been inverted: "Instead of the culture standing under the judgment of God to whom religion witnesses, religion (that is, conventional religion) tends to understand itself as an expression of, and a utility for, the culture" (Herberg 1962, p. 153). "The goals and values of life are apparently established autonomously, and religion is brought in to provide an enthusiastic mobilization of human resources for the promotion of the well-being of the individual and society" (Herberg 1955, p. 270). In this framework, it is the operative religion which prescribes what conventional religion must be. The operative religion prescribes that conventional religion, "no matter which (of the three faiths) it is", is a good and necessary thing. This religion, however, is "a way of sociability or 'belonging' rather than a way of reorienting life to God. It is thus frequently a religiousness without serious commitment, without real inner conviction, without genuine existential decision" (1955, p. 260). "The American Way of Life [can] best be understood as a kind of secularized Puritanism, a Puritanism without transcendence, without sense of sin or judgment" (1955, p. 81). Indeed, "under the influence of the American environment the historic Jewish and Christian Faiths have tended to become secularized in the sense of becoming integrated as parts within a larger whole defined by the American Way of Life" (1955, p. 82).

To conclude, how can we assess Herberg's contribution to the secularization debate? The answer is rather complex. On the one hand, in spite of the very wide use Herberg made of this term, and of the importance it took in his theoretical framework, we are still a far cry from the secularization theories of Berger and Wilson. In fact, in Herberg's hands, secularization remains largely a descriptive term, which functions at three different levels. The first refers to an evolution in society, in which the conventional religion ceases to function as the operative religion. The second is a movement inside conventional religion, in which this religion progressively becomes more similar to the operative religion. The third (which we have not discussed) is, in accordance with the traditional sect-church typology, a movement in conventional religion, in which a religious cult becomes a denomination - thereby moving from the margins to the center of society (Herberg 1962, p. 37). Obviously, the three processes are interrelated: Pushed further, the cult-denomination movement results in the transformation of the denomination in the "trifaith system" (1962, p. 39) - but at this point, we are no longer witnessing an evolution in religion itself, but at the global, societal level. Thus, secularization does not refer so much to historical process as to a movement, which can occur in a wide variety of circumstances, along a continuum between "conventional religion" and "operative religion", or "secularity".

But on the other hand, Herberg's definition of secularization as the process by which the operative religion separates itself from the conventional religion constitutes one of the most important building-blocks of the CISR paradigm. This definition makes it possible to see privatization* and generalization* as complementary processes: Conventional religion becomes irrelevant socially, or privatized*, but at the same time, religion becomes generalized* as the operative religion. Herberg's view thus gives us the key to a coherent understanding of the CISR paradigm: Although Wilson (decline of conventional religion) and Parsons (rise of operative religion), for instance, might appear to be arguing in different directions, Herberg's definition of secularization encompasses both views in a broader consideration of secularization as a process of reorganization of the forces of religion in modern society.
The later Parsons (1957)
Herberg's views on generalization* were further developed by Parsons. In a paper published in 1957 in French, Parsons explains that we must reconsider the problem of secularization. Developing the logical implications of his refusal to consider secularization a stable product (Parsons 1937) and of his positive assessment of the place of religion in modern society (Parsons 1944), he contends that this term cannot mean that religion has been eliminated, but only that its place in society has been redefined: "It is a 'secularization' [The quotation marks were probably added by the translator because this term was, at this time, very unfamiliar in France (see 1957, p. 22, n. 2)], not in the sense that it would aim at the elimination of all organized religion from social life, but rather in the sense of an effort toward a reconsideration of the place religion is entitled to hold in social life" (1957, p. 23). This formulation marks a very important break in the thought on secularization: For the first time, secularization was used in a positive sense by a sociologist. This reevaluation was achieved by placing the problem of religion in modern society in the framework of functional theory, more specifically, of structural differentiation: "The particular place religion occupies in American society is determined by a structural differentiation" (1957, p. 25). The key to this new understanding lay in a distinction between the "primary" and the "secondary" functions of religion. Examples of the primary function of religion, as given by Parsons in his article, revolve around the problems of meaning, while the secondary functions are social (social control, social integration, etc.; Parsons 1957, p. 26). Through structural differentiation, there is a progressive transfer of some of the secondary functions to other social institutions: politics (1957, p. 27), the intellectuals and the professions (1957, p. 28), and rights (1957, p. 30). Therefore, even though religion seems to have lost much in the process, this is not as dramatic as it may seem: "Has it lost everything? My own point of view is that, all considered, it still has much left" (1957, p. 31). This positive reassessment of secularization is fundamental in the genesis of the CISR paradigm: Together with the reflections of the theologians, it opened the way for a new view of secularization, in which religion would not necessarily be the loser.

Drawing on Weber, Parsons concluded his paper on a generalization* thesis very close to that developed by Herberg two years before: "The values of contemporary American society have roots which are fundamentally religious" (1957, p. 32). Parsons expressed his views on secularization as generalization* in several subsequent papers, most notably in his criticism of Sorokin's position on secularization (Parsons 1963). As this formulation is part of the CISR paradigm, I will defer discussion of it to the next working paper. For the present purposes, it was sufficient to note the shift in the interpretation of secularization by Parsons, as early as 1957, and to note again the link between functional theory and a more positive evaluation of the role of religion, and even of secularization.

Harold Pfautz (1955-1956)
Harold Pfautz's two papers on secularization - which appeared in AJS and in Social Forces - are not very important from a theoretical point of view, but rather because they provided one of the first empirical studies on a form of secularization. Pfautz's views on secularization are more a variation on the old sect/church typology than a move in the direction of the CISR paradigm. The development of his theoretical framework was motivated by the problems he encountered in his Ph.D dissertation on Christian Science to isolate the "sociological aspects" of secularization (Pfautz 1955, p. 121, n. 1).

Pfautz disposed his types in the following order of increasing secularization: cult, sect, institutionalized sect, church, denomination¹⁵³. This classification of religious movements represents an improvement over previous typologies in that it takes into account more dimensions. The five dimensions are: the demographic, the ecological (geographical diffusion and segregation), the associational (basis of recruitment, degree of differentiation and basis of solidarity: symbols or feelings), the structural (type of leadership), and the psychological (Weber's types of orientation to social action).

The characteristics of the cult - the "most primitive form of religious organization" - are mainly the confinement to the local community, the lack of internal differentiation, non-symbolic (affectual) forms of social interaction, isolation, and a low power and prestige relative to other groups in society (1955, p. 123). Christian Science, in its early days, had the status of a cult. The appearance of the sect is marked by an increase in size and by geographic expansion, by the appearance of an internal differentiation (charismatic leadership), and by an increase in symbolic elements (ideology; 1955, p. 124). It is not necessary to pursue the detailed presentation of the evolution of religious bodies as they move further along in the direction of increased secularization: Roughly speaking, the more they become secularized and the more they become differentiated, the more the symbolic elements - versus the affectual - play an important role and the more the prestige of the religous body in society grows. The sect can become either an institutionalized sect, a church, or a denomination. Examples of an institutionalized sect are the Christian Scientists or the Mormons (1955, p. 126). The prototypical example of church is the Catholic Church of the thirteenth century, while the denomination is best represented by the modern American Protestant religious bodies. The author concludes his article by briefly mentioning that "secularization is a general social process", and is as such also relevant to the study of other institutions, as well as to larger societies - as in Becker and Redfield, who are mentioned at this point - and that "the religious denomination is functionally related to the more inclusive form, namely, urban society" (1955, p. 128). But this view is not developed further.

In a later paper (Pfautz 1956), Pfautz simplified his analytical framework for the sake of an empirical study of Christian Science, retaining mainly the psychological dimension. He redefined secularization more simply as "the tendency of sectarian religious movements and groups to become both part of and like 'the world'". (It is interesting to note that this definition was apparently inspired by a discussion of worldliness* by Parsons, in which Parsons himself did not use the word secularization; Pfautz 1956, p. 246, n. 4, referring to Parsons n.d.¹⁵⁴, p. 27.) The paper focuses on the "social psychological aspect of this process", and is designed to test the idea that, in this process, the psychological structure of the group "change[s] from that of an emotional to a rational association" (1956, p. 246). To achieve this aim, Pfautz examines testimonies of healings as they appeared in Christian Science official publications from 1890 to 1950. The content analysis of these testimonies sought to distinguish mainly the "affectual motives" (e.g., benefits of a religious nature) on the one hand from the "traditional motives" (e.g., "my parents were Christian Scientists") and the "purposeful-rational motives" (e.g., benefits of a material nature)¹⁵⁵ on the other hand (1956, p. 248). The data indicated a decrease in the proportion of testimonies indicating an affectual motive (from 53% to 19%) and an increase in the proportion indicating a traditional motive (from 0% to 21%) as well as of those indicating a purposeful-rational motive (from 4% to 32%). The author thus concluded: "The data [...] clearly indicate that a process of secularization has taken place in Christian Science as far as its social psychological structure is concerned" (1956, p. 251).

Milton Yinger (1946-1967)
Yinger was probably the first lay American sociologist to offer a complete picture of sociology of religion¹⁵⁶. His contribution, unlike Herberg's or Lenski's, did not center on a single study, but rather comprised a great number of articles and books examining the different problems of sociology of religion from a single theoretical point of view. This point of view was functionalism. Yinger thus fit very neatly into the general trend of his time, in which, as we have seen, the rise of sociology of religion was closely related to the emergence of the functional view. Yinger himself stressed this point. In a chapter devoted to the influence of anthropology on sociology of religion, he asserted that the strong reliance of early sociologists of religion on evolutionary theory "may actually have slowed the development of a sociology of religion by reducing interest in the study of religion among students of contemporary society". This trend was reversed by the rise of functional theory: "Religion had been seen as an anachronism by many sociologists. Functionalism has helped to refocus interest on religion and relate it to the whole sociocultural complex" (Yinger [1958] 1963, p. 124). And, it must be added, Yinger himself was a very active advocate of the recognition of sociology of religion, as is clear from the opening statement of his 1957 Religion, Society and the Individual, which sounds like a manifesto: "This book is an expression of the belief that the student of society must be a student of religion" (Yinger 1957, p. xi).

What was the place of secularization in this framework? Yinger explicitly discusses two meanings of the term secularization. The first, which he calls the "usual dictionary meaning", is very close to differentiation*. The second, which is a use of which he disapproves, is very close to the worldliness* of denominations. Here, instead of secularization, Yinger proposes that we use the term "religious change". Yinger discusses these terms not because he is interested in the secularization issue; as a matter of fact, he does not seem to like it very much. Rather, he is attempting to show that some of the recent analysts of religion - most notably Will Herberg and Peter Berger - were wrong in considering that the increasing worldliness of American denominations represented a secularization. This trend, Yinger explains, is not new: "The churches of those who are comfortable in a society are almost always well accommodated to that society. This is not secularization. It is instead one of the most persistent functions of religion - whether we applaud or lament the fact" (Yinger 1963, p. 71).

This discussion is of particular interest to us since it represents one of the first attempts to clarify the terminology in the secularization issue. Let us turn again to the first of the meanings addressed by Yinger, that of secularization as differentiation*. According to Yinger, this is the "standard definition of the concept of secularization - the separation of religious motives, feelings, and decisions from other aspects of life". It implies that "many of life's decisions are carried out without reference to religion" (1963, pp. 69-70). Yinger does not develop this analysis further, and it is not really clear what he thinks with respect to this issue¹⁵⁷. The second meaning of the term - which he proposes to replace by "religious change" - he discusses in much more detail. He defines it as a "decline in orthodoxy or a reduction in the acceptance of traditional beliefs". Another way of expressing the same phenomenon is to say that "persons [are] acting religiously in a way that does not express directly the faith they profess", that "religious movements [...] lack 'authenticity'". (This is the way Herberg presented his "American paradox".) This apparent discrepancy occurs because people redefine their religion "while disguising or obscuring the process by holding, somewhat superficially, to many of the symbols of the earlier religious system. Religious change is usually a latent process, carried out beneath symbols of nonchange". In this process, the tension between religion and the world diminishes so much that it all but disappears: The church abandons its "'prophetic task' of reforming the world", and instead "supports American institutions and values" (1963, pp. 69-71).

Finally, it should be noted that one of Yinger's prevailing interests was in the study of religious pluralism. In connection with this, he brought out the relationship between pluralism and secularism (in the sense of differentiation*): "Pluralistic societies that do not have a substantial amount of secularism are generally unable to develop the mutual tolerance that religious diversity requires" (Yinger 1967, p. 27). In a pluralistic society, the traditional religious assertions usually lose some of their force, while the burden of social integration shifts to a new unifying system of belief - for instance Herberg's American Way of Life (Yinger [1961] 1963, p. 106). Thus generalization* is itself one of the consequences of pluralism.

Although he does not propose any global view on secularization, Yinger is very close to the CISR paradigm in two respects. First, he separates the idea of worldliness* of denominations from the more global process of secularization by renaming it "religious change" (this distinction will be institutionalized in the CISR paradigm by Karel Dobbelaere in his 1981 trend report [Dobbelaere 1981, pp. 22-26]). Secondly, he establishes a link between pluralism - one of the exemplars of the secularization paradigm - and differentiation*.
Charles Glock and Rodney Stark (1959-1965)
Glock and Stark's Religion and Society in Tension (1965) was another very important landmark in the development of the sociological study of religion. Indeed, it was a self-contained sociology of religion in itself, starting with a discussion of problems of definition and measurement, which were then applied to the empirical study of religion in America. It included one of the most important methodological proposals in the development of the discipline: the five-fold classification of the dimensions of religiosity in terms of the experiential, ideological, ritualistic, intellectual and consequential dimensions (Glock and Stark [1962] 1965). This methodological refinement, which substantially improved the measurability of religion, certainly played an important role in the recognition of the new sub-discipline in the heavily empirical environment of American sociology.

Glock and Stark repeatedly use the term secularization in this book. The most important discussion of secularization appears in the chapter discussing the "religious revival" in America. Some analysts, Glock and Stark explain, have noted a religious revival in contemporary America, while others, in contrast, think that the trend is one of increasing secularization ([1959] 1965, p. 68). The authors note that the proponents of the revival view rely very heavily on the ritualistic dimension: The indicators of revival used are essentially religious practice and affiliation. On this dimension - even though many technical problems cast doubt on the reliability of the statistics used - the authors conclude that "within the limits of the indicators used, some postwar growth in religiousness has occurred" ([1959] 1965, p. 78). But what about the opposite claim, concerning the increasing secularization? Authors presenting this thesis tend to emphasize the ideological and consequential dimensions ([1959] 1965, p. 82). Proponents of this view - who are "likely to be clergymen, church administrators, theologians or journalists" - either argue that there is a "watering down of belief", which manifests itself in the trend toward ecumenicity and the tendency to consider all religions equally good (ideological dimension, or decline in belief*) or that "religion is playing less and less of a role in family life, in education, in social welfare, and in economic life" (consequential dimension, or autonomization*). The authors conclude, without any critical elaboration, that "there is nothing in the literature that would constitute a serious and systematic defense of the secularization hypothesis" ([1959] 1965, pp. 82-83). The easy dismissal of the secularization thesis by Glock and Stark contrasts with their careful discussion of the revival thesis. This contrast can be at least partially understood is one considers that the revival thesis could be brought under the scrutinity of an empirical investigation, whereas the secularization thesis, as presented in their book, called for a discussion on a broader historical plane. But Glock and Stark are correct in the sense that, at the time they were writing (in 1959), no modern sociologist had presented such a historical analysis. And the dismissal of the other advocates of this thesis as mere theologians or journalists reflects in an interesting way the growing professional self-awareness among sociologists of religion.

In a later chapter, however, the authors reverse themselves: they now claim that there was indeed a trend of secularization¹⁵⁸ in the sense that "the mystical and supernatural elements of traditional Christianity have been replaced by a demythologized, ethical rather than theological, religion" (1965, p. 116). How can we account for this reversal? Between 1959 and 1965, Glock and Stark had been conducting several important empirical investigations in the framework of the Research Program in the Sociology of Religion, at the Religion Survey Research Center, Berkeley, and were thus able to substantiate empirically the hitherto unfounded claims that belief was declining: "In light of the data we have just examined, it seems clear that important changes of this kind [i.e., secularization] have indeed occurred in some American denominations" (1965, p. 116). Members of the most secularized denominations, Glock and Stark found, no longer believe in tenets such as the Virgin Birth or Biblical miracles. Glock and Stark's attitude toward secularization is thus rather ambiguous: their tendency to stick very closely to empirical findings and to refuse any generalizing theorization compels them always to waver, in their appreciation of secularization, between different conclusions, depending on the definition they seek to operationalize and on the particular questions they ask in their research.

To summarize, the situation at the beginning of the 60s was very different from that which had prevailed 20 years earlier. The sociological study of religion had become an empirical discipline, which increasingly drew the attention of lay sociologists. Concomitantly, the distinctiveness of Catholic sociology tended to disappear, bequeathing to lay sociology a professional association, which was put at the service of empirical investigation. In parallel fashion, the theoretical status of religion had been drastically altered: With the rise of functionalism, religion had become one of the institutions present in any society, and with the American religious revival, it had asserted itself as a force that must still be reckoned with. But although it appeared to be used slightly more frequently, "secularization" remained a descriptive term. The interest in this period was focused on revival, not on secularization; thus, secularization appeared as nothing but the negative counterpart of the revival thesis.

Let us now turn to Europe again. The old continent had witnessed no religious revival, and functionalism was not nearly as pervasive as in the United States. But the new climate originating in America rapidly crossed the Atlantic, and started putting Catholic sociology under pressure. This pressure, together with other factors, finally led to what, in a Kuhnian framework, must be called an intellectual revolution in the CISR, and to the end of the era of religious sociology.