reflection paper

CHAPTER NINE

Historicizing the Secularization Debate

An Agenda for Research

Philip S. Gorski

The trends are quite clear: In most parts of the West, Christian belief and practice have declined significantly, at least since World War II, and probably for much longer (e.g., Ashford and Timms 1992; Davie 1999). The variations are also quite clear: In a few countries, such as Ireland and Poland, levels of belief and practice are still very high; in others, however, such as Sweden and Denmark, they are quite low.

But what do these trends and variations mean? And how might we explain them? Current thinking on these questions among sociologists of religion is dominated by two opposing positions. The first is classical secularization theory, which sees the recent decline of Christian religiosity as part of a general trend toward greater “secularity” and an inevitable consequence of “modernization.” The second is the “religious economies model.” It argues that transhistorical and cross-national variations in “religious vitality” are caused by differences in the structure of “religious markets,” and, more specifically, that the freer religious markets are, the more vital religion will be.

Who is right? The diehard defenders of secularization theory? Or their upstart crit- ics from the religious economies school? In my view, the answer is “probably neither.” I say “neither” because there is now a great deal of evidence which speaks against both of these theories – against the view that modernization inevitably undermines religion and against the view that “free markets” (in religion) generally promote it – evidence, moreover, which seems better accounted for by other theoretical perspec- tives that have been forgotten or ignored in the recent debate. But I would add the qualification “probably,” because the accumulated evidence is still too thin historically and too narrow geographically to allow for any credible judgements: as sociologists of religion, we know a great deal about the twentieth-century West, but relatively little about anything else.

For those interested in advancing the current debate, then, two tasks would seem to be of especial importance. One is to revive and/or elaborate alternative theories of religious change. In what follows, I will discuss two perspectives that I regard as particu- larly promising: (a) a sociopolitical perspective, which focuses on conflict and competition between religious and nonreligious elites and movements; and (b) a religiocultural per- spective, which focuses on the relationship between religious and nonreligious values and worldviews, both within different religious traditions, and across different stages of religious development. The second task is to contextualize the postwar developments,

110

Historicizing the Secularization Debate 111

both historically and sociologically. This means studying the ebbs and flows of sec- ularity over the longue duree, and examining the interactions between religious and nonreligious actors and institutions.

THE RECEIVED ORTHODOXY: CLASSICAL SECULARIZATION THEORY

The roots of classical secularization theory can be traced back to the early nineteenth century and the writings of Henri Saint-Simon and Auguste Comte.1 Although their analyses differed somewhat in the details, both argued that human history passes through a series of distinct stages, in which the power and plausibility of traditional religion are gradually and irreversibly undermined by the growing influence of the state and of science (Saint-Simon 1969; Comte 1830–42/1969). In their view, modernity and religion don’t mix. This view was later echoed in the writings of sociology’s “found- ing fathers” – Marx, Durkheim, and Weber. While each viewed Christianity somewhat differently, all agreed that its significance was definitely on the wane. This became the dominant view within Anglo-American sociology as well. With the notable exception of Parsons (1963), postwar sociologists of religion all agreed that the public influence of religion was shrinking, and many thought that private belief itself was bound to decline or even disappear (e.g., Berger 1967; Luckmann 1963). During the 1960s, the “secularization thesis” was integrated into “modernization theory” and became one of its central axioms. As societies modernized, they became more complex, more ratio- nalized, more individualistic – and less religious. Or so the argument went. Today, of course, modernization theory has few adherents – except among sociologists of religion. While the rest of the discipline has moved on to other approaches, present-day defend- ers of secularization theory continue to use the old modernization-theoretic framework (e.g., Dobbelaere 1981; Wilson 1982; Bruce 1996), a framework that still bears strong resemblances to the classical theory of secularization propounded by Comte and Saint- Simon.

From the perspective of classical secularization theory (henceforth: CST), then, the decline in orthodox Christian beliefs and practices in most parts of the West is inter- preted as a part of a more general decline in the power of religious institutions and ideas and explained with reference to various social processes (e.g., differentiation, ra- tionalization, industrialization, and urbanization), which are loosely bundled together with the rubric of “modernization.” As social institutions become more differentiated and social life becomes more rationalized, the argument goes, religious institutions and beliefs lose their power and plausibility.

In support of these claims, defenders of secularization theory usually point to two well-documented developments. The first is the establishment and expansion of secular institutions in the fields of social provision, education, moral counseling, and other fields of activity once dominated by the church, a development they characterize as a “loss of social functions.” The second is the long-term decline in orthodox Christian practice and belief noticed by contemporary observers beginning in the late nineteenth century and subsequently confirmed in opinion polls throughout the postwar period. The fact that these declines have been especially pronounced among industrial workers

1 For a more detailed discussion of the development of secularization theory, see especially Tschannen 1992.

112 Philip S. Gorski

and educated city-dwellers – by some standards, the most “modernized” sectors of society – seemed to underscore the connection between secularity and modernity.

There are two main sets of objections one might raise against CST. One regards evidence and interpretation. As we have seen, secularization theorists view the recent downtrend in orthodox Christianity as part of a long-term decline in religiosity per se. However, it is not at all clear that the twentieth-century downtrend is really part of a long-term decline, and proponents of CST have not produced much hard evidence to suggest that it is. The usual way of “proving” this claim is to assault the reader with a barrage of twentieth-century evidence, and then confront them with a romanticized portrait of the Middle Ages, in which Christendom is all-encompassing, and all are devout Christians – a portrait that is no longer credible.2 Unless and until better ev- idence is forthcoming, the hypothesis of long-term decline must remain just that – a hypothesis. And even if such evidence were forthcoming, it still would not suffice to prove the broader claim that religion per se is in decline. After all, the simple fact that orthodox Christianity has lost ground does not necessarily imply that religion itself is on the wane. For example, it could be that Christianity is in a transitional phase, similar to the one that occurred during the Reformation era. Or, it could be that other religions will eventually take its place, in much the same way that Christianity supplanted “pa- ganism” in late Antiquity. Or it could be that the very nature of religiosity is changing, as it did in the Axial Age transitions that occurred in many parts of the world roughly two millennia ago. And even if religion per se is really on the wane of late, there is no reason to assume that the decline is permanent or irreversible. The history of religion is rife with ebbs and flows, and Christianity is no exception to this rule. Maybe the recent decline is really just a cyclical downturn of sorts. To make a strong case for long-term decline, then, secularization theorists would need to extend their analysis back beyond the modern era, something they have not yet done.

This brings us to the second set of objections. They concern the theory itself and, more specifically, the claim that the recent downtrend in Christian devotion can be traced to the effects of “modernization.” If this claim were correct, then we would ex- pect to find a strong, inverse relationship between the various dimensions of modern- ization (e.g., industrialization, urbanization, differentiation, and rationalization) and various indicators of secularization (e.g., levels of religious belief and participation). In other words, we would expect to find strong correlations between modernization and secularization across both time and space. As we have seen, there is some evidence that seems to support this claim. When we begin to compare different countries, however, the picture becomes more complex – and less clear-cut (on the following, see espe- cially Höllinger 1996). Take Scandinavia and the Benelux nations, for example. Despite their late industrialization and sparse population, and the existence of unified state churches, the Scandinavian countries, have long been, and still remain, the least de- vout and observant countries in Western Europe. In Belgium and the Netherlands, by contrast, where urbanization and industrialization began much earlier, and a higher degree of church-state separation prevails, orthodox Christianity is relatively stronger. Nor are these the only anomalies of this sort. Why, one might ask, are the Italians more observant than the Spanish? And why are Americans generally more observant than

2 For a typical example of this rhetorical procedure, see Bruce 1996. The classic critiques of this romanticized view of the Middle Ages are Delumeau (1977) and Thomas (1971).

Historicizing the Secularization Debate 113

Europeans? It is not at all clear that these differences in religious observance can be traced to differences in modernization.

There is also another anomaly that is worth noting: the difference in Protestant and Catholic rates of observance. Based on the classical theory, we might expect “supernat- ural” forms of religious faith such as Catholicism and fundamentalist Protestantism, to decline more quickly than more “rational” types of religiosity, such as liberal Protes- tantism. But in fact the very opposite appears to be the case. Throughout the West, Catholics are more observant than Protestants, and fundamentalist and evangelical Protestants are more observant than their “liberal” and “mainline” coreligionists. Thus, there are important variations – cross-national and interdenominational variations – which do not readily conform to the expectations of secularization theory. It is pre- cisely these variations that the next theory – the religious economies model – claims to explain, and it is to that theory that I now turn.

PRETENDERS TO THE THRONE: THE RELIGIOUS ECONOMIES MODEL

Why do levels of religious belief and practice vary so much from one country to the next? As we have just seen, classical secularization theory does not provide a com- plete or satisfying answer to this question. It is this deficit that the religious economies model (REM) seeks to address. Drawing on neoclassical economics, proponents of the REM argue that “religious vitality” is positively related to “religious competition” and negatively related to “religious regulation.” More specifically, they argue that where “religious markets” are dominated by a small number of large “firms” (i.e., churches) or heavily “regulated” by the state, the result will be lethargic (religious) “firms,” shoddy (religious) “products,” and low levels of (religious) “consumption” – in a word: Religious stagnation. By contrast, where many firms compete in an open market without govern- ment interference, individual firms will have to behave entrepreneurially, the “quality” and “selection” of religious products will be higher, and individual consumers will be more likely to find a religion which is to their liking and standards. If there are variations in the level of “religious vitality,” they conclude, these are due not to “secularization” but to changes in the “religious economy.”

Since the late 1980s, proponents of the religious economies model have produced a steady stream of books and articles that appear to confirm the theory (e.g., Finke and Stark 1988; Stark and Iannaccone 1994; Finke, Guest, and Stark 1996; Finke and Stark, Chapter 8, this volume; for an exhaustive bibliography, see Chaves and Gorski 2001). Most of them have focused on the effects of religious competition, rather than reli- gious regulation. The most pertinent of these studies examine the relationship between “religious pluralism” (operationalized in terms of the Herfindahl Index, a standard mea- sure of market concentration) and “religious vitality” (operationalized in terms of re- ligious belief, church membership, or church attendance) (e.g., Finke and Stark 1988, 1989; Finke 1992; Stark et al. 1995; Finke et al. 1996; Hamberg and Pettersson 1994, 1997; Johnson 1995; Pettersson and Hamberg 1997). These studies generally find a positive relationship between religious pluralism and religious vitality.3 Based on these

3 There is also a second and smaller group of studies that examines the relationship between the relative size of a particular religion – its “market share” – and its internal “vitality” (e.g., Stark and McCann 1993). These studies show that minority religions receive more support from their

114 Philip S. Gorski

findings, the leading proponents of the REM claim to have disproven the secularization thesis and argue that the term “secularization” should be “dropped from all theoretical discourse” (Stark and Iannaccone 1994: 231). Is their claim justified?

The work of the religious economies school has been challenged on a number of different fronts. Some scholars accepted the empirical findings, but questioned their theoretical significance (e.g., Lechner 1991; Yamane 1997; Gorski 2000). They pointed out that secularization theory is a theory, not only of individual behavior, but also, and indeed primarily, of social-structural change. In their view, secularization refers first and foremost to an increasing differentiation between the religious and nonreligious spheres of life, and only secondarily to its effects on individual behavior. Since the REM focuses exclusively on individual behavior, they argue, it does not really address the core claim of secularization theory and speaks only to those versions of the theory that postulate a direct connection between increasing (social-level) differentiation and decreasing (individual-level) religiosity.

Other scholars have challenged the reliability and validity of the findings them- selves (e.g., Blau et al. 1992; Breault 1989a, 1989b; Olson 1998, 1999). Using new datasets of their own, or reanalyzing REM data, these scholars often obtained null or negative correlations between pluralism and vitality. Defenders of the REM then challenged these results on methodological grounds (Finke and Stark 1988; Finke et al. 1996). The ensuing debate was long and complex, but the key issue was Catholics. Many of the analyses that had yielded a positive correlation between religious pluralism and religious vitality also included a statistical control for “percent Catholic.” Advocates of the REM defended this procedure on the grounds that the Catholic Church displayed a high degree of “internal pluralism,” and that treating it as a single denomination would therefore distort the findings. Critics pointed out that the positive relationship between pluralism and vitality usually disappeared or became negative when the control was removed (see especially Olson 1999). More important, they showed that removing any group with the characteristics of the (American) Catholic population – large in overall size but varied in local presence – would automatically result in a positive finding, and for purely arithmetic reasons!

On the whole, then, the REM’s claims to have disproven the secularization thesis and laid the foundations for a “new paradigm” in the sociology of religion are somewhat overblown. As we have seen, the central findings of the REM do not really address the core concerns of secularization theory, and are themselves open to dispute. Indeed, in a recent survey of the literature on “religious pluralism” and “religious participation,” Chaves and Gorski (2001) found that the balance of evidence actually tips against the REM, once we exclude analyses that employ inappropriate measures of competition or statistical controls for percent Catholic.

Of course, it is possible that further research could tip the balance back the other way. But this seems unlikely to me. For even a cursory review of the comparative and his- torical evidence reveals two large and potentially troubling anomalies for the REM. The first regards Catholic-Protestant differences. As we saw earlier, overall levels of Christian practice in various Western countries are closely related to the proportion of the

members than majority religions, a finding that also has been replicated by scholars working outside the religious economies perspective (e.g., Zaleski and Zech 1995; Johnson 1995; Perl and Olson 2000; but see also Phillips 1998).

Historicizing the Secularization Debate 115

populace that is Catholic: The highest levels of religious participation are to be found in homogeneously Catholic countries (e.g., Ireland, Poland, Italy, Austria), while the lowest levels are in homogeneously Protestant countries (e.g., the Scandinavian lands), with confessionally mixed countries (e.g., Germany, the Netherlands and Britain) gen- erally falling somewhere in between. This state of affairs is very much at odds with the competition thesis – the thesis that greater competition is always correlated with greater vitality – and, indeed, statistical analysis suggests that this thesis cannot be sustained for Western Europe as a whole (Chaves and McCann 1992).

Now, it could be that the relationship between Catholicism and vitality is actually spurious, and that the actual cause of the observed variations is religious regulation. In other words, one could argue that the differences between Catholic, mixed, and Protes- tant countries are really due to differences in the level of state control over the church. For it is true that the Catholic Church often has more institutional and financial au- tonomy than its Protestant rivals, and it is also true that the Protestant Churches in the confessionally mixed countries of North Atlantic Europe are more autonomous than their Protestant brethren in the Scandinavian countries. And, in fact, this hypothesis – that religious regulation is negatively related to religious vitality – has withstood statistical scrutiny (Chaves and McCann 1992). Unfortunately, it is not clear that the regulation hypothesis can withstand historical scrutiny. If the regulation hypothesis were correct, then one would expect that the historical declines in religious vitality that began during the late nineteenth century and accelerated during the 1960s would have been preceded by increases in religious regulation. But this does not appear to have been the case. In most countries, levels of religious regulation actually declined during this period. What is more, there is some evidence that suggests that these de- clines in regulation were actually preceded by declines in vitality. Thus, both the sign and the direction of the relationship between regulation and vitality appear to have been the opposite of those predicted by the REM (see Gorski and Wilson 1998; Bruce 1999). Why?

I now turn to a third approach that suggests some possible explanations for these anomalies.

A THIRD APPROACH: THE SOCIOPOLITICAL CONFLICT MODEL

Different as they may be in most other respects, there is at least one important similar- ity between classical secularization theory and the religious economies model: Neither pays much attention to politics. For classical secularization theorists, of course, politics plays no role whatsoever: “Religious decline” is the product of deep-rooted, socio- economic changes, such as urbanization and industrialization. As for the supply-side model, politics do enter in to some degree, but only as an exogenous and secondary factor, that is, as state “regulation” of the “religious economy.” There are other schol- ars, however, for whom politics has loomed larger and been more central, in both the explanans and the explanandum. They see sociopolitical conflict as the master variable in the secularization process, and changes in church-state relations as a key part of the outcome. But these scholars have played little role in the recent debate over sec- ularization, perhaps because most of them are historians. This is unfortunate, since their work speaks directly to the problems at hand, and may help to resolve some of the anomalies generated by classical secularization theory and the religious economies

116 Philip S. Gorski

model. I will refer to their approach as the sociopolitical conflict model (henceforth: SPCM).4

In the English-speaking world, the best known and most cogent proponents of the SPCM are probably David Martin (in sociology) and Hugh McLeod (in history) (Martin 1978; McLeod 1995, 1996; see also Höllinger 1996). On first reading, their views may seem very similar to those of Stark et al., insofar as they stress the effects of “competi- tion” and “pluralism.” And, in fact, members of the religious economies school often cite proponents of the sociopolitical conflict model in support of their own positions. On closer inspection, however, the resemblance between the two models proves to be superficial, for when Martin and McLeod speak of “competition,” they mean competi- tion not only between different churches, as in the REM, but also competition between different worldviews, both religious and secular. In particular, they argue that Protes- tant, Catholic, and Jewish religious communities were competing, not just with one an- other but also with “political religions,” such as socialism, liberalism, nationalism, and, later, fascism. Similarly, when McLeod and Martin discuss the effects of “pluralism,” they understand them in political rather than (quasi-)economic terms. Their central line of argument could be summarized as follows: In situations of religious monopoly, church and state will tend to become closely identified with one another, and social protest and partisan opposition will tend to evolve in an anticlerical or anti-Christian direction; a high level of religious disengagement is the result. In situations of religious pluralism, by contrast, in which some churches and church leaders are institutionally and politically independent of the state and the ruling elite, opposition to the existing regime did not automatically translate into opposition to the religion per se, and could even be expressed in religious terms; here, the degree of religious disengagement is likely to be lower.

The advantage of this approach can be seen in its ability to account for one of the major anomalies generated by the religious economies approach, namely, the paradox- ical combination of decreasing “vitality” with increasing “pluralism” and decreasing “regulation,” which can be observed in many parts of the West beginning in the late nineteenth century.5 From the perspective of the SPCM, the decrease in “vitality” – in orthodox belief, belonging and participation – was the result of competition, but the competition came, not from other churches, but rather, from nonreligious movements, which offered many goods previously monopolized by the church: Comprehensive worldviews, a social safety net, and communal and associational life. One of the things that these movements often fought for was a loosening of ties between church and state – that is, a decrease in religious “regulation.” In this, they were sometimes aided and abetted by “sectarian” religious movements, who bridled at the privileges of state churches. To the degree that they were successful, these campaigns against religious

4 It should be emphasized at the outset that the “sociopolitical conflict model” is not a model in quite the same sense or the same degree as secularization theory or the religious economies approach, since it is not rooted in a general theory of social change (e.g., “modernization theory”) or human behavior (e.g., “neoclassical economics”) and is not associated with a par- ticular “school” or discipline. Rather, it is an interpretive framework that has emerged out of the historical researches of a loose-knit group of scholars.

5 Interestingly, there is now some research that suggests that the recent increase in religious nonaffiliation in the United States may be partly a reaction to the close ties between Christian fundamentalists and conservative Republicans. On this, see Hout and Fischer (2002).

Historicizing the Secularization Debate 117

regulation created a situation more conducive to the growth of religious pluralism, that is, to the emergence and growth of alternative religions, and thereby reinforced and ex- panded the constituency which supported decreased regulation. From the perspective of the SPCM, then, the combination of decreasing vitality, increasing pluralism, and decreasing regulation is not paradoxical, and the fact that decreasing vitality preceded decreased regulation and increased pluralism is no longer anomalous.

But what about the second anomaly facing the REM, namely, the greater religious vitality of contemporary Catholicism? To my knowledge, this problem has not been explicitly addressed by advocates of the SPCM. But the SPCM does suggest a possible answer: One might hypothesize that varying levels of religious vitality are bound up with varying responses to the secularist movement. In most places, Catholics responded to the socialist and liberal “threat” by building social milieux and political parties of their own. The result of these efforts was Christian Democracy, a movement that re- mains powerful even today in many parts of Europe (e.g., Hanley 1994; Becker et al. 1990). Similar responses can be seen in some Protestant countries, such as Norway and the Netherlands (Scholten 1969). But the resulting movements and parties may not have been as broad (socially and geographically) or as deep (organizationally and polit- ically) as their Catholic counterparts, perhaps because the Protestant Churches lacked a centralized leadership structure capable of coordinating the various movements, or perhaps because the Protestant churches were more (financially) beholden to, and thus less (politically) autonomous from, the state. But these are no more than tentative hypotheses. Historians have only begun the task of identifying and explaining these cross-national differences, and have not yet brought quantitative data or comparative methods to bear in any systematic way. Clearly, this is one area in which historical sociology and the sociology of religion could contribute to the study of secularization.

One also might extend this general line of argument to explain intraconfessional variations in religious vitality, that is for the varying levels of religious vitality that we observe within the Catholic and Protestant blocs, between Italy (high) and France (low), for instance, or Norway (low) and Sweden (very low). One could hypothesize that these variations in religious vitality were because of variations in the relative suc- cess of the Christian Democratic movement and its various Protestant analogues, and one might attempt to explain these latter differences with the standard tools of social movement theory (i.e., “resources,” “political opportunity,” “frames”). Here is another area in which sociologists – especially political sociologists – might be able to add to the debate.

The SPCM is also superior to its rivals in another respect: It provides a concrete ex- planation for macro-societal secularization, that is, for the diminution of religious au- thority within particular institutions or sectors of society. Proponents of the REM have either ignored this second, macro-societal dimension, or defined it away, by insisting – quite wrongly! – that secularization refers only to a decline in individual religiosity. This cannot be said of the classical secularization theorists or their present-day defenders, of course, for whom the sharpening of boundaries between religious and nonreligious roles and institutions, and the declining scope of religious authority within various sectors of society has always been a – even the – key aspect of secularization. But they have tended to explain macro-societal secularization in a vague and often tautological fashion, as the result of other macro-societal trends, such as “modernization,” “differen- tiation,” and “rationalization,” which are closely related to secularization. By contrast,

118 Philip S. Gorski

the SPCM suggests a much more concrete and clear-cut explanation of macro-societal secularization, an approach that focuses on battles between religious and secularist movements for control of particular institutions and sectors, such as schools and edu- cation, or marriage and moral counseling. Indeed, scholars working within this tradi- tion have already produced case studies of societal secularization for specific countries and contexts (for overviews and references, see Bauberot 1994). What they have not produced, at least not yet, are systematic typologies and comparisons, which would al- low one to classify and explain the forms and degrees of macro-societal secularization across various countries and contexts. This, too, is an area in which sociologists might be able to contribute.

Unlike its rivals, then, the SPCM suggests clear and plausible answers for one of the key questions that confronts contemporary sociologists and historians of religion, namely: What explains the recent historical trends and cross-national variations in both Christian religious practice and macro-societal secularization? There are at least two other sets of questions, though, which the SPCM does not answer – or even begin to address. We have already encountered the first. It concerns the theoretical interpre- tation of the historical trends, whether they point to decline, downturn, transition, or transformation. These are not the kinds of questions that are susceptible to a definitive answer; the social sciences are often poor at predicting the future. But it would be pos- sible to shed some light on them, by situating the present more firmly within the past. Thus, one of the key tasks for future research will be to put what we know about the modern trends into historical perspective. The first step in this process would be to trace out the ups and downs – for ups and downs – in Christian practice and ecclesiastical authority as far back as the historical literatures and sources allow. In the case of eccle- siastical authority, this should not be a difficult task. The institutional history of the Western Church and its involvement in politics, education, charity, art, the family and other fields are well documented and well studied. Tracking the level of Christian belief and practice across time would be a more difficult undertaking, but not an impossible one. Early modern and medieval historians have unearthed a great deal of evidence on the religious practices of the premodern populace, some of it quantitative in form. By mining local and regional studies, and combining them with modern sources, such as census data and survey research, it should be possible to piece together some sort of picture of religious participation for various parts of Europe perhaps as far back as the late Middle Ages.

The next step in the process would be to put the patterns themselves in context – to figure out what they tell us about changes in religiosity per se. In this regard, it is important to bear in mind that variations in religious participation are not necessarily the result of variations in individual religiosity. They also can be – and sometimes are – caused by social factors such as the geographical proximity of religious services (a se- rious problem during the Middle Ages) or laws requiring regular church attendance (a common provision in the Reformation era) or influenced by the presence (or absence) of nonreligious incentives, such as access to church schools or eligibility for religious charity. Variations in religious participation also may reflect changes in the quality of collective religiosity rather than the quantity of individual religiosity. Caeterus parabus, a religion that sees ritual life and priestly intervention as a sine qua non of individual salva- tion (e.g., Catholicism) is likely to generate higher levels of religious participation than one which sees individual salvation as the result of individual faith (e.g., Lutheranism)

Historicizing the Secularization Debate 119

or predestination (e.g., Calvinism). Thus, it could be that the observed variations in religious participation are due less to changing levels of individual religiosity than to changes in the character and context of religious belief.

This brings us to the second and deeper problem which confronts the SPCM: the roots of the sociopolitical conflicts themselves. The SPCM treats these conflicts as a given and focuses on their dynamics and effects. But it says nothing about their under- lying causes, about the social and cultural conditions of possibility for the emergence of political religions and secular ideologies. From the vantage point of the present, this development has a certain self-evidence. But it is important to bear in mind that in many and perhaps even most times and places, sociopolitical opposition was expressed through religion rather than against it. This was particularly true in late medieval and early modern Europe, where biblical doctrine was the lingua franca of upstarts and mal- contents of all stripes from the Hussite Rebellion through the Revolution of 1525 to the English Civil War. In modern Europe, however, revolutionaries learned to speak other languages as well, languages such as nationalism and socialism, which were un- or even antireligious. What is more, large numbers of people were willing to listen to them. But where did these languages come from? And why did they resonate so widely? These are important questions for which the SPCM has no answers. To address these issues, we need another set of conceptual tools.

A FOURTH APPROACH: NOTES TOWARD A SOCIOCULTURAL TRANSFORMATION MODEL

Classical sociological theory suggests two possible approaches to the preceding ques- tions. The first is inspired by Durkheim’s writings on the division of labor (Durkheim 1893/1997) and the sociology of religion (Durkheim 1912/1976). For most of the last two millennia, one could argue, intellectual labor in Western societies has been monop- olized by the priestly classes. Since the Renaissance, however, the number of nonpriestly intellectuals has grown steadily, and various groups of experts and professionals have taken shape (e.g., jurists, bureaucrats, scientists, and psychologists). In order to estab- lish their jurisdiction over areas of knowledge and practice previously controlled by members of the priestly classes, they have had to draw sharp lines between religious and nonreligious domains and institutions. The result of this development has been the gradual removal of religious language and authority from an ever-expanding swath of social life, and the articulation of nonreligious sources of moral valuation (on this, see especially Taylor 1989).

The second approach derives directly from Weber’s sociology of religion and, more specifically, from his essay on “Religious Rejections of the World” (cf. Weber 1919/1946). In traditional societies, argues Weber, religion and “the world” were of a piece. The divine, however conceived, resided within the world, and “salvation” con- sisted of worldly well-being (i.e., health, wealth, and progeny). With the emergence of “world-rejecting religions” in South Asia and the Middle East roughly two millennia ago, this original unity of religion and world was broken asunder, and individual salva- tion and the divine were catapulted into another realm, a transcendental beyond. The implications of this transformation are difficult to overstate. Wherever it took place – in India and China, Persia and Palestine, Rome and Mecca – religious and nonreligious values and activities now existed in a state of tension with one another. The demands

120 Philip S. Gorski

of the divine were not easily reconciled with the realities of the world: blood-kin ver- sus coreligionists, the Sermon on the Mount versus raison d’etat, brotherly love versus the profit motive, revelation versus reason – these are some of the stations along the westward branch of the road that Weber wishes to describe. It is not a straight path, but a spiralling one, in which the ongoing conflict between the religious and the non- religious leads not only to ever sharper institutional boundaries between the various “life orders” but also to greater and greater theoretical consistency within the individ- ual “value-spheres” (political, economic, aesthetic, erotic, scientific). The consequence, says Weber, is an ever growing differentiation between the religious and the nonre- ligious, both institutionally and intellectually, a tendency that, for various reasons, Weber believes has gone further (so far) in the West than in other parts of the world.

These two approaches are not necessarily at odds with one another. In fact, they might even be seen as complementary. For each addresses a question which the other leaves unanswered. The neo-Weberian approach explains why religious and nonreli- gious spheres of knowledge came to be separate, something that the neo-Durkheimian approach takes for granted. For its part, the neo-Durkheimian approach identifies the actors who drew the boundaries, something that Weber (uncharacteristically) omits from his analysis. Nor are these approaches at odds with the SPCM. On the contrary, they might deepen our understanding of the “secular revolution” of the late nineteenth century.

CONCLUSION: AN AGENDA FOR RESEARCH

I have pursued two aims in this chapter, one critical, the other constructive. On the critical side, I have tried to identify the empirical and theoretical shortcomings of the two perspectives that have dominated recent discussions of secularization: Classical secularization theory (CST) and the religious economies model (REM). One problem that is common to both, I have argued, is that they are insufficiently historical, albeit in somewhat different ways. The problem with CST, historically seen, is that it is premised on a truncated and romanticized version of Western religious development: Truncated, insofar as it tends to juxtapose the modern era to the Middle Ages and ignore the intervening centuries; and romanticized insofar as it adopts a rose-tinted picture of the Middle Ages as a period of universal belief and deep piety, a picture that is very much at odds with contemporary historiography. As I have argued elsewhere (Gorski 2000), once the Reformation era is inserted back into the narrative, and a more realistic view of the Middle Ages is adopted, the story line of Western religious development becomes more complicated, and the classical tale of an uninterrupted decline in religious life beginning in the Middle Ages becomes very difficult to sustain. For what we see is not simply (quantitative) decline, but (quantitative) revival (in ecclesiastical influence) and (qualitative) transformation (in individual religiosity) – a multidimensional ebb and flow.

The problem with the REM rests on a somewhat different but equally flawed picture of Western religious history, a picture that is at once foreshortened and anachronistic: foreshortened in that it focuses almost exclusively on the nineteenth and twentieth cen- turies, thereby ignoring the medieval as well as the early modern period, and anachro- nistic in that it tends to see earlier historical periods through a twentieth-century lens. This leads to some rather egregious errors of interpretation. Consider the claim that low

Historicizing the Secularization Debate 121

levels of church membership in colonial New England indicate a low level of “religious vitality.” This ignores the rigorous standards for church membership then in force, and the large numbers of “hearers” who filled colonial pews. Or consider the claim that widespread “superstition” among medieval parishioners indicates a state of religious stagnation. This emphasis on knowledge and belief ignores the ritual and communal dimension of religious life in the Middle Ages (Gorski 2000). Once we correct for errors of this sort, the antisecularization story that underlies the REM – a story of ever increas- ing religiosity since the Middle Ages – becomes just as hard to defend as its classical rival. In my view, then, both CST and the REM are based on implausible narratives of Western religious development.

This brings me to the constructive aspect of the chapter, which is the attempt to outline some possible alternatives to CST and the REM, which I have dubbed the so- ciopolitical conflict model (SPCM) and the sociocultural transformation model (SCTM), and to suggest some possible directions for future research. In their present forms, both of these models are open to some of the criticisms I have leveled against CST and the REM. For example, the SPCM in its current form might be accused of a foreshortened historical perspective. With the exception of David Martin (1978), researchers working within the framework of the SPCM have focused mainly on the late nineteenth and early twentieth centuries. This is unfortunate, because there is good reason to believe that a more generalized version of the SPCM could be used to analyze other episodes of secularization, such as the privatization of religion that occurred in the wake of the Thirty Years’ War (Kosselleck 1988) or the process of disaffiliation (Entkirchlichung) that followed the upheavals of the 1960s (Hout and Fischer 2002). In both of these instances, religious ideas and institutions suddenly found themselves confronted with ir- or antireligious world pictures and social movements. And it seems likely that a more serious engagement with the historical record might turn up other episodes of structural or cultural secularization.

For its part, the SCTM (à la Weber) might be accused of a truncated historical per- spective, insofar as it focuses mainly on the beginning (antiquity) and end (modernity) of the secularization story, with little attention to anything in between. This is also unfortunate, because Weber’s analysis of the growing tensions between the religious and nonreligious “value-spheres” contains allusions to numerous episodes of conflict between priestly and nonpriestly intellectuals and their respective supporters (conflicts over religious mission and raison d’état, Christian charity and capitalist imperatives, sexual morality and erotic experience, and so on), which could be analyzed for their contribution to the secularization process, using the conceptual tools that have been developed for the study of “boundary-formation” in science studies and other subfields of sociology (Gieryn 1999; Lamont and Fournier 1992).

Despite these narrative gaps, the SPCM and the SCTM, in my view, are still more historicized, and indeed more sociological, than their predecessors and rivals, CST and the REM. For unlike CST, the SPCM treats secularization as a historically variable and contingent outcome, rather than as a universal and inevitable developmental trend, thereby leaving open the possibility that secularization is an episodic, uneven and perhaps even reversible process. CST, by contrast, is still framed by a high modernist meta-narrative that sees religion and tradition as inherently opposed to science and progress in a way that even many modern-day progressives and scientists would now find hard to swallow. And unlike the REM, the SCTM treats religion as something that

122 Philip S. Gorski

varies not only in its quantity but also in its quality, thereby avoiding the anachro- nisms that often plague the REM (e.g., equating seventeenth- and twentieth-century church membership as operational equivalents that “mean” the same thing). Proper interpretation of quantitative variation requires greater sensitivity toward contextual – and sociological – nuance. And proper analysis of secularization processes requires greater attention toward macro-societal transformations.

In closing, let me sum up what I mean by “historicization” and, thus, what I think would be involved in “historicizing the secularization debate”: (a) adopting a longer- range (and fully encompassing) historical perspective that extends well beyond the modern era; (b) engaging in a more serious and sustained way with the relevant histor- ical sources and literatures, so as to develop a clear sense of the temporal and spatial contours of secularization in all its dimensions; (c) viewing secularization as a con- tingent outcome of particular events involving particular actors; and (d) being more sensitive to changes in the context and content of religious practice and belief.

I do not think historicization is a panacea, nor do I wish to denigrate nonhistorical strategies of research. But I do think that the literature on secularization could stand a dose of history, and that greater attention to the past might shed new light on the present. Only by contextualizing the recent episodes of secularization will we be able to assess their larger significance.