Secularization Theory’s Differentiation Problem: Revisiting the Historical Relationship between Differentiation and Religion

Abstract

Much theorizing about secularization tells a “differentiation story” that puts a historical process of structural differentiation at the center of its understanding of secularization. The heart of the story is the claim that the increasing differentiation of social spheres over time freed the “secular” spheres of life (politics, economics, etc.) from religious control or domination. This conceptual framing has been widely shared by scholars in the field, not only by adherents of the classical secularization paradigm, but also their leading critics in the supply-side and historicist–revisionist schools. While the story sometimes serves a purely descriptive function, at other times it is used to explain secularization (i.e., differentiation causes secularization). A close examination of the differentiation story, however, raises questions about the historical accuracy and theoretical plausibility of some of its core assumptions. Aspects of the differentiation story that require critical reconsideration include the empirical accuracy of its historical generalizations, its underspecified notion of “spheres,” and its explanatory assumption that some spheres are innately or properly nonreligious.

1. Introduction

For half a century and more, social scientists have debated the nature, extent, and causes of the secularization of Western societies (Gorski and Altınordu 2008; Swatos and Christiano 1999). Classical secularization theory, which since the beginnings of social science posited a general pattern of declining religious influence driven by modernization, has faced sustained criticism since the 1980s, though it still has vigorous defenders and is enjoying something of a renaissance in the work of scholars like David Voas (Voas 2008, 2009; Voas and Chaves 2016; Bruce and Voas 2023). Its most ambitious challenger in the 1980s and 1990s, the supply-side or religious economies school associated with Rodney Stark and his colleagues (Stark and Iannaconne 1994; Stark 1999; Finke and Stark 2005), explains religious vitality or its absence in terms of competition between religious firms and disputes the claim that modernization leads to a general decline in religious vitality. A looser group of revisionist approaches, represented by scholars like David Martin (1978, 2005) and Philip Gorski (2000; Gorski and Altınordu 2008; Nelson and Gorski 2014), differs to some degree from both of these schools and provides alternative analyses of secularization that emphasize cultural differences and historical contingencies.

Given the intense disagreements between and within these three broad approaches, it is easy to overlook the fact that they hold certain concepts in common. This article identifies and examines one group of shared concepts, which I call the “differentiation story”. The differentiation story pictures secularization as involving a historical process of structural differentiation in which social spheres like politics and education became distinct from a religious sphere and were thereby freed from religious control. The choice of the term “story”, unlike “theory” or “model”, indicates that these elements are more often assumed as part of the conceptual background of theorizing about secularization, rather than serving as the objects of investigation or argument themselves. The differentiation story tends to function as a kind of master picture or metaphor that is invoked rather than argued.

In what follows, I demonstrate that this differentiation story is found in the works of theorists on all sides of the secularization debate as it has played out in the English-speaking world since the 1960s,1 reproduced even by some of those who are highly critical of other aspects of the inherited secularization paradigm. The article then takes a closer, critical look at some of the core elements of the differentiation story, in terms of both their empirical accuracy and their theoretical justification. I argue that there are potential problems with the story’s accuracy as a historical generalization about Western societies, its underlying master concept of social spheres, and its assumptions about the relationship between social structures and values. I conclude by suggesting that secularization scholars should stop relying on the differentiation story as a common-sense “given” and instead should relax its assumptions and subject its elements to critical analysis and debate.

2. The Differentiation Story

In his compelling essay “Historicizing the Secularization Debate”, Gorski (2000) shows that all major variants of the classical secularization paradigm—the “old paradigm”—agree that “religious and nonreligious institutions have become increasingly differentiated over time” (p. 140). This, Gorski argues, forms the “central claim” or “core” of the old paradigm. Its main challenger, the “new paradigm” of the religious economies school, is concerned with different issues, leaving this core differentiation idea not only “unchallenged” but “as solid as ever” (p. 143). Casanova (2006) likewise writes that “the differentiation of the secular spheres (state, economy, science), usually understood as ‘emancipation’ from religious institutions and norms”, is the “core component of classical theories of secularization” (p. 7, original emphasis). In fact, he claims, “this understanding of secularization as a single process of functional differentiation of the various institutional spheres or sub-systems of modern societies remains relatively uncontested in the social sciences, particularly within European sociology” (Casanova 2006, p. 9). In what follows, I substantiate and extend Gorski and Casanova’s observations by showing the prevalence of this differentiation story in works from the major schools of thought about secularization since the 1960s.

2.1. Classical Secularization Paradigm

Olivier Tschannen’s overview of the leading theorists of the 1960s and 1970s—Thomas Luckmann, Peter Berger, Bryan Wilson, (early) David Martin, Richard Fenn, Talcott Parsons, and Robert Bellah—confirms that despite their important differences, they represented a common “secularization paradigm” with shared features. The most important shared feature, according to Tschannen, was the concept of differentiation, which he says “in one form or another, is absolutely central to all the secularization theories, without exception” (Tschannen 1991, p. 404). Tschannen summarizes this concept as follows:

In the course of history, religion becomes progressively differentiated from other domains of social life, eventually emerging as a very specific institutional domain within a new type of social structure made up of several such institutions (education, politics, economy, etc.). For example, Church and State become clearly differentiated… At the same time, the different non-religious institutions born from this process of differentiation start working on the basis of criteria that are rationally related to their specific social functions, independently from any religious control or guidance.

(Tschannen 1991, pp. 400–1)

Nearly all of these theories, Tschannen adds, conceive of differentiation as “autonomization”, which is the liberation the of social spheres from the control of religious values or institutions, allowing them to function autonomously according to their own norms. In describing Berger’s approach, for instance, Tschannen gives the example of “the emancipation of education from religious tutelage” (Tschannen 1991, p. 407; cf. Berger 1967, p. 107).

More recent reformulations and defenses of the classical paradigm also fit this description. Bruce’s (2011) Secularization: In Defence of an Unfashionable Theory, for example, defines secularization as including “the shift from religious to secular control of a variety of social activities and functions” (p. 2), and his diagram of the factors involved in secularization gives a central place to structural differentiation (p. 27). He argues that “Increased specialization directly secularized many social functions that were once dominated by the church: education, health care, welfare, and social control. Either the state directly provided such services or, if churches remained the conduit, the work became increasingly governed by secular standards and values” (p. 30).

Karel Dobbelaere, also a proponent of the classical paradigm, is likewise committed to the differentiation story. Dobbelaere distinguishes between three levels of secularization. At the “societal” level, his key explanatory concept is what he calls “functional differentiation, to wit, the differentiation of sub-systems in society (e.g., economy, education, polity and law) on the basis of their specific functions” (Dobbelaere 2002, p. 19). Elaborating elsewhere, he explains that “Modern societies are primarily differentiated along functional lines and have developed different sub-systems (e.g., economy, polity, science, family and education)”, each of which has its own characteristic organizations which function “according to the values of its sub-system and its specific norms”. He identifies a process of “autonomization of the sub-systems” whereby they “affirm their autonomy and reject religiously prescribed rules” (pp. 165–66).

2.2. Religious Economies School

In the 1980s and subsequently, the classical paradigm was challenged by the religious economies or supply-side school associated with Rodney Stark (e.g., Stark and Iannaconne 1994; Stark 1999; Finke and Stark 2005). Armed with a well-articulated alternative explanatory framework and striking (though disputed) statistical evidence, the religious economies school put advocates of the classical paradigm on the defensive and sparked a theoretical and empirical debate between the two schools that has continued into recent years (e.g., Thiessen 2012; Bibby 2012; Bruce and Voas 2023). What the religious economies school did not do, however, is call into question the account of structural differentiation assumed by the old paradigm. As Gorski (2000, p. 143) points out, the new approach did not challenge the idea of a differentiation of religious and secular spheres; instead, it sought to explain “the internal structure of the religious sphere and the effects that this structure has on aggregate levels of individual religiosity”. In fact, Stark explicitly accepts the differentiation story, which he describes as “a decline in the social power of the once-dominant religious institutions whereby other social institutions, especially political and educational institutions, have escaped from prior religious domination”. He adds, “If this were all that secularization means, there would be nothing to argue about” (Stark 1999, p. 252). The religious economies approach, therefore, takes the differentiation story as given, disputing only the claim that differentiation leads to a decline in individual religious beliefs and practices.

2.3. Historicist Approaches

Alongside these two grand paradigms, the classical paradigm and the religious economies model, various scholars have developed alternative accounts of secularization that belong to what Nelson and Gorski (2014, p. 4) identify as a third school following “a contingent and historicist approach”. The pioneer in this school is David Martin, whose book A General Theory of Secularization (Martin 1978) emphasized different historical patterns of secularization in different cultural regions. Nevertheless, he endorsed the differentiation story as a general pattern in modern societies (e.g., pp. 3, 279) even while stressing cultural differences in its expression and results (e.g., p. 205). In more recent work, Martin (2005, p. 49) continued to stress historical-cultural differences in the secularization process but also continued to accept structural differentiation as a “master trend” involving “the freeing of sectors of social life such as education and welfare from ecclesiastical oversight”.

Riesebrodt’s (2010) theory of religion also critiques several elements of the classical secularization paradigm. In his view, differentiation does not necessarily lead to a decline in the influence of religion in people’s lives (agreeing here with Stark, among others). He emphasizes the need for attention to historical case studies of the actual relationships between differentiation and other types of secularization and draws attention to the variable nature of differentiation between church and state in even Western countries. But Riesebrodt’s definition of the term “secularization” reproduces the differentiation story in familiar language: “The concept of secularization refers to a transformation of social orders, namely, to the process of freeing social institutions from religious control. This process results in the emergence of relatively autonomous spheres with their own standards and rules: on the one hand, secular institutional complexes such as the economy, politics, and culture; on the other, a rather clearly delimited religious sphere” (p. 175).

Casanova’s (1994) Public Religions in the Modern World likewise calls into question several key claims of classical theories, especially the claim that modernization must or should lead to a privatization of religion, but affirms the differentiation story. “To say that in the modern world ‘religion becomes private,’ refers also to the very process of institutional differentiation which is constitutive of modernity”, he writes, “…whereby the secular spheres emancipated themselves from ecclesiastical control as well as from religious norms. Religion was progressively forced to withdraw…. This forms the unassailable core of modern theories of secularization, a core which remains unaffected by the frequent assertions of critics” (Casanova 1994, p. 40).2 In a later essay, Casanova does question whether what happened in Western societies can be summed up as “a single teleological process of modern functional differentiation” and also questions whether “secular differentiation” is a necessary component of modernity everywhere in the world (Casanova 2006, p. 10).3 He does not, however, develop these questions into a full critique of the differentiation story.

Finally, Gorski’s (2000) article mentioned above presents a sophisticated critique and reconstruction of the classical paradigm’s account of differentiation and secularization. Gorski comes closest to a direct challenge to the differentiation story, by highlighting its role in secularization theory and showing that it oversimplifies historical events. Gorski’s specific objection to the differentiation story is that it tends to assume differentiation was a more or less linear process. Against this assumption, he argues that the era of confessionalization following the Reformation amounted to a massive “de-differentiation” of society (pp. 158–59). In this period, early modern states and territorial churches became closely institutionally linked and cooperated in governing education, social welfare, and the like. Yet, Gorski stops short of a full critique of the differentiation story and appears to accept its core assertion that there are essentially “religious and nonreligious values and institutions” which can become, and in the modern West have become, structurally differentiated (p. 161). He also affirms an idea, stemming from the work of Max Weber, that there is an inherent tension between the values of the religious sphere and the other “value spheres” (Gorski 2000, pp. 159, 161). This concept of value-spheres is an important constitutive element of some versions of the differentiation story, to which we will return shortly.

2.4. Elements of the Differentiation Story

From the foregoing examples, it is clear that scholars representing all three major approaches to secularization draw upon a common differentiation story with some or all of the following elements.

• Society is made up of various distinguishable spheres (whether called “spheres”, “sub-systems”, “institutions”, “domains”, “functions”, or “sectors”) (e.g., Tschannen 1991, p. 400; Bruce 2011, p. 2; Dobbelaere 2002, p. 19; Stark 1999, p. 252; Martin 2005, p. 49; Riesebrodt 2010, p. 175; Casanova 1994, p. 40; 2006, p. 7; Gorski 2000, p. 159).
• These spheres are now much more clearly differentiated, at least in modern Western societies, than they were at some point in the past (e.g., Tschannen 1991, p. 400; Bruce 2011, p. 27; Dobbelaere 2002, p. 165; Martin 1978, p. 3; Casanova 1994, p. 40; Gorski 2000, pp. 159, 161).
• Religion is something distinguishable, always conceptually and often also institutionally, from the other social spheres. Thus, religion is a sphere (e.g., Dobbelaere 2002, p. 166; Martin 1978, p. 205; Riesebrodt 2010, p. 175; Casanova 1994, p. 21; Gorski 2000, p. 161). Often the religious sphere is identified with the church institution, ecclesiastical authorities, or clergy (e.g., explicitly in Martin 1978, pp. 3, 278–79; Martin 2005, p. 49; implicitly in Tschannen 1991, p. 400; Bruce 2011, pp. 2, 30; Casanova 1994, p. 20; Stark 1999, p. 252).
• The other spheres have innately nonreligious values, functions, norms, or purposes. Although they were under religious control at some point in the past, religion is not (so to speak) indigenous to them: their true natures or purposes are secular, that is, nonreligious. This point is implicit in the typical descriptions of these spheres as “nonreligious” or “secular”, but for more explicit discussions see Gorski (2000, pp. 159, 161 on “value spheres”) and Tschannen (1991, p. 401 on “criteria…rationally related to their specific social functions”).
• Religion previously controlled the other spheres but has lost this control through the differentiation process, which allowed the other spheres to autonomously pursue their innately nonreligious functions. Scholars often characterize the emergence of these “secular” spheres from religious control as emancipation or liberation (Tschannen 1991, p. 407; Martin 2005, p. 49; Riesebrodt 2010, p. 175; Casanova 1994, p. 40; 2006, p. 7; see also Durkheim [1869] 1933, p. 169). This is the conclusion to which the rest of the elements of the differentiation story point.

Not all of these components are necessarily invoked by each theorist, but the references provided show that most of them are explicitly present in several major works. Moreover, the components are logically interdependent.

While these theorists share the assumptions of the differentiation story, they disagree about its implications for the place of religion in modern society. Those working in the classical paradigm typically think that differentiation reduces the social importance of religion in some way, though they disagree about whether they conceive of this as a privatization, decline, or disappearance of religion or even an institutional privatization alongside a generalization of religious values (Gorski 2000, pp. 140–41). Supply-siders, in contrast, see differentiation as likely to improve the effectiveness of religious organizations and therefore to increase levels of active individual religiosity (Stark 1999; Stark and Iannaconne 1994). Revisionists take a variety of positions; Casanova (1994), for example, accepts differentiation but does not believe that it necessarily leads to the privatization or decline of religion. Agreement about the differentiation story does not require agreement about its implications.

Likewise, we can distinguish between two ways in which scholars employ the differentiation story. On the one hand, they use it in a purely descriptive way. The claim here is simply “this is what has happened in certain secularized societies”. In such societies, as this descriptive use of the story goes, people have come to think of religion as a separate “sphere” distinct from other domains of life like politics and to see the latter in purely secular terms. In other words, in such societies a process of differentiation of spheres happened to be at the same time a process of secularization insofar as religion came to be seen as merely one sphere among others. Used in this way, the differentiation story is nothing more than a description of a particular historical pattern that may be highly contingent on specific circumstances in a specific cultural setting (e.g., Western Christendom). This version of the story does not go beyond the first two or three of the components of the story in the bulleted list above.

On the other hand, scholars also employ the differentiation story in an explanatory mode by assuming the remaining points in the bulleted list. In the explanatory use of the story, not only has religion come to be seen as distinct from other spheres in a particular cultural setting, but this situation has come about precisely because these other spheres are in fact innately not religious; that is, they are secular by nature. In this use of the story, the emergence of distinct religious and secular spheres is not a contingent and possibly unique historical development but is the actualization of underlying distinctions between different areas of life that are inherent in social reality itself. The secular social functions are, in a sense, striving to achieve their potential, to emerge from the undifferentiated state and fully come into their own— especially, to escape the imposition of religious control. This is why theorists do not talk about secular spheres being created or constructed, but rather of secular spheres achieving “autonomy” from religion (e.g., Dobbelaere 2002, pp. 165–66; Tschannen 1991, pp. 400–1) being “free[d]…from religious control” (Riesebrodt 2010, p. 175) and “escap[ing] from prior religious domination” (Stark 1999, p. 252). The secular social spheres are not invented in the differentiation process; rather, the differentiation process actualizes a distinction between religious and secular spheres that was always there in principle, even though it was obscured by the religious domination of society. The implicit claim is that, insofar as a universal process of increasing social complexity produces greater differentiation of social spheres, we should expect to see a secularization of most social spheres as they are emancipated from religious control in any sufficiently developed society. This use of the differentiation story goes beyond describing something that has happened in a particular society to explaining secularization as a result of apparently universal principles of differentiation.

Although descriptive and explanatory uses of the differentiation story can be distinguished in principle, it is not always clear whether a given scholar is employing the story in a descriptive or explanatory way. Often, however, explanatory claims are at least implied by language about the “emancipation” of social institutions from religious control, because it implies that religion is somehow an alien imposition on said institutions, which in turn implies that greater differentiation will free them to assume their true, secular character. As shown above, scholars very commonly employ such language when invoking the differentiation story, which suggests that much of the time the differentiation story carries explanatory implications.

Before moving on, a word of elaboration is in order concerning the idea that most social spheres are inherently nonreligious. This idea has a long pedigree in the social sciences, including the work of Weber ([1920] 1946, 1968). Weber saw an inherent tension between the ethical demands of religion and the inherent requirements of activity in “autonomous spheres of life” (Weber 1968, p. 578), especially what he called the “non-ethical or unethical requirements of life in the political and economic structures of power within the world” (p. 598). In his view, secular institutions, especially political and economic ones, have an innate tendency to organize their activities according to functional or goal-oriented rationality (Zweckrationalität), while religion remains governed by value-oriented rationality (Wertrationalität) (Dobbelaere 2002). Thus, Weber concluded that

To the extent that a religious ethic organizes the world from a religious perspective into a systematic, rational cosmos, its ethical tensions with the social institutions of the world are likely to become sharper and more principled; this is the more true the more the secular institutions (Ordnungen) are systematized autonomously.

(Weber 1968, pp. 578–79)

Thus, Weber pictured a differentiation process whereby various areas of life, including economics, politics, sexuality, and art, shake off religious norms and pursue their autonomous ends.

Although recent theorists rarely discuss this Weberian notion explicitly, it appears to lie at the root of the characterization of the differentiation process as emancipation of the nonreligious spheres from religious control and its corollary, the belief that the autonomous functioning of these spheres will necessarily result in them operating in a secular way. It only makes sense to talk this way if these other spheres are inherently or naturally secular. That is why only outside control can bend these spheres to religious norms and purposes; freed from such control, they disclose their true nature. Thus, drawing on Durkheim and, implicitly, Weber, Bruce writes that the process of differentiation freed various areas of life to follow their “own values”, which Bruce assumes are “rational”, i.e., non-religious, ones (Bruce 2011, p. 30). Likewise, Tschannen, summarizing the 1960s and 1970s theorists, explains that the newly differentiated “nonreligious” institutions “start working on the basis of criteria that are rationally related to their specific social functions, independently from any religious control or guidance” (Tschannen 1991, p. 401). Again, this implies that religious criteria are not rationally related to the functions of these institutions, so must be imposed from without.

Among recent theorists cited above it is Gorski (2000) who most explicitly draws upon the Weberian notion of value-spheres. In posing hypotheses about what led to a “re-differentiation” of society after the “de-differentiation” of the Reformation era, he suggests that “the principles of the churches inevitably conflicted with the principles underlying other ‘value spheres’ (political, economic, aesthetic and so on)”, such that “the attempts of church leaders to impose [religious principles] on society met with varying degrees of resistance—from monarchs inspired by raison d’état, from merchants who traded with heretics and infidels, from artists who appropriated religious imagery, and so on” (Gorski 2000, p. 159). Even though Gorski is critical of other ideas of Weber’s, such as his account of the relationship between differentiation and rationalization, he appears to endorse the notion of value-spheres and sees it as a promising starting point for future theories of religious change (p. 161).

Returning to the differentiation story as a whole, it appears to enjoy widespread support across the opposing parties of the secularization debate as it has played out in English-language scholarship since the 1960s—perhaps even canonical status. Of course, there are scholars outside the social sciences, including historians and philosophers such as Hugh McLeod (2000, 2007), Callum Brown (2009), and Charles Taylor (2007), who bring a different set of questions and disciplinary approaches to the study of secularization.4 And some social scientists have offered accounts of secularization that do not engage significantly with the differentiation story one way or the other.5 But in terms of attempts to provide a systematic, social scientific account of the secularization of Western societies, the examples earlier in this section suffice to show that the differentiation story is widely assumed in classical secularization theory and that even leading critics of the classical paradigm retain the story’s core assumptions. Indeed, part of the power of the differentiation story is that it seems like common sense to many Western observers—something that can be taken as given rather than established through argument. This is all the more reason, then, to take a closer, critical look at the empirical accuracy and theoretical plausibility of the story.

3. Problems with the Differentiation Story

Without attempting a comprehensive critique of differentiation theory in general or even the secularization-focused differentiation story in particular, this section highlights some apparent problems with the differentiation story as it is commonly employed by secularization theorists.6 First, scholars who invoke the story are assuming that it is an accurate summary, albeit in simplified form, of the historical experience of secularized societies. It is far from clear, however, that this assumption is warranted. Even in the case of secularized Western societies—to which the differentiation story should apply if it applies anywhere—the historical record is not easy to square with a master narrative of increasing institutional differentiation leading to the divergence of value-spheres. This is a problem for both the descriptive and the explanatory uses of the story. A second problem, more theoretical in nature, applies to the use of the story as an explanation of secularization: the crucial and apparently common-sense notion of “social spheres” as employed by secularization scholars turns out to be deeply unclear when closely examined. This lack of clarity leaves secularization scholars without a theoretical justification for identifying what exactly we should expect to be differentiated by the differentiation process—including religion. This observation brings us to a third problem, which is that the assumption that most spheres are inherently or rationally nonreligious is not well justified theoretically. In conclusion, I ask whether, in the face of global evidence that religion and the “secular” spheres have not in fact been differentiated historically in this way in non-Western societies, it is more plausible to assume that these societies were devoid of significant differentiation and thus failed to actualize the inherently secular character of those domains or that the universalist assumption of the innate secularity of these spheres is mistaken.

3.1. Inadequacies as a Generalization about Western History

As noted, the master theme of the differentiation story is one of increasing specialization and autonomy of various social spheres, but Western history calls into question the generalization that major social institutions have followed an overall trend of increasing autonomy over time. For the sake of argument, I am here assuming that the “spheres” (and equivalent terms) referenced by scholars invoking the secularization story include, at least, major social institutions such as the university (higher education) and the state (government), an assumption that is warranted by the way scholars have employed these terms as documented above. It is far from clear, however, what should and should not count as a “sphere” at the conceptual level in the differentiation story, a point explored further below. In this section, I examine whether the differentiation story works as a historical generalization if we assume that the university and the state, inter alia, count as “spheres”, a reasonable assumption in light of the explicit examples of spheres given by secularization scholars. Furthermore, secularization scholars have generally not explained in detail what they mean by terms like the “emancipation” and “autonomy” of spheres from outside (usually religious) control. It is reasonable to assume that (for example) the legal, administrative, or financial dependence of one institution upon another and the ability of other institutions to supervise, direct, and limit the activities of the institution in question are conditions of low autonomy, while legal, administrative, and financial independence and freedom to determine institutional priorities and activities without outside approval or regulation are conditions of high autonomy.

We shall consider some general historical difficulties with the overall picture presented by the differentiation story but will begin with the history of a particular social institution, the university, an institution of great influence in the modern West that is sometimes implicated as a key driver of secularization (e.g., Berger 1999). The differentiation story would imply that universities began their life under the church’s domination but were gradually emancipated from ecclesiastical control through the differentiation process, achieving greater autonomy over time to pursue purely educational and scholarly goals—secular goals—apart from the control of religious authorities, implicitly, the institutional church.7 This is not, however, the story that emerges from the work of historians in recent decades on the history of Western universities, a large and complex story that I briefly summarize here. Instead, the historical literature identifies decreasing autonomy as one of the defining features of the institutional development of the university from the thirteenth to the twentieth century and reveals that the periods when universities were the most autonomous were also the periods in which they were the most religious.

Universities have never enjoyed greater autonomy than they did in the earliest period of their existence, the high Middle Ages, due in part to competing patronage efforts by papal and princely authorities who sought to outdo each other in guaranteeing university rights against outside interference (Nardi 1992; Howard 2006). Thus, members of medieval universities were exempt from civil jurisdiction and, to a surprising degree, from local ecclesiastical jurisdiction; they decided who could teach and what they could teach; they set their own internal regulations and selected their own officers (Gieysztor 1992). Likewise, a combination of low costs and a wide variety of funding sources meant that they were not financially controlled by any single outside body (Rudy 1984).

Far from increasing over the centuries, university autonomy was progressively curtailed during the period stretching from the late Middle Ages through the Reformation and the early modern period, when a combination of centralized state-building, territorial confessionalization and religious competition, and increasing building costs allowed state and church authorities (the latter also now under state control in Protestant lands) to impose stricter ideological, curricular, regulatory, and financial control on universities in both Protestant and Catholic territories (Rüegg 1996; Hammerstein 1996b; de Ridder-Symoens 1996). The power to appoint professors, for example, came more and more to belong to state officials, even though in some cases this required abrogating pre-existing rights of universities or their faculties: professors became in effect civil servants (Vandermeersch 1996).

From the eighteenth century onward, church authorities became increasingly irrelevant to the universities, but state control over finances, curriculum, and the faculty intensified (e.g., Gerbod 2004a, 2004b). Nowhere was this truer than in France, where the revolution and Napoleonic era transformed university education into a centralized, state-run affair. For example, the revolutionary and subsequently Napoleonic state confiscated the land endowments that had provided French universities with a measure of financial independence, abolished the very legal existence of the 22 universities then in France and replaced them with a centrally controlled and administered system of professional schools, determined the ideological and philosophical character of what was taught in the higher education system, and prohibited non-state institutions of higher education or church involvement in the government-run system (Anderson 2004, pp. 40–46; Brockliss 1997). Similar developments occurred in places like Austria under Joseph II and Portugal under the Marquis de Pombal (Rüegg 2004; Hammerstein 1996a). Even the Humboldtian model adopted in some German lands and widely imitated elsewhere combined its vaunted Lehrfreiheit and prominent role for the scholarly guild with state financing, state exams, state limits on self-expression of scholars (especially in the realm of political criticism), and a state right to appoint and dismiss professors. By the twentieth century, intensifying financial reliance on the state, close regulation by ministries of education, and state policies related to professional training meant most universities were far from autonomous (Anderson 2004; Gerbod 2004a; Neave 2011). If there is a master trend in the history of universities, therefore, it is one of decreasing institutional autonomy and increasing integration with the state.

Significantly, the periods of greatest institutional autonomy also tended to be periods of religious orthodoxy for universities, while the most crucial periods of secularization—in the sense of a departure from Christian assumptions and constraints—were initiated by state intervention or control (Rudy 1984; Anderson 2004). As late as the eighteenth century, universities in the Western world were deeply committed to Christian orthodoxy (in its respective Catholic or Protestant versions) as the basis for teaching and scholarship, a commitment that had been present since the Middle Ages and had if anything intensified under the watchful eye of the confessionalized state in the early modern period (e.g., Hammerstein 1996b, pp. 119–21; Brockliss 1996). For example, faculty and students at Oxford and Cambridge, the only universities in England until the nineteenth century, were required to subscribe to Anglican confessional standards until the 1870s (Brooke 1993). French universities were so committed to Catholic orthodoxy in the eighteenth century that later observers often have accused them of intellectual stagnation (Rudy 1984, pp. 83–84; Hammerstein 1996b, pp. 126–27; cf. Anderson 2004, pp. 5, 17). German universities, at the outset of the eighteenth century, similarly understood their task in “profoundly religious terms” and “Oaths of confessional fidelity were the rule of the day” (Howard 2006, p. 46).

On the other hand, the key junctures in the secularization of universities resulted to a large extent from state intervention in the eighteenth and nineteenth centuries. Just to give a few examples, the shift of the University of Halle from Pietism to rationalism was partly driven by the priorities of Frederick the Great, who reappointed the previously dismissed determinist Christian Wolff to the faculty, allowing him to rise to the position of chancellor. As noted above, prior to the revolution, French universities were deeply committed to Catholic orthodoxy (Brockliss 1997; Hammerstein 1996b); it took the wholesale abolition and recreation of the nation’s universities by the revolutionary government and Napoleon to build a secular system of higher education in France. And it was parliamentary fiat that finally overcame the exclusively Anglican character of the ancient English universities in the nineteenth century (Bebbington 1992; Brooke 1993). Though the American situation is somewhat different, in that a leading role in secularization was played by an academic guild imitating the already established German model, here too the period of most rapid secularization corresponded to the period of greatest intensification of state control (Marsden 1992, p. 20). (For an overview of the secularization of universities in four countries further substantiating the connection between state control and secularization, see (Flatt 2020)).

Someone might object that, whatever the institutional arrangements, the modern state-controlled university is more “autonomous” than the medieval university in that modern states, at least since the mid-twentieth century, guarantee a wide freedom in the content of teaching and research, while the medieval scholar was beholden to a system of dogma that limited the range of permissible doctrines and conclusions. This is an important objection. Granting, for the sake of argument, the premise of the freedom of teaching and research in the modern university in comparison to the premodern situation,8 my response is twofold. First, in the periods of most rapid secularization in the late eighteenth and nineteenth centuries in the countries whose university systems served as global models (France and Germany), academic freedom was not absolute but in fact tightly constrained by state and university authorities (Hammerstein 1996b; Rüegg 2004; Anderson 2004, pp. 8, 43–47, 58–61; Rudy 1984, pp. 100–4). The periods of intensive secularization thus do not correlate with the periods of the greatest autonomy of teaching and research, suggesting that other factors beyond autonomy must be at play.

Second, a key question here is whether there is a connection between institutional autonomy of the kind I defined above and the claim being made by secularization theorists in the differentiation story. Do “emancipation from religious control” and “autonomy to pursue its own purposes” in the case of the university mean nothing more than “at some point universities stopped orienting their activities according to religious ideas”? If so, the objection holds, and the institutional history of the university given above (as opposed to its intellectual history) becomes largely irrelevant to the question of secularization. In that case, the differentiation story also ceases to say much interesting in terms of providing an explanation of secularization and becomes rather tautological. Yet, the differentiation story, at least in its explanatory versions, seems to be making a stronger claim: formerly, the church was able to control other social domains (states, economies, and, yes, education), but when it lost that control, it “freed” these institutions to pursue the “rational”, i.e., secular or nonreligious, purposes proper to them. In other words, when employed in an explanatory fashion by secularization theorists, the differentiation story posits a causal connection between institutional autonomy from religious (implicitly or explicitly, ecclesiastical) control and the waning hold of religious considerations in those institutions. It is precisely that causal connection that is directly challenged by the actual historical experience of universities outlined above.

In sum, the emergence of universities as distinct and originally largely independent institutions in the Middle Ages was unquestionably a process of differentiation, but universities maintained their highly religious (Christian) character at least into the eighteenth century and were only secularized over the course of the late eighteenth, nineteenth, and early twentieth centuries. Differentiation and secularization, in the case of universities, therefore, were separated by five centuries or more. The history of this major Western social institution thus calls into question both the differentiation story’s generalization that there has been a master trend of increasing institutional autonomy and the accompanying causal claim that increasing autonomy has been the driver of secularization. In the history of the university, the precise opposite appears to be true, on both counts.

Broadening out the perspective from the university specifically to more general considerations raises further questions about the adequacy of the differentiation story. To the extent that the differentiation story deals with the control of one sphere by another, the emphasis is on the (past) control or domination of secular spheres by “religion”, usually implicitly identified with the institutional church, and the liberation of the former from the latter through the differentiation process. This account tends to obscure the increasing subordination of many institutions and areas of life, including but certainly not limited to the university, to the growing role of the state over the course of Western history (cf. Cavanaugh 2009; Gregory 2012, pp. 129–79). From the centralization of royal power in some medieval kingdoms to the emergence of the omnicompetent twentieth-century regulatory state, the story of Western history could just as easily be summarized as the increasing consolidation of diverse social functions under the control of this single institution. The last two centuries in particular have witnessed the extending reach of state financing, regulation, and policy direction into education, science, medicine, social welfare, and even artistic creation. But if distinct social spheres have their own intrinsic values rationally related to their purposes, as the explanatory use of the story assumes, what consequences arise from this subordination of ever-increasing areas of life to the state? In particular, how is state expansion and control related to secularization? In at least some cases, the growth of state power seems to have played an important role in secularization (e.g., Peker 2019) and was wielded to bring about secularization as part of a conflict between status-seeking secularist groups and religious authorities (e.g., Smith 2003), but these points are obscured by the differentiation story’s central emphasis.

Similarly, insofar as it rests on the Weberian idea of a differentiation of value-spheres over time, the differentiation story hides the remarkable convergence of values across different domains of life in the late modern West, in particular convergence on the core value of the autonomy of the individual. There is an obvious alignment between the citizen of liberal democracy—the free and equal individual considered apart from any embedding communities or groups—and the imagined mobile economic agent of free market capitalism, who singly and autonomously enters and exits contractual agreements of employment, production, and consumption. This convergence is not limited to politics and economics, though, but can also be seen in other domains in what Taylor calls “the Age of Authenticity” (Taylor 2007, pp. 473–86). Consider the twentieth-century shift in the goals of public education in many Western countries from the inculcation of Christian or republican values to individual self-realization, in which students are encouraged to become self-aware as autonomous moral agents and choose their own system of values (e.g., Gidney and Millar 2001). Consider also the recent shift in sexual mores from an emphasis on restraint and subordination of sexuality to other purposes (e.g., marriage, child-rearing, and social discipline) to an ethic of individual self-fulfillment, reproductive choice, and sexual and relational consent (e.g., Witte 1997, pp. 194–213). Similar observations could be made about the rise of self-expression in defiance of constraining norms as the central value of the creative arts and even about the increasing “Sheilaism” (Bellah and Madsen 1996) of contemporary religion and spirituality. Late modern Western societies, far from experiencing increasingly sharply diverging value-spheres, appear to have achieved a degree of agreement about secular, humanistic ultimate values that in terms of its cross-sphere coordination rivals the quite different shared values of medieval Christendom. Both in the particular example of the history of the university as a social institution and in terms of the more general considerations of the growing role of the state and the convergence of values across spheres, it is far from clear that the differentiation story presents a good summary of the relevant trends in Western history.

3.2. The Mystery of the Spheres

As we have seen, the idea of a differentiation of social spheres over time is foundational to the differentiation story as a whole. But what is a sphere? The foregoing analysis has left this question open, but it merits close consideration. In practice, secularization theorists universally assume the existence of spheres (or a terminological equivalent) but do not rigorously define or justify the concept. The first clue to this conceptual fuzziness is the wide range of seemingly distinct words that secularization theorists use to refer to spheres. Tschannen, in his summary of 1960s and 1970s theories, refers to both “domains of social life” and “institutions” (Tschannen 1991, p. 400). Bruce talks about “specialized roles and institutions” (Bruce 2011, p. 29) but also gives a list of what he calls “social functions” (p. 30). Dobbelaere uses various terms—“institutions” (Dobbelaere 2002, p. 13), “spheres” (p. 30), “sub-systems” (p. 165), and “systems” (p. 169)—apparently as rough synonyms (education, for instance, appears in all four lists). Gorski (2000, pp. 159–60) and Casanova (1994, pp. 6, 40) sometimes say “spheres” and sometimes “institutions”. Riesebrodt appears to use the terms “social institutions”, “spheres”, and “institutional complexes” interchangeably (Riesebrodt 2010, p. 175). These examples could be multiplied.

One might assume that these various terms were simply acceptable synonyms for a single coherent concept if the actual lists of examples of “spheres”, “institutions”, and so on given by theorists were roughly consistent with each other—but they are not. Rarely, if ever, do these secularization theorists attempt to give a comprehensive list of “spheres”. What we find instead are incomplete and mutually inconsistent lists of spheres or equivalents. Thus, Tschannen’s examples include “education, politics, economy, etc.”, but also “Church and State” (Tschannen 1991, p. 400). Bruce’s list of social functions includes “education, health care, welfare, and social control” which he distinguishes from both “the church” and “the state” (Bruce 2011, p. 30). Martin’s list of institutional spheres includes “justice, ideological legitimation, the state apparatus, social control, education, welfare” (Martin 1978, p. 3). Dobbelaere gives perhaps the widest variety of lists, listing “culture, education, law and medicine” as well as “the constitution” as institutions (Dobbelaere 2002, p. 13), elsewhere naming “the economic and political spheres” alongside “science” and “the educational system” as spheres (p. 30), and in a third place giving a list of sub-systems including the “economy, polity, science, family and education”—which he distinguishes from “organizations” including “enterprises; political parties; research centers; families; schools and universities” (p. 165). A few pages later he lists “religion and the juridical, the educational, the economic, the family, the scientific, the medical, and the political systems” (p. 169).

Clearly, there is no agreed standard of what the spheres are that are differentiated through the differentiation process; sometimes authors are not consistent themselves. Nor is this an easily resolvable problem. On the one hand, it is not clear at what conceptual or organizational level the boundaries should be drawn. Using examples from the theorists above, take “politics” or “the political”. Is it a sphere? Or is “the state” a sphere? Or should we think, as some of these authors do, in terms of distinct spheres of “law”, “the constitution”, “social control”, “justice”, “ideological legitimation”, and “welfare”, all of which might by other theorists (or even at other times by the same theorists) be lumped in as part of a “political” sphere or a “state” sphere? Furthermore, not only is there a problem in terms of conceptual lumping or splitting, but the various items within particular sphere-lists sometimes appear not to be conceptually equivalent phenomena. Do “culture”, “medicine”, and “the constitution” (Dobbelaere) really belong to the same category of analysis? Within and between lists there is a shuffling together of what could be distinguished as organized groups (e.g., government, family, business firm, school, and hospital), modes of activity (e.g., administration of justice, sexual relations, scientific research, artistic creation, education, and medical care), conceptual systems for organizing meaning (e.g., law, science, and art), and sites or networks of formal interaction between individuals and groups for specific purposes (e.g., the economy).

There is good reason to think that these inconsistencies and ambiguities were not created by secularization scholars but are more deeply rooted in the work of influential theorists of functional differentiation such as Spencer, Durkheim, and Parsons (for a detailed analysis and critique, see Taylor 2020). Whether satisfactory solutions to such problems have been or could be proposed in the general differentiation literature, that is, beyond the work of these secularization theorists, is not a question that can be pursued here. For present purposes, the key point is that not only do secularization scholars disagree with each other about which sphere-word should be used or what domains should be included in the list of spheres; they do not agree about what the concept of a sphere entails. More is at stake here than the exact language used to describe spheres. To accurately explain or even describe historical phenomena such as secularization in terms of differentiation of spheres, it is essential to have an idea of what spheres are. An explanation of the secularization of politics, the economy, science, and art as a natural or logical outcome of the differentiation process, for example, takes it for granted that politics, the economy, science, and art, on the one hand, and religion, on the other, are inherently distinct spheres. Thus, despite the apparent obviousness of the differentiation story, there is a problem with its foundational concept: what, exactly, is being differentiated through the process of differentiation?

Even if we grant for the sake of argument that it would be possible to formulate a coherent concept of “sphere” and to come up with a defensible comprehensive list of spheres—again, the literature on differentiation is extensive, and this article does not attempt a critique of it beyond the use of certain concepts by these secularization theorists—it does not necessarily follow that religion should be considered a “natural” sphere in this sense. Nor is it obvious that “religion” and “church” can be used interchangeably, as some theorists do, such that the institutional differentiation of church and state (for example) is also necessarily a differentiation between religion and the state.

3.3. The Problematic Notion of Value-Spheres

A closely related theoretical problem in the explanatory use of the differentiation story is the concept of “value-spheres”. As discussed earlier, this concept posits that each sphere has values, functions, norms, or purposes that are logically proper to it, and that in the case of most spheres—the “secular” spheres—the appropriate values, functions, norms, or purposes are intrinsically nonreligious ones. A religious orientation of politics, therefore, is an unstable situation that is ultimately corrected through the differentiation process by the separation of politics from religion as distinct spheres. Such a view has been challenged in similar terms by Taylor (2007) and Asad (2003). Taylor critiques “subtraction stories” in which “the underlying features of human nature” are regarded as having always been essentially secular, but long obscured by religious constrictions and illusions until they were finally “sloughed off” in the course of time (Taylor 2007, p. 22). Asad similarly argues “that ‘the secular’ should not be thought of as the space in which real human life gradually emancipates itself from the controlling power of ‘religion’ and thus achieves the latter’s relocation” (Asad 2003, p. 191). These critiques raise serious questions, yet we should not dismiss the notion of intrinsically secular spheres too quickly, without properly addressing the theoretical basis its advocates put forward for it in the notion of value-spheres. As Gorski (2000) points out and Dobbelaere (2002) discusses at some length, this way of thinking goes back to Weber, who posited an inherent tension between the value-orientation of religion and the goal-orientation or functional rationality of political and economic activity oriented toward Realpolitik and profit, respectively (e.g., Weber 1968, pp. 584–85, 590, 598, 600–1; see also Weber [1920] 1946, pp. 331, 334). For Gorski and Dobbelaere, as for Weber, different values (terminologically these authors do not necessarily follow Weber’s distinction between values and goals) are inherent to the different spheres. As with the concept of spheres itself, this seemingly obvious idea of value-spheres requires closer scrutiny.

First of all, the organizations allegedly defined by political and economic functional rationality, the state and the business, are hardly unique in pursuing the acquisition and preservation of power (raison d’état) and resources (profit). Any organizations, including religious ones such as the church, could orient themselves—arguably, have sometimes oriented themselves—primarily to one or the other of these purposes. Someone might object that in the case of the state, the pursuit and maintenance of power is the natural goal (Weber [1920] 1946, p. 334), whereas if it becomes the primary goal of the church, it is a perversion of the church’s natural aims. But on what grounds can we claim that the “natural” aim of the state is or should be the acquisition of power? Why is it power rather than the pursuit of justice, or representation of the people before the gods, or the elimination of the vestiges of the capitalist mode of production, or the preservation and expansion of individual liberty, or any of the other purposes that actual historical states have claimed and directed their actions toward? What Weber portrays as the defining goals of politics and economics are in fact simply two possibilities from a long list of goals human beings have sought to achieve in those realms. Relevant here is the argument of Carl Schmitt (at least in one reading) not only that putatively secular notions of the state are in fact derived from theological concepts,9 but that political doctrines necessarily stem from strong theological or philosophical assumptions (Schmitt 2005, pp. 36–52)—that there is, in other words, “a metaphysical kernel to all politics” (p. 51). One need not accept Schmitt’s specific analysis to take his point that even secular political orders require thick notions of the aims of politics and the nature of reality that transcend Weber’s arbitrarily narrow definition of functional rationality in the political domain and are not so easily divorced from larger questions of values.

To be sure, states and businesses must attend to the maintenance of their power and resources if they are to continue to function. States and businesses are not unique in this regard, however. The same is true of any kind of organization, including “religious” ones, especially in an environment of competition for power and resources. The question is, power and resources for what? To what end? In that sense, the maintenance of power and resources is no more the innate goal of the state and the business than of the church, the art gallery, the family, or other organized groups and institutions. All of them require power and resources to carry out their purposes—but all of them require purposes to carry out. Weber’s formulation therefore confuses ends and means, telos and techne, rather like politicians who claim to be “pragmatic” rather than “ideological”. Pragmatism is always pragmatic with respect to some goal, and goals are variable. Dobbelaere (2002, p. 87) levels a similar criticism at Weber’s distinction between value-oriented and (functionally rational) goal-oriented action: “goal-oriented action is invariably based on at least implicit value options…In extreme cases efficiency and expediency become the core values. As a result, the evaluation of rationality will in the last resort depend upon one’s values and consequently be relative. If categories of persons differ in value-orientation, they will see each other’s actions as ‘non-rational’ from their own point of view”.10

None of this requires a denial that there are distinct characteristics which identify an institution or sphere as “political” or “economic” or “artistic” or that these differences imply different specific realms of activity and different sets of concerns.11 The point here is that the realm of activity (political, economic, or artistic) does not alone dictate the goals pursued, nor does it require that the goals in different spheres be ultimately in tension with one another. These goals always reflect values, which may ultimately be compatible with or at odds with the values directing goals in other institutional contexts. In fact, one of the key claims of certain overarching religious or secular systems of meaning is that they provide a sort of master set of values that can orient activities across all realms of human activity. Islam orients all spheres to submission to the revealed will of God; Confucianism orients all spheres to ethical harmony reflecting the heavenly order; influential contemporary Western value-systems orient all spheres to the maximization of individual human autonomy. A secular state whose core value is individualism may pursue goals that bring it into conflict with those of an Islamic school, but in such a scenario the conflict is just as likely due to rival conceptions of life and the world rather than the necessary character of government or education per se. The fact that the concrete realization of overarching values takes a different form in politics than in economics or art does not negate the possibility of the ultimate congruity of values and goals across spheres, as noted above when discussing the convergence of values across social spheres in the modern West.

This line of analysis leaves open the possibility that the values directing what the differentiation story calls “secular” spheres or institutions may—even in highly differentiated societies—be religiously derived values, without there necessarily being an inherent tension or conflict between such goals and the essential nature of that sphere or institution as such. In other words, the values guiding an organizationally autonomous “sphere” may be religious ones, even if it is not under ecclesiastical, i.e., church, authority.12 If we define religion, as many scholars of secularization would (e.g., Bruce 2011, p. 1; Dobbelaere 2002, p. 52), as values or beliefs or practices predicated on the existence of superhuman persons or powers with moral purpose, religious values/beliefs/practices can be distinguished from churches/religious institutions and are not necessarily tied to, or excluded from, any particular differentiated structural entity existing in a society. Other organizations or fields of activity can be oriented according to beliefs or values that are “religious” in this sense, even under conditions of high levels of structural differentiation. And therefore, the observation that a social institution or other entity has become organizationally differentiated from ecclesiastical authority is not in itself a sufficient explanation for why that institution has (or has not) adopted a “secular” guiding purpose or set of values.

3.4. Taking a Global View

The foregoing questions about the differentiation story’s theoretical basis gain force in light of the argument that the modern West’s peculiar way of distinguishing between secular spheres and religion—both conceptually and institutionally—is alien to the historical experience of most human societies. This point has been a major theme of recent scholarship (Asad 1993; Assman 2014; Carnesecca 2016; Cavanaugh 2009; Fitzgerald 2014; Goossaert and Palmer 2011; Horii 2018b; Josephson 2012; Joustra 2018; Masuzawa 2005; Nongbri 2013; Sun 2013; Van der Veer 2013), some of which groups itself under the umbrella term of “critical religion”. Carnesecca (2016), for example, notes the inapplicability of Western Christian distinctions between politics and religion to Confucian and neo-Confucian China. Asad (1993, p. 28) likewise warns against “the attempt to understand Muslim traditions by insisting that in them religion and politics (two essences modern society tries to keep conceptually and practically apart) are coupled” (see also Asad 2003, chp. 7). In much of the Islamic experience, political, legal, and religious authority were seen as part of a single fabric rather than fundamentally distinct (e.g., Lapidus 2012, pp. 48, 53–54). A full exploration of this point is of course beyond the scope of this article, but examples from around the world could easily be multiplied based on recent scholarship (e.g., Assman 2014; Fitzgerald 2014; Goossaert and Palmer 2011; Josephson 2012; Masuzawa 2005; Nongbri 2013; Sun 2013; Van der Veer 2013). In light of this literature, some scholars have concluded that the concept of religion is thus unworkable and should be abandoned, while others maintain that a critically chastened but globally applicable definition of religion is possible (Riesebrodt 2010; Schilbrack 2010; Smith 2017). In either case, this literature suggests that the differentiation story’s understanding of religion as an aspect of life that is always distinguishable from other social spheres even though it may dominate or mingle with them simply lacks historical equivalents in most cultures.

These arguments have been built upon but also partly challenged by a strand of research under the heading of “multiple secularities” (Wohlrab-Sahr and Burchardt 2012; Burchardt et al. 2015; Kleine and Wohlrab-Sahr 2020, 2021). The multiple secularities literature accepts much of the critical approach to religion against older approaches that regarded religion and the religious–secular distinction unproblematically as universal concepts. It agrees that notions of the religious and the secular vary historically and culturally and builds on a modified version of Eisenstadt’s (2000, 2003) argument that there are multiple modernities, not just a single path of modernization that necessarily converges with the Western European or North American examples (see also, for example, Tu 2000; Wittrock 2000). Where the multiple secularities approach differs from some of the critical religion literature is that it hypothesizes that binary distinctions, both conceptual and structural, that map at least partially onto modern Western distinctions between “religious” and “non-religious” or “secular” domains can be found in many cultural settings. Researchers associated with this project have begun to identify historical candidates for such analogous distinctions, such as the distinction in much of East and Central Asia between the “nomospheres” of the ruler and the Buddha (Kleine and Wohlrab-Sahr 2021, pp. 58, 61–62) or the distinction in some Islamic societies between shari‘a and adab (Salvatore 2018). Such emic distinctions in various cultures help explain why, in the process of encounter with the West and the establishment of modern legal regimes that referenced concepts like “religion”, certain phenomena (e.g., Buddhism in Japan) were easily classified as “religions” while others (e.g., Shinto in Japan) had a more ambiguous status (Kleine and Wohlrab-Sahr 2021, p. 57).

The multiple secularities research convincingly pushes back against extreme versions of the critical religion approach that would, for example, reject the possibility of comparisons between Western and other societies and their categories. It is also generating careful comparative research on historical and contemporary conceptual and structural distinctions between social domains in many societies and the ways in which those distinctions have been mutually influential—especially how they have shaped contemporary global appropriations and rejections of originally Western notions of religion and secularity. It would be premature to conclude, however, that this line of research has established that notions of secularity closely equivalent to modern Western ones are common in the historical record prior to the era of global interaction. The directors of the multiple secularities project regard it is an open question, in fact, whether it is appropriate to use the term “secular”, with its substantive connotations of distancing from religion, for historical circumstances which may even lack a corresponding terminology (Kleine and Wohlrab-Sahr 2020, p. 24). In recent work they explicitly reject the “naïve universalist stance…that the differentiation of religion as a specific social domain is an anthropological given” and state “we do not claim that ‘religion’ or ‘the secular’ are universal categories” (Kleine and Wohlrab-Sahr 2020, pp. 51, 65).

What are the implications of this discussion for the argument of this article? If critical scholars are correct that Western-style religious-secular distinctions were not present in most societies prior to engagement with the West, the explanatory use of the differentiation story is forced into the position of arguing that these non-Western societies had simply not yet experienced the differentiation process that is necessary to disentangle religion and the intrinsically secular spheres. On that reading, the experience of non-Christian societies reflects the undifferentiated character of those societies, not any underlying problem with the differentiation story. It is true that the absence of religious–secular distinctions or fully developed secular spheres in such societies does not disprove the claim that there is a universal social reality (the intrinsically secular character of most social spheres) that until recently was only recognized, named, and actualized in the West. But the claim that modern Westerners alone reached correct conclusions about the true character of religion, politics, economics, and so on should at least raise suspicion about the universal validity of these conclusions. In other words, if politics, economics, and education are intrinsically, rationally, or properly directed to purely nonreligious ends and if an inexorable social logic leads in this direction, why did a similar secular/religious split not emerge in other highly urban, literate, organizationally complex societies in the early modern era, such as those of China, India, and the Middle East? Certainly it was not for lack of wealth, philosophical acumen, or administrative sophistication.

Here, the multiple secularities perspective would argue that there may indeed have been differentiations, conceptually and structurally, between domains that could appropriately be termed “religious” and “nonreligious” in some of these societies before the encounter with the West. It must be kept in mind, however, as these scholars themselves and their critics (e.g., Horii 2018a) point out, that such distinctions cannot be assumed to match the modern Western dichotomy. Taking the example of East Asian distinctions between the jurisdictions of the monarch and of Buddhist monasteries, it is easy to designate the former as “secular” because it is clearly distinguished conceptually and structurally from another domain that registers as obviously “religious” from a Western viewpoint, but the risk is that we then fail to notice the ways in which such ostensibly secular rulership is often related in such contexts to spiritual beings, cosmic norms, and superhuman powers that would definitely not be considered “secular” in a modern Western context. Indeed, by far the typical human experience seems to have been that states and rulers, despite being conceptually and structurally distinguishable from (for example) monasteries and priests, have been no less “religious” for all that, given that they have presented themselves as possessing superhuman, heavenly legitimation or even (semi-)divine status. There is a sense in which an emperor who claims descent from a goddess, rules by virtue of the mandate of Heaven, or performs rites to honor deities on behalf of the people can be described as falling on the “secular” or “non-religious” side of an endogenous binary distinction in his society, but it is a type of secularity with non-trivial differences from the secularities seen in modern Western political thought and practice. In the latter case, “secular” means precisely the exclusion of superhuman considerations.

Are there, then, innately secular spheres in the substantive, modern Western sense of “secular”? The preponderance of evidence still seems to suggest that the notion assumed in explanatory versions of the differentiation story—that some social spheres are intrinsically secular such that considerations related to divine beings, spirits, sacred texts, or superhuman moral orders are properly out of place in those spheres and that structural differentiation thus necessarily “secularizes” them as they are liberated from religious control—is a peculiarly modern and Western notion. Combined with the preceding two critiques—the fuzziness of the concept of “spheres” and the lack of justification for the concept of secular value-spheres—this observation lends plausibility to an alternative view: that the emergence of the type of distinct “secular” and “religious” spheres seen in the modern West (conceptually and in practice) is a product of unique, contingent historical developments in Western societies rather than the actualization of universal, necessary truths about social reality.

4. Conclusions

Since at least the 1960s, many of the leading scholars of secularization, from all major schools of thought, have been making similar assumptions about the historical relationship between differentiation and secularization. These scholars invoke a shared story in which a long-term historical trajectory of differentiation has separated various social spheres from each other, resulting in a differentiation between “religion” and the “secular” spheres like politics and education. This differentiation story can take on a purely descriptive form, which merely attempts to describe the experience of societies which have come to see certain aspects of life as secular and distinguish them from religion. It often carries a stronger explanatory claim, however, that there is an underlying, innate distinction between religion and the other spheres, such that the growing autonomy of these other spheres freed them from the alien influence of religion and allowed them to pursue their rationally proper secular purposes, thus contributing to the secularization of society.

Although the differentiation story seems like common sense to many Western scholarly observers, if we make it the focus of critical attention, its empirical accuracy and theoretical plausibility look less certain. Its accuracy as a generalization about Western history is called into question by major phenomena that do not fit the narrative of increasing structural and value differentiation over time, including the growth in the scope of the state, the convergence of many spheres on the value of individual autonomy, and the historical experience of the university as a particularly influential social institution. On the theoretical level, the apparently simple notion of social spheres is plagued by a lack of conceptual clarity that is difficult to resolve and that undercuts the explanatory value of the differentiation story as a whole. The notion of secular value-spheres, which sees most spheres as necessarily rationally oriented to nonreligious purposes, likewise lacks solid justification, which in turn compromises explanatory uses of the differentiation story. Finally, the historical absence of a substantive secular/religious split like that seen in the modern West in most societies lends plausibility to the idea that the type of differentiation that emerged in Western societies is a contingent product of unique historical circumstances rather than the necessary consequence of the underlying essence of social spheres. The time is ripe to put aside uncritical invocations of the differentiation story and reconsider the relationship between differentiation and secularization, employing more rigorous concepts and drawing on a historically and culturally broader range of data.

It is beyond the scope of this article to propose what an alternative to (uncritical versions of) the differentiation story would look like, but it is possible to sketch out some suggestions for future directions for research and theorizing on this topic. First, future discussion of differentiation by scholars of secularization needs to employ more clarity and precision in its use of sphere-terms and refine generalizations about differentiation with more empirical attention to historical and cultural variety. On the theoretical side, comparatively recent alternative approaches to understanding the differentiation of social domains should be taken into account (e.g., Luhmann 1997, 2013; Bourdieu 1993; but see also the critical perspective of Taylor 2020). In any case, scholars should consistently distinguish between structural and conceptual differentiations, on the one hand, and value-orientations, on the other, remaining open empirically to many different configurations rather than assuming that a shift in one (e.g., a new structural differentiation) necessarily leads to a shift in the other (e.g., a new value-orientation). Such an approval is more compatible with the multiple modernities paradigm than the classical modernization theory assumption of a single developmental trajectory of all societies.

Second, taking a cue from the multiple secularities research and bringing together several observations from this article, future work should recognize that while differentiating practices and concepts may exist in many societies between domains that obviously code as “religious” in a Western framework (e.g., monasteries) and those that do not (e.g., governing and agriculture), such distinctions do not necessarily imply that the “non-religious” domains are “secular” in the substantive modern Western sense of being predicated on what Taylor (2007) would call exclusively humanist assumptions. In other words, the emergence of differentiations in any given society between domains that the modern West considers religious and other domains does not necessarily imply that the latter domains will therefore operate without reference to (for example) divine powers, cosmic moral frameworks, or sacred texts. Indeed, even in earlier periods of Western history clear distinctions between ecclesiastical and political authorities were quite compatible, for centuries, with political authorities seeing their task in deeply religious terms. As an analytical concept “religion” can meaningfully and appropriately refer at the same time both to particular differentiated domains of social life and to an overarching system of beliefs and values that still fully encompasses and directs activity in the other social domains. In brief, “religious” should not be assumed to be coterminous with “ecclesiastical”; both church and state (for example) may be highly “religious” even in settings where they are clearly differentiated.

Finally, in addition to mapping differentiating practices in various cultures and periods and comparing those to the Western religious–secular binary, as the multiple secularities project is doing, it might also be illuminating to look for potential similarities and differences across cultures in terms of integrating concepts apart from Western ones. For example, Philip Rieff argued (Rieff 2006, 2007; Zondervan 2005) that it is a near-universal feature of human cultures that they attempt to “transliterate” sacred order into social order. While Rieff’s elaboration of this idea was idiosyncratic, unsystematic, and not thoroughly substantiated, this framing is potentially quite promising as a way to draw intercultural comparisons of how phenomena conventionally considered “religion” relate to various social domains without relying on the conceptual category of “religion” as such. For example, as noted in a few places above, concepts and practices that subordinate a broad range of social domains to some notion of morally authoritative, superhuman cosmic orders and powers seem very widespread across periods and cultures, including in societies that are internally differentiated. Suggestive examples of points of comparison among many include concepts of cosmic law (dharma, ma’at, ius naturale, and some interpretations of Dao) as well as practices invoking divine or Heavenly blessing in non-priestly, non-monastic, non-ecclesiastical settings.13 Provided that investigation of such questions proceeds with an eye on both theoretical formulation and empirical complexity, open to both similarities and differences, this framework could be helpful in comparing situations in which internal conceptual and structural differentiations coexist happily with overarching notions of sacred or cosmic order—as in, perhaps, premodern East Asia—and situations in which such differentiations result in the rejection of sacred/cosmic order as guiding principle in some of the differentiated domains—as in the modern West.