Exploring further debates in the secularization paradigm. Debate on Jörg Stolz’s article on Secularization theories in the 21st century: ideas,evidence,and problems

The ultimate causes of secularization

Stolz spends a large part of his address exploring theories of what is driving the decline of religiosity indicators in many countries over the past decades and centuries. More specifically, Stolz examines four of these potential root causes of secularization: increased existential security, the propagation of higher education, the multiplication of sources of secular competition, and growing pluralism. Other factors also discussed in the existing literature include heightened values of individualism, choice and antiauthoritarianism, as well as processes of de-traditionalization and global mobility. Much recent quantitative research has used international datasets (such as the International Social Survey Programme, World Values Survey, and European Values Study) and hierarchical linear modelling to measure country-level characteristics when it comes to these factors, and how well they associate with levels of (non)religiosity. Stolz provides notably the example of Höllinger and Muckenhuber (2019) who show a correlation coefficient of –.66 when it comes to their aggregate religiosity scale and the Human Development Index (HDI), based on 2010–2014 World Values Survey (WVS) data. Höllinger and Muckenhuber see in this and their other findings support at the macro level for development and greater existential security driving religious decline.

However, there are many problems with trying to identify the causes of secularization in this way: notably, problems of spurious correlation, of small sample size at the country level, and of ecological fallacy. Since the religiosity indicators measured by these international datasets (such as religious affiliation, frequency of religious service attendance, belief in God, and salience of belief in life) score much lower in many parts of Europe, any aggregate indicator that is more commonly found in Europe is going to be highly correlated with lower levels of religiosity. For example, when I measure the association between Höllinger and Muckenhuber’s country-level composite religiosity scores and national average annual temperature ¹ with the same group of countries used in their 2019 study, I get a correlation coefficient of .63. Based on this finding, I could then argue that colder climates lead people to believe that there is no God because, let’s face it, why would such a being create a winter during which one cannot feel their own face when walking outside. Or I could come to the more boring but sounder conclusion that indicators trending in Europe (such as lower temperatures, compared with the global South) are going to be associated with lower religiosity. Many of these associations will be spurious, and this cross-sectional data will not usually be able to identify the spurious from the causal. Try and control for all factors in your models, and you quickly run into problems of measurement, multicollinearity, and small sample size. Stolz does not explore this critique in any real detail in his address, but he does hint at its existence:

[. . .] the high correlation of insecurity and religiosity on the country level is highly confounded with other variables that also correlate with religiosity: GDP per capita, urbanity, literacy rate, educational attainment, access to mass media, and others [. . .] the data permit various other interpretations besides the insecurity account.

The small sample size of countries in many of these international datasets is also an important issue. In the Höllinger and Muckenhuber study, the number of cases included at the country level is 49. That is a small sample size on which to base conclusions. It also makes the data sensitive to which countries were measured and which were not. As an example, the WVS data that Höllinger and Muckenhuber use do not include the Czech Republic: a country with one of the lowest levels of religiosity in the world, but only ranked 27th for HDI in 2017. ² This may have a significant impact on measured associations. I do not want to single out the Höllinger and Muckenhuber study here either: all the multilevel quantitative studies that Stolz cites in his address, including my own, grapple with these issues.

Most of these studies also make the assumption that factors tied to between-country differences are also the same factors that would impact secularization trends in one country over time (existential security for example), an assumption tied to the secular transition framework that has yet to be systematically tested. Is the secular transition process set off by only the one same factor, or could it begin in different contexts at different times for a number of different reasons? Stolz as well as Höllinger and Muckenhuber also point to another limit when it comes to the findings on the causes of secularization from these multilevel models: that relationships between aggregate-level variables at the macro-country level often cannot be reproduced at the micro-individual level. There is an association between HDI and lower religiosity at the country level, but it is not necessarily the most materially secure individuals within a given society who are the least religious.

I wholeheartedly agree with Stolz’s message throughout the address that social context is crucially important to understanding (non)religiosity trends and effects, and that hierarchical models with international data have and continue to push our understanding further on the impact of socio-environmental factors. However, these types of studies will also not be resolving the ‘What exactly causes secularization?’ debate anytime soon. Strong theory remains crucial to this enterprise.

Individualization and secularization

In many parts of his address, Stolz heavily criticizes rational choice/religious market theories, the secularization paradigm’s main foe since the 1980s. I share this disapproval of rational choice theory, and have much evidence to support critiques of this theory in my own work. Yet, there is little mention of another source of argumentation made against the secularization paradigm, perhaps more popular in recent years: individualization theories as named by Stolz when he does briefly mention them. In contrast to the secular transition framework, individualization theories (or spiritualization/spiritual but not religious) argue that religion and spirituality are changing, rather than declining (see, for example, Ammerman, 2014; Davie, 1994; Fuller, 2001; Heelas and Woodhead, 2005; Houtman and Aupers, 2007; Roof, 1999). Instead of narrowly defining and measuring religion against conventional institutional markers, such as church attendance or other communal-oriented religious activities, the individualization framework stresses ongoing private spirituality among most individuals. According to this framework, a vast majority of individuals draw on a number of beliefs, rituals, and practices from a variety of sources, some of them linked to religious groups and some of them not, to build and maintain their own personalized spiritual and meaning systems away from church doors. These more personalized spiritual and meaning systems may infuse and inspire special activities (such as meditation or prayer), as well as the ordinary aspects of life at home, at work, in relationships, and in health matters (Ammerman, 2014). This framework understands the trend of declining religion then as a shift to more innovative and individualized forms of believing and spirituality in people’s quest to continue to answer their fundamental spiritual needs.

There has been very limited quantitative research so far testing the secularization and individualization frameworks against one another. This also reflects a wider issue of limited survey measurement: most existing measured religiosity indicators focus on belief, behaviour, and belonging tied to organized religion (notably Christianity), and not on less traditional forms of individual spirituality. In the conclusion of his address, Stolz makes the bold statement that ‘Alternative spirituality may be a form of intermediate religiosity that equally undergoes a trend to ever more secular forms’. This may very well be the case, but so far, there has been little empirical evidence produced to support this claim. Much more research is needed to test the secular transition hypothesis against the individualization one in Western societies.

Substantive elements of nonreligion

The quantitative secularization studies that Stolz reviews in his address are all what philosopher Charles Taylor (2007) refers to as subtraction stories. Their focus is on showing the decline of key religiosity indicators and its potential causes, which is an important endeavour. However, Stolz does not mention another growing body of research focused instead on the substantive elements of nonreligion: studies that examine how religion is translated into or replaced by new values and worldviews in more secular societies, the content of these forms of nonreligion, and how (non)religiosity’s impact on socio-political attitudes and behaviours shifts in areas where active religious adherents form a minority (notably when these active adherents are largely coming from non-Western immigration).

As one example, Stolz explores the literature showing that an absence of religious socialization often leads individuals to have no religion as adults. Yet, irreligious socialization is not simply a similar socialization as generations past minus religion. It is an alternative way of raising children, often favouring a hands-off approach to parenting when it comes to religion, spirituality, and nonreligion, leaving learning and choice of these matters up to the child. As Zuckerman et al. (2016: 217) note, ‘Secular people tend toward nonconformity, independence, and antiauthoritarianism [. . .] to base their maturational goals on personal independence, and their childrearing philosophy emphasizes autonomy rather than obedience to authority’. Modelling this ‘hands-off’ approach to religion in the home does, in fact, pass on a particular individualist and secular orientation to religion and the world more generally. Such an approach strengthens the likelihood that someone who identifies as nonreligious will raise children who also say they do not identify with a religion.

There is a whole new subdiscipline of nonreligion and secularity studies emerging, led notably by researchers such as Lois Lee and Phil Zuckerman. This emerging subdiscipline is inspiring a new generation of scholars interested in showing not just the decline of religion in many societies but also the impact of this decline and the new forms social phenomena have subsequently taken.

Concluding remarks on the risks of grand theories

I will conclude my comments here by first saying that I greatly appreciate the secular transition framework. I appreciate that it has added much-needed complexity to many of the neo-classical secularization theories of the past, notably by highlighting the importance of context in the religious decline process, as first proposed by David Martin (1978). For these reasons, I use this paradigm extensively in my own work. This said, I also caution against making some of the same mistakes as our neo-classical forebearers. Many secularization theorists have made large and overstated claims beyond what the data have shown them, arguing that secularization is a worldwide process affecting all countries as they modernize, that there is only one process that everyone goes through at the same speed once begun, and so forth. These overstated claims arguably led to the contentious phase in the 1980s and 1990s as Stolz names it, a phase which in turn has set the secularization paradigm back many decades in my opinion, and that we are still grappling with today. Many scholars, notably in the United States and in the discipline of religious studies, consider secularization theories as illegitimate due to these overstated grand claims.

Figures 4 and 5 that Stolz reproduces in his address provide important ideas and data regarding the secular transition, but the temptation is to jump to more simplistic overstated conclusions from there about secularization. I would instead encourage all of us to keep building complexity into these models, to better reflect and advance our understanding of the social world. Does the idea of a secular transition, begun at different times, but then subsequently experienced by countries in a similar way, hold up when we add more measures of religious and spiritual phenomena to the model? Or do we instead start to see different national and regional configurations emerge when it comes to secularity? What drives important exceptions to the secular transition trend, such as in France and East Germany (that Stolz points to in his address) as well as in Russia? Stolz does an excellent job of reminding us of this important complexity in his address, notably with his micro–macro model in Figure 9, and we can use this address as an impetus to continue rich and diverse research within the secularization paradigm.