Ideal Types, Semantic Anarchy, and the Study of Atheism (etc.)

By Christopher R. Cotter, in response to an interview with Chris Silver on “Atheism, New Religious Movements, and Cultural Tension”. Listening to Chris Silver’s recent podcast on Atheism, New Religious Movements, and Cultural Tension was a thoroughly pleasant experience. I enjoyed hearing a colleague who I first met in 2011, and who quickly came on board the nascent RSP team (as interviewer, editor, writer, and more), taking his well-earned place on the ‘other’ side of the microphone. Like Chris, I recognize the problems associated with treating atheists, agnostics, ‘nones’, etc., as distinct groups with coherent attitudinal correspondences. I’ve also used grounded theory and social constructionism in my approach to ‘non-religion’ and recognize the issue of one’s work being taken out of context. I welcomed hearing of broad-ranging, quantitative work being carried out within the psychology of religion with an eye to debates in the broader, critical study of religion, and the new developments in his work relating to status loss, New Religious Movements, and more. Despite the many positives of the approach outlined by Silver,  critical problems with it—exemplified in my own work, too—concern the construction of ideal types, and the proliferation of idiosyncratic terminology. In what follows, I’ll discuss each of these issues in turn before proposing a discursive solution.

Ideal Types

Referring to the ‘insubstantial’ secular, Lois Lee pithily observes that ‘it is not possible to organize absence into types’ (2015, 51). However, this isn’t the case with substantive understandings of ‘non-religion’ and in 2011 I set out with similar ambitions to Silver, producing a typology of non-religion in my Masters dissertation (Cotter 2011, 2015). This project was conducted amongst the undergraduate student body of the University of Edinburgh taking a grounded theoretical approach which elicited narratives via electronic questionnaires and qualitative interviews. Two of the key insights suggested were that:    
  • The ways in which students negotiated (non-)religious terminology throughout their narratives allowed the development of five ideal types which were seemingly independent of established religious categories: naturalistic, humanistic, philosophical, familial, and spiritual.
  • Regardless of the salience of the students’ (non-)religious identifications, they appeared to be keenly aware of where they stood when religion or non-religion were perceived to interact with what mattered to them
A significant reason I have felt unease with these insights as they stand is my somewhat lofty and obfuscating attempt to provide an exhaustive ideal-typical account of non-religion. Ideal Types are commonly understood as ‘analytical tools to be used to facilitate comparison’ (Barker 2010, 188 fn. 2), as ‘pragmatic constructs’ which can in no way ‘be regarded as essential categories or ontological realities’ (Cox 2006, 83). Recent relevant examples would be Lois Lee’s five types of ‘existential culture’ (2015, 161–72), John Gray’s Seven Types of Atheism (2018)Silver and colleagues’ ‘six types of non-belief’ (Silver et al. 2014), or my own five-fold typology of ‘non-religion’. Even when writing my dissertation, I was clearly uneasy concerning the correlation between these types and my data:

“These narrative-based types cannot be assumed to be constant, and must be understood as firmly rooted in the context in which they were revealed. However, they reflect what individuals actually say, and give priority to individual self-representations, providing a much ‘truer’ representation of what (non)religion means to these individuals than wide-ranging, quantitative, typologies which suffer from the same contextual constraints.” (2011, 76)

Although I, and most others who employ ideal types, have never claimed that classifying individuals is a simple matter, and have emphasized that there is much overlap between types, it is almost invariably the case that their artificial and constructed nature becomes lost in translation (see Silver’s comments about misrepresentations of his work in the media), giving the false impression that individuals can be easily boxed off into discrete types. Whilst such work is valuable for macro-level analysis, it fundamentally breaks down at the level of the individual where heterogeneity and contextuality abound: there is ‘no such thing as a perfect or ideal-typical form’ of difference to ‘religion’ (Lee 2015, 44).  ‘Religious, spiritual, secular, and non-religious identities are not stable, unitary formations’ (Hoesly 2015), but rather what Jean-François Bayart refers to as ‘operational acts of identification’ (2005, 92). However, in relation to the study of ‘Atheism’, Ethan Quillen argues that prevalent ideal-typical generalizations are perhaps ‘nothing more than a product of the current scholarly study of Atheism’s predominant focus on the social-scientific attempt at making sense of “Atheism-in-general,” rather than “Atheism-in-specific”’ (2015a, 30), the ‘attempt at finding an identity in the numerous applications of an ambiguous word’ (Baird 1991, 11).

Semantic Anarchy 

Going further, these problems are clearly connected to issues surrounding terminology. The contemporary situation has been described as verging on ‘a situation of semantic anarchy, in which individual scholars work with idiosyncratic definitions’ (Jong 2015, 19). Indeed, Quillen argues that the whole discourse on ‘types of atheism’, or indeed the very terminology of ‘ir-religion,’ ‘un-belief,’ ‘non-religion’, etc., is ‘not unlike that which complicates the definition of “religion”’ (2015b, 132), and falls foul of much of the critique levelled at the term (see RSP podcasts with Timothy FitzgeraldRussell McCutcheonJames CoxBrent NongbriTeemu Taira and others). In addition, ‘The conceptual balkanisation that results from the proliferation of idiosyncratic definitions makes […] fruitful collaboration more difficult’ between scholars doing empirical work in different contexts (Jong 2015, 19). We need not despair, though. This situation of proliferating ideal types and idiosyncratic terminology can be remedied by scholars being ‘vigilantly specific about the aspect of “nonreligion” that they are interested in’ (Jong 2015, 20), restricting themselves to very particular contexts—historical, textual, ethnographic, etc. Alternatively, my preferred route is to take a discursive approach.

Discursive Approaches

As I’ve previously argued, we can fruitfully 

“conceptualize non-religion as part of a religion-related field comprising ‘all phenomena that are generally (or according to a certain definition of “religion”) considered to be not religious, but stand in a determinable and relevant relationship to the religious field.’ This relationship can take the form of criticism, competition, collaboration, mirroring, functional equivalence, interest, etc. Such an understanding sees […] ‘descriptions, claims, reports, allegations, and assertions’ about non-/religion [as] the topic of the analysis, rather than ‘religion’ or ‘non-religion’ themselves. Thus, we arrive at a critically-engaged, relational concept of ‘non-religion’ which can be operationalized empirically in a non-stipulative manner and which emphasizes that religion need not be a dominant, normative, or positive term in the contexts studied.”

Although discursive questions were not the driver for my Masters project, it can clearly be interpreted as having focused on the way ‘religion is organized, discussed, and discursively materialized’ (von Stuckrad 2010, 166) in a particular context, by individuals who self-described aspects of their individual practice, beliefs, attitudes and/or identity as different from their subjective self-definitions of religion. It can be viewed as discerning a range of discourses, which could be classified as spiritual, familial, philosophical, humanistic and naturalistic, surrounding a variety of negotiated phenomena—identities, practices, attitudes, beliefs—in a field of discourse with boundaries dictated by the logics of the research project, i.e. substantiating the ‘non-religiosity’ of Edinburgh students, in 2010–11, whose self-descriptions were ‘non-religious’, and so on. Returning to the two key insights of the project, this discursive re-reading allows them to be reframed as follows:  
  • In these narratives, these students primarily invoked five types of discourse when they engaged with topics related to religion, non-religion, and related categories. These discourses appeared to operate at a level independent of the specific terminology—the discursive objects—in question. 
  • Regardless of the salience of these discourses in individuals’ lives, they were invoked when the students were confronted with phenomena that were deemed to be related to—i.e. which meaningfully intersected with—the field of discourse on religion, non-religion, and related categories. 
This move from types of atheism/non-religion to types of discourse shifts the focus to ‘the social effects of the way people talk, rather than the apparent meaning of their words’ (Martin 2017, 104). People employ multiple discourses, situationally; they can say things in many different ways, depending on the discursive resources available to them in a particular cultural context.  A shift in focus from the individual to the discourse they employ, from the person to what they say, and how they say it, allows the individual to be incorporated analytically into the wider societal conversation of which they are inherently a part. As Bayart argues, no speech act can be attributed to a single individual, but is a product of the broader complex social situation in which it occurs, as well as its historical context: ‘every utterance is related to earlier utterances’ (2005, 112). There is much more I could say here, and I’m not claiming that discursive study is the only way to study ‘non-religion’ critically. However, it is one way in which we can usefully sidestep some of the issues associated with the construction of typologies in the social sciences. Whether this helps us to avoid the ‘misinterpretation’ of our work when it takes on a life of its own is another matter. One of many Challenges and Responsibilities for the Public Scholar of Religion.


  • Baird, Robert D. 1991. Category Formation and the History of Religions. 2nd ed., with A new pref. Religion and Reason 1. Berlin: Mouton de Gruyter.
  • Barker, Eileen. 2010. “The Church Without and the God Within: Religiosity and/or Spirituality?” In The Centrality of Religion in Social Life: Essays in Honour of James A. Beckford, edited by Eileen Barker, 187–202. Farnham: Ashgate.
  • Bayart, Jean-François. 2005. The Illusion of Cultural Identity. London: C Hurst & Co Publishers Ltd.
  • Cotter, Christopher R. 2011. “Toward a Typology of ‘Nonreligion’: A Qualitative Analysis of Everyday Narratives of Scottish University Students.” Unpublished MSc by Research Dissertation, Edinburgh: University of Edinburgh.
  • ———. 2015. “Without God yet Not Without Nuance: A Qualitative Study of Atheism and Non-Religion among Scottish University Students.” In Atheist Identities: Spaces and Social Contexts, edited by Lori G. Beaman and Steven Tomlins, 171–94. Dordrecht: Springer.
  • Cox, James L. 2006. A Guide to the Phenomenology of Religion: Key Figures, Formative Influences and Subsequent Debates. London: Continuum.
  • Gray, John. 2018. Seven Types of Atheism. London: Allen Lane.
  • Hoesly, Dusty. 2015. “‘Need a Minister? How About Your Brother?’: The Universal Life Church between Religion and Non-Religion.” Secularism and Nonreligion4 (1). 
  • Jong, Jonathan. 2015. “On (Not) Defining (Non)Religion.” Science, Religion and Culture2 (3): 15–24.
  • Lee, Lois. 2015. Recognizing the Nonreligious: Reimagining the Secular. Oxford: Oxford University Press.
  • Quillen, Ethan G. 2015a. “Discourse Analysis and the Definition of Atheism.” Science, Religion and Culture2 (3): 25–25.
  • ———. 2015b. “Everything Is Fiction: An Experimental Study in the Application of Ethnographic Criticism to Modern Atheist Identity.” Unpublished PhD Thesis, Edinburgh: University of Edinburgh.
  • Silver, Christopher F., Thomas J. Coleman III, Ralph W. Hood Jr., and Jenny M. Holcombe. 2014. “The Six Types of Nonbelief: A Qualitative and Quantitative Study of Type and Narrative.” Mental Health, Religion & Culture 17 (10): 990–1001.
  • Stuckrad, Kocku von. 2010. “Reflections on the Limits of Reflection: An Invitation to the Discursive Study of Religion.” Method & Theory in the Study of Religion 22 (2): 156–69.

How Meanings are Made and Taken Apart: Reflections on Discursive Analysis

In an interview with the Religious Studies Project, professor Kocku von Stuckrad outlines interesting possibilities for discursive analysis. He describes an approach that “goes beyond terms” and also beyond examining political power structures. The interview brought up many important, broad themes that are discussed in the study of religion. This essay is an examination of some thoughts the interview brought up and provoked, also in relation to some practical realities of the academic world.

Discursive analysis has become an important theoretical approach in the study of religion. Seen through the discursive lens, religion is a concept that is being used by different people in different settings in a number of ways. The content of any concept is always changing, always negotiated and contested. Still, there is some room for confusion. This is, in part, because discursive analysis is not exactly a unified approach but rather a collection of approaches.

The type of discursive analysis von Stuckrad speaks of does not only include texts and usage of certain terms, (i.e. how different terms and themes are linked to one another so as to produce knowledge), but also includes institutionalised and materialised products of this knowledge. Von Stuckrad refers to “discourse of practices”, which definitely is a welcome link between language and the material reality, acted and experienced.

This approach goes beyond certain styles of critical discursive analysis, but power relations are not forgotten. As one becomes more aware of how academic knowledge, for example, inevitably shapes the discourses on almost any given theme, and these discourses in turn may shape or create actual practices and institutions, it becomes evident that scholars may actually hold a tremendous power. The next responsible thing to do is to turn a critical gaze to our own institutional links and what kinds of “knowledge agreements,” discursive compounds, we, together with our research, are standing on. As scholars of religion, or of any other subject for that matter, we should pay critical attention to our own position. When we as researchers pick up a concept and use it, we must be aware just how far from sterile, self-evident and unpolitical they are. They come with underlying assumptions, a whole history of negotiations and selection processes.

We must also be aware of how our participation in certain discussions may shape the world around us. In our view, this does not mean that researchers should shy away from these discussions, but that they should enter them understanding the possible weight. Academic knowledge, or language at the very least, will leak into the surrounding society one way or another. Studying topics under some political crossfire can especially attract expectations. For example, studying the Church of the Flying Spaghetti Monster has gained attention from groups and individuals that promote the cause of this movement, and it is entirely possible that all the scholarly attention given to the movement will be pointed at as evidence that it is a movement to be taken seriously also in political debates – this has already been seen in the case of some neopagan groups.

A whole other problem is the fact that scholars from many fields, often cultural studies, have in recent years become quite aware of many possible audiences to their research, not all of which are of the good kind. We can only hope that hate mail and anonymous threats are not a growing phenomenon. And that this sort of publicity will not drive researchers away from the public sphere.

Examining these usages and often especially the power relations – who has the right to define the content of a concept – is at the heart of discursive analysis. And keeping an eye on power relations is more or less a necessity if one wants to dig into how concepts have evolved in time. The very concept of religion is a great example of the historical, setting-specific nature of language. All the more illuminating is to think about how the concept carried its Judaeo-Christian underpinnings into academic research and was used to conceptualize cultural systems that had no such concept in their own reference system.

When scholars start paying critical attention to their particular position and the load of their concepts and ideas, research becomes a consciously two-way process. In order to adequately examine the subject of our research, one must also take a good look at one’s own instruments. We must know them well in order to know what kind of information they can offer us about our subject. This sort of critical perspective should more or less go hand in hand with all research, not only studies that explore discourses.

As for practical applications, there are probably many different ways in which the genealogical point of view von Stuckrad suggests can be incorporated in actual individual research projects. As he points out in the interview, not all research projects need to be discursive analyses. Within a broader framework of discursive understanding, a wide range of methods can still be applied. Still, the discursive reflection should be described in the actual research. As researchers are always making decisions from a particular point of view, they should make an effort to make themselves more visible in the research. Apart from reflecting critically on one’s position and terminology, for example, it is important to report these processes. Only then can the reader examine the way the researcher has reached his or her conclusions.

But what exactly would be the most constructive way to incorporate this reflection in research papers and reports? We have heard warnings about using the chapter titled ‘reflection’ for pouring out all sorts of affiliations, engagements and other caveats, then going on with our research without giving these questions a further thought. This is hardly the kind of critical thinking we are looking for. Another question is, might there also be a risk that research papers become more massive and complex as more of the process is made visible in writing? Simultaneously, other kinds of demands are on the increase in the academia, such as writing as clear, succinct, and reader-friendly academic papers as possible. More transparency, fewer words. Luckily, we at least see developing academic writing further as a meaningful challenge.

There’s More Than One Way to Skin a Cat: Comments on Kocku von Stuckrad’s Discursive Approach

There is more than one discursive approach in religious studies. In his interview with the RSP, professor Kocku von Stuckrad outlines some of the key issues that are relevant for constructing a discourse theoretical framework for religious studies. As I have tried to contribute to a similar framework in some of my publications, it comes as no surprise that my overall approach to von Stuckrad’s interview is highly appreciative. Therefore, I shall simply add some nuances to the interview by demonstrating how von Stuckrad’s approach could be located within the field of discursive approaches as being only one, though very relevant, version.

Von Stuckrad begins to delineate his approach by distinguishing it from more linguistic discursive approaches, including so-called critical discourse analysis (CDA) (represented by Norman Fairclough and Ruth Wodak, for instance), and talks about a ‘sociology of knowledge’ approach. The distinction itself is pedagogically useful, although we are not really talking about two different approaches, but rather a variety of approaches in a continuum – some are more linguistically oriented, designed for detailed analysis of texts, whereas others are oriented more historically, designed for thinking about large-scale historical organization of a plurality of knowledge through discourses (represented by Michel Foucault, among others).

I would even suggest that despite the obvious and deep theoretical differences and disagreements between Fairclough and Foucault (see Taira 2016a; 2016b), the choice between them – if they are taken to represent two common options – often depends on the research design. In my experience, it is often the case that students prefer linguistically oriented versions simply because they find it easier to apply to the analysis of texts. Indeed, in this context it would be better to refer to Fairclough’s model as ‘textually oriented discourse analysis’ (TODA), as this properly highlights the difference with a Foucauldian approach. To simply talk about ‘critical discourse analysis’ does not convey the difference, because ‘critical’ could be applied to both Fairclough and Foucault (despite the fact that Fairclough’s understanding of what counts as ‘critical’ is not the same as Foucault’s).

Von Stuckrad correctly emphasizes that discursive research is not simply a study of a particular word or text, but a study of practices and institutions. As I see it, this applies – or at least should apply – not only to the ‘sociology of knowledge’ approach von Stuckrad prefers, but also to critical discourse analysis.

Research Perspective, Method, or Both?

The interview does not say much about how to use discursive methods. Indeed, von Stuckrad highlights the historical narratives in how discourse on religion operates and prefers to write about a ‘discursive research perspective’ (von Stuckrad 2014: 15) rather than discourse analysis as a method. It is telling that in The Scientification of Religion (2014, 18) he lists eight methods appropriate for (his version of) discursive study, but he does not mention discourse analysis. I agree with him that, as a theoretical framework, a discursive approach can utilize many methods, depending on the research design, but discourse analysis is certainly one of them.

There are other examples within the discursive study of religion in which the methodological dimension of discourse analysis is addressed in a ‘how to do’ manner. One is mine (Taira 2013a) and the other – with a slightly different emphasis – is written by Titus Hjelm (2011). This is also a good opportunity to advertise a forthcoming volume edited by Frans Wijsen and Kocku von Stuckrad (Wijsen and von Stuckrad 2016). Its contributions, including original chapters by von Stuckrad, Hjelm and myself, among others, deal with many of the details von Stuckrad mentions in the interview, but they also help us to situate von Stuckrad’s approach as only one useful opportunity among a diversity of options with different nuances.

One of the interesting aspects in the interview is that, given that the method part is neglected, von Stuckrad talks about epistemology. As is often the case, the discussion uses – perhaps in order to bring in the listeners – the traditional terms and distinctions, such as realism versus relativism (often realism versus anti-realism). Von Stuckrad even says, approvingly, that American philosopher Richard Rorty provides a very intelligent form of relativism.

The problem is that practically no one self-identifies as relativist; it is a label addressed by opponents. This is also the case within discursive approaches: some versions cling to a realist standpoint and criticize others as relativists. This applies to Fairclough’s criticism of Foucault and I guess some might say the same thing about von Stuckrad. My question is, then, should those whose versions are accused of being relativist stop talking about realism and relativism and invent new ways to talk if the standard terms do not do justice to their position? It is true that Rorty wrote about himself as a ‘so-called relativist’, but he never seriously defined himself as such, simply because the term is embedded in a discourse that he wishes to discard. As Rorty (1999, xviii) stated: ‘We must repudiate the vocabulary our opponents use, and not let them impose it upon us.’

Are analytical definitions of religion necessary?

In the interview, von Stuckrad talks about three scholarly debates relevant for discursive research: crisis of representation, situated observer, and essentialism. These complex and significant issues have been addressed carefully in many disciplines in recent decades. However, it seems to me to be perfectly possible for someone to agree on the problem of representation, highlight the importance of reflecting on the situatedness of observer, challenge essentialism and still show no particular interest in problematizing analytical definitions of religion. Therefore, I wish to highlight the distinction between studying ‘religious discourse’ and discourses on ‘religion.’ (Taira 2016a, 2016b.)

The first approach defines what counts as religion and examines its constituent parts, its key distinctions, and its functions in society at large. An example of this kind of discursive approach is the work of Bruce Lincoln. Religion, according to him, is ‘discourse whose defining characteristic is its desire to speak of things eternal and transcendent with an authority equally transcendent and eternal’ (Lincoln 2012, 1). Lincoln starts the analysis of ‘religious discourse’ with a definition of religion. It allows him to write about the relation between religion and culture and use concepts such as ‘religious conflict,’ ‘religions of resistance,’ ‘religions of revolution’ and count how many ‘religious words and phrases’ there are in bin Laden’s and Bush’s speeches (Lincoln 2003).

Many people find Lincoln’s research so useful that standard definitions are worth pursuing, and I have no problem with that. However, in order to understand the variety of discursive approaches, I find it relevant to emphasize the difference between discursive approaches that use analytical definitions and discursive approaches that study how the category of ‘religion’ is used, defined, and negotiated in organizing social practices. The starting point of the second type of discursive approach is to avoid analytical definitions of ‘religion’. Rather, the focus is on studying how something comes to be classified as and called ‘religion’ and how it is connected to social practices and power relations. For an example of how the analysis is different in the context of ’deliberately invented religions’ depending on whether an analytical definition of religion is used, see Taira 2013b.

To be fair to von Stuckrad, he talks a little bit about the distinction in the interview and provides examples of how some strands become part of discourse on religion. However, I find the distinction worth underlining when navigating between different versions of discursive approaches. It may not be the most important distinction for someone who is interested in applying discursive approaches in the humanities and social sciences in general, but it connects such approaches to core debates concerning the study of religion by asking, what are the analytical terms we use and which terms need to be defined?


Fairclough, Norman 1992. Discourse and Social Change. Cambridge: Polity.

Foucault, Michel 2002. The Archaeology of Knowledge. London: Routledge.

Hjelm, Titus 2011. Discourse analysis. Michael Stausberg and Steven Engler (eds), The Routledge Handbook of Research Methods in the Study of Religion. London: Routledge, 134–150.

Lincoln, Bruce 2003. Holy Terrors: Thinking about Religion after September 11. Chicago: The University of Chicago Press.

Lincoln, Bruce 2012. Gods and Demons, Priests and Scholars. Critical Explorations in the History of Religions. Chicago: The University of Chicago Press.

Rorty, Richard 1999. Philosophy and Social Hope. London: Penguin.

Taira, Teemu 2013a. Making Space for Discursive Study in Religious Studies. Religion 43(1): 26–45.

Taira, Teemu 2013b. The Category of ‘Invented Religion’: A New Opportunity for Studying Discourses on ‘Religion’. Culture and Religion 14(4) 477–493.

Taira, Teemu 2016a. Doing Things with “Religion”: Discursive Approach in Rethinking the World Religions Paradigm. Christopher R. Cotter and David G. Robertson (eds), After World Religions: Reconstructing Religious Studies. London: Routledge.

Taira, Teemu 2016b. Discourse on ‘Religion’ in Organizing Social Practices: Theoretical and Practical Considerations. Frans Wijsen & Kocku von Stuckrad (eds), Making Religion: Theory and Practice in the Discursive Study of Religion. Leiden: Brill.

Von Stuckrad, Kocku 2014. The Scientification of Religion: An Historical Study of Discursive Change, 1800–2000. Boston/Berlin: De Gruyter.

Wijsen, Frans and Von Stuckrad, Kocku (eds) 2016. Making Religion: Theory and Practice in the Discursive Study of Religion. Leiden: Brill