Theoretical-methodological aspects of researching the area of religion and public administration

Abstract

This article, commissioned by the editors of Public Policy and Administration, contributes to the theoretical and methodological groundworks for the development of research in religion and public administration. It does so in three ways. The first part contrasts different approaches to defining the concept of religion, those from the vantage of academic observers against those from within religious traditions themselves. Several theoretical and methodological challenges for public administration follow. Part two sets out two theoretical concerns; first, the role of religious belief in intentional explanations of public administration activity; second, the institutionalised position of secularity in the recently freighted concept of a religious regime. Part three identifies a clutch of methodological concerns associated with these theoretical issues. A final section concludes with a summary of the contribution of the article to the nascent research programme in public administration for a post-secular world.

Introduction

Of the many post-concepts in academic use to describe a changing, contingent and uncertain contemporary world, post-secular is perhaps the most unsettling for the social sciences because it dents confidence in progress. Post-secular is not a label for a novel social phenomenon but rather a signpost backwards to the pre-secular, religious world. Leading scholars have recently begun to take on the challenge of building a field of research on religion and public administration (PA) (Ongaro and Tantardini, 2023, 2024; Dreschler et al., 2024). This line of inquiry recognises that religion has not faded away under putative moral progress of a secular age guided by enlightenment thinking and needs to be included in PA as an essential component in a process of decentring the ‘Western’. In sociology and anthropology, research programmes are established that investigate post-secular or de-secularised societies in which religion contributes to the legitimacy of the governing system (Bowie, 2020; Berger, 1999; Habermas, 2008; Taylor, 2007). However, this work has yet to attract sustained attention from PA scholarship.

The first part of the article considers different approaches to defining of the concept of religion. This is contested terrain; religion escapes easy definition in terms of set of elementary characteristics or key attributes that are common to the domain of the concept. Several theoretical and methodological challenges for a PA research programme follow. Part two sets out two theoretical concerns; first, the role of religious belief in intentional explanations of PA activity; second, the institutionalised position of secularity in the recently freighted concept of a religious regime (Ongaro and Tantardini, 2023). Part three identifies a cluster of methodological issues associated with these theoretical concerns. A final section concludes with a summary of the contribution of the article to the nascent research programme in PA for a post-secular world.

Defining religion

Reflections on questions of essentialism, as well as its anti-variety, have gained prominence in debates about the philosophical foundations of public administration (Ansell, 2011; Ongaro, 2020, 2021). These are useful in considering efforts to map the conceptual domain of religion and identifying theoretical and methodological issues for the study of religion and PA. Essentialism is a view of ontology and refers here to the position that there are essential properties that a social phenomenon must possess for it to be a religion. Ongaro and Tantardini (2023: 11) label this as a religionist approach to defining religion, which can be contrasted with a social science approach that takes an anti-essentialist line by establishing a set of characteristics that define religion as an abstract type from the perspective of a researcher qua academic spectator, but then allows for flexibility in making observations of concrete instances, or tokens, of religion. In terms of a menu of characteristics, there is a debate about which are individually necessary and jointly sufficient for something to be a religion, but none refers to the essence of religion. For example, the insight that religions have some metaphysical reference to the divine is frequently accepted (Tylor, 1871; Tillich, 1957; Turner, 2011; Lincoln, 2006) or, relatedly, as Geertz (1973) holds, religions are constituted by religious beliefs, held by followers about the divine, that are accepted as real and literal.

While social science approaches to defining religion have generally avoided essentialism, religious followers usually take an essentialist line to explain their own religion and its traditions (Hedges, 2021; Smith, 1978, 1982, 1995). For example, the field of Islamic Studies wrestles with what Muslims think and how the label Islamic is applied in different contexts; while Islamically-trained scholars - and much of the Muslim faithful - are able to state the essential elements of Islam clearly in Islamic, rather than social science, terms (Esposito, 2016). There is a difference between the set of behaviours, attitudes, opinions that researchers may attribute to Muslims, and accounts of the essence of Islam based on the interpretation of scriptures and law. This divide between the researcher and those who are being researched is an important element in questions how to bring religion into the weal of PA studies.

Smart (1996, 1998) identifies seven dimensions to the domain of religion: practical and ritual; experiential and emotional; narrative or mythic; doctrinal and philosophical; ethical and legal; social and institutional; material. Although not all these dimensions are immediately relevant to the study of religion and PA, this article highlights three that support multiple and flexible descriptions of religion and facilitate conceptual portability across time and space with theoretical and methodological implications for the emerging religion research programme in PA. Firstly, unlike social movements, religions have followers rather than members. These form a moral community of devotees who believe their shared religious experience in ritual worship will deliver progress or benefits for that group. Second, there is some threshold level of belief in transcendent that is supported by some combination of reason, emotion or faith that marks off religion from other social phenomena. Third, religions also have their distinct ethical norms that are demonstrated and promoted by rituals of worship that marks off religion from secular phenomena.

Theoretical issues

The analytical framework of Ongaro and Tantardini (2023: 17) (OT) sees religion as shaping personalities through motivational and ideational influences on human behaviour. In the framework, these influences may be observed at different levels: micro, meso, and macro scales. OT introduces the novel concept of a religious regime as a mediating factor on the influence of religion on PA systems. This concept covers the character of religion in a jurisdiction – both its nature and demographic distribution – as well as the relationship of religion to the state.

Although the framework may support different sorts of explanatory accounts, such as general covering law seeking, systems analysis, or propensity analysis, this article focuses on its use for mechanism-based explanations. Mechanisms are frequently occurring causal chains where similar causal effects between different factors, entities and activities are observed across space and time. In turn, a process is a contingent combination of different mechanisms which may or may not follow some regular observable pattern (Tilly and Goodin, 2007; Cartwright, 2017). Cataloguing the relevant mechanisms and how they concatenate into processes in different contexts is essential to theory building in the area of religion and PA.

Ongaro and Tantardini (2023: 20) conjecture that motivational and ideational mechanisms will likely occur together in most processes relevant to the influence that religion exerts in PA. The mediating factor in OT - the religious regime - is a cluster of environmental mechanisms. These mechanisms are not focussed on actors or their activities, individually and collectively at different levels of observation, but rather operate in the background context, or environment, that may enrich, expand or mollify other agent-focussed mechanisms in operation (Tilly and Goodin, 2007).

The second part of the article does two things: first, it defines intention as a distinct causal mechanism related to religious beliefs to complement the OT framework which relies on ideational and motivational mechanisms for the causal analysis of religious influence on PA; second, it considers secularism as an institutional device that is an essential component in the religious regime mechanism.

Religious beliefs and intentions

Both motivation and intention may play a causal role in explaining decision-making and action by agents under scrutiny, but they are distinct causal mechanisms. Raz (2017) elaborates the argument that an intentional mechanism consists of an agent acting for a reason because of the belief that there is value in that action. Motivation covers a range of psychological and social mechanisms that explain the desire of an agent to act in a particular way at a specific time. Although distinguishing the two types of mechanism may seem an exaggerated refinement, it is useful for the discussion here because it allows religious beliefs to enter more directly into explanatory accounts of religion influencing PA.

There are theoretical issues with the causal role of religious beliefs as distinct from secular ideas that usually inform and guide an intention to act in PA accounts. In theology, religious experiences raise questions about the existence of God, but from a PA academic spectator perspective, religious experience is simply the testimony that one thinks one has experienced the divine in a particular way. Whilst a straightforward position, there is an emerging strand of research in PA on the significance of phenomenological considerations to unpacking religious experience and its role in beliefs and intentions (Elías, 2020). Theoretically, there is a distinctive phenomenology of religious experience and the extent to which data on reported accounts of religious experiences in case studies can capture that phenomenology is moot. This is discussed methodologically in the third part, but the theoretical dilemma for PA scholars is how the phenomenology of religious experience and belief performs explanatory work through intentional mechanisms of agents valuing certain activities.

Specifically, phenomenological perspectives present questions of how actors experience a relationship with God as well as other actors in a moral community of religious devotees, and how these then inform religious beliefs which give agents reasons to act. These questions present formidable theoretical challenges. Without understanding and accepting a theological narrative, it is not obvious how the experiences of religious actors, as responsive and reflexive beings, can be understood properly in terms of PA and there is a risk of misrepresenting them through the application of secular labels.

In one way, this is unhelpful line of inquiry liable to lead to a version of self-defeating relativism that denies the possibility of any social scientific knowledge of agents’ interior worlds. It does, though, for the purposes of PA research designs, counsel against easy analogies to existing work on social movements and encourage more ethnographic approaches to certain research questions in religion and PA (especially at the more micro-level of street-level bureaucracy).

Relatedly, religious beliefs are not the same things as ideas as usually understood in PA and that relate to ideational mechanisms. Holding religious beliefs about the value of actions involves commitment, trust and faith that do not usually feature in PA approaches to the role of ideas. Specifically, faith involves accepting what cannot be known with reasonable epistemic certainty, for example through empirical observation. Religious beliefs about PA do not rely on evidence, in the way that secular ideas do, and are justified on grounds of faith, and enter explanatory accounts in PA despite being fully resistant to disconfirming empirical evidence or argument.

In prophetic religious traditions, truth is revealed by God, and the source of the revelation takes a scriptural form – ‘the Book’ – which is subsequently the authority on which religious belief rests. Many religious beliefs in the Abrahamic traditions are not about God directly, but they are held to be true, and the object of faith, because they are revealed by God (D’Costa, 2009). For example, this is position of Thomas Aquinas, influential in both secular philosophy and Christian theology, who argues that claims to truth in religion are made based on human testimony but are accepted because they are seen to backed by divine authority (Davies, 1993).

The dilemma for PA scholarship is how to model agents accepting as true and acting upon beliefs in intentional mechanisms that lack justification in the sense of being beyond directly accessible human experiences. The presumption of much of social science has been that religious beliefs will wither as social scientific knowledge expands. The notion of a constantly diminishing ‘God of the gaps’ is inherited from Comte which bases a perspective on theology in terms science; only what science cannot explain is left for theology (Gane, 2006). PA scholarship about religion needs to take on a position on the issue of whether theological claims should count as hypotheses about facts and tested against a hierarchy of evidence as established by the standards of good social science.

Secular ideas, institutions and religious regimes

In different PA literatures, ideas are often conceived as narratives that stitch together events in causal order and render situations intelligible. For example, ideas about public problems are often part of framing contests in which different and contentious attempts are made to develop an authoritative understanding of the problem. Such framing contests take place within established institutional configurations which influence the ideational process and shape the extent to which a particular idea affects the policy process (Béland, 2016).

The place of religion in democratic governance has been typically managed in the context of Western PA by ideas of secularism. One aspect of a religious regime is the prominence of secularism as an idea underpinning institutional configurations that serve to limit the impact of religious views in the public sphere. Although there are multiple varieties of secularism and with importantly different implications, all imply institutions to enforce separation between the political and religious worlds and a carefully circumscribed role for religion in the public realm (Audi, 2011; Taylor, 2007).

The theologico-political problem, originally posed by Leo Strauss, is about limits to the authority that religious beliefs and ideas can command in PA systems (Audi, 2011; Meier, 2007). Many religious regimes function to constrain religion extensively on the grounds that governing by popular consent in a democracy is not compatible with claims to divine authority. This gives a normative edge to secularism and the need for religious constraint: the trend of religious beliefs and arguments becoming private matters over time is desirable and contains the corollary that they should no longer acceptable in terms of public governance.

Taylor (2007) considers multiple types of secularism and argues that we may live in a ‘secular age’ even if religious participation is visible and fervent. This moves questions of secularity on from theories of how religion will fade as social scientific knowledge advances to the issue of how religious pluralism may be established and maintained in liberal societies. For example, Rawls (1993) argues that religion should be excluded from deliberative discourses that relate to constitutional affairs and issues of basic justice in a liberal democracy i.e. secularism should be the precursor rather than outcome of such deliberations. This position softened subsequently to legitimate public action cannot be based exclusively on religious grounds, there must be some complementary secular rationale (Rawls, 1997). The corollary is that institutions in a religious regime do not need to exclude religion from public sphere but instead limit its influence. As such religious considerations may be influential in contentious politics and citizens and groups may advocate for their interests and policy preferences on exclusively religious grounds.

The constitutional separation of church and state as institutional manifestation of a secularist religious regime has a long history (Jung, 2007; Audi, 2011; Mill, 1871). The underpinning idea is that the comprehensive nature of religion threatens the legitimacy of a democratic PA system because it may require coercive public action covering non-believers. Less prominent, and from the point of view of religious devotees, is the idea that secularism is equally comprehensive and exclusionary; in this case, religious belief and identity is being prevented from being fundamental in public life.

Cases of the global revival of religion since the 1970s have presented questions to secularist settlements in terms of public policy and management without really attracting sustained attention in PA scholarship. Marty and Appleby (1992) argued that religious fundamentalism produced a cluster of movements in the last quarter of the twentieth century that reject secularism explicitly and promoted instead comprehensive PA structures and religious institutions covering every aspect of social and political life. As the concept of a religious regime is developed further in PA through empirical application, uncovering the variety of these fundamentalisms is important for the success of any non-Western PA research programme (Kay, 2024). For example, Voll (1994) describes contemporary Islamic fundamentalism as a complex cluster of different movements emerging from a wave of Islamic revivalism in the 1970s, but all are concerned with foundational Islamic beliefs and principles and the political project of organising government for their reaffirmation such as the responsibility of an Islamic society for forward thinking (islah) and ‘reform’ (tajdid, in the sense of renewal and revival).

These movements often reach back to Islamic modernism of the nineteenth century, influential in many anti-colonial movements in the twentieth century. Thinkers such as Muhammad ‘Abduh, Jamal al-Din al-Afghani, Rashid Ridha, Hassan Al-Banna and Sayyid Qutb stressed the centrality of Islamic law, theology and ethics in the structure and organisation of society supported by an Islamic PA. The goals of an Islamic society (maqasid al shari’ah) required an institutional separation of the Islamic world from the non-Islamic (jahiliyya, or ignorance). This, rather than secularism, is the starting point for the relationship between Islam and PA seen in many contemporary Islamist movements, the Muslim Brotherhood for example, which have developed sophisticated PA programmes to provide medical, education, finical and social services run on perceived Qur’anic lines (although these have often been challenged within Islam).

Methodological issues

The challenge of distinguishing intentional mechanisms that make sense to actors from reconstructions of intentions by researchers using a social science theory of PA presents a methodological problem in researching beliefs and intentions in religion. There are different trade-offs to be navigated between an explanation achieving capturing fully the relevant beliefs and intentions of religious followers or the goodness of fit, and the PA research standards of parsimony in causality, or reconstructing an account intelligible to PA readers of the structure and pathways by which internal mental life such as classification, reason, belief, intention, motivation and rationalisation may shape actors’ decision-making and action.

While this is a general PA problem, it is more likely to acute for this context because of distinctive features of religion and religious beliefs. Well-established PA research programmes may seek to explain the behaviour of public administrators and managers by reference to the benefits to be gained for their position, their own power, their capacity to hold the loyalty of their followers, and their influence in the wider world. Yet religion may require something beyond such rational actor logic: a constant and unfolding reflection by its followers, less about the rationality of any choice or pathway of action, than upon the mind of God and arguments of a kind that western PA may dismiss as exclusively theological. For PA to apprehend a post-secular society and its influence on PA, can scholars simply dismiss this and stick to observed behaviour to infer decision-making in terms of payoff, maximisation, constraint and the limits and costs of information? If religion and belief in God’s will are prior intentions and motives for religious followers, they are only obliquely related to the issue of individual calculations of interests within institutional constraints.

The categories of belief and intention that religious actors think directly relevant to their decisions are likely to be significantly different to the things that PA’s own standards of explanation would regard as directly relevant. Obliquity and indirectness in explanation is a methodological well as an empirical issue. The standard of a satisfactory explanation is determined as much by arguments about which of the rival virtues of explanation PA should privilege, in any given trade-off. After all, being empirically right – privileging goodness of fit – will not be enough for explanatory satisfaction if we cannot also show a clear causal path – privileging causality – by which reconstructed beliefs and intentions work, to rival the coarse directness – parsimony – of the conventional and straightforward intentional explanations and restricted menu of PA terms for public manager behaviour.

The key point for this part of the article is that what PA may count as a relevant belief or intention for understanding what someone does, in studying the influence of religion on PA, is determined not only by the empirical considerations of evidence and fact. Instead, there are series of methodological issues to be identified and considered for the purposes of PA research designs in this area.

Methodological agnosticism

The methodological challenge in researching religious beliefs and intentions mirrors the theoretical: determining which types of data are valid and reliable as evidence of religious beliefs in cases where religious followers may view the co-created data from the research process differently from those undertaking the research; specifically, in setting the standards of what counts as directly relevant religious beliefs and intentions in an explanation of the influence of religion. For example, the religious followers may not divide their statements along a secular divide into public and private ones in the same way as the PA researchers and readers of research may do in their research designs; or if they do, they may place the divide in different places or attach importantly different significance to it. This is likely to be as true within societies as it is between them, as well as between different temporal stretches being investigated; indeed, the same biological individual may work with different identities and related practices of what counts as public and private, if anything does, in relation to their work, their home or their religious life.

Work by Berger (1967) and Smart (1962) developed the notion of methodological agnosticism, endorsed by Bell and Taylor (2014) for use in the study of organisations, as a promising way of tackling this issue. It does so by insisting researchers avoid ontological and epistemological commitments concerning the nature and truth of religious belief. Many scholars in the wide and varied field of religious studies recognise the merits of this approach. For example, in supporting commitments to religious pluralism and interfaith understanding. Additionally, in terms of research practice, avoiding a position that either confirms or denies the existence of the gods that research participants may hold as foundational beliefs may help reduce the risk of offence or undermining research conversations between believers and non-believers.

However, it is not clear, methodologically, how this helps PA and its emerging work on religion. The attractions of agnosticism as a research practice are obvious, but it is less obvious how it undergirds a philosophical position to investigate the role of religious beliefs in explaining decision-making as claimed by Bell and Taylor (2014). Its counterpoint is methodological naturalism underpinned by a staunch secularity: for an explanation to meet the standards of a social science, it cannot refer to any divine activity, variables or mechanisms that are overtly religious or theological. Observations of such things are only meaningful as proxies for underlying social reality. On this view, PA research must stick to the solid ground that the long-established sociology of religion has set out.

Martin (2018, 2023) argues that methodological agnosticism cuts off possibilities for critical scholarship such as emancipatory ambitions associated with commitments to human rights in PA (Roberts et al., 2024). For example, recent debates over marriage equality include claims that marriage is a socially constructed, oppressive institution while religious claims insist that marriage is a sacrament and gift from God. There are several normative features relevant to a liberal commitment to the sovereignty of individual persons and their human rights that are highly controversial from a religious viewpoint.

Further, methdological agnosticism does not solve the issue in explanatory scholarship about the epistemological status of claims about the divine and revelation in social science. For example, religious beliefs about the truth of four Christian gospels are not amenable to falsification and cannot be tested empirically. It is not obvious how these can enter explanatory accounts in PA using tools of social science inquiry. The possibility of agnostic description of such beliefs as having the same status as secular rational beliefs that enter intentional mechanisms in explaining decision-making risks undermining many other aspects of existing PA knowledge and collapsing fully into a theological position. The role of methodological naturalism is an important one for the development of PA seeking to offer insights into the influence of religion.

Such consideration requires peeling back the debate to its underlying ontology. Martin (2018) argues warranting inference to description of religious influence requires ontological commitments. For example, any description of Islam as a tradition in terms of stability and change over time requires ontological commitments to what is immutable in Islam; the essence that must be unchanging for something to be Islamic. Similarly, limiting descriptions of Christianity to claims by those who identify as Christian at a moment in time may well be a useful approach but requires an ontology distinct from theology.

Translation

As discussed above, there is a methodological dilemma in the translation of terms from those used by the people being studied into those used by researchers: how to sustain explanatory power without becoming unnecessarily reductive by squeezing phenomena into descriptively thin schema. Such risks are unlikely to be mitigated by a single position covering all PA research ambitions; instead, encouraging a multi-theoretic approach to the study of religion in PA (Whetsell, 2013), varying case by case according to different ambitions such as developing explanations against the standards of intelligibility and authenticity as well as the objectives of goodness of fit, parsimony, and causality.

Among social sciences approaches in the study of religion, anthropology has been the most insistent on a distinction between an explanation drawn in the terms that would be used by the people being studied, and one drawn in language that will be recognised by the readers of the research. In that field, the conventional way of contrasting these two things is to use the label of ‘emic’ for the terms used by the people studied, while those of the researcher and the wider research field are ‘etic’ ones (Bowie, 2020).

For the researcher, of course, it will not be enough to rely on either alone. An explanation of PA actions that made no reference at all to the ways in which those actors framed their decisions, how they classified their risks, their expectations, how they described their intentions in deciding on the development of their reform model, would be too schematic. It would detach its answer to the ‘why’ question so far from its answer to the ‘how’ question in ways that raise methodological concerns about explanatory satisfaction.

Multiple, vague and ambiguous beliefs: Combining religious and secular identity

One implication of the OT framework is that researchers may be more interested in actors’ motives and their rationalisations than intentions on the grounds they are more likely to be informative and useful for research purposes. Specifically, religious beliefs about PA may lead to multiple, vague or ill-formed intentions, particularly when combined with secular beliefs. However, they may instead contribute motives that are rich, complex and multiple, and where actors’ own rationalisations of their decision-making activity using religious beliefs are the best available evidence. There is a gap between an intentional explanation, pairing of prior beliefs and intentions to explain an action, and a sense-making explanation, where the beliefs and motives reconstructed by the actor in retrospect after the action, as part of a process of narrative ex post rationalisation of why they did something. In the latter case, the absence or at least the inchoate character of a prior intention may leave motivational mechanisms, linked to religious beliefs, to fill the explanatory void. Research in PA is sometimes sense-making of this kind where explanatory accounts of public sector reform programmes discover that governments may launch these with quite vague or incoherent intentions and that explanatory accounts are superior with motivational mechanisms and rationalisations than standard intentional explanations.

The OT framework allows researchers to decide whose religious beliefs and intentions matter on the micro-meso-macro spectrum; specifically, whether research designs are concerned with individuals, public organisations, or institutions. An important concern is that what researchers may count as a relevant religious belief or motivation for explaining observed activity is not a straightforward issue of data quality but also partly shaped by methodological trade-offs made in research designs for investigating religion empirically. There is a need for further discussion in PA on different virtues of explanation in different research designs when studying the relationship between religion and public administration.

For example, inferences about the status of what counts as a public or private statement of religious belief and what difference this makes, must logically underpin our inferences about religious beliefs and intentions. However, deciding what counts as evidence is itself an inference rather than a basic fact, and one that rests on the whole set of data and the inferences researchers may make from the set using the OT framework. This circularity is an important methodological issue to be acknowledged in researching beliefs and intentions.

Relatedly, it is not an empirically open question whether intentions based on religious beliefs attach to individuals, clusters, groups or institutions; instead, a prior methodological position needs to be staked out. In studying public organisations, PA has made significant progress elaborating the effect of institutions in shaping the beliefs and intentions of the individuals who work in the organisation. But in terms of religion, PA researchers may also want to consider variable measures of collective belief and collective intention at the meso-level. These shared beliefs and intentions may well be inchoate and/or incoherent combinations of the religious and the secular.

Collective beliefs are potentially more important than individual ones in investigating religion. In many religious traditions, revelation and religious laws, over time debates about original intentions and literalism prevail. In deciding such cases in Islam for example, combinations of imams, muftis, qadis use a consultative institution (shura) to provide interpretations of what is the Islamic thing to do backed by a collective Islamic intention. There is a rich, sophisticated and long-standing tradition of interpretation in Islam must therefore construe the words of the Qur’an, Sunnah and Hadith in the light of other ethics and values and to impute a collective and determinate intention.

The attribution of belief and intention is a methodological inference rather than a discovery. Moreover, it is not just an inference because of the problem of limited or missing data and the need to make a potentially risky inference. It is a methodological level inference: the intentions that researchers attribute to religious people are only specifiable and only intelligible as part of a broader network of explanatory propositions that are being conjectured. These attributions are, in turn, only defensible if other hypotheses within the OT framework can also be evidenced. Critically, the trade-offs PA is prepared to defend on methodological grounds as meeting the standards of a satisfactory explanatory account limits the menu of types of beliefs and intentions, religious and secular, that PA can plausibly and relevantly attribute to actors being studied.

Conclusion

This article has made three steps. The first part looked at the defining the conceptual domain of religion from an anti-essentialist perspective. While this approach helps the portability of the concept of religion across different jurisdictions to assess its influence on PA, it does present a set of theoretical and methodological challenges associated with the inclusion of the beliefs, experiences and intentions of religious devotees. Specifically, it risks explanatory accounts becoming schematic and unsatisfactory because religion is reduced to preexisting PA categories.

The second part of the article elaborated two central theoretical concerns. First, the problematic role of religious belief and how far the inclusion of insider, or emic, perspectives of religious followers can be sustained in the use of intentional mechanisms to support explanatory accounts of PA. Second, existing PA knowledge about the relationship between institutions and ideas is used to consider secularity in the recently freighted concept of a religious regime within the OT framework. Secularism is the default position of the PA field with the corollary that the religious and secular worlds can be separated with varying degrees of completeness. Although this may well be the best position for PA to work from, it comes with its own legacy in terms of liberal, enlightenment thinking and arguably limits the ability of non-Western PA research programmes to apprehend religious influences in other societies that do not accept a priori church-state and public-private divisions in PA.

Part three has identified a cluster of methodological trade-offs that do not present any easy options for PA scholarship. For example, while methodological agnosticism imposes limits on critical and explanatory PA scholarship, its alternative – methodological naturalism – is liable to miss key aspects of religion and its causal influence on PA. The translation of terms from religion to social science is demanding and PA needs to able sustain analysing of combinations of secular and religious beliefs in different contexts. This is a central methodological challenge for PA scholars seeking to distinguish the role of ideational, intentional and motivational mechanisms in causal accounts of the influence of religion on PA.

The nascent research programme in PA for a post-secular world is important and promises much in enriching the theoretical scope and empirical richness of the field. This article has sought to contribute a discussion of some issues that need to be acknowledged, if not fully resolved, for its successful development.