Religion in European constitutions – cases of different secularities

Despite all predictions about the disappearance of religion, we recently have observed the revival of the ‘religious argument’ in Europe's public and political spheres: this concerns the public display of religious symbols and buildings, the emergence of new religious states, religiously motivated terror acts or the religion-related discussions about boundaries of freedom of speech. These empirical developments give reasons to re-consider once more the relationship between societies and religion beyond the well-known theoretical argument of secularised European societies (critically summarised by Pollack and Rosta 2015: 20 ff.).

In order to do so, we treat religion as a multi-dimensional concept that manifests itself not only in individual beliefs and practices or in public discourses but also in legal documents.¹ The analysis of religion within legal documents provides not only perspectives on freedom of religion, independence of churches and the right of worship, but at the same time, the religious (in)dependence of the state and its government (governmental secularity). Constitutions, in particular, regulate the relationship between societies, the state and religion(s) in their most fundamental legislative way. They are a centrepiece of nation-building and expression of modernity (Blaustein 1994: v).

Sociologically speaking, constitutions embody the power struggles at the time of their enactment. They incorporate basic rules, regulations and values of a society. In this regard, they mirror the nations’ self-understanding (Murphy 1993). They also formulate arrangements of religious and governmental organisations and constitute the nation-state as ‘religious’, ‘laicist’ or ‘secular’. They provide the legal framework within which corporate actors act; they affect policies as well as individual understandings, interpretations and normative perspectives. They make religion relevant even for the non-religious members of society. In turn, they are also a reaction to political discourses.

Constitutions turn out to be contested not only on the level of nation-states. The debate about the reference to religion in the Preamble of the European Constitution around the year 2000 exhibited how powerful the topic of religion is in relation to constitutions in the political arena of the EU (Grötsch and Schnabel 2013). The aim of this paper is to explore how religion and society are intertwined institutionally in the member states of the EU, since such institutional settings provide conditions for private and public religious understandings, beliefs and practices. More specifically, we focus on how constitutions comprise, integrate, relate to and contextualise religion. Because we consider constitutions as a centrepiece of the institutional (here: legislative) setting of modern nation-building, our analysis is guided by two classical sociological approaches that help to understand (the development of) modern institutional arrangements. First, neo-institutionalism stresses the influence of world culture leading to institutional isomorphism as an adaption of worldwide shared institutional scripts (Meyer et al. 1997); and second, theorists of historical sociology emphasise the emergence of institutional arrangements as path dependent on particular historical events (Pierson 2004).

Based on quantitative, explorative document analysis, we developed an original data set in which we were able to describe similarities and differences concerning the manifestation of religion within the constitutional texts of the EU-member states. Thereby, we focus on two text-immanent aspects of religion-related topics and patterns of phrasings. From a neo-institutionalism perspective, we expect more similar phrasings and topics in more recent constitutions while theories of path dependency indicate similarities only in cases of similar historical events. Subsequent to the discussion of neo-institutionalist and path dependency theories as most helpful tools to understand the historical formations of constitutions (chapter 2) we provide a brief motivation of why constitutions are sociologically relevant (chapter 3). We continue with a presentation of our data set and with the empirical analysis (chapter 4). The article concludes with a discussion of our empirical findings. Our analysis of constitutions provides a deeper insight into the religious self-understanding of the EU-member states and the way religion is societally embedded into the legal system beyond the dimension of individual religiosity.

As a particular kind of statute, constitutions are part of the Enlightenment's ‘enterprise of subjecting human conduct to a governance of rules’ (Fuller 1975: 96). Although it is not necessary for a state to have a constitution in order to be recognised internationally, most modern states are based on constitutional documents (Go 2003: 71). By now, 154 out of 193 member states of the UN-system have codified constitutions (UN 2015). The American idea² of drafting constitutions for every nation-state of the world seems to have prevailed (Blaustein 1994: v). Constitutions are constantly debated and re-drafted, not only when states collapse or regions declare independence (e.g. Elkins et al.2009, though such processes will not be of interest here.

Modernisation theories help to understand how enacting a constitution became best practice of founding a state. Among sociological theories, the two well-discussed approaches of path dependency and neo-institutionalism provide explanations of similarities and differences in institutional settings.³ These two approaches are well established, empirically tested (Meyer et al. 1997; Pierson 2000) and – although not previously applied to religion in constitutional texts – allow for developing expectations concerning the position of religion in constitutions.

Sewell (1999) suggests that path dependency means, ‘what happened at an earlier point in time will affect the possible outcomes of a sequence of events occurring at a later point in time’ (Sewell 1999: 262–63). This perspective indicates that ‘history matters’ (Pierson 2000: 252) – the formation of institutions does not only depend on the immediate process of coordination and negotiation but also on longer historical paths of events, political developments and opposed interest formation caused by social, political or economic heterogeneities (like cleavages) or they may even depend on cultural characteristics such as religious cultural teachings, as argued by Pickel (2009). As Mahoney (2000) points out, such conditions are handled in a specific way and produce in turn unique outcomes from that point of time onwards. In this perspective, constitutions can be read as ‘power maps’ (Duchacek 1973) in which a specific constellation of opposing groups and interests is inscribed.

There are ‘critical junctures’ (Collier and Collier 2002) such as revolutions, wars and violent regime changes or technical and social innovations that enable change of paths. But once an institution is entrenched, it is stabilised by continuous reproduction as the costs of a transformation or modification would be too high (Mahoney 2000: 510 ff.). Paths are re-enforced by the mechanisms of positive feedback and increasing returns:

Neo-institutional perspectives suggest that political actors – such as organisations or states – aim to appear as rational actors to gain national and international legitimacy (Meyer and Jepperson 2000) and to be part of a world culture⁴ (Meyer et al. 1997; DiMaggio and Powell 2000). For Meyer et al. (1997) world culture is a mono-centric, hegemonic knowledge system, providing scripts that concern interpretation and action. States are integrated into this world-system of expectations that constitutes rational actors, as well as particular forms of legitimation and understanding. These expectations, legitimations and meaning patterns create an independent, irreducible social structure according to which adaptions of national legal systems are conducted. Examples for the adaption of world culture are institutional scripts comprising, for example, state power and individual rights, mass schooling systems, rational economic and demographic book keeping, formal equal opportunities for men and women (Meyer et al. 1997: 152), or – in this case – constitutions as documents of nation-building. Such elements are implemented by coercive, normative or mimetic practices. However, they do not necessarily provide the intended consequences on the micro-level of everyday reality, which is described as loose coupling or decoupling. Decoupling refers to the divergence between ‘formal models and observable practices’ (Meyer et al. 1997: 155). Such divergence implies that the mere existence of constitutional texts may indicate some effects but it does not ensure which kind – and that such effects may be different for varying nation-states (and may differ from the intentions of the law-makers). The ‘me-tooism’ – as Duchacek (1973) calls such institutional isomorphism – implies that there is an important difference between the written constitution and the constitutional reality: ‘New and would-be nations need constitutions to discuss and display regardless of their expected uses, simply because the established nations have them’ (Duchacek 1973: 11).

These two historic-theoretical approaches give rise to two different perspectives on European constitutions: path dependency implies that constitutions are institutions that are established at a particular point in time with consequences for the actual state. But not all states necessarily have a constitution and considerable differences between the constitutions can be expected. Following the theory of neo-institutionalism, constitutions are institutional replications of global blueprints.⁵ To reduce the costs of establishing a ‘new’ institution, the isomorphic strategy offers a tool to adjust form and content, therefore we expect that states will have constitutions and that these constitutions will show considerable similarities.

Before we start our analysis, we want to establish briefly how constitutions matter as sociological data. Constitutions are legal texts of a particular kind (e.g. Blaustein 1994: viii; Finer et al.1995: 2ff.; Culic 2003: 43ff.; Go 2003: 71). Literature on constitutions suggests that they matter with regard to a formal, a functional and a declarative dimension.

(i) Formally, constitutions are put down in writing (as explicated in Grey 1984): they are not an ‘imaginary thing’ (Grey 1984: 15) since they are in fact written documents that entail legal titles or claims which develop independence from those who wrote the text in the first place. Moreover, constitutions are fundamental and superordinate compared to other legal texts: ‘In the formal sense, a constitution is a written document containing legal rules and principles claiming priority over other rules and principles’ (Tschentscher 2011: 3–4). They lack explicitly expressed sanctions. From a sociological perspective, constitutions therefore, consist of agreements embedded within a dispositif of sanctions, but not of tangible norms. They are transacted in other kinds of laws, regulations or policies and they can only be amended by different procedures: they are – in some way – institutions grounding institutions.

Additionally, constitutions can be understood as the contractual result of negotiation processes between different interests at a particular point in time about how the social life shall be arranged: constitutions are ‘meant to express an arrangement vastly more complex than those underlying most legal documents: the web of society's basic institutions and ideals’ (Grey 1984: 16); they make the ‘unwritten constitution’ of a favoured society visible and compulsory and, following Alexander (1998: 2f.), reveal the ‘meta-rules’ of an imagined people. Their writing is a forceful process: ‘Constitution writing requires vivid process of deliberation, discussion and consultation’ (Culic 2003: 39), without implying that they are necessarily amended in a democratic process (Elkins et al. 2009).

(ii) Functionally (or substantially, Tschentscher 2011: 3), constitutions produce consequences. As ‘performative acts’, they postulate the material and symbolic foundation of the significant political entity (state) and its actors. They comprise of rules and regulations concerning the governmental arrangements and assign rights and duties (Finer et al.1995: 1). By that, they formalise a relevant context for governmental and individual actions in the private as well as in the public sphere and legitimise relations and regulations of the fundamental political and social order.

(iii) In their declarative dimension, constitutions are descriptions of self-understanding manifested in formulations as ‘we the people’ or ‘we’ and ‘us’. Such wordings declare and describe the political and social unity of the political entity. Additionally, constitutions declare the normative regime and arrangements of this entity. Because constitutions are ‘snapshots’ of a negotiation result they are partly ‘aspirational’ (Blaustein 1994: vi) and comprise of a utopian element (Eckstein and Apter 1964). They are always subject to contestation, re-negotiation and opposition and thereby ‘under constant construction’ (Elkins et al. 2009). All constitutions share some basic features: ‘Not only are they all packed in a single document, they all specify in one way or another the organisation of political power, the division of governmental labour, the major principles and goals of governance’ (Go 2003: 72).

In the following analysis, we focus on constitutions as (i) written-down texts that can be considered as (ii) performative acts describing (iii) ideal societies and their governmental arrangements. They do not represent societal or political realities and they are embedded in constant processes of re-writing, interpretation and implementation. They are, however, a social reality insofar as they are an authoritative description of anticipated societal and political order (Torres and Guinier 2011). We thereby distinguish between constitutions as texts (as our research subject) and constitutionalism as the way constitutions are set into reality.⁶

Because of their uniqueness regarding the dimensions developed above, we expect constitutions to guarantee religion a distinctive place within the societal framework including the religion-related self-understanding of the people, as well as the assignments of rights and duties concerning individual freedoms and governmental arrangements.

This gives rise to the question of how constitutions help to understand the revitalisation of the ‘religious argument’ in Europe's public and political sphere. While the longstanding tradition of secularisation theories assume that religion has lost its significance as an essential part of the social during the modernisation process in Europe, some studies on social capital and social cohesion imply that religion still plays an important role in the public and the private sphere (e.g. Putnam 1993; Greeley 1997; Pickel and Gladkich 2012; Schnabel and Grötsch 2014). We suggest that whether religion is relegated in a rather insignificant position or a prominent part of the constitutional text can serve as indicator for the degree of secularity (as a current state of affairs; as overview see Cianitto et al. 2013).

(i) Compared to other regions of the world, European societies are descried as uniformly secular: Berger et al. (2008: 8) stated that ‘[t]he United States is a religious society, Europe a secular one’ while Davie (2006: 247) referred to Europe ‘as an exceptional case’ compared to other regions of the world. Without intending to go deep into the discussion, secularisation can be understood as a loss of importance of religion. All secularisation theories refer to it as a component of the modernisation process of Western societies (Tschannen 1991). Here, secularisation goes hand in hand with modernity, in particular with an increase in wealth, social security, urbanisation and individualisation (Pollack and Rosta 2015: 28 ff.). However, European secularisation relates to different societal levels (as argued by Dobbelaere 1981; Casanova 2006; Pollack 2008). Following Dobbelaere (1981), we differentiate: secularisation on the individual level relates to pluralisation and individualisation of religion (Luckmann 1967; Davie 1990) and to people losing their religious faith (Berger 1967). Such trends are detected all over Europe, mirrored in particular in a decline in church attendance and decreasing membership numbers; but these trends show different extents and dynamics in the different countries (Pollack 2008, Pickel 2009). On the organisational level, secularisation involves religious organisations (such as churches, congregations or parishes) losing their importance; they become less essential as employers and are increasingly subordinated to worldly legislation and policies and lose their functions as social associations. However, there are vast differences among EU-member states concerning such state-church relationships, ranging from state-churches to financial support, or coexistence (Alber 1995; Fox 2007, 2011a). On the level of society, secularisation refers to the ‘enchantment of the world’ (Weber 2006 [1922]) when religious explanations are replaced by scientific ones, and to progressive functional differentiation that reduces religion to one among other functional spheres in society (Luhmann 1977).

But the decreasing relevance of religion is contested: it is argued that religion does not necessarily lose its vitality and that an increasing desire for new and alternative forms of spirituality and New Age-movements indicates that this applies only for the church-related forms of religion. Additionally, it seems a still open question how religion and modernity are linked (Pollack and Rosta 2015: 28). Although empirical evidence can be found that the secularisation thesis holds best to explain developments in Europe, we want to question whether Europe as a geographical and political area is as uniformly secular as often assumed: different levels and different nation-specific developments seem to suggest otherwise. As Pickel (2009) argues, contextual factors and path dependencies such as denominational legacies and the developments of different social security systems have to be included into an informed analysis of secularisation.⁷ Additionally, as argued above, the different levels on which secularisation takes place may lead to a differentiation in secularisation processes. While the idea of a homogeneously secular Europe ties in with theoretical considerations about developments towards a European isomorphism, such different nation-specific developments suggest nation-specific paths of secularisation.

Although our data set does not allow for tracing developments we may locate secularity in constitutional texts: in the lesser importance of religion in the people's self-understanding in more current constitutions, in a loose state-church-relationship, in restored individual rights and freedoms and in higher administrative independence.

(ii) Despite all secularisation tendencies, empirical studies repeatedly recognise religiousness as still important for social capital and societal integration (concerning deviant behaviour: Hirschi and Stark 1969; Stark and Bainbridge 1987; concerning social capital: Putnam 1993; Smidt 2003). Social capital seems to evolve through religion because it fosters a high level of value consensus, creates longstanding familiarity among the community members and suggests stability through a high level of reciprocity (Welch et al. 2004: 319). Pickel and Gladkich (2011), however, showed that the relationship between social capital as a factor of societal integration and religion is two-faced: besides bridging different religious groups religiousness can encourage discrimination against alternative social groups. Gross and Ziebertz (2010) and Yuval-Davis (2011) argue similarly that religiousness develops exclusive powers and that religion operates as a marker of group boundaries between ‘us’ and ‘them’. In modern, pluralistic societies, religiously generated social capital may promote social conflicts and lead to discrimination and even violence (Pickel and Gladkich 2011: 104f.). For Europe, it has been argued that Islam became the most prominent religious ‘other’ during the turn of the century (Casanova 2006: 5) – most visible in the debates about Turkey's EU-accession, Islamist fundamentalism and violence within and between nation-states. If religion is still present, it has effects on social cohesion and it is a significant marker for accession and exclusion. Constitutions as self-descriptions of the people and as ‘grounding institutions’ provide the framework within which religion (or its absence) reveal such outcomes. If religion has such far-reaching effects for societal order, we would expect that religion is considered worth being regulated significantly within different areas within constitutional texts.

If neo-institutionalism is correct, nation-states are no longer self-contained units. Independent of the year of enactment we expect that recent major constitutional reforms would have led to a harmonisation regarding content, placement and formulation of religion in constitutions. There should be – at least – a kind of European script of religion in constitutions. In contrast, path dependent theories understand nation-states as independent entities that are subject to developmental paths manifesting themselves in constitutional texts. Therefore, religion is expected to be articulated and positioned differently within the individual European constitutions. However, there are historical events that might constitute critical junctures. As far as these events are not only of national provenance, we expect similar paths for those countries that were affected. We consider the end of the Second World War, the date of EU-accession and the collapse of the Soviet Union as such influential international historical events.

There are two major reasons to focus on Europe: firstly, we want to establish whether (traces of) secularity can be found within the institutional legal framework of the EU-member states and if such traces are similar enough to justify speaking of the EU as one (exceptional) case and – in this regard – of European isomorphism. Secondly, the attempt to formulate a joint European Constitution for the EU failed because religion was considered a highly controversial issue. The comparison of the member states’ constitutions may shed light on the reasons of this failure. We centre not as much on how constitutions protect religious freedoms (as already done by Fox 2007, 2011a, 2011b; Fox and Flores 2009) but on how countries and their people perceive the necessity for legal regulation of religion. In addition to other studies we are interested in the different thematic fields in which such necessity for regulation occurs.

Constitutions provide particular data: because constitutions are written documents comprising similar basic elements and functions, easily accessible and more or less stable, they are especially suitable for comparison. Because of that, we do not include other legislative documents regulating the state-religion- or state-church-relationship (e.g. the Church Act in Denmark that constitutes and regulates the Evangelical-Lutheran Church as state-church of Denmark).⁸ We do not read the constitutions judicially but strictly sociologically (Cotterrell 1998). We also provide neither an analysis of the formation context, the history of the constitutions nor do we analyse any effects or outcomes. Our analysis is a strictly text-based analysis.

According to theories of path dependency, we want to contextualise our analysis by the year of enactment, last constitutional modification and significant historical events implying changing political cultures (as (i) the end of the Second World War, (ii) the date of EU-accession and (iii) the collapse of the Soviet Union) in order to see if these historical events affecting more than one nation-state are associated with similar regulations regarding religion.⁹ Additionally, we want to take religious cleavages (degree of religious heterogeneity measured by the Herfindahl-Index)¹⁰ and dominant denomination in the year 2008 into account since they reflect differences in social cultures.¹¹ Unfortunately, for most countries we do not have data about the religious cleavages at the time of enactment. Although our theoretical framework is argued historically, the lack of such data does not devaluate our analysis because we focus strictly on showing how religion is part of the current legal framework of nation-states. Moreover, we want to assume that if more recent changes in the almost stable distribution of religious cleavages¹² would have had any impact, constitutions would have undergone changes within the last five years (2008–2013). Remaining differences between constitutions can be interpreted as indicators of certain national particularities being considered to be matters of continuing significance.

Although we do not provide a hypothesis test in the strict sense our empirical analysis of constitutional texts is guided by the following assumptions:

If the suppositions of neo-institutionalism about institutional isomorphism are accurate we presume:

If the assumptions of the theory of path dependency are correct, we presume:

In order to address our research question of how constitutional texts intertwine religion and society, we focus on the constitutions of the 26 EU-member states of 2013.¹³ For the analysis we concentrated on articles that refer directly and explicitly to the field of religion; other topics that might be related indirectly to religion without being explicitly mentioned are not taken into account (e.g. public welfare or donations). In order to facilitate cross-country comparability and to provide an original in-detail-text-based analysis we are explicitly not interested in the drafting processes or interpretation processes leading to changes (as done by Melton et al. 2013).

To generate a detailed and reliable data set we proceeded as follows¹⁴: on the basis of a first reading of the constitutional texts we generated a codebook comprising all aspects and dimensions which touch upon religion. The coding by Fox (2011b) served as a starting point and was modified according to the distinct features of European constitutions in order to receive a more precise, detailed and tailored instrument that takes individual freedoms and administrational arrangements into account. As argued above, literature on constitutions as sociological subjects suggested that religion could be regulated (at least) on three dimensions (or thematic fields): self-understanding of the people, assignments of rights and duties concerning individual freedoms and governmental arrangements. The concrete dimensions as well as the indicators were then developed from comparison of the constitutional texts through identifying patterns and common sequences.

Our codebook constituted the basis for the coding process. Thereafter, we coded all passages of the constitutions concerning religion according to these variables. In order to secure intercoder reliability, the three authors coded the constitutions independently and compared their coding afterwards. Cases of doubt were discussed in detail in order to receive an exhaustive and reliable data set. We obtained a quantitative data set of 37 variables for 26 countries that we analysed in an exploratory way and for this article mostly descriptively.

In all European constitutions religion is mentioned and regulated in some way. Religion thereby seems to be a topic that EU-member states see worthy of being included in their constitutions. However, there are striking differences regarding the regulation density. Regulation density indicates how important the regulation of religion is for a given society. We measure regulation density in two ways: firstly, we enumerated all religion-related topics (variables) that were to be found in a constitutional text (regulatory depth). This indicator has a minimum of ‘2’ (for Latvia) and a maximum of ‘14’ (for Cyprus, Germany, Poland and Romania) out of 37 possible entries (variables). Secondly, we counted how often the words ‘religion/religious’ and ‘church’ are mentioned in a constitution (word count). The word count indicator for ‘religion/religious’ varies between ‘2’ (Denmark and Latvia)¹⁵ and ‘30’ (Germany). The word count indicator for ‘church’ varies between ‘0’ (Austria, Belgium, Bulgaria, Cyprus, Czech Republic, France, Hungary, Ireland, Netherlands, Romania, Slovakia and Slovenia) and ‘8’ (Greece). This indicates that – compared to the broader term ‘religion’ – the organisational dimension of religion seems to be less regulation-relevant in most countries.

The word count of ‘religion/religious’ does not show any statistical association either to the year of enactment or the year of last modification nor to religious homogeneity (see Table 1). Constitutions of post-Soviet states tend to mention the word ‘church’ significantly less often than countries of the former Western block (Pearson's r: −0.34, p < .1).

Our former discussion of constitutions as data established three main areas of regulation: described self-understanding of the people, assignments of rights and duties concerning individual freedoms and governmental arrangements. Reading constitutions with regard to religion revealed that religion is regulated in relation to these topics. However, we had to refine the scheme slightly: constitutions contain described societal self-understanding (89% of the constitutions), regulations of individual rights and freedoms of religion and worshipping (92%), regulations of the relationship between religious organisations (e.g. churches) and the government (65%), of institutional regulations concerning marriage (42%) and education (54%), and of the relationship between the governmental administration and religion (58%).

These dimensions of provision, however, are targeted in quite different ways.

One way in which the importance of religion is signalled is in its position within the constitutional document. If religion is mentioned in the constitution's Preamble, we can conclude that it is considered as an essential part of the self-understanding of a given society due to the declarative and self-descriptive character of the Preamble. Out of 26 constitutions, 14 include a Preamble, but none of these refer to the society as being secular. But three constitutions describe the nation as secular in other parts of the constitutional text (Czech Republic, France and Slovakia). Six constitutions refer explicitly to ‘religion’, ‘faith’ or ‘God’ within the Preamble (Germany, Greece, Hungary, Ireland, Lithuania, Poland and Slovakia), in the sense that the people are bound to either of them: the Irish constitution, for example, states:

Religion as part of the Preamble is associated with higher regulation density¹⁷: regulatory depth as well as the word count for ‘religion/religious’ are positively associated with religion being mentioned within the Preamble (regulatory depth: Pearson's r = 0.47, p < .05; word count: Pearson's r = 0.51, p < .05).

European constitutions do not show any distinctive pattern when we focus on individual rights of equality and antidiscrimination: in 17 of the 26 constitutions, antidiscrimination is mentioned and equality with regard to religion is mentioned in 16 cases. The phrasings of antidiscrimination and equality do not differ significantly: ‘No person shall be favoured or disfavoured because of sex, parentage, race, language, homeland and origin, faith, or religious or political opinions’ (German Constitution, Art. 3). Only 10 constitutions refer to both (probability that a constitution refers to both: 38%). Their appearance is neither associated with date of enactment, date of last modification, religious homogeneity nor with the state-religion-relationship.

While antidiscrimination regulations are associated with regulation density (Pearson's r: 0.41, p < .05), equality is not. Because all EU- member states have to acknowledge the European Charter of Fundamental Rights, it seems plausible to assume that antidiscrimination and equality regulations concerning religion are not considered as an important topic to be developed any further within national constitutions.

However, religious traditions and religious minorities enjoy different rights and regulations in European constitutions. Six constitutions explicitly support religious traditions of the majority religion (Bulgaria, Cyprus, Greece, Hungary, Ireland, Malta and Sweden). Concerning denomination, no clear pattern was found: majority's support is enacted in constitutions of three countries with orthodox majorities, in two countries with a catholic and in one with a protestant majority. We may understand such support of the tradition of the majority religion as a constituting aspect of the society's self-understanding and as a feature of national identity. The governments’ support of religious minorities is granted only in four constitutions: Cyprus (Armenians, Maronites and Latins), Poland, Romania (both no specific minorities) and Sweden (Sami) acknowledging religious heterogeneity.

Most constitutions regulate individual freedom of religion and worshipping independent of antidiscrimination and equality. Both kinds of regulations are often located in separate articles and in different parts of the constitutional main text. Individual religious freedoms are regulated differently: some constitutions have no explicit regulation (Austria and Belgium), some constitutions confine individual and/or public worshipping with regard to public order (e.g. as formulated in the Constitution of Bulgaria: ‘The freedom of conscience and religion shall not be practised to the detriment of national security, public order, public health and morals, or of the rights and freedoms of others’ (Art. 37).¹⁸ The occurrence of individual religious freedoms relates to the religious majority of a nation-state: out of all constitutions of countries with a catholic majority, 66% restrict either religious freedoms or their public exercise. All countries with an orthodox majority restrict the public exercise of religion.

Antidiscrimination/equality regulations concerning religion seem to emanate out of a different regulatory logic than individual religious freedoms. Freedom and worshipping concern religious practices which have the potential to interrupt public order (and therefore needs to be restricted) while antidiscrimination and equality concern religion as an individual ascriptive characteristic referring to the Human Rights framework.

Fox (2007) sees the preference to particular religions or to significantly regulate religions as government involvement in religion (GIS) but not necessarily as a violation of the separation of religion and state (SRAS).

Out of 26 constitutions in Europe, 17 explicitly contain regulations concerning the state-church-relationship. ¹⁹ The state-church-relationship can be characterised by its closeness – referring to whether a state is secular or if it grants privileges to religious organisations or even comprises of a state-church (indicating increasing closeness). While nine constitutions do not include any regulations of the state-church-relationship, three states claim explicitly to be secular within the body of the text (Czech Republic, France²⁰ and Slovakia) and only Denmark still has a state-church.²¹

Most of the 17 constitutions regulate particular church-related topics: the governmental regulation of church property as one historical core issue of secularisation can be found in the constitutions of Cyprus, Germany, Greece and Ireland (Italy has a special arrangement with the Vatican). Only in Romania, churches as organisations enjoy explicit governmental support.

Financial privileges are granted in Belgium, Denmark, Luxembourg and Sweden.

The state-church-relationship does not depend significantly on the historical context at the date of the constitution's enactment. Transition of a nation-state (be it into a post-Soviet state or EU-membership) does not produce similar national patterns (see Table 1). Only the year of the last modification correlates with the state-church-relationship: Pearson's r: −0.35, p < .1. The negative sign indicates that more recent changes work in the direction of a less detached relationship. A closer state-church-relationship involves a higher regulatory density.

Societies with a higher degree of religious homogeneity seem to prefer a closer state-church-relationship (we cannot, however, establish the direction of influence here). This correlation is significant (p < .1) for all former non-Soviet countries, but is not significant for the subgroup of the post-Soviet states.

There are two areas of the social sphere that are highly regulated with regard to religion in most of the constitutions (Table 2): the generational order (including marriage) and education – both involving the ‘reproduction of society’. Both are areas where the private and the public sphere are linked. They are targeted by government and churches alike and tend to end up regularly as bone of contention in ideological conflicts (as most recently, the legal recognition of same-sex marriage).²²

Nine constitutions establish a generational order in which parents retain responsibility for the religious and moral instruction of their children. This guarantees the parents’ sovereignty over their children and protects it against interventions by the state and/or religious organisations.²³ Out of these nine constitutions, seven are enacted in countries with a catholic majority (78%). Five constitutions explicitly establish religious marriages as equal to civic marriages (Belgium, Cyprus, Lithuania, Luxembourg and Romania – all countries with a catholic majority). This guarantees sovereignty of the state to legally justify (reproductive) relationships. The societal importance of this topic becomes visible by the controversial debates about same-sex marriage, which are still condemned publicly by the Pope.

Referring to education, 14 out of 26 constitutions regulate the relationship between religious education and the state. In five out of eight constitutions, which guarantee religious education as a matter of the state, the state-church-relationship is regulated in favour of separation. In the remaining three, privileges are granted to religious organisations. This combination of regulations seems to indicate some defence mechanism of the state against religious influences (Stepan 2000). Four of these constitutions are enacted in countries with a catholic majority, two in orthodox countries and two in countries with a non-religious majority.

It seems that particularly in countries with a catholic majority, constitutions have the function to secure the state's sovereignty concerning generational order, marriage and education. Additionally, we may conclude that constitutions articulating a loose state-church-relationship contain slightly more regulations concerning generational order, marriage and education in order to sustain the state's power.

Constitutions regulate how governmental administration is related to religion. Such regulations are found in 58% of the European constitutions. They comprise of regulations allowing some (or all) members of the administrative staff to formulate their oath without a reference to religion or contrariwise: pledge them to a mandatory reference to God. Additionally, in some cases the possibility to refuse military service due to religious beliefs is regulated (only 7 out of 26 constitutions). Furthermore, regulation of the head of state's religious belief is found in the constitution of Denmark, the Netherlands and Sweden (referring to the monarch). Such articles regulating the relationship between religion and government are significant indicators for secularity. They show directly how religion affects these states’ administrations. Most of the European constitutions restrict the influence of religion on administration.

The intertwining of religion and government is positively associated with the date of enactment, regulatory depth and the word count for ‘religion/religious’.

Our analysis provided an original up-to-date data set on how religion is negotiated within the constitutional texts of the EU-member states. The analysis focused on where and in which thematic fields religion is regulated. The analysis refers to how the people of a country perceive the necessity to legally regulate religion. It signifies the legal framework for individual opinions and for the degree of secularity in all EU-member states comparatively.

The most important point of our analysis so far is that within the constitutions of the EU-member states, patterns regarding religion are difficult to detect. If patterns emerge they present themselves in the following ways: all constitutions mention religion; none of the constitutions’ Preambles define an European nation-state as secular; all constitutions regulate and limit the degree to which churches are allowed to interfere within governmental affairs (e.g. education or governmental administration); all European constitutions are bound to the European Charter of Fundamental Rights via accession. Religion appears in all constitutions at least in one of the following five dimensions of provisions: (i) religion as part of societal self-understanding, (ii) regulations of individual religious rights and freedoms, (iii) the relationship between religious organisations and governments including (iv) regulations of marriage and education and (v) the relationship between governmental administration and religion. In this rather superficial way, we might talk of similarities as European isomorphism. In this regard, none of the EU-member states’ constitutions enacts an explicit separation of law and religion.²⁴ However, the comparison revealed that Europe is not characterised by a similar degree of secularity within the constitutional texts of its member states.

In particular, the increasing variance of regulatory depth in more recent constitutions indicates that isomorphism as explanation falls short. Concerning our expectations derived from neo-institutionalism, it also challenges theories that religious heterogeneity plays a role at least when it comes to governmental arrangements. Additionally, major denominations are important with regard to individual religious freedoms (protestants), state-church-relationship (protestants and orthodox) and family and education (catholics) – all indicating that religious power relations matter. One might say that the power dimension of historical paths becomes visible here.

Hence, it cannot be concluded, that path dependency theories are any more suitable: historical events (the Second World War, date of EU-accession and collapse of the Soviet Union) are only associated with regulatory depth and state-church-relationship. The increasing variety of regulation density and the relationship between state and religious organisations imply that constitutions in Europe ‘react’ increasingly to national particularities and power relations. The influence of denominational majorities point into the same direction: different protestant and catholic social teachings and the influence of a trans-national Catholic Church might lead to competition concerning the legitimacy of norm setting and in additional regulatory requirements. This highly relates to findings by Pickel (2009: 115) that the denominational legacy of a country is of great importance for the vitality of religion.

Despite this, the analysis does not support convincingly the understanding of constitutions as descriptions of states as self-contained units (particularly with regard to the influence of the Catholic Church). Paths, however, can be traced clearly within the constitutional texts: occurrence of religion within the Preamble is followed by a higher regulation density; regulations of generational order, marriages, and responsibilities for education are associated with intensity of regulations referring to religious organisations, while regulations of administrative arrangements are associated with antidiscrimination, equal opportunity-regulation and regulations referring to religious organisations. Such intra-text-paths suggest that constitutions answer to both – nation-specific power relations and intra-text logic. Intra-text logics, however, may point again into the direction of international isomorphism as described in the checklist by Blaustein (1994).

If Europe is to be described as an exceptional case, the analysis suggest that at least the constitutions of the member states show quite different degrees of secularity, with regard to quite different dimensions of provisions. The increasing variance of regulation depth in more recent constitutions can, however, be read as an indicator of an increasing degree of secularity: while older constitutions include religion as a silent ingredient, the relationship is regulated explicitly in younger constitutions. In contrast to classic/usual interpretations, a more explicit regulation might indicate the resolution of religion as implicitness – meaning that regulation may take place in countries in which the silent ingredient of religion is not silent anymore and societal conflicts make it necessary to regulate topics legally that have been unquestioned so far.²⁵

Considering such diversity among European constitutions it is not surprising that the reference to religion within the Preamble of the European Constitution has led to such heated debates. In the face of our analysis, religion as a common denominator for European integration appears to be unlikely.

Although we are not able to provide an analysis about the effects of constitutional texts, we are able to conclude that constitutions as texts provide a helpful insight into the legal framework within which policies are developed and within which people believe and act. The anchorage of religion in constitutions helps us understand why religious symbols and worshipping in public, the moral public interventions of representatives of the Catholic Church or religiously motivated demands for the restriction of freedom of speech become significant (even heated) issues in public debates. With regard to religion, constitutions even provide a quite fine-gained guidance concerning different dimensions of the social sphere on five dimensions of provisions: this indicates that a reasonable analysis of religion in the public sphere has to take into consideration that religion is a multi-level and multi-dimensional phenomenon affecting social order in different domains. In this regard, constitutions even provide insight into how the people of a nation-state envision their ‘we’ involving religion.

No potential conflict of interest was reported by the authors.

Annette Schnabel is a professor for sociology and social theory at the Heinrich-Heine-University Duesseldorf. Her research interests cover, among others, religious and national identities, welfare state research, gender and social movements.

Kathrin Behrens is a PhD student and research assistant at the Institute of Social Science, Department of Sociology, Heinrich-Heine-University, Duesseldorf. Her research interests focus on social theory and methods of empirical social research, sociology of religion and sociology of national identity.

Florian Grötsch works in the field of European research funding and resources management. His current scientific interests are focusing on the influence of Religion in Europe, the dynamics in the religious field and the European (dis-) integration process.