Multiple secularisms


back to issue

WHAT is the variety of possible, and actual, democratic patterns of state-religion-society relations? I want to suggest six responses, not all of which I can fully develop and defend that I will weave through this essay.

First, patterns of state-religion-society relations that happen to co-exist with democracy at any given time are best seen as contingent, socially constructed, normative political arrangements, rather than as fixed, normative models.

Second, religions are transnational and we are living in a time of global movement of populations who are able to stay in contact with the sources of their religion, and influence the interpretation and practice of their religion, as well as being influenced by it, via the media, the internet and actual religious leaders from their countries resident in the diasporas. Thus, most contingently settled patterns of state-society-religion relations are becoming contingently unsettled and in need of some social, religious and political reconstruction.

Third, modern political analysis of democracy, while it absolutely requires use of such concepts as voting and relative freedom to organize, does not necessarily need the concept of secularism. However, religion matters. Democratic institutions need sufficient political space from religion to function, just as citizens need to be given sufficient space by democratic institutions to exercise their legitimate right to participate in public life. I call this mutual space giving, the ‘twin tolerations’.1

Fourth, if we want to use the concept of secularism it should be conceptually reformulated and used as ‘multiple secularisms’, for many of the same reasons that Eisenstadt reformulated and used the concept of ‘multiple modernities’. Western Europe and North America have three quite different patterns of secularism, all under some pressure. Many analysts and advocates think that the ‘separatist pattern’ found in France and the United States is the norm in modern western democracies. It is not, and the two other patterns are not strictly separatist.

Fifth, if we closely examine the three countries in the world with large Muslim populations whose political systems receive the highest ranking for the quality of their democracy – India, Senegal and Indonesia – none of them follow any of the three dominant models found in Western Europe or North America. Their circumstances were different and they socially constructed, or invented, variants of a new pattern, which I will discuss.

Sixth, all four state-society- religion patterns I discuss that have coexisted with democracies respected the ‘twin tolerations’; thus the ‘multiple secularisms’ of modern democracies. However, all four patterns, including separatist secularism, have instances of ‘twin intolerations’.


Is the concept of ‘secularism’ necessary to analyze democracy? Robert Dahl, Arend Lijphart and Juan L. Linz (the first three winners of the Skypte Prize, often called the Nobel Prize of Political Science), have not felt the need to include any discussion of secularism in their definitions of modern democracies, much less of secularism being a ‘necessary condition’ of a democracy. Dahl, in his elaboration of the ‘eight institutional guarantees’ that must be created for the functioning of a democracy, or what he prefers to call a ‘polyarchy’, nowhere mentions secularism.2 Neither does Lijphart in his analysis of the thirty-six long-standing democracies in the modern world.3 Linz (and I), in our book, Problems of Democratic Transition and Consolidation, also decided not to use the concept of secularism in characterizing any of the five basic types of political regimes in the modern world because each type includes some regimes that call themselves secular.4

Linz and I are, of course, aware that many countries that are now democratic could not have become so without processes, often called ‘secularization’, that facilitated the reduction of religious prerogatives in the polity. But it did not seem to us conceptually or empirically correct to use secularism as a defining or distinguishing characteristic of any regime type per se.


Despite my general reservation about the term secularism, in my current research, building on S.N. Eisenstadt’s and Sudipta Kaviraj’s work on ‘multiple modernities’, I use the concept of ‘multiple secularisms’ to get round some of the difficulties of a single meaning of secular, and to help me identify, and analyze the great variations in, religion-state-society relations that exist in modern democracies and autocracies.


Many people view secularism through a narrow prism, defining it as the end result of the narrative of modernity. Some simplistic versions of modernization theory imply that there are at least four reinforcing and compounding dichotomies related to modernity and religion: traditional versus modern societies, high religious practicing societies vs. low religious practicing societies, little separation of church and state vs. strict separation of church and state, and non-democratic regimes vs. democratic ones. According to this view, properly modern polities are on the right hand side on all four dimensions. In a strict sense, however, perhaps only France between 1905-1959 is found on the right hand side of all four dichotomous sets.

S.N. Eisenstadt, in his important Daedalus article on ‘multiple modernities’, draws attention to the USA as the first western case of an exception to the modernity thesis because the USA, though obviously a modern democracy, falls on the ‘high religious practicing’ side in the dichotomous set listed above.5 In the same issue of Daedalus, Sudipta Kaviraj correctly notes that India in some respects has been modernizing, while parts of Hinduism have been ‘retraditionalizing.’6 In a later magisterial work Kaviraj develops a convincing argument as to how and why different ‘modernities’ were socially and politically crafted.7


In general, I find it more useful when discussing democracy and the world’s religions to speak of what I have called the ‘twin tolerations’. By this I refer to the minimal degree of toleration democracy needs from religion and the minimal degree of toleration that religion needs from the state for the polity to be democratic. Religious institutions should not have constitutionally privileged prerogatives which allow them authoritatively to mandate public policy to democratically elected officials. The minimal degree of toleration religion needs from democracy, if a democracy respects Dahl’s eight institutional guarantees, is not only the complete right to worship, but the freedom of religious individuals and groups to publicly advance their values in civil society and to sponsor organizations and movements in political society, as long as their public advancement of these beliefs does not impinge negatively on the liberties of other citizens, or violate democracy and the law, by violence or other means.8

After a period of self-secularization, the Christian Democratic political parties of Europe, as Stathis N. Kalyvas has shown, became autonomous democratic parties.9 A legitimate question to ask is whether the current Muslim-influenced ruling party in Turkey, the Justice and Development Party, will be allowed by Atatürkist forces to complete a similar process of self-secularization without being subject to a 1997 style ‘soft coup’ in order to defend secularism.


My concern in this essay is to very briefly examine the patterns of ‘secularism’ as they relate to state-religion relations in modern European and North American democracies, with the intention of reflecting on what these patterns mean, and do not mean, for religion, peaceful societal pluralism and democracy in polities such as India, Indonesia and Senegal, each with sharply different histories and conditions.

Empirically, if we examine contemporary western democracies, two different patterns, or models, with widely accepted names, immediately emerge if our focus is on actual state-religion relations. France, the USA and Turkey are close to the ‘separatist’ pole, and the Scandinavian polities, the UK and Greece are close to the ‘established religion’ pole. Moreover, particularly in Western Europe, a third pattern (so varied it is not a model) exists with no widely accepted name. There are countries such as Germany, Switzerland, Belgium and the Netherlands that are neither near the ‘separatist’ nor ‘established religion’ poles. All three patterns coexist with democracies. But as I shall document, they can also coexist with, and support, non-democracies.

Following a variety of quite different historically negotiated agreements concerning state-religion relations, and different historical sequences, countries that are now democracies have distinctive ways of respecting the ‘twin tolerations’ and of constructing and reconstructing patterns of democratic secularisms in public life.

As I argue later in this essay, while most citizens of countries such as India, Senegal and Indonesia think democracy is appropriate for their countries, ‘complete separatism’ and ‘official state religions’ as democratic models have been in the main, and in my judgment correctly, seen as inappropriate to their societies. Thus new models of secularism were appropriately invented.

Let me begin my analysis of these various state-religion possibilities by first looking briefly at what are the clearest models of these multiple secularisms in Western Europe and North America, the ‘separatist’ and ‘established religion’ models.


The fact that both France and the USA are close to the ‘separatist’ pole, makes many commentators assume that separatism is the normatively preferable, and empirically predominant, form of modern secular democracies. But, for comparative purposes, particularly for USA-centric readers, it is important to be aware of how many existing Western European polities violate USA norms of a ‘wall of separation between the state and religion’, but are nonetheless strong democracies. If we exclude the United States, a survey of state policies in twenty-five other western democracies documents that all of them fund religious education in some way, 76% of them have religious education in state schools as a standard offering (many, but not all, with the option not to attend), 52% collect taxes for religious organizations, and 36% have established religions.10 See Table 1.


Percentage of Western Democracies With State Supports for Religion (excluding the US)

Form of State Policies of Support ( or monitoring) of Religion


Government funding of religious schools or education


Religious education standard (optional in schools)


Government collects taxes for religious organizations


Official government department for religious affairs


Government positions or funding for clergy


Government funding of religious charitable organizations


Established/Official religion


Some clerical positions made by government appointment


Table 1, by itself, should make it absolutely clear that complete separation of religion and the state is not a necessary condition for democracy to function, and that complete separation of religion and the state is not the empirical norm in modern western democracies.


We should also be aware that ‘separatism’ as a model, even in its democratic variants, can historically be established for completely different purposes, with sharply different results. France and the USA are the two long-standing democracies with the greatest legal separation between religion and state. However, the forms of separation of religion and state are polar opposites in their origins and consequences. In France, the Catholic Church had been an intrinsic part of the pre-revolutionary regime and continued to be a powerful part of the anti-Republican coalition in the Third Republic. ‘Laicité’ was thus created in France in 1905 as a clerically-hostile form of ‘freedom of the state from religion’. In sharp contrast, the First Amendment in the USA was passed as a clerically-friendly form of ‘freedom of religion from the federal state’.

But the US story is more complicated than this phrase implies. By and large of course, American societies, and American colonies, created a very supportive context for religion. However, the U.S. Constitution’s First Amendment, which states that the ‘Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof’, is misunderstood by many contemporary U.S. citizens. The amendment did not prohibit the thirteen original states from having their own established religions. The First Amendment was meant to ensure that the federal Congress could never establish one official religion for the United States as a whole. In fact, on the eve of the revolution, only three of the thirteen colonies – Rhode Island, Pennsylvania, and Delaware – had no provision for an established church. Even after the revolution, the South Carolina constitution of 1778 established the ‘Christian Protestant Religion’. Four New England states continued for some time with state subsidized, largely congregational, churches.11


Moreover, even if it exists, complete ‘separatism’ may produce its own tensions and inflexibilities in democracies. The classic French and U.S. models have not changed fundamentally since they were created, but their societies certainly have in ways that challenge their models’ capacity to sustain the twin tolerations. France, because of its model of republican representation of interests, finds it difficult to manage the ethnic and religious demands of some of its second and third generation Muslim citizens. Affirmative action is virtually illegal. The census cannot collect much information about religion. French citizens with Arabic-originated names have been in a disadvantageous position in their job applications, despite all the rhetoric of republican equality.

Moreover, the French state has had more restrictive policies toward Muslims than almost any other Western European state.12 Such policies include the ban on students wearing of headscarves in public schools, the bureaucratic barriers against mosque construction, and the very low level of funding for Islamic instruction.13

The USA, because of its great toleration of virtually all activities of religious groups, finds it politically, and to some extent even constitutionally, difficult to control some of the demands of rapidly growing and politically assertive fundamentalist religious groups from virtually all religions.


Separatist secularism should be analyzed both as a dichotomy and as a continuum. There can be separatist secularisms with very low state controls on minority and majority religions, and secularisms with such high controls on minority and majority religions that the label ‘separatist autocracy’, or ‘authoritarian secularism’ might be appropriate. Low state controlling separatist secularism is fully consistent with the twin tolerations; state controls at the high end of the continuum are not.

Turkey is seen by many scholars as following the French 1905 model of strict separation of church and state (laïcité), and thus democratic secularism. The fallacy of this approach is its exclusive focus on the separatist versus non-separatist dichotomy and its implicit modernist assumption that since Turkey is on the separatist side, this implies its secularism is fundamentally democratic. With dichotomous lenses France and Turkey are both classifiable as forms of laïcité separatism, and are thus seen by many as politically similar. But from the perspective of a continuum we can, and should, classify political systems by how much, and how, they control minority and majority religions.


Consider the following six sharp differences between Turkey and France. First, in Turkey, every single publicly read prayer for the Friday ceremonies in mosques is written by one state office, the Diyanet (The Department of Religious Affairs). There is nothing comparable for majority or minority religions in France. Second, in Turkey, all mosques that are allowed to have public ceremonies are controlled by the state and clerics in these mosques must be authorized and approved state employees. No similar nationalization and control of religious establishments and of the clergy is found in France.

Third, in Turkey, two large groups of citizens – Sufis, who are Sunni Muslims, as well as Alevis, who self-identify themselves as not traditional Sunni Muslims – are excluded from the financial support of the Diyanet. More importantly, from the perspective of twin tolerations, neither Sufis nor Alevis are allowed to practice their religion in public spaces. It is illegal for them to construct religious spaces. While it is somewhat difficult for some religions to build places of worship in France, no mainstream religion is prohibited from constructing such buildings.14


Fourth, in Turkey, the graduates of public Islamic Imam-Hatip schools are not allowed to attend universities, except the departments of theology, whereas in France the graduates of Catholic schools have no such exclusion. Fifth, it is banned in Turkey to teach the Qur’an to those under age 12, while in France there is no prohibition over teaching the Qur’an or Bible. Sixth, the Turkish state still does not fully recognize the rights of association and temple construction of non-Muslim minorities, despite the reforms the current AKP government has pursued to meet European Union norms.15 Each one of the above practices violates one of Dahl’s seven institutional guarantees that he argues are necessary for a democracy.

Jonathan Fox has for over a decade been working on a data base that examines laws that are passed by states which control religion. The lowest score is zero if there are no controlling laws (but of course there could be controlling state behaviour). When one examines the constitutions of Turkey, France and Senegal (which was a French colony that from 1848 until independence in 1960 sent elected deputies to the French Assembly every year that elections for this chamber were held in France), one notes that all three declare their commitment to separatist laïcité. However, Senegal has negotiated its relations between Islam and Catholicism, and between the four major Sufi orders, while passing only one law that controls religion and thus receives a score from Fox of 1.


France, given its origins in a hostile separation of church and state and wanting to have ‘freedom of the state from religion’ scores 6 on the Fox scale. Turkey, following a policy I would call ‘control all religions, but financially support (and control) the majority religion’, scores 15 in Fox and in our continuum.16 I should note that this overall combined score for control of minority and majority religions puts Turkey in a higher religious controlling position on our continuum than some non-democratic Islamic Middle-Eastern political systems such as Egypt (14), Algeria, (9), Morocco (12), and Tunisia (11). Most revealing is that only four of the nineteen Muslim majority countries in the Middle East and North Africa are more controlling of minority religions than Turkey; Turkey (9) Qatar (10) Iraq (12) Iran (13), and Saudi Arabia (14).17

With this caveat about continuums stated, let us return to misreadings of Turkey’s ‘control all’ separatist secularism by two of the most influential interpreters of modern Turkey. Bernard Lewis writes that Atatürk’s goal was to ‘reduce Islam to the role of religion in a modern, western nation state’, while Dankwart Rustow argues that the ‘new [secular] identity was fully [and democratically] accepted’.18 However, such comments obscure the fact that Turkey’s secularism was imposed from above by Atatürk, and the fact that the military, since the experiment with democracy began in 1950, has imposed three formal coups (1960, 1971 and 1980) and staged one ‘soft coup’ in 1997.


The military, the self-appointed guardians of secularism, allied with other Kemalists, particularly the Constitutional Court, to ban four major Islamic parties.19 In the summer of 2008, the Constitutional Court came within one vote of closing the Islamic influenced, but by its practices, democratic, ruling AKP party.20 The 162 page indictment concerning alleged attacks on secularism advanced no evidence of actions against democracy but was largely devoted to (the still unrevealed) ‘hidden agenda’. If the Constitutional Court had closed the party on the basis of such evidence, it would have been a case of ‘secularist fundamentalism’.21

The current 1982 Constitution was drafted by a committee vetted by the military during a period of military rule. The final version was revised by the National Security Council. The constitution was approved in a plebiscite, but it was forbidden for anyone to campaign against ratification of the constitution.22 Article 2 reasserts that the Turkish Republic is secular. Article 4 says Article 2 can never be changed even by a constitutional amendment. Article 24 reasserts that ‘education and instruction in religion and ethics shall be under state supervision and control.’ Article 24 further says, in a key clause that was used to constitutionally ban the Welfare Party in 1997, that ‘No one shall be allowed to exploit or abuse religious systems.’


The Turkish Constitutional Court (along with the military and the Republican People’s Party, a core institution of Turkish ‘hard line’ secularism) prepared the way for the 1997 ‘soft coup’. In their January 1997 ruling, the Constitutional Court said that secularism was not merely ‘separation of religion and state’ but ‘separation of religion and worldly affairs... [Secularism] means separation of social life, education, family, economics, law, manners, dress codes, etc. from religion.’23 This Constitutional Court ruling is a clear example of state imposed secularism that so ‘polices’ citizens’ religious practices that it violates Dahl’s institutional guarantees for citizens’ rights and my concept of the ‘twin tolerations’.

It goes without saying that many authoritarian regimes, like contemporary Syria, call themselves secular. Indeed, many of the leaders in the Middle East in the immediate post-independent period were not Islamists, but like Egypt’s Gamal Abdel Nasser, Algeria’s Ben Bellah, or Tunisia’s Habib Bourguiba, secular nationalists who used authoritarian measures to control and repress many religious leaders in the name of modernity. Separatist secularism is obviously neither a necessary, nor a sufficient, condition for social peace and democracy.


Officially established religions – varieties and problems: I next turn my attention to the second pattern or model, the ‘established religion’ model. Given the data in Table 1, which shows that more than a third of western democracies have such established religions, we need to analyze the possible interrelationships and tensions between democracy, the twin tolerations, and ‘official established religions’.

A polity with an established religion can respect the ‘twin tolerations’. For example, all the Scandinavian states – Norway, Sweden, Denmark, Finland and Iceland – have constitutionally embedded Evangelical Lutheranism as their established religion.24 However, church officials in the social democratic welfare state era by and large did not enjoy prerogatives to block policies of democratically elected officials or to constrain the individual religious freedom of the members of the majority or minority religions. The United Kingdom has established churches in England and Scotland, and Greece has established the Greek Orthodox Church, all of which have more prerogatives than the Scandinavian official churches; and as their populations change, they also will have to change.25


Obviously, some ‘official or established religions’ violate the ‘twin tolerations’ and some not. Thus, if we want to assess the impact of ‘established religions’ on individual religious freedom and democracy, it is best to analyze established religions not as a dichotomous variable (established religion versus no established religion) but as a continuum of degrees of state control within the polities that have ‘established religions’. For example, the valuable ‘Religion and State Database’ allows us to construct a continuum of state control of majority religions.26

This continuum has a four point scale for eleven variables concerning state control of majority religions. Composite scores on this continuum range from zero (Sweden, Denmark, Norway), to 1 (Bangladesh), to 8 (Pakistan), to 11 (Egypt). We can also construct a continuum for state control of minority religions based on scores for sixteen variables. The composite scores for control of minorities in polities with ‘state religions’ range from 1 (Denmark), to 2 (Norway and Bangladesh), to 13 (Pakistan), to the two most controlling (Iran and Saudi Arabia), with scores of 34 and 38 respectively. See Table 2.


Some Established Religions in Democracies and Non-Democracies:

Their Range of Controls Over Majority and Minority Religions


Restrictions on Majority Religion

Restrictions on Minority Religions

Total Score

Saudi Arabia
































United Kingdom








Sources: All data are collected from the ‘Religion and State Dataset’ gathered by Jonathan Fox, Department of Political Studies, Bar Ilan University. The data represent values calculated for the year 2002. The first column shows a composite score for restrictions on the majority religion, the second column shows a composite score for restrictions on minority religions, and the third column shows the combined total. All values are calculated within the dataset. Please contact Jonathan Fox directly for access to the dataset and a codebook for the data. In order to view a discussion of these data, please see his ‘World Separation of Religion and State into the 21st Century’, Comparative Political Studies 39(5), 2006, pp. 537-569 or the previously cited ‘Separation of Religion and State in the Twenty-First Century’ by Jonathan Fox and Shmuel Sandler in Comparative Politics.


So far, we have clearly documented that some countries have established religions, but are nonetheless inclusive democracies in which the rights of religious minorities and majorities are respected. But can we make reasonable estimates about the societal and religious conditions in a society that would be relatively supportive, or relatively unsupportive, of such outcomes? The following proposition seems plausible. The greater the degree of religious homogeneity in a polity, and the less the intensity of religious practices, the easier it will be for an established religion, and an inclusive democracy, to coexist.

Let us compare Denmark and Northern Ireland. Denmark – the country with an established religion that used to (but no longer does) score best on tolerance towards minority and majority religions has high religious and denominational homogeneity and low intensity of religious practice. For much of the 1980s, roughly 95% of Denmark’s population were Lutherans and 3% were other Protestants. Between 1973-1999, the percentage of respondents who said they went to church at least once a week was the lowest in Europe, 3-7%.27


Northern Ireland is near the opposite pole; 53 % of the population are Protestant, and 44% are Catholic.28 Also, between 1973-1992, the percentage of all respondents in Northern Ireland who said they went to church at least once a week never fell below 54%.29 Given this combination of religious heterogeneity, religious intensity, and ethno-religious conflict, the Anglican Church was disestablished in Northern Ireland in 1920.30

John Madeley has tables and maps of historic ‘mono-cultural’ zones in Europe. The only country in Western Europe that has a ‘state religion’ and is not in one of Madeley’s historic ‘mono-culture’ zones, the United Kingdom, has witnessed historic conflicts over Catholic rights. Now, it will be interesting to watch how the prerogatives of the official religion (its monopoly of entitled representation the House of Lords, in Anglican schools, and indeed its veto on some matters of state school curriculum) is increasingly coming into tension with new socio-religious migrant populations, and thus has more contested issues concerning the status of the established religion than in the Scandinavian countries.


Religious homogeneity in the United Kingdom was never as high as in the Scandinavian countries nor religious practice as low. However, in the last few decades, migration of committed Muslims from Pakistan and Bangladesh, Hindus from India, evangelical Protestants from Africa and the Caribbean, and a new wave of Catholics from Poland, has reduced religious homogeneity and increased religious intensity. The state-society-religion formula has undergone great changes in its societal and religious components but not enough in its overall formula.

In the year 2000 Sweden started a process of disestablishment of the Evangelical Lutheran Church and an opening up of numerous lines of contact with Muslim immigrants. Denmark did neither. Sweden managed the cartoon crisis much better than Denmark.

There can be a Muslim majority country that is a democracy with a state religion. Note that Bangladesh (which the Polity index and the Gastil index both ranked as a near democracy for much of the 1993-2006 period) has the same overall score in Table 2 as Finland concerning control of minority and majority religions.


Many authoritarian regimes in Muslim majority countries with official state religions such as Pakistan and Egypt do of course impose numerous constraints on the freedom on individuals from both the majority and minority religions. However, the reasons for this, in these two cases, may have their origins in military efforts to simultaneously gain legitimacy by supporting some political Islam policies, while also using their state controls over the official religion to repress any form of independent Islam. In the case of Pakistan, causation is further complicated by geo-political factors; the U.S., for its own security reasons, has been supportive of some jihadist groups based in Pakistan so that they could fight against the Russians in Afghanistan.

A polity with an official religion whose leaders establish a ‘theocracy’ of course violates both aspects of the ‘twin tolerations’. In Iran, anti-fundamentalist forces, despite having limited access to the media and having most of their leading political candidates barred from running for election, won around 70% of the vote in the parliamentary, municipal, and presidential elections between 1997-2001.31 This meant that for this period Iran was a ‘diarchy’, in that it had two different principles of legitimacy and authority coexisting and competing within the same state. Nonetheless theocratic judicial and security forces were able to impose a reign of ‘twin intolerations’. The theocracy blocked all attempts by elected officials to change the system. Likewise, the theocracy repressed any form of religious freedom for ordinary believers of all religions and put the majority of ayatollahs under house arrest where they remain to this day.

Like the separatist model, purely as a model of state-religion relations, the ‘official religions’ model can, and does, vary greatly, in this case from inclusive democracies to exclusive theocracies.


The hypothesis that ‘official religions’ fare best in a context of high religious homogeneity and low religious intensity is borne out when we consider major European countries that in the democratic age have never had an established church. Since World War I none of the historically most multi-confessional polities in Western Europe – the Netherlands, Switzerland, Belgium or Germany – have established religions.32 Rather, they are the home of intensely negotiated ‘consociational’ power and ‘space sharing’ arrangements in politics and in religion which we will discuss briefly.

There exist several important countries such as Germany, Belgium, Switzerland and the Netherlands whose pattern of state-religion relations falls between the ‘separatist’ and the ‘state religion’ models. I discuss them briefly in the ‘Twin Tolerations’ chapter in Arguing Comparative Politics (pp. 219-223). None of them are separatist. Indeed all have politically negotiated spaces and prerogatives for religion in their society that do not violate the ‘twin tolerations’. It may be that these countries, in spirit and practice, are closer in some respects to the patterns I will discuss in India, Senegal, and Indonesia. I do not, however, have time or space in this piece to go into these ‘mixed model’ European countries.


Respect all, support all ,secularism is found in India and Senegal (and with some reservations in Indonesia).. India, with its Muslim minority of 140 million citizens, and heavily Muslim majority Indonesia and Senegal, are some of the highest ranked democracies in the developing world. In the most recent ranking of all the countries in the world on a democracy scale in Ted Gurr’s Polity, of the 43 Muslim majority countries ranked, Senegal and Indonesia received the highest scores and India has been ranked at that level for over thirty years. Polity uses a 21 point scale with plus 10 being the most democratic and minus 10 the least; India, Senegal, and Indonesia all received a plus 7 or plus 8. The highest ranking Arab majority countries, Jordan and Yemen, were at least nine points lower at minus 2.

The other frequently cited survey is the Gastel Index of Freedom House which despite its different panel of experts, methodology and political orientation, arrives at the virtually same overall rankings. On the Freedom House’s seven-point scale for political rights (with 1 being the most democratic and 7 the least democratic), in 2008, Senegal, Indonesia, and India all receive the score of 2. None of the 16 Arab countries in the same survey received a score better than 5.33

What is theoretically interesting for our examination of varieties of democratic secularism is that none of these three democratically exemplar countries – India, Senegal or Indonesia – have either a U.S. or French-style notion of secularism, meaning a strict separation of religion and the state, or an ‘established religion’. All three have variants of ‘equal respect, principled distance, and equal (and substantial) support for all religions.’34 In Senegal, they call their model ‘Laïcité Bien Comprise’; in Indonesia, ‘Pancasila’ (state of five principles); and in India, simply ‘Secularism’.35


The financial support of religions on the part of the state in Senegal, India and Indonesia certainly violates French or U.S. ideas of a strict separation of religion and state, but does not violate citizens’ human rights, or violate the necessary spheres of autonomy that I have has identified as the ‘twin tolerations’ that modern democracies need. Indeed, the strong majority of religious leaders and followers alike in India and Senegal, and to a lesser extent Indonesia, have arrived at a mutual accommodation with, and even support of, a democratic polity and their own version of a ‘secular state’.36 All three countries have versions of the secular state that can impose some normative and constitutional constraints on religious majoritarianism, and on possible religious violations of human rights.


Senegal is an intensely religious country. During the critical period from 1880 to 1920, when systems of accommodation and rituals of respect were developed, a high percentage of the Black Senegalese population in the four urban communes were Catholic. Thus Senegal developed its patterns of respect when it was religiously heterogeneous and intensely religious. A French or Turkish style of religiously hostile secularism would not have produced the patterns of policy cooperation between the Sufis and Secularists I document elsewhere.37 Thus, even the French radically altered their religiously hostile variant of secularism and tried to be seen as a religiously friendly ‘Islamic power’.38 This French policy change facilitated the turn, by the founder of the Mouride Sufi order in Senegal, Amadu Bamba, away from any idea of complete fusion of religion and the state, and toward not only Sufi-Secular accommodation, but toward Sufi-Secular mutual respect.


For modern analysts of comparative democracy, one of the greatest questions of our day is how India has persisted as a democracy despite enduring inequalities, great poverty, and linguistic and religious diversity. I will argue that part of the answer has to do with the adoption of the ‘respect all, support all’ secularism similar to Senegal’s and so brilliantly discussed for India by Bhargava.

Let us briefly look at the inappropriateness of either an ‘official state religion’ or a complete ‘separation of religion and state’ for India upon independence in 1947. Certainly, even more than Senegal, India with its great religious heterogeneity and religious intensity was extremely far from being an example of Madeley’s ‘mono-cultural, mono-confessional zones’ he associates with inclusive and non-conflictual ‘official religions’. Despite the desires of some Hindu fundamentalists in such organizations as the RSS, and even some Hindu nationalists inside the Congress party, fears of religious conflict moved the Constituent Assembly away from any form of a single state religion.


Likewise the religiously hostile, extremely controlling, Turkish secularism, or even French secularism, with its focus on ‘freedom of the state from religion’ might have made it difficult to have retained the support of deeply religious citizens of the new state. Moreover, the U.S. style of secularism, with its focus on ‘freedom of religion from the state’ would not have allowed Nehru and Gandhi to have used democratic state power to insist that Hindus open temples to untouchables. Such prohibitions against untouchable temple use were a violation of the ‘twin tolerations’. But the ‘multi-value’ Indian model that Bhargava analyzes facilitated the legitimate use of democratic state force in the name of inclusion.

In many countries religious linguistic and religious minorities feel persecuted, or at least marginalized, by the state, and thus do not identify strongly with the polity. In answer to the question ‘How proud are you to be American, Austrian, Canada etc?’, of the eleven long standing federal democracies polled by World Values, only the USA and Australia have a higher percentage of respondents who say they are proud of being members of the polity than in India. In ‘all India’, 61% gave the answer ‘very proud’ and 28% said ‘proud’. The responses by ‘Indian Muslims’ were virtually identical, 58% and 31%. More importantly for our concern in this paper with democracy, all the major religions in India, unlike in Sri Lanka or Pakistan, are strongly supportive of democracy. See Table 3.


Support for Democracy in India as a Whole and for Minority Religions and Marginal Groups


All India





Scheduled Caste

Scheduled Tribe

Very Poor


Democracy is always preferable



















Sometimes authoritarianism is preferable










No difference










Don’t Know/ No answer






























Note: All figures in column percentages. Figures in parentheses in the first row are for per cent of valid responses if the DKs are treated as missing data.

Source: National Election Study (India) 2004.


Not only were the positive responses of Hindus and Muslims concerning ‘Indian identity’, ‘trust in the Indian state’, and ‘support for democracy’ statistically similar, but by world comparative terms, quite high. In striking contrast, in Sri Lanka, differences between the majority Sinhalese-Buddhists, and the minority Tamil-Hindus, on similar questions were sharp. Due to ethno-nationalist policies that privilege the Sinhalese-speaking Buddhist majority, religious minorities feel much more unequal and threatened, in Sri Lanka, than they do in India.

The most counterintuitive finding in the questions on religion formulated by Linz, Stepan and Yadav for the State of Democracy in South Asia Surveys (SDSA, 2008) was that in India, for both Hindus and Muslims, the greater the intensity of religious practice, the greater the intensity of support for democracy.39 Interestingly, the major survey analyst in Indonesia, Saiful Mujani, found somewhat similar results for Indonesia.40 Unfortunately, we have no comparable survey analysis in Senegal, but the similarity of the findings that high religious participation correlates with high support of democracy in India and Indonesia suggests that ‘respect all, support all’ model creates space for democratic and religious actors consistent with the ‘twin toleration’ perspective.


In Sri Lanka, Buddhism is the majority religion and official state religion, but the opposite relationship holds – for Sri Lankan Buddhists, the greater the intensity of religious practice, the less the support for democracy. Indeed, in India’s federal system, with its numerous official languages and its ‘respect all, support all’ religious policies, we find majorities with more inclusive attitudes, and minorities with more trust in state institutions, than we find in Sri Lanka with its unitary state, its privileging of Sinhalese language, and its official state religion of Buddhism.


Let me conclude with a few words on Indonesia. A key aspect of the Indonesian version of state religion relations is the doctrine of Pancasila. The five principals of Pancasila were included in the Preamble to the Indonesian Constitution of 1945. The principles are: (i) belief in God, (ii) a just and civilized humanitarianism, (iii) national unity, (iv) Indonesian democracy through consultation and consensus, and (v) social justice. Despite numerous challenges from Islamists who wanted an Islamic state, and the military who often challenged these five principals, they have remained there ever since. The five principals were developed by Sukarno, the nationalist leader of the independence movement and the first President of independent Indonesia, with the active collaboration of some Islamic leaders, including the father of the three time elected president of Indonesia’s largest Muslim organization, NU, Abdurrahman Wahid.41

Indonesia’s Pancasila has flaws, such as an ambivalence about using state force to protect against Islamist violations of human rights in some parts of Indonesia, and the fact that for most of the Suharto period , the military defined and orchestrated Pancasila, flaws which I will discuss in a later version of this essay.42 But, Pancasila has some political virtues for a society such as Indonesia’s. It persistently helped ward off demands for an Islamic state religion that would, given Indonesian society, have exacerbated inter-religious relations in its highly diverse society.43


Also, in an intensely religious society, Pancasila officially recognizes and gives some support to five religions in addition to Islam – Buddhism, Hinduism, Catholicism, Protestantism, and with the advent of democracy, Confucianism. The state officially recognizes that these last five non-Islamic religions have at least 27 million adherents. The Indonesian state does not, however, officially recognize the Jewish religion, the numerous animists within Indonesia’s estimated 400 ethnic and language groups, or the variant of Islam called Ahmadiyah and, indeed recently, the state was very slow to dispatch police to protect Ahmadiyah from mob attacks.

In the ‘twin tolerations’ I argued that all religions are ‘multi-vocal’. What this means for Islam is that officially implemented systems of sharia law would necessarily have a strong element of ‘state sharia’ because one side of the multi-vocality would be state privileged and have the coercive powers of the state behind it. Given the deep differences between ‘traditionalist’ Muslims in Nahdatul Ulama (NU), and ‘modernist’ Muslims in Muhammadiyah – and their political and cultural sensitivity to the existence and rights of Hindus, Buddhists, Christians and non-practicing Muslims – leaders of both these massive organizations, with members numbering over 25 million each, are now opposed to an Islamic state which they know would lead to the non-consensual imposition of a single group’s vision of ‘state sharia’.


Amien Rais, a former President of Muhammadiyah, Speaker of the Consultative Assembly, and presidential candidate, advances variants of the following argument: ‘First of all, the Qur´an does not say anything about the formation of an Islamic state, or about the necessity and obligations on the part of Muslims, to establish a sharia or Islamic state. Second, the Qur’an is not a book of law but a source of law. If the Qur’an is considered a book of law, Muslims will become the most wretched people in the world… We should not establish Islamic justice as it will create controversy and conflict. Indonesia should be built on the principles of Pancasila to be a modern state, and to allow every citizen of Indonesia to pursue his or her aspiration.’44


Abdurrahman Wahid of NU in particular, would reject any Rawlsian idea of ‘keeping religion off the agenda’. Precisely because he knows that in multi-vocal Indonesia there are religious advocates of an exclusionary approach to religion and politics, he articulates alternative public discourses. He is a constant participant in public arguments making the case why Indonesia, given its great social and religious diversity, which he sees as an empirical fact, should make the normative political choice for a pluralist polity – a tolerant inclusive Islam in a tolerant inclusive Indonesia.45 He also works to create religious schools and organizations that advance these religious and democratic goals not only inside religious spaces, but in civil society, and in political society.46 He could not have carried out these agendas in a context of Turkish, or even of French, secularism.


* This paper draws on the draft prepared for the 2-3 October 2008 meeting of the SSRC Working Group on ‘Rethinking Secularism’.



1. The idea of the ‘twin tolerations’ is developed in more detail in my chapter ‘The World’s Religious Systems and Democracy: Crafting the "Twin Tolerations’’,’ in my Arguing Comparative Politics, Oxford University Press, Oxford and New York, 2001, pp. 213-255.

2. See Robert Dahl, Polyarchy: Participation and Opposition, Yale University Press, New Haven, 1971.

3. See Arend Lijphart, Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries, Yale University Press, New Haven, 1999.

4. See Juan J. Linz and Alfred Stepan, Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-Communist Europe, Johns Hopkins University Press, Baltimore and London, 1996.

5. See S.N. Eisenstadt, ‘Multiple Modernities’, Daedalus, Winter 2000, pp. 1-30. The reference to the USA as the first case of a multiple modernity in the West is on p. 13. Also see his ‘The Reconstruction of Religious Arenas in the Framework of "Multiple Modernities",’ Millennium: Journal of International Studies 29(3), 2000, pp. 591-611.

6. Sudipta Kaviraj, ‘Modernity and Politics in India’, Daedalus, Winter 2000, pp. 137-162.

7. See his ‘An Outline of a Revisionist Theory of Modernity’, European Journal of Sociology 46(3), 2005, pp. 497-526.

8. Op.cit., f.n. 1.

9. See his The Rise of Christian Democracy in Europe, Cornell University Press, Ithaca, 1996.

10. All data are collected from the ‘Religion and State Dataset’ gathered by Jonathan Fox, Department of Political Studies, Bar Ilan University. The data are reported in Jonathan Fox and Shmuel Sandler, ‘Separation of Religion and State in the Twenty-First Century’, Comparative Politics 37(3), April 2005, pp. 317-355. For a more detailed analysis of this data see Jonathan Fox, A World Survey of Religion and State, Cambridge University Press, Cambridge and New York, 2008. Please contact Jonathan Fox directly for access to the dataset. The countries comprising the non-US western democracies are the following (itals denotes countries with official established religions): Andorra, Australia, Austria, Belgium, Canada, Cyprus, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Liechtenstein, Luxembourg, Malta, The Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, and the United Kingdom. Sweden began a process of disestablishment in 2000.

11. For the history of the establishment of churches in America and for debates over the First Amendment, see A.J. Reichley, Religion in American Public Life, The Brookings Institution, Washington, D.C., 1985, pp. 53-167.

12. See for example, Joel S. Fetzer and J. Christopher Soper, Muslims and the State in Britain, France, and Germany, Cambridge University Press, New York, 2005.

13. Ahmet T. Kuru, ‘Secularism, State Policies, and Muslims in Europe: Analyzing French Exceptionalism’, Comparative Politics 41(1), 2008.

14. Some religious groups in France, such as Jehovah’s Witnesses and Scientologists, have experienced problems on this issue.

15. Ahmet T. Kuru, ‘Passive and Assertive Secularism: Historical Conditions, Ideological Struggles, and State Policies Towards Religion’, World Politics 59(4), 2007, pp. 568-594. See also Ahmet T. Kuru, Secularism and State Policies Toward Religion: The United States, France, and Turkey, Cambridge University Press, New York, 2009.

16. See Fox, A World Survey of Religion and the State, Table 5.4 for France, 8.4 for Turkey and 9.2 for Senegal.

17. Ibid., table 8.4.

18. See Bernard Lewis, The Emergence of Modern Turkey, 2nd edn. Oxford University Press, Oxford, 1978, p. 412 and Dankwart A. Rustow, ‘ Transitions to Democracy: Turkey’s Experience in Historical and Comparative Perspective’, in Metin and Ahmet Evin (eds.), State, Democracy and the Military; Turkey in the 1980s, Walter de Gruyter, Berlin, 1988, p. 246.

19. For the ‘exclusionary’ and ‘authoritarian’ qualities of secularism in Turkey see Stepan, ‘Twin Tolerations’, pp. 242-246.

20. Since 1963, the Turkish Constitutional Court has closed down 24 political parties.

21. See my ‘Secular or Islamic Fundamentalism’, Today’s Zaman, (Istanbul) 24 March 2008. This syndicated article has been published in many European newspapers, such as the Guardian, 25 March 2008.

22. For a close analysis of the context of Turkish constitution making in 1981-2 see Murat Akan, The Politics of Secularization in Turkey and France: Beyond Orientalism and Occidentalism, PhD Dissertation, Department of Political Science, Columbia University, 2006, Chapter10. Also see Kemal H. Karpat, ‘Military Interventions: Army-Civilian Relations in Turkey Before and After 1980’, in Metin Heper and Ahmet Evin (eds.), State, Democracy and the Military: Turkey in the 1980s, Walter de Sruyten, Berlin and New York, 1988, pp. 239-248.

23. Quoted by Ahmet T. Kuru, ‘Reinterpretation of Secularism in Turkey: The Case of the Justice and Development Party’, in The Emergence of a New Turkey: Democracy and the A.K. Parti, edited by M. Hakan Yavuz, University of Utah Press, Salt Lake City, 2006, p. 144.

24. In the year 2000 Sweden began a process of disestablishment. Finland also has an established Orthodox Christian Church to service its small Orthodox population.

25. I will develop these remarks in later versions.

26. I have just received Jonathan Fox’s 2008 book which uses the same 2002 data but employs a different weighting system. Analytically, all the relationships are the same. In this section I will use his 2005 and 2006 articles, not the 2008 book.

27. Pippa Norris and Ronald Inglehart, Sacred and Secular: Religion and Politics Worldwide, Cambridge University Press, New York, 2004, Table 3.5, p. 72.

28. The data are from the 2001 census of Northern Ireland.

29. Norris and Inglehart, Sacred and Secular, p.72.

30. See John T.S. Madeley, ‘A Framework for the Comparative Analysis of Church-State Relations in Europe’, West European Politics 26(1). 2003, pp. 23-50, esp. p. 45.

31. See Mirjam Künkler, Democratization, Islamic Thought and Social Movements: Coalitional Success and Failure in Indonesia and Iran, Columbia University, PhD Dissertation, Political Science, 2008.

32. See the maps of mono-confessional and multi-confessional polities in Western Europe in Madely, op. cit.

33. See the 2007 Freedom House Survey, Journal of Democracy 19(2), April 2008, pp. 60-67. See also my article with Graeme Robertson, ‘ An "Arab" More than "Muslim" Electoral Gap’, Journal of Democracy, July 2003 and our response to the controversy this ensued ‘Arab, Not Muslim Exceptionalism’, Journal of Democracy, October 2004, pp. 141-146.

34. For this model in India see Rajeev Bhargava, ‘The Distinctiveness of Indian Secularism’, in T.N. Srinivasan (ed.), The Future of Secularism, Oxford University Press, Oxford and Delhi, 2006, pp. 20-53. For the moral and political theory behind India’s secularism, see Bhargava’s ‘Political Secularism’, in John S. Dryzek, Bonnie Honig and Anne Phillips (eds.), The Oxford Handbook of Political Theory, Oxford University Press, Oxford and New York, 2006, pp. 636-655. Also see the volume Bhargava edited, Secularism and its Critics, Oxford University Press, Oxford and Delhi, 1998, especially the articles by Bhargava, Akeel Bilgrami, and Amartya Sen. Also study the wrong headed, completely U.S. normative vision of strict separation of religion and state, in Smith’s article in that volume. By this inappropriate measure India falls short of U.S. style secularism. Smith misses the point that this U.S. model was inappropriate for India and in fact, was never the intended model.

35. The five principals of Pancasila were included in the Preamble to the Indonesian Constitution of 1945. The five principles are: (i) belief in God, (ii) a just and civilized humanitarianism, (iii) national unity, (iv) Indonesian democracy through consultation and consensus, and (v) social justice. Despite numerous challenges they have remained there ever since. They were agreed upon by the nationalist leader of independence and the first President of independent Indonesia Sukarno with the active collaboration of some Islamic leaders, including the father of the three time elected president of Indonesia’s largest Muslim organization, NU, Abdurrahman Wahid. Two basic books for the history and evolution of Pancasila are Azyumardi Azra, Indonesia, Islam, and Democracy: Dynamics in a Global Context and Douglas E. Ramage, Politics in Indonesia: Democracy, Islam and the Ideology of Tolerance, Routledge, London and New York, 1995.

36. For numerous tables supporting this assertion for India, see chapters 2-4 in Alfred Stepan, Juan J. Linz and Yogendra Yadav, Democracies in Multinational Societies: India and Other Polities, The Johns Hopkins University Press, Baltimore and London, (forthcoming).

37. Alfred Stepan, ‘Rituals of Respect: Sufis and Secularists in Senegal’, in Thomas Banchoff and Robert Wurthnow (eds.), Religion and the Global Politics of Human Rights, Oxford University Press, Oxford and New York, forthcoming. Available on my website and on the SSRC sharepoint.

38. For a documented and convincing discussion of how and why French colonial authorities in Senegal came to the conclusion that the religiously hostile 1905 version of French laicité would be inappropriate as state policy in Senegal, and their effort to construct religiously friendly, partly pro-Islamic policies, see Chapter 4, ‘France as a "Muslim Power",’ in David Robinson, Paths of Accommodation: Muslim Societies and French Colonial Authorities in Senegal and Mauritaniaa, 1880-1920, Ohio University Press, Athens, 2000, pp. 75-96. Also see, Donald Cruise O’Brien, ‘Towards an "Islamic Policy" in French West Africa’, Journal of African History 8, 1967, pp. 303-316.

39. Linz and Stepan, and Yadav, at the invitation of Yadav, wrote some of the questions on religion and democracy for the State of Democracy in South Asia, Oxford University Press, New Delhi, 2008. The study included surveys and reports covering India, Pakistan, Bangladesh, Sri Lanka and Nepal. Many of these tables will be presented in later versions of this article.

40. See his important doctoral dissertation Religious Democrats: Democratic Culture and Muslim Political Participation in Post-Suharto Indonesia, Ohio State University, Political Science, 2003.

41. Two basic books for the history and evolution of Pancasila are Azyumardi Azra, Indonesia, Islam, and Democracy: Dynamics in a Global Context, Solstice Publishing, Jakarta and Douglas E. Ramage, Politics in Indonesia: Democracy, Islam and the Ideology of Tolerance, Routledge, London and New York, 1995. An important work on the development of democratic Islamic thought and practices in Indonesia is Mirjam Künkler’s previously cited dissertation, Democratization, Islamic Thought and Social Movements: Coalitional Success and Failure in Indonesia and Iran.

42. For a review of the question of violence in Indonesia since 1999 see Greg Fealy,’ In Fear of Radical Islam: A Critical Examination of the Evidence’, paper given at The Australian National University, 24 November 2005.

43. One of the biggest historical and political problems for the potentially democratic pancasila was that for much of the Suharto military regime Suharto appropriated the control of pancasila. Ramage is particularly interesting in analyzing how and why Wahid was able to recapture, and transform, the pancasila discourse into more democratic and more consensual directions.

44. Mohammad Amien Rais, Putra Nasantara: Son of the Indonesian Archipelago, Singapore Press Pte, Singapore, 2003, p. 11. Also see the interview with Amin Rais by Mirjam Kunkler and Alfred Stepan in the Journal of International Affairs 61(1), 2007, pp. 205-216.

45. For diversity as a ‘sociological fact’ and pluralism as a ‘political choice’, in general and in the speeches and actions of Wahid, see the chapter by Abdullahi Ahmed An-Na‘im, ‘Indonesia: Realities of Diversity and Prospects of Pluralism’ in his Islam and the Secular State: Negotiating the Future of Shari’ah, Harvard University Press, Cambridge, 2008. For an analysis of Wahid’s political discourse also see the chapter, ‘Abdurrahman Wahid: Scholar-President’, in John l. Esposito and John O. Voll, Makers of Contemporary Islam, Oxford University Press, Oxford, 2001, pp. 199-216 and the chapter, ‘Abdurrahman Wahid and Nahdlatul Ulama’, in Ramage, Politics in Indonesia, pp. 75-121.

46. Robert W. Hefner and Muhammad Zaman have recently edited an invaluable book that reviews madrasas in eight different countries, see their Schooling Islam: The Culture and Politics of Modern Muslim Education, Princeton University Press, Princeton, 2007. One of the most inclusive and tolerant systems described in the volume, and the one that now works most cooperatively with a democratic state, is in Indonesia. The chapter shows how NU and Muhammadiyah, and the negotiations required by Pancasila, have made substantial contributions to this outcome. See the chapter by Azyumardi Azra, Dina Afrianty, and Robert W. Hefner, ‘Pesantren and Madrassa: Muslim Schools and National Ideals in Indonesia’, pp. 172-198.