"Separation of church and state" is not a switch but two dials that turn independently: how far the state intervenes in and supports religion, and how far religion shapes state decisions. The common combinations include strict separation, cooperation, an established church, state administration of religion, and official recognition of several faiths.
The deeper fork is whom the separation protects. One version holds that it shields religion from the state; the other that it shields the state and its citizens from religious domination — churches long controlled schooling and the registers of birth, marriage and death. Both call themselves "separation," and both grow into entirely different institutions.
| Country | Arrangement | Cost |
|---|---|---|
| France | The 1905 Law on the Separation of the Churches and the State: the Republic neither recognizes nor subsidizes any religion; Article 1 of the 1958 Constitution declares the Republic "secular" | Alsace-Moselle belonged to Germany at the time and still operates under the 1801 Concordat, where clergy of four recognized faiths are paid by the state |
| Germany | Cooperative model: Article 140 of the Basic Law incorporates provisions of the Weimar Constitution; religious bodies holding public-corporation status may levy a church tax (9% of income tax in most Länder, 8% in Bavaria and Baden-Württemberg), collected by the tax authorities | The state becomes the financial infrastructure of religious organizations |
| Indonesia | The first principle of Pancasila is belief in one supreme God; the Ministry of Religious Affairs recognizes six religions | Not separation but plural recognition — faiths outside the recognized list sit at the margins |
The strongest case for strict separation: public money flowing to religion makes taxpayers fund doctrines they do not hold; and for the state to decide "which of these counts as a real religion" is to make a theological judgment, the one thing it is least equipped to do.
The strongest case for cooperation: zero contact is a fiction. Armies and prisons need chaplains; religious hospitals and schools already deliver public services. Since contact is unavoidable, open cooperation under rules beats an unaccountable understanding.
The trade-off: strict separation buys formal neutrality and sells visibility into how religious bodies actually operate; cooperation buys oversight and sells the purity of staying out of their internal affairs.
Treating the legal degree of separation as a measure of how secular a society is. They are independent variables: some countries keep an established church while everyday religious practice is thin; others forbid establishment outright yet rank near the top of the developed world in religious participation. The law governs what the state does, not what people believe.
The word "neutrality" conceals two incompatible solutions. The accommodating kind: treat all faiths alike and let religious expression be visible in public space. The excluding kind: keep religious markers out of public space altogether, because once expression is allowed, visibility tracks numbers and the majority wins by default.
For the majority the two answers converge; for minorities they diverge. The accommodating rule lets a headscarf into the classroom, the excluding rule does not — but it also takes down the crucifix. Which better protects minorities depends on whether you think the threat comes mainly from the state or from their own community.
France's laïcité sits near the excluding end; Turkey's laiklik goes further still — the state also runs a department that administers religious affairs. That is control, not withdrawal.
The strongest case for exclusion: religious markers inside a public institution signal to minorities and non-believers that they are guests in someone else's house. And since family and community pressure is real, a uniform rule hands anyone who wants out an external reason they can point to.
The strongest case for accommodation: making "no religion" the default is not a zero point but a substantive position. The burden of a ban falls very unevenly — a garment understood as an obligation cannot simply be put back on at home, while majority symbols have often been reclassified as "tradition" and go untouched.
The trade-off: exclusion buys formal equality and sells an even distribution of the burden; accommodation buys freedom of expression and sells the appearance of institutional neutrality.
Taking laïcité as the general definition of secularism. It is the product of a specific French history — a long contest between the Republic and the Catholic Church over schools and civil registration. Mistake one country's historical solution for the concept itself, and you will read "solved a different problem" as "insufficiently secular."
Religious nationalism is not the same as devout nationalism. Its core operation is converting religious affiliation into the boundary marker of national membership.
It does two things at once: it gives "who are we" a line that is clear, inheritable and hard to fake; and it translates divisible conflicts of interest — land, budgets, seats — into indivisible conflicts of value. Interests can be split down the middle; sanctity cannot.
It tends to appear during modernization rather than before it: urbanization scatters place-based communities, and religious identity supplies a portable belonging. Berger predicted in 1968 that by the 21st century believers would survive only in small sects, then revised that judgment in The Desecularization of the World (1999); Stark argued the same year that secularization theory belongs in "the graveyard of failed theories."
The strongest case that religion should ground nationhood: a political community needs a shared source of meaning, and purely procedural citizenship struggles to sustain high-cost cooperation such as taxation, military service and sacrifice across generations; religious tradition is a resource already present locally, and easier to internalize than an imported abstraction.
The strongest case against: tying membership to belief permanently places local minorities who have lived there for generations one rung down, with no change of conduct that could change their standing. The structural problem is the correction channel: the authority behind a religious criterion sits outside secular procedure and cannot be adjusted by an election.
The trade-off: grounding a community in religion buys mobilizing power and sells internal pluralism and negotiability; procedural citizenship buys inclusion and amendability and sells some of the emotional bond.
Assuming it grows out of rising devotion. The opposite pairing is just as common: mobilization along religious identity works quite as well among populations whose religious practice is declining — by then religion has shifted from a demanding set of obligations to a low-cost badge of identity. Identity and practice are two separate variables.
The first three sections asked how the state treats religion; this one asks what religious bodies actually carry. Three functions are near-universal: welfare provision, social capital and organizing capacity, and moral language in public debate.
The third is the hardest, because it turns on the acceptability of reasons. One position holds that on fundamental political questions citizens should offer reasons others could in principle accept; critics answer that this asks a large share of citizens to present only a translated version of their reasons, which makes actual motives less visible rather than more.
A further interface is easy to overlook: personal law, where marriage and inheritance follow the law of one's religious community. It shifts the unit of legal application from "an individual on the territory" to "a member of a community."
| Country | Practice | Cost |
|---|---|---|
| Germany | Churches and their welfare arms run a large share of nurseries, hospitals and care homes | Standing tension between a religious employer's staffing autonomy and equal-treatment law, drawn case by case through labour litigation |
| France | The reverse path: under the 1905 law, public finance does not as a rule subsidize religious bodies | Funding shifts to private and foreign channels, and transparency falls |
| India | Religious communities follow their own personal laws on marriage and inheritance; Article 44 lists a uniform civil code as a Directive Principle (not judicially enforceable) | Community autonomy is respected, but the rights of its weakest members rest on the community's internal power structure, with little option to exit |
The strongest case for accommodation: religious bodies already supply services the state cannot reach; severing the arrangement does not remove the need, it shifts the cost onto the weakest. As for public debate, excluding faith-driven reasons does not remove them from decisions — it only lets them enter in disguise, and a reason has to be stated openly before it can be rebutted openly.
The strongest case for limits: public money must carry public obligations — non-discrimination, curriculum standards, financial accountability. And personal law, in the name of respecting the community, hands the adjudication of individual rights to the community's existing authorities, when the point of rights protection is precisely to protect those with no voice inside it.
The trade-off: accommodation buys coverage and social trust and sells a uniform standard of rights; limitation buys consistency of rights and sells the service network. Most countries split by subject: fundamental rights uniform nationally, the manner of delivery left to communities.
Equating "religion taking part in public life" with theocracy. Nearly every live dispute sits in the borderland — whether public funds may reach faith schools, whether religious institutions may hire by belief — and that is not the same order of question as religious authorities holding legislative power directly. Collapse the two and you lose the vocabulary for the middle ground, which is where the choices actually are.