Secularisation

Capitalism, cities, factories and citizens arrived together (3.2.2–3.2.5). So did something harder to measure: a change in what religion governed, in what people believed, and in whether the world was experienced as enchanted at all.

This lesson is also an exercise in something Part 1 kept demanding: taking a famous thesis apart into its separate claims before asking whether it is true. Do that here, and a debate that has run for a century becomes tractable.

Two illnesses, four hundred years apart.

A village, 1600. A child develops a fever. What is done? The mother makes a preparation from herbs whose properties are known and are not sharply distinguished from their virtues. Prayers are said; a vow may be made at a shrine. A local person with a reputation for such things is consulted, and may use words as well as substances. If a neighbour has recently quarrelled with the family, someone will wonder aloud. The priest is told. If the child dies, there is a framework in which the death is intelligible: a trial, a judgement, a mystery, a wickedness, the will of God.

There is no domain here labelled "medical," separate from a domain labelled "religious." There is one world, in which illness, morality, weather, harvest and misfortune are continuous.

A city, now. A child develops a fever. The parents check a thermometer, consult a website, take the child to a clinic. A doctor makes a differential diagnosis, orders a blood test, prescribes an antibiotic if bacterial. The family may also pray, may make a vow, may consult an astrologer, may light a lamp. If the child dies, there will be a cause of death recorded, and possibly an inquiry into whether the treatment was correct — and there will be a funeral, and someone will say it was God's will.

Notice what has changed and what has not.

A separate domain has appeared. Medicine now has its own institutions, training, authority, jurisdiction and standard of evidence, and its claims are not settled by religious authority.

And religion has not disappeared. In many families it is fully present — in the same room, at the same bedside, doing different work.

The secularisation thesis — that modernity produces religious decline — was for most of the twentieth century simply assumed by sociologists. It is the discipline's most confidently held prediction, and it comes from the founders themselves.

Then two things happened.

The evidence went in several directions at once. Church attendance in Western Europe collapsed. The United States, the most modern society on earth by most measures, remained highly religious for decades. Pentecostal Christianity grew explosively across Latin America, Africa and Asia. Islamic revival movements reshaped politics across a wide region. Religious nationalism became a major political force in India, Israel, Turkey and elsewhere. Global religious populations grew, driven by fertility differentials.

And the thesis turned out to be several claims wearing one name. Which is why it could be simultaneously obviously true and obviously false, depending on which claim the speaker meant.

Peter Berger, who wrote one of the most influential statements of the thesis in The Sacred Canopy (1967), later published a public reversal, writing that the assumption that modernisation leads to secularisation was mistaken and that the world today "is as furiously religious as it ever was." He was more candid than most, and 1.4.6 asked you to value exactly that.

So: pull the thesis apart, then look.

The idea, in plain words

What the founders said

Worth stating carefully, because they are routinely misquoted.

Comte (3.3.1) predicted straightforwardly that the theological stage would be superseded by the positive — and then tried to found a Religion of Humanity, complete with a calendar of secular saints, which tells you something about how confident he was that societies could do without one.

Marx (4.1): religion as "the sigh of the oppressed creature… the opium of the people." The full passage is more sympathetic than the slogan: religion is a response to real distress, and Marx expected it to fade when the conditions producing the distress were removed. A prediction conditional on revolution, not on modernity as such.

Durkheim (4.2) is the most misread. He did not predict religion's disappearance. His argument in Elementary Forms is that religion's real object is society, and that societies therefore continue to generate sacred objects, collective rituals and moral communities — which is why he thought modern societies might develop new sacred forms, and why the concept of civil religion (9.3, 9.5) descends from him. What he expected to decline was the authority of particular traditional religions, not the sacred as such.

Weber (4.3) is the source of the deepest version, and it is not a prediction of decline in belief. His disenchantment of the world (Entzauberung) means the retreat of magic — the belief that the world can be influenced by ritual manipulation — and the spread of the conviction that all things can in principle be mastered by calculation. Crucially, Weber thought this produced not the end of religion but a loss of shared meaning in the public sphere: science can tell you what is, and cannot tell you how to live, so ultimate values retreat into private commitment and into what he called the war of the gods between incommensurable value spheres.

That is a subtler and more defensible claim than "religion will fade," and it has aged much better.

What the evidence shows

Take the three processes separately, and the picture becomes clear rather than contradictory.

Differentiation: strongly supported. This is the most robust element of the thesis. Across a very wide range of societies, medicine, law, science, education and welfare have acquired autonomous institutional authority, and religious bodies no longer determine their content. Even where religion is politically powerful, it typically operates upon these differentiated spheres rather than constituting them. Note that this can happen without any decline in belief at all.

Decline: true in some places, not others, and not unidirectional.

Western Europe is the clearest case: church attendance, affiliation and belief have fallen substantially across most of the region over the twentieth century, and continued falling.

The United States was for decades the counter-example, sustaining high levels of belief, affiliation and attendance in a highly modern society. That picture has changed recently: the share of Americans reporting no religious affiliation — the "nones" — has risen very substantially since the 1990s, which suggests the American exception was less permanent than it appeared.

Elsewhere, growth has been the story: Pentecostal and charismatic Christianity in Latin America, sub-Saharan Africa and parts of Asia; Islamic revival across many societies; and the political mobilisation of religious identity in India and elsewhere. Post-Soviet societies saw religious resurgence after state suppression ended.

And globally, religious populations are growing in absolute terms, principally because religiosity is higher in societies and groups with higher fertility. This is a demographic fact rather than a story about conversion, and it means that even where individual-level religiosity declines, the world's religious population can grow (2.3.8, 11.3).

Privatisation: substantially disconfirmed. This is where Casanova's contribution bites. He documented the deprivatisation of religion in the 1980s: religious actors entering public life across very different contexts — liberation theology in Latin America, Solidarity in Poland, the Iranian revolution, the American religious right. Rather than retreating to private conscience, religion re-entered public debate, often on questions of rights and justice, and often in modern organisational forms.

The religious economies model. Rodney Stark, Roger Finke and colleagues proposed inverting the classical account using rational-choice reasoning. Religion, they argued, should be analysed as a market: demand is relatively constant, and what varies is supply. Where a state church holds a monopoly, it becomes complacent and under-serves; where there is pluralism and competition, religious "firms" innovate and mobilise, and participation rises. On this account, Europe is secular because of monopoly and regulation, and America religious because of competition.

How it has fared. The model made testable predictions, which is to its credit, and the central prediction — that pluralism raises participation — has been tested extensively with mixed and much-disputed results, with critics arguing that the correlation depends heavily on how pluralism is measured and on which cases are included. Treat it as a productive framework that has not been confirmed, and note that it did useful work in forcing the field to specify mechanisms.

The existential security thesis. Pippa Norris and Ronald Inglehart's Sacred and Secular (2004) offers the best-evidenced synthesis. Their argument: religiosity declines where people experience existential security — reliable food, health, physical safety, income in old age — because religion's demand is driven substantially by vulnerability and the need for reassurance in the face of unpredictable threat.

This explains a great deal at once. Western Europe's welfare states produced high security and falling religiosity. The United States, with high inequality and weaker social protection, retained higher religiosity than its wealth alone would predict. Poorer and more insecure societies remain highly religious. And — their striking demographic point — because secure societies have low fertility and insecure ones high fertility, the world contains a rising proportion of people raised in religious societies even as the richest societies secularise.

Why this is a good example for a course on method. It takes a vague thesis, replaces it with a specified mechanism (security, not modernity as such), derives predictions that discriminate between cases, and accounts for the outliers that defeated the original. That is what 1.5.3 asked for. It is not the last word — critics question the measurement of security and the direction of causation — and it is a considerable improvement on what it replaced.

Go deeper

The categories themselves have a history

The deepest criticism, and it belongs in this Part.

Talal Asad argues that "religion" and "the secular" are not natural kinds waiting to be measured but categories produced in a specific European history — the wars of religion, the settlement that confined religion to belief and conscience, and the construction of a secular state that defines what counts as religion and what may enter public life.

The consequence: applying the category "religion" to societies that never separated a religious domain — where practice, kinship, obligation, medicine and cosmology are continuous — can distort before measurement begins. Asking whether such societies are becoming "more secular" may be asking a question whose terms they do not contain.

This is the same problem 2.1.10 identified with Murdock's list and 1.4.3 with disciplinary grids: the categories are not neutral, and asking where they came from is part of the analysis.

India: a different secularism

Essential for this course, and a case where the European model does not transfer.

Indian secularism is not separation. The Constitution does not erect a wall between religion and state on the American model. The state funds religious educational institutions, administers major temples, legislates on religious practice, maintains distinct personal laws for different communities, and intervenes to reform practices within religions — most notably in the abolition of untouchability and in temple-entry legislation.

Rajeev Bhargava's influential account describes this as principled distance: the state maintains a non-sectarian relation to all religions but may engage or disengage with each differently, depending on what equality and freedom require in that case. It is a distinct model with its own logic, not a defective version of the Western one.

T. N. Madan (6.2.3) made the sharpest critique from within Indian sociology: that secularism as conceived in the West is a product of Christianity's particular history, that in South Asia religion is constitutive of the moral and social order for the great majority, and that a secularism which requires religion's privatisation is therefore socially unrealistic and may itself provoke fundamentalism. Ashis Nandy argued a related case, distinguishing lived religious tolerance from an ideology of secularism he regarded as alien and brittle.

Both are contested — critics reply that this underestimates indigenous traditions of tolerance and pluralism, and that it concedes too much to majoritarian claims. The debate is live and is one of the most substantial in Indian sociology (9.3).

And note the empirical point that anchors it: religious identity in India has become more politically salient over recent decades, not less, which is a straightforward counter-example to the privatisation thesis and to any assumption that modernisation reduces religion's public role.

What actually declined

To close constructively, here is a defensible summary of what the evidence supports.

Religion's institutional jurisdiction over other spheres declined nearly everywhere — differentiation is real and robust.

Magic and ritual manipulation of natural events declined substantially where scientific and technical means became available — which is Weber's disenchantment, and it is the best-supported version of the classical thesis.

Individual belief and practice declined sharply in specific places — Western Europe above all — with existential security the best available explanatory mechanism, and did not decline globally.

Religion's public and political role did not decline, and in many places increased.

And the world's religious population is growing for demographic reasons.

Anyone who states "secularisation" without specifying which of these they mean is not making a testable claim.

"Secularisation means religion is dying." It names at least three distinct processes — differentiation, decline, privatisation — which move independently. Specify which.

"Durkheim predicted religion would disappear." He argued religion's object is society, and that societies keep generating sacred forms. Civil religion descends from him.

"Weber's disenchantment means people stopped believing." It means the retreat of magic and the loss of shared meaning in the public sphere — a subtler claim that has aged well.

"Europe shows the future." European decline is real and is not the global pattern. Treating one region as the norm is exactly the error Part 6 identifies.

"America was the exception." It was, and the rise of the religiously unaffiliated since the 1990s has changed the picture. Exceptions can expire.

"Religion is retreating into private life." Casanova documented deprivatisation across very different contexts, and India's trajectory is a clear counter-example.

"Indian secularism is a failed version of the Western model." It is a distinct model — principled distance — with its own logic, and it is the subject of a serious internal debate (Madan, Nandy, Bhargava).

Why it matters

It is the discipline's best worked example of a thesis being disaggregated. A century-long argument became tractable the moment somebody separated three claims that had been travelling under one name. That operation — pull the thesis apart, then look — is the most transferable skill in this lesson.

It is a case of a major figure changing his mind in public. Berger's reversal is worth more as an example of intellectual conduct than most confirmations are worth as findings.

It corrects a persistent Eurocentrism in the discipline's self-understanding. Sociology grew up in the one region where religious decline was most dramatic and treated that experience as the universal trajectory of modernity. Norris and Inglehart's mechanism, Casanova's deprivatisation and the Indian debate together show why that generalisation failed — which is 6.7's argument, arrived at through evidence.

And it prepares Part 9. The sociology of religion (9.3) is one of the discipline's most active fields precisely because the classical prediction failed in interesting ways, and failed predictions are where research programmes come from.

Secularisation names three distinct processes (Casanova): differentiation (religion loses jurisdiction over other spheres), decline (falling belief and practice), and privatisation (retreat from public life).

The founders: Comte predicted supersession; Marx expected religion to fade with the distress producing it; Durkheim did not predict disappearance but the continued generation of sacred forms; Weber's disenchantment means the retreat of magic and the loss of shared public meaning, not the end of belief.

The evidence: differentiation is strongly supported; decline is sharp in Western Europe, was long absent in the United States (where the unaffiliated have risen substantially since the 1990s), and is absent or reversed across much of the world; privatisation is substantially disconfirmed — Casanova's deprivatisation.

Two revisions: the religious economies model (supply and competition rather than demand) made testable predictions with mixed and disputed results; Norris and Inglehart's existential security thesis explains the pattern better — religiosity tracks vulnerability, and because insecure societies have higher fertility, the world's religious population grows even as rich societies secularise.

Asad: "religion" and "the secular" are historically produced categories, which distorts measurement in societies that never separated the domains.

India: secularism as principled distance (Bhargava) rather than separation; Madan and Nandy's critiques; and rising rather than falling political salience of religious identity.

Berger, the thesis's most influential advocate, publicly abandoned it.

Secularisation — the umbrella term for religion's changing place in modern societies; better analysed as three separate processes.

Differentiation — the separation of institutional spheres (medicine, law, science, education) from religious jurisdiction.

Privatisation / deprivatisation — religion's retreat to the private sphere, and its documented re-entry into public life (Casanova).

Disenchantment (Entzauberung) — Weber's retreat of magic and of shared public meaning under rationalisation.

Civil religion — the sacred symbols, rituals and narratives of a nation or state, descending from Durkheim.

Religious economies model — the analysis of religion as a market in which pluralism and competition raise participation (Stark, Finke).

Existential security thesis — Norris and Inglehart's argument that religiosity tracks vulnerability, with demographic consequences.

Believing without belonging — Davie's description of retained belief alongside collapsed institutional participation.

Principled distance — Bhargava's account of Indian secularism as differential engagement rather than strict separation.

Communalism — in South Asian usage, the political mobilisation of religious identity against other communities (9.3).

One — disaggregate a thesis. Take any broad claim you hold about social change ("families are breaking down," "people are more polarised," "communities are weaker"). Break it into three separately testable claims. Then ask which the evidence you have actually bears on.

Two — two generations. Ask someone forty years older about religious practice in their childhood household, and compare with yours. Note separately: belief, practice, affiliation, and the jurisdiction religion had over decisions.

Three — find the differentiation. Identify one domain in your society where religious authority once governed and no longer does — or the reverse. Date the change if you can.

Four — test existential security. Compare two groups you know that differ in economic security, and consider whether their religiosity differs in the direction Norris and Inglehart predict. Note honestly what else differs between them (1.5.3).