The secularization thesis held that modernity necessarily brings religious decline, that as societies industrialise, urbanise, and become better educated and more scientifically informed, religion loses its social significance, its institutional power and eventually its plausibility. It was close to an assumption of the social sciences for most of the twentieth century, and its ancestry runs through Comte, Marx, Durkheim, Weber's "disenchantment", and Freud.
It was substantially revised from the 1980s onward, and the revision is one of the more instructive episodes in the history of the social sciences, because the evidence against the strong thesis was neither subtle nor contested. The United States was highly modern and highly religious. Religion was expanding rapidly across the global South, with Pentecostalism the fastest-growing religious movement in the world. Religion had returned conspicuously to politics, in the Iranian revolution of 1979, in the American religious right, in Solidarity in Poland, in political Islam and political Hinduism. And global religious adherence was rising in absolute terms. Peter Berger, who had been among the thesis's most prominent proponents, publicly retracted it in the 1990s, describing the assumption that modernisation produces secularisation as mistaken and the world as "as furiously religious as ever".
What survives the revision is real but narrower. Institutional religious practice has declined sharply and durably in Western Europe, in Canada, Australia and New Zealand, and more recently in the United States, where the religiously unaffiliated have grown very substantially since the 1990s. Differentiation, the separation of religious institutions from the state, education, medicine, law and welfare, is a genuine and general feature of modernity, and it is analytically distinct from any decline in belief. And religion has in many societies become a matter of individual choice rather than of inherited communal obligation, which changes its character even where its numbers hold.
Current positions divide accordingly. Jose Casanova distinguishes three claims that the original thesis bundled together, differentiation, decline in practice, and privatisation, and argues that the first holds generally, the second regionally, and the third not at all, since religion has conspicuously re-entered the public sphere. Charles Taylor's A Secular Age (2007) reframes the question as concerning the conditions of belief rather than the number of believers, arguing that what has changed is that belief has become one option among others rather than an unquestioned background. Rational-choice theorists argue that religious vitality tracks market conditions, state-supported monopolies producing complacency and pluralism producing vigour, an account that fits the American and European contrast and has been criticised on other grounds. Grace Davie's "believing without belonging", and its inverse, describe the European case specifically.
The general lesson usually drawn is that Europe is the exception rather than the template, and that a theory built on European experience was over-generalised into a law.