In this paper, we show, using a panel of developed countries, that there is a long-run negative association between church attendance and total factor productivity (TFP) with predictive causality running from declining church attendance to increasing factor productivity. According to our preferred estimate, about 18% of the increase in TFP from 1950 to 1990 is caused by declining religiosity. In order to explain this phenomenon, we integrate into standard R&D-based growth theory a micro-foundation of individual cognitive style, which is either intuitive-believing or reflective-analytical. Under the assumption that R&D productivity is positively influenced by a reflectiveanalytical cognitive style, we find that secularization leads to an increasing labor share in R&D and gradually increasing productivity growth. We use these insights to reflect on trends in religiosity and R&D-based growth in the very long run, from Enlightenment to the present day.