The author insists on the need to take the social-structural context into account to evaluate the existing types of definitions of religion. Proposing a definition for the Western hemisphere, he then checks his definition against sociological studies referring to 'new types of religiosity' and concludes that sociologists should differentiate between meaning systems - of which religion is a sub-division - and spiritualities. In a second step, he discusses two sociological theories about religion and insists that sociologists should be more careful in using such theories by taking into consideration the type of religion the theory is concerned with. Finally, he suggests that rational choice theory and secularization theory might well be integrated and applied in a European context if we move to the level of the competing existing meaning systems.