The line of reasoning to be examined first is found in its most pronounced form in the writings of social scientists who label themselves ethnomethodologists, but may be encountered elsewhere as well. It should be stressed at the outset that ethnomethodology is only a loosely-defined approach; its adherents vary greatly among themselves, inter alia in the extent to which a constructivist position is adopted. Here, I examine the most extreme version to be encountered, not only because this is the one that most clearly represents constructivism, but also because it is the one that best safeguards the distinctiveness of ethnomethodology. Weaker versions will be found to overlap with related positions such as symbolic interactionism. But it should be stressed that the position examined is probably not shared by the majority of ethnomethodologists. Its merits lie in the clarity of the position, which is what interests us here, whereas a detailed empirical mapping of the current convictions of social scientists (an opinion poll of the discipline’s practitioners, as it were) is not our concern. Those who object to identifying ethnomethodology with this position may simply substitute the phrase ‘the radical constructivist wing of ethnomethodology’ wherever I write ‘ethnomethodology’.