This thesis investigates the use of rules in the analysis of language mastery and human action, which are both viewed as social phenomena. The investigation is conducted through an examination of two analyses of the use of language in everyday social life and documents how each formulates a different understanding of rule-following in explaining linguistic and social action. The analyses in question are ‘Speech Act Theory' and 'Ethnomethodology'. The principal idea of speech act theory is that social action is rule-governed, and the theory attempts to explain the possibility of meaningful social interaction on that basis. The rigidities imposed by the notion of rule-governance frustrate that aim. The thesis then turns to an examination of ethnomethodology and conversation analysis and contrasts the notion of rule-orientation developed by that perspective. From that examination it becomes clear that what is on offer is not just a greater flexibility in the use of rules, but a restructuring of the concept of analysis itself. It is argued that re-structuring amounts to a reflexive conception of analysis. Its meaning and implications are enlarged upon through a close scrutiny of the later philosophy of Wittgenstein, particularly his concern with the nature of rule-following in his ‘Philosophical Investigations'. The thesis argues that his concern with rules was motivated by his insight that their use as ‘explanations’ of action said as much about the formulater of the rule as the activities the rules were held to formulate. The thesis concludes by outlining the meaning of this analytic reflexivity for social scientific findings.