This book proposes a fresh approach to sociological analysis and, in particular, to the analysis of scientific culture. It moves away from previous studies, which have tended to focus on scientists' actions and beliefs to show that analysis of scientific discourse can be productive and revealing. The book demonstrates that scientists produce varying accounts of their actions and beliefs in different social situations. Rather than attempting to extract one coherent interpretation from these diverse accounts, the study identifies two basic scientific repertoires and shows how scientists use them to create their discourse. This provides a point of departure for more complex analytical topics. Discourse analysis is applied to show how different degrees of 'consensus' can be ascribed to the same group of scientists at a given moment in time through the application of standard interpretive techniques. Finally, discourse analysis is used to explore scientists' humour, a neglected topic that is shown to provide important insights into the normally hidden interpretive regularities which underlie the cultural diversity of science.