Emotions have increasingly become the object of systematic sociological inquiry. Due to their special nature, a number of methodological problems accrue to the task of making emotions accessible to analysis. This article presents the results of an inquiry into the ways in which emotions are expressed in narratives and conversation. Conversations involving positive and negative self-feelings are analyzed in regard to paralinguistic markers. These markers are presented and discussed, and the ways in which markers ofthis kind can be interpreted in regard to the analysis of interview material are also considered.Emotions have become the object of systematic sociological inquiry within recent decades (Kemper, 1990;Collins, 1975;Hochschild, 1983; and Scheff, 1990, among others). Systematic inquiry, however, gives rise to a number of methodological difficulties related to the special nature of the object of analysis. Emotions have somatic characteristics, their expression can take nonverbal forms and their structure of meaning be far from clear-cut. In social research the language of conversation, including that of the interview, remains one of the most important tools of social analysis, ä means whereby insight is gained into everyday life, äs well äs the social and cultural dimensions of our own and other societies. A broad spectrum of data is provided by the .verbal reconstruction of experiences, episodes and situations undertaken by respondents during interview. Whether reconstruction of this kind, undertaken in interaction with an Interviewer, can also yield insight into the emotional dimensions of social life is less clear.