In 1997 Emirebayer published a manifesto for relational sociology. Social network analysis (SNA) figured prominently as a methodology appropriate to his program. This paper argues that Emirbayer has overlooked a major tradition of social research in SNA that examines networks from a bottom-up perspective using 'ego-centric' or egonet methodologies. This paper urges sociologists and social researchers to look at the benefits of egonet research methodology. It argues that egonet methods remain sensitive to the qualitative dimensions of social actors' involvements with their immediate social context, a sensitivity that is lost in other SNA approaches. The paper describes qualitative dimensions of the egonet methodology - 'name generator' questions that generate a list of 'alters' and follow-up questions built on name generator responses. I identify a crucial qualitative issues associated with this methodology as that of finding accurate and meaningful, grounded but theoretical categories for describing relationships. I then describe three studies that derive usable grounded theory concepts that deal with this crucial aspect of social network research. In conclusion I return to the Emirbayer's manifesto for relational sociology and suggest some of the ways that egonet research might also fit into its programme.