Knowledge Communities and Socio-Cognitive Taxonomies

Abstract

Social network analysis (SNA) typically appraises social groups by relying either on interaction patterns or on affiliation similarity. The former case represents the bulk of SNA approaches and relates to the so-called one-mode networks, which are by design blind to actor attributes. The latter case relates to what is denoted as two-mode networks and corresponds to a less abundant literature which uses actor attributes, yet eventually tends to focus much more on actor rather than attribute groups. This chapter aims to show how approaches such as formal concept analysis (FCA) make it possible to appraise actors and attributes on an equal footing. In the particular case of knowledge communities, where actor attributes represent cognitive properties, we deal with joint social and cognitive taxonomies, or socio-cognitive taxonomies. We further demonstrate that FCA also addresses several of the key traditional challenges of community detection in SNA—namely, overlapping groups, hierarchy, and temporal evolution and stability.