| Abstract |
Network motifs are fundamental to many structural social phenomena, such as structural holes, triadic closure or coalition formation. Furthermore, motif occurrence varies across social networks of different genre, suggesting connection to diverse function and origin of these systems. However, social network theory lacks a generic model capable of explaining observed variability of motif occurrences. In this paper we focus on two basic social forces: homophily (similarity) and hierarchy (popularity) and we demonstrate how their interplay influences the statistics of motifs observed in social networks. Through the embedding of social networks in hyperbolic space we estimate the effects of similarity and popularity in various systems. This allows us to classify social networks based on identified balance between homophilic and hierarchical tendencies. |