Concept Modeling by the Masses: Folksonomy Structure and Interoperability

The recent popularity of social software in the wake of the much hyped “Web2.0” has resulted in a flurry of activity around folksonomies, the emergent systems of classification that result from making public the individual users’ personal classifications in the form of simple free form “tags”. Several approaches have emerged in the analysis of these folksonomies including mathematical approaches for clustering and identifying affinities, social theories about cultural factors in tagging, and cognitive theories about their mental underpinnings. In this paper we argue that the most useful analysis is in terms of mental phenomena since naive classification is essentially a cognitive task. We then describe a method for extracting structural properties of free form user tags, based on the linguistic properties of the tags. This reveals some deep insights in the conceptual modeling behavior of naive users. Finally we explore the usefulness of the latent structural properties of free form “tag clouds” for interoperability between folksonomies from different services.