To understand the micro-foundations of human be-havior, we need to move beyond log-files and also engage in analysis of dialog among individuals engaging OC that takes place in chat rooms, on pin walls, etc. Today, analyzing the evolution of dialog is a time consuming process. Usually, there is no predefined set of codes available as researchers often need to engage in inductive and axial coding to make sense of data. Learning-oriented algorithms are needed that would allow one to build a rich lexica and meta-data related to par-ticular dialog themes[33]e.g. the phenomenon of perspective taking and distancing in knowledge creating dialogs.