I have been thinking a lot lately about the underlying dynamics of big data, and how most discussions around online privacy and surveillance are functions of absent or simplistic taxonomies for social data. This is a lecture I gave last week to a 100-level convergent media class, where I tried to synthesize these ideas in a more or less coherent package illustrating the dynamics. I start with a list of Pompeii graffiti, which look surprisingly similar to tweets, illustrating two features of social data: we generate a lot of it on the go, and it tends to outlast the context for which it was generated. I then move through artifacts such as Raytheon’s RIOT software, the numbers on big data and and the way they bring forward the notion of flow management, the FinFisher spy software, and the Camover anti-surveillance game. I end with Bruce Schneier’s proposal for a working social data taxonomy.