Characterizing a collaboration by its communication structure

I present first results of my analysis of a collection of about 24,000 email messages from internal mailing lists of the ATLAS collaboration, at CERN, the particle physics laboratory, during the years 2010–2013. I represent the communication on these mailing lists as a network in which the members of the collaboration are connected if they reply to each other’s messages. Such a network allows me to characterize the collaboration from a bird’s eye view of its communication structure in epistemically relevant terms. I propose to interpret established measures such as the density of the network as indicators for the degree of “collaborativeness” of the collaboration and the presence of “communities” as a sign of cognitive division of labor. Similar methods have been used in philosophical and historical studies of collective knowledge generation but mostly at the level of information exchange, cooperation and competition between individual researchers or small groups. The present article aims to take initial steps towards a transfer of these methods and bring them to bear on the processes of collaboration inside a “collective author.”