Temporalities in the spread of knowledge

Socio-epistemic networks are an analytical and methodological framework that has been developed to explore the interplay of the social, material and semantic dimensions of the structural changes in knowledge systems by using the concepts and tools of network science. This framework is suited to analyzing the development of scientific fields throughout the 20th century, which traditionally presents several challenges: the scales of change are often gradual; the definitions and boundaries of these fields are typically constructed retrospectively; and the perspectives on the structure of a field at any given time can vary significantly among key actors. Additionally, traditional scientometric techniques used for historical analysis have major limitations, as they often rely on presentist assumptions regarding the metadata format of available online publications repositories. In this methodological contribution, we aim to overcome these limitations by proposing a workflow that uses temporal co-citation clustering of publication corpora. This workflow is part of an iterative process involving cooperation between digital humanists and historical domain experts. The goal is to extract meaningful, temporally extended clusters of publications that represent evolving scientific fields. We utilize the Leiden community detection algorithm to delineate these clusters. For the time scales of evolution of each retrieved cluster, further analytical steps are conducted to examine the temporal shifts in centrality measures of specific institutions or actors in the field, thereby revealing their changing roles within the research field. The approach is exemplified through a case study on the evolution of astrophysics and related research fields, and the role that the Max Planck Society played in these dynamics.