Scientific discourse is vital to make informed decisions about pressing societal issues. The recent COVID19 pandemic has driven an increased public interest in scientific findings, demonstrated by a proliferation of science communication as part of wider societal discourse in social and online news media. Recent research has argued that such trends lead to a disruption of the traditional scienitific knowledge order, most notably the phases, context, roles and hierarchy of knowledge production. Risks emerge from the over-simplification, generalisation or political instrumentalisation of scientific insights which can have adverse effects on public perceptions, decision-making and polarisation of society. This is even more problematic given that false and controversial information spreads more rapidly than the truth. This project aims at advancing the understanding of the changing role of science in the emerging knowledge order of the digital society. Through an interdisciplinary consortium combining expertise from science communication, cognitive and social psychology, and computer science, we will address three specific goals: 1) Understanding the erosion of the traditional knowledge order in dynamically evolving crisis situations, specifically, the dissolution of phases and contexts of production, examination and dissemination of scientific knowledge, 2) Understanding the perception of roles, sources and hierarchies in scientific online discourse and their impact on trust-worthiness assessment and knowledge gain of individuals and 3) Providing computational methods in the field of natural language processing (NLP) for detecting and classifying roles and hierarchies in large-scale online discourse. The project will focus on discourse about COVID19 in online news and social media, specifically Twitter, and tap into unique data sources of large scale online discourse with a particular focus on discourse in the DACH countries and anglophone regions.