Abstract

How can the harm caused by waves of fake news or derogatory speech on social media be minimized without unduly limiting freedom of expression? I draw on an eighteenth-century debate for thinking about this problem: Hume and Smith present two different models of the transmission of emotions and ideas. Empathetic processes are causal, almost automatic processes; sympathy, in contrast, means putting oneself into the other person’s position and critically evaluating how one should react. I use this distinction to argue that the architectural logic of social media should be improved to prevent cumulative harms and to facilitate sympathetic processes.