As Ezra Klein says, Twitter, and other social media platforms, have come to function as our contemporary “Town Squares”. A huge proportion of political life now occurs online. According to a standard, rationalist view, the goal of political discourse is to share information and arguments. If that were true, Twitter, which makes this easier than ever before, should be a boon to political life. And yet, it is a familiar fact that online political discourse is highly dysfunctional: Jonathan Haidt characterises it as filled with “anger, contempt, insults and obscenity”.

Many recent discussions focus on the role of social media algorithms in amplifying toxic content. I want to explore a different line of explanation – the moral psychology of the human beings who use social media. A large body of psychological research indicates that physical, and especially facial, cues are vital for promoting empathic responses. When we debate on Twitter, these cues are systemically absent. The design of Twitter creates a low-empathy environment. If we accept an emotivist view of political discourse, according to which a core goal of political communication is to create emotional connections, it is no surprise that Twitter should be disappointing as a forum for political life. Empathy is central to political life, and we shouldn’t assume that it will function the same way in a world of avatars.