New Frameworks for Approaching the Study of Discursive Polarisation

It’s the second and last day of the ECREA PolCom 2023 conference in Berlin, and it starts with a panel on polarisation that I’ve had a hand in organising. We begin with Michael Brüggemann, whose focus is on discursive polarisation. He begins by pointing out that polarisation is often ill-defined, and the communicative dimension in particular is often under-conceptualised and under-researched.

Discursive polarisation is when debates break apart: a multi-dimensional divergence emerging in and through communication. There is also a more intuitive aspect to polarisation, as is demonstrated for instance with the German debate around the Letzte Generation climate activists (whom some commentators call ‘extremists’ and ‘terrorists’).

Polarisation outside of communication is generally described as a multidimensional process with an ideological and an affective dimension (but this breaks down even further into smaller categories). Political science research in this field is dominated by US perspectives, and often survey-based and focussing on people’s political attitudes rather than communicative acts. What might a more communication-informed perspective on discursive polarisation look like, then?

We might build on Dan Hallin’s idea of the three spheres of discourse: a sphere of consensus, of legitimate controversy, and of deviance; polarisation then shows that there is an eroding and dissolving common ground, which also erodes shared assumptions about reality. This is also overlaid with the model of four interconnected polarisations: institutional polarisation leading to false (perceptions of) polarisation leading to affective polarisation leading to ideological polarisation. This is necessarily driven by communication.

News media play a critical role in this too. They are both an arena for public debate, and structured by journalistic practices – yet there appear to be no comparative studies of polarised or polarising news content across countries. Polarising journalistic practices would be expected to simplify debates as battles between two camps, which are represented by their most extreme voices, and therefore further increase perceptions of polarisation; the same may be true also for patterns on social media.

Discursive polarisation thus consists of both ideological and affective polarisation; in each case these need to be studied both through content and on interaction patterns. In ideology, these produce polarised framing of content and polarised issue networks; in affect, they produce polarised group content and polarised group networks.

We also need better empirical methods to study this. Here, there is a critical question about the extent to which state-of-the-art computational social science methods can be employed to help with these tasks – e.g. for topic modelling or framing analysis. This isn’t a purely quantitative and computational exercise, of course: manual, qualitative interpretation of the results of such computational steps is critical, and mixed methods are always preferable.

In the case of German coverage of Letzte Generation and Fridays for Future, incidentally, framing appears to be largely uniform across a broad range of news outlets: there is no evidence of fundamentally diverging polarisation of these different climate action movements. Retweet networks for these two movements show much clearer polarising tendencies, however.

Finally, it’s important to note that polarisation isn’t always a problem: kept within limits it can be productive of meaningful debate in democracies, too.