Modelling Social Complexity in Argumentation

Argumentation theory studies how the structure, content, and sender of an argument influences how persuasive that message is to the recipient. This has been studied normatively and descriptively, theoretically and empirically, or qualitatively and computationally. These explorations typically take point of departure in one recipient in isolation who receives an argument.

Aside from a few important exceptions in argumentation theory (i.e., pragma-dialects, informal logic), empirical studies on argumentation have largely ignored the social component of argumentation. A focus on studying argumentation in isolation is methodologically sound: it allows researchers to isolate and control experiments, explore individual pieces of argumentation, and adjust and compare key variables. However, argumentation normally does not occur in a vacuum. Rather, it is almost always a social phenomenon – whether at dinner parties, in a pub, or online, we engage and argue with others.

Social networks and the flow of arguments between peoples is complex –on social media people choose who they follow, whether to engage with a person on a particular topic, seek out information as they see fit. As people increasingly contribute to public discourse and argumentation, the flow of information changes from a traditional top-down media landscape to a bottom-up mass media landscape. In recent years, researchers have developed models that can account for these complex information systems. So-called agent-based models represent the engagement between peoples in dynamic and adaptive systems.

With the advent of social media, it is critical to understand how arguments can flow in these systems, as this influences the choice of political leaders and shape the public understanding of climate change, and other key issues in democracies. The workshop brings together argumentation researchers and agent-based modellers.

The aim is twofold. For argumentation theorists, the workshop provides a space to become acquainted with ABMs and dynamic modelling. For agent-based modellers, the workshop provides a space to implement and expand upon computational techniques and methods – typically, agent-based models are used in environmental systems, but have only more recently been applied to cognitive and information-related aspects (such as echo chamber formation and information diffusion). Given this relatively new expansion of ABM application, the workshop provides an opportunity to learn about computational approaches to argumentation that are directly implementable in ABMs and to test these with novel methods. More broadly, the workshop aims to foster collaborative work between the argumentation and ABM communities and to exchange ideas and methods.