Computational modelling of social cognition and behaviour
Theodoropoulos, Nikos C.
Philosophers have always been interested in asking moral questions, but social scientists have generally been more occupied with asking questions about morality. How do people differ with regards to their morality? How frequently are moral values inconsistent, thus resulting in internal conflicts? How likely are people to revise their moral beliefs? The aim of these questions is to explore moral reasoning and identify patterns of moral behaviour between people. Simultaneously, social scientists have moved beyond the exploration of small-scale, static snapshot of networks onto nuanced, data-driven analyses of the structure, content, and dynamics of large-scale social processes. This drives researchers to use far more elaborate tools, such as automated text analysis, online field experiments, mass collaboration, machine learning, and more generally computational modelling, to formulate and test theories (e.g., Evans & Aceves, 2016; Molina & Garip, 2019; Nelson, 2020; Salganik, 2019). It is fair to argue that social sciences are on the verge of a new era, an era in which computational methods and large-scale data are the primary tools/sources of gaining information and knowledge. In this dissertation, I will focus on developing formal models of the cognitive dissonance involved in moral values conflicts within individuals, and how this might be reduced. I will also attempt to extend this to connect with research linking moral and political psychology. Then I will try to explain echo chamber development, as a socio-cognitive phenomenon, arising from dynamics described in chapters 2 and 3. Finally, I will focus on moral belief updating, as an alternate (class of) response(s) in chapter 6. I try to explain these phenomena by bringing together cognitive and social theories. The three principal theories we build upon are Festinger’s Cognitive Dissonance, Bandura’s Moral Disengagement and Haidt’s Moral Fountations Theory. As it is detailed in the forthcoming paragraphs, the union of these theories, alongside with computational modelling, sparks off some interesting hypotheses. We now go ahead and discuss why computational modelling is a powerful tool in social sciences, and then present a historical background for each of the aforementioned theories.