Shaping Science Policy

Abstract

Agent-based models are models that simulate individuals behaviors with the aim of analysing, evaluating and understanding relationships between micro-level behaviors and macro-level phenomena of complex systems. In recent years, the agent-based approach has found effective entry into social epistemology in the form of so-called SOSR models, which are used to examine relationships between socio-epistemic sys- tems and their epistemic outcomes. An overarching aim of social epis- temologists is to use these models to provide normative suggestions for redesigning, restructuring, or reorganising real-world socio-epistemic systems. This thesis defends an instrumentalist view that accounts for this aim in terms of model-based instrumental norms: norms that prescribe socio-epistemic systems as means for achieving epistemic outcomes, which are taken as ends that we have reason to pursue. In laying out this view, we identify three factors relevant to justifying model-based instrumental norms: model evaluation, argumentative context and instrumental transmission. A notable consequence of this instrumentalist view is that the normative aim of social epistemolo- gists cannot be successfully achieved using exclusively philosophical or exclusively modelling methodologies. Instead, the success of this aim depends on a wide collaboration of epistemologists, modellers, moral philosophers, policy-makers and the like to ensure that reasons for particular epistemic outcomes produce undefeated reasons for socio- epistemic systems.