Seoyeon Park
University of Pittsburgh
Models are a useful instrument in science, enabling us to explain the behavior of complex systems. However, the explanatory power of models is puzzling given that models typically contain idealizations, false assumptions about real-world systems. My goal is to clarify the role of models in scientific knowledge. I claim that models enable the regularities underlying real-world systems to be epistemically accessible. It turns out that idealizations are necessary for the model to play this role. Accordingly, I argue that models make a distinctive kind of representation.
First, I clarify the role of models by exploring the distinctive features of scientific knowledge. The target of scientific knowledge is the underlying regularities. We explain a system behavior by capturing its behavioral pattern within our theory, rather than specifying its accurate description. Hence, we are entitled to scientific knowledge when we can answer what-if questions about how the system would behave under certain counterfactual conditions. Models enable scientists to apply the current theory to complex real-world systems and answer what-if questions. So, the role of the model is to increase the epistemic accessibility of the underlying regularities. I propose that models can play this epistemic role due to idealizations, which help us focus on regularities. I analyze two kinds of idealized models. Galilean models can be de-idealized, so they increase epistemic accessibility in a weak sense of enabling us to answer the what-if questions more easily. Non-Galilean models do not permit a de-idealization. Targeting are real-world systems to which our current theory is inapplicable, non-Galilean models enhance epistemic accessibility in a strong sense of allowing us to answer the what-if questions that otherwise remain unanswered. Lastly, I argue that models represent the real-world system in a sui generis way and thus a special kind of justification is required. Models do not make alethic representations whose correctness standard is truth, for models essentially contain false idealizations to reveal a general regularity. Models neither make fictional representations as models have intentional objects in reality. Thus, there is a distinctive kind of model-based representation. I briefly explore how this representation can be justified.
Chair: Niklas Parwez
Time: September 11th, 18:20 – 18:50
Location: SR 1.004, online
SOPhiA 2026 