17th International Congress on Logic, Methodology and Philosophy of Science and Technology

JULIÁN REYNOSO; SOFÍA MONDACA

Congreso; 17th International Congress on Logic, Methodology and Philosophy of Science and Technology; 2023

Scientists devote a large portion of their working careers to testing, comparing, and revising models as they have become paramount for producing knowledge. In this vein, it has been argued that models perform different epistemic functions ranging from understanding [(de Regt et al., 2013)](https://www.zotero.org/google-docs/?REzExU) and explanation, (not only of their target systems but also about models themselves and instruments) to concept creation and formation in different contexts, such as pedagogical or as training devices. Yet the construction of these models is not without issues. Idealizations, parametrization, and several other tools are used to tailor the model's behavior to observed data, rightfully bringing forward questions as to exactly how reliable models are and to what extent we can trust in their outputs. George Box famously summarized these issues: “all models are wrong, but some are useful”.We gain knowledge about models by interacting with them. Thus, attending to the general practice of scientific modeling is a promising direction to approach the question of the capacity of models to provide knowledge. Studies on human expertise provide a fertile ground for analyzing such practices and with them, some important aspects that allow us to understand our confidence in the epistemic power of the models we use. This relationship has been scarcely studied in the literatureIn this paper, we set out to argue that a pragmatic view of models as put forward by Boon & Knuuttila [(2009)](https://www.zotero.org/google-docs/?pTz1Jo) provides a better understanding of the role that models take when dealing with trust-assessment issues. From this perspective, models are dynamic epistemic tools rather than static representations of systems. This perspective might offer a better view of actual scientific practice, in which the scientist's expertise is considered a crucial dimension product of socialization into a community of experts  [(Collins & Evans, 2018)](https://www.zotero.org/google-docs/?5Kbjtz)As inquiry moves forward, new elements appear that allow us to analyze and evaluate scientific knowledge. We do not focus anymore on the ability of models to represent as accurately as possible the represented world, but rather on our epistemic goals and the practical ability to develop them through models [(Elgin, 2017)](https://www.zotero.org/google-docs/?2uh8nR). Thus, analyzing models from a pragmatic perspective allows us to understand scientific knowledge from a practical dimension, emphasizing its close relation with scientific expertise. It is not only about acquiring precise information but also about developing abilities that allow constructing, ordering, and directing such information in function of epistemic objectives determined by scientific communities or specific contexts, such as policy, decision-making, or risk assessment.