Title: Analyzing and comparing models according to their functions, natures, principles, and uses
Dr. Franck Varenne - University of Rouen – Normandy (France) & ERIAC (UR 4705)
Aix-Marseille Université
INS - Faculté de Médecine, 27, Boulevard Jean Moulin
Marseille, 13005, FranceIn view of the huge variety of model types, how do modelers come to prefer one model to others to which they have compared it? In addition to the specific problem of validation, the question of the actual comparability of models arises. But are they comparable? Don't they all complement each other, as it is often said? But here again: in what way? And from what point of view? In reality, the question of preference still arises: even if they are said to be complementary, why should we now prefer to work on this type of model rather than another? What can we expect from each of them? How can we more precisely characterize the limits of our model's contribution to scientific knowledge and practice? What we need here is an analytical and epistemological grid for comparison.
An epistemological approach that is both historical and comparative can show, for example, that contrary to what is often assumed, the nature of a model does not in itself determine its knowledge function, nor does this knowledge function determine the use (heuristic or social) that is made of it. In fact, it can be shown that a minimal conceptual grid for analyzing and comparing models requires distinguishing between at least four of their essential characteristics. These are: their functions, their natures, their principles, and their uses. As we can see, these characteristics are also varied in nature. This certainly adds to the complexity of the subject. But recognizing this fact also provides us with a more precise, more adaptive analysis tool, more likely to follow the model through its development and life cycle.
Above all, it appears that these four characteristics, while not unrelated, are not mutually reducible, nor do they systematically determine each other. As a result, their mutual articulations, while often rationally motivated in context, are themselves varied. In this talk, I will propose a definition of each of these four characteristics. And I'll give several examples of how they can be both contextually and rationally articulated. I will conclude by mentioning the supplementary degree of epistemological complication added by the development of certain new types of formal models: simulation models, multi-models, multi-scale and/or multi-process integrative models.
Selective bibliography:
Frigg R., Hartmann S., 2020 (1st ed. 2006). “Models in science”, in Zalta E.N. (Ed.), The Stanford encyclopedia of philosophy,https://plato.stanford.edu/entries/models-science/
Morrison M., 2015. Reconstructing reality. Models, mathematics, and simulations, Oxford, Oxford University Press.
Varenne F., 2018. From models to simulations, trad. K. Turnbull, London, New York, Routledge, Chapter 8 in particular. Open access: From Models to Simulations.
Varenne F., 2022. « Comparer les modèles à l’aide du vecteur caractéristique : fonction, nature, principe et usage des modèles »,Natures Sciences Sociétés, 30, 1, 93-102.https://doi.org/10.1051/nss/2022014