Abstracts

Scientific Models: Against Fictionalism and Representationalism

Guilherme Sanches de Oliveira University of Cincinnati

In recent years, many philosophers have drawn comparisons between scientific models and artistic products such as paintings and works of fiction, bringing insights from aesthetics to bear on questions about the relationship between models and the systems and phenomena they are used to investigate (e.g. French (2003), Godfrey-Smith (2006, 2009), Frigg (2010), Contessa (2010), and Toon (2010)). For them, the comparison between science and art is meant as more than mere analogy, rather revealing something deep about the fictional nature of models as, like paintings, imperfect and limited depictions of the world or, like literature, descriptions of an imagined world or of what this world might have been like. In this paper I examine these accounts in light of Morrison's (2015) moderate perspective, which, rather than calling all models "fictional," reserves this characterization for the latter kind, namely, models of imaginary targets. As I argue, both "extreme fictionalists" and Morrison's moderate position fit in a broader conceptual framework that can be identified as the "representational view of models," holding that models are used for indirect investigations of target phenomena because models represent those targets (Morrison & Morgan 1999).

Despite its popularity, the representational view of models is riddled with problems, both general (e.g. accommodating idealizations and abstractions) and particular (e.g. the anything-is-similar-to-anything-else-in-some-way objection to pure dyadic representation, and the related anything-can-represent-anything-else-for-someone objection to triadic representation). A powerful but still widely neglected alternative to the representational view is the artefactual view of models, according to which models are not truth-bearing representations, but rather autonomous tools or devices. As such, models can be more or less useful for certain purposes, but they are useful by virtue of what they present rather than what they represent. In articulating the (anti-representational) artefactual view of scientific modeling, I reject fictionalist approaches and the comparison between science and art, favoring instead the analogy with engineering: like scaffolds, models function as tools which facilitate our engagement with the world rather than trying to represent it. Thus, models are never fictional, but can play a mediatory role in inquiry depending on what uses they afford.