Full Program »
Reconsidering Scientific RepresentationThis poster features my research project engaging with recent literature on modeling in the sciences. It investigates one of the slogans commonly adopted by authors working in this area: the idea that the empirical success of models consists in their being involved in a particular relationship with their target real-world system—a representational relation. Models are said to represent their targets; they do so in such a way that they allow researchers to draw inferences about their targets by considering the features of the model as they compare to the target. On the face of it, this is a perfectly reasonable claim to make. A given model must have some specifiable connection to its target in order to be a reliable source of information about that target. However, the manner in which this connection is usually characterized calls for greater clarification.
The organization of the poster would reflect the following three-part structure of my initial intervention into the literature. The poster would supplement the main points and supporting arguments on display with imagery that allow viewers to visualize the flow of arguments, better understand the mathematics and experiments of the case study, and see the representation relations I consider in graphic form:
- First, I present a critique of the way the representation relation is characterized in the current literature. The focus of this critique is an assumption common to various theories of models. Authors assume that there is an account on hand of the mathematical structure or relevant features of the target system to which a model may be compared, but they do so without specifying how such an account was arrived at. I argue that in many cases it is likely that the available descriptions of a target depend on conceptual resources that are intimately tied to the relevant theories and models that are available, and that this poses a problem for the notion of ‘comparison’ employed by authors in their accounts of the empirical success of models.
- Second, I present a historical case study examining how a typical model was used in the context of scientific discovery. The focus here will be on Francis Crick and William Cochran’s use of a mathematical model of a discontinuous helix in the quest to determine the structure of DNA.
- Third, from this historical case study, I draw some conclusions that I generalize in order to take a first step toward an alternative account of the empirical efficacy of models. The main claim is that a model primarily represents a target system in virtue of its ability to license correct inferences about the results of interventions on the target. While the primary inferences drawn from models concern predictions of data, they also allow for new interpretations of data. In other words, they give the conditions under which the representative status of a model is confirmed or disconfirmed, and so the extent to which the features of a target system can be understood in terms of the features of the model.
History and Philosophy of Science
University of Pittsburgh