Abstract

Philosophy of science has characterized scientific knowledge as fundamentally propositional . This account leads to an inability to recognize and articulate the significant role of non-propositional, visual representation in the practice of science. Toward the development of a more productive framework for understanding visual representation in science, the present study critiques the standard philosophical view, reviews the literature on visual representation in science, and examines the scientific case of neuroscience. Specifically, the study looks at current research known as "functional mapping of the human brain" which uses neuroimaging technologies such as positron emission tomography in the localization of particular functions in the human brain. This case study suggests that, contrary to previous accounts in philosophy of science, visual representation and propositional representation work together as parallel, interacting, interwoven practices in the discovery/construction of scientific knowledge. Visual representations are not reduced to--translated into--linguistic and mathematical propositions, but instead are used in their multidimensional forms. The present study focuses on the external, materialized visual representations used in the practice of science, as opposed to internal, mental images that might be studied by cognitive psychology. Several hypotheses are proposed which describe aspects of the role of these visual representations in the practice of science. These include the following: a primary function of visual representation is the representation of structure which is made possible by the use of actual space in the process of visual representation; visual comparisons among visual representations of real objects and processes, visual representations of theoretical models, and visual representations of integrations of the two are used to assess the fit of the theoretical models to the real objects and processes; the construction of visual prototypes or prototypical visual representations of objects of scientific research are used to categorize and stabilize--hold stationary for the purposes of scientific examination--such objects; and the construction of material models make it possible for scientists to interact with and/or manipulate materialized structures as representations of real systems