Scientific modelling with diagrams

Abstract

Diagrams can serve as representational models in scientific research, yet important questions remain about how they do so. I address some of these questions with a historical case study, in which diagrams were modified extensively in order to elaborate an early hypothesis of protein synthesis. The diagrams' modelling role relied mainly on two features: diagrams were modified according to syntactic rules, which temporarily replaced physico-chemical reasoning, and diagram-to-target inferences were based on semantic interpretations. I then explore the lessons for the relative roles of syntax, semantics, external marks, and mental images, for justifying diagram-to-target inferences, and for the "artefactual approach" to scientific models. Keywords Representational models • Mental images • Notation • Mechanism • Physical models • Syntactic symbol manipulation • George Gamow • Francis Crick • Protein synthesis 1 Introduction Many scientific diagrams purport to depict features of the world. Some diagrams can also be employed as "representational models" (Frigg and Hartmann 2012), i.e. as systems that aim to represent the world and, in addition, enable scientific investigations to be carried out on the model, rather than on reality itself. As representational models, diagrams are manipulated in lieu of their representational targets (Giere 2002; Downes 2012). Chemical formulas have received particular attention in this respect. Berzelian formulas like 'H 2 O' depicted chemical substances and allowed that marks on paper, like 'H' and '2', be rearranged in order to learn about the chemical reactions in which Electronic supplementary material The online version of this article (