Scientific Models in Philosophy of Science

4 Theories

I    Faraday, Kelvin, and Maxwell, models were central tools for the development of scientific accounts. These models were predominantly mechanical models, and they frequently guided the mathematical treatment of increasingly less mechanical phenomena, such as electromagnetism (see chapter 2). Mathematical treatment of phenomena, in turn, is one factor that furthers the emergence of theoretical concepts. Although mechanical approaches maintain the appearance of familiarity so that it would seem as if the mechanically explained process in question is directly observable, this is blatantly not the case when the approach becomes more mathematical. Indeed, the mathematization and abstractness of physics went to new heights— for example, with the non-Euclidean geometry employed in Einstein’s Theory of Relativity or with quantum theory. Modern physics seemed to require that physicists focus on interpreting empirical phenomena in abstract and theoretical terms, marked by the extensive use of specialized mathematical methods in physics, worlds apart from the mechanical models that could either be physically built or at least be observed before the mind’s eye. The goal of this chapter is to analyze why, despite their practical importance in nineteenth-century science, models did not receive any (positive) attention during the first half of the twentieth century. This 81 1 2 3 4 5 6 7 8 9 S C IENTIFIC M OD E L S M EC H ANIC AL MOD EL S AN AL OGY THEORIE S PARAD IGM S AND M ETAPH ORS THE S EM ANTIC VIE W AND THE S TUDY OF SC IEN TIFIC PR ACTIC E PH EN OM EN A , DATA , AND DATA MODE L S REPRE SE NTATION CONCL USION Bailer CH4:Layout 1 7/5/09 2:21 PM Page 81 is very much a negative piece of history of philosophy of science regarding models, emphasizing what did not get said about models. Yet the almost total disregard of the role of models for a fifty-year period needs explaining. For this reason, this chapter talks much about the role attributed to theories. The received opinion was that good theories rendered models theoretically and practically redundant. As a hangover from the early days of philosophy of science, even now the term “model” very often still implies that a chosen description is in a merely preliminary version, to be confirmed later. “Theory” has the connotation of being well established and well abstracted—that is, free from the little failings and inaccuracies of models. According to this common perception, models are mainly of temporary benefit; they come and go, while theories last. Models are thought to be accessible at the expense of the correctness, generality, and precision of theories. As a consequence, it requires less commitment or conviction to call something a model simply because it is “only a model,” still awaiting correction and improvement. The underlying expectation is always that in the longer term the model will be replaced once a definitive account has emerged in the form of a theory. However, this received attitude toward models not only conceals the benefits and advantages scientists derive from the use of models, but it leaves important philosophical issues ignored that, after a long period of disinterest and even ridicule, led to a philosophical reevaluation of scientific models. The historical lesson of this chapter meets its philosophical response in chapter 6, where I argue how and why the balance between theories and models needs to be redressed. In section 4.1, I begin with the successful application of mathematical descriptions that supported theory dominance, picking up a thread from chapter 2, section 2.1. Theory taking axiomatic form and being reconstructed according to the “hypothetico-deductive method” is introduced in section 4.2. Pierre Duhem, who made appearances in chapters 2 and 3, has reached notoriety for playing down models as a tool for lesser minds, but he is also a proponent of the hypotheticodeductive approach to theory. His influential views on the matter are reviewed in section 4.3. Formulating theories also involves postulating theoretical entities. Section 4.4 introduces the operationalist and Logical Empiricist responses to the problem of linking theoretical 82 Theories Bailer CH4:Layout 1 7/5/09 2:21 PM Page 82 postulates to observational evidence.₁ The Logical Empiricist emphasis on the context of justification, rather than the context of discovery, is a further building block in accounting for the disregard of models...