Models, Simulation, and 'computer Experiments'

Introduction: As Nelson Goodman famously observed, “Few terms are used in popular and scientific discourse more promiscuously than “model.” (Goodman 1968, 171) Writing more than 30 years later, much the same might be said of the term “simulation.” Yet this was not always the case. Both words have ancient histories, but until very recently, the meaning of “simulation”, at least, was manifestly stable: it invariably implied deceit. Usages offered by the OED prior to 1947 include: “false pretence”; “A Deceiving by Actions, Gestures, or Behaviour” (1692, OED); “a Pretence of what is not” (1711, OED). Evidence provided by the OED, in short, suggests that it was only after WWII that the word took on the meaning that brings it into its current proximity with models: “The technique of imitating the behaviour of some situation or process ... by means of a suitably analogous situation or apparatus, especially for the purpose of study or personnel training.” Here, the valence of the term changes decisively: now productive rather than merely deceptive, and, in particular, designating a technique for the promotion of scientific understanding. The shift reflects a crucial change not only in the perceived value of simulation, but also, as others have already noted, in the means of production of scientific knowledge. Furthermore, it is this new sense of the term that encourages its use in much of the current historical and philosophical literature as either interchangeable with the term model, or as one part of a single composite noun (as in “models and simulations”). An obvious question arises, however, and it is this: do the actual uses of simulation in contemporary scientific practice in fact warrant such facile assimilation? Or, to pose the question somewhat differently, does the use of simulation in post WWII science add significantly new features to the range of practices that had earlier been subsumed under the term “modeling”? My answer is yes, but I argue that the novelty has been multi-layered and cumulative in its effects, requiring a more nuanced history than has yet been made available. The rise of simulation in post WWII science is not exclusively associated with the advent of the computer—in fact, the earliest invocations of the term relied primarily on the use of electrical and electronic analogue devices designed to mimic the behavior of real world