Here I analyze the anatomical thought of the French physician and naturalist Félix Vicq d'Azyr (1748–1794) in order to bring to light its importance in the development of comparative anatomy at the end of the eighteenth century. I argue that his work and career can be understood as an ambitious program for a radical reform of all biomedical sciences and a reorganization of this whole field around comparative anatomy, on the conceptual as well as the institutional level. In particular, he recommended a close connection between anatomical and physiological studies, and a generalization of the comparative approach towards organs and functions in man and animals. This conception led him not only to reform the scope, the methods, the style of description, and the vocabulary in anatomy, but also to construct a new classification of living beings and to pursue a quest for laws of organization. This strategy was successful, since Vicq d'Azyr was able to promote his thought as well as his institutional position efficiently. The Revolution and his untimely death prevented him from achieving his program, but his attempt would serve as an example for younger scientists like Cuvier.

We continue our analysis of modeling practices that focus more on qualitative understanding of system behavior than the attempt to provide sharp forecasts. The argument here is built around three episodes: the ambitious work of the Princeton Meteorological Project; the seemingly simple models of convection in weather systems by Edward Lorenz at MIT; and then finally analysis of the dripping faucet by Robert Shaw and the Dynamical Systems Collective at UC Santa Cruz. Using the Princeton Meteorological Project as an argumentative foil for the later chaos work of Lorenz and Shaw, we first show how the epistemological interest of modeling came to shift from issuing predictions to probing the very meaning and limits of prediction. The second step of our argument shows that what may be seen in one context of use as a modeling technology that is error ridden, imprecise, or inadequate, may be parsed completely differently in another context. This argument about technology and practice, we argue, feeds through to epistemological conceptions of error. Far from being something that can be defined in the absolute, the notion of error is shown to be contextually plastic.

This paper is the first part of a two-part examination of computer modeling practice and philosophy. It discusses electrical engineer Jay Forrester's work on Industrial Dynamics, later called System Dynamics. Forrester developed Industrial Dynamics after being recruited to the newly-established School of Industrial Management at the Massachusetts Institute of Technology (MIT), which had been seeking a novel pedagogical program for management for five years before Forrester's arrival. We argue that Industrial Dynamics should be regarded in light of this institutional context. Unlike economics, as well as operations research and management science (twin fields which were also gaining traction at that time), Industrial Dynamics was not meant to be applied by mathematicians and technical specialists in consultation with managers, but by managers themselves. This concern shifted the emphasis in modeling from the specialist act of analysis of a crucial problem to the managerial act of choosing which problems were crucial and developing effective policies around them.