Limits to Modeling: Balancing Ambition and Outcome in Astrophysics and Cosmology

Computer simulations of complex physical objects and processes for which data are very sparse or inexistent have become a major tool of scientific investigation in astrophysics and cosmology. However, one must ask how these simulations acquire their epistemic credentials and whether their realistic ambition is legitimate. A close look at two model-building processes—one in galactic astrophysics, the other in cosmology—reveals heretofore underappreciated features of simulations, such as path dependency. This article argues that such features undermine our confidence in the outcomes of the simulation. Case studies presented here reveal a general tension in computer simulation between realistic ambitions and the possibility of empirical confirmation. The analysis will thus lead to a reassessment of the epistemic goals actually achieved by composite models of complex astrophysical and cosmological phenomena.