| The widespread use of models and simulation modeling in scientific practice reveals a scientific methodology in conflict with many of our basic epistemological intuitions. Scientific modeling strategies deliberately employ falsehoods: idealizations, abstractions, approximations, and other fictional elements. Nonetheless, models are successfully used by scientific inquirers, trusted to perform diverse inquiry-based tasks, and thought to provide genuine explanatory power. Contemporary philosophy of science faces a major challenge: how to rethink the methodological norms that guarantee scientific success in light of the persistent use of fictive strategies in scientific inquiry. This dissertation answers the challenge by arguing that reliability should replace traditional epistemological standards, such as truth or empirical adequacy, as the norm in model-based inquiry.;To illustrate the methodological power of scientific models and the fictional strategies used in modeling techniques, this dissertation examines two case studies from nanoscience. The case studies demonstrate that nanoscale inquirers use models not because they are 'true' or even empirically adequate, but because they are functionally fit. This dissertation shows how the concept of reliability gives functional fitness the necessary epistemic dimension to secure sound methodological practice.;The philosophical importance of the commonsense notion of reliability was first acknowledged by John Dewey. In this dissertation, Dewey's notion of reliability is transformed into a fully articulated epistemic norm. In my account, reliability is a property assigned to a variety of scientific items (such as theories, instruments, strategies, models) in virtue of their performance in problem-solving activities. Recognizing how the principle of reliability operates in practice both illuminates the structure of model-based inquiry and justifies the use of fictions in scientific practice.;Finally, this dissertation argues that nanoscale inquiry illustrates the increasing explanatory relevance of scientific models. The challenge for the philosophy of science is to demonstrate how model-based explanation is possible. This dissertation responds to the challenge by showing how my proposed conception of scientific methodology leads to a novel account of scientific explanation. On my account, explanation is best characterized as an activity---not a form of reasoning---and successful explanatory strategies are governed by the principle of reliability. |
