Dr. Humphreys is investigating the ways in which computer simulation methods can and cannot be adequately accommodated within existing philosophical accounts of scientific method. He is placing particular emphasis on the relationships between laws and specific conditions, on the role of analogical reasoning in the construction of simulation models, and the effects that the absence of an explicit deductive structure in the models has on our understanding of these models. The research will focus during the first part of the grant period on models in physics and astronomy, especially those that use Monte Carlo estimation techniques. During the second half of the grant period, he will focus on identifiability issues in social science models. While carrying out the research, he will make various points of connection with current work in cognitive science that uses simulation methods.
