Friday, September 12, 2014
Wes and Me, and Search
Carnegie Mellon U., Dept. of Philosophy
3:30 pm, 817R Cathedral of Learning
Abstract: The first part of this lecture will offer some remembrances of my years with Wes Salmon. In the second part I will suggest that the logical tradition in philosophy of science of the first half of the 20th century was implicitly an attempt to characterize a universal search space for all of science. There is, so far as we know, no such universal search space, but science historically and today has many local, miniature variants of the logical empiricists' grand schemes. The automation of statistical search methods is moving the "creative" part of science away from conjecturing particular theories and towards framing hypothesis spaces of alternative models and their connections with kinds of data so that feasible methods can extract true information--if it is in the hypothesis space at all. I suggest (well, something stronger than suggest) that the chief way in which philosophy of science can actually aid science is through the characterization of "search spaces" and the description of methods to search them. The results include reconstructions of how various scientific theories could be rationally justified, methods for addressing contemporary scientific problems, including a host of strategies for searching for causal relationships. I will give examples motivated by or directly applying philosophical work to: particle physics, space physics, classical cognitive neuropsychology, genomics, phylogeny, climate science, and brain imaging. Or at least some of these examples.