Philosophy of Social Science
Are inanimate objects like billiard balls governed by the same kind of laws as social beings with intentions? Can we explain features of the social world, like driving conventions and tribal cooperation, with the same sort of science we use to explain features of the natural world, like covalent bonds and superconductivity? Can we characterize the difference between what we can learn from active experiments vs. passive observations about the causal relations in a social system? Can we discover something about which features or properties of social agents underlie their observable behavior, for example whether they have multiple intelligences and if so, how many? Under what conditions can collectives of rational agents act rationally?
These and other questions like them constitute the philosophy of social science as we practice it at Carnegie Mellon. Using primarily game theoretic tools, Peter Vanderschraaf, and his students investigate how we might explain features of social systems like conventions and cooperation. Combining graph theory and statistical models, Clark Glymour, Peter Spirtes, Richard Scheines and their students investigate the methodology of causal social science, that is, what can and cannot be learned about the causal relations governing a social system from statistical “studies” under a variety of assumptions. Teddy Seidenfeld and his students explore the nature of rationality, and apply it to collectives as well individuals. Seidenfeld and Kevin Kelly explore the foundations of statistical methods, perhaps the most important feature of social science.