Inferring And Possibilities, Rather Than Natural Laws, In Robust Climate Modeling

One concern raised about sciences which rely upon simulation models (such as climatology) is that the nature of simulations calls into question the soundness of the inferences that can be drawn from them. I argue that this concern stems from a belief that simulation models must provide laws in order to count as rigorous science, when in actual practice simulation models can investigate a variety of types of possibility with differing inferential potential, and that these differing inferences are necessary parts of the experimental process. I appeal to philosophical work in epistemology to make the argument that simulation models in general - and climate models in particular - explore different kinds of possibility, from logical possibilities to physical possibilities. I then argue that not only is this plurality of inference compatible with robust modeling practices, but that it leads to stronger inferences for climatology.