Hanti Lin (University of California, Davis)

The problem of theory choice and model selection is particularly hard when the standard of statistical consistency is too high to be achievable, that is, when no inference procedure is guaranteed in a statistical sense to eventually identify the true theory given what is known or assumed. This paper studies a crucial instance: the problem of inferring causal structures from non-experimental data without assuming the so-called causal Faithfulness condition or the like. A new account of epistemic evaluation is developed to solve that problem and to justify proceeding as if the Faithfulness condition or the like were accepted as true.

Kino Zhao (University of California, Irvine)

At its strongest, Hume's problem of induction denies the existence of any well justified assumptionless inductive inference rule. At the weakest, it challenges our ability to articulate and apply good inductive inference rules. This paper examines an analysis that is closer to the latter camp. It reviews one answer to this problem drawn from the VC theorem in statistical learning theory and argues for its inadequacy. In particular, I show that it cannot be computed, in general, whether we are in a situation where the VC theorem can be applied for the purpose we want it to.

Corey Dethier (University of Notre Dame)

In spite of its massive influence, Duhem's argument for testing holism rests on a mistake: it conflates the assumptions necessary for the derivation of an empirical consequence with the assumptions necessary for that consequence to be evidence. Using examples from physics and biology, I argue that these come apart and the nature of the latter depend on the details of the case. This hundred-year-old mistake has surprising relevance for contemporary discussions of the epistemology of models, simulations, and testing.

Klodian Coko (University of Western Ontario)

I argue that the epistemic strategy of multiple determination (i.e. the epistemic strategy of using multiple, independent procedures to establish "the same" result) is not a form of robustness. There are many characteristics that distinguish multiple determination from robustness. They are all, however, related to the same core difference: whereas the different robustness variants can be considered as involving some invariance to different types of perturbations, multiple determination cannot. Multiple determination is better distinguished by its ability to support a specific type of a no coincidence argument. Namely, it would be an improbable coincidence for independent procedures to establish the same result and yet for the result to be incorrect. No such argument can be construed from invariance to perturbations