Final thought on the reference class problem

Final thought on the reference class problem

And if we reject the counterintuitive advice to Adam & Eve et al. then there also are cases committing us to:

There is room for diverging reference class definitions within these limits. At least as far as the arguments provided here are concerned, there is thus a class of reference definitions that are rationally acceptable. One cannot rule out that there are other arguments that will be discovered in the future that will narrow down the range of permissible choices further. But for the time being, what we have is a firm lower bound (Non-minimalism) and some considerations that may give one to reason in some cases to pick a reference class that is much smaller than the universal reference class.

My suspicion is that at the end of the day there will remain a subjective element in the choice of reference class.⁹¹ I think there is a subjective element in the choice of an ordinary Bayesian prior credence function over the set of (non-indexical) possible worlds. I do not hold that every such possible prior credence function is reasonable; but I think that after everything has been said and done, there is a class of credence functions that would all be reasonable and perfectly defensible in the sense that intelligent, rational thinkers could have any of these credence functions even in an idealized state of reflective equilibrium. (This could perhaps be said to be something of a “received view” among Bayesian epistemologists.) What I am suggesting here is that a similar and additional subjective element exists in regard to credence assignments to indexical propositions, and that this is reflected in the fact that there are many permissible choices of reference class. And isn’t this just what one should have expected? Why think there is no room for rational disagreement regarding the indexical component of our belief-formation while there is very considerable room for disagreement between rational thinkers in regard to non-indexical belief-formation? The theory presented here puts domains, the indexical and the non-indexical, on the same footing. In both, there are constraints on what can reasonably be believed, but these constraints don’t single out a uniquely correct credence function. Anybody is of course free to try to find arguments for additional constraints. These would then simply be added observation theory we have proposed. The formula OE itself (along with its quantum sibling) is neutral with respect to choice of reference class.

In the two chapters remaining, we apply the observation theory to the cosmological fine-tuning problem and to the problem of how to model imperfect recall in game theory.