ABSTRACT
Aerts and Sassoli de Bianchi offer a measure theoretic meta-indifference they call the universal average. Examining their solution, we find that the universal average is applied to distinct questions, each addressing a kind of random process rather than an instance of a random process for choosing chords. In doing this they do not fall into van Fraassen’s trap in the way that Marinoff does: they fall into it in a different way. For their universal average does not average over all the kinds of random processes, which is what is required to answer Bertrand’s question by a meta-indifferent instance of Well-posing, but is taken only over each kind separately. Nevertheless, their universal average is capable of being generalized and I show how, so giving the universal average as an instance of Well-posing. An extensive analysis shows a number of significant problems for the claim that the universal average is truly a kind of meta-indifference, including a triviality threat. I show that in the end, the claim that their universal average is genuinely meta-indifferent can be sustained only in some very etiolated sense and that even in that sense, the paradox still recurs.
