Readers of this list might be interested in a viewpoint similar, in
its foundations, to Kyburg's. This broader viewpoint starts by drawing
attention to the common - and extremely obfuscatory - Bayesian mistake
of not having the event space distinguish between the value of an
external object and one's guess for that value.
I outlined this viewpoint in my article "Reconciling Bayesian and
Non-Bayesian Analysis", in Maximum Entropy and Bayesian Methods,
G. Heidbreder (Ed.), 1996, p. 79
David Wolpert