Expert Judgement Combination using

I enjoyed this paper, though disagreeing with its theory. For years I have tried to understand the Bayesian linear methodology; intrinsically some of the ideas seem so sensible. Ultimately, however, I have failed because for me judgements of uncertainty seem more primitive than judgements of prevision. The approach stemming from Savage (1972), DeGroot (1970) and others seems more natural to me than that stemming from De Finetti (1974; 1975), Lad (1996), Goldstein (1981) and others. Thus I am much more comfortable with expert judgement approaches based around probability than about judgements of moments. Moreover, as I have argued at various times (French 1985; French and Rios Insua 2000), I see the expert judgements as data rather than probabilities and thus tended to be against approaches such as the linear opinion pool which begin by treating their judgements as probabilities and then looking to axioms that draw them into a valid synthesis, where I use valid in a weak sense to capture whatever motivation lies behind a set of axioms.