ON A POSSIBLE ANTINOMY IN THE BAYESIAN MODELING APPROACH

The well-known probabilistic Bayesian modelling approach has become a standard and most useful tool in the probabilistic safety assessment of nuclear power plants, in order to combine generic a priori probability density functions with plant-specific information into plant-specific a posteriori density functions. In this brief Paper, a possible antinomy (logical contradiction) in the Bayesian methodology is presented, leading to the conclusion that this approach is insufficient to deal in a consistent way with subjective propositions as they are conventionally employed in Bayesian analysis.