The measurement of subjective probability distribution has shown a marked revival of interest recently primarily due to the following two factors: the first is the advancement in the research of cognitive schemata that produce subjective probability judgements, typically represented by the works of Tversky and Kahneman, and the second is the growing concern in the practical use of experts' probabilistic judgements concerning the occurrences of uncertain events, like in meteorological forecasts, where no strictly formal methods of probabilistic inferences can be applied. As being one of the few who had deeply been involved in the initial stage of technological developments in the arts of measuring subjective probability distributions, the present author describes in this paper the three major classes of measuring methods, i. e., direct numerical statements, scoring rules and the two-person gomes, in addition to his comments on the nature of subjective probability with a special reference to the inplications of calibration curves.
