Last modified: 2018-06-20
Abstract
The innumerable extensions and generalizations of this theorem have led to an unfinished succession of attempts at a solution to Bernoulli's problem, fervently discussed during the last 300 years. This succession began with the classic proposals of the nephew Niklaus, de Moivre, Laplace and Poisson, continued with Borel's in-measurement convergence, von Mises's frequentist axiomatics, deFinetti's subjectivist interpretation, his representation theorem and his connection to ergodic theory, the axiomatic approach of Kolmogorov and its vindication to frequentist approach because of an empirical deduction of these axioms and, finally, led to a methodological approach based on the theory of martingale.
The purpose of this research is not a detailed and exhaustive anthology of these proposals to identify a probability, but an interpretation aimed at seeking a rational review of the metaphysical assumptions on which the theoretical proposals were developed. The result of this unitary treatment attempt on the theory of probability, carried out more from a conceptual rather than mathematical point of view is the one presented in this paper. As it is demonstrated, the exclusively subjective nature of the premises on which the authors have based their proposals for a way of identifying the true value of a probability, it allows to conclude on the impossibility of such identification and the ratification of the Finettian principle that postulates that "probability does not exist".