Abstract

The idea of ensembles which are both pre- and post-selected was introduced by Aharonov, Bergmann, and Lebowitz and developed by Aharonov and his school. To derive formulae for the probabilities of outcomes of a measurement performed on such an ensemble at a time intermediate between pre-selection and post-selection, the latter group introduces a two-vector formulation of quantum mechanics, one vector propagating in the forward direction in time and one in the backward direction. The formulae which they obtain by this radical generalization are vindicated by a rigorous derivation using Bayes’s theorem together with standard quantum mechanical predictions regarding ensembles that are only pre-selected. Their own two-vector derivation, however, suffers from a serious lacuna