In our communication presented at the International Congress of Philosophy in Prague, we proposed, along with Reichenbach, a connection between the probability calculus and a kind of multi-valued logic (I take the liberty of using the neologism introduced by Cavaillès). However, the attempt to bring the theory of probability into contact with this logic has aroused some objections. Although Reichenbach replied immediately to these objections, we desire to return to this question, since our point of view is not completely identical with that of Reichenbach. The first objection indicated that this combination was impossible a priori, since, in the logic of propositions, the values ofthe sum and of the product are univocal functions of their propositional arguments, while the probabilities of the sum and of the product are not univocal functions of their arguments. Different probability values can correspond to the same values of the argument, hence the former values cannot be considered as new logical values, in addition to true and false. The second objection can be summarized as follows: why should one construct a new logic if the theory of probability can be accommodated into the frame of ordinary (two-valued) logic?