According to Nicod's criterion of confirmation, a hypothesis of the form ‘All P are Q’ is confirmed by its positive instances, i.e. by entities that are P and Q. It is well known, however, that the criterion gives origin to counterintuitive results. For example, it sanctions the confirmation of ‘All nonblacks are nonravens’ by a nonblack nonraven. Since this hypothesis is logically equivalent to ‘All ravens are black,’ it follows that the criterion indirectly sanctions the counterintuitive confirmation of ‘All ravens (...) are black’ by a non black nonraven, e.g. by a white shoe. (shrink)

One of the important problems that have to be dealt with in the Philosophy of Science is to account for the high reliability of a very large part of our inductive inferences. This reliability is noteworthy because not only were many of these inferences made by ordinary people, even by children, but they were very often based on the observation of just a few positive instances. In the present paper, I deal with this problem by treating it as a normal (...) empirical problem. This approach enables us to make an important step towards an explanation of the phenomenon and, consequently, towards the formulation of a general theory of scientific inquiry. (shrink)