Search

In his paper “Explaining Deductive Inference” Prawitz states what he calls «a fundamental problem of logic and the philosophy of logic»: the problem of explaining «Why do certain inferences have the epistemic power to confer evidence on the conclusion when applied to premisses for which there is evidence already?». In this paper I suggest a way of articulating, and partly modifying, the intuitionistic answer to this problem in such a way as to both answer Prawitz’s problem and satisfy a requirement I argue to be crucial for any epistemic theory of the meaning of the logical constants: the requirement that evidence is epistemically transparent.

I will be concerned with the following question: are there compelling arguments for postulating a distinction between the truth of a statement and the recognition of its truth, when truth is conceived along the lines of a suitable generalization of the intuitionistic idea that it should be characterized as the existence of a proof? I will argue that the distinction is not necessary within the conceptual framework of intuitionism by replying to two arguments to the contrary, one based on the paradox of inference, the other on considerations concerning the content of a statement.

Suppose we want to conceive a thought not, a la Frege, as something that can be true or false, but as something about which we are capable of establishing what a justification to believe it is. Putting this characterization together with another one that can be found in Frege, according to which a thought is all that is necessary to know, or better to grasp, in order to understand a sentence, the idea emerges that understanding a sentence, grasping the thought it expresses, is tantamount to having a criterion for establishing what a justification to believe it is. From this we can extract the starting point of a theory of meaning: the meaning of a sentence is a classification criterion that enables us to establish whether something is or not a justification to believe that sentence. Of course, the notion of justification for a sentence A must be introduced as a theoretical notion, and my general strategy is to characterize it by induction on the logical complexity of A. The basis is a definition of the justifications for an atomic sentence of the form “P(n)”. My contention is that this notion should be defined in terms of two others: the notions of authorization to use n to refer to a given object, and of authorization to concatenate P with a name of an object. In this paper I wish to concentrate on the former one.

On etudie les relations entre verite constructive et verite classique d'un point de vue unifie