on this by first saying something about a quite different kind of statement, namely about analytic statements. I will then compare these to the "confirmations." In the case of analytic statements it is well known that the question of their validity constitutes no problem. They hold a priori; one cannot and should not try to look to experience for proof of their correctness for they say nothing whatever about objects of experience. For this reason only "formal truth" pertains to them, i.e., they are not "true" because they correctly express some fact. What makes them true is just their being correctly constructed, i.e. their standing in agreement with our arbitrarily established definitions. 

However, certain philosophical writers have thought themselves obliged to ask: Yes, but how do I know in an individual case whether a statement really stands in agreement with the definition, whether it is really analytic and therefore holds without question? Must I not carry in my head these definitions, the meaning of all the words that are used when I speak or hear or read the statement even if it endures only for a second? But can I be sure that my psychological capacities suffice for this? Is it not possible, for example, that at the end of the statement I should have forgotten or incorrectly remembered the beginning? Must I not thus agree that for psychological reasons I can never be sure of the validity of an analytic judgment also? 

To this there is the following answer: the possibility of a failure of the psychic mechanism must of course always be granted, but the consequences that follow from it are not correctly described in the sceptical questions just raised. 

It can be that owing to a weakness of memory, and a thousand other causes, we do not understand a statement, or understand it erroneously (i.e. differently from the way it was intended) but what does this signify? Well, so long as I have not understood a sentence it is not a statement at all for me, but a mere series of words, of sounds or written signs. In this case there is no problem, for only of a statement, not of an uncomprehended series of words, can one ask whether it is analytic or synthetic. But if I have misinterpreted a series of words, but nevertheless interpreted it as a statement, then I know of just this statement whether it is analytic or synthetic and therefore valid a priori or not. One may not suppose that I could comprehend a statement as such and still be in doubt concerning its analytic character. For if it is analytic I