Some writers, for example Carnap in his “Logical Syntax of Language,” treat the whole problem [of defining logic] as being more a matter of linguistic choice than I can believe it to be. In the above-mentioned work, Carnap has two logical languages, one of which admits the multiplicative axiom and the axiom of infinity, while the other does not. I cannot myself regard such a matter as one to be decided by our arbitrary choice. It seems to me that these axioms either do, or do not, have the characteristic of formal truth which characterizes logic, and that in the former event every logic must include them, while in the latter every logic must exclude them. I confess, however, that I am unable to give any clear account of what is meant by saying that a proposition is “true in virtue of its form.”