as, in principle, corrigible, truth can consist only in a mutual agreement of statements. III This view, which has been expressly formulated and represented in this context, for example, by Neurath, is well known from the history of recent philosophy. In England it is usually called the "coherence theory of truth," and contrasted with the older "correspondence theory." It is to be observed that the expression "theory" is quite inappropriate. For observations on the nature of truth have a quite different character from scientific theories, which always consist of a system of hypotheses. The contrast between the two views is generally expressed as follows: according to the traditional one, the truth of a statement consists in its agreement with the facts, while according to the other, the coherence theory, it consists in its agreement with the system of other statements. I shall not in general pursue the question here whether the latter view can not also be interpreted in a way that draws attention to something quite correct (namely, to the fact that in a quite definite sense we cannot "go beyond language" as Wittgenstein puts it). I have here rather to show that, on the interpretation required in the present context, it is quite untenable. If the truth of a statement is to consist in its coherence or agreement with the other statements, one must be clear as to what one understands by "agreement," and which statements are meant by "other." The first point can be settled easily. Since it cannot be meant that the statement to be tested asserts the same thing as the others, it remains only that they must be compatible with it, that is, that no contradictions exist between them. Truth would consist simply in absence of contradiction. But on the question whether truth can be identified simply with the absence of contradiction, there ought to be no further discussion. It should long since have been generally acknowledged that only in the case of statements of a tautological nature are truth (if one will apply this term at all) and absence of contradiction to be equated, as for instance with the statements of pure geometry. But with such statements every connection with reality is purposely dissolved; they are only formulas within a determinate calculus; it makes no sense in the case of the statements of pure geometry to ask whether they agree with the facts of the world: they need only be compatible with the axioms arbitrarily laid down at the beginning (in addition, it is usually also required that they follow from them) in order to be called true or correct. We have