There is still another point, which would require a long and technical discussion once it was introduced, without substantially changing the outcome. Speaking of pure (not applied) mathematics, one might say:
6 "Arithmetical propositions are not analytic by themselves: they are analytic only in the context of an arithmetical system. If you accept Peano's postulates, you can generate all of arithmetic as a logical consequence of these postulates; but you must first accept the postulates.
Peano's postulates are as follows:
1. 0 is a number
2. The successor of any number is a number
3. No two numbers have the same successor.
4. 0 is not the successor of any number.
5. If P is a property such that (a) 0 has the property P, and (b) if a number n has P, then the successor of n has P, then every number has P
Using three undefined terms- number "0," and "successor"-he was able to generate an infinite series of numbers from these axioms. The axioms yield the entire system of integers. Are the axioms themselves analytic? If they are taken as definitions and statements of defining characteristics, they are; and since any propositions deduced from analytic propositions are also analytic,the propositions of arithmetic are as analytic as before.
For the sake of accuracy, however, we must remind ourselves that the postulates can be construed (and were intended to be construed) not as propositions but as propositional forms, like the p,q and r of our tautologies (pp 164-65). Peano left the terms "number," "0," and "successor" uninterpreted. We might then give the postulates an entirely non-arithmetical interpretation: for example, we could take "successor" to mean offspring, and "number" to mean chicken, and then by Axiom 2 we could derive the conclusion that the offspring of a chicken is a chicken-which is a synthetic statement about the world, and although it happens to be true, it is a contingent truth, not a necessary truth. The axioms become arithmetical only when the terms number. "0," and "successor" are interpreted in accordance with customary arithmetical usage (as is done for example in Bertrand Russell and Alfred Whitehead's Principia Mathematica and Gottlob Frege's Foundations of Arithmetic). The only point that concerns us here is that when this is done, the axioms become analytic, and in consequence all arithmetical propositions deducible from these axioms are also analytic.