The project of grounding mathematics in logic › The Price of the Hierarchy
PhilosophyAll chapters Reading tools Notes & comments
The project of grounding mathematics in logic › The Price of the HierarchySection view · Sections in this chapter7 / 9 ↓
The restrictions of the ramified theory can make mathematical statements difficult to recover in the form in which the authors want to use them. If a property was defined by quantifying over a wide range of functions, it may be of higher order than an inference requires. The issue arises even if the property has perfectly clear applications. The calculus needs a way to reach those same applications using a function admitted at the relevant order.
Chapter II, §VI introduces the axiom of reducibility. In the single-argument case, the axiom says that a given function is coextensive with a predicative function of the same argument. The latter occupies the appropriate lowest order. They agree as to which arguments satisfy them, however differently the original and the predicative function may be specified. The authors use this guarantee in their treatment of classes, among other places. It does specific work that the restrictions by themselves would leave undone.
Reducibility is a substantial commitment. Explaining why it is useful in a derivation does not decide whether it is a law of logic. The treatment of infinity makes a related problem plain. In Volume II, *120, the authors use infinity as an explicit hypothesis when it is needed. It ensures a supply of distinct elements for results that depend on it. They acknowledge that logic gives them no basis for settling its truth. This lets a proof make good on an important part of the plan even where it falls short of an unconditional reduction. The reader can see exactly where a special hypothesis is used.
This is one reason to resist judging the whole project by the success of a single derivation. The proof of a cardinal sum may need an analysis of disjoint union and cardinal identity. A theorem asserting the availability of arbitrarily many distinct terms brings a further issue. Asking what each proof assumes gives a more precise assessment than asking whether all of mathematics is really logic as an all-or-nothing question.
Comments
Add to the conversation
A question, a correction, or an example from your own work.
No comments yet. Be the first to share your thoughts!